You are on page 1of 65



y&i§w[i^gg¥¥ ®F m m
\f\ Reonomic C.r.otfigy Vol. 64. I%9, i>i.. 538-550


/ A Simplified Statistical Treatment of Geochemical Data * by Graphical Representation'

Abstract In the of a mineral exploration spon.sorecl by the United Nations Development Programme in two selected zones of Guatemala, a stream sediment reconnaissance was carried out, and graphical methods of interpretation were attempted in the seai'ch for a simplified statistical. treatment of about 25,000 geochemical results. The data were grouped by di-ainage and lithological units, and the frequency distributions of the abundance of Cu, Pb, Zn and Mo were studied in the form of cumulative freciuency curves. The four elements appear to be approximately loguormally distributed. Background, coefficients of deviation and threshold levels were graphically estimated. Examples are given of simple and comple.x populations. Mineral associations were studied by correlation diagrams. Contents

Introduction • Difliculty of the statistical approach in the case of stream sediment survey Adjustment to a lognormal distribution Definitions Construction of the cumulative frequency curve Comparison with histograms Information given by cumul-ative frequency curves Background .T.~".-. Deviation Threshold Elxamples Advantages of cumulative frequency curves . . . . The coefficients of deviation '. Correlation diagrams Conclusion References ...Introduction

538 539 539 539 539 542 543 544 544 544 544 545 546 546 548 550 550




T H E United Nations Mineral Exploration P r o gframnie in Guatemala relied heavily on geochemical prospecting. D u r i n g one year (1967) 60 percent of tlie total Project area was covered systematically by a geocheinical recoiina'i.ssaiice carried ont in the drainage systenis. Nine thousand stream sediment samples were collected over about 12,000. km(rounded figures). All the samples were analyzed for copper and zinc, ar.d the total number thinned out to appro.ximately 4.000 before being run for lc;id and molybdenum. Finally about 2.S.000 geochemicnl results were availahlc for compilation and iiiteq>rctatioii. A s tbcy acctiinulated, it became apparent that liigli-contrast anomalies wliicli are obvi* This article is publi.'shcd with the authorization of the United Natioii.":. The opinions c.Kprcsscd arc not necessarily endorsed by this Organization. " *

ous targets for follow-up operations would not I-encountered but rather more subtle features not «-. easy to pinpoint and interpret. T b e interpretation phase of the survey was ch-.iracterized by two essential features: tbe great aiiumir of data to be analj'zed and the lack of precision n; these data. Sampling and analytical inethods must sacriiWiprectsion for speed due to the nature of geoclu-niici! prospecting, and the first consequence of this '\:\c. is tbat an isolated result bas little meaning in .L'I'"chemistry. It must be part of a population :i. numerous and homogeneous as possible. Indeod if. all kinds of phenomena, individual inaccuracies shad!off progressively when observation is e.xtendid i" larger and larger populations. T h e first phase of geochemical interpretatioii i . < i" condense large masses of numerical data and <•.'•' tract from them the essential information. The lU"'; objective and reliable way to do it (and sonKimu'.tbe only one) is statistically. Large sets of minihers, cumbersome and ditticult to interpret, may Inreduced to a useful f-jrm by the use of descri|Hiv(statistics. This is best done by the graphic;il rein'''sentation of the frequency distribution of a i;iv''-ii set of d a t a ; then the average value, an exprcs.-;i«>i> of the degree of variation around the average, am. tbe limit above whicii the anomalies start an- "Umediately and precisely determined as well as ti>i' existence of one or several populations in tbe .-^u''veye<l area. This treatment of the data also simplifies tbe ci'uiparison of the geocbemical behavior of an cloiuii' in various geological surroundings or of sovcr'il elements in tbe same lithological unit. I am grateful to Mr. TTeiiry H. Meyer. Project Manager of the Guatemala and El Salvador Mini.-i'.il







.Surveys, and to Mr. .Stej^lien S. .Stcinbauser for technical criticism and much belpfui discussion. Difficulty of Statistical Approach in Stream Sediment Surveys A reliable statistical interpretation reepiires that a great quantity of data be treated and that these data be homogeneous.In drainage reconnais.sance surveys, tbe first condition is easily filled hut not the second. As a matter of fact, the importance of sampling technique is sometimes overlooked in this type of prospecting. But even if given the appropriate attention, too many types of rivers and too many lithological units are generally sampled to result in a homogeneous collection of samples. The way to limit the inconvenience of the heterogeneity of the samples (particularly pH, organic content and grain size) is to split the survey area into drainages and lithological units, when possible, and to make the statistical interpretation. for each of them separately. However, even if this is done, tbe same degree of precision cannot be achieved as in the case of a soil survey where good homogeneity is possible. Adjustment to a Lognormal Distribution Definitions When dealing with a large mass of geochemical data, the first step is to find what sort "of distribution pattern best fits the various sets of observations. And, thus far, the lognormal distributioir pattern appears to be the one most applicable to the results of most geochemical surveys (Ahrens, 1957). In geochemical prospecting, we study the content of trace elements in various natural materials, and to say that the vaiucs are loguormally distributed means tbat the logarithms of these values are distributed following a normal law (or Gauss' law) well known as the bell-shaped curve (Monjallon, 1963). Many natural or eccinomic phenomena can be expressed by a value varying between zero and infinity, represented by a skewed distribution curve. If, instead of the actual value of die variable itself, we plot its logarithm in abscissae, the frequency curve takes a symmetrical, bell-shaped form, typical of the normal distribution. This hapjiens when a phenomenon is subject to a proportional effect, that is to say when independent initial causes of variations of the studiixl value take efTect in a multiplicative way. It is the case, for instance, for the distribution of trace elements in rocks, for the area of the different countries of the world, for the income of individuals in a country, for the grain size in samples

nf as a re n5.

'I •e

of sedimentary rocks, and others (Coulomb, 19.i9; Cousins, 1956). ' In all these examples, the character studied follows the lognormal law, which is pi'obably more common than the normal one. It is interesting to note here tbat tbe lognormal law fits very well in the case of low-grade deposits like gold but for high-grade deposits, iron for instance, the experimental distributions are generally negatively skewed because of the limitation towards tbe high values. G. Matheron gives a thermodynamic interpretation of the jn-oportional effect in the case of ore deposits and relates it to the Mass Action Law (Matheron, 1962). To the extent in which geochemical anomalies are extrapolations of ore deposits this theory should apply to geochemical prospecting. Cottstriiclion of the Ciiiniilative F-reqitcucy Curve A lognormal distribution curve is defined by two parameters: one dependent on the mean value, and the other dependent on tbe character of value-distribution. This latter parameter is a measure of tbe range of distribution of values, that is whether the distribution covers a wide or narrow range of values. The two parameters can be determined graphically as will be explained on following pages. For practical purposes, we work on cumulative frequency curves, and Uieir construction shall be explained by means of a concrete example. The various steps of this construction are the following: (a) Selection of a precise set of data ("population") as large and homogeneous as possible. (b) Grouping of the values into an adequate number of classes. (c) Calculating the frequency of occurrence in each class and plotting it against the class limits; this gives a diagram called the "histogram." (d) Smoothing the histogram to get the frequency curve. (e) Plotting the cumulated frequencies as ordinates gives the cumulative frequency curve, which is the integi-al of the frequency curve. (f) By replacing the arithmetic ordinate scale with a probaI)ility scale the cumulative frequency curve is represented by one or more straight lines. Examples of kignormal frequency curves are shown in Figure 1. Some brief comments on the dilTerent steps follow: (a) The larger tbe population to be analyzed, the more precise and reliable tbe results. If necessary, as few as 50 values may be treated statistically but

voidd npt be itrres not so ey was charjreir amount pjedsion of UH sacrifice gfijchemical of diis fact
lirii' "^ geo-


. as

HiiLiced in

•aifes shade -•cttnded to tai33n is to a and e.xThe most srtnetinies
s "if iium-

!t, may be iercriptive icai; repref a; given gression raje, .md : ire im11 as the tke surthf conitiement ieveral Ftoject Mmeral

4 > - ^ T - - i , ' ; ' >-''••,--•,'-'--'r^, - - ' - - ^ i . : ' . - - : -



Figure I . Lognormal distribution curves

, FiHjuencyX


Frepucncy Y,

-. I
' !

».Value Afllhinetic scale LoeaiiUimic scale

• VeiiK

Cumulated tieiiuency % 100


Cumulated fieiguency J 99.99'




ii f2C

/ f / I I I I I
' — . — i — . — • • • . • — »Value

- * Value Logarilhmic scale



Logariltimic scale

the confidence limits must be calculated to see if the analj'sis is meaningful. (b) A correct grouping of the values is mandatory if some precision is to be achieved in the statistical interpretation; too few classes will result in shading out important fealiires of the curve; too many in losing significant details amidst a cloud of erraticones. The results are distributed in classes, the modulus of which should be proportional to the precision of the analyses: the more precise the analyses, the smaller the modulus. The logarithmic interval must be adapted to the variation amplitude of the values and to the precision of the analytical niethods (Miesli, 1967). ' ' In statistics, working witb 15 to 25 intervals (or classes) is recommended. As a rule, the width of a class, expressed logarithmically, must lie kept equal to or smaller than half of standard deviation (Shaw, 1964). For geochemical ]>urjx>scs, it is convenient to work with 10 fo 20 points on the cumulative frequency line, tbat is lo say with 9 to 19 intervals or classes. There are three variables to consider; the number
• \

of points {n) necessary to construct a correct Iim;; the range of distribution of the values {R), expressed as the ratio of the highest to the lowe.-^t value of the population; and the width of the classes expressed logarithmically (log. int.) which has to be selected in function of the two first paramctors.Tbese three variables are linked by the relation: log. int. =


In most of the cases R varies from 6 to 300 (exiicriinental average values), then, with (») varyin.i; from 10 to 20, log R from 0.78 to 2.48, the extreme vaiucs for the logarithmic interval will be; log. int. = - ^ = 0.039 . • 2.48 „^. log. int. = T Q - = 0.2J The 0.10 was selected as tbe best suited logarithmic interval for the classes because it suits most distri-



i k I • 1:1 ' i1 i i! • -1 looo 10 tk.ia>m-.95.tTr-iirtiofj.1 log. the histogram-frequency curve step is ski])pcd and the cumulative frequency directly constructed. The most useful and commonly employed in geochemical work is the 0. ppm.I fN.iak^MiJ^ A S I M P L I F I I W S T A T I S T I C A L T R E . 2.5 1 O. I'l I i 1 1 '^4.t «"»> 1 1 1 ! ! +^ .4.. 50 _.7 . (c-d) After selecting the class table.on \ . for instance. i N "M li! '!l \\% •K-I SO ' 16 10 5 2. classs table. in the case of copper. 0.4TA 541 butioii. .34. Usually diis results in an excess of 2. 5 ppm values.5 95 90 Bk ! . This is of importance for a correct construction of the frequency curve. 0.irrr-^-a^^ . _ :..05.47..^ J A kS -> 1 1i \ ^ • \ J ''' i1 i ' \ 1 1 ^ J .9 99.5 9T. 7_ b. giving reasonalile number of classes and a good definition of the curve. However. m. — 1 — 1 — 1 — n ._ ~T . 4 and 7 values.i_. Tbe only precaution is to avoid starting witb a round value so that no analytical results will fall on the limit of two classes.•> . ' l T M E N r O F G E O C H B. » [ 1 ~1 k ) 1 s 500 0.Ut. 1 ' " I T " . 3.S 0. < . a part of which is given below._.J T j ^ . 1 . 1. the lower part of the standard colorimetric scale reads 0. When the logarithmic interval is selected.. H T 1 . 0. •O. For instance. 2. and the raw values must often be corrected by extrapolating tbe general'shape of the curve.57 class limit (ppm) .1 0. t : 1 1 ^. giving a histogram which is smoothed to a frequency curve.V'. the normal procedure followed by many authors is to start cumulating the Cu a i .s 20 U I S . being strongly affected by slight changes in class intervals. and where to plot the cumulated frequencies? .:^::t<.'.48.).Jl. the values are grouped and the frequency calculated for each class (in percentage). •\ > \ ! ir-\ = = I 1 i % ^ _ %J r i-N f.17. it is easy to calculate a table giving the class limits in ppm. int. 0.72 It can be extended in both directions as far as necessary. U • 1" w. class limit (log) . As for tbe first point.ioi. 0. \ SJ> 1 !> 1 1 1 ^ 'i'i' 1 ''t r~ 0.5 iI I ^i 1 >s.27. 1 : i ! ! ii IM . experience shows that there is an inevitable concentration of the readings.1 r •• —• 1 i J '^ i i" 1 •" i • >M "^V \ . and frequency curves are difficult to draw and handle. j t 1 . and it is the one used in geochemical presentation and interpretation of the results. it may be necessary to use 0.*XSH:Ai£l^t:i^y:s*i•r.t'" 100 200 ppm . ] I.. one obtains the integral curve of the preceding.^-^i^. As a matter of fact.>^ 1 \ . ItJias the form of a straight line when using the appropriate graphpaper (probability-log). Tltitrm 2 Practically.37.86. M J. on tbe values actually represented in the colorimetric scale. I . « no " 1 r. whoever the analyst. -.di'&:-3&i fj. and if the disi)ersion is specially large. 0..2.« ' 1 \ 1 — .07.t.2 will be chosen. i! Cuj -_ -" • • 1 lili M.Cuaulatlv« FVtqucttcy Dlttrlbutlon for Zn end ZZ 99. • O. 1. .01 0.I r'nat' .M I C A L D. Then two questions bave to be answered: where to start accumulating the frequencies. note here an advantage of the histogram: it clearly illustrates the effect of tbe sensitivity of the analytical method and more precisely tbe bias brought to the low values by the use of colorimetric scales of standards.•ft-- i Mi. it is necessary to determine tbe inflexion points of die curve in order to evaluate the standard deviation. then the frequencies are plotted against the class limits (the latter being logarithmically calculated. 0.2 . 1 1 Hi.-»S:kgat-Xiyii~n. M. (e-f) By plotting the cumulated frequencies as ordinates instead of tbe frequencies. and a conspicuous lack of 1. In case of very reduced dis])ersion of the values around the mean. But histograms are often misleading.17. Dll lb"pm 1 l l i . ordinary arithmetic-arithmetic paper must be used). 1 A 1 i >•' J \ . 3. 1. m .

-. 1) (Hubaux.o-j '4 classes used and is easily calculated ( 6 % for iln. for practical purposes.feife^^^A. then. the curve heing an integral one. If the clas.1 ln^j. and this value is impossible to plot..Piiit. (iidliiaVUUn. P j l at O. int.).ii the dispersion parameler (coefficient of devialiun i.2 log. classes can be directly compared widiout correction. As for the second point. Confidence limits ( P j .^giiia^»A^ii^^ 542 CLAUDE LEPELTIER F i g u r e 3 . •:Qj. limit is used. . then. I consider it much better to ciiiiiiilatc the frcqiicitcics front the highest to llic loti'cst 7utliics.>^:i«>^. 1961.-^-w^'.a^t*f»h'-. Uernv. This error. or difference. 2 ) .i .1 log.. the values zero and 100% are rejected at the infinite: it does not matter for zero because zero% never occurs. the histogram-frequency curve step is skipped and the cumulative frequency diagram directly constructed.White. 3 ) .i •?i JOOO lOOQO logi? 1.. The width of the confidence channel is inversely proportional to the iinportance of the population considered: the bigger the population.faiitijtiaipntiquf j Ij ibEijIigDC. frequencies from the lowest values toward the highest (Fig. one should use the Pearson's test (Rodionov. .^.05 logarithmic class interval. Tlieii considering the lack of precision in the low values and the iinportance of the high ones for the determination of the threshold level. Using the class ci-nii-r . population: N ~ 989 range: R = 230 = 23 10 The best class interval is selected as explained above: i '.t'«r>»-K. interval will give 14 intervals.. Ikfore enumerating nnd coni- '""""^^ . Figures 2 and 4 jircsent the same data in two diflcrent ways.4 0. in Figure 4. 1960). curves constructed from difterent !i.36 „ ^ „ . the 100% will correspond to the lowest class and be eliminated.OS probability level- CuffluUlive Fitiiuency ino/o against the lower ciass limits..will oitail an error of exciess on the central teiuleiicv pannneters (background and threshold) but not <. "' In Figure 2. Comparison with Histograms For comparison jiurposes the cumulative freciuency curve for Cu in the Motagua drainage (Fig. hut this longer operation is generally not warranted in this type of iiUeriiretatiou and. int. the ordinates must be plotted at class limits and not at class center. This confidence interval has been drawn on Figure 2 by using a graph (Fig.». one has to consider a property of the probability scale used as ordinates. but in each case the last cumulated frequency is .^ — = . cnmukited frei|uencies are to be plolted A 0. ScKjru: A. 127& for the 0. thus.ilAi-^^nl-tittr. but this does not matter provided they stay in a channel delimited by die confidence limits usually taken at the 5% probability level. However. the points fit fairly well along a straight line. 2) was also constructed. = 0. To check that a distribution fits a lognormal pavtern.-^. log.-'• -. 1959). since one cumulates the fre<iueucics from the highest values to the lowest. • w .Actually. the poinls never fit the line exactly. and 267o for die 0.097 . varies with the tvjK. 'i. int. which is acceptable. There arc 989 results ranging from Hj to 230 ppm. the grajihical control described above is satisfactory.100%. . int.. . 14 «. Wl. lost for the curve. 1965: Vistelius. the narrower the confidence interval. suggesting a lognormal distribution oi zinc in the alluvial deposits. the corresponding histograms and frequency curves for Cn and Zn. Tennant and . Let us take a concrete example: the distribuiioii of Zn in the quaternary alluvial deposits of Block 1 (Fig. Usually.--*ii <iTfag»4^y:^v>. whicii avoids fastidious calculation and gives a fairly good precision for the cumulative frequency values between 5% and 95%. = .

-f' . 8 1 58.86 2. s") and threshold level ( 0 .:.-Xt>fi.5 2 3 rn 4 5 C 7 8 S 10 Mill -1 IS 20 30 40 50 60 70.-. 4 .H. coefficients of deviation {s.-.3 117l 1 « Iti 7 i 4 ??'^ menting on the advantages of the former presentation over the latter.^-^^.--:. W ^ : . This effect is illustrated for copper in Figure 4. . an interesting feature of the histogram sbould be mentioned.. Cumulative frequency curves are of easier construction and more precise than ordinary frequency curves.17 l . s'.^^^^ A SIMPLIFIED STATISTICAL TREATMENT OF GEOCHEMICAL DATA ) 543 Figure 4. wbere the classes including a colorimetric standard are shaded and the value of the standard itself is given as a larger figure ( 1 ..eo 109 ISO TX 3^ il li (L93 1. one sees immediately that it is easier to compare two straight lines than two overlapping bell-shaped curves.-. in the case of colorimetric determinations made in the lower range of sensitivity of the analytical method. Information Given by Cumulative Frequency Curves The main purpose in constructing the cumulative frequency curve for a given population is to check if it fits a lognormal distribution.Xj-.72 4.2 « .1 93.-. particularly if it is started from the high values. it is supposed to be die upper limit of the fluctuations of (/') : it depends on (6) and ( s ) .-^..f4i.-:.. The values equal to or higher than (t) are considered anomalous.6 23.7 J O 18. but it may be necessary to bring some corrections to the low value frequencies in order toconstruct a precise distribution curve..4 29.5 37. wc found that trace element contents in stream sediments appear also to be loguormally distributed.. to estimate graphically its basic parameters: background (l>).I.95 3. (b) gives an idea of the average concentration level of the elements in a given surrounding. and if it does.i£^ii. « 1.>.W^ T.^Jifcfflifgaff.3j 11."^i?M*!!!ti'P?.68 5. 2.. it is simiiter to draw a line that fits a set of jKiiiits than to draw a bell-shaped curve with iiillexioii points. many more populations can be presented on the same diagram by using cumulative frequency curves than by using histograms. ppm) . Histogram and frequency curve for Zn and Cu I TH 1..9 74.iy:<^:-. (t) is a complex notion which might be termed "conditional": statistically it depends on the probability level chosen.69 7. (s) expresses the scatter of the values around ( b ) : it corresponds to the sjiread of the values and their range..---.i>. .:.. from the lowest to the highest.V^v-. Adjustment to tbe lognormal law is generally tbe case when soil samples are considered: in tbe drainage reconnaissance survey in Guatemala. geologically.41 9.. the histogram shows clearly the bias introduced in the readings by the human factor and by tbe accuracy and sensitivity limits of the method. and for practical purposes. Comparing Figures 2 and 4.-. the cumulation of the frequency reduces this effect.J-•.U 2.

a function of the two former. Because all the reasoning is made on logarithms. b curve) must be recalled here: (6) being tbe median value and (s) the standard 0.Ition (from b — s to h -^.is conventionally taken as the threshold level (/) ground. 68.2 X 5. It is exjircssed as a percentage: the deviation no dimension.ii':-. H . it h. it is also necessary to exjiress the deviation by n logarithm.s or that 32% is outside log / = (log b) -j..28 Then multiplying or dividing tbe background v.i:ir-iar:it^:tiisii'r--:--sf^^^ 544 Background CLAUDE LEPELTIER of tbe deviation: it is inversely proportional lo ih. to consider a third deviation inde.\iK'.2 ppm b (Zn) = 48 ppm X i9-5 — )9-5 — 9f — / .-i t . roiinding off the above.j : . one Practically.2 = 47.28 s = log s' = 0. This geometric mean is a morc significant value that the arithmetic mean.<). the third important parameter is the threshold and b -V 2s 99. It is also a more stable statistic. 2). 9. In the illustrated example.f') . "In the examples given in Figure 2. s' = 2. 10 and 50 ppm are taken as reasonably good approximations of the background levels. ii:!i. we have: Ixickground value for copper background value for zinc b (Cu) = 9. In Figure 2. these values must be rounded off.x: tbe relalin: Deviatio-n deviation (j") sometimes called coefficient of viiiiBefore explaining how to determine graphically . which is.5. slope of the line.5% ordinate.. It has been seen that in tbe case of symmetrical disand b -V i s tribution (either normal or lognormal) 95% of the This holds true in the case of the lognormal dis.36 3. the vaiucs /' -h s and b — . 50 — I ZOIC.74% of tbe population falls between b — Zs level ( t ) .^ (.xis. is read directly bas to consider tbe ratios and not the absolute values on the grajib as the abscissa of the intersection of thus established.-iluc by the geometric deviation will give the upjicr and lower limits of a range including 68% of tbe IXJJJUI. 167o of or to avoid using logarithms: the values will fall above b -'r s and 16% below.exceeds the uj)j>er limit b -\.v will be I = h X i'obtained by projecting the intersection of the dist = 9.s. In (Fig. the background thus calculated .2^ these limits.and b — 2s.5% of the j)oi)ulation nonnally distributed.e. or A'A on the figui-e).26% of the population falls between b — s Threshold and b -i.corresponds to the mode (most frequent) and median (50% of the values above. we can say that 68% of the population falls above which the values are considered as anomalies: between b — s and /. fitting the "bell-shaped". one determines the points P (at the 16% this example one reads 47 jijim. i .a. tribution since tbe logarithms of tbe values are that is to say that only 2.ici ''•p^ifS^s'!''r'g^M^!'-^^?»y«-»^'^**«^g^^^^ . the background value (b) is given by the intersection of the line with the 50% ordinate.s. The distribution curve being symetrical around an axis of abscissa ( i ) (Fig. In this simple case.>. it is a factor obiaiu.i>i~i- ^*> the slight difordinate) and //. . of the values (b — 2sio b -{• 2s).36 Of course.i.-af. 50% below it) values.-r.. This upjier limii mentioned percentages and taking (/)) as the back. OA is the geometrical expression ference is due to the rounding olT of the ex. l[ o.individual values fall between h -f 2^. -t.2s. -r\iii. We call it the gcoitictric tlci-inii. 4 ) . Working with logarithms. 2.. Then. (/) as well as (b).s After tbe background and the coefficient of devi95. the coefficient of deviation (s) is the logarithm (base 10) of the geometric deviation {. Multiplying or dividing by the square of the geometric deviation gives a range including about 95/.il l)y dividing the value read in // liy the value n-ad ni 0 : ' A straight line denotes a single population lognormally distributed. and is ihe geometric mean of the results.}^:.ll 12 a9 \ ul It will be seen later tbat it might be interestin. it will be illusory to imply a precision far out of reach of the analytical methods.44% of the population falls between b — 2s ation. less subject to change with the addition of new data and less affected by high values.2.ation.8 ppm tribution line with the ordinates 16 and 84% on the abscissa a. In the case of a perfect frequency distribution curve. 2. Taking tbe same example of Cu die distribution line with the 2. an essential propert}' of the s" = 100 normal distribution (i.9% 100 deviation then.b — s.

liackgrouiid and coefficients of deviation a.. This is the complex pojnilation or of different ones. in the lognormal distribution). decreasing frequency order: In this case. the material sainpled. "in the directioii of the high logical unit may include two types of mineralization values (jMsitive skewness). Ji---\. a superimjiosed Oc is lifted to Oy which means that instead of having mineralization related to ore. 1 1 \ \ \ 1 1 < ! — . the. 0 .. A SIMPLIFIED STATISTICAL TRE.»ia^^ r^L'.. Tbe abscissa of the breaking jioint.tillii^^: ^J^>2iic^^gS^J^ii--. the threshold These three cases are represented grajjhically in is five times the background for Cu and only 2.d a t a . different thresholds if their coefficients lation. there are 17% of thciil.5 20 100 2O0 500 ' i ' .irTii'-^''"-' >:*^jr-*A .v should sentative of the normal or background content of extend as a straight line in 0 ..^ ^ \ 1 • ! • 1 si •'•3 ' 1 1 K 0.5% of tbe values 30 ppm or greater. for instance. a.r. If the jjopulation was showing up in soil or sediment samples. a^T^^.4&.j^ k.skewed to the right. a mixture of two populations in a given set of" ordinate 2.9. an excess of low values in the considered popunevertheless.°fc.^.xcess of high values in the sets of data selected so as not to include more than jjopulation.whereas. \ . / — \^ a\ ^ ? 1 \ \\ o\> V i \ \ V N^ • s.. E. 2.-. This shows the importance of the deviation in the estimation of the threshold. ! ' • ^rX \ \ \ \..5%. The abscissa of the bre.. and the other.. broken line is frequently obtained sugcumulative frequency line (Fig. an excess of high values in the considered are calcubited witb the main branch O. in this case. 1\ > • --I—i--! '• i i ! 0.J. the dia.i=i^t. In Figure 2.t.ikiiig jioint may be conveniently jjopulation. 1000 PP" ?5i b. an excess of high vaiucs. gram. encounetred in the Guatemalan drainage survey and In all the foregoing.01 0.^. the histograin will give a frequency curve two different distributions. when constructing cumulative frequency how to interpret such lines. Some indications are given below showing How-ever. the interpretatioii is made on case when there is an e.I5 R( S'T \ i 1 ! 1 ' 1 i i \ \ ) ' 1 l ! ! • \ \ ^ i l "' ^ 70 10 j i 1 . ^j|gwf^!aMSaiiiiaiii{?^j»i^^!^?<^ .__ 1 ! • i 111 1 1 ' .e. and two populations may have the same background but.which there is a dejiarture from the norm (i.ajijjear as solid lines witb slojie breaks on the diagranimadc expression of which is-a straight line..S ^ V- — — rCu • Mo -— % f-i'l^ •-. \ 1 ^.4TMENT OF GEOCHEMICAL DATA 545 99>! 99. \ \ 1 1 .5 1 — .biiant£fc.7 Figures 5.yi.-:Ti^^j^-Li:^-^ff^j. Whenever possible in practice. \J _3 X \ \ \. from tribution tbat are the most likely to occur are.lognormally distributed. one repre. Copper Dislribtitiun (in a lithological unit).1 ! : : ' i f\ *. of deviation are different..1 pr" 0. 2. \ ^ \ p-.Vf ji >. c.2 \.1 \ 10 'i 1 0... '^Al' N V. 5) shows a break gesting that the set of data considered consists of a to a flatter slojie at the 30% level.- ''•'^-^-.. \ I 'A \ ^ N ilL 7n .vainl>lcs (in this case 18 jipni) indicates the limit above The three main cases of noii-homogeneous dis."-^~L \ 50 \ .>iM-.28% to 2.main branch Q. The curves. They correspond to real distributions times for Zn. a single lognormal population.-. I have considered the simplest case.^.:»':K.

and (3) analytical errors.Furthermore. Distribution heterogeneities will be spotted and the corresjionding units selected for further investigations. as a logarithmic coefficient s. :£: r a: c .B. Zinc Distribution (in a drainage unit). There is also a relationship between die background (/>) -and the coefficient of devialions (s) whicI iiatc jiroj^ very tive I'ollo' l. By splitting the data at a value taken around the middle oi A -^ B (at 4 ppm for instance).and b on the diagram. toward these low values (negative skewness). in this case. This is conveniently done by plotting the corresponding distributions on the same diagram—for instance Cu distribution in three or four different drainages in the case of a stream sediment reconnaissance.5%. to poor sampling (for instance. which is done on the main branch of the distribution line in the uaual way.Mfail'.-VWm. however. suggesting the existence of two distinct pojiulations in the set of data considered. Positively broken distribution lines are the more interesting because they indicate an excess over the background mineralization.J. . The negatively broken line on Figure 5 is the expression of an excess of low values in an essentially lognormal distribution. Confirming this assumiition would require very extensive geological-statistical studies encompassing all metaliogenic cases. however. first a Jiositive.should be taken as usual (abscissa of point P ) . though the threshold of population a may also be considered. 2.-n-ftir-i^-r. The term "deviation" is preferred to "dispersion" which might be more expressive. and we lacked computing facilities to calculate theoretical distributions. The coefficients of deviation must be calculated separately for distributions a and b. many people use the term "disjiersion" in statistical interjiretation of geocheniical data. then a negative one. 5).can be readily compared.5% level (at jjoiiit p for instance) die threshold .-r-'^'^-ffi^''- jJiJJ^ 546 CLAUDE LEPELTIER 'tia !as \%.«JA»UJUmJau«e-t". On the. All tbe characteristic jiarameters of the distribution can be estimated without cumbersome calculations. it is possible to separate the total population into two elementary ones appearing as a. the geochemical behavior of trace elements in a given geological environment from different countries or metaliogenic provinces. more often. presenting Figure 5 as histograms and frecjucncy curves will result in an overloaded and illegible diagram). branch A corresponds to the main or normal jjopulation. On a broader scale. the break occurs lielow 2. This excess of low vaiucs may be due to the inclusion ill the population of a low-background lithological unit or. the adjustment to a lognormal distribution can be checked graphically. may indicate similar geochemical processes in their formation. the histogram is skewed to the left. . diagram (Fig. collecting an imjiortant set of sediment samples that are too coarse).>-^^ l» i-'-'i ^^'l-Mi'trr^ . they do not interfere in the interpretation. Comjiarison betweeii various pojiulatioiis are easy and comjilex distributions are clearly identified. If. We shall consider here only the most frequent case of a main "background" jjopulation mixed witb a smaller one of higher average value. The general background will be taken with branch A and the threshold as the abscissa of tbe middle of branch A -\. more commonly. When rocks are ' considered. branch B to the anomalous population and the central branch A -{. Molybdenum Distribution (in a lithological unit). but we have not enough examples of such complex distributions to make definite recommendations. a similarity in the coefficient of deviation.B to z mixture of the two. The cumulative distribution line shows two breaks. i wl taken as threshold value if die break occurs above the normal threshold level of. because there is no genetic imjilication in the concept -of statistical dispersion whereas there is one in the notion of geochemical disjiersion. Advantages of Cumulative Frequency Curves Plotting the distribution of an element in a selected unit as ciinuilative frequency curve on probaliility grapli jiajier is tbe easiest and most jirccise way to present a great amount of data (for instance. It has been seen that the absolute deviatiott can be expressed as a geometric factor s' or.<HllB»J. This is an approach to a better understanding of the distribution laws of trace elements in naturally occurring materials. When the results do not fit a lognormal distribution. (2) complex geology (imjirecision in the lidiological boundaries). Provided their proportion is not too high (20% or less or instance). The Coefficients of Deviation A lognormal distribution is completely determined by two Jiarameters: the geometric mean (6) and the coefficient of deviation (s). It gives a double-jjeaked histogram. (1) lack of homogeneity in sampling. Such a graph is the expression of a dual distrihution. It should also be kept iu mind that some elements in some surroundings may nol be lognormally distributctl.)laiiation may generally be found among these diree factors.». au exi. The coefficient of deviation is a dispersion index specific for the distribution of a given eleinent in a given environment and expresses the degree of homogeneity of diis distribution. the two of them being lognormally distributed. It is possible that a given value of s corresjionds to each tyjie of mineralization in a lithological unit. Comjiaring the geocheinical features of the various units of a survey area is important in assessing their mineral potential. together with similar average»~»"a" .

35 0.a^WtHfltjAn. the extreme values for b and s". This is expressed very clearly liy the relative dispersion / ' (or relative deviation).-ITMENT OF GEOCHEMICAL DATA which is the exjiression of the geochemical law which states that the dispersion of an element is invcr. h Zn Cll J s" = 100 { 0 s" Block 11 Zn Cu Pb Mo b J s" 0. \ \ Cu\ 10 1 1 1 d ^ • S T " N^ . \ L ^s~ \v _lN.rillni rx.^ 1.p.H 4. •*-K X s V 1 'J >.. \ • .^Sr-?i. Pb and 1 o. Tiie weighted mean values of b. for Cu. thus determining jiarallelograms including all the individual values. 1 . for each element.38 0. s and / ' for each element have been calculated separately for Blocks I and I I : Block 1 .2. the lower the relative deviation. 0....S All liUiolotic'l uii'l^ —E '~" - IDO 9 B 1 t.^XW'i*r'''iiJ*l^^ ..24 3. 6.\fv"^v.^fex'..17 8.42 0..< at oit 1 2 1 « i < ) 1 i 10 V Cu PbMo ^ I 3 'y't 5" t f • 1 ICD • •ail o. ::---j^.8 0.<^ely proportional to its abundance.^'^^B.a.'^v. Figure 6. Figure 6.8 0.. The diagram has been constructed by taking.1 0. The higher the background. 8.3 o.iffltW'ii«H->-^-"--^'--^^^ A SIMPLIFIED STATISTICAL TRE.2 125.7 0.8 5.40 The fact that the absolute deviation for Pb is equal to or slightly lower than tbat for copper is due to two factors: (1) the sensitivity limit of the analytical method for lead. which entailed a number of assumptions and extrapolations in the interpretauon—determination of b and s.BxV o. One sees immediately tbat there is an inverse linear relationship between b and s" (which is evident from tbe definition of s") and that the Pb Mo 55. This is best shown on a log/log correlation diagram by plotting s" as abscissa and b as ordinate.37 97. Zn.5 70. Carrel(itlon diagram b/s" for bloclcs I and 11 (ppm) )«0 _. for instance.T. 6) also decreases when the abundance of the element increases.32 4.» 4 •'' r. and (2) the existence of some Pb mineralized zones in the survey area where b was high and s low. 0. show-s the varialion of s" in function of b in the different lithological units of Blocks I and II.-. Xs 1 1 Z« ai u 0.2 0.^ PN — N ^"^ \ \ .i^»Vflyl'l!ff^il^. 547 average absolute deviation s (graphically estimated in Fig..30 0.30 5. a jierceutage related to b aiid s as follows.

. and consequently might give an indication of the type of anomaly encountered. it is probably related to the type of geochemical disjiersion.^Bgs^HB!^n«}^^ .1 0. mechanical or chemical.1 . 1962). ^J. Much remains to be done in this field.-iiJ^-.'j.::. 5 > 10 2ii(re«J Dj •= 98 " 1 "2 "3 Id ii» "1 ies 193 2 N. p measures the degree of dependency of two lognormal variables namely the tenors of two elements in a sample jiojiulatioii (Matheron.fiii. n .-4ji'JArwi»it---'tr'i. ??^g^^^gipgfM^p«^?. .ites (Fig. ..--i.l.-.® ai b — . ' •c • • k:' 1 • u : : • I r' -1 0 9 -- ^ _ V i£ ^ •Hr . it is also interesting to note the variations of the disjiersion of the same element in different lithological units whicii is particularly noticeable for copper.^::. .>-.• ' c ' ^<1« 5 < ] } ' ** "^ '-—. Each ... ..1 .1^ir\-f. syngenetic or epigenetic. .i'. In the ca. : . direct or inverse. .it:^'jr--f '!si^±<. the coefficient of correlation "p. The coefficioit p always falls between —1 a)id -f 1. coor(lin.r / i • "K 1 V ~W 1 y _ y _3.11 \ = V*"4" In Figure 6. . — —— .-1 • ^ s. I. .a jliai^Cti^iiiuieiUL. It ajijiears that a higher coefficient of deviation indicates a jirejioiideranlly inecbanical disjiersion.sanijile of tbe iiojiiilalion under study is Jilottcd following its two coordinates: its tenor in eleinent .'/ and its tenor in element B and the total jiopnlation ajijiears as a cloud of jioints. •.a precise factor. 8). -: ) « fl i ( 1 I J 10 4 1 t 7 1 9 li L t I I s i i .j>. /> — o means a comjiletc iudejicndence between the two elements. Corr«lallon diogram Cu/Zn (Id (ppm) vo V -t • / ) 1 1 <?l s J -/ •^f\ \ -. constructing a correUttion cloud in full log. The coefficient of deviation is a yery important character of the distribution of an element in a given surrounding.Cj.^ . which gives a rigorous measure of their degree of dcjiendeiicy..) Oi OS a* 0.i..^^aWiHi-tt-^r-. for instance between lead and zinc. but this has not been proved.. p = ± 1 indicates a functional relationship. Correlation Diagrams In the case of a jiolynu'tallic mineralization.—There is a graphical way to estimate p... :^.se of geochemical prospecting. belween them (it is a linear relationshij) between tbe logarithms of tbe tenors).^ .frf.--. Simplified Calculation of p.i. . slightly less jirecise but much faster than the comjilete statistical calculation. . This geologic concept of a relationship between two types of mineralization (only qualitative and rather vague) may be substituted In..Zi^-'-i^:^-. with two or morc elements lognormally distributed.5. a samjile high in Ph is commonly also high in Zn. .. t HjJ 0.---A«t-. ^ . there is generally a jiositive correlation belween them._ K_ 1 0. 1 .^^ 548 CLAUDE LEPELTIER Figure 7. 7. -the width of each parallelogram indicates the range of variation of s for each element.

. are then drawn... p is never equal to ± 1 (which would be the case if all the points were on a straight line) and the jioints form an elliptical cloud..'ITA .... Then p is given by the formula: The correlation cloud is in fact a two dimensional histogram. Tbe points falling in each in the same direction. correlation coefficients ' ^''''\.::. A^ = number of points in first and third quadrants N2 — number of points in second and fourth quadrants.95) (Matheron..)..--.3 the same sign as p (if p > 0 the two elements vary jipni. In Figure to the coordinates.•.:.M.. Two cases may happen.^. •. The axes jiassing by die and the two variables arc indcjiendent. £)z„ = 75 pjim). in the second... In geocheinical prospecting.:-. (2) or p is clearly different from zero and the gravity center {b^.r. if p < 0 the two elements (juadrant are summed up and counted as follows..m^'m^'l^s.». ) I S i M IIOIKO I n (ppm) :: r. the axes will jiass through the jioint (6cii = 5. they split into 2 or several attraction ceniers and form several ellijitical clouds more or less overlajiping. the jioints tend lo group in a single elliptical cloud.•a•4^i^4^:J•^?^'AkS'. then it is likely that a geologically based chemical equilibrium exists between the two elements considered. b.'!^t'^^Ji>''f>J^^>^ . Correlation diagram Pb/Zn Pb (ppm) tl..549 (1) eitlicr p is equal or near to zero: the ellijitical Practically.•.!i-V-tt^^t^^^^^ -iiiKM-'r''^'--^-''-'- A SIMPLIFIED ST.--•.--'. 1962) .:}'-"M-" !S^?««-: ...rri. that is to say b y d i e jioinl whose coordinates are the background values for the cloud is an cllijise whose axes are inclined rekilivc two considered elements. it is the best and simjilest way to establish whether a jiojnilation is homogeneous or heterogeneous: in the first case. I!.^-l^-~^...Y'N7VN^\ Practically.. Matheron jioints out that the relation expressed by p is an expression of the Mass Action Law if p = ± 1 (or of tbe order of ±0..>ml^^^. Figure 6. vary inversely). The slojie of the main axis has 7. this jirescntation of the data i s ' very convenient because it gives a geometric image of cloud has its axes parallel to the coordinate axes the distribution laws.STICAL TREATMENT OF GEOCHEMICAL D. G..

B. ?i> 7> ul Its low absolute value indicates a nearly complete independence of the two mineralizations. y-<s-L. 1956. an ellijitical one. 54. A.xes. Traite de geostatistique appliquiie. marked Cu. L. and I think it might be useful lo any geologist involved in geocbemical prospecting.the statistical analysis of the data.. inthe Suchiate drainage. one element may be used as a pathfinder for the other. Geological Survey Professional Paper 574-B. H. In this drainage.g%ipi^^^ . but much morc information can certainly be extracted from the analytical results by a more thorough. 14 du Bureau de Recherches Geologiquos et Minieres. Pb and Zn are ' associated: the correlation cloud is an elongated ellipse whose main axis bas a 45° slope and die correlation coefficient p = -^0. "Vistelius... B. 1959.vM'& 5...example of direct relationship between two types of mineralization..^-. 1967.. including the majority of the points which are spread more or less equally among the four quadrants.. Belgique. 1961.i'^^t^^:^«^y^^ 'i^L-<iils!-m'. in Figure 7. T. The lognormal distribution of llie elements—a fundamental law of geochemistry: Geochim. In similar geological conditions. Study of the distribution of some geocbemical data: Ecox.t/wrfi4»i^'>^ri>' . South Africa Trans. Cciilrc d'Etudes Nuclcaircs de Saclay. and a third one. 1969 REFERENCES Ahrens. G. 7) and the relationship of Pb/Zii in the Kio Grande drainage (Fig. ^"-if^?ti^^»sji^'t^^^j?^:. is intended only to illustnite tbe lack of relationship between two types of mineralization. A. Jamiary 20.<. The first example. On die contrary. • The graphical methods described above have the great advantage of being quick. Conclusion In the Guatemalan geochemical reconnaissance. 1964. p. but it can be divided into three zones: one around the intersection jioint of die a...--V.. although elementary. The correlation diagram shows whether two elements are .. March 2S. 1959. C. there is no relationship whatsoever between the Cu and Zn mineralization. Interpretation gcochimique des elemenis en trace dans les roches cristallines: Masson et Cie. M. that the Cu anomaly is more important than dial for Zn and that tbe two anomalies are well separated spatially... All this is expressed by the coefficient of correlation: p=-0. The cloud of jioints has no definite • shape.MALA.s d'oligo-elements: Ann.50 CLAUDE LEPELTIER may be used to a. V. Introduction a la niethode statistiquc: Vuibcrt.U.x. UNITED NATIONS MI.XERAL SURVEV. lead and zinc anomalies will have the same pattern and will be spatially related.. M.ff^tei^..^^A•. L. Geol..\. and White.. Hubau.^^aj^ifctti^t. 1281-1290. GUATEMALA CITV. Rodionov. 11. Contribution a la Geochimie dc ruranium dans les granites intriisifs: Rapport C.^:.-. LXXXIV — Mars 1961. Coulomb.fc.87. Paris. including only a few high-Zn/background-Cu points. no. 1962.itical knowledge. V. 1960.:>. The skew frequency distributions and {nndaniental law of the geochemical processes: Journal of Geol. GUATE.. Soc.. 1957.-:iiiie>^. .^-. v.^4. Jan.<-. Paris. v. In the Rio Grande drainage. LIX. computeroriented. 4.--y*^^^lt'^A-^trf^. Tbe value distribution of economic minerals with special reference to the Witwatersrand Gold Reefs: Geol. It is a convenient and synthetic way to present a great amount of geochemical data.scss mineral associations of elements in natural samjilcs.E.. Methods of computation for estimating geochemical abundance—U. treatment.^':k.jiiw. GEOL. Tennant. Monjallon. T. Representation graphiquc des distrjbutioii. Distribution functions of the elcmcnls and mineral contents of igneous rocks: Consultant Bureau. D.-j'. Miesh. cheap and easy to use in the field without any sjiecial niathem. 8 ) . This shows tbat. in the range high-Cu/backgroundZn >^lues. .surrounding.'.sjiatially associated and if one may be used as a jiathfindcr for the other. C. lonie 1: Memoire no. A. Paris.^.. 1173. . Matheron. A.-. Figure 8 shows an. 1960. 1963. ct Cosmochim..A. S. New York. Shaw.. A.. Soc..11 •iS was useful in outlining subdued anomalous jxitteriis in a coinjilex geochemical . D. R. 1965. France. Acta. with a tendency to inverse relationship (negative value). Cousins.«s-^-^iV^:i' . Let us consider two examjiles: the relationship of Cu/Zn in the drainage of the Suchiatc River (Fig..s.

minerals and reactions Appendix B: Glossary of symbols used in WATSPEC Appendix C: Program listing and input arid output examples ™»™Tfiw IT 'i-i!.L.T. Knapp.Norwich. Wigley. ISBN 0 9.I. Key to species. and others. J. ISBN 0 90224 6 8 9 5 An instrument system for measuring soil movement.llBRll" OR-C V u c / T O T Ji . 1974. ISBN 0 902246.. ISBN 0 9 . Leopold i Thoraas Dunne.. England US' TABLE OF CONTENTS page 3 4 •5 7 7 8 C3] Ion Pairs 9 [ 4 ) Calculating the Concentrations of Individual Species 10 (5) Extended Debye-HUckel Equation 11 ( . 1970.A. 1973."T. Andrews. Finlayson. 2 Drainage basin morphometry. Kirk.1970.SUBJ .Group ucuinui pnoxogicaa Kesear. University of East Anglia. Hanna. NR4 7TJ. . ISBN'O 902Z45 11 9 Techniques for the tracing o£ subterranean drainage David P. Peter A.and Activity Cpefficients (2) Law of Mass Action [5] Alkalinity and Carbonate Systems (7) Mineral Saturation Indices 1 . with Denys Brunsden and John B. ISBN 0 902246 46 1 Instruments for measuring soil creep. 1 ) Activities .M. E-J^' Isaac. Hosier. 1969. 0 2 2 4 6 90 9 WATSPEC: A computer program for determining the equilibrium speciation of aqueous solutions. Finlayson. G.ysas. ISBN 0 902246 02 X Techniques of till fabric anal. England. Ada Phillips.K.chemical and Fluororaetric Tracer techniques for streamflow measurements. 197S. ISBN 0 902246 07 0 A systera :for the field raeasurement of soil water movement.246 12 7 The deteTmination of the infiltration capacity of field soils using the cylinder infiltrometer. 1970. 1971. Ariderson & B. ISBN 0 902246 91 7 lished for the British Georaorphological Research Group by Geo Abstracts Ltd. 197 7. 1971. Osmaston.HilLs.e in the field. /1977.L. ISBN 0 902 246 13 S The use of the Woodhead sea bed drifter. A. Gardiner. 1975. Rodney G .Shorter Technical Methods-(IT).02245 40 . 1977. Fetch.THODS HOW T i D USE WATSPEC REFERENCES Appendix A. INTRODUGTION WHY THE COMPUTER? UNITS THERMODYNAMIC BACKGROUND •. McDonald S' J. ISBN 0 9. Tr. ISBN 0 902245 04 6 Field raethod for hillslope des. 1969. James. L. ISBU 0 902. 1975. frudgill S B. High &. 01 .02246 4 7 ' X Shorter Technical Me.F.cii I f l V i A ^ ^ mefiift^s of water hardness determination ^^^„*eouglas. 1973.i 1 e survey . Robert M. B. 197 5.M. 1 A raethod fdr the direct measui-einent o£ erosion on rock surfaces. 197 5.T. Eawards. tuna B. ISBN 0 902245 37 2 The measuremerit of spil moisture. ISBN 0 902246 34 e Slope pr6f.l. S. 1970. Brian J. V.cription.G. 9 ) Reduction-Oxid a tion Reactions ITERJMIVE ME. ISBN 0 902246 42 9 The use of electrode instrumentation for water analysis. F-inlayson & H. ISBN 0' 902:2^16 33 X An instrument systera for shore process studies. Drew 4 David r7 Smith.M.'COMP WCP WATSPEC: A COMPUTER PROGRAM FOR DETERMINING THE EQUfLIBRrUM SPECMTION OF AQUEOUS SOLUTIONS by T. E.R. ISBN 0 90224 5 48 8 . WIGLEY_ Climatic Research Unit University of East Anglia Norwich NR4 7TJ.W.udgill. Michael i:hurch. Anthony Young.P. T. B. Curtis & S.thods (P) M. ISBN 0 902 24 5 06 2 The measureraent of soil frost-heay. F. Thornes 1974 ISBN 0 902 246 36 4 Electro. 2 15 16 17 20 24 27 29 37 39 I (8') Charge Balance ( .L. . Finlayson. A. .

Uriiversity of East Anglia.e. . to ha ve a source of standardised information relating to increasing!y sophisticated methpds of data collection and data handling.. The authors and editor have endeavoured to ensure accuracy and where necessary the mahuscripts are being fbrwarded b. a word of caution is in order. In raost cases this juggling is best done. England. School of Environmental Sciences. The prograra uses a siraple computational algorithra which' can l i e easily understood arid so adapted = to suit individual needs.Before proceeding with a d etailed' description pf WATSPEC and its theoreti cal fouridatloris. Furthermore. expressed by the Group as a whole. applied successfully to many situations . in sorae detail in the following sections assuming very little background knowledge on the part of the reader. for the views expressed. such analyses do nbt imm'ediately reve'al the thermodynamic state [or saturation level) of a water sample with respect tp various common minerals such as. magnesium. and can be used to analyse water saraple data from the most dilute waters to those 2 or 3 tiraes more saline than seawiiter. in a work of this length. ITT" m7!'^vm:Mmm^L .C. Like all computer 'package* programs . 7TJ. way WATSPEC can' act as an introduction tq the basic chemistry necessary for a proper understanding-of the chemical processes which occur in natural waters. The user should bear' in mind EDITOR:Dr. in this. in solution) ra-cher than the. lakes. In other respects the authors alone are responsible. NR4. Some underst anding of this chern istry is essential before one can reliably use and interpret the program's output. dolomitey gypsum arid the clay minerals. fundamental concepts must be left unexplained or only partially explained.order to find'out the detailed chemical makeup of a water sample. The cheraical and mathematical bases of WATSPEC are e x p l ' a i r i e ' d . is a relatively simple water chemistry manipulator. The most commonly occurring cationic and anionic species are included as Input variables.EDITORS' PREFACE INTRODUCriON | The publication of the Techriical Bu lletins by the British Gecmorphqlogicai Research Group arises o ut of a common need. which makes consid erable physical sen se. WATSPEC. and its thermodynaraic state some sort of iterative mathematical juggling is required to obtain the values of more fundamental variables^ from those. Atkinson. The need for such a prograra arises becaus'e standard chemical analyses only tell the gross chemical inakeup of a natural water sample (i.calcite. The thermodynamic data and s ' p l ution chemistry use d in WATSPEC is the result of many decades of e Xperimental and thep ret ical development wo-rk by m'any authors. In. some. The computer program'given heres.raph describes the development and use of a computer program for interpreting chemical data f r ' o r a the aqueous environment. WATSPEC: is a tool which should not be used blindly. the total amount of calcium. which are-usually measured in the laboratory or field. Norwich. For more detail the reader is direct' e d towards the excellerit texts of Garrels" and Christ [196S) and Stumtii and Morgan [19 70)'. This is pa rticularly iraportant he-re because a greiat deal 0f bas ic cheraistry is hidd'en within the program. It is hoped that. groundwaters and the ocean. rivers.efore publication to interested commercial firms for critical evaluation. for the dissemination of i riform'ation concerning techniques. a nd for manuals which stress particularly the uses and Iimitat ion's" of the techniques in the field of geomorphology. The theoretical approa ch used is one of a number of possibie app roaches: although it is one which is popular. sulphate etc. It has a wide range of applicability. concentrations of the individual free ions-. Naturally. particularly where this spee ds up and facilitates the execution of a research programrae. Th e deraand for such Bulletins is three-fold: for comparability of results among workers in the field. This monog. and which can be. a computer.

Finally. milligram. One might. . It would be p OSsible to modify WATSPEC to accep t and translate any type of input units. free ions.g. constituent parts. and additionally allows one to conjecture abou t the past and future evolution of the water sample. The prefi 'm' is comraonly used' for molality. Once the. total Mg to total Ca ratio) and other directly measured variables such as pH. (The relation between concentration and activity will be defined later). this was. moiari ty arid molality bear the most common WATSPEC accepts iriput in mi lliraoles per litre of solution [i. ways: parts per million.).re. ask the more fundamental question: why not use the raw.t to do this partly in order to fcee p WATSPEC simple. deterraine saturation levels. the'thermodynamic behaviour of a species in solution is. UNITS WHY THE COMPUTER? ' In this application the coraputer is basically a time-saving device. the density of seawater is about 1. .fjj molarity (M) CFW (volume of solution i r i litres) These three quantities are related (since mass of solution^ mass of water • ^ total raass of solutes) by the following: M ^ pf [4) (S) where p = density of solution (gm/cc or Kg/l) . _ _ total mass of solutes ^ total mass of solution For most aqueous solutions the density is close to 1 (e. Direct chemical analysis [e. loosely combined with other species. In these circurastances M> m =and f are almost equal. Perhaps-. In aqueous solutions its analogue is molality defined as the number of moles of solute per Kg of water. either experim entally or graphically: but such methods lack generality and usual iy haye other•drawbacks . the validity of the assumption of equilibrium should always be examined before results from WATSPEC are interpreted. In principle the calculations performed by WATSPEC could be perforraed 'by hand': indeed. so Ithat mHC0.. incongruent solution The results of chemical analyses can be expressed in many different. ity.e mlllimolarity) and convers ion of iriput data into these units is essential. by titration) usually gives the result as molarity or in units easily converted to molarity (such as parts per raillion or railligrams per litre). WATSPEC. WATSPEC accounts only for commonly observed inorgaiiic species. the raost fundamental concentration variable Is the raole fraction. Y "GFW [mass of water in Kg) ^ " J * mass of solute in gms . The presence of dissolved organic material may have similar effects.e.g. Computer analy.g. calculates the carbon dioxide partial pressure of the sample (strictly the partial pressure of a Cpi gas phase which would be in equilibrium with the' wa'ter saraple) arid determines the charge balance error. c t and Correct gross er rors early in the analysis.2'-\. therefore. A mole is a mass unit equal to the gram-formula-we ight (GFW).7 denotes the bicarbonate raolaiity andraXwould deno. total hardness. Formality and molarity are closely related to raola. mineral precipitation and dissolution. for example. equival'ents or millie. Finally. WATSPEC determines the detailed corapo sition of a solution. s per lli. forraality (f) 'GFW (mass of solution in -Kg) ••••\ mass of solute in gms _. in gms .with the data and able to det e . I have chosen no. from cation or anion ratios [e. in the past. n J^mmmmmmm\^\. The presence of significant amounts of other more unusual inorganic species raay Invalidate the results. computer programs like WATSPEC forra the basis for more. Most chemical species are present in solution in a number of forms. the only means available.complicated programs which can be used to synthesize or model natural processes such as mixing of waters of different eoraposition.. data for interpretation? In some circumstances this is possible. (i. a useful parameter for detecting errors in chemical analysis. I r t addition. and partly because some per sonal data raanlpulation keeps one 'in touch' .that WATSPEC Is essentially an equilibriura thermodynamic model oi the aqueous phase. considerable insight into natural chemical processes can be gained from an examination of changes in total species concentrations [e.tis the molality of a general.g.are deifined below: raolaiity ( * m ) mass of solute.035 Kg/1) and the solution itself Is dominantly water so thst S is very much less than 1.' and straightforward. In some cases it is possible to.. the concentrations and activities of all s pedes.quivalents per litre. evaporatiori 'and GOj degassing. in some ways this is like trying to understand the mechanics of a car engine from its outside: WATSPEC breaks a water sample down into its. . simultaneous dissolution of one mineral and preeipitatio?i of another).sis also facilitates the calculation of other derived variables used in wa'ter sample interpretation. species 'X'. determined by its activity rather than by its concentration. The three . potential determines whether the solution has the th for further dissolution or for precipitati on of that mineral. detailed composition of a solution is known it is possible to determine in a fundamental and precise way the level of saturation of the solutiori with respect tq ariy raineral. Saturation level then erraDdynainic. or as various ion pairs.

is a measjjre of the therraodynamic concentration or chemical potential of that species.60953) = 1.--41Tfti^^^ii* I 0 3 ~ i m / J 8 J S mjoA i ^ ^ m . but in gene. except for a f a c t o r 1 . 1975). Division of mg/ 1* by 100.Q883S converts these units to millimolarity.g. however. Howeyer. Sometimes these will become attached to the original ion for a significant time period and form. less than its true concentration. A conversion is necessary and this requires a knowledge of both . or mllliarams per lit-re. Using the factor 0. solution mass is rarely raeasured and chemical analyses are usually based on.* ' •• " • • .ih Mf. parts and have adtivity coefficients less than 1 (Reardon and Langmuir". however. This cloud tends to shield the ion frora the bulk of the solution sp that it is partly hidden from other species in solution.01732 is th e GFW of HCOy). * 1 In general) although the approximation is a very good one.e'. WATSPEC does not incorporate this recent result. y WATSPEC to be 1. based largely on gas solubility studies._. Frequently equivalent p e r l i t r e a r e a Iraos t.jim.. The f i n a l comm only-used cbnc entration raeasure is equivalent..(8) . gram foTmula weight of solute species gram formula weight of CaCOj although there is an important exc eption to this rule (see below).is reported. * . For some uncharged ion pairs. faking alkalinity to be equal to bicarbonate ion concentration. but there can be large d i f f e r e n c e s i n some cases. For ex ample. identical to molarities (i. Frequently p p m . P. t h e GFW o£ hydfog e n ) mrlliori CaCOa equivalent.cies in solution ts generall-y less than I and decreases as the total dissolved? solidsj^ of ther solution increases (although there are iraportant exceptions to this generalization). this has been questioneii. one would expect the conversion frora ppra* to ppm to be effected by raultiplying ppm* by Most chemical calculations involve activities rather than concentrations.ral the original ion tan be pictured as raoving about with an attendant cloud of oppositely charged ions.T 75":rfl-'. CaSoS and MgCOj appear to behave like dipoles in ai^ueous.^7 ' ^ iuijuj "as tnerefore.'"••IFi'-^'111 . In carbonate waters concentrations are often. To convert ppm* to ppm one. is not the . Ca** • + 2HC0I HaO . However. when the units actually employed a re milligrams-CaCOj equivalent per litre (mg/l*). and the correct fac tor is 1. To con-vert mg/l to millimolarity one need only divide the concentration in mg/l by the gram formula weight of the solute.-. WATSPEC calculates S and uses equation (5) to convert the input into molalities. Frequently units reported as ppm are in reality milligrams per litre mg/l mass of solute in mg volume of solution in litres THERMODYWAMIC (1) Activities and Activity BACKGROUND • cn Coefficients To convert ppm to rag/1 one must multiply the concentration in ppm by the density (gm/cc or K g / l ) .alkalinity • and a large positive charge balance-error. mole of carbonic acid (HjCOj) p r o d u c e s 2 moles of hydro gen so its equivalent weight and f o r m u l a w e i g h t d i f f er by a factor of approximately 2. j. S and p.60963 is quite The activity coefficent of a charged spe. one.equivalents pe r litre of solution where one gram .21927. In. volume.ase.08835 gms. gcomorphoioEy arid hydrology chemical data are often given in ££tvts_ (pf the specific ion) per million (of solution) . If density is not specified then it is -assumed b . As an empirical result. This. Alkalinity (which will be defined later. The GFW of CaCOa is 100. but which is roughly equivalen'^t t o bicarbonate ion concentration) is sometimes.qui v a l en t i s the nuraber of grams of substance which will r e a c t w i t h or p rodu ce one gram of hydrogen. By virtue of electrostatic attraction any charged species will attract oppositely charged species."^ ( o r m i l l i . In this ease the calculations are Only approxiraate (since p .enerally multiplies ppm* by the ratio.e q u i v a l e n ts) per litre of solution. concentration as far as other species are concerned is therefore."-^^^Tt.e. 'The activity^ or effective concentration of a species. The reason for this can be seen with a simple picture of ions in solution. WATSPEC uses molality for internal calculations. '^^.expressed as ppm-CaCOs: equivalent (ppm*). The two are only identical when the solutioh density is 1.>. its •apparent or effective. been chosen for data per million (ppm) Ts a airect concentration variable defXned by ppra mass of solute mass of solution X 10= (6) a common error " a n d leads to an underestimate of the. ».by a factor called the activity coefficient (iTi^) : •aX = (rvraX (9) roa^gfisls' o-^os63 (61. Activity is denoted here by ^he prefix 'a' and activities and molalities are related. This is defined as tiie nuraber of g ram. This anpmaly irises because the dissolution of calcite raay be written as the reaction CaCO. This shielding effect is accounted for by the activity coefficientFor uncharged species the cloud will not occur and the shielding which produces activity coefficients below' 1 will not result. expressed as ppm*.ccording to_f8) dissolution of one mole of CaCOj leads to two oles of HGO3• The appropriate conversion factor in this case is herefore 2(0. activity coefficients for uncharged species are takeri to be slightly greater than 1. CO. 0 0 7 9 7 .21927.. an ion pair (see below).

In terms -of . J I L 4 . A + B t -C I . Thus. If the reaction were written HCO3 = H + C O " then the corresponding. for example. or products. l .A. As a direct result of equation (11).In(aA) . |i "•. concentrations) will decrease and bicarbonate ion acti-vity (and concentration ) will increase until equilibrium is reached.* itiTOT a i a s i y . from left to rights Hydrogen and carbonate ion activities (and.J.ion between chemical potential and activity involves teraperature (see equation (10)) equilibrium constarits are in general functions of teraperature. By convention species on the left are called reactants and those on the right are called products).:&i-. most cheraical analysis procedures (excluding some direel electrochemical techniques) do not distinguish between free ions and ion 10 ). e . Consider as an example the reaction where species A and B Freactants) react to produce species C and D [products]: A f B * c • ^ D col *• U HCOJ (Note.e.-. is actually the sura of the constituent concentrations (13) In raany situations a significant proportion of calcium raay exist in the paired forra. carbonate CaCO^) . and for calculating saturation levels. The reaction is in equilibriura when the chemical potential of A • • • B is equivalent to that of C • ^ D. the product species and K HCO 3 IS the association or formation constant for HCO7.. to form bicarbonate ion. species C and D will also react t. ( ' 3 ) Ion Pairs Equation (11) is the basis for separating measured concent trations into their constituent parts. The cheraical potential (prefix g) of species X depends on the (natural) logarithm of its actlvit> i i X = RT In(aX) •^ constant (10) aHCQi aG0j. b. forward' or backward) of a chemical reaction is determined by the relative chemical potentials of reactants compared with products. . . occur together. calcium in solution occursvas the free ion C a " ^ " * " or as pairs with sulphate (CaSO?) .ty to react) is analagous i r iraanyways to raechanical potential energy (or capacity to do work). hydroxyl CCaOH''') and other ions (tive four mentioned here are the most iraportant calciura ion pairs in natural waters and are fhe only C a ' * " * ' pairs considered in WATSPEC).In(aB) ? constant which can also be written as aC aP aA aB K (il) where K is the equilibriuin constant ^for the reaction.(12) and aH are such that aH^ a C O " > t ^ the reaction (12) will proceed "^HCO. ^ .nf?ether to produce A and.G and D are all present'.and T is temperatuTe in °K.iir iS»-fel. which is what would be measured by titration with EDTA for instance.T8i-'l:. As a specific exaraple consider the reaction between carbonite and hydrogen ions. Thus. The chemical potential of A and B is therefore uA + (iB. At equilibriuin.H.. the law of mass action necessitates" that *^HC0l 10.U3 All chemical reactions are reversible. nX = Y + Z) then the activity of that species must be raised to the power n in the law p£ nass action.icarbdnate (CaHCO|).. i .'h. l . Most dissolved species exist in solution both as free ions or in association with other species as ion pairs.production of reactants.6 to 94.'-^«.(2) Law of Mass Action Activities detennine the direction of chemical reactions through the fundamehtai law of mass action. lower values corresponding to cases where sulphate raCa mCa** •^ raCaSO? + raCaCO? *• mCaHCO:^* * mCaOH* where R is the gas constant. The net direction (i. Equation (11) is a statement of the law of mass action. Just like potential energies. V f l i e n aCO) ML. A + B The two reactions will. in general..V^tf^^ 08-?/WC/TOT.D) and backward ( . the law of raass action is always written in the form J .and would be called the" dissociation constant for HCO. or vice-versa. B: " C ^ D' * .^ . Cheraical potential (ca pa ci. ! .V 13.aH-" constant .activities [through equation (10)) this leads to the matheraatical statement In(aC) • ^ In(aD) . _ products reactants In the example above HCO^ is. > - . "Thus.5.g. if a species occurs n times in a reaction (e. the equality sign is used to indicate that no particular reaction direction is specified. if .of -K„„-. equilibrium constant would be the inverse of K„p„T. Sirice the rela-t. The coraposite or net reaction is said to be in equilibriura when the forward (i.-. By convention. C » • D + A + B) reactions proceed at the sarae rate sb there is no net . In addition. cheraical potentials are additive. free: ion accounts for from 70.329 (at 2S°C the value. the total calcium concentration (or molality).B. Wigley (1971) gives examples in which the.% of th'e total calcium concentration.e.

key species for aluminiura. b. It is defined by = ii=l s .e. aluminium. results can be written for all of the other species. relating total molalities (e. bromide. sulphate.g. for example. In (14) and (IS) the Ks are. not be in equilibrium with other phases such as solid minerais or gases with which. ironj strontium.was present in substantial amounts. (ZO' . is the charge on species X^. Equilibrium within a single phase is termed homogeneous equilibrium.g. The activities in (14) and (IS) can be replaiced by products. while. "'^" "^CaSO^ '^Ca** ^ C O r ' ^ " %aCO? "^^CaSO? ' C a 'HCO... although generally very good model's.7) activity co-' efficients of charged species can be determined using the Extended Debye-HUckel equation (see. is the molality of species Xj. For dilute solutions (ionic strength less than 0. for reference. 1970. /I = mCa T ^Ca*^ 1 t Sisor "jSO. the latter are. l i i X .over all charged species. Such prograras are therefore! orily equilibrium raodels of the real situation. *^°03 ^CaHCot ^ca^. In this and in all other cases the key species is that which dorainates over other species related to it in 'normai' natural waters. and HjBO] are the. of activity coefficients and molalities (see. solve these . they are collectively called aqueous models. lithi.1 it is necessary to furt. key species here is HCOJ. sulphide and ammo tii um species Carbonate species are determined by alkalinity (see below) • The>..(14) nitrate. borate. chloride. WA'TSPEC (and other computer programs sirailar to it) only gives th'e equilibrium chemical speciation of a water sample. the former are determined by the temperature and ionic strength of the solution. for example. and all the equilibrium constants. the inverse of the dissocatipn constants). in general. depends on an empirica 1 rule called the Maclnnes convention which gives excellent result 5 up to high ionic strength. (4) Calculating the Concentrations of Individual Species In natural waters it is gerierally assumed that dissolved (or aqueous phase) species are in mutual equilibrium within the solution. The assumption of" homogeneous equil ibriuiii is iised to calculate the detailed speciation.mX. bariura.The . and the restated mass balance criterion Xi^6') (which is derived from the mass balance equation (13)) together form a complicated set of non-linear algebraic equations which is solved by WATSPEC to calculate the molalities of all the individual species in solution.method which WATSPEC uses to..^m'^^^'h. mCa**"*) . WATSPEC accepts as input data total molalities of calcium.[IB) 1 • > • a-B/T 1 A and B are known functions of temperature and a^ is a parawhere meter called the effective ion radius (which has known values deterra ined from experimental data for most comra'ori ions). HuSiOi.. that between different phases is termed heterogeneous equilibrium. mCaT) fo the molalities of key species (e. and the following modifi'cation of it.then substituted into equation (13) one eventually obtains mGa + 1 where Z.) to A I. silicate.r. (-S) Extended Defaye-HUckel Equation In order to use equations like ( 1 4 ) . /I *b. carbonate. potassium.equation (9^)). Equations like the equilibriura relations (14) and (IS) .1.. For example. FOr exaraple A l ( O H ) ^ . although the solution as a whole may. ' PTSST A Z. sodium.(17) aCa 10 •log. the free ions Ca'^'^ and S O " and the ion pair CaSO° will"be in equilibrium (although calcium and sulphate species need not be in equilibrium with. :and the summation ds . known functions of temperature and pressure. Ionic strength is a measure of the total concentration of charged ions in solution. r . magnesiura. For ionic strengths above 0.:aOH • i r GaOH (16) .equations is described below in the section on Iterative Methods. (IS) and (16) it is necessary to know all the activity coefficients. 1 (19) 1 • ^ a^ B/T 11 tmr^fwrji 08-e/Hg/l8T# uuod lllt''r% . *^CaSO? aCaCO^ aCa** aCOl" "^GaCQ? -CIS) etc. the ionic strength of seawater is approximately 0. This equati on. The assumption of homogeneous equilibrium within the aqueous phase is a fundamental one.' -log JT^ Similar. Stumra and Morgan. ^CaGO? mco. the mineral gypsum). silicate and borate. If the results are.association constants for the ion pairs (i.her 'extend' the Debye-HUckel equation erapiri cally (introducing another is in contact. The appropriate mass action equations can therefore be used to relate the ion pair activities to the free iori activities: aCaSO° .

. At equilibrium these species are related by the cheraical reactions raH (2S.' 1 ' ^ no_r /T .MgHCO) etc.ionic strengths as predicted by the simpile shielding explanation given earlier. (6) Alkalinity and Carbonate Systeras 1 The other variables are difficult to measure individually. C o n s i d e r a solution containing dissolved carbonate species. For many species tht value of bi is zero." • ^ H* = HCOj for which" the following mass action equations raay be written aHjCO? aCO? aHiO •• h Q l aHiCO? aH* aHCOi aHCOl aCOv" aH"^ ° "^HCOa (23) "HiCO.'H'J'-H. the carbonate systein is comjpletely determined by specifying pH and alkalinity.0 (26) Equations (25) and (26) introduce two new. . HCQj (bicarbonate ipn) and C O " (carbonate i o n ) . CaHCOj. H* and 00^' together with (21). At ysj-y high ionic strengths activity coefficients may begin to increase with increasing ionic strength..) from the following discussion. They differ from ion to ion. and b. defined for this system only (see below for a more general 'definition) by ALK = raHCOj + ZmCOj Note that mOH' and mH aH + mOH It has already been noted that alkalinity . and for the activity coefficient of water where . a^ and bi values used In WATSPEC are those determined by Truesdell and Jones (1974) and used in WATEQ and WATEQF. Equation (19) generaily gives a decreasing activity coefficient for increasirig ionic strength up to moderate. GPi and HjCOi.017(XT0T). ignoring activity coefficients. included. (Plummer et al. although the calculations become a little more coraplicated than the outline"given above when activity coefficients and ion pairs are.o will be enough to specify the whole set. the ion pairs CaSO? and CaCO?). both In practice and according to the predictions of (1*9).^ (22) (21) The molality of water is unity so that. axe determined of the input variables for WATSPEC.Alkalinity is'an important variable in carbonate systems. variables (ALK and mOH ) enlarging the set of equations to five with seven variables. In practice it is not necessary to know both CO? and H^COi •and they are invariably combined as one variable. for solutions up to about three times as concentrated as seawater). is based on raeasured activity coefficients and is valid . e.ll 10 (ZO) except for free carbon dioxide (defined below) where WATSPEC use! a raore complex relationship.g..up to ionic strengths of two or more (i. For simplicity I will omit ion pairs (such as CaCOf.. and these systems are most easily examined experimently by determinii alkalinity and pH. (22) and (23) deterraine HCO.. alternatively written a. UfUl^Ti]!^^ T! J^'i^ 1976). The parameters a. .WATSPEC uses C0° ••^ H^O = Hi CO 3 \C ^ HCOJ = HjCO? GO. and the value of a^ which is appropriate when using equation (19) is not necessarily the sarae as that which would be used with equation (13).'. pH I -log(aH ) (24) XTOT being the sum of the molalities o£ all dissolved species. and that this is a measure of (and often approxiraately equal to) the bicarbonate ion concentration. determined by Wigley and Plummer(1976). T h u s .e.) are related by aOH aHiO ^H.Equation (19). (Note that the mola1ity of HiO is 1 by definition). and instead it is easier to raeasure alkalinity (ALK). Their advantage is that alkaliriity is easily measured by titration with a strong acid. The carbonate species then are: C0° (dissolved carbon dioxide. For uncharged species (such as. which WATSPEC uses. the above is a set of three equations in five unknowns . for example. Hydrogen ion activity can be directly measured as pH which is defined by HiO s aHiO . Thus. the free carbon dioxide content: 12 13 . WATSPEC determines the activity coefficients using the erapirical relation O. COiiaq)' HjCO? (carbonic a c i d ) .. . any tw.

The free carbon dio. the thermodynamic state of a solution relative tp equilibrium with a specified solid-'phase. dissolution will occur if one neglects any kinetic limitations to reaction progress)> and if TAP > Kc then precipitation will occur. in theory it can te d etermined by titration with a strong acid (p. + ZraHiSO?) is a m easure of departure from equilibriujn.mineral.7' Kr since solid-phase mineral activities are taken by convention to be unity. SI^ < 0 means lAP < KQ and unders aturation witK respect to calcite.. is a measure of. For exaraple. Ther e are. (7) Mineral Saturation Indices (28) This notation is useful because PCOi is generally a small quantity. • ^ mAlOH*"*^ + + 2(riiAl(0H)t) • ^ 4(mAl(0H)O • ^ m H S ' •f-2mS"'+ mNH^^aq.. titration alkalinity and total alkalinity.!.' + niHS" mMgCOV + + 2mS -^ mHiSiOi.0 to 3. This term. is a little misleading because CO? dominates over HjCO? by a factor of about 1000.-. therefore. well established.* • mHgHCOa raNaCOO •^ mNaHCOj + 2(iriC07" * mCaCO? • < • + mAl(OH)i. It is necessary.reaction kineti cs raay allow a . in fa'ct.PGOi and HjCOj are related by Henry's Law: "H PCO 3 where K | . . The litter two are of concern here. In practice such titration is usu.'-'i(!i'r.e. and the" index SI. If the product aCa*"^ aCOJ" (called the ion activity product or lAP) is less than Kj. 'p' = '-log') so that pPCOi -log(PCOi) aHjGOl (2 7) This expression only differs from the total alkalinity for the WATSPEC aqueous raodel in that iron-hydroxide species.further it is often useful to employ the 'p' notation used in the definition of pH ( to theraethyl orange.Q. 1976.(31) where subscript 's' denotes solid-phase.xide content is.raOH + mliOH (3p) The difference between ALK* as defined by equations (29) and (30) is usually very small. To illustrate. frequently replaced by the GO2 partial pressure of a hypothetical coe'xisting gas phase: PCOi. or saturation corresponds to lAP = K Q .S and for most groundwaters pPCGj ranges frora 1. 5 ) .4-.mLiO. l°g i lAP (32) mCaOH* * mMgOH* + mSrOH* •^ raBaOH* -i. (Note that this is only a hypothetical reaction and doesn't precisely describe the meehan^ isra by which calcite dissolves or precipitates: see"Pluramer and Wigley.i- .and Plummer et al. IS 14 T E^i7F'"t"'-":*'• I''• '""L •'.mHiGOj = mCO? + raHjCO? The notation used here. To go one step. + 2raHiSiOi. For a solution to be in equilibrium with solid-phase calcite the following reaction must be in equilibrium. The notation i s . I will consider the raineral calcite. slow. in calcite. carbonate al kalinity.mNaHGOj + 2(mC0J. Other programs use different definitions of titration alkalinity. to di stinguish a pra tical titration alkal inity from the ideal concept of total alkaUnity WATSPEG accepts titration alkalinity (in railliequivalents per litre of solution) as input. mMgHCot i . [ is a function orily of temperature and which is usually given in units which determine PCOi in atmospheres. Equation (3 2) defines the sa turation index (SI) for calcite. The. however. while SI^ > 0 raeans lAP > Kc and the solution will be supersaturated with respect. end-point (pH ... mH + mHjBOi -t. ^ 2 m H i S i O " + raHiBOi. . Total aIkalinity is a.!. measure of th e ability of a water to neutralize acids which are stronger than carbonic acid.wfpf. three different types of alkalinity." f mCaGO? * raMgCOj * mNaCaj) * raHjSiO.-^ mOH" + (mH* -i. Other saturation indices are defined in a similar way.raHCl° -i(29) WATSPEC calculates saturation indices with respeet to a nuraber of minerals. which follows Stumra and Morgan (1970).H° TiHSO. 19'76.!f. The alkalinity' d eflnition given above (equat ion (25)) IS only valid in an idea lized carbonate system. have been* neglected. 1977). Equilibrium.then the 'reaction"" (31) will proceed from left to right (i. Note that s aturation indices only determine the thermodynamic potential for di ssolution and precipitation. and not a ll theoretically neutralizab le weak a cid species are neutraliz ed during titration because of slow c hemica reaction rates.IN HCI is often us ed) to the carbonic acid equ ivaIence point (see Stumm arid Morgan . EaCO s. For the -species corisidered by WATSPEC titration alkalinity is defined by ALK* = = mHCGl • • • mCaHCot ^ .= 3. for further details-) .solution to raaintain a super^ or urtder•satura ted state for a v e r y long period of time.mass action equation for (31) is aCa aC0. which react slowly. For the atraosphere pPCGi. In practice. an abbreviation of 'degree of saturation' index. WATEQ uses ALK* mHC07 •^ mCaHCOj.s Ca** + col' .

301) reduction: oxidation: redox reaction. QBE V^mX. I 1 1 lOOi (34) Y I ^ X Z.all these pairs the oxidized form is given first. Because sulphate concentration is difficult and time consuming. Such reactions are conceptually analagous to acid-base reactions. p. 2. t. sulphate. other units mentioned earlier can produce CBE values up to l l \ . 0^ to HjO (equation 3:7) into two parts (equ ations 3 5 and 35).e 2. or at least to-exaraine it closely for errors in chemical analysis or conversion of units. (33) 1 I i=i wher. redox. In equation (34) superscripts C and A oyer the summations indicate summation over cations and anions. Jiist as the proton activity is corive'riientl. so the electron activity is determined by the pe. (36) . r. Large positive pe implies st rongly oxid i^ing conditions with a predominarice of oxidized f orms. (37) IJ DI ^ 4Fe** + 4H* = 4Fe*** " "iW iO Charge imbalance is often measured using the charge balance error which is defined by raX. In the field or laboratpry pe can be determined by raeasurihg the.. I n . pH = -log(aH*). as redox reactions so that free electrons (e" in C3S1 and C36» do not actually occur in solution. Even larger errors can arise from negieeting to analyse for species such as sulphate or chloride when these' species are actually present. condition s with a preponderance of.y dete. and/or the existence of species which raay be preserit in a saraple but not measured. iron occurs as Fe''"* and F e ' * " * " ' " . Z. For instance. S O r and di ssolv ed HjS. the possible error in converting alkalinity frora ppm-CaCOj -equivalent to. Because of chemical analysis errors . The-probable level o f charge imbalance depends on the corapleteness and accuracy of the chemical analysis procedures used: the acceptable level de pends on what use is to be made of. The method is to assume that any charge imbalance is due to (unanalysed). potential (Eh) which is the potential difference (in volts) between an inert (Au or Pt) electrode immersed in the solution and a standard hydrogen electrode. while sraall or negative pe implies strongly reducing. pe and Eh are related by Eh P^ ° 1. and NQJ and dis'solved NHj. dissolved O2 and HiO. Conclusions or wha t iraplications are to drawn from data which shows a ppreciable charge Irabalance must be regard ed as rather suspect Certain species can occur in different oxida-tion states: for example. Oxidation and reduction reactions always occur in pairs. Nevertheless. it is spmetimes estimated 'by" di'fference' . sorae irabalance of charge will invariably occur when one compares the t otal nuraber of positive and negative charge s. HCOJ and dissolved CHM.rrained In logarithmLC form by pH.reduced forras> Exaraples o f redox p airs or cou £le^ are Fe*** and Fe**. to determine.(8) Charge Balance ( ' 9 ) Reduction-Oxidation Reactions A balance between the total nuraber of positive charges on positively charged ions and the total number of negative charges on negatively charged ions is an essential condition for any water saraple. OJ * 4H ^ 4e" = 2HiO 4Fe = 4Fe + 4e -. reductants and oxidants can be defined as electron donors and receivers. CBE values outside the range ± SV are usually sufficient for one to discard the data.9B422 10"" T 17 16 . n I mX. The following example has been taken from Stumm and Morgan (1970. in the same way that acids and bases may be interpreted as proton (or hydrogen ion) donors and receivers. b e drawn from the'data. Reactions involving different oxidation states are called reduction-oxidation (or redox) reactions. • ^ ^ m X .is the charge bri spe cies Xj^ with inoi al ity mXj and siimmation is over all charged spe cies. pe -log(e") „ where e i s the elect ron activity. This splits the simultaneous oxidation of Fe** to Fe** *-and reduction of. I . The pe is a key variable in determinin fi the relat ive araounts at equilibrium of all species which can occur in di fferent oxidation states. when redox reaetions are split up in this way the electron can be consid ered as a 'virtual species' whose concentration determines the p otentifil for oxidation or reduction and the relative concentrations of oxidized and reduced forms fsuch as Fe*'*"^ and Fe-**) of 'any given species.(3S) . This is definitely not a recommended procedure since the-charge balance criterion is "autoraatically fulfilled and no longer provides a check on the quality of the data. or for omitted species. This charge balance condition can be written-mathematically as .

Oj . except for "the amm pnium species. concenl i ri u r a "thermodynaraic data s are determined by using. Eh is less coramonly measured than pH for a nuraber of reas. The partial pres'sure of oxygen in the atmosphere at sea-level is O.5 the pe is 1 3 M 1 . By GO nventldn t he reaction 2H ^ 2e = HI •2 ^^'•^ ~ 4 0'-. Gas-phase partial pressures and aqueous phase activities are reiate.9 10"". . the concentrations of many oxidiEable and/or reducible species are frequently far from the values predicted by equilibrium thermodynamics. As a second. Reaction (35) can be written as Fe + e for whi ch.. to quote Stumm and Morgan (19 70.the equilibriura constant at IS'^C is 10'"^^ •°^'^. t and many others For . Th e redoX pairs . The law of mass action gives aFe e_ aFe** so that Fe 10 •13. One can • use this to calculate the pe of a solution which is in equilibrium with the atmospheric oxygen concentration.5376) = 20. •21. Now.The equilibrium constant for this reaction at 25 C is 10~'l-''^'^. quant itative interpreta'tiori with respect to solution coraposition i s often-not justified'.ctrbeheraically.013 At pe = 13.49S 10 (39) Using aO. NH I and ass pciate d species. rs are the five pairs. Eh is actually measured with a probe containing an inert electrode and a reference electrode (usually calorael. HS'.with the atraospheric oxygen reservoir and these ex:amples are only given to illustrate calculation procedures.l.all the'se pairs Fe(OH)? . At pH = 7. 1 .aq)^ aH* e' (aHiO).i^aq = 2. and taking negative logarithms of both sides of equation (39) gives pH pe 2 .2." for a gas 'X' is aX = Kj^ PX Here PX is the gas-phase pressure arid K„ ^ is the Henry ' s Law constant for species X. Two quantitative examples may help to clarify the issue and to illustrate the type of calculations which WATSPEC perforras.cies in a redox pair occur in very low concentrations so that many important details of the speciation of a water sample can be determined.wh ich' WATSPEC conside given a bove WATSPEC al so considers a nurab er of other species related to th e indlvidu al pair members.61 The pe of a solution In equilibrium with atmospheric oxygen therefore depends on the pH.) which has a known potential. 1. a. Because of point (iii) WATSPEC (following WATEQ and WATEQF) can i f certain redox pair concentracalculate Eh f rom the details o tions (special cally th e SO. example I will consider reaction (36) above an-d calculate the ratio of the activities of the free ions Fe-*** and = Fe** at equilibrium in a solution with pe = 13.".where T is in °K. (ii) since many redox reactions are slow. 9 TT . (aOa. withoutrieedirig'Eh "data. For oxygen at 2S°C. the O.097} if redox equilibrium existed. p 350). 4 . 18 1 . K H is 10"2 so that the equilibrium oxygen activity in the aqueous phase is 2.teraperature correctiori may be necessary.11.ons including: (i) quite often either or bpth spe.11 the ratio aFe to aFe. from equation (38).Zl atmospheres. Reaction (3S) above may be written as aFe aFe pe 13.I (3. which. Even when u sing this method one should note that these pairs need not be in e quilibrium.013 (40) >gas log is arbitrarily assigned an equilibriura constant value of 1 and equilibriura constants for other redox reactions are deterrained relative to this value. FeSOt. (iii) Eh is a difficult pararaeter to measure ele. HiS pair. HjO pair is frequently not in equillb rium. would therefore be. when re producible results are obtained .25) = 0. H2O pair).2 5 (note: log(. and the dissolved. Because the refer^ ence potential deperids on temperature.aH^O = I. sp ecies such as . Sine e free electrons dp not actually exist such ma ss act ion equati ons cari only be int erpreted in terms of the rel ative activities or concentrations of different oxidation states.d by Henry's Law. The reference electrode potential must be added to the measured potential difference to obtain the Eh. equil l tra t ion and the law 0 f mass action for reactions s uch as equation (35) and equ ation (35). Actual water samples are often not in equilibrium. 'Eyen. relative to the standard hydrogen electrode.9 10~ H^ (38) where subscript 'aq' denotes the dissolved species. 9 5 . Furth ermore.

In many natural waters. The bas"ic equations are tho se of mass balance (e. the re are 18 such equat ions in WATSPEC) and a number of mass action equations (e. much more complicate d equa-tion.e ionic strength of the sample is known and where empirical equilibrium constants. incorporating activity coefficients within them. To illustrate this." ^ " .mP (cf. This need for less generality in orie sense and ^increased generality in another has led to the development of a large number of programs specifically designed for the study of chemical equilibrium problems in natura waters.. One su ch method is to use rae-thod. ITERATIVE METHODS WATSPEC calculates the detailed chemie al composition of'a water saraple (i. In addition. There are 20 To illustrate the cheraical equilibriura problem and the raethod of solution employed fay WATSPEC. Examples three unknowns which can be solved for the. equations (41) j (42) and ( 4 3 ) constitute a set of three eouations. HALTAFALL. 1973). both published and unpublished .'•:. Ah early application of this approach w as to deterraine the detailed chemical composltion of seawater (Si l l i r i . Other similar programs coriside r more species.mP .. This gives. van I eggeren and Story (1-970') and Wolery arid Walt ers (197 5). Dyrssen et al (1968). Although' developed independently by the pre sent author (the earliest version was written in 1970) the p . 1976) and SOLMNEQ (Kharaka and Barnes. and anion 'A') and a single ion pair ("P'). to solveC*. equa tions (12).and of varying de.. a first estimate for the act ivity coefficients whic h allows. . 1961. mC' and mA^. mC* and mA ) . (14).'(i0. and the programs WATEQ (Truesdell and Jones. WATSPEC works within the limited f raraework of an aqueous raodel containing up to 18 total con centrations (70 i n = dividual species) and assuming homogeneous equilibrium between all species. since„ activity coefficients are relatively insensitive' to raoderate' errors in ionic strength (and hencfe in..(4 3) mA = mA t .:. The species in solution are (assuming singly charged major species) C* and A" and the uncharged pair P wfose activity is related to that of C* and A" by trie mass action equation aP aC -aA ^--1mC mP (41) jr.g..rogram has been modified to parallel WATEQF (Plummer et al. 197' 6) a program which is an expanded version of WATEQ (Truesdell and Jones.mA The followirig mass balance relations raust also hold = mC* -1.including a general formulation of activity coefficients.extreme flexibility of input co'iiiponents is not required.. and use these guesses to estiraate ionic strengt h. but WATSPEC has been designed to be a compromis e 'between extreme generality and conciseness and -simplicity WATSPEC considers the most common species found in natural waters and the prograra is sraall enough and' fast enough tb be used ." • ' " • «*•???!. which in turn depends on m C * " and mA'. therefore.raany others. WATSP. 19.62) . however. 1974). even in this a first guess of the unknown concentrations ..(4 2) T + .such programs are those used by Wigley (1971) Langmuir (1971).••''[•: •'"i. Morel and Morgan (1972) with REDEQL. WATSPEC uses S3 mass action equatio ns). the range of possible constituents is known and . the a ctivity coefficients) depend s on the unknown concentrationsi Bit her equation (41) must b e replaced by a.e. the main need for using an iterative solution is not due to the activity coefficient terms in equation (41). Thrailkill (1972) and van Br. Equation (4l) becomes fflP = K mC raA (•4 l a ) mC 21 'TP^^'-'^fvT'^"™" ' . For computer analysis of natural watersi.eeman (1973). and the field has been reviewed by Zeleznik and Gordo (1968). gener al coraputer prograras have benn developed by Ingri et al (1967) wi th HALTAFALL. mA and mP.r*n*it. can be specified. The proble m of finding Individ ual species concentrations from total cone entrations is often referred to as a 'ehernical equilibrium pro blem' ..unknowns mC .on quite small computers." i . mC* and mA~ . the concentrations of al 1 free ions and ion pairs) from a knowledge of the total concen trations of various species.some sort of iterative method is necessary even without the ionic strength/activity coefficient difficulty.EC solve s this problem by using a r t equilibriura con stant approach which was pioneered by Brinkley (194 7) . (15) etc. equation (13). In general. or one must so Ive the set of equation s by an itera tive simple example. a built-in formulation pf activity coefficients is highly desirable. 197 4).More recently. Since the activity coefficients are determined by the ionic strength. These the equations to find a better estimate of ra values can be used t o determine a better ionie strength estimate and the procedure ca n be repeated (iterated ) over and over until negligible. are most useful in cases'where th. equation ( 1 3 ) ) . Such programs. I will simplify equation (41) further by assuming all activity coefficients have the value one..'t"''!iii'i!'r'•*.-^ grees of coraplexity. In any realistic problem there is a rauch larger set of equations (over 70 in WATSPEC) and. mA" and m P. and others.. . REDEQL and some other very general programs are limited by not. Unfortunately t he ionic strength (and.changes o ccur between each itera tion. ionic strengths are not known a priori arid can cover a wide range.. I will consider a simple example in which there are only two major species (cation ' C . In this case routine cheraical analysis would determine the total molalities. WATEQF (Plummer et al. Garrels and Thompson.g.

617188 •727273 S 6 7 . = mC ^ K mA*'^} • JI • mA^ = I K = -"i Ti-1 mAj. .73'2051.731707 . m C ' * ' = .K mC mA n n •*• - (4 7) . t h e s e e s t i m a t e s y i e l d a first f mP 1 ^ K mC The table below compares the results of the two iteration schemes outlined above for the example with mCT = -mA^ = 1 and K = ).m A 1 • ^ K raC' 1 • ^ K mA"* . w h i c h is p e r h a p s t h e m o s t ibvlous o n e .K raCi mAl 1 g e n e r a l t h e '[n-i'l)-th e s t i m a t e can be found from t h e n ..fraction) 1 1 1 2 O.666667 3 0.i mAi T T + _ iriAji = raA .mployed by W A T S P E C .3991 . and a c o n t i n u e d f r a c t i o n m e t h o d .T' raC 1 ^ ^ K [rac'^'-mA''') i .4) g i v e s raC / I . raCi mC mP I T t m../(l-K(raC'^-mA''')) ^ t . (42) and solution given by (43) actually have an analytical mA' (-44) mA 1 ^ mP iiA = mA- 1 ^ K mC (SO) 2K and s i i n i l a r r e s u l t s f o r mA" and raP. In b o t h m e t h o d s we s t a r t w i t h f i r s t 'uesses of mCi niAi = jnC raA .4 K.. n+..000092 It takes the brute force raethod 27 iteration's before its error is less than this latter value. (4'5) estimate mA n t h e b r u t e f o r c e m e t h o d .= .S 0.732051 mPi K mCi mAi hen (42) and (43) can be applied directly to give second estiates of mC* arid nlA. estimates will be given by mCj.= raA 1 f K mC and the general iteration relations are (compare and contrast with (4 7) arid (48)) { raC n-n -1 rac'^ 11 + K mA^ i {l .875 0. in (51) a continued fraction n..1 = 0..75 4 .. To raake t h e problem l e s s a b s t r a c t I w i l l t a k e a r b i t r a r y n u m e r i c a l v a l u e s of mC'. mA'''' and K: raC Substitution into (4. . I w i l l c o n ^ lider two o n l y . A''' If the same first estimates are used as before (equations (45) and [46)) then second.In this case (41a).809540 ..K mA (49) 23 n O .": .732143.t h es t i lt e using ""^n+i.C .tr /-rn- •litiifti 'm-\ .733333 .(S2) To s o l v e the problera by i t e r a t i o n on a c o m p u t e r ( n e c e s s a r y lecause an a n a l y t i c a l s o l u t i o n w i l l g e n e r a l l y n o t be p o s s i b l e ) rhere are a nuraber of d i f f e r e n t ipqssible a p p r o a c h e s .-C48) After seven iterations the brute force method is still out by .}"^ .(4 5) .. equation (l6)) mC mC mP 1^ mC* = mC 22 1 I .041940 while the continued fraction raethod is only out by . Iteration 7 should be corapared with the exact analytic result.77. . etc. K mC mA n n .672523 .1 C mC. C n-1 V ' to give o m -. w h i c h is the m e t h o d .rst juggled a little to give (cf.r lT. a b r u t e f o r c e m e t h o d .mPi = raA . I n t h e f o l l o w i n g t h e s u b s c r i p t d e n o t e s t h e irder of the i t e r a t i o n .i 1 + K..K mC.(51) . .mAT The name of this method comes from substituting successively for AI . nm C (tanite force) m G '(continued. Krac. ' ""^ mA n-*-! mA T I t e r a t i o n number. It should be obvious frora this example that the speed with which the correct answer is approached (and hence the computer time used in getting a satisfactory answer) depends critically In the continued fraction method equations (42) and (43) are .

species. More . modified slightly to include activity cb efficients.•ence. Optionalraino. ionic specified In any of five different ways. (Forraat. l blank) indica tor card to indicat e the extent of the data. Output is selective arid contains a summary of input details. With this option EZ (columns 12 to 21) raust also be specified. theh they may be left blank or sped fied as ze ro on t heraajor spec les card. PGOi. FLA. sorae of which are mathematically qui te sophisticate*d and are a little raore efficient than the continu ied fraction raethod.(If there is no r eduction-oxidation data these will be the 3rd and 4th card s). Ehraay. FLAG3' = I. chlo ride and sulphate. or omitted. total sodiura. is given on a separate card (CARD 3 ) . (211. species... the brute force method never gets to a satisfactory answer. total calcium. -sod ium-. FLAG3 FLAG 3 t Z l ' is the electrochemically measured field Eh value.FIO. (Il. Eh can be specified in ariy of 5 ways. values must be eithe left blank or specified as zero. total magnesium. [6(r2. 3. All coricentTations must b .0. Wolery and Walters (197S) and others. Optional Eh card (Format.0)) Column 1 gives the v alue of FLAG! which indicates the presence (FLAGl = 1) or absence [FLAGl = blank. The program accepts. Minor species data a re given on CARDS 4 and (if necessary^) S. calc ium.0. (up to S per sample). Unknown. PO2 and PGH^ (as negative logarithms). The type of Eh input is specified by the value of FLAG3 given in cpluran 1. species molalities for 69. A car d with 99 in columns 1 arid 2 must be placed aft er the la st saraple to. Each minor species card can have up to 6 rainor species s o FLAG2 can be 0 (or blank). EZ is the Eh of a stand ard Zobell's solution measured at the same teraperature as the sample. corrected for sample teraperatui 2ZZ is the uncorrected field Eh (Calorael refer.on the type of iteration scheme used. total^' Mimssmms . The siraplest data iriput requires two ca rds per sampl e. :ZZ2 is the negative logarithra concentration (in moles of Oi tion). s should be made in interp re ting the' ou extensive dat a input r equires mo re inpu t cards. FLAG3 = 4: FLAG3 ZZZ is the negative logarithra of the total sulphide concentration (in raoles of HjS per litre of solution). Reduction-oxidation data. CARDS 4 and 5.rspecies cards.) . pe. species not included on CARD 2). Golumn 2 . Format. Other input data required are temperature and pH (both essential) and density (optionai). exce ss cation charge. total dissolved sol ids. titration alkalinity and total sulphate. I or 2.0)) Temperature (°C). 2. ZZZ is the uncorrected field Eh (Calorael reference). Major species card (Forraat. magn esiUm. Columns 12 to 21 (EZ) are only used If •FLAG3 = 3 ' . pH. if the v alue . a ( p o . and a' major species "card giv ing detail s of total concentrations of the. Input is given on 2. in volts. 4 pr 5 cards per sample The nuraber of saraples wh ich may b e analysed at any one time is arbitrary and need not be specif ied. total potassium. (9F6.tal inorganic carbon molality.0)) HOW TO USE WATSPEC WATSPEC is wri tten in standard FORTRAN I V . s : s l ] ) .. The data must be in the following order.2F10. ion balance error.4)))Details of rainor species total concentrations are given on CARDS 4 (if less than or equal to six minor species) and 5.'Eh is calculated using the sulphide-sulphate redox pair. 0. to. Due allowance.G10. in some circurastances. Ot her methods are available. Eh is calculated using redox pair (see equation (38) of the oxygen per litre of solu the water-oxygen in the text).-WATSPEC uses the contin ued fraction method. tput). up to 18 to tal concentrations as input. in fact. and is quite efficient in its use of computer time. More detailed discussions of the numerical a spects of chemical equilibrium problems are given by van Zegge ren and Story (1970). e asrailliraQlarities(millimoles per litre of solution) except for 'titration alkalinity which is in milliequivalent per litre of solution. A source listing and sample input and output are given in Appendix C.G2 is the number of minor species cards (i. tog ether wit h titratio n alkal inity. in both cases WATSPEi: assumes density » 1.-8) of reduction-oxidation data. pH arid temperature [If any major species concentrat ions ar e unknown. Temperature. CARD 1: Indicator card ( . po tassium. pH and seven raajor species total concentrations are given on this card.indi cate the end of the data.of FLAG2. CARD 3 . FLAG3 may have the values 1 to S.e. All these concentrations must be specified in milliraoles 25 24 -. and minera I saturation indices for 40 rainerals. Columns 3 to 12 give solution density in gms/cc or Kg/litri If density is unknown then this may be left blank or assigj ed any value less than 1. The simplest data input will therefore require a blank care for CARD 1. CARD 2. Columns 2 to 11 (ZZZ) give the value of the Eh variable.

^ Garrels.R. W. . minerals and e q u i l i b r i a .99 and -99. WATSPEC is not infallible..99. 28. R ' . Input of data on s pedes which partidpate' in redox reictions requires furthe r explanation. Dissociation constants of calcite and GaHOOt from'0 to SO°C. Ceochimca et Cosmochimica . then the saturation index for tlvat mineral will be printed out as -99.. 1261-1286. Eh could be calculated using the sulphide-sulphate redox pair (redox input option S). N. R. 9. New York). 1965. WATSPEC til detect and signal the presence of the following probajile (consistencies in the input data: .program may fail if thermoynamically inconsistent or incomplete data are given as input. This is signifi cant for the sulphide-sulphate and ammonia titrate redox pairs. High-speed computers as a supplement tp graphicalraethods. PB 215-899. 16. Data .. The G input forraat is a general format which accepts both F and E input formatted" data.M. ijnle ss total sulphide is given o .per litre of solution (same units as CARD 2). ) Input o£ total sulphide with no Eh data given.on rainoT spe. 1967. Data must he given as pairs of numbers with the species number (IZ) followed by the concentration (G10. The . -log(POj) and -log(PCHn. s c t a . U. Ingri. (sj Subroutine PRINT in WATSPEG only prints out a selection of the data generated by the program. i rerun of WATSPEC with total sulphide omitted frora the minor species cards would be needed to calculate the sulphide specia.E.. theii pe. D.) Input of total iron with no Eh data given. very low pH and large positive alkalinity) then WATSPEC will generally discover •this at the first iteration by obtaining a negative value forraHCbl. WATSPEC performs these calculations if they are losslble and prints out the results. M . Calculation of the-equilibrium cdmposition of systems: of niany constituents. Speciesra.S. In such a case there.The species numbers and the appropriate molal reference formulae are: 8.pedes card. In the event of this extremely unlikely occurrence WATSPEC signals the fact. 7$0. Geochimica et cosmochimica Acta. I. WATSPEC uses the minor species data. 17. and if any coraponent of a mineral has zero concentration. 1974. nor left blank.. Y.L. Sl. but does so independently f the main speciation calculations because this particular redox air is rarely in equil ibrium. then WATSPEC will print out a warning and change density to 1. Fe.g.4). & Wamquist B.) will be printed out as no way to calculate the relative amounts of Ferrous (Fe**) to Ferric (Fe***) species.. and the program will accept some lirly bizarre (though thermodynamically sensible) inputs such . Langmuir. a general program for calculating the ccmposition of equilibriura mixtures. Density less than 1. Journal of Physical Chemisiry. 1968. The program can easily be modified by the user" to print out other details." V . Dyrssen. For example. & Langmuir D. If density on CARD 1 is neither specified as greater than or equal to"1.given parameter. Al. The geochemistry of carbonate groundwaters in central Pennsylvania. S Thompson M. Solution-mineral equilibrium computations. If Eh data of any form are given then sul)hide species can be caIculated from sulphate and pe. & Barnes. or is unspecified. 19!*7. 26 (4) Negative bicarbonate ion molality. below '4. B.If so. Sr.ion implied by suiphat e and pe. 14. To obtain this information one need only replace the X In line D 52^ ( WRITE(6. IB. Ll.11.cies cards :akes precedence over d ata on the Eh card in calculation of spec-r ation. 301-318".70) etc. Since tot'dl sulphate should almost invariably be an input variable.redox pair could be us^ed to calculate Lwmoniura species from n itrate and pe. Kakolowicz. i£ np Eh data is given. NHi.99. NO. Finally. Talents. outputs all details! as at the SO-th iteration and then proceeds to the next sample.0. A chanical model for sea water at 25 C and one abnosphere total pressure. REFERENCES Brinkley. & Christ. Jacobson. Junericaa Journal of Science. WATSPEC automatically gives a warning and then calculates iron speciation using the arbitrarily chosen pe value of 10. thfe user may wish to know individual species activities. allure is unlikely in the 'normal' pH range of natural waters.0.) by l-rnr . The fact that this option has not been specified probably indicates an input error arid WATSPEC signals this and then proceeds to the nex-t sample. rather than molalities.. In this. WATSPEC signals a warning and theri proceeds to the next sample.S.. 1973. with up to six pairs per card. For example. L.0.. Jagner. No convergence after 50 iterations.K.. 1023-1046i 27 ) •fli-i^tfii-r: •'Itvi'AJ^HS'!'*!""!!] AQ — C li.C. Ba. Silleri. .G. ipproxiraately. HALTAFALL.Dept...*'.99 are used in the output to indicate that there was insufficient data to calculate a . 57-66. Kharaka. D. If Eh data of any forra is given ihcn the ammonia-nitrat-e. (John Wiley and Sons. Report. 10. Br. SiOj. pH S to 11). 13.e.I. IS. SOLMNEQ. D.L. S Wengelin F.5). N^T. Only direct nitrate arid/or ammonia oncentrations (on mino r species cards) are used in the main peci ation calculat ions implied concentrations are never used. HjS. Activities will be then printed instead of molalities. 1971. . 35. Computer calculation of ionic e q u i l i b r i a Arid t i t r a t i o n procedures. of Commerce. 99. 14. 12. Garrels.aybe in any order. 1952. negative alkalinity at v'ery low pH (i. This is r i a minor (hat WATSPEC does. (Harper and Row. c ase. New York).. S. 107-110. Solutions. or nitrate from tota"! immonia and pe. R. If alkalinity and pH are-^ thermodynamically inconsistent (e. although the output will still show the heading '**.** INDIVIDUAL SPECIES MOLALITIES **.

F. Geochimica e t Ccsniachimica Seta. F. Wigley. Thrailkill.M. X(i9) is the ion pair CaOH* and K(19) is the association constant for CaOH*. ig*?*. L.J. Geochimica e t Costnochimica Acta.6) is HCOj and K[6) is the assqciatiori constant for HCO'j.G. 1970. S Wigley T. it is ihvilid above 100°C. Van' Breeman.77 "(although not all values of 1 are used).1 3 . JournaJ of Hydrology. 101-107. i Parkhurst D. (Wiley-Inters.of chemical eguiiibria.H. L. Mineral dissociation constants (KMIN(I). 233-248. a ccmputer program for calculating cheraical equilibrium of natural waters. in Table Al . ies are labelled A ( I ) . W. se are converted to molalities e ariy in the program. and Z(l) is thB total calciura molality. Z(I) used for input total conceritrat ions as rai1limblar ity. 2 .. Chemical evolution of the system calcite-gypsum-water. 1976) with the following. and total concentrations ) . 191-202. (1) The Debye-HUckel constants (A arid B in equ. 40. IndustriaJ and E n g i n e e r i n g Chemistry. for K(73)).O atm OOi." "&Waltei-S I. (in press).. oceanography. The WATEQF value will be used by default if line. 1976.L. E. G e o l o g i c a l Survey. U.73. 989-995. nor are X (I) equivalents to A(73) and A ( 7 4 ) . Nurabers 6 7 . 60. ed M. G e o l o g i c a l Survey. rainerals and reactions activ are Z XCl) pondi of Ca iimlal is al but t list! are n is on value WATSPEG species molalities are labelled X(I) . "(2) (3) WATSPEG does not account for the pressure dependence of equili- brium c o n s t s n t s . A-H. N. X(l} is the raolaiity . The kinetics of calcite dissolution in COj.. 19. . 197S.J.S. T. Wigley. exceptions. 7. (Cajjibridge University Press). Physical .M. 1=1.. A(I) and Z(i) a re keyed together so that d A(I) refer to the same species and Z(I) is. 1965) or using an analytic expression. H. Washington. F. Sa-67. A(75) used in subrout ine PRINT where it is assigned the fixed All equilibrium constant values are calculated. L Truesdell A. The dissolution of calcite in COjsaturated solutions at 25°C and 1 atmosphere total pressure. L.0001 for B. C ) . T. S Plummer L.0003 for A and to withiri . 1977.-log(K). T.N.72. Plumrner.J.0~i. 99-115.^brel. 1973.H. and are all association constants . The fit Is accurate to within . B. 4 0 . Calculation of ionic activities in natural ivaters. For exarapl e. S49-SS. 7 6 .N.-. Sille'n. in.L. APPENDIX A Key t o species. i Morgan J. & Storey S. 1970.J.. B 45 [ PKMIN(17)= etc) in subroutine EQUCON is removed or bypassed. 1=1.J. 10.L. Geochimica et Cosznochimica A c t a .. Canadian J o u r n a l o f E a r t h S c i e n c e s . 1975. MJIthemjitioal CTcVjiogy. Reardon. A of species is g iven. U. WATSPEC calculates the temperature dependence of species and mineral equilibriura constants using either the van't Hoff equation (se. S ^k3rgan J. -loloiiiik.water systems at S-60?C and 0.S.chemistry of sea water.A. Truesdell. 29 28 ' p j trrf3"*Ti^'F"j ^HSi . S . WATEQ. 1976. A nirierical method for conputing equilibria in aqueous chemical systems. 1=1.cience. 1971. 1 = 1. (A. NAter-Resources I n v e s t i g a t i o n s . Dilculatlbn of equilibrium distributions of chemical species in aqueous solutions by me. rhe computation. 1972. 306-315. Sears. T. In prac: tice KMIN values are not actually calculated and WATSPEC only uses the logarithmic equivalents.A. 1972.J. ' S Jones B. L. IfiATEQF: a FORTRAN IV version of W^TEQ.. i irfrdori S. X(. SO. 93-104. MLxing of carbonate waters. 468-476• Wigley. The activity coefficient of HiCOI is that calculated by Wigley and Plummer (1976).. Geochlmica e t Cosuhchiiidca A c t a . Wigley. The cheraical reactions for all equilibrium eohstants are given iri Table A l .iF.Tns of monotone St\^ue]ltCS.M. Journal" o f R e s e a r c h .S. Jr. a computer program for calaalating chemical equilibria of natural waters. Jones.40) are determined for the gi-^fen temperature in subrbutine EQUCON. Envirohrae/itaJ s c i e n c e and Technology 6.74. & Langmuir D. Stianm.ion pairs' a ' s a function of ionic strength.chd!iical equilibria.Canadian J o u r n a l of E a r t h S c i e n c e s . Ion pairing and water qualityraeasuranents. Carbonate cheraistry of aquifer and stream water in Kentucky. J. for the given terapera. 1961. J7. Calculation of canplex. 69 and 70 used. Ki77) is the Henry's Law constant for COi.the correstotal concentra tion.American J o u r n a l of S c i e n c e . New York}. T.. 1=1. and 2(9) is the to tal iron raolaiity. The constants are keyed to the species.(except. X(9) is the y of Fe[0H)?. (19)) are calculated using a quadratic fit to data in the range 0°C to 100°C.4. For example. Van leggeren. 196S. X ( I ) .L.e Garrels and Christ.L. 16. PKMIN(l) where pK . 27-57.L.18. Aquat^ic-chemistry.N.M.M. Plumraer. Mineral names and dissociation equations are listed in Table A 2 . The dissociation constant for gypsum is that given b-y Wigley (19>5) . Pluraraer. D .N. Activity coefficients of MgCO* and CaSO^ . 1976.tute in subroutine EQUGDN and are denoted by K ( I ) . The thermodynaraic data used in WATSPEG corae from a number of sources and are the same as those used in WATEQF (Plummer et al. I i t e i l c r y ..

'"TI-. f ->-:r-.'-T-'IS-y• --:*)''5?M1SIgJrjji"'". Ca CaCO? + HCOl = CaHCOT C a * * •^ S O r = CaSO? M g * * * OH" = MgOH* Mg*"^ •^ c o r = MgCO? 30 U •r. WATSPEC chooses to include the ion p a i r CaHCOj in i t s c a l c u l a t i o n s and uses the a p p r o p r i a t e d i s s o c i a t i o n constant for c a l c i t e . Ca -I.. M&'iks&immi- (fTlT'JIf. NUMBER 1 ?. H^SiO? Ha BO? Br" .frr'. Fe(0H)1 A l * * ^ * 40H" = Al(OH). . 3q NHJ CaOH CaCO? CaHCot CaSO" MgOH* MgCO? . 3 4 5 SPECIES Ca** Mg**' Na* K* REACTION cr HCO.>K. CO.OH = CaOH Ca** + c o .H = HCOj t6 7 8 9 10 11 12 13 14 15 16 17 18 19 t20 +21 22 +23 +24 sor Al(OH). I . -t.TABLE A l For c a l c i t e two options are a v a i l a b l e in the l i t e r a t u r e (see Jacobson and Langmuir. 1974). F e * * + SHjO = Fe(OH)? -f 3H* * e ' Sr" Ba** Li* NO.

5 H2O = 0.|I|'IJ-- v^-— •'I •y...OH"= BaOH* Li* ^ OH" = LiOH° Li * SOu LiSO^ co. H -I. KC1° Mg + HCO. 49 SO 51 MgHCOt MgSO? NaCl° NaCor NaHCO? NaSO..g * 3HjO NOl + IOH* + 8e" = NHt 1 .-- . c r Feci** + e" Feci? FeOH** Fe(OH)* Fe(OH}: FeSot 73 74 75 76 Fe** + 3 C r = Feci? + e" Fe** + HjO = FeOH** + H* + e" Fe** + 2HjO = Fe(OH)* + 2 H * + e" Fe** + AH. HC1° [ > 58 59 60 61 Sr** f OH"= SrOH* Ba** I .3q + H* NHt t .j^. 2so.Hso: HjSO? HjCOt H t .2H* s" NHj.SiO? = H j S i o r -i.!'•.-iyyfi .SO.SOl.r = FeSot + e" r - expre used in calculating equilibrium constant. Fe** .Fe*** . HjSior Al*** AlOH** Al(OH)t Also* A1{SOO.. • A1(S0.continued 52 53 54 55 56 57 i Fe"* FeClt FeOH* Fe(OH)? FeOOH" FeSO? SrOH* BaOH* . Fe f 201" = FeClj + e' Mg** * S O r = MgSO? Na*-^ c r Na Nacr Fe** t .'!. H2B0.S"" NHt = NHvSOZ Oz. = NaHCO? Na* -f s o r = NaSO.3H* Fe I . 0.9H" ( • Be = CH^.s o r = HSO.2H2O = Fe{OH)? + 2H* Fe** •¥ 2H2O = FeOOH" i .i. ] 70 7 1 72 H* OH" H2O e" H2O = H f OH sor ^ IOH* + Be" = HzSgq + 4H2O HCO3 t . = FeSOl + c o r = NaCOj Na* + HCO. = MgHCOt TABLE Al ..'wi. 'UiiriPilJ .0 = Fe(OH).. H .. A l * * * + OH.3H2O Fe -i"^•^ f j f J Fed -^-f Fe** .rnnfs nued +25 26 27 28 29 30 31 f32 33 34 +35 36 +37 +38 39 40 41 42 43 +44 45 46 47 48 ..• ill-:>"i. K* + c r = KC1° ^* + s o r = KSO: H* -f c r = HC1° Kso.l t V ^ ^ .= AlOH** A l * * * + 20H" = Al(OH)* A l * * * -1. • +62 63 +64 65 66 67 +68 69 HS" HjSaq = H" f HS" HS" = H* i . + denotes analytic analy tic expression 32 33 TT: "•i'.s o r = AlSot A l * * * .HCOj = H2CO3 H3BO? = H* + HjBO. + 4H* + e" Fe** -1.HiO = FeOH* + H* Fe** I . LiOH° Li so.s o r = NH^SOi. .">j» '.S — : • --I. r .TABLE Al . =!.25 Oj^aq < • H* + e" 2M^ • < • s o r = H.. e " .SO? H^SiO? = H3SI0: + H* H.

.2H. 2 « ..i 5 6 7 8 910 11 JZ 13 14 15 15 17 18 NajSO^ = 2Na* ^ S O " Mg(OH)z> = Mg** + 20H' A1(0H)j = A l * * * -1 30H' SiOi -i.O. 3 A l ( 0 H ) . A l ( Q H ) . 4H.SOi. ( 0 « ) .SiO°. ) . ) . 3 K .-+10HzO FeS + H* = Fe** * HS' CaM9.= 3Fe** .(0H).0s(OH).0. ^ Phlogopite Adularia (K-Feldspar) Muscovite (K-Mica) CaAl2Si20. . .2H Mg. = Hg** + ZOl' fJaCl = Na* + C l " SrCOj =-Sr*^ •^ COr SrSO^ = Sr** -i S07 BaSO^ = Ba** -i.2HiO HgCOj-SHiO = Mg** + C O " + SHjO 19 20 21 22 23 24 25 26 27 •28 29 Hirabilite Mackinawite Huntite . 0 .SiO?^2H Pyrophyllite „ g .1DH20=2Na*. S i 0 . 0 . KAlSi30. .2HiO = H„Si0?' •.+ 3H20 Hydromagriesite Magnetite Fe. + 1 0 M .. 2 A H 0 H ) .0i. S i 0 ." BaCOj = Ba** -t.COl" CaSOi.(0H). ( C 0 . * ^0H20 = 3Hg** .2 C 0 r CaS0. .Mg** -1. B 0 .0? CaH9(C03)2 = Ca** -i. + ^ IMn** + AKOH).8H20 = H a * .2Al(OK). 3 K . = Ca** + SOr MgCO. ^ i ^ Anorthite TaU Kaolinite Halloysite Chlorite A l 2 S i . 2 H 3d 31 32 33 34 35 35 . a O H ) .-2HjO = Ca** * SOZ" -1. + 3 H . 3 H 2 0 = 4 .*8H2O = K * . FeO(OH) + H2O = Fe*** + 30H AIO(OH) + H2O = A l * * * + 30H Fe^Oa ^ 6H* + 26" = 2Fe*^ ^ 3H2O NaAlSi30. Calcite Aragoni t e Anhydrite Magnesite Halite Strohtianite Celestite Barite Witherite Siderite Thenardite Brucite Gibbsite Chalcedony Quartz Dolomite Gypsum Nesquehofiite 2 3 4 .. a A n C . ..7K2O.(C0O.AUSi30.'?.c o r ..4H..2. 4H2O H g 3 S i . A l ( 0 H ) .Si0?.H^ .TABLE A2NUMBER ' 1 MINERAL REACTION CaCO 3 = Ca** -»• COl" CaCOj = Ca** -». . K 8 H n 2e. SOH^ AUSi.Vl2H20 = K * .-0H20 = 5 M g * * .3 M 9 * ' * C a * * * 4 C O . * 1 2 M = 2Al(OH).(0H).SiO? .(0H). 3 K . 3 C q r * ^ 0 H . S i D ? KMg3AlSi. Goethite Boehmi t e Hematite Albite Na2S0. .SO. S i O ? KAV3Sl30i.2H.WH2O=-2Al(0H). ( 0 H ) .FeC03 Fe'^ + CO. < • 3H-S-1O" + 60H =K +3^9 * AHOH).Al2Si. . g ..SiOi + 2H2O = H^Si.+eH20 = C a * * .. .

0 E 4.. b.) C OJ J3 0 I- u 0. ( l/l «. OJ crt X «-i+ + <u Li. A in equation (19) coefficient for calculation of K(I2) coefficient for calculation of K(I2) coefficient for calculation of K(I2) extended Debye-Hiickel parameter. — . CVJ + 3: •*• C/l m •n t/1 + l/> at u. ro II — . Debye-HUckel constant. + rt -7- X •f C M X 0 e 0 + •t•t- QJ U. » o» 0 + Glossary of symbols used in WATSPEC A(I) A(7S) AA A1(I) A2{1) A3(I) D(I) BB C CATXS CHG(I) CHH(I) CTOT DELH(I) DENS E ELIMIT ERROR EZ F(I) FLAGl FLAG2 FLAG3 G GFW(I) GMH20 GMSOLN ITER 11(1) 12(1) ISl(I) ISZ (I) IS3(I) activity of I-th species (I ^ 75) 1. i.r X 10 c: OJ OJ X 0 r*. in equation (19) Debye-HUckel constant.O OJ N + •t- : > ^ 0 II + +18 •f CJ II ro OJ -0 > > 0 0 M 1 CO e>j x* 1 1/1 X 0 CM •a c ta IA OJ •rk.^ . < CTl C M 1^ E **: 0 Irt >D 01 u.^ O) 1 fO 0 g E 3 C < 10 Z r^ ro CO ro ro 0 1 - 36 . >. 01 +J OJ *J C3I 1*- OJ 4J L.< * Z:. B in equation (19) temporary variable label excess cation charge charge on species I charge on species I total inorganic carbon molality enthalpy values for calculating K(I) ysing van't Hoff equation solution density (Kg/litre or gm/cc) electron activity (=-A(74)) test criterion for convergence charge balance error (equation (34)) Eh of standard Zobell's solution activity coefficient of I-th species flag for signalling redox input flag for signalling minor species input flag for type of redox input temporary variable label gm. >s m •^ in d tf) N -o 0 s + f3 .J 3 c 0 1 1 CU CM OJ CVI •I- .7 X 0 OJ CO 3 0 Il- • ^ +CT» s KO CM ^ ~ «I E ro ro c<J . II 0 0Irt f. OJ I_ CJJ . •r4.i' ' (O 0 f0 1. used in subroutine PRINT . « . formula weight of I-th species mass of water in one litre of solution (in gms) mass of solution in one litre of solution (in gms) counter for number of iterations K(I1) are calculated using van't Hoff equation K(12) are calculated from analytic expressions species number used in calculating lAP and SI species number used in calculating lAP and SI species number used in calculating lAP and SI 37 + 1 ..O' 0 lO i n + X CNJ APPENDI'X B T3 01 3 X n .. xt OJ IQ j : CO CM « « + ro •1- > 1 .

containing 9 in columns 1 and 2. is required to indicate an end to the data input. Eh data and nine minor species. calcium. a typical execution time for a single sample is 3 seconds. temperature and four of the major species (viz. Only two input cards are required. and this accuracy is certainly adequate on ICL 1900 series computers (which have 11 figure accuracy for real arthimetic). ELIMIT.lOl(I) 102(1) K(I) K(77) KMIN(I) NDATA NX(I) N1(I) N2(I) N3(I) N/|(I) PE PH PK(I) PK2S(I) PKH02 PKMIN(I) PPCH4 PPC02 PP02 R(I) SI(I) SUMAN SUMCAT T TDS TEMP TEST TKEEP U X(I) XTOT X13 XIS Y Z(I) ZZZ Z17 OlIU Ol species number for determining order of output mineral number for determining order of output association constant for I-th species Henry's Law constant for CO2 dissociation constant for I-th mineral number of input data sets analysed minor species number for input coefficient used in calculating lAP and coefficient used in calculating lAP and coefficient used in calculating lAP and coefficient used in calculating lAP and •pe (£ -log(E)) pH -log(K(r)) PK(I) or PKMIN(I) at 2S°C -log (Henry's Law constant for O2) -log(KMIN(I))I -Iog(CH* partial pressure in atmospheres) -log(C02 partial pressure in atmospheres) -log(02 partial pressure in atmospheres) effective ion radius. On this computer.2 seconds per sample (using the G compiler). and cards 6 and 7 specify data for sample 2. a. The sample itself has a chemical composition similar to that of seawater (although the Eh value is unrealistic). Storage space required is approximately 19K on the ICL 1903T and SOK on the 370/165. which is relatively small and slow. All test cases run on an IBM 370/165 in single precision did. density. WATSPEC automatically assigns a value 1. The first card is a blank card which te] the program that there are no Eh or minor species data. SI SI SI SI CTOT value stored from previous iteration temperature of previous input data set ionic strength molality of I-th species total solution molality NO3 molality implied by Eh and NHi. Compilation takes approximately 40 seconds so that the total execution time for ten samples would be about 70 seconds. magnesium.0 for the density in this case. The input cards are shown over. WATSPEC is written in single precision. and execution time is about 0. The second sample only contains information on pH. used in subroutine PECALC 38 39 -Xn—ifT KIT" . Note that the last equation (19) saturation index for mineral I total anion charge total cation charge temperature in °K total dissolved solids (gms per litre of solution) teraperature in "C (not used) APPENDIX C Program listing and input and output examples WATSPEC has been developed on an ICL 1903T computer. On IBM 360 and 370 series computers (which have 7 figure accuracy in single precision) double precision may be necessary. and thai density is unknown. The first is a water sample with a very detailed set of chemical data including all major species. WATSPEC has also been run isuccessfully on an IBM 370/16S computer. Compilation time on this machine is approximately S seconds. The program listing given over is followed by output for two examples. is reduced. give identical results to those obtained on the ICL 1903T. Cards I to 5 specify data for sample 1. NHJ molality implied by Eh and NOJ temporary variable label total molarity or molality keyed to X(I) Eh input variable value total sulphide as specified on Eh card. sulphate and alkalinity] This is an example which is close to the minimum amount of data needed to specify the character of a water sample. but this may not be the case if the convergence test criterion. in fact. Five input cards are required.


LT. GMSOLN-IOOO.l).EQ.AND.10. (DENS.Q.I-l..0.99 ' CTOT-1000.GE.r .9))GO TO 130 REAO(5. E-lQ.GT.ANO.)) WRITE(5.-f(X(23)-»X(24)->X(25)-»X(26))/X(2)) ^F(X(3). REDOX EQUILIBRIUM IS NOT ASSUMED C " •• FOR THESE SPECIES.LT.ANO.72 60 U-U->.NE.GT."rtjjj. Z ( N X ( I ) ) ) .(DENS.150) X18 IF((FLAG1.'X(47)+2. 1-1. 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 S3 A A A A A A A A A A A 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 A A A A A A A A A 80 81 82 83 84 35 l(F(40)'A(72))4K(41)/F(41))-fA(7)'(K(42)/F(42)-»K(43)'A(7)/F(43))) C-A(39)'A(72) A(8)-K(8)*C'A(72)"3 A(40)-K(40)*C A(41)-K{41)*C*A(72) A(42)-K(42)*A(39)*A(7) A(43)-K(43)'A(39)'A(7)'A(7) IF((FLAGl. lF(X(l).ANO./F(^4)•^(K(37)/F(37)-^K(38)/F(3e))/A(71)) A(37)-K(37)«A(14)/A(71) A(3B)-K(38)'A(14)/(A(71)«A(71)) • BORATE SPECIES.•X(63)•^X(64) l-i-X(72)+X(19)-»X(23)+X(58)->X(59)*X(60)-X(71)-X(33)-X(35)-2./F(15)-t^K(44)/(F(44)«A(71))) A(44)-K(44)'A(15)/A(71) • ALUMINIUM SPECIES.0.65 X(I->7)-A(I-f7)/F(I+7) 100 •• KEY SPECIES.. IF(Z(18). CH4 AND SULPHIDE SPECIES.71) A(I)-F(I)«X(I) 70 XTOT-XTOT+X(I) C-TEMP'TEMP F(68)-10.? / w c / T P T y .l). + (X(60)-i-X{61))/X(l2)) C-X(37)-»2.Q.6) IF{FLAG2.5*. X(6).0004'C)) 1/T) '• ACTIVITIES OF H20.1.O.140) X13 IF((FLAG1.(FLAGl.-f{X(21)-tX(25)-<-X(29)->2.ZZZ.'-(U'(33.AND.250)((NX(I). A(15)-Z(15)/(l. OH.-»X(58)/X(10)) IF(X(ll).K(47)"Y'(A(5)"3) A(48)-K(48)*Y/C A(49)-K(49)'Y/G A(50).).109'TEMP-k.0.)) WRITE(6.1)C-10.E0.(FLAG2.GT.A(6)/(K(6)'A{71)) '• TEST FOR CONVERGENCE.-^(X(^2)-^X(^5)-•X(30)-^X(32)-^X(35)-^X(36) l-fX(42)-i-2.99 i PE-99.0)G0 TO 120 ITER-0 50 IFdTER.E0.1)CALL PECALC • SILICATE SPECIES.230) TEMP.Q.)X(12).250) ( ( N X ( I ) .NE. 91 < 92 93 94 95 A 96 A 97 A 98 A 99 AlOO AlOl A102 A103 A104 A105 A106 A1Q7 A108 A109 Alio Alll A112 A113 A114 A115 A116 A117 I A118 A119 A120 A121 A122 A123 A124 -.GT.GE.)X(10)-2(10)/(1.(2(13).«*(-PH) X(71)-A(71) X(72)-K(73)/A(71) A(73)-l.)X{l)-Z(l)/(l.GT.180) .ELIMIT)GO TO 110 ITER-ITER-H '• CALCULATION OF P£.-f2(l)'K(21)-^Z(2)'K(25)*-Z(3)'K(29)-^ 12.E0.'X{43)+X(51)-*X(57)->X(51)*X(65))/X(7)) IF{X(10).GT.NE.1*U) IF(CHG(I).LT.(Z(9).21 90 A(IPAIR(I))-K(IPAIR(I))'A(IH1{I))«A(IR2(I)) A(36)-K{36)*A(71)-A(71)*A(7) DO 100 1-1. A(73)-1..GT.(Z(9). DO 60 1-1.)X(4)-2(4)/(l.0)F(I)-10.ACTIVITY.2)REA0(5.6) l l f R i r £ C 6 .GE.15 IF(ABS(TEMP-TKEEP). IF(FLAG1.GT.1)REA0(5. TIT"' 0 9 .-^(X(27)-^X(^8)-^X(29)+X(30))/X(3)) IF(X(4).AND.(FLAGl. 02. U-0.).0 T-TEMP+273.l).y':w«" .-<-X(59)/X(ll)) IF(X(12).'OENS 00 30 1-1.0.GT.'PE*3.-»(X{19)+X{20)-fX(21)-fX(22))/X{l)) IF(X(2).(2(1).'AL0G1Q(A(73))) IF((Z(13).0.)) WRITE{6.1)REA0(5.•X(41)-^4.'{X(20)-fX(24)fX(28)-»X(34))) 1/X(6)) GO TO 50 110 IF((FLAG1.EZ IF(FLAG2.DENS 1F((FLAG1.5Q) WRITE(6.NE.Q. GWH20-GMSOLN-TOS DO 40 1-1.GT.AND. QUDj I. DO 70 1-1. NDATA-NOATA+1 READ(5.O.^ .v rjJiJji JjT .SO) GO TO 120 C " •• CALCN OF IONIC STRENGTH AND ACTIVITY COEFFICIENTS.O.) A{18)-Z(18)/(1.0.GT.E0.-fBB'S0RT(U)'R(I))->U'B(I)) IF(I.O.O.-nr-rwt-tM^titr-.)) GO TO 80 •• IRON SPECIES. I'S-.GT.1)CALL EQUCON TKEEP-TEMP PPCH4-99.GT.9). •X(63)*X(72)-fX(60)-X(71) C-X(37)+2.240)FLAG3.D.O.GT.0)F(I).GT.017*XT0T A(72)-K(73)'A(73)/A(71) X(72)-A(72)/F(72) A(34).'X(36) X(6)-(Z(6)-C)/(l.AND. TEST-CTOT CT0T-X(6)•X(2O)^-X(21)-^X(^4)-^X(25)+X(28)•X(^9)-^X(34)-^X(68) IF((ABS(TEST-CTOT)/CTOT).GT.-. '•• FIRST ESTIMATE OF HC03.(2(18)."(PK(76)*10.EO.O..)) PE-IO.Z(NX(I))).A(52)/E A(9)-K(9)'Y/(C'G) A(45)-K(45)*Y A(46)-K(46)*Y'A(5) A(47).18 30 TDS-TOS*Z(I)•GFW(I)/1000. AND.7) IF(FLAG1.(2(6)-tX(71)-X(72))/(l. A(I)-0.D) GO TO 120 A(71)-10./F(18)+K(64)/(F(64)'A(71))-»K(65) 1/(F(65)-A{7))) IF(FLAG1. K ^ .E0.AND.015*TE'MP-.->(X(27)-fX(31)->X(33)->X(46)-»3.•(l--•Z(l)•K(20)-^Z(2)•K(24)•Z(3)•K(2B))/(K(6)•A(7^))) IF{X(6).NE.(FLAGl."(-C/(l.•X(38)•^X(44)-^X(40)-^^.E0.18 Z(I)-Z(I)/CMH20 X(I)-Z(I) IF((Z(17).)X(5)-Z(5)/(l.)X(2)-Z(2)/(l.).72 C-SORT{U)'CHG(I)'CHG(I)-AA IF(CHG(I).1))X18-A(13)/(C*F(18)) IF((Z(18).AND.160) CALL PRINT GO TO IQ IF(ITER. 1 7 0 ) NDATA IF( (DENS.AND 0 0 3 .GT.GT.1.'PH*8.))WRITE(6.NE.EQ.K(50)'Y/(G*G) •Y'A(7) A(51)-K(51)' A(53)-K(53)*Y*A(5)*A(5) A(54)-K(54)*A(52)/C A(5S)-K(55)*A{52)/G A(56)-K(56)'A(52)/(G*A(71)) 80 A(57)-K(57)«A(52)'A(7) C " •• NITRATE AND AMMONIUM SPECIES.). TDS-0.)A{64)-K(64)*A(18)/A(71) IF(A(18).GT.20 X(I)-0.PH.FUG2.I-1.'-(-PE) A(74)-E C-A(71)/A(73) G-C'C A(52)-E'Z(9)/(E/Ft52)->K{9)/(F(9)'C'G)-»K(45)/F(45)-»K(46)«A(5)/F(46) 1->K(47)-A(5)*'3/F(47)-K(48)/(F(48)'C)*K(49)/(F(49)'G)*K(50)/(F(50)' 2G*G)-»K(51)'A(7)/F(51)->K(53)'A(5)*A(5)/F(53)-t-K(54)'E/(F(54)'C) 3•K{55)•E/(F(55)•G)-^K(56)•E/(F(56)•G•A(71))-^K(57)•E•A(7)/F(57)) Y.5-.O.E0.GT.0.«'(0.5'X(I)'CHG(I)*CHG(I) XTOT-0.210)DENS IF((DENS.0.220) FLAGl.)A(65)-»C(65)'A(18)/A(7) •• ALL OTHER ION PAIRS.Z(12)/(l.)X(^)-Z(3)/{l.l).0014'C-U*(1.).0.1).1))X13-A(18)'C/F(13) IF(A(18).Q.GT.E0. DO 90 1-1.LT.(Z(9).)X{ll)-Z(ll)/(l. F(I)-1.)X{7)-Z(7)/C1. Afl4)-2(14)/(l.• • i i .A 87 ^V 88 A 89 9(J.GE.))DENS-1.*(X(31)4-X(32))/X(4)) IF(X(5).0. •X(38)+X(44)-»X(8)->X(62)->2. . IF((FLAG1.GT.' 1X(53))/X(5)) IF(X(7).•X(8)•X(62)•^2.

NE.0184/. 49.l)) WHITE (5. 13.755.-12. 2 5 .58. 15.478. 5 ) GO TO 10 217-10.(FLAGl. 1 9 .3. 2 .. 7 .F(72).0133 2.50.92 .-.2^2. 0 5 4 .CTOT.-2 4. 5. .011055. 9 0 .71.DENS.29.2--2.5 . 2 2 .li-'. -6.9.TOS. O . OATA DELH/ -9.3*73.-8. PP02-99.43. 3 .-TEMP)/(T'l.32 .-40.-10.2'73.43..5^8. 11..1)EH-222 I F ( F L A G 3 .54.39 PKM1N(I+1)-PK25(U39)*DELH( 1*39) • (25.55E-4*TEMP'TEMP'2.FLAGl.N2(38).7.4)PE-PK(66)-.2(72).••(-222)•F(66)/{GMH20/1000.3.2^-7.))A(66)-10.PKMIN(40) C-AL0G10(A(73)) IF(FLAG3. 67.-3405. 12.CTOT. 28.211. 8 .47.34.PKMIN(40) DATA 11/8. -5.11.0 .0.76 9 .*PH-AL0G10(A(6)) RE-OIRN ENO SUBROUTINE PRINT •• THIS SUBROUTINE CALCULATES SATURATION INDICES AND OTHER OUTPUTS. 9 9 5 .886.75.51.86 11.4 9 .0401'T->-3OO0.Ooq. -8.))A(17)-10./F(62)-kK(53)/(F(63)'A(71)))/A(71)) 10 10 20 SUBROUTIKE EQUCON •• THIS SUBROUTINE CALCULATES ALL EQUILIBRIUM CONSTANTS ANO •• DEBYE-HUCKEL CONSTANTS. -3. AND.4.I02(40) INTEGER GHG. -2.f .34.N3/15'0./F(17)-fK(62)^(l.0) F0RMAT(I1.5.IS4/23^75.4. 1AND. 5 .82 .3*l.38.A3/2902.I01(59). R 43 6^4 B 45.61. 2 5 .LT.41 5.45.Gl*20.115. 4 .51 41. 19. • 302366.E2.53 20.6.17.PE. 4.00557.2F10. 3 9 .U.52.7.97 .39.55.55 .3)EH-222*.918..BB. 3 2 .41. 13'4.KMIN C0I*«NA{75). 9 0 5 .48 17.3.0 .48863*TEMP'7.Nl/23*l.3.OOO101*TEMP) AA-.).2'2.-99. 12/6.N4/23'Q.E0.'PH-PK(74)) IF(FLAG3.154(38). PH AND PE IMPLY NO3. 52'-l.-. 4 .57.. 4 .DENS.7 18. 3 9 .ANO. E 0 .GT. 3642378) PKMIN(l) —13.582.12(15).F10.2 -0.3/ OATA 1 0 1 / 7 2 .34. 8 9 5 . 26.572.AA.0 .2*72.It .429+. 7 .34.2.^C+PK(75)-i-9. 42.0.FLAG2.99)SI(26)-SI(26)-f3.ZZZ.TEMP.U.26.Q") F0RMAT(1H1. 5 .032786.0.0.0. 2 3. ..11 21.4114. 2 ) E H .473.43 37. 20.PH.-2835.28.33.(G.-29 .557.-10.6. ."G IF(2(17).0051295.59.-27. .99)SI(28)-SI(28)-f2.FLAG2.4 -2.268. G T .) IF(FLAG3. -5. .173.X{72).U. .2 2 2 . 4 5 .2'62/.5 7 .58.O. 3 1 .95.f .0 20.PK(77).l. 13.-5. 2 * 1 . 3 0 .. 3 4 . 2 .71.68 .6535-TEMP'(. 245.99 SUMCAT-0.2385. 2 6 .-8. . 1 1 .1573.0-) F0RMAT{2I1.1 0 33. 29.NE. -1.CHG(72).0 25.012078.3'74/.CHG(72).29 .565.0 .02379.71.:' .65.15 . 8 3 2 . 5 2 3 . 00 10 1-1.61. 1.19. 54.E0.615./F(62)-i-K(63)/(F(63)' 1A(71)))/A(71)) A(62)-K(62)'A(17)/A{71) A(63)-K(63)^A(62)/A(71) G-4.0. 0 5 5 -2.1 .644.MOL lALITY .458.3.E0.48E-4*TEMP'TEMP'3.64 .-2.28.22Z.35.894. O .631.8435. 10.FLAG3 REAL K.4.12. 02.0 .319.9.-18.15 PK(I2(I))—Al(I)-A2(I)'r-A3(I)/T IF(I. O .238.CHG(72).987.24 -1. 9 1 1 .: VALUE CHAN IGEO TO 1.2.001225.) A(17)-217/(l. 1.KMIN(40).40.2.2.K(77).3.-3279.AA. 32.A2/.14.152(38).-8.2. .18.769. .153(38). -2. . 2 . 2 ' 1 0 . 344.'2.2.GT.KMIN C0I*»0NA(75).6059.57. 4 . 3 6 1 .'AL0G10{A(2))f6.68.543t.077 8.8.PE. 17. 5 2 .47.-75. 1 4 1 . POSSIBLE PH QR ALKALINIT IY ERROR-) F0RMAT(1X.26l' -4. 13. 54.07 .FLAG2.TDS.106. PH AND PE IMPLY NH4-1 MOL lALITY ••.F6.925. 63.2.6322.373/ . 10.0183412.-TEMP)/(T'l. SUMAN-0.4/.GT.32415+TEMP'1. 3 . 18.37.-2.95 .)G--PK(74)*^AL0G10(A(7))-10.1.565.3^1. 12. 2 .33.62.938. 2 1 ..76/.75 .215. • SAMPLE REJECTED : HC03 < 0..29812E-05-T'T IF(I.2. SAMPLE NUM8ERM4) FORMAT(//.N2/lQ'1.(A(IS4(I)).73/ DO 10 1-1. 1.GT.'AL0G10(A(72)) IF(SI(28). 7 5 9 .13.2 -0.602. 33. 7 .-TEMP)-E2 IF(FLAG3. N E .. ) . .7.12. 9.T.-l. .TOS.53.18. 3 12.4. 15.32 10.2-34. B 46B 47 B 48 B 49 B 50 C C 1 SUBROUTINE PECALC C " " ' THIS SUBROUTINE CALCULATES PE.-10. 2.4) FORMAT(//.£11.2^72.1 1.44 .TEMP.*C IF((A(7).55. 2 3 .40.N3(38).373/ DATAA1/-6.-TEMP) IF{FLAG3.GMH20.85 .' N03 ANO PE OATA GIVEN : N03.15'73.3^3./T PKMIN(17)-4.40.48. 2 7 .40 20 SKD—99.-14. RESULTS AT THIS S ITAGE ARE •) F0RMAT(1X. 4 0 .065927.56 1.36.44.-12.309.0. 36. 2 ' 1 1 .57 49.21..26.6..7..T.393.KMIN C0I*I0NA(75) . ) .56.498.l.FLAGl.59. 5 2 .3. 234. .*PE-4.7.30.SI(40).15.02 .2^2.'PE+3.1571.51.O..25 •0.GT.LT. . 5. A3(15) INTEGER CHC.ANO.2 .-.-14.99 00 20 1 .DELH(78).3'52/.9.8SE-6 88-.63.77 3 .190) IF((Z17.64 -0.2 4 .82 . 14.FLAGl. F ( 7 2 ) .ANO. 18.PH.47 4.- 30.£Z.998.IS2/2*34.29 .41 .19.-673.8.AA. . 3 8 I F ( ( A ( I S K D ) . 1.5*1.3505.G10.60.K(77). 34.KM1N(40).72. 5 2 .'"|UT. INTEGER GHG.6)PK(I2(I))-PK(I2(I))-2.15.PK25{78).76) K(I1(I))-10. 30.N1(38).71.959. 2 0 . 6 . CH4 AND SULPHIDE SPECIES..82 -32. ."G PPCH4-8.015264/.065.49. 0 0 0 8 6 * (25.991.3 8.11.ll)PK(I2(I))-PK(I2(I))+13.K(77). 18.71.E0.98422E-04'T) I F ( F L A G 3 .29.99 .10.PKM1N(40) 2 3 c C 4 C 5 C 5 C 7 C 8 9 c 10 c C 11 12 c C 13 C 14 15 c C 16 C 17 18 c G 19 C 20 c 21 22 c C 23 c 24 25 c G 26 c 27 0 D 0 D D D D 0 D 0 0 0 0 D D D D 0 D D 0 0 D 0 0 0 0 0 D 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 15 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 OATA I S 1 / 3 ^ 1 . 1 . 5 38.1 . •25. 2 ( 7 2 ) .37.Pe.152. 2 8 .GT.GT..0 1. ANO.169.24. . 5.41 .23.25*AL0G10(A(66)) IF(A(7).FLAG3.39.3.66/. DENSITY SPECIFIED AS'.42. 2 4 4 .19 1. .PK{77).21.4)) STOP END A153 A154 A155 A156 A157 A158 A159 A160 A15I A162 A163 A164 A155 A165 A167 A168 A169 A170 A171 A172 A173 A174 A17S A176 A177 A178 30 DO 3D 1-1.F(72). 1N4(38).4.38.53. 610^14.64. -1. 3642378) IF(I1(I).T.25.-. -0. 102/27./F(17)-fK(62)'(l.(G. G T .741]-119.2.684.4-3.2258*ALOG10(T)-fpK(73) K(I2(I)).25.17.A1(15).<>-l|ii.72 1.8.DENS.179 11.62.GT.48.-6.' WARNING.2 .) WHITE (6.39.22 .77/ DATA PK25/-33.GT.0. ( A ( I S 2 ( I ) ) .35.10.39 PIC(n(I)J-PK2S(I)-fDELH(I)'(25. 8 .i'?'-'. lPPCH4.0) FQRMAT(9F6.222.£11. 45. 1 .10.25.0024*(25. 1 5 .50.495. 6 5 . 39 1. 0 0 5 . .1 ..A2(15).37.••(-Z22)/(GMH20/1000.31.54.3'-2/. 2 4 . 6 1 .1 .66. 4 3 5 .3.EQ.30.0 .). 1 8 .46.EQ.684 JH.«'(-PK(I1(I))) DO 20 1-1.' FE TOTAL > ZERO BUT NO EH OATA CARO i PE ARBITRARILY lASSUMEO EQUAL TO 1Q. 8.FLAG3 REAL K.998.-12355.PK(77). 221.FLAGS.FLAG3.34.345.GMH20.72 3. 1 .PH.3^71/.457.33.91 .-.10.76 25.4*8.11..0 . X ( 7 2 ) .5^C+PH) IF((FLAG3.22.551.555. 3 .NE. 1.i i i l i i w i i K ! fa'Jw.50) CALL PRINT GO TO 10 CONTINUE FORMAT(//.4) A(66)-10.))SI(I)-N1(I)'AL0G1Q(A(IS1(I)))*N2(I)'AL0G1Q(A 2(IS2(I)))*N3(I)^AL0G10(A(IS3(I)))-i-N4(I)^AL0G10(A(IS4(I)))*PKMIN(I) IF(SI(26).27. 1PPCH4.3.4) PE-EH/('2.4.918. .-4/.68.:i.'PH-8.319 20.74. 149.6 11. 2 .67 .328.!tSi:iiri-( i T ' t-'ris'i-'i'?y s i R ' / i .218.50. 17. • NH3 ANO PE OATA GIVEN : NH3.113 4..2^71.3. DIMENSION 11(39).4) FORMAT(//.)A(17)-2{17)/(l.548.24. 32.3404.0) F0RMAT(6(I2. .3.31..' NOT CONVERGED AFTER 50 ITERATIONS. -3.0E-7 RETURN END .15 44.16. . 9.130 140 150 160 170 180 190 200 210 220 230 240 250 IF(X(6).0 . 1-27 .20.-.• POSSIBLE OATA INPUT ERROR : TOTAL SULPHIDE SPECIFIED 1 BUT NO EH DATA CARO') F0RMAT(1X.53 45.9/ A(75)-l.13.42.-8. 9.72.200) IF(ITER.BB.Z(72).64.86 .2. 2 9 .-99.52 68.368.«-(-PX{I2{I))) B 1 2 3 4 5 6 7 8 9 B 10 B 11 B 12 B 13 B 14 B 15 B 16 B 17 B 18 B 19 B 20 8 21 B 22 B 23 B 24 B 25 B 26 B 27 B 28 B 29 B 30 B 31 8 32 B 33 B 34 B 35 B 36 B 37 B 38 B 39 B 40 6 41 B 42 PE-(-AL0G10(A(17))-4.CTOT. 13. ( A ( I S 3 ( I ) ) .BB.-C-fALQG10(A(7))-10.20.FLAG3 REAL K. 8. DIMENSION IS1(38).05617. 13.57 .1.'AL0G10(A(71)) ^.09 .73.41. '1PPCH4. 6.l. 46.0 .'(PE-PK(66)->.013 11.33 62. 65.X(72).22. •0 .35.974.34. -1.OO4545-.E2. .82 .1 8 7 . -8. ) . 2 * 1 4 . G T .IS3/15'75.5'C-PH<-. 432.74.2^2.4)..

' O > (-> .93 MACKINAW -58.*"-! .* . ..37 l U I T E -1.2573E-02 ION BALANCE ERROR • 0.O Z i^ I IO O O lO-^ o • ».(—Ti^-.:!5^iS^^ZS^^f'f'So:a^:i^:i'^:i.w + i . ^ .^J= > .7000E-04 BTOT O. -1 > m m t . .-nwih .44.3486E-09 0..m .1 * -* > IK* '"^ ro o» O ^ ^ A w .cnx M. t^ I J --^i O) o ! .• ot O'— > o .> •w + O — • O • X ..-£.33 -5.A m .t .1453E-62 0.«.82 -3..O T i • Z -J . *m * ..4038E-08 O.1984E-04 0. . . • i i .68 -0.* < ^ Q l/l—*^-.65.2641E-18 ALS04+ FE0H4FES04-* FES040 NH4S04 HS040.lOOOE-06 FETOT 0.1155E-11 0.^ O / O O T l P .J Ol —I t n • r7 C w £ m M n I H .44 10.. »-«•. • Z * .^i-.55 1.1753E-14 AL0H4. A z -4 rn A m w ^ » < A w .1498E-03 CLNACLO KCLO 0 5S06E 00 0 2536E-02 0 4830E-04 CA++ MG-f-> NA-» K+ BAt+ SR-t-t LI* H4SI04 AL-f-MFE-i-MFE-MNH4«' H+ H2SAQ H3B030 0. > w .at) m M A M + fcrf 1^ w .5000E-01 NATOT 0. .2000E-04 NH4T0T O. ' .+• •!-• .9052E-04 0.4266E-07 O. o * •'^ n i n l n o i S < $ a s O O O O O O O O O O O O O O a o o o O O O O . . >-• S s Off) H.5672E-11 a. „ ^ ^ . o o -1 w .1013E-03 ND3. > • ». ui + \ m V m *•_ _ •-' o > ro o > .18 -0.Xk • O m h .• " • z ..l ' .v.. o . i ^ .S865E 18 INPUT OATA (HOLES PER LITRE OF SOLUTION : EQUIVALENTS PER LITRE FOR ALKALINITY) CATOT 0 lOOOE-01 WGTOT 0.6906E-10 O. ^ 1 ..2088E-06 0.19HE-02 0. I c c > w a»'-» o 35 • ' — t-l Ul >*-*!-• r* *-* (O m w r m • »-« o w f . o I £>«—<— J * .2012E-07 0..2928E-05 0.n°. O _ .r i * .8000E-03 H2STQT O.-gpb-.1779E-16 0.5573E-05 HC03CAHC03 MGHC03 NAHC03 0.550OE 00 S04TQT 0 3000E-01 ALTOT O.0.7089E-08 0. H A ro . r X.4162E-07 0.10 -1.9860E-02 0.< -H.18 -4.— niTim • r " .* -t. •..•X O o E + >.' ^ • S-ni33 O .74 0.37 -2.1 » -vJ -< -H n . • -* . M J> + . .35.3857E-03 BAOH-* SROH+ LI OHO H3S'.74GM/LITRE SOLN TOTAL INORG CARBON MOLALITY . T l 2 . t .5914E-62 HSH2B03..4 BATOT 0.— Ti-Tj .ikOM . . . 2 ..4O0OE-03 BHTOT 0. m-r-. .2833E-04 0. .. o 1 .0.89 TALC -0..9000E-0.83 PYRITE WITHERIT -0. • ..9808 INDIVIDUAL SPECIES MOLALITIES OHCA0H+ MGOH-i0.2422E-05 C03-CAC030 MGC030 NAC030.. .4 — o * . 5 ' .( T i m w z u i i : I .0.0 .I• O i-> 3 g g \ t Zl Sl lo 2 • * '.1577E-17 0.^ *-* t o o o > Ol O) > o> »-• l-l O — M ^ ^ *D IO u* o ro ot'-^ • o o +^M-.OOOEG C DENSITY .I -1 O O • ..^ . ( J _ .S T 5 : r -n-or-.2.47 HALITE 0.90 -2.1.* X H > .' ' " " . + . w C> • A—' o ro • o + -l'^ • • — » ro • •u > o ro • o u* r" • ^-o tn T» o z o « x*— o t.58 0.9964E-07 FE0H30 0.66321 TOTAL DISS SOLIDS .fo——* ..f * TII ^ w ro H* r* Tl m Tl Ct CA A W fU »-• > Tl « X < • cn z o f . f ^ x . > 1.M .7020E-04 0. • O.Z T l »-<. " : .8391E-16 H2S040 0.X *mA to > o * . p .I > w .OOOOE 00 LISQ4.• A F ..5426E-02 0. o • • -4 Hro»-«Hro»-t.> . > m o n i -t/l cn .* i n m p" X*.30 GREIGITE. 2 2 j .030GW/rc -L0G(PC02) .^„'^.* s ° .& • K' A Uf I-.• O 30 I >J • w f .M ' .- . A A . & O T ) « * .04 ALOH++ FEOH*-!FECL2+ FEOH+ NH3AQ 0.»:.OOOOE 00 ALK 0.-.7997E-16 0.^ "^ ^ ' . . i'§ 5S ?„^.• !">»-•»-•> + " (Jl "-n • ? .3151E-12 0.^ . O • Z Tl O m 33 •-••-« - 4- 7 -< O .1091E-02 O. .n • : j " ' S . .0.. <-^ • O ' ^ < .M ^ H ..52 ALBITE CALCITE GOETHITE HYOROMAG UIRABILT STRONTNT -0.0 5029E-05 HCLO 0 1417E-14 BR0 S046E-03 02A0 0 9534E-47 MIN :RAL SA TURATION I WICES ADULARIA BRUCITE GIBBSITE HUNTITE MAGMETIT SIOERITE 0.i:iPi1RZZS?'<s°'?'^°'''"'"'"'"'"'"'"'"'<-'^'^'^'^^'^'^'^^'-'^'^'-'i-ii-'i-'i-'i^ ^-'S SAMPLE NUMBER N03 AND PE DATA GIVEN : N03.41 CHALCEON * ro >j ." .^ X •-« •-I O Z E -< H 3 O O u* • > ^ ^ O l ..• (D tA Z TJ Ul M • • > • • M > O > 7^ ' ^ O ' — E —--^ > X 0 ' ^ > r^ A > fo*^ 1 O Tl I-.2500E-02LITOT 0.61 CHLORITE HALLOYST -2.. .4861E-12 0.4456E-01 0.• ^ 2 .74 0.91 BARITE 0.2341E-67 H2C03.03 .m m rn c/l m M »-• »-• s j — . »•«•»• k-* .66 -L0G(P02) .*-i>.T I T I X H * I T I r A .' • ." ' ' ' ^ ' .4 • >-•--* ~ * ' ^ 2 .3977E-03(CHARGE'M0LES) H20 ACTIVITY . P 3 .' K m > .OOOOE 00 PH . . Am + >^XmMA.1621E-01 0.w 0> CJ •HI-'. ' ' * ^ .o-n a • t> - o>-x M X o. 5 0 PE . m w 2. ^ o ^* -I • m • AO »-'<mr-o' • K^ ro + »-• A »H A w ^ ^ w w u o . X > m t mr • O ' 0 .89355-02 0.20 -L0G(PCH4) .o •-. .1531E-06 0. ~ g .13 -2.2136E-14 0.• -H Tl 7^ m • 00 Tl t-1M -w .U l « .I n « • O -1 ^ • -I»-.30 PYROPHYL -3..x . j ^ • •-"-1 tn > E ( niTi »-. X • .^ j > • .o . .29 ARAGONIT -0. — O Z CD > 2 » i 5 m 3 9 " - ^ • ' ' _ i k X O I w i m j k ^ H U D M > v - ..t.*^r-"-"""-''"'-''"">'""'-"ss? -CTl cl «S o * i ^j T I r m A t . 2 .M 7^I ro^^ • o> ro • •m ' ^4 f~ H " ^ O u * .1 .ro . o — z N. .0 » I .o Tl ro*-* •f o z • •.19 BOEHMITE DOLOMITE HEMATITE MAGNESIT QUARTZ -0.> X o > o -<*— A O X X 0 > — » > .O 3) aj 3 A o • ol • i-s.0. . X H •.00 ANORTHIT CELESTIT GYPSUM KAOLINIT PHLOGOPT THENARDT -5.i .X t. .m .87 IONIC STRENGTH . *-^ O ^ .153SE-10 0.* o •m A c o r r H m • .*-< • r O M .80 K-MICA -91.M -__ • ^ ^ S i .7295E-02 0.</>xro• tn fn •-« r . i ^ ^ C ' i " m >^ ..m J rui:-' ^ M S .I -< H ) tn r.5000E-07 SRTOT 0.* ' .• t-* Z -H A a Z ^ -HfOp.n ™ ' ^ * 0 ^ . TJ 3* fM ) > TJ T» X a s ' .I »-t • ' 3 ••• * W 9 *f .fe _.-^ • • fc-..i .^ > t-.uif^ o .TCwrniM'-n o t-* A N.~ 7 .2000E-07 N03T0T 0 5000E-05 SI02T0T 0.W Tl (n X •-^ U t O — ' • O < V I ' — • • rsi IO .25 7. -.4762E 00 0.. " P 5'-n m Q ^ .1 . _^ lf^ • » . ZI02>. Uf -H ." — I . O O O O O a o a O O O O O O O O O O O O O D O O O O O O O O O O O O O O O O O O O O O O O D O O O 't °^ -O' i i A u 't ° r 5 l^ D '^ ^U '^ 'P . igv.2490E-12 0. 5->.64 0.OOOOE 00 0.o > o o to ro J) WM I O • • I-* o + • z E ro X o • i-^w • tn » r~ o r ^ ' ^ o fr'. o — .2138E-03 S04 — CAS040 MGS040 NAS04KS040.• + • .03PERCENT CATION EXCESS • 0. Q w .-^—.• 13 : fm » - o- o •• o *11 T ip o* p •_ .• » .2776E-06 H2SI04 AL0H2-f FE0H2-> FECL30 FE0H20 0. — . OTI T) H. + J> .^ .1562E-04 0.1491E-10 0..' w > f c .I - « H .68 NESOUEHO -2.& m 33 X .9622E-08 ALS042 0 5576E-13 FECL*+ 0 3142E:16 0.56 TEMP . -.0. . O A i-i.10 CA-MONT 3.c5 .. " z 03 T) ro . — CJ ..l / O C -HfOZ m ^ » z o . mTlT)»H. ' ^ .0.9218E-09 0.6184E-05 0.0.-< C I r« . . PH AND PE IMPLY NH4* MOLALITY . Q • to .4835E 00 KTOT O.^ ^ • * • • ! « .1 .•o a 09 » Ul w» » O IO O I— • o o o w o ^ O o hj .5292E-15 0.1158E-04 0.• .> w o IQ .U l ^ .z to O '-*-' w o t/l •-• o i.--_m .OOOOE 00 0.48 -1.M ro ru *« • ^ M > U> • <--• U l • —t o o w .* X A • O O • O ?^ .m .0..I s i • .6375E-09 0..32 -0.3.o .84 -ANHYDRIT -0.2925E-03 0.76 -3.»-.lO.i 7^ ~i l-t > > o ' .IOOOE-Ol CLTOT O.2531E-08 FEOOH.218.>v. O O 00 A ^ .

125-137. No responsi bility for content rests with the Group.4.7.J 3 J5 o r> > X X > -< m z m n o X P" o > Q . Details regarding membership and activities of the group may be obtained from The Secretary. j 2 2 m H O H o o 33 rn > rn m z . fully documented with references and submitted in duplicate in doublespaced type Numerical d ata shall be in the metric system.^. in final form. Knowles G. o t J Ul o v j t^ a> » M w K5I • is is is ig ig!:.S page si ze. 1. •>'••-. b) in a journal Lester W. : i^iSaifiil STTO'iPW^V" . n mmm gggs S 66 6 • A A A I O O to S to to to o .l m O O O I i J CJ CJ I < o o O O O O O W A M O 0> M >J o ro ro (D O Uf 0> A m m m m z r oo £ e £o O O O U» 88Sj ro Tl z P Please addr ess all communications concerning publication to the editor (add ress on page 2 ) . All journal titles and titles of conferences to be given in fulJ Authors receive 25 copies free of charge. ed S. & Woodward G.1 . 1969.O M 1 m m m m O O O H • • -< If bibliography is annotated. I . Captions should be t yped separately. Illustratio ns shall be drawn in Indian ink on tracing paper.M. & Clarke A. The prediction of the distribution of dissolved oxygen in rivers. (O s J UJ go O O O O O t ^g O o o o o o O O O O iii O O O ( m m m 1 to to (S fTimmrnmnirnrTimiTi 8pr z £ . Copyright i s vested in the British Geomorphological Research Group. in: Advances i n Water P o l l u t i o n Ros-earcb.. O O O 4 O O o < O S Q I 8 88888 > O > O O O U O Q Q O O S88i 5 55o O O t^ O o > CJ Ul O O O ro to CJ CJ cn O D OD to sJ M A Ul ro t/1 X . Responsibil ity for the content of Bulletins rests with the authors. Reduction from this size will be 1 : 1. .^. Jenkins. London.. > o o o O X 1. 72.H. CPergamon) .F.I .1 H -< . 30 3] H I H H O O O -H • S U3 to O tS to IO i S i o . P P m m m S O Q 33 r S3^5o I I I tn r m i x Tl Tl Tl > X 2 C? C? C ^4 I gPgg2 !S ik S to < So . in a form suitable for publication and capable of reduction to A.NOTES TO AUTHORS Pi •ki. References shall be prepared in the following style:a) from a book Owens M. Kensington Gore. water PoiJuii-on Control.: 8 88 O il. c/o Institute of British Geographers.-i C Ul .W. the notes to go on line below citation. Water quality monitoring in the United Kingdom. -. Preferred format of original illustrations is within a fr ame 203 mm x' 159 mm (full page) or 90 mm x 159 mm (half page) .-. Manuscripts shall be typewritten. 289-298.l . 1972. m m m m m <o <5 to to tS (O to to S to 388888 8 ISP SSSg PSS 8 (O to I m m m O p o O O O A H X So O O O ro m H ftl nl nl • I I '8 8g §88 mmm 8888 to M to M to 88 8 88! £ Ol 18 0918 MEMBERSHIP of the BRITISH GEOMORPHOLOGICAL RESEARCH GROUP is open to all persons with an active interest in geomorphological research.32q ^ > to o to e e I 88S88 8SSS8888SS 5. S.

specl lo a solid phase Or to a g.-^ilions are indifferent to ibem. . 6. Lafon. niermodynamic stabilities of aqueous species.-jpccics of major aiid imporiant minor elements in naliira] waters iL^ing the cliemiea] analysis and in situ inca. One simple .Manuel Nathenson for checking the thermodynamic data. Tbe examination of reaction slates may sujjgcst tbe origin of dis. Sillen and Martell (1964). WATEQ. Washington. minerals. 138.M 360 computers.. aqueous .lour. Menlo Park. Ha.85 per paper copy. The .s. From tliis model. 2.-1 for IB.-:cribed by niass-aetion expressions and to a. M.S.-LS. J.Az-z^/C-vT-C^ # m m ' t f & f UTAH .-Apr. 1965. L. the ionic strength.:umc tbe existence of complexes whose formation is de.nls of lemperalure. p. VA 2 2 1 5 1 .33—24B SUBJ COMP WATQ WATEQ. as document No.i'-fficients of the chemical species represented by the ii|jperi:a. Tlic pro-am is u'riltcn in PI.-ipproucb for muliicomponent water i-olutions is to a. For each possible reaction of the form.. Tbe complete computer program is available from tht. A COMPUTER PROGRAM FOR CALCULATING CHEMICAL EQUILIBRIA OF NATURAL WATERS By ALFRED H. for the calculation of chctnieal equilibria in natural waters at low temperalures. 2. Gcol. 1974. 1955. and redox potential. 342. —Tlic coiiipiilcr iiroyrain. y /46.'Amendment 1.-. and minerals and the use of iteration for the toliilion of sinuil- tancous eijuations and for tbe ealculation of activity coefficients practically necessitate the use of eoinpiitcr methods. Afjknou-ilt'il^nicnls. J. Mar. Jtime for data and for corrections lo the [irogram. The tliermodynamic apjiroach bas been influenced by Carrels and Christ (1965).ed b. that may be expanded and updated by the user as additional stability data on complexes and minerals become available. D. Karaka and E. Allbougb tbe use of inorganic eijuilibriuin models for tbe processes of mineral solution and precipitation cannot jiroducc a complete description of tbcsc proces. and Denbigh (1955). and to C. {•^) -233 . This repori is an attempt to provide a general computer program. Calcnliilions of tbe states of saturation of natural waters with minerals are complicated by llic consideration of all Ibc factors wbicb affect the activity of tbe ions involved in the solution eipiililiria. JONES. an eijuilibrium model is a useful reference.-ites may be calculated from an cipniibriuni clicmical model of tbe water and from tbe stabilities of pluiscs wilb wbicb it may react. there is a mass action equalion of the form. calculates iJic equilibrium disiribulion of iiiorijanir. PB-220 464 at a cost of -$1. A. MASS-ACTION EQUILIBRIUM EQUATIONS In a mixture al TRUESDELL and BLAIR F. It can indicate wliieh processes arc imjiossible for a given water-rock syslcm and suggest wbicb processes may control water compositions and wbicli proees-es arc so iiindfrod by kinetic factors tbat tbe water compo. p.issist in tlic (irediction of tbe cbcmical effects of groimd-watcr production. Many readers find the approach familiar. G. The rcaelioii rt. National Technical Information Service. Calif.«(riict. Research U. Tbe number of possible ions. complexes. aA-ibB-cC+dD.45 per microfiche and 34.solved consliluenls aiid . Pearson. Wc are especially grateful to .TS al a cerlaiii pressure). Y. and irrigation. Sur\-cy Vol. No. tbe activities of tbe chemical S|>ecics Jiresent are related by a set of mass-action eijuilihriuin equations (Garrels and Christ.xperimental data.-i5iimc tbat the activity coefficients of simple ions and complexes can be described by cijuations depending only on tbe teni])eraliire and a function of the waler composition.. Christ. . 2.—Our tlwinks arc extended lo Ivan Banics. the stales of reaction of the water with solid and gaseous phases arc was financed in part by the Defense Advanced Research Projects Agency o f t h e Departmenl of Defense under Order 181-3. Springfield. F. (1) in which lowi-rtascd letters arc tbe stoichiomelric ti. Denbigh. p. Jr.-iiirenii-.-ll<-rp. The clicinislr)' of walcr-rock iiilcractions is ilctcrniined in part by possible reactions willi regard to tbe stales of llic water (undi'rsaliiraterl or siipersaliirntcd willi rf. whose earlier program suggested the format. plf. and gases have Ixren selected from a careful consideration of all ai-ailablc e. reebarge. and they may wish to omit the next sections in which the minimnni lhermod)-iiainic theory necessary to explain the calculations is prescnted.C.

not necessarily as printed here.Mg(. Ma.:ill ami otli..c h a n g e of r e a c t i o n i.MPLTKIt PltiKHtAM. equilibrium use of these cfjualions. be related thermocbeiciical derived f r o m c a l o r i m e t r i c inea.1 .920 18.ient ai.013 -15.4 . Wagman a n d thermoothers.-i-3H* Fe+'-fSO."-l-2CI=FfCI*-frF e . Data source 1 FeOH*' Fe+' + HjO=KcOH"-l-e+H + -13.7). Do.'ul. logA'=-2. thi. . Sources for ACf and A//f of individual ion species in solulion are all from 270-3 or 270-4.-OH*-HI* Fc*'-f3llj()=Ke(OII).ai-s.((:().ilso b e d e r i v e d from t h e s l a n d a r d free e n e r g y c h a n g e o f t h a t r e a c t i o n .'.Useful conijiilations of s l a n d a r d free e n e r g i e s of f o r m a t i o n ( a n d o t h e r Standards (Rossini and o t h e r s .=iire gase.' rj. C + d A C .115 AC f and A//f from 2 7 0 4 . (heat of reaction at 298 K). „ . . Do.29tl -3. F e * ' + ll.-CI°.1)1)1) 8. D-(aAGf.-rii.-^cFe + '-^C|-=^FeCI"•l•eFr. -. H^SiO" is from Helgeson (1969).-pre.. 19(1'/). and t b e b r a c k e t s p. a " . FeOII"^ 3 4 .. . _ . 2. l-"g A . and H-*" plus e"are 0 by dermition.319 29.lg.995 15.| 7 .).0-F. .-SO. M g " + 2 ( : o : ' CaCn^ Ca^' fC.stead o f a c t i v i t y .-int of t h e r e a c t i o n b y t h e e q u a t i o n . ill this e q u a t i o n . (logarithm of equilibrium constant at 298 K) and A//.l'l()(ll.0=FcUll'^-l-H+. t h e s u m of t b e s t a n d a r d energies of formation.'M CO. I-. and Ellis ( 1 9 6 6 ) . -. R o b i e and W a l d b a u m ( 1 9 6 8 ) a n d I b ^ c s o n ( 1 9 6 9 ) .' t.2'MI(ll.suremcnts. Fe'^'-fH.fiO Log A=-2.5C0 (l-/. KOUII.s. ACf and A//f fri. F o r e q u i l i l i r i a in'vuKiiii' low-prc5.') -H. 270-3 and 2 7 0 .M^.si„i.-. •¥ VFr + '-(-SO. I')(i9). d r o l ) s i 5 ( S w e e t i m and I'. 3 0 3 y ? n o g A' pressures i'^) 7' is the absdlulc dala exjierimental In which R is By data the the may gas constant and to m e n t a l m e a s u r e m e n t o f c o n c e n t r a t i o n s in a series o f eijiiilihrium m i x t u r e s of d i f f e r e n t lalion to infinite dilution.O.-.. concentrations may total coneentration wilb extrajioAlternatively.i .• 8 .-i. " + :>C|. \VATI-:i!S T h i s s l a n d a r d free energj. I9<.U. CALCLLATION OK CIIKMICAI.l 6 7 8 9 II) Fe(01l)i FeSO: FeCI"' FeCI* Fi-CI° FrSl)^ .-. Mineral o r .55 fj. Sr-^'. Useful compilations of e x p e r i m e n t a l l y derived equilibrium c o n s t a n t s have b e e n m a d e b y Sillf.= F. .20. . a n d activities equation. Carrel.473 9. Ba"^'.029 17. l-'>g K „ . S°. I'll. tbe same A C ° = . n i i g M i l i l r ( l 9 6 9 ) . . 1969).:iii-. T h e e q u i l i b r i u m c o n s t a n t for a r e a c t i o n m a y . and C h r i i l (196.IIJKIA.Sidrrile . Data source values arc given for the reactions as considered by the original reference.2. A + b A C p B ) .r . . 1 9 5 2 .'^TEQ ( t a b l e 1) are from a c o m j i i l a t i o n in prejiaratioii b y t h e a u t h o r s o f this p r o g r a m and iVlanucI N a t h e n s o n .. 1 9 7 0 ) .'=Fe. ACj-. Lamb and Jacques as quuti-d in L a n g m u i r ( 1 9 6 9 ) .919 1 1. respectively.m 2 7 0 .700 20. . A / / . A / / f from 2 7 0 . Doldiiiili* .'=F.o. .S8 8.2 .t refer to Wagman and others (1968) and (1969). of the jiroduets times free their s t o i c h i o m e t r i c c o e f f i c i e n l s less t h a t o f t h e r e a c t a n t s t i m e s t h e i r s t o i c h i o m e t r i c c o e f f i c i e n t s is t b e s t a n d a r d free e n e r g y c h a n g e of r e a c t i o n : AC^-cACf.i . T b e d a t a c o n t a i n e d in W.NATUKAI.-r. R and W refer to Robie and Waldbaum (1968). ACf rri. unless otherwise noted.ll.. A//j^.52 From A// a n d A. ^ ( . A / / f f r o m K ami W.~3 a c t i o n o r eipiilibriun: e o n .and I. i / Z p -«.i sobiliuii may be u s e d in cipiilibria.livities.-r..SO: F e C O j . . (3) e q u a t i o n s will be c o n s i d e r e d in a l a t e r s e c t i o n . The effect of t e m p e r a t u r e and pressure o n mass a c t i o n be corrected calculated a c t i v i t y c o e f f i c i e n t s (see l a l e r d i s c u s s i o n ) . Ca"^'.. Tbe equilibrium constants may be derived from experiand partial for gas-aqiieou. . except that Mg'^'.~ili-.5).s r e l a t e d lo the c t p i i l i b r i u m consl.8I!6 11.2I8 32. ' t a n t .46.200 •10.7). tbe experimental of t o activities b y m e a n s temjieraturc.n a n d M a r t e l ( 1 9 6 4 ) .. i .l')ll l .--=.sprcics name Reacliun log A' Ml.- . F o r t h e r e a c t i o n given b y e q u a t i o n 1. N o single s o u r c e o f e q u i l i b r i u m c o n s t a n t s o r t h e r m o c h e m i c a l d a t a is o f s u f f i c i e n t s c o p e o r of r e c e n t e n o u g h jinblication t o include all the dala relevant to near-surface rock-water r e a c t i o n s . m a y be used in.3211 -fi. . £tC and A// values are given in calories] Reaction No.1. Ellis ( 1 9 6 7 ) a n d Helgeson ( 1 9 6 9 ) .i"^ are from Latimer (1952). H e l g e s o n .92:'.p a r t i a l p r e s s u r e o f t h e ga-.5. ."? of niagm-tite liy-. standard state of entropy. Table \.-il..3711 J3. . ( ) ( l ! i .—Reactions and thermodynamic dala [Log R'J .3. 3 7 (ll.Mg"-f(:<).4. A is t h e ma. ' 9. B a r n e s . are calculated from free e n t r i e s and enthalpies. c h e m i c a l d a t a ) have b e e n m a d e by t h e N a t i o n a l B u r e a u o f 1 9 6 8 a n d 1 9 6 9 ) a n d b y L a t i m e r ( 1 9 5 2 ) .F e " I-COi' .00(1 10. . .t 2 7 l ) .-l. Uo.

959 4. (011)^ +5H._.0l|. Log A'=2.I'O.. .Si0:=21l* + l l . ACf and A//f from Itobie and lli-iningway(l9:2).-2ll.Source ll.=n*-fll.929. S i 0 .520 3n.800 -8.240 29.<.\iliilarij i:a*'-llll'0. 0 . 1964).0 .= 29. S'-pi'ilile Talr llyilroiiiHgiii-. ANDJONKS lullli. M g * ' + 4ll.fll. ACf and A//f from R anil W.i5 Log ^'=-9.I.Mg.f 2 0 l | - -9.553 .619 12.769 •261 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 HI'O.' CaSO..Si.6(IO:.309 2. J O I I ) .s.8 (Iloslcllcr and Christ. ACf from llo.'--Mglll'0: 2. and . 0 .). I.0 .09 oblaim-d by calculation from data of Smith and Alberty (1956) in a . ACf and A//f from R and W.511.600 -9. Klll'Oi Ca.757-log RW.7I4 from log A'(7") e\prc.379.82.s 1968). Log K=2.065 -25.'-(:. H i i * .-1-2211.530 -4. 0 = C a * ' + Mg*' -f 21l.' 11.<jg-K=1. ' = T f S 0 .. 19.300 by analogy to CallPO° dala of Clniglilai.SDKI.'VlL.c<ioii(Ky-/.>:ion (Ry/. H ( ) l l .300 1.<iiililllirriiiiid\ii.650 (Bell and Ceorge.5.SiO'.Si0° + 60HCaC0.' II.H70 3.4. H 0 . .Si0°-f40HClino.MgF* CaSO: .200 12.+ 4011D-iopside C a .305 1.532 45.300 :u 35 36 37 38 39 Calll'O" IICO. ' . ACf and A//f from It and W. Mar.Sit)° .106 16.-f4H.976 26.090 90.lirnko.iiiiii-ilritii ('iiiiliiiiiril 23.976 from log KCT) e. Na*+llP0. I{riii:li(in. 30 NallPO. 1969). A//j=l.650 2.Mg+'+ 2H.(). A//^-I. + l l ^ 9.(:o° = i i L t ) . 9 3 5 from log K(T) e. ' ) " ' " ' Aapprox ^"'^ °V a s s u m i n g 7 1 1 1 .-iim Hriiiile Cliry.SiO. .6 (llostetlcr..'cpri.0^2Mg*' + 3ll. 1972).4. Log K=1.Mg(OII). ' M I .\lg*'-^OH=MgOH* H.300 (Clmglilai and ollier. .•.58 1. A//f from R and W.-. 0 + 4.762 20.637 .848 -11. ACf and A//f fnim 2703. A / / ^ = 8 .820 2.Mpi-SiJ). II ll. i\H° For.739.933 36.Ki(in (Ry-/. =7lNV. 7.426 1.585 -2.252 27.^io.0=2Mg*'-1^211. S i O . .)|^^i from log K{T) expression (Mesmer and olheriN. Mg.l l . 1967).l((:iiri.3 . 0=2Ca'^' + 5Mg*'-i-81I. + 6 l l . Do.<iO°440l|.imilar manner lo NallPO.0=Ca+'-i-SO.573 3.Mg*' * lll'O.0-5.0=Mg*'-l-ll.714 -3.739 -6. 6 I 7 .100 6. Do.«otile Aragonite . 0 .(<:0.<terite ll*r().316 . Log K'=4.l(15. Tremolite .lieiiko.'=CaS0: .675 U g A = .674 1.BO. 1953). M g S i . ACf^-l.079 60. Do.. Log ^. .TKUK.I. A//^=-10.Mg*'l4 ( : 0 .l21 -(A//|.0-25. Ix>g A=-51.20 obtained by calculation from dala of Smilh and Alberty _ (1956) by using A'e<. .309.520 -3. ACf and A//f fn. ACf and A//f from R and W. 31 32 . Ri-airlion log A' A//.MgF'' Ca*''+S0.=Ca*=-I^S0.' . Dala . 850 27. ACf and A//f from 270-3.S0.\pre.215 33 Mglll'O" . 196-4).* + lll'Oi'=MgllPO: -139.\lg*' -t-F•^. SiO.6.<letl<-r and others (1971): A//f froiii-R and W.'-l-211. Log K'=-6.-ili' .694 36. N 2011-1411.S'° = 9().\aiicollas(l968).Si0. and TiNaHPO.Na*--0.HJ0 .*=--lll'0.204 -51. A//.930 ()/j.411.SiO: + 1 4 0 H :. AW^=2.480 4. ' -21.2 I .-1-311.-K*^-A!(Oll). 1967).ii.Maiid olhir.140 3. ^HII. A//r=24 (Sillen and MarteU.-^ 2.'179 40. Si.0=3.0 KAISijD.Mg"-i-20l|Mg. 1968). 0 .870 21.224 iMi:=NII>ll* Mg..'^=2.SiO".MgOH* HjBO. 13 Miin-ral or .illl'0° ii. A//j=3.4 6()l|Mg.uiv=7NaHP0-/ ( T N a ^ T H P O .0.-liall.H0. 1963).^iiecics name II.MgSiO.87 (Sillen and Marlrll.lieiiko. Do.5 Kearlion No. Aiih>drile Gyji.=C3"'-I^CO. ACf and A//f from R and W. S i 0 ° .s 1973)..7. Log A'=2. A//^-3.' 2II* + 1 ' 0 .(OII).-lalite . 1963).-'-NalirO.iii It and W.(OH).Si. Ca. Log A'=1.I40 (Helgeson.

C O .. Al.6711.+ H . S i ( . ACf and A//f from R anil W.0=2AI(01l).(Oll)^ = K * t 3 A r ^ + 2 S O . C 0 j from A//|. + 121l.Vi. ' -0. A//f from R and W...680 60.2 1 1 . 0 .0 + 911. anil A//f of llirrnionatrite eompuli-d in tlii.0.SiO:+ -32. 1 7 6 1 " 1-2.3 6 . i i i I I I . O .701 KAl3Si. Dc i 49 50 51 52 53 54 Chlorite Alunite Gibbsite (crystalline) Doehmile .802 113. .-iilr -89..-^ .36 CO.+1211.-: ayi-rage of llelge.102 3II. S i O ° (Na.K. .N OF CHK.ll.f r . :. + l 0 l l .820 14.0 (g).355 K a o l i n i t e + 6ll*=-2AI*'+2II. 54.rill. .of Lilimer (1952) and S" of Walerfield and otliii.. .511.+4ll.iili-+ 0. A//^-1.860 Do. 0 + 8. s o . .3 l l .795 3. O .C:i*' + 2 F Cl lilDiiliiii.. 0 . ' + 6 0 l | A1(0H). No data.fJUll. 0 .470 11.. 1 .SiO! -12. Si.f r o m R and W.Si. .+311.218 (Walerfield and ollirr. SiO° -19.on (I 969) and Zen ('(972).C0. 5 H .0. 0 = N a * + AI(OII). N a .s (1969). 40 41 42 43 44 45 46 . C O . 0 .0(g)-.'^i. AC.Mg-''+2Al(OII). / . p. SiO: + 2 i r (Na. Data source NaAl?i.002 3ll.from llilge. „ ( O l l ) .K. 1965. and trniia-naliculile-.944-. l i o i i i It ami W. 25. (H. 4 and 5 in Kce.„(01l).+3.179 918 572 A C . N a A I S : .Si. . ACf and A//f fniiii II ami W.774 -3. . l ° C ( L i i i k e and Seidell.AI. .V. +ll.31 I 1.sKo. S i . Pyrophyllile. 0 . 0 + 6 I I .760 29.(SO. 2.3(i\a. 1011.0=5. .0 -1.Min (1969) for Na end iin-nibi-r.ATIO. ' +1011.MICAL F. 0 . • II. „ ( O I I ) .582 0.267 + 0.424 i^'^AI. + 5 l l ..(l96. ) > 2 i r .'AMg)„„AI.-^tiidy).s 1968).5AI. 925).027 1. . O. SiO: NaHCO.iiilih.1.0.soda in ei|iiilibriiini at 2 l .34 -32.+2II. Si.'/7Mg)* + 2.SiO:+2ir K M g .896 17. ( 0 1 1 ) .spar from R and W. O = 0.SiO° KAl.() =Na.-H8II.33 A l ( ( ) l l ) ..+3.684 •19. 0 = 2 A I ( 0 H ) .K.905 2ir ^'^'o.037: Waler- 62 63 64 65 Fluorite G.1 0 1 1 .\ N a ' l C l Tli.720 -18.-^on (1969).3.206 67. Na.^AI*'+30HA10(OH)+ll.548 -0.'=0.0=Ar'+30l|AI.830 -45.-11.3. O I I (I'169).+211.2 1 1 . Al.511.000 ACf and A//f from Latimer (1952). 9 2 1 A C f and A//(. 9 0 0 from dala in table.M.0^:Na%AI(0ll). ..HJRIA. O = 0 .Mineral or .SiO° (K. AC^ = 2 0 . +211.0=(K.. ACf and A//.-^lgo. 0 .S. CALCL'I.. 0 -. + 7 1 1 . Mordenite Nahcolite Trona .+ -49. . .+12II.ltl5(Kiliriek.. .-ii. C O .T416 -48.016 45. 1966). ACf and A//|.SiO: 55 56 57 58 59 Erionite Clinoplilolile .Na)AISi5 0 „ . 0 .563 -85.S.milr Qi„ .jAISi^O.( i \ a . 0^ 2i\a* + C 0 .l l . SiO° + 2H* Mg. + 3.25. + 2 1 1 . ' + 1 1 .-. A//^--^ 60 Natron Na.0-K*+3Al(0H).+3II.n).314 -19.5Na* +0. 1952).-.745 61 Thermoiulrile Na.Si. O . nahcolite (Li timer. Do.=Na* + llCOi N a l l C O / N a . CO. C O j . ..taken a. . Si. O = K* + 3 M g * ' + Al(OII). 4 3 5 .ililiX. No data. + 3. .3 l l . SiO: +1.373 field and othrr. 0^2:Na* + C 0 . O : ACf of K-fi-lil..5ll.from lleiiilev and other. .(1968): ACf of N a . l o g y i = 7. 6 K * -40. ( 0 1 1 ) . Do.5K* = K-feld.211* Al.530 18.2.S•()'^^2ll• Il. A C ^ = 6 5 . 0 = 2 A I ( 0 H ) . Do. . 2 l l . A I S i . 0 0.J0ll).0= 0.874 54.i* 1 .-species name AlbiliAnorthite Analrime K-mica Phlogopite Illite Kaolinite Rearlion logfC etll. 1966).Na)* + AI(OH).'^it)^ . .CaAl. .M). 0 + 8 H .K). Log A.| \ a * + AI(01l). and A / / .530 58.. .150 A C f and A / / f f r o m llelgr.67ll. + l l ..3AI(OH).33AI(OII).511. + 3II.5K*+ AI(OII).3 .„Si. A C f anil A . 47 48 Halloysite.125 -2. ACf and A//f from R and W.272 44.Mg'"+2. .C0.7 lor rearlion pliiHip. 0 + 511.+8ll.0. K ) * + AI(OII).-.N'.sniaii and Keller (1968). Phillip.'+HCO- -0. Dn. j .+4. Do.MPIJTKR PROCKA.5Na* + l l .-. O t ) .0.L T . „ ( O I I ) .).-^liiily.i \ a .-l -111.5..11.-cpar+0.„(0ll).SiO. AC^=2.-. ACf and A//f. — Hcnctiuns nml tlfrnioilyiminic t/n(o —Conliniird Ri-acliiin No. 0 .„(OH). iN'ATURAL WATKIIS Table \. + 1211. + 7 1 1 . 0 =Ca*'•t-2AI(011).+IOII. Beidellite .AlSi..0 + 711. S i O : + 6 0 1 1 ''^o. 0 + 3^3*+ CO. From dala on natron (Ihi. .

71 N'aSp. Log A'=-0:672 (Carrels and Chri. .25 (Garrels and Thoiinisoii. 106R).238 4.i30 1. Al{( from 270-3. 1068).-l£». Na 11C0°. MigA=!0. (16 Al F .-r..'=KSO.i. 1969)I. 1068). -CaOH* .235 33.(182 73 74 .135 IK.50. A//^=-3.+2ir*. Log A--5. Utg A--^().506 (Rirm-r. Af.0 FeS+H*--Fc*' +HSiic().()/0. . 0.1011 3 ) .'11 8 21l. A//f--l. I:«gA'-^7.=0.i|. .(i.33l (Lafon.76 (Hem nnd others.'-AIS0* . fonO). iSiya? Data source 237 69 NaCO. p.r90 (Sillen and Martell.and olhers.(. Al/f-308 from log A(T) e\pres~ii>n ( L i f o n anrl Truesdell. 6 5 . •1062):. .4fl -ld. .-Jt. 1964).W»J-i'/'*. Ijdg A-0.398 ' 0.-ir+C0.500 .H. -»W-.MgGO°-.GaHCOJ Ga^' + OlL^CaGH"^ 1-400 C .226 3.990 79 . .o-/r.200' 0. .-m. .847.990 (HelBeso'n. ' = MgCO° M g * ' + H C O j = MgHCO* 3.CaCO" .'-NaC0i ^^* + lICO. 1961): A / / ^ = 5 8 ( l j r o n . 1963). "03 04 H. loi.-tler..7ii 65.Na.938 .aH .370 75..614 (•••m 12. Mg*'+SO.7»0 .SO. ACj-aiid A//f from 270-3.250 A//r=-t 10. Log A'.LL ANUJONKS Tabto 1. .so. lOfil ).. AI*'+2S0-'=AI(S0.(I Roberson. A//^---6. . l > i\" -. 1962).S--H^-(I(S:11S'II**S'' HCO-. A'--7. * ' + H C O j -GiHCO* 1. 1968): i t l l . Liji.-~illll (l. 1071).720 3. .H. . A1{0H).'.550 fi. 2.from 270-3. .114 A//.100 1.l30 (Helgeson. S * .0 H.t.'=MgSd°. 1969).. . 1909).(:.50 7.020 10.3. l..238 (1 lamia and others.t 10 5. .226. S HS. .iin 370-..—---. {Hem and 20.87 Jf=-12. 1909).CO. .. . A//^=-4.260.-. .ke olli...-USF-|0II^ CH. 1961). 1967). i / /r • LOI tl irom log f\ (T) e\pre. .-m.afoil. LL-K^ A' = .370 (Lafon.2. 67 M'jckiivjwiic 68 . 1060). 8 0 .229 72 KSO. i.-. O = A10 H * ' + I I * . .75 (Hem. .86 {Ho~t.. 1073). 2..AI(0H)7+H*. No. . MgHC0*=.0l (Hem.13t) „ 1..AIK'*..iel/.911 (I. Ca*'+CO.M g * ' + C O .070 •L9I0 -ir+so-'iiso.. a llfl olln-rs.A'=-5.308 {Garrels . Jif-'l-jiirj Ji//|. Na* + C0.. J 971). A//.672 8. 1965. 0=..C0°.).„) .i9o 6. .S* +411.\iso: AI(. .26 ( C a r r e l s and Tlioint>. K: = 3 2 .-j.-IOH. b i g ^'=-1. 1968). natne 66 Minibilite.(-.' log/f -1.. A1*'+40H=AI(0II). O=Al{bl0.AJOH*' 2Na*+C0-'=Na. antl 90 91 02- . MgHCO. . «7 88 89 .md otliers.'=CaCO° 3. Al*'+3F"iAH" .268 -0.-^on.320 A l " -)-2H. 109). .t>g A' = 1. 1062). . . I3II. .og A=3.40.0 l .00 (ileni and olliers 1973). 1967). .1 7. Ar'+4F'--AIF. 3/f^ -'LiirO (l/.. .' '.847 3. ' / l l .7. ' = NaS0- -. ( m i c r o c r y s t ) + H .71.). ..^ +1 o n * l 8e". . . lAig A--=-2. Ai(bH)j log (microcryst)=Al*' +3 OH.0-2iVa" + 50. A/fj=-8.. 1969).I / I W (Izalt. Log K--Sr2 (Garri-ls and 'I'hemp.)i H*-K- 10.MgC0°+ir.0 20 (Helgesun..-<^Ji. 157 57. log 83 AIF*' 8 4 . Log A'--17.lir.Ar'+20HV. .ono 2. 58 10. 1969). 81 82 Al(OH)t AI(OH)-. .0. A//.—Ki-iirli(ins nm! tlii-rihiMlyiiuntic cffifd —Coiiiiiiued Ueattion Mineral or species. A l * ' + H . :iiiil -1//|. .200 3.. friuj) jrsl-.• . .i'¥*^' Ti{l:]l-:S1)I-..02 (H.-_ -9.1 (Hem.010 12.928.SOj: H^O: .C0° Al*^ +01 r -AlOH*' Dii. 2. 1909).-. log K=-9. -.268 (Carre l.. -AIF* AI*'+F%A1F*' A t * ' + 2 F =A1F* 7. 1068). .331 7 8 . N 3 * + S 0 .920 76 . 77 .-i.ind oilier.. 0. . = NaHCO'. AI*' + S6.7if-'ip"r\-. 1 og .OH2 frnm liig K { T ) exjire^siim (TruesdeU iind I lost l-t ler. .087.:i(M) 12.2 tll. . 18.. A/J^=3.' Reaction i\'a..i. Log K=-i. ff Illll 27IK3. K~~ 1. SO.Jil(0H)t .-200 5.'9l I 70 . 1068).'l. A//^= 1. ••- K*+SO. A'-3. -MgSOV. .998 3. .'-11011. Lo'g A'-M. Ill rill I tin II.

. . 1073). K ) .0=5C.'*^+7H..243 -3.n-||. „ Mg„ .iinurph) r:illli-r tli. Magi let ile 108^ .l I. GaMgfCf).i.'+3UIL+FSiO.. Maghemite + 3HjO = 2Fe*' + 6011-:!og.. .-uktd 111 eipiilihriiiin with Fi-(()H)j {-. Si...570 28.523 19. . .Na. AC|. .+3H.. d = F e ( O I ( ) * + 2 i r + e - . . J069).i> . .-t-2l ML IT.660 -30.). 1060) !{i.lliite FeO(QII) + H . ludrogen inogitiiuiriMiiiiite wa. Mimtmorillf.0-Ni. . . UU . J. .^•i(v'.108 (i!r.Miriefat or .+Sn.O.(i. + 8 H'*'=:3 F e * ' +'411.-t 1.0-. .K)„ „ ..!*' f ICO.<1H apalile .nile. .J2(ri. Do. 1969)Froiii A//j and A. SijOsCOH). 0 = F r * ' + 3 0 H - -41..Kitnmrilliimle. Log A and A/Zj.960).. O .K)* + 0.Magadiite 14. . Ruugli estimate from Fe'*^ + 4 0 H ^ F e ( 0 H ) .-54.89.0ii7).5(:a*' t 3HP().At(nii). Au..\j.* + r- -.220 103 . • F f 3 { P 0 . . 95 .iii In'iir. +3.68it IIO .50.0= Fe(011)* + H*. . ACf and &}l\ from Uolierjoii (1066). 1072b). +3. I Oil. ( O H ) . Foilfnier and J. 0 . () H).(l'0.p(OH)i +1.(nH}+31!. NATURAL WATKRS TaWe l.. 109 . AC j. .(.Sij „0.23H COMPLTLR I'KOGKAM.t5.SiO. llerali-nlaleil from i b l j in talile 2 of Kiltfiik (107ns) a-'i. .0..300 .29.No.Fe.-ti'r>ns. .r+ 8H.695 4. 0 .7(>n —- . log A>34:11" in 3 .i*' ^ 3I)P0:'+-H)1|Cai(l'U.6 . 2 goethite = hematite + J l . \Aiiih and JaCfjues as qimlcd in Langtimir(1969).-(. 2 l l .173 5.555 III 1 12 Green J lite Fr(On).n4H*MI. Log A=-I4. flii-lli. Hemalite .3H-'. .8-45 IK) Goe.\lg^' + (i.iite . Fe.0-11.itlli-. .Foiirehe) •+ HLILH I.-{0H}°. . Silieagel SiO. .7.-2'25 Data source (Hi apatite ^ 5 i : a " ' i 3 P ( > J ' ' O H :log A".Gi log A=-36 {Nriagu. NaSi.8211.. Hecalenlaled ffoiii data in table 2 of Kittriek {I 971 a) a. ..571 10 i . M I C A L KOUIMHit'A.0+ H* + 9H. Fe(QII)° Fe*'+2H. . - —- 117.angmiiir. . . 06 . .500 .. FeOH*' +1 !. {Alierdem) i 9'.. . . ..645 18.f e- -26. 0 .Fe*^j Al.S. . . splution (Langmuir. I...M.+6H*=2Fe*'+3H. Do.28(H.».6.Oj+6H*=2Fe*^+3H. . . .018 -4.k)* 10.58 A l(() I)). 0 . f ihiileedoriv 67. .U^ from R and W. Log A' -. . 1. Fe (0H)° = Fe {Ol-[)++Oi l-. (H.+6011Fe{OH').!iii-(1<i::^).0311.0--K* + 3Fe*'+AI(OH).). f : i L A i l O . F e * ' + 3 l l .u-l-'.(011).„. Fe.!pg A=7.0) 3 62i48() 11.6 (Doyle as qiioted in Langmuir. Vtvi. I-V.species name . ) .23Fr*' -20. . {amcirplion.0 + e . + 2 l l i Q . .inlili.i070>. A/f^ from II am! W.+211.000 107 .> ili.47. . . (aitiorplious)=Fe** . „ ( l ) H ) . . A//f froni R and W.. .Mg„.440 6. A C ( and Af/j fi-(iin R and W.alite . l o g A' = -37.615 98 .1 (Langmuir. . .'i) .U i^iaC10.300 -3.»vii. Do.\lg* = fO.tnite = 3Fe*'' + 2 P 0 .^ii. + 3 H * = l ..S^ of niagrielite !iydfotys!S'(Sweetou and Bats. . OO IOO 11)1 Ul'2 . .SiO: FeS. ( H .. .AISi. \ OF C U F .3 H * . -3H.0. F jpjtile .F e ( d H ) ° + . ACf and A//f froin Helgeson (1969). .iiiiiiing hydros gen iiioiitniorillonite was di.OH.6=3Fe*''+.. I Icm aud Cropper as qtjoled in Lniigninir (I960. log A=-4. Fc(OH). . iMaghetnite -9.0 . .-isoKed in eiiiiilihriiiiii with Fe(()ll)j (amurpli) rallier lltaii Innialite.-Ki-<i<.+2H.\a. ^.34.0-F.Si():. Gri.b -36. 3Mg*' iC. No data. + 4H.. .008 '6. SiO".3 (lirieker. .0(V!I UV05<J 25.2P0. C A l .„ Fe*'.006 20..-*' t l.05 . I -Hi* t •!< Log A 7t). ACf atid A//f from R aiid W. 23 Fe -^' + 1.). 07 . ' N a .A'--77. ..a"d 6.r * ' + 3 1 l .i. Pyrile KFe.2m) 25.4 2ll.Rowe (in Fimrnier.-565 -4. -8H. Fr. Fc-^' + . Crvi. F e * ' + 4 l l i 0 = F e ( 6 H ) ..SiO'.slol.1 tllS 20... . .306 -40.oii(f thermoifynnitiir drttti-Coiitiinied Kfaelioi . .from I l.+-2H* -20. Fe(0!l). S i O . A C J = 545 ([. . Al.iiii/.. SiOt Fe{OH)t . .S'iO^ Viiartz SiOj. olit:iined from data of H.587 -3.4. G=U.l ' 3 0 i r . 1071).0--3Fc*' + 211.)jl-+311.891 I 13 114 115 .'siO'. ' + 8H.565 106 Vivianite . A//j. Reaction Ca. + 0. 196(1)..+2[r*+"2e^-rr*'+2HS^ 85.470 3-1. 6 4.. Fe(OH)°.. .l(l!i-fiter.-[iiiitii.

P O .53 0:820 0. 1904).400 -'60. —'{(-(led'oN. 4 (^'riat^l. .Marlcll. P O . Log A=0. FeHP0° FcHPO*• H^+cr^Ha"* m .141 0. . l.and A f / f f r o m Z e n ( 1 9 7 2 ) . 0 = C a * ' + 2Al(Oll). . 1068).452 3 600 -7. log A = . L o g A = . . . 0 + 1 7 M .513 3. 1067). . 1 . A//j-:3. Fe*' +HPO-' =FeHPO° Fc*' + HP0. 1 1 0 {Sillim and Mart c i l .70 f r o m equation ( 1 ) o f Sato (1900). 127 .750 ..611 -11. • 4 H .030 -187. by nsiilg analogy belween G:il I P O .000 -11. 1968). {Sillen and M a r t e l l . 7 4 ) . ' =Hj S0° 0. . I. SiO° 142 .585 (Hanna a n d others. . 5 1 2 . . .l j ( ) ( i 2 (ITamia and others. 0 147 . (aqueous) H.L o g A''-0.0 = H .589 A/t Da 1. S i .. 1071). 7 5 0 {Sillen and .=3. Do.3 . Mg*' + l'0-'--MgPO. A C f and A / / f f r o m R and W. H * + H S O . o ( O H ) . L o g A ' = I .=.015 2.585 -1. Sttontiaiiite . 6 3 0 {Helgeson..950 -2. 0 4 2 from liig K ( T ) e x j i r e f t i o h i n I.M c l ' o . ( .4 ( N r i a g u . Mgll. A'= 1. PO.6 . Jog K . . l-'ell.l'!) 121 . A / / r " 2 .385 -1. . . . 0 + 8 H .= N i l * + 31l.512 -5.. 138 . . 123 .150 1. 132 .SijO. . and M g H I ' O " . . that is.Mineral o r u p e d c s name Read i o n i-'. ¥ ' . Lf>-g A = . ' + 2 1 1 .1 . 8 3 3 (Re rner. / N H J 128 . . L o g K adjii.390 2. .n f i i • TKU.. A/f^=3. 1971). 100 -1. .400 (Gin ight ai aiifl olherii. + cr ——. Log. .' = N H .2 . . A l . 122 . .+4H.. . . 136 . A C f and A/Zf f r o m Lalitner { 1 0 5 2 ) . / j = l . NaClKCl" I I j SO^ H j O / 0 .i v * ' + [ l . Fc*' + HPO. ' = N 3 . +811^ 0 + 2H*=2Ca*' +2A1(0H).070 -2. ' 145 Wilherile EaCO.. A l . .700 6. 0 = 2 C a * ^ + 4A!(OH).\lgl!.Mraight line i n log K ys. A / / f = l 8 . = A I * ' + 3 0 H 141 . P O . + 3H.' 143 .-s\ II*-.640 1. thai is.361 -1. " . L o g /C=-1.200 4. 1909). 7 4 ) . . . 134 .602 -1. a l ' 0 .789 -6. . Eh=0.000 18.' 146 .-+8[I. A / f f = l . . =Sr*' + C 0 .64 {Sillen and MarleO.055 39. . A C f ' a f t d A / / f f r o m R and W. A l ( O I I ) ^ {amorphous) .335 -26. = S r * ' + S 0 . I j T plot o f data i n Sillen and -Martell (196.A 3.613 -31. ' 1964). . . . b y u.610 "•""IVISO 1.* 1.4).2 7 3 . . . A C f and A//'f f r o m 270-3. C 0 ..O .SiO°Sr*' +OH-==SfGM* Ba*' + 0 i r . 5 8 9 . SrCO. f r o m 2 7 0 .1 source 1 . ! ' o : . . .'=LiS0. . 133 .(g)+H. . .iOO (Ghnglitaj ' ami ollii-r.'i. 1 9 7 2 b ) . FeS (iiri-i-ipi lale) 120 . A .C a l l .i* + Gn-^LiOH° 0. .V=FeHP0.'=FcHPO*. 8 7 .B a O H * Nll* + s b . .5113 0=0.41)0 6. l o g A = 0 . s 6 . .Hlell { 1 0 7 1 i 24 . 1971). BaOir .. A C f = .Pu. 6 (Nriagu. Celtstile SrSO.l'0. S i . „ 3. .400 3. . C a H . Laiinionlile ! 29' .S H .( ' o i i l i n l t r d Reactifm No. • 7 H . l o g A'=5.077 -31. BaSO^^Ba*' + S 0 . 1 9 7 2 b ) .1 . = H j S O i .2 6 .:. .lg A .408. . 8 3 2 obiained by f i t t i n g best .6 .400 125 .?-GaPO.. P 0 . . O . 4 5 9 + ( 2 . .736 1. 4 5 9 . SOt 12. Do. L o g A = 0 .-FeHJ'().100 L o g A'^2-7 (Nriiigii.408+ ( 2 . ..F c * * + P 0 . 1964). A C f a n d . ' 144 . A C (.. . 1072 b).^•o.^ . -. . . s Fe*'+!l. 6 4 . . 8 7 . . ..NH. S0° Ca.'t (irn/ (fieriinji/ymiitn'c i / f t f i t . l S O ' ( S i i l e n and M a r t e l l . .4nO by analog)'Avi111 G a H . 135 . M g * ' + n . 0:640 119. Ca*' + H . Log A . Ca A l ..832 126 .SiO.^led f r o m C a H . . (:a*'+PO. .100 3.042 .25 OjCaq)+lI^+eC0.sing analogy between GaHI'O" and M g H P O t . L o g / C = ..349 -9-773 -13. Slreiigilc FePO. r964).Log K = .LHUFLL AND JONKS 239 'faljle 1 . A C f and A / f f from R and W. 2 N a * + S 0 .3 .1 6 .82. . 5 1 3 .: 3 . 1964). IIO .' f i r . . N 0 . 0 . . Li*+s6. Aff.. A//r=3. .SO.1110 b y aiialogy w i t h GaPO. A / f f f r o m Zen ( 1 9 7 2 ) . 0 .licr .054 6.a f o i l and Triie. Barite. .A I { O M ) . Leotiliarrlitc Na. . C a . A C f and AY/J. . . .030 90.2 .110 -6.—— 140 . 139 . 137 ... A / / f = 4 . P r d i n i t c . ..0. A l l f f r o m R and W. = B a * ' + C 0 . Log A ^ 0 . . ..095 -11. .:-. SFUH*": 130 . . 131 . L o g K adjusted f r o m GaPO.990 10.l . 148 . Ug. -LiOir :.iSO.+ ei-^NaCf K*+cr=Ka° 2 H * + S 0 . N 0 ^ + r 0 1 i * + 8 e . ..C0.

0 . H.PO. NATURAL WATLKS Table!.iotiic cliiirgu. Wc [nust.s the ratios pf activities to con cell trations) approach unity.212 (Wollast and others.5lIjO=0. CALCULATION OF CHKMICAL EQUiLtHR/A. . . 1955.619 150 .405 •13. 0=2. The convention that activitif^s are dimensionless requires tbat single-ion activity coefficients bave dimensions of molality "\ Stiiglt>ion activities and single-ion activity coefficients cannot be defined therniodynarnically or exactly measured or calculated because measure men t of the activily (iuid therefore the cheiiiical potential) of a single-charged ion would retjuire tbe nicasdrenieiil of the finite free energy charige of the solution resulting frpm a finite change in ctincSnlration of the siu'du cltarned ion wliile the concentrations of all oilier ions o O The Debye-Hiickel theory considers the effect. Data source A C ( and A l l f from Itobie and I Icm- -4.. Fe-"'-(-H. The Debye-Hijckel theory -^ °rV"r (5) in which a-. 229) states that.d form of tin.Debye-Hiickel equation. 1968).-vt be i:. . ' + 3 I I .=FeH.=FeH.mass action expressions are written in ivhicb concentration's may: be extrapolated to infinite dilution to yield ctjiitlrbrium consiants. . .Artinite ISl 152 153 154 155 156 MgCO J •Mg{OH).0= 2Mg'^'+3ll«Sib°+4bHDiaspore AlOOH+H. on the free energy of a single ion. .s half lhe sum of the produets of the iTipbdily . 0 log. 149 Mineral or species name Nesquchonttr Reacti(in M^G0. Two in (ide Is have been uaf. the activity coefrit-ierit. (Iciisily. FeH.A 4. .s in the solulion) and C()iil.ill iiin. ulali on.0. 2Al{OH). Mg. tin: "liydr.+4n. Can be extended with adjustable parameters tp-mPre concentrated solutions. f j . ' one jiaraiiieler. -311. log K= -5.PO.r2l -26.i[i:d If.212 A//.e that in WAT ICQ for the cab. IQfifi. P.nrf. ACf. a conse que rfce of the deftnitioii of tbe standard state for idiis in solution is that all iotiic acti'vities approach i o n i c concenfratjpns and activily coefficients (defined a. tin. • if siit'i^lf. use npntliennodynamic inpdel. Log A=-37. dielectric constant. Z ts tbc. Fe*'.PO*'.980 -21. activity..H .0 + 10H. it is necessary t o ijse single-ion activities and slngle^ion activity coefficients: These are formally defined by the equation.i-ijtniiiun (Ivobiiisiin and .St(iki:s.PO.405 26. of electrical interactions ivi tb other ions by assuming thai oppositely charged ions'can be.0=Ca*'-j.-l!i where A and li arc co(iBlaut=: dcpeudtng only oil llu. .u«d tlie Bijutire of the charge of -. and m .d from eYjii:rim<:i]lal dala.»n -si/. tlnE-refpre. Uctiye-Ilne. Si. (aiiucous) 0.25O. AGf and Af/f from Hemingway atij RobiK{i972).SiO.are respectively the. The Maclnnes assiunptioh prondes inforinatipn on the behavior of single-ion activities al higher concen trations with wbicb to fit the parameters of the •extended.708 -7. a. . CaAljSf^O. .2:ii).Mg*' +Cb-'+20H-+3il.457 13.s!uti[!tii)ri.-i.ijnati(jii and tin: Maeliiiii-s a^. .ui:ffi(:ii:iil.PO*' +6' 498 33. Tliis a-ss ump tion is valid only for veiy dilute solutions.tivity c.' .-iiin ae. A C T I V I T Y COE F F t o t ENTS In the^ limit of infinite dilulion.0=Mg*' + C g .sIy cannot be made.—/C*tieti'iirL*pni/ thi-rntodynitrTiic Jtiia —Cfmtinued Heactiisii Ko. The original equation (Rpbinson and Stokes.583 - ingway (1972).'d a. a n d / is the ionic strength (defiuf. Thi. This property is usefulih experimental studies where . ACf and Af/j-from 270-3: Do. and temperaiure... available but are perhaps the most widely use^d auJ are. 0-H*+OHSepiolile (precipitate) . 0=A1*' +30HWairakite. ACf and A//f from Zen (1972).'Nlini:itf.and Af/f frotn 270x3. Tlie reader sboub! be aw.909 -17.(PH)'3U. Tbi'-si: are init the uniy iiiiidi.008 -37. .345 — 15.43 {Nriagii. . 1072b). . . The DcbyeIliickel theory provides art equation which describeslsingleiori activity coefriGieut beliavior of ions in dilute solutions and which. 0 H.s measu re men I obviou... 0 / 0 .240 COMPUTKU I'ltOGHAM.-3il.s to evaluale siiigle-ion activity coefrieienL=j.tibns used t o correct experimental deterrni nations to infinite dilulion.140 — -35. loKlr=- (0) H-DaVT and thi: eleetrical jiOlcnlial of llie pha^c are held cpristanl. of the addilional unuertninties intrndufiud by this ajijiruach.(ati)+ll* + e. . b u t the property gives'no c l u e t o activity coefficients in real solutions of finite conceniration.5H.-2H. .+H. p. and the niolalily of the itb ion. exli'udr. . considered as forming a spherical shell around the ion.O+4. In real solutions of more than a few components. gerierally consistent -with the func. nnd activity coefficienls derived from the theory deviate increasingly from experimental results as the conccnlration increases.l i-.

...5 3. . . ihc mpsl reasonable values of TSOa can be obtained from T'+Cs-SO ' "^iCsCI. 1970). -y±..:i ^ . and.Mgil. .5 H.jlipn*') and for minor ions with vahies of a frorn Kiclland (1937) and b set to zero.1* Fi--". . . density.iiion is used in WATEQ for major ions with ff and b vVdues calculated from uxperimeulal mean salt single-ion adivity coefficients (sec " T h e Maclnnes A. AlfSO.b*.. -75Q-3 = T t C5j504 "^^ KCl ..4ldy associated.'i. The Maclimes assnrripHon (Maclnnes. Hi S O . *a yalii)-. AISOJ..?rcr.s have bi. L i * .singlc solution.CaOI..5 3.).SO. —riiromelerj o/tfie Di'b. OH.ivoid spliiliuiis in which the ions are lii. A1(0H)*.T±KBr T!k and f. 7±CaGh T± KCl (13) . \ HPO. fi. in coticcTilralion of solvent in concentrated SGhitlbifs. which allows f«ir the effccl of the dccrea.^ ' r l If (10) (11) " ..^.-i-n ralciiluli. 3. i'. the T+. SrOIL*. it is necessary to know the variation of singfe-ion activity coefficients with ionic strength iri a. : . HSO. • -. 1956.' MKF-^.T>GI+-^ FeCIJ. n* TCa .'.S 4. Nll^ SO.S 211 Jug7=-il^&+W.- . if the activity coefficient pf one ion can be calculated.-. Fell.Maryott.. . 'T'±KiS04 cannol be used because of llic formalion of lhe KSOa" iun -3 pair. (15) B 5 0 2 9 1 6 X 1 0 * / ' . „ .': from Kiiliaiid {1037): 6 value.rcsuUs must be used with caution betausc Cs and c r may be weakly as.?0+. .165 . 5-'t 54 -0 .md DIIUTS. . of many salts. NH+ NO. HjSiO. Akerlof and Oslicry. CalljJ'O:. KaCQjNaSp.0 4.r.6.cfiinjiari- . BOJ. a second adjiislaldc parajnctcr. 1968) ^ ^ L 8 2 4 8 3 X 1 0 ^ / ^ ..015 -. 1936).l. 1939) that tbe single-ion aGtivity coefficients of K and Cl' are equal to each other aud to t h e mean activily coefficient of KCl allows lhis to be done.. KeOH*.. and e is the dielectric constant of •ivater (Malmberg and . 7+ T .ye-Htictelcgmition Major I'ofij' Ca+^ Mg-^' Na^K* Cr SO. Sr*'.lES)Ji--LLANDJONl".PO.a ffirlhT± KCl (14) .'^'.' wbcrc c?is-tHe density of water (Keenan and Keyes. ^ ^ Q . then p.3 .iclivily ccieffieieiit. : : . c m .f %-^v-rr'!i^-»^' 'nU..0 Ti KcrTK-*-.\ LiSO. . AIF. ''.'»».KeCUH). •.NallPO. A J { 0 H ) .-t[ji-rliiii'nlal nieiin salt .sueialed aud TSO4 values derived in.20 ..lliis way may be somewbal l o o Iilgh al high ionic slrciigths.. ^ „ . Tbe Cfjostanfe >1 and B u t e cakutated froth ibc dieleclric constant.-i: e.frKii'Mt. .0 b 0. ' . *'.5 .0 3. 0)'/^ (8) In deriving these mean sail . ^ . 1950). Ki-OH*'. Singbi-inii aeti\-ily cot. .Minor loni^ HCOi COi' .0 a 2. Wlieri. H. AIF*'." " ' ' ^tKCI ^Y " ' ^ relatibn.4.thers may be derived from il..l'0+ (12) fV.HS" • MgHCOt. 2Tabic 2. K H P O ' .. This cqu.'fl for ctniecntrali'd siuglr-jall spiiitifiiis jiy u*c of the StukesUohiii^fju riiiialiou (liuti. 5. . . In calt-nlaling ^RO"^ ' ^''^ csamjdc.'*„jGalU.04 In order to assign the adjustable paraineters in eijtialipn 7. The Maclnnes assumption a 5. Experimental values are available for the mean mptal activity coefficienls. r-J7y. CaPO. Vaiucs of a and' fi for major ions obluiriCjd from cornptiler fitting of calculalcd mean salt activity coefficients as well as values of o for rntiior ions derived from KieHand (1937) arc shown in table.NaCl TNa+" 7±KC1 Alt-"?. and tcmjierature by the eqinitions (Hamer.' inole-''^'(10% H^O)^^ (9) Even here.)!i-ii!aled rriiin i-. p 6 . K. Fe. A l " .0 3. \ . In lhis calculation.s. ..'isuinj. ." '« and b valtie. A I Q H " " .0 5. MgPO.. F i s the aljsolute t'eihperature. FeHP0+. S .aOU*.-el in /ern.'-l be careful to . 'Jni: nm.v .-iiigU-iini activity 1 iii-ffii ietits.. By definition.. Ha*'.PO|. .. UBaiiT (7) adds./ .075 -015 .

. „ . „ „ -t-„ .salt caleiilation.. Table 3.604 ." / r^ TniiH t.^i.406 .671 -668 "^SO^^ '"Si total '"H^SiO. i .604 .870 1. Tin: di.i . these deviations have proved t o be' insi gii ificant „ .150 /IS.(-.716 .6-19 .-s and otlii-r.'«^ . .297 -293 .^ . anil the activity f.586 . see Truesdelland andcomiiarisoti Jpnes (1969). lO'P" .and/«j[ i ^ j g ..214 .li(nt:ili!« parameier... further discussion of activity coeificient equa/inf. T b e u a e o f a n y i n o d e i of single-ion activily coe fficients based on cx [teri me ntal measuremenis made on single salt soliilions requires the assumjjtion that.-en From [iaramelers of Debyc-Hnckel cfju.573 .562 . „ . b. 11 11 recti011 for other weak a d d raflical.389 S .i^ .779 .. Kl • .642 . values • of single-ion actinty coefncienlo used in WATEQ arc compared Willi mean salt coefficients and vrith tho. In table 3. . . Stokes Robinson.567 . are jiO--wiM''.155 . activity coefficients in simple solutions are.591 . (19) '''Na* . ..1 0. .i .. 1972). .380 .DHJ' .265 . . Tfl^ in tliirjiriigram).654 .577 .!iat.715 "0.i lion (table 2). . liie p n .• ticienls ot the species.t c-f\-7 . r .905 W8J*.^ . equalion for total silica (silicic acid and ^„u:p."^ ^ " 4 I ^ S i 0 4 '^'^''*-''^''''^''''J''''''"*" ^''^'"''-^^ ^'^^'°''*''l''"''"''s ^o solve for m[[. . n 2 p ! l \ t^^^'^'i / / : . .563. W«J .247 . CNij.289 S .389 .^„^ ^{ ^ chemieal model (GarTcIs and Tbompion. . / T ^ LO ' \ 1" ^ T|j^SiQ4( + " \ ^'j^iO. ma.108 . b u t limited experimental work in concen tra led ( > 1 molal) mixed electrolyte sol uti uns-indicates that it Is not always true.)f-ffif'i'-nts ba... J J Q . .Mi .0 4..„ . at a given tcmjieratiire. ..stribution of anionic weak acid si.670 ..iti.247 . . llu.i i h e concentration of each species IS calculated Irom tlie ' • '. t.600 .01 0.470 . . -•685 SK' " ^ G I ' .hyiia:^.''GI"from NaCl solu. KQ UILIU It IA.654 .763 . rt i .... and other slnffle'-idn activity coefficients Ionic strength 0.770 . . Davies^.. e. and the nieasured |»H.-bicarl>oriati.692 ..'085 . .0 and '"H^SiO. . ^ -!-m. . NAT ll K A L W ATK HS SOLUTION OF MASS ACTION AND MASS BALANCE EQUATIONS Compulalion of-solution species distribution is aecomjdi.350 .mite. For ^ ' .^^e calculated by Bates and others ( t J / U ) . .620 .nls of Individual sjiecies.623 .230 .-.2 .s iitustraled by silicate cfiuilibria. a.-.674 ^ M S V .627 .te inn^^ is ^"''=-'"^*= ^°"^} '^ _ =n.708 . .pavi. . equal to those in cpinptex sobitib'ns of the same Ionic strenglli.653 ''HGOj D H ' .506 . G A LG IJ LAT 10-\ O F Gl 1 K M IG AI.: Ffinn Walki-r. 901 SH . Tlie exten Lof deviiitipn from ionic strength dejiendprice is small except for ions that differ greatly in si/eaiid hydration sucb as H and Gs .782 0.135 . .mj. ^...659 .iite. ^ ' . or analytical concentration..i.600 .„ nt-i^ 1.>is calculaied firsl from total analyzed coneeriiraHons. . TH^SiO. tbcsi: v a l u « agree reason ably with activity Ci...JL„^ -n (Pytkowici and Kester.756 .-.563 .1 .s.^ /2n V^^J ^ • • : : . .900 MS.205 . ibl. muhility :iiiil lb.386' . H4Si04=H*+Il3Si04' .376 :Mi .n\ ^.626 .560 . .. MS^ SR D H ' ." « ..606 . .ss balance cijualiuns. dislribmif. ii CjO"' =[]* +11 SiO '^ (16) (n\ son. alkalinitv has b. . D H ' ... ^Meins'dlG Corrected during tin: cliCitiicai analynw.283 .697 .. first ilis^oi'iallfiii eiiieilaiit tif carbtiuii: acid. and Joliiisiin {1027)..s {if tin.a. .562 .713 .MjierimenUiI Sfjlulibn cquilibriuni cpnslants.786 . (18) . . a „ j activity coefficif.945 . liraj-. DH' .292 . .266 :234 .570 .590 .077 . t r o m the [(receding cijiiatjons.649.661 . ff.cR!i:.^onable iti dilute solutions.003 0.763 . ' ^\ .388 .0 •2:0 3.659 509 "lio ..569 . 1962) which" uses analylieat concenlralioris. .901 1. .^. i r . i total ' -t coef- .209 . MS.r j = rl-2^i\J4 'll2iuU4 A '-—^ ^^ ™H3Si04 TH^SIO.062 . '^ '"Si l o t a r ^ H 4 S i Q 4 + ' " H 3 S i 0 4 + ' " H i S i 0 4 ' • :L /om i~^) '"Ca^^ The mass action equations can be combined with the mass l^^'-^ce equation to s 6 I v e : f o r .126 Tlie mass balance.. n^r. 1969.368 .242 CO.184 . . Yea tts and Marsh all.570 .573 610 . Tti<.5 1. l O P " A . However.708 0.-.—Sin^le-ion actltitty coefficients at 25°C from o ftwo-^arsjmeter eq'uatjcin used in WAIEQ compared with mean salt. t h e saine nielhod is..! ljie:iri. .372 .667 MS .s.i • • T* .783 . .043 j^jgl „„./. after cor- tr.'* stiqi may be ^In elilof ide solutions {Ilal. jill.900 901 . '^<' . .i ' . .. from jiII and the alkallnily dclermination. com-entration of lljCOs k Cioii ^No a.661 "'Mg*^-.toycd for [diosphatc.D H ' . 1970)..371.789 0. ..770 -773 ..-1 y.600 .d pti mean . .-. ..11.692 .773 ..6-12 . .554 . ^ .. . and siilfide sjiecies anil for the carlnin.155 . „ ^ S' i 40 4 ' .070 .229 .hea l y „„.790 ..103 .230 . . .I I -•.M P L i K K P H 0 GK AM.623 . This assumption is' n.701 . caieuiated from the r.806 SK t '' .. The activily coefficients ^'^'^ caleulatecl from tin: ionic strength by an itcn'iUve jirocj:dure. .i i. ..091 .-<:i. for models in wbicb all ion associations are con side rpd (as in WAtE{5). 0.. . ^ . 004 .904 ""COj^ . tions.577 . fni U-..822 .400 -390 .580 .038 Tk* DH' .D H ' . r . 1062).

K.22—25.. SO4 .C u l l C p 3 ^ C a ^ H C O a " ^ =(iiCG3. and the e\|^rl•^.a "H:a+'"'Ga total + . of the free anions HGOs" CO3 .ilii)ii. + ''CaGH' _^> g p H " ' " C a * ^ TCa TCaOll* (30) A". aivimi cprieeiitratioiLS are ulill-/. complex idnsi and free ions for all aiilons agree >i'ith the analytical values withiii 0.''CaS04 ''Ca+2'''S04' (29) Erom these qua tions the expressions.-. If the .iiid eqiiilibrium a-ii^ociation conslaiiLs are emjiloycd.-^laiil ufsulubilily.suggest possible origins of dissolved consli lu ents and possible controls by mineral reactions.SO^ From equations. ION RATIOS When the cbenilca! model is comjilete. 1965.-=c Willi lii'i ^aqueous .ohllion can be cllHraclerizetl by a mass aclion equation.•ilculiiliiii) pf the ciiii'.into l!ie mass action cxprcsKIOUS to solve for the concentrations of ipn jjairs. The major ion-jiairing react ioiis arc 'Cll tf. aiid l i complexes: are repeated..5 percent. acrom-: J ll i. iLs acliviiy is e.coiislanl n-duces lo lhe "tolitbilIl\ jiroditel.solid Is a pure ^llb.-lji^|^^^n fpr the eijuillbrium. ^'^I'^filf ^ ' i ^ ' i i r h : ^ ' a ^ c O a ' ^'"^SOa' L-l_i—f—-_1 -m ) WaOH* T(.SO: (32) where fC i:. the program reduces.35) ATi (28) «ca*^ '^co.LL AND JOM-'S (.s (if iun jiair. Ffir a solid of fGrmula AX which (lissplvcs lp fffrm tons A and X".— "CaOH* (26) "Ca-*-' ^oir /c.the i-fjuililiriiiin (-oit. The calculations may be illusIrated for the calcinm ipn sjiecies. H C O j '"Ca+^ TCa*=' '"CaHCOa TCaHCOt* (31) "'CaCOl Tc..tar"'('. F ~ ." (33) may be substitliled liilti llie mass bal<iii<:i: tqiKiliuii for caleiiim in jilace of the jill .it:0* C'a+^ + OH''=CaOH\ (22) (23) (24) (25) Ca*' + HCO'3 . + /^ i ° a.. lhf:so compulations in WATE(ji also include jibosphate species. 5).. In each iteration. il is useful to calculate molal eon ten tration ratios and ioii activity ratios for (ilolting on waler coniposition and mineral slability diagrams^ respeclively. ion.^ A. The concentrations assigned lo ion pairs and weak acids reduce the concentrations of the free ions and change tlic ioiiic strength and ihercfpre the activily coefficients. llic ileraiion is slopped. equilibrium epnslan. and Ca^'-l-S04"^=CaS04. TCaGOr TC. the rfiolafitics.i^)nim-'^- •nUJK.When the sums of all weak acids. p.»)nijiiil(. Then the calculations df free Ca and Ca complexes along with similar calculations for Ma. The compuled concentration of free calcium ifin.g uul li> Que bi'caiise it is In itsj^tiiii(IiinLilale_(fiarrels aiu! Cbri.COj ^'•^S04-^'rtCa^'TG/^ "'CaSO: TCa.SIJi.^.d vahies for lin. if neces. A C T I V I T Y PRODUCTS AND SOLUBILITY PRODUCTS The equilibrium nf a-r^ulid ji!ia.G1".'i. this exjiression'is K (.Coniparispii of these ratios with those of relaled waters can .-y.^.= "CaHcot (27) «Ca*^ "HCO3 _''GaC03 In actuably. and P04 and recalculates the Ignic strength and the activity coefficients. The eorrceted vaiucs arc calculated b y iteration.-j! ied by a i>roeedurc simtkir lo thai fnr weak acid i^jiecie.[s for thiv association reactions are A-. but llic analyzed ur e..a''^^i is substituted back .-tt follows: ^I-^A'^X" (36) ._jKJl a sollil ^ohilidn. Fe.a-'' ' '"CaDH^ •••"'(JaSttS ' '"fJaHCLll^ ^•^"CuCifj !43 lu obtain all (•xpn'-ssion for free (i'iie"nijile\cd){.sary. A numlier of these ratios are calculated in WATEQ. "'{-. 4 .


COMI'L'TKH PItOGRAM, CALGrLATlOX OF GII K.M IGA L LytilLlHmA; NATilKAL W..VTKKS W A T K Q and [iri:.-;enl(-.d in aildllipn lo the logarithm of the tnti.^t jirubable value for visual comparI.=iiin wilb the logarithm of the aetiWty prudiicL Beciiiise uf sjiaee liniitatiiin.^ only the most probabli; soluliility product is used lit calc id :i ling values of AP/A' lag (AlV/f), AG^, arid AG^ [.er ecjuivaleni. Rnlhat|>y values and solubility products used in tbe prograin, logether ivith tbesoiirecs of all data, are given in talde 1. EFFECTS OF TEMPERATURE AND PRESSURE In the; rclationsliijis devclojied in the previous sections, temjieratiirt; and jiressure have been .as.sntu(:d to be (;pnslant and their effect on the cijiiitibria IUL? n o t been discussed. The great majority of cxpcrimenLiI determiiialiiins of ccjiillibrium constants and of free energy values has been m.-ide at SS^G and, particularly for solution equilibria, dala at othe.r tein peratures may be entirely lacking. If experiments have been made over a wide range df temjieratures or If complete thermochemical data are available for al! sjiecies o f a reaction, then the equilibrium con.stant may be^expressed as a-ipovver function ofthe absolute temperature log K=A*BT*C/T+D log T, (39)

In hvi!rti!\>i> ri-actiims, wuti-r is con^iiti-ri-il explii-ilK, a.- jiurt fif the reaction- In tht- .solulion of ijiiart/ to form .^ilicit: acid, for r.\am[i!i-, Si"^(f,uart.)*2H,0^H4SI04, (37)

tlie water is written as part of the reautipn, and its activity ajipeara In lhe equilibrium cxjires.sion. f\ water siunjile when collected is u.sually n o longer in contaci wilh min eral phases, and these phages may nut be acces.=:Iblc to observation. It is of Interest then lo determine •with what, mineral phases the water is saturated or nearly so. The calculated activities of tbe dissolved ipns in a water may be combimnl to produce the apjiropriate a(;livjty product •which may be compared with tbe solubility efiuitllirium constant to show the degree of saturation of the water with each mi neral, con side red. This comparison may be made by means of the ratio of the activity product to the equilibrium sbhibillty product wliich Is given in the program as " A P / K " and " L O G - A P / K " and by means of tbe free energy change pf the reaction ^C, (which is zero at equilibrium) and Is given as "DELGR",in the program. These quantities are related by the expressloh A(S^=2.303/;ripg(AP/K). (38)

Sorne niineral formulas cdritain a relatively large'number of atoni.s, and the Af? values for these minerals will deviate from zero more rapidly with dilution of concentration than will tliose values; for minerals "tvitb Eimple formulas. This deviation can be illustrated b y comparing the activity product of dolpmite, .flCa*^ " M v ^ ^ ^ C O ' ^ ^^''th 'li^t of calcite, oCa*^ aCO"?If a water iiiitially'saturafed with bolh minerals Is diluted with pure water, A(>j. dolomite -ivill be twice ACj- calcite. To correct, this, ACj- values are divided" by the number of negative charges in the formula of the mineral and presented as (for want o f a better label) "PER E Q U I V " ACf. The compilation'of a con:si!;ten t s e I oPsta bility constants for minerals suffers from several mice rt ain tics. The standard enthalpy of formation and standard eutrojiy of most minera!-; have been mea.sured by calorimetric inethods, and lhe standard j free energy of formation calculateJ from these quanrities is often refereiiced to the free energies bf formalion pf the elements railitir than to tbe free cnergit'.s of fonhatiun of the ion.s formed on solulion of tbe niineral. The cd ml iin a tibn of such values with those for solulion species involving aqueous ions liiav lead lo erroneous stability constants. The use of i;N[MT!nit-nlal solubility prdducls or resulting free cticrg)' values is free rrom lhis incbn^IslehcyL The ni'ain uuccrtai nly in the USI: of thi-ii; data lies in llu: [ireci-se (lefiniliuii of reacLiiilsarid jinidneLs iitvptviid in lhe cvjiiriiin-nt and iti the difficulty of rrvi.-r.iuig tin* ilijiiibhrium. Ht-i-uu'se of tliese iinet'rtaliilles, the lugaritlimn of the m;i\imuni jlid mitiitiMitti sobihiht) jutiduels are ealcuhiluil in

in which one Gr mure ePefficIetiLs riiay be zero. Wlie'fe this type of expression was available in the- lileriitiire, it has been used In \V.-\TEQ (lable 4). If experimental dt-terminationsat otily two or three teinperatures are -available, a liiiearji dejiendence of log K with the recijirocal of the ab.-iolulej lemperature may be Indicated (that Is, B and D arc zero, in eqJ 39) wil ich is equivalent to a cpnstatil value of the enthalpy] (heat.'coiilent) changeof tbe reaction, AH, This is exjircssed by the VarrtH'gff ttJalldii. .v

lot; K^loir K ' ^ ••^'" 2.3K 2.3 K
• 2. 5 u ^ ^


Trj '

.1(40) 33,

r i C ° !"7>7 -, J .
^ . J , i , i ^ ''^


in which Tr is the reference tcmjierature (29fi.l5 K (= 25 C) in WATEf}) and ithe constanLs A and C in equation 39 are equal lo • -^ .-^- , ^ . ^ j j ^ - ^ ' : ^ f•'"i\/
_a /•

lo" K Tr


^h'f 2.3 «

res jicc livelyThe enthalpy change, of rcaiition can be (ibtaliied by determining the slojie o f a jdol of cxperlmcnLiI values of luj: K versus ( 1 / 7 ) , from labiilated values of tin- slanilarit eiillililjiy of forma tibn of lhe species iu the reac lion by using a rehiliun analugoiis lo eijuitlion 3 or from direet measuremenis. 1 bien thai py of ri'ucliiin al 25 C h:is beeii ealciilaled for iiiusl of t!ie-rqiiililiria li^.ed in W.-VTI'IO (l:ible 1), and eqiiatluii -10 i.s used to ealciihile lhe value of lhe eqnililiriiim I'uli^Liiil fur llie



i\ t - L G ^




THIIK.SIIKLL AND JUNES :iliti- 1. ,.1ii.,fvlii-ii/ i-n/ii.-«e,i!,v f.n li-A' KCl") ii,«i-i( m 11,1 7'?-.'y I T, ill Ki'lvitis]
lilrlllJIjiT MiMi-liini


K';(jri-,s.-!iin fur lijj; K{T) 6.36l!-(l.<n6346 r-34(l5.0/r ,'!'>.4 7!y).{)65927 r-123.S5.1/r J 573:21/T+28.6()50-t 0.012078: r -13,22-18 los; A'JK 0.6322.0.(ini 225 T-2!i3.'-i.76/7' 8.153-0:02194 r-2 382,3/7 5.3«n-();(12!00 7-2730,7/7' 3,106-67 3,6/r -5-35(l54i(.0l 03412 1+557.2401/7 ) 1.17:0.112386 r-3'279/r

It.;fiIlv-/,li,-nko (1067). Dn. Mcsincr, H;u-s. and ?wi-W«ir( 1972)1 ^ fisilil, L!nii?av, iiiid D™i;a{]96j). Ityi'.licnkoClQBS),' Do. Truoclfll and Hii,sti'ller (,1968).; Stnil(;litnn, nnd' VoimK(1961). D'vaclikova and khoiliikiivstiy{l968).

KT(13y KTi 14) KT(23) k'r{26j KT(,35> K'r{68) KT(72) Kr(89) K'l'{01)

n,sio,''-H;sio/-iirll.,SIG;°l!,.'^iO.-'-i2ll H,,lJO/-lL,HO,--i|M' M1,*--M1/ + 11*

liGO:--lt^^GOl•' K* i - < 0 / ' = K S O /

H+4S0,-' = HS0^ll,S°=ll*'iHS-

'In lugri: rs-cent practice, itic cxpicssiptis lia.-^cd on llic.fjrigitijl wurk of llariietj and co-"'orki'rs (HariiecJ and Owi:n, 1.9.58)'.cnutili'/.id, de.'.-Iiitt' beiiifj limiled to 50°G inaxliiiuin.tcmjjtralurc.

teiiiperalurc of the water. For a few reactions in which data at Icmperatnres otlier than 25 C were npt available, the 25 G value.of the eijuilibrium conslani is used at all temperatures. The effect of pressure has nol been calculalcd in WATEQ because the necessily of in|nitiiig a measured pH value virlually titnits WATBQ lo surface and ncar-si;rface waters and mucli nccessarj' dala Is nol available for Ion jiairs. Correlations suggested by Ellis and McFadden.(1972) allow the calculation of the pressure effect pn equilibria involving i pnly minerals, and ,';imjilc Ions (not ion pairs) to be, made for lempcralures to 250°G., These calculaUons suggest lhat pressure effects are not large for pressures Icss'tliaii a few hundred aliiicispberesREDOX.REACTIONS Oxldatton-rediiction equilibria havebeen treated in the same manner as dthtr reactions in WATEQ. To achieve ihisj the incasurcd_Jf,h value, or the Eb value calculated from the measured concenlralion of dissolved oxygen, is converted to the negative logitrilbm of the conventional activity of the e!celron'(or pE) by the n;lation.

aqueous ,=Glu[ion are zero by convcnlion. The convetiliojsal electron actl\'ily thus ranges from ICT^" lo 10 ^^ while the actual cteelron aclivilv is about 10"*° lo 1 0 " " " ' . These , . .. . , conventions are discussed by Sillen andMbrlcIl (1964) and by--W' Truesdell (1968). " ' All advantage of the use of f;Iectron activity is lhat it is not . +3 necessary lo set separate ei|uilibrlum cxjires-sions. For example, llieupeiju iii briumredox between Fe *^ and Fc is expressed by a cohvcnlional equilibriifiireoiistant.


"F. .+3 a„



E^Eh/(2jQ3RT/F), j


in which F is lhe fa rad ay iind 2.3Q3RTfF is lhe NcnisL Klqpe; pE is related to the conventional acti\-ily (jf the electron by rT„-=lO-I'K(42)

and the value of the equilibrium constant may be calculated from G^", Fe* , and Cj-", Fe*^ (*^f°' electron = 0 by convenlidii). Oliicr redox equilibria,arc treated similarly, and the inclhod of calculation of the con ecu I ration of ion pairs involving iron is the same as that involving melals lhat arc not redpx active. In natural w-aters that contact lhe almo.sphcre the dis.iolvcd oxygen (DOX) conlcul may have'been riica.sured in addition tp or in jilacc of the Eb. If ihii dissolved oxygen has been nieasured, it info thc-program after the n'drmal data as a slaleincnl, "DOX = (jiarLs jicr million of dissolved oxygen),". Two valufiS uf jiE arc calculated in W.ATEQ from lhe relation. 1 1»E?-logA'-pH-0.,^logg[j^Qt.Q.2.1 l o ^ ' g n o x - J (44) in wliieh log K value; are froni tlii;rmudynamic f!iila("PE (^..-Vl.C ()'*) and frum the ••mpirteal Eh-p11 relation fur n a l e r s in (-untaet with the almo.-jibere (.'^ato, l^fiO). ("EMf'IR P E O ' ) a u d DOX aclivitif.j an; on j nudal .st.-ilc. [fal'JO.X inea.-nrciii,t'iit >.given \%illlciiil an Eh value, the value uf i'E ('.'\I.G 0 is used thriiii^litiul the |iriigraiii. If itifte:id, EMl'lH I'E ' t is Io be

This eijuation is .'^tinihir l<> thai assumed for jill, and because, bolh mca.siiremi'tiits liave an unknown liquid jimelipn polenliaf, lhe relations of jiE to cleetron activity and uf. pH to by(Irf*g<:n ion aciivily an: equally linriirt.iin. Il i.<; Mece.-Har)', however, to use ret;tliuits flcsjiile the uiicertainly. The slaiulanl free energy and enllialjiy of lhe hydrated ele* Irun in


C 0 i - l 7 . T T . l l PKOGKAM, GALGIILATIUN OF GHFMICAL F.OUIIJIJRSA, NATLilAL ^VATF.HS a-i well a.-i s u m e trace irlemi:nt a n u h ses rnjy hi.- included. In o r d e r l o allow t h e inctiijipn of dala. t h e Ia~t space on t h e fir.-,t carfl is coded with 1STD.-\T.-V wliicli ii tlie i i i n n b c r o f cards containing lhe ncec.=.sarv dala inchiding the normal c h e m i e i J aiinlysl.-; and the s a m p l e dejcrijilion. C a r d s after t h e c h e m i e a l analysis are .used for ojitioiial data. .-V blank card jmist b e inchidr.d after each d a t a set t o f c p a r a l e d a t a s e t * . T h e r e q u i r e d d a t a Is c o d e d in free field ( t h a t is. o n e ^[lace bf tween (45) each number) in tlic order indicated below-. Sec li.-^l o f identifiers iti t h e c o m p l e t e com p u t e r p r o g r a m for detailed w i t b each and

a d i i | i l e d , the s L U e m e n t " : ^ l l ' O X =- 1 " is a d d e d t p t b r ojilioiia! d;it:i. .Scjiarale analyses o j ri-r.:iv-ed a n d o s i d i / . e d sjiecies allow the ealeubitioii o f p E valu-':- - " h i e h m a y be comjiared redox equilibrium. Tw.-. s u c h iilher o r tbe tnea.-iired r E l b e s l l i n a l e t h e di-gree o f i n t e r n a t [lalrs a r c .suifide-sulfate a m m o n i a - n i l rale. T h e ei;:: i l i b r i u m b e t w e e n sulfide a n d sulfate c a n , b e WTillen, H , S + 4 ! 12.0 - 5 O4 - +1 OH * +8e-,

and t h e mass a c t i o n exprrr>.sio]i,can b e r e a r r a n g e d t o give I.E=(log A'+log t i ^ f y , - l o g fi! [ ^ s - l O j f H - 4 log ",,^ y ) / 8 . (46)

d e s c r i p l t o n s , Samjile scLs o f d a t a are given w i t h t h e resulting p r i n l o u l a f t e r ihf: p r p g r a m . Card 1 Card 2 Snmplc ncsCTlptitni (79 sp.acis) aiid ISTHATA (space SO). T t M P , PI I, E!1M [in volts, corfi.- 9.9 if daia is nol avaibbic),


the equilibrium





FLAG ( = ' P P M ' . ' M G / L ' , ' N 1 F Q / L ' or ' M O L ' ) .

yields tlie, ex J > re.>s i.o 11 pE=(-log / f + l o g a : ^ Q - - l o g 0 ^ 1 9 + - l O p l W l d g , c i , j ^ o V ^ (47)

T h e s e q u a n t i t i e s , P F C A L C S a n d I ' E C A L C N, are c a l c u l a t e d in W A T E Q .

Although gas p a r t i a l pres,siires are seldom mea.>ure,d in

n a t u r a l w a t e r s , in s o n i e s o l u t i o n s t h e y m a y b e c a l c u l a t e d from t b e gas solubility constant,-; a n d t h e W a t e r analysis. T h e jiartial pressures of C O j , O i , and CH4 are calculated from the foil o\vi n g ' e q u a ti o n s , ' o s ^ C O r ' ° ° ^ ' ^ ' " ^ ' • H C 0 5 + ' " S , ' ' H + - '°S'°.H3 0 ' ^^^^

Card 3 CFiL-mieal analysis in PPM, MG/L, and MEQ/L or MOL/L (set FLAG) in the order Ca, Mg, Na. K, Cl, SO4. HCOj. Fc, H , S , CO3. SiOi, NH4, B, PO4, Al, F . N O j . Succeeding card.-.- . - - - Other data (iiienlirie!-, equality- sign, numerical value, and coninia) Including; "DF,NS=D'' (if not specified, density- is set equal tp 1); if alkalinity is corrected for noncarbonate alkalinity " C O R J A L K W I , " (omitted if not corrected); el ectri lia! potential, (volts) of the Eh cell including the'calpmt! reference electrode "EHMC=a,"; electrical potential (viilts) of the Eh cell with Zobell's solulion for calibration, "EMFZSGE=n,"; parts per million of dissolved oxy-rgen,. "DOX=[3,"; and certain trace elements including Li [1=50). Sr (1=S7), Ba (1=8?) in the form, •'CliNlTS(I)=D,''. A semicolon in place o f a comma follows the last-data statement. Last card Blank.

Akftrlof, G. C.. and Oshery, iL L. 1950. The dielectric constant of water at high tempe futures and in equilibrium with its vapor: Am. Chem.;Sdc. Jour., v. 72, p. 2 8 4 4 - 2 8 4 7 . -t.Bamesv IL L , Helgeson, H. C , and Elli?, .\. J.. 1966, Ionization const:inli in aqueous solutioiis: Ceol, Soc. America Mem, 97, 401-413. Bates,"R. G., Staples, B. K., and Robitiion, R. A„ 1 9 . 0 . Ionic hydration and slni^c ion activities in unissociated chlorides at high ionic strength: Anal. Gliemislfy, v. 42, No- 8, p. 8 6 7 - 8 7 1 . Bell. R, P„,and Georyt, J. H. B., 1953, The ificomphte dissociation of some thaltous and calcium sails at different lemperaiures; Faniday Soc, Trans., v. 49, p- 6 1 9 - 6 2 7 , Berner, R. A., 1.967, Comparative dissolution diaradfirijlics of carbonate minerals in ilir; presence and ali.-^i-nLf. of..iquroiu masiiciiuni ion: AuL Jonr. Si:i., v^, 265, p. 4 5 - 7 0 . 1967, Tlicrmtidynamjc stability p f iron siiiridi:*: Am. Jour. .Sci;, v. 26'5, p. 7 7 3 - 7 8 5 , Bricker, 0 . P., 1969, Stability constants and Gihli^i free fiur^t-= of formatiini of ma^ailiilc and kciiyaitc: .Am. Mincraloyist, v. fi4, p. 1026-1033. Brown, W, E,. 1960, Behavior of .-^ligliliy .^oliibSi- caleijiin,phn.^j.hates as revealed by pha.vr-cquilibrium calrntallrms: Soil Sri., v. 9LI,.p. 5 1 - 5 7 , Q L , HoHteller, P. B., and Sifberl, R, M., I97n, Slmlies in [lie system Mi;0-SiOj-GO, •11,6(111); the aitivity-proflncl i.on.^t:ml of ,w[jidit..-:Am. Jonr, Sei.. No, 2 9 3 . p . 6 5 - 8 3 . Ghui^htai, A„ Mafsliall, U., am! Namonas, G. 11.. 1908, Giiml.ti-.Ni.^ in talciuni piiiiiiiliate solntiiiiis: Jonr. Phys. Glietiti-lry, v. 72. p. 208-211. Davies, C^ W., 1962, bm As.-im-i:itiim; Wa.'iliiii^tDU, li!illenyiiflli>. I'll) p.

log P Q I =1OE^'+2 lognj,^_o+4pH+4pE,


' ° " ^GH4 "-"=^'" ' '"^°HCOi" 9ldI-9pE-31og.iy^O.


T h e - activity of w a t e r is c a l c u l a t e d in W A T E Q b y a p p r o x i m a t e relation (Carrel,: a n d C h r i s t , 1 9 6 5 , p . 6 6 ) the

° 1 1 , 0 ^ 1 - 0 - 0 1 7 Em-.


w h e r e £ m . is t h e s u m o f t h e m o l a l i l i e s of dissolved an Idns, c a t i o n s , aiid neutr;il sjiecies. T h e e q u a t i o n yields r e a s o n a b l e val lies i f Ziii- is- IC.KS lb ai I imila!. ~

I n p u t t<i W A T E Q e u i i s i s l s o f u c(iiiqite:le chc mica! analysis of t h e water sainph; and fieh! measurf:miints pf its tiunjieralure a m ! p l l . If :iv:iil:ili!e, m e . a s n r e m e n t s o f E!i ;inil dis.solved u x y g e n

hydrotliennal systems at IJtike. Inorganic Chemistry. Rc.. W. P.. Form and stability of .raelions of aluminum w t l i aqueous silicaat 25''C: U. 52. 4 7 6 .6 6 Gcopbys. Belle Fourche and Clay Fdiimicr.S)e. . p. and Iheir j.ition of kaolinite from f^ hyilrulhermul sulu I ion. C . Harned's mle behavioV of Slime . J „ 1968. v.4. I n v e s t i g a t i o n s . B." II. 1675—1678.iiesiiim wilh biearbonalc: in Geolo[^cal Survey S o c .]: Am. p. M. A.. S o c America P r o c . Jour. V. Some quantilalive aspects of the chemical . Dielectric constant of Water-Supply Paper 1827-B.idden. 1).3 0 0 ' ' G tiniiienilure rani.M. S. 5. B. Fej(P03. 2 6 1 .aqueous ' i'n B. ft. Gisicliemislry of liytlrolhtrmHl ore di-posits: splutions. and Trui:.. M.equilibria: New 1-969. of n e s f i i i e h o n i t e .T. 59.. 5 3 5 .3b. p. V. " 3H. water from 0° to IOO°C: U. 1967. A calorimetric 1972a.. y.-il. p. Sci. and the analy. J. The oxidation slates of lhe elements and their Hanna. Robie. p.\lSj(G03)^. v.Mesmer.. 450 p. Vp. Prentice-Hal]. v. 1969. Clarkr. Jour.5 4 1 .. Paper 650-B.36.. .537—543.slry. M(. l-l Cosinpi liim. anil ter-valetit melal O'yai-hkova. /. Ki-uni-lh.-'Vel. computed from solubility electrolytic solulions [3d e d .. 1955. W. P. D. lon-associalton potential in aqueous . U. 35.S. 1 1 .tliirimi-lrii-ally ili-ti •rmini'd I014 K. Roberscm. Sci. f'. 478 p. Slabiliiy. rd. J. 1457—1466. of ions in Killrick. 19.. p. {^bi-mii-lry. Sci. 'I'lirrfiiudynainic equiliiiria in IIHI sysli-iii.ati. 3. GhiiMl. Solubilities of inorganic and elevated Icinperalure. 662. Geol. R. Nriaj^u.. E. Slandard Reference Data Ser. 1 ... He. Survey Pi*of. and Robie. J. Surrey Jotir. Standards Jour. 35. of liionlmorillonilcs.als.0 j .. and liarlholDim-vv. A.'. Gcpl. 1965. Iiii::teller. 1965. W.sphcre total prcs. 33 p. Pytkowic/.. NaCl-NajSOj sohilions explained by an ion-a..taiici-s at 29H. p. H. A.. C F. v. p.S. hydroNide coinplcxcs iii dilute solulion: U. E. evohilion of the oceans: Northwestern Univ. Tbe Hamrd. 803 p. No. C..). Siin-cy Malmberg. Particle si/-c effect on the reaction geothite-bcmatite -1fJ ect rol)'Ics in aqueous . 0 . hi^b-almnina and clay mi nerals: of ehrysiitile: 543-547. Solutioii. . Jour. 1968. v.'^ 11..sea I»afon. The !'riiu:i[]lrs cif fllii-mlra! Kiinilihriiim: l/. I. and Roberson. p. Ceok Sun-cy Prof. 53. 7 2 0 ^ 7 2 1 . disserti. spdiOin sutfale complexing in aqueous solutions [abs.\mi. i!. from 0 to lOO'C: U. v. Researcli. R. and Mmili. I n c . 0 .v. Prederiek. R. ItiJivIdiial acliviiy eoi:fficients of ions in.Surrey Jour. F. Glirislrn.0"G) and inie Tale-ebrySolile f-qiiilihriiun in aum-nus--uliiliim.S. JPUlis. D. p. and Me. Jr.S. 413-426. .5 1 4 . BUT. C F.M„ and Christ. AMI) JO. 1958. p. E. M. and Waldliaiiin. Ph.ii: and prcs-surcs: Am.. dala: Jour. S o c America Abs.NHS 247 niiihipb. Pajier jeiir. . Graphical methods for . v..i!ll)a1pies of formation Am. WiiLnton. p. 190.Clu-m. Phase equilibria of some minerals in Ibc Ree.--: Giot.e energies rif formation: U. Inc. Abs with Priigrams V.. Thcorc!ica] mean activily eoefticicnl. U . 19*62. 1972. PHein. Stability of vivianite and ion pair formalIiVii in the sysleni statvfard Gililj-i fii. p. Harper & Row.3. ] : New York. Jour.M. Lniversity Pfi.. R. and Iheir Sei.9 4 2 . M. 1968. j ( 0 U ) i C O .. Solubility etiuilibrium cdpslanl of slrcnjiite: Am. A-.iioiJ[. Solubility studies on vivianite and strengile.. el Co-i-moehim.i'<itJictnjcHl iiilirprclalfjns: Keenan.2 5 . G.. No. Jonr. A. A.. 1 2 2 . arid Kester.jN'AjO-AljOj-SiO.s: GIK-NI. v.itre and at higher lenipiralufes: U. Ik. . aild Hemiii^vay.. Geol.riea aliiu>>jibi-re (I. 267. M i .. R S .-11. p. 1973. Complexing uf ma. D. Am. G. J.sdell.slabiliiy rela I inns of alunite: -Econ. B 1 8 0 . H. Am. T.ssoeialion model. Jour. pL I. 271 p. and Thompson. system F e . P. fL A-..07. 2 4 .. Garrcls.5 1 1 . 1969. 9 2 9 .si.s: Am. Ccol. 3. Aberdeen D . of steam: New York. Ganlbridjie. and Owen. Reiiiliold..55pL mi-3. Am. Hamer. M:. 1909.conslant from 25 to 225° . Kielland. and Young. Grochita el Cfismochiin. Am. tVi P r o c e e d i n g s of I i i t e r n a t i o n a l S y n l j i o s i n m o n . U.Soc America Proc..r li chemi lit fy . D. 0 .U c l / k c . 1 4 7 .l-'. S-.irch 1966: U. 190. (ieiicliim.'.=ejreJi. 5. 1968. C L . 1973. B. 0 1 7 8 .5"K {2:1. Ghem. Raymond. 1971. Survey Waler-Supjily Paper 1827-E. C E.: Am..soIiitlonR [2d e d . (if minerals and relaled ~tili-. p.. The principles of electrocbcrtiislry: New York.. F. a//".lyslein SMgO'JGQ. Ttiefmodynamics of.Hemley..: Acta. iNatl. and I-jfon. Chemical itile.Mineralo^i. The CQihs free energies of subslances in tbe of'carbtiiiale solubility by carbonate complr. 392 p.lL. W. Phys. kl-311is. 197.diiiniiiiim 56. v. 5 7 . v. 1966. and a rli nilc. IL. Sci. 260.S. F.. No.2 2 9 . Hustiller. solubilily iile. Studies in tlie . p. U. 19. 89 p. 1961.s. Jl. 4 6 5 . Am.. p. 11. 47—53. Jour. p. Hy d ro.B 1 8 3 . p. J. B. J. .. 36. v. 1971a. . . S«i(-. J.^en.1 5 6 . with Proj^anis..-6S.*.-md AlS'' valins for lhe itile raet ion of suljihate ion wi|li sivi-ral hi. 267. 599-61Z Jour. 1961. v. Rineharl. 2 2 4 7 . .U(5-497.-H3PO^-IIjPi Geoehiin. Union Trans. W. — 1971b.. A d a . l b . v.y of f(irm. C. The physical diemistry of beulfale acid . B.H . R.and Biocliemi.C 0 j al25''C. and 1914 729-804. O.M. 1952.140-1-15. Research.S. . The i. 2 5 . and Siever. M.3a..36. Roberson. 1. Baes. Ceol. . S u n e y Hull.0 system at 25°C and 1 almospliere. J.. v. 1937. . and Sweeton. The. J... 32.8 .a. and hydroniiignesile. Erec eiier(.S. 2 7 1 . p. 11. 677-686. Latimer. . 1971. Japan 1970: \Va. 362. p. Helgeson. 16.surements at elevated tcmpcrahires: VI.. p. v. C. Set. explored (jcollu-rm. Sci. J. Slability of monlmorillonile^: I. Pet liy bridge.S. P. E. Aria.siii of prneise eptiductimelrii: data: Elcclrochim. J. 0 . 550-D.1 3 9 montmorilliinile: Soil Sci.3 0 9 . GiiHci A. J. Boric acid equilibria: Hem. H. I. IL C .t. A. 0 : U. A. p. Rein bold. Gan-rls.. and Khodakqvskiy. p.SiOj-COj-HjU (1>: Tin' acliviiy-prod net ciin. G.ya. . 2 3 0 . M.s miiii:r.ld Spur nioiilmoriltoiiilcs: Soil Sci.TlHILSDLLl. HnSlellrr. Stilubility implications of apatite iu ^l•a water. Stnugbtoii. i*Jall.. 1972b. D. v.. Tlieidii-. • Ca. 1960.0 in. 2.. 2. D. ] . Geoloe'cal Survey reitarch 1969: 259. J 969.. A.urritinil.605. Parlia! iimla! vfilimir.-|. 459-470. Hem.-!. Bur. (Ixnidoii] J'ltir. 136 p. D. 1259. J „ |V7I. C H.0 1 8 5 . Geoliigy.-ilabilily and surface energy of briieile in water al 25''G: Robie. J. 64. Jour. 2 1 7 . GeoL Ceol. J o h n Wiley aild So. Huridc and sulfate complexes. Aqiieous solubility htmliesof KJ O. Acidity Waicr-Supply Paper l827-A.65. Survey .. 272.Soc. Tbiriiiodjnainic properlies lliL-li-ller.iurc: Am. waler.. p.S. Mitierabigisl.. Temperature dependence of waler al 250°G and one almo. A.\L !^. p. J. ami Key. A clicmical model for . Thermixiyiiiimie properlies Ccoeht'inistfy liili:rn«l. 3. C-.. Dellietl.. B. and ^Minmtjoy. and Seidell. MgO-. 264.xe.systi-. and Pob. R. W. J. . M-. Sci. New York. 0 . 1971. v. .studies of aqueous aluminum Maclnnes. p. Silica in thermal waters. and Chri. 1072. v. Control LangniuirCj Donald. Hemley. Laboratory and rn. v.S. (iarre. 2 8 9 . Si-HjO and Ti'-H.2 2 4 9 .4 8 4 . v.2 5 8 .. Chemistry.s. Lind. 1973. IL .inislry of so mi'. I'hys. 1939. J.. J. Natl.fw.solution.. Hull. Standards. G.|iinglon. J. New York.IM3 b:irs) pri-.c. C. IVfi'J. E:. Heminsiway. p. and Keller. IL.-3Hj0. J. . Thompson.sK. p. 1968.s S-lljO.. . 1955.M. J. 256 p.-e. p. A. v. T. and MaryoH. 1. 57 p. v.sinan..s of Strong 1971.. tlie ion.-.M. v. determination of I be slandard en ihal pies of formation of hunt iic.mi<-. l^'iit. 1967. W. IL A. fL E. 1069. v. 820—823..C0. inCtal-orgLinic com[iounds: Wasliinylon. 1108—1125. R. 491 J>.. 5 1 ..

s of chrinical U S . Baleji. ajid Martell.s at elevated lemperaiures as derived from p. Acta." from 10°" to 50°C: Geochinv et Cusmdcliim. 4.^. R. 75. Ro. The advantage of using pE rather than Eh in dihydrale atid association equilibria in . C'l'VH. 1968. a Tniejidell. H.Scariionie acid and the decree of Kyilrei]ysis of ihr G O . 1968. 1964. andEng. li-. A-. 1. H.. diserepancics. 4 . C temperalures. ions in . 1. Bur. 559 p.shini. Note 270-4. Ion . p. v. IJ. A. and Jaffi:. G. v. Gibbs' e!ier<. 1970. F. Selrcled value. Evans.pioltte al Earlh-surface Smith. ('i-olo. 4. 1960. R. 1952.3. The apparent. Data.T. p. Pari I. Delerminalion of tbe bydrotysis of . li..siieiation in n a l u i a l l. 1972. $ ) c . R--. lUir.-i: U. ' ' aiid llCflj' Nall.lliltlA-. U. L B..C. 1972.. Si'leelid values of chi-mieal lliermodynamie projierties. I). F. N.. U. L..s. Education..ilc.'•IH CtniPiri-ER I-KOGRAM. 1963. 17.. Pub. Ycatls. v. Ccol. L A. NATURAI.S.nt c-ilions. Ilalow.mai!.. l').. p. and Dru^a. p. W.^ W. Chem. 5 1 . Standards Teck Note reprinlrd in 1961. 264 p. and Baes. Wa). v. No. WATl-KS KidiiiiMin.. 16. ThcrtTiodyriamics. R & D rept. 1 8 0 .. 49. v.-md. Jolm. 479-500. F.lon. and Ho^leili-r. 2. [ L o n d o n ] Sjiec.>! of clieiniral llierinodyiiamir jirb|irrlies: and . Mineralogisl. E.s: Geoebemislry. 32 p. rock-forming minerals. K.. V.son. M . 60. B.iim sidfatc Truesdell. Natl. 55. E-an. v. M„ and Alberty. Evan.iso'ciation constants of . Bray. complexes: Cliem. E. T.•illl 1 1 1 ions of alkali rarbonales: Am. K.. Irring. Jour.stry Jntetnat. 99—107.r.. p. 1958.. Srlieled valiic. Jr. v. 1 .. V. v. Ihilleruorlhs . U. I).-. v.i. Go:illiy.ioti of dis. p.. and Marsbali. 1 9 6 9 . S. . Elyafd.sorialion constants of — T969. 151-163. 1235-1256. Parker.Iratij-laled from . Sainiiel. T. W.ile and caleiilaticfnofdissociation coiLSlatils of orlhusilieic acid al elevated Wali-rneld. (168-874.. L G. Baily. T.Seierilirir I'ubliealioii^. r-Jectroiyle salt .. p.slabiliiy conditions: Am.. No..GHFMIGAL EOllM. Solubibly of caici. 1968. l-'li-eimlyle soliilioiLs: ^ V m s d e l l . A„ anil Stokes. . p.. 32. H. T„ Jr. Lindsay.sol 1 1 lions of alkali carbonaJes and Iiiearbonates at elevated Wijlki.C. R.. Tran.. 1956. Jour. tiilhalpy. Mackeiiy. D. Slabiliiy complexes of melal-ioh Wollast. A. Ry/henko.s.Slatidartls Ciru 500. H. A. R..8 2 2 . R. SlamlardsTecb. 0.-^siiiL. Rpl. W. p17-20. ^ and Bricker. IL. v.sodium Kilie. and enlropy of len KSO. M. J. 9 2 8 . 17.Seliimim. S. Dcle>tnin. Gi-ochi-iiii. 6-4. M: . WAPD-TM-204. impl ica tiotis: Ant 1019-t022.. T.. G. Molaki.1 8 4 of tieclroiylic Sol uti on. and Jones.( Zen.. Geology. coiiduciance measurements: Wa.inperatures. Di.ibies. Sillen. Chem. P.. consianLs of iofiic eoinplcxes of various adenosine phosphates . A. 196B.. p. ealculalioiis.P. B. pGeokhimya. Nail. ExpcrimenUl precipitation and ECiicsis of se. IX. Tlie behavior moimv. Tbesolubility of magnetiic.4 p. and Johii. M-. Wa^imn. 1961.C. A.1 2 3 1 .ATION Ol'.sodium tarbonale: Faraday Soc. 1927. 141 p. disorder in tbe hydrates of .53. 524-55. l/viiie. Jour. Albinlc Energy and ll yd rol ysis of ferrous iron in aqueous solulions at elevated Comm.sysleirts at 25°C*. 11. atid Slaveley.268 p. It.with Wrigbl. Ihrrmodyiiainit projierlie. H. 163-168. p. ll. v. p.Iliir. Pliy.S o c Jour. C . Liiifurd.1045—1662. Tlu-rmody ha mic investigation of Salo. 270-3.y. brines: Glieiii.."'ri.. 2 . Cbemiilry.. B. Mineralo^st.U. 57.. CALGUI.. Equilibrium in tcinjieralKre. D. .several aqueous mixed redox equilibrium ealeiilalions: Jour.6 2 .. A. D.. F. Oxidalion of sulfide orebodies: Eeiiii.

B.814 15. The method of calculation in WATEQ is back-substitution for the cations and successive approximation for the anions with convergence on mass balance for anions.1978. J. p. and the oceanic sulfur budget. and Berner." Nobel Symposiim 20. Geochim. drew on the models of Barnes and Clarke (5. Denmark). such as the addition of pollutants. and Thomsen. John~T7iley & Sons. 16. ed. pyrite formation. I. M.R. 814-831 (1977). 20.. a i E M I O L MODEUNC IN AQtffiOUS SYSTEMS 18. C. SOLMNEQ C6). and bioturbation in Long Island sound sediments: Report of the Foam Group. The sulfur cycle. Beyer." 433 p. 2) evaluate solubility hypotheses.A. R. Goldhaber.C- p 36 SUBJ COl'lP UAT2 WATEQ2—A Computerized Chemical Model for Trace and Major Element Speciation and Mineral Equilibria of Natural Waters JAMES W. Almquist and Wiksell.These models.B. J. CA 94025 DARRELL KIRK NORDSTROM Department of Environmental Sciences. Madsen. Behrens.B. Berner. 5. The sulfur cycle of a-coastal marine sediment (Limfjorden. the earlier water-mineral equilibria model written in Pl/1 by Truesdell and Jones (2^. drew on the preceding model of Barnes and Clarke (5^).S. in Dryssen. Oceanogr. C S. U / c i i l nsjrr b. University of Virginia. 19. 17. P. BALL and EVERETT A.. Gardrter.S. diffusion. 1974. Massachusett:s.E.. } ) . M. ed. 22. (fi) The WATEQ program contains an extensive thermodynamic data base which was carefully selected for use with low-temperature natural waters {^. 0-8412-0479-9/79/47-093-815$05. Allen..K. 569-655. Sulfate reduction. Sulfate reduction. FOAM (i'riencrs of Anoxic Mud): Goldhaber. Earth Planet. "The Sea. Charlottesville. Addison-Wesley Publishing Co. 37_. D. 21. In addition. Stockholm. manganese speciation is Included in WATEQF. Ceological Survey. copyright .• ( 0 . Toe.G. CUcryi. 163-198 (1975).).R." vol.D. -Berner. Litmol. 3) account for sorptiondesorption processes.A" Sulfatie reduction and the rate of deposition of marine sediments. Martens. Acta 37. C ... ONiyERSITy OF UTAH RESEARCH i m m m E A R T H scimcilm The protection of ecosystems. and A) incorporate chemical kinetics. Water Resources Division.G. New York. in Goldberg. D. Lett. R.. Amer. Ecological Modelling 1.B. 1969. in turn.) and a prepublication version of Truesdell and Jones (2. Sci.B.. The prediction of trace-element behavior in ecosystems requires a multicomponent model by which one can: 1) calculate aqueous speciation of the trace eleraents among both natural organic and inorganic ligands.. Froberg. Cochran.) as well as the thermodynamic data treatment of Helgeson (7^) and Helgeson £it £1^. WATEQF changed to-the more rapid backsubstitution method for anion mass balance convergence. Reading. and from WATEQF. and Jagner. 22. L. 277. and additional modifications are explained below. the Fortran version of Plummer ^ £!. requires that we understand natural processes and develop the capability to predict the effect of changes. Rosenfeld. p. and Kaplan. 193-237 (1977). 347-361. E.D. on these ecosystems. These activity coefficients are considered more-reliable than the standard Debye-Huckel equation or the Davies equation for high ionic strength solutions (up to 1-3 molal). The present model has evolved from WATEQ'. RTX". JENNE U. Activity coefficients for the major ions are calculated from a computer fit of an extended Debye-Huckel equation containing two adjustable parameters (^. C .. Cosmochim. The related PL/1 model. 1972. O. 53-68 (1973). and an optior for calculating activity coefficients by either the Debye-Huckel or the Davies equation is provided. Some aspects of modelling the long-term behaviour of aquatic ecosystems. 492-498 (1978).K. Sci. "The Changing Chemistry of the Oceans. "Initroductiot^ to Numerical Analysis. WATEQ2 retains most of these features. upon which our health and lives depend (j. J .25/0 This chapter not subject to U. VA 22903 I 1 'k RECErvED November 16. Jorgensen. Jour. R . 3^). Menlo Park. J. This paper documents a chemical model that partially accomplishes the first two of these four goals. 1 ) . Chemical models for sulfate reduction in closed anaerobic marine environments.

To the selected set were added six minerals for which thermochemical data are unkno-wn. hausmannite. MnS(green). Alkali metals complex with 0H~. i f l spite of the near absence of solute complexes or pure solids. However. or for one ligand group. and Zn. pstlomielane (Ba[)^7aCaQ^13Ko_o3Hn2'h«n^"''Oie-2. and the HMn02 complex.. However. birnessite. . such complexes have been dropped from the model. EX AL. Fb. The partitioning of an unexpectedly large portion pf an element into a particular comple.t^l^Mn^s6*taS^Og. have materially assisted in this modeling effort. Elements. Cd. B. ngutite. Cu. Gs.. H. and sevetal ion pairs of major constituents to the model. all reactions in the present model which have been added or revised. complexes of Br and I." L. rhodochrosite. it would be desirable to enter at once into the chemical model the reactions and therraodynamic data for all element B whose indus ion. The six minerals are^ cryptomelane (Ko^MnJ^gMng OJ7). our experience is that the greatest reliability la obtained by adding only the data for one element. F. Computerized Chemical Model 817 We have added ten ad. at a time. Cs and Rb have been added to the model. The following subset of the mineral species was selected: p'yrolusite. and operational-information to facilitate input and output of the data. Other changes include'reorganization of the computer code into a series of external subroutines and changing the mode of convergence to decrease the number of iterations required. activity product to solubility product ratios (AP/K) via computer terminal printer or tape-driven plotter. Similarly. whom we have frequently called on for specialized knowledge and infonnation. Colleague criticisms (constructive and kind for the most part) of studies in press and in preparation have prompted us to make specific testa and proceed Immediately with some change ox adtlition. we have developed some adjunct computer programs. Because of the Interactive nature of aqueous solute speciation calculations. Collaborative studies have often provided the impetus to add some specific element. led us to make an expanded compilation for the complex or to consult with colleagues to assist in selecting best values. Additions and Modifications to the Model The thermochemical reaction values vary according to the way the reaction is written. As.x has. together with the selected thermocihemical values. For mass I •i • \ . Mn2(S0it)3. The solute portion was utilized in Its entirety. and Zn). BAU. some sparingly soluble salts. which would otherwise have awaited a "more opportune time. todorokite (Gao. Hn3(P0ij)2 and MnHPOj. These include AACALC (Atomic Absorption and emission spectrometry CALCulatiori) . but the MnOK*" and Mn(0H)3 as soc iat ion reactions were expressed in terms of H2O and. Mn(0H)3. MnO. The merit of including"Cs and Rb In the model. For these reasons. Mn. manganite. AACALC (FORTRAN) reduces atoKiic absorptioti or emission spectrometry data to concentriatipiiB. Therefore. severel metastable BOlide. Cd. they have not been included In the model. then test data sets and real world water SMiple analyses are run before making further additions to or changes in the model.ditional elements (Ag.3H20). are available in an adjunct report (10). Jones and A. in order to obtain log AP values of the individual minerals for different waters. Cl . HnClz. or group of solid phases to the model. EQLIST (Equilibrium computation LISTing). a duplication of the Kn(0H)3 complex. MnClj-ZH^O. or. AppareiSoveraaturation with one or more solid phases of an element has often prompted us to seek out and add data for additional solute complexes or more soluble solid phases. All solute reactions are written as association (. there are data on alkali metal compounds such aS' Rb2S Which might' have been included. MnSOi^. EQLIST (PL/1) constructs tables from the WATEQ2 (input) card file. Pb. the pure alkali metal sulfides are " known to be highly unstable and/or deliquescent (11) and are rarely found as"minerals. and NO3 only at such high ionic strengths that the basic assumptions of a multicomponent ionassociation model are no longer valid. and EQPRPLOT (FORTRAN) constructs ratio and scatter plots of dissolved constituents. In our effort to collect the appropriate data and develop the requisite understanding of geochemical processes.393Hg[3^H73Hnftl3i*Mn^'*'Oi2-2H20). blxbylte. thermodynamic data for these complexes are highly uncertain. pyrochroite. is that in the future the •activities of the uncomplexed alkali metal Ions can be used to compute their probable substitution into certain silicate minerals and to examine ion^exchange processes. bight affect the computed activity or equilibrium solubility ef other solute species. Plinnmer provided frequent consultation and supplied a prepublication copy of the reactions and associated thermodynamic data for the manganese section df the WATEQF chemical inodel ( ± ) . activity products (AP). Therefore.5H20) . Rather large sets of solute complexes and mineral phases have been added for Ag.78^6^57 Mnl^SjMn^ Ols) . Various associates. Truesdell have also been particularly helpful on many occasions. The following subset was excluded. Hn. Cu. and EQPRPLOT (EQullIbrium computation PRlntlng and PLOTing). Aqueous Complexes.formation) reactions whereas the solid reactions are written as dissociation (dissolution) reactions. MnCl2*AH20. Additionally. In addition. lithlophorlte (Ll2AleMn|-^Mn-f^03s-14H2O).816 CHEMICAL MODEUNC IN AQUEOUS SYSTEMS 36. Rb. K*" instead of 0H~. Ni. As. The manganese sections of WATEQF (4_) were heavily utilized in preparing a Blmllar section for WATEQ2 (1^) . Mn2Si0i4 and MnSiOs. hollandlte (Bao. K. HhClj'HjO. and rancleite (Ca{). Ki. ligand group. was excluded.

halogen-rich hot springs (42_. stability constants are needed for Iff*. For the reaction: HF° + F * HF^ U) the stability constant has a larger uncertainty owing to the competing HF° association equillbriiBO. (3) The equilibrium constant for this reaction has been measured at 25"C by Roberson and Barnes (40) and the enthalpy la ei|tlmated from the data of Wagman ^ Jil.58 ± 0.have been shown to exist by Warren (35) who measured a log K of 0. and (HF)| species. as cited by Fisher (^). by Broene and DeVries (29_). For equilibrium calculations in acid solutions. Although the reliability of this estimation cannot easily be determined. Several measurements have been made on the dissociation of HF* at 296.that using data from one .speciation. e) analytical chemistry.source would help raaintain internal consistency. (10) was calculated frora these investigations after dropping the high value of Patel _et ^ .09 ± 0. FeF2+. the remaining four values average 0. it certainly ia better than assuming iH • > 0.ineral8 are more soluble tn natural waters having high fluoride concentrations and low pH values than in other waters. Reaction 3 la important in many applications: a) chemical procesaea involving volcanic gases and condensates (40).. Careful checking with published literature values . The log K for the dissociation of HF* and for reaction 1 have been measured between 0 and lOO'C . These species become important when the pH drops below 4. so that all constants are for overall rather than stepwise reactions. unweighted) given in Ball et al. IT) show good agreement {± 10%) except for NaSOu. BALL i r r AL. ^ ) which is identical to the value selected by Smith and Martell (26) .15'K and the agreeraent is excellent (29-37). Slei)ert and C. 1978) has suggested that the low value should be discounted and that the formation of a sodium sulfate Ion triplet is unlikely.(HF) 2* (2) I Enthalpy values for the calculation of temperature dependence.226 value of Lafon and Truesdell ( _ 1 £ ) to the 1.05 (2Z. ^ . The enthalpy of association has not been published but it has been estimated by assuming that the difference between it and FeSOt Is equal to the difference between Al(S0it)2 end AISO^. Christ (Continental Oil Co. all scilute reactions are written in terms of the free form of the parent species. M.showed no serious discrepancies and it was felt .The sulfide reactions have been rewritten in terms of HS~ rather than S^" since HS~ is the dominant sulfide ion inraostwaters. and 2) polysulfidea are in equilibrium with bisulfide. Aqueous HF dimers.. b) chemical reactions in acid. was one exception. such as in the determination of either fluoride or dissolved silica (46) and f) the formation and hydrothermal alteration of ore deposits (47. 36) and the weighted mean value Is 0. ^ . BF(0H)3. 15". 1976) after comparing their values with those reported in the literature. the F:e(HS)2 and Fe(HS) 3 complexes (13^) have been In' eluded to increase the rigor of the sulfide. S. Survey.010 (lo.'^ are not available for reaction 2.05.70 ( _ 2 9 . Calculation of the activity of polysulfIde lone requires the assumptions: 1) the quantity of S|•(free_sul£ut) is not a limitation on its reaction with bisulfide'(HS') to form polysulfides.17 value of Pytkowicz and Kester (1^).70 + 0. L. If the one low value of Lafon and Truesdell (18) and the high values of Fisher and Fox (^l) and Fisher (^) are dropped.. The log K for this complex is the average of the results of Izatt £t^ el. c) watera receiving effluent from phosphate processing plants (44). 30_. personal communication. Ellis (3a) and Hamer and Wu (33). M. (25) and Mat too ( ^ 8 ^ ) which differ by less than IX.43 ± 0 . who may have used the same evaluation technique. U. (41).30 kcal mol"^ for reaction 1. 0. FeF|.5 and fluoride concentration rises above 5 x 1 0 ~ % .. The polysulfide complexes of Ag and Cu have been added to the mpdel in an attempt to reduce the apparent oversaturation with Ag2S(B) calculated for San Francisco Bay waters: (12). 12. Various''published values fpr the association constants and association enthalpies of metal sulfate ion pairs and triplets (14. L. 15.Silicate m. Reported log K values range from 0. Vasll'ev and Kozlovskii (37) have also obtained enthalpy data for • these reactions by calorimetric titration. The average iH -3. h } ) . Computerized Chemical Model 819 balancing purposes.818 CTIEMICAL MODELING IN AtJUEOUS SYSTEMS 36. ^ . The Ion triplet Fe(50^)2. FeFt. ^ . . personal communication. The weighted meas value of log K • = • 3. BF2(0H)2 and B F j O H ) " have been evaluated by Nordstrom and Jenne (50) and are In good agreement with those selected by Smith and Martell (26) who used a different evaluation procedure. U_. d) fluoridation of water supplies (4^5). Enthalpy yalues were also selected from the . K. G.j + 6 F " + 4H'^ -^SIF^" + 4H2O. Folyaulfldes and Sulfide. The log K and AH for the formation of the aqueous complexes CaF*. This addition to WATEQ2 . High concentrations of dissolved silica may he maintained by the formation of hexafluorosilicic acid: Sl(OH). Most of the other association•constants were obtained from R.46 ± 0.49. Fluoride. HFj.preliminary data of R. Lafon (Johns' Hopkins U. 49). Geol. . Christ which had been evaluated by the Fuoss equation ( _ 2 7 _ ) . Siebere and C. Log K values for the formation of NaSOiJ range from the 0. 0 5 for the reaction: ^HF" • * . In addition to the sulfide complexes of the added trace elements.75 kcal mol"^ for HF" dissociation and ^ 1.169 ±0.). ( _ 3 4 j and the low value of Vasll'ev and Kozlovakil (37) which is necessary in order to niaintaln consistency with the kinetic data of Kresge and Chiang Q8_. 24.

unpub. 1977) and agrees with the value obtained by Kashkal et^ a^. (62) and Kashkal . ( _ 5 5 . and all pertinent output data are flagged. In the absence of supplied values. Christ (personal communication. 6 0 ^ ) must be discounted because of very large uncertainties in the results. Solid Phases. Organic Lljtands. R. Solubility product constants and free energies of formation for the jarosite mineral group (Jarosite. is r ' ^ l . presented evidence. The iH for the reaction 8 has been estimated to be -31.0 for the reaction 1 communication. (61) for the entropy of Jarosite. natrojarosite. Only one log K value is available for hydronium Jarosite (13). L.from the work of G. were so high that the two species accounted for virtually all the As5"'" In several water samples. Vlek . If the dissolution reaction is written as: 6H" + KFej(30^)2(OH)g(s) t K* + 3 F e ^ + 2S0j" + eH^O (8) A I ^ + SH^O J A I ( O H ) ^ + SH"^. Nat'l Bur. data. 'the previously cited investigations for the mean iG° value and W a g m a n . (4.. personal then the log K is -14. as the hydrogen form is termed in the older literature) have been compiled by Nordstrom -(58). 1976). Therefore. Considerable. for these two ligand groups.832 fpr H^AsO' + F " = H A S O J F " + H^O (6) and log K . well below that at which water-decomposes (0 to -0. ) have. Jenne. The E*' calculated from. written in the same manner as reaction 8. Sulfates. Siebert and C. uaa assumed for the jarosite group since no data are available. pies of the ions. From the original tiata of Dutt and Gupta (56) the log K = 2. If a concentration for either substance ia used as input data.. We have therefore included both neutral species in the model. without an accompanying analytically determined molecular Weight. Jacobs and E.Is less than -12 and this agrees with the generally accepted value of -13.-lie. A. Clifton (Continental Material Co. revised AG^ data (10) from NBS (B. A mean value of -98. There is also a lack of consistency between values for the jarosite solubility product constant. partly because different complexes were used.037 for H^AsO" + F~ = AsO^F'" + H"^ + H^O * „2- (7) Thus. Reported equilibrium constants for these metal-ligand complexes also vary widely and should therefore be user aupplled.. C4) . anti hydronium jarosite or carphosiderite.. a wartilng message Is printed. Important new species added to the model are the polynuclear complexes Fe2(0H)2 and Fes(OH)4 * Some controversy has arisen over the existence of Fe(0H)3 and A1(0H)|.6 (_5}). the activities of As'^ and A s ^ under these conditions -was near -4 volts. .1 for the log Kgp has been selected for HATEQ2 from the works of Zotov et al. respectively. It' is real. The model has been expanded to permit sensitivity analyses of naturally occurring organic ligands. with a log K < _ -15. Equilibrium association constants calculated from free energy 'data (41) for two aqueous arsenic fluoride species. 1977) and for'hydronium jarosite a linear correlation between free energies and enthalpies . A. e t _ al. Neutral and_Polymerlc Aluminiaii and Iron. The enthalpy for natrojaroeite dissolution is derived from the entropy and free energy values given by G.83 volts from pH 0 to 14). Recently.820 CHEMICAL MODELINC IN AQUEOUS SYSTEMS iM 36.80 + 1. Brown's results (.. Staples. Differences between the two data sets are negligible and the final selection was from Baes and Mesmer (51) because data on more complexes are found there..1. discrepancies occur between different Investigations because the solution equilibria are very complicated: several strong complexes are formed and attempts are seldom made to account for the effect of hydroxide and sulfate complexation of the cations involved on apparent solubility. Clifton (personal communication.^ • ^ . The log K for natrojarosite. These composite ligand groups are referred _to as fulvate and humate. Pb2(0H). Stand. practically irrespective of the fluoride concentration.59. Of the four investigations made on jarosite. (63^). (13). Z + 1. there appears to have been a computational "error in converting the stability data of Dutt and Gupta {56} to scandard free energies of formation. 1978) agree very well with the original al^. Baes and Mesmer (51) have Indicated that although the formation constant of A1(0H)| is only known from one measurement (52) and has a large uncertainty. The aBSociation constants and enthalpies of aluminum and Iron hydroxides have been evaluated by comparing the critically selected data of Baes and Mesmer (51) with that of R.3Cl) did not agree with those of the original authors (52).for the existence of Fe(OH)^ and reconfirmed the value of the log K..8 + 1. The model defaults to molecular weights of 650 and 2000. BALL KT AL. . AsOjF^" and H A B O F~. (11).1. li) for the enthal. i Baes and Mesmer ( _ 5 l ) also suggest that the log K for Fe^"^ + SH^O t Fe(OH)° + 3H* (5) . al. Reported molecular weights for these substances vary widely (S. personal communication. Equilibrium log K^ values calculated from free energy data (41) for cwo lead hydroxy chlorides (PbOHCl.0 . the model defaults to data from Smith and Martell (^6) for oxalic acid (10). 1976).-3. However. Others. Byrne (54_) and'Kester .. Computerized Chemical Model 821 makes a fairly complete inventory of aqueous fluoride "coiiplexea.28 kcal mol'^ by utilizing the data of Zotov et al.

(76) for the reaction: 2Fe(OH)2(s) + Fe 2+ + 20H + Fe2(0H)g(s) (12) and log K values for the ionization of water and ferric hydroxide dissolution. For the log K and AH° of potassium alum solubility. having alog K of 3. a bestfit line to field samples from the Sierra Hevada is described by the equation (75): log K .). we calculate a log K of 20. and phlogopite (KMg^AlSisOio(OH)2) dissolution were taken from Plummer _et. SrFj) have been added to the model. x is the mole-fraction of silica and is equal to 1. controlled by the mole .r 0. Unfortunately. cuprous ferrite and cupric ferrite.47. Although a total uncertainty of ± 0. 2i) so that the uncertainty in the log K at 298. (11). Thermochemical data for the ferrous' chlorite.5 and -11. (13) (14) .7H2O + O.i' V.5 was assigned to the log K (to include analytical and computational errors).. fraction of silica and dissolved by the reaction: [Al(OH) 2] ^^_^^ [SiO^.-21. This expression describes the linear variation between pure amorphous hydrous alumina and silica as a function of pH (75). yielding an expression for the equilibrium solubility (75) of: log r .24'. but have as yet not been found to be equilibrium controls on copper or iron solubility. 'Solubility calculations were added for two allophanes. He interpreted this as an indication that a metastable aluminosilicate more soluble than kaolinite was controllit^ the concentrations of alumina and silica in these waters.]^(a) + 3 (1-x)H"*" .15"K Is +0. 7H2O . they are less likely than their respective' sulfates or carbonates to be solubility limiting phases. •. The calculated activity products for the two minerals-.e.SCl".. The enthalpy of dissolution is 2. (4^) who used the free energy values of Eugster and Chow (72) for greenallte' and Bird and Anderson (73) for phlogopite.04 + 0. In fact. This aluminosilicate was further hypothesized to be of varied-chemical composition.7010^3) which has been seen to precipitate from sea water.-5. We prefer the value of .7H'^^ Fe^"^ + 2. However.86 kcal mol*^ based on iH| . offering evidence that most iron (II) in reduced soils other than acid sulfate soils is present in this form. The log K and AH" valuea in Ball et_ al. The revised log K (10) agrees quite well with the upper limit of fluorite ion activity product calculations of many.05 (51) for the reaction: Fe(0H)2 . for which the equilibrium constants and formulae are a function of pH. JO i Bledermann and Chow (77) describe a ferric hydroxy-chloride (Fe(OH)2. values were obtained from the free energies and enthalpies of Wagman jEt A t . Ponnamperuma & Z _ al. al. (17. are characteristically severalj orders of magnitude oversaturated when compared to their respective equilibrium constants in a wide variety of surface waters.0 (19-..8 ± 0. (41) and Parker et al. FeSOi. Cl-x)Al^^ + xH^SiO* + (3-5x)H20 . Paces (74) found cold groiind watera collected from springs in granitic rocks' of the Bohemian Massif of Czechoslovakia to be supersaturated with respect to kaolinite while being unsaturated with respect to amorphous silica.56 from Ponnamperuma et al. t I (9) I i s In equation 9.822 CHEMICAL MODEUNC IN AQUEOUS SYSTEMS 36.jCl^ ^(s) + 2. its solubility and related thermodynamic properties are not well established. (10^) for the dissolution of epsomite have been obtained from the free energy arid enthalpy data given in Wagman et fl. At this level of deviation the analytical and computational uncertainties inherent in the calculations of the ion activity product are likely to be greater thati those in the thermodynamic properties.25.' BALL ET ALi Computerized Chemical Model 823 We have observed melanterite.52pH. (10) Under field conditions the solubility of this material should be lower due to the large difference in the speed of crystallization of amorphous alumina versus amorphous silica. Other alkaline earth fluorides (BaF2. more recent investigations indicate that the log K falls between -10.68pH.135pH. (11) Copper ferrites have been Included In the model. geothermal waters in the western United States. The equilibrium constant for this aubstance waa calculated by combining two endmember constants from the literature and incorporating the pH-dependence equation into the resulting expression. i.4 + 1. (76) describe a ferrosoferrlc hydroxide (Fe3(0H)a).5 (65) which results In a log K of -2.1 kcal mol"^ for Fe^ from Larson and Hepler (65). The log K for melanterite dissolution has been derived from the free energies of formation of the constituent species ^nd the greatest source of uncertainty lies with the iG| for Fe? .7 + 1. greenallte (Fe3Sl205(OH)(t). '20.-5. (41) and Kelly ez_ al^.222 for the reaction: Fe-(OH)-(s) + 8H'^ Z 2Fe-^ + Fe^"*" + 8H-0 . Using a log K of 17. 2 to 3 • ' orders of magnitudei Nordstrom and Jenne (50) utilized simultaneous multiple regression analysis (6S) to evaluate these thermochemical data. to be one of the common sulfate minerals produced by the oxidation of pyrite during weathering. Others. Fluorite.-22. The fiolublllty and related thermodynamic properties of fluorite have had large uncertainties.

the platinum electrode provides a reliable and accurate estimate of the ferrous-ferric redox potential.8PO^" . Stauffer.„ 3+1 in acid mine drainage waters. of the species. In order to evaluate the effect on the modeling calculations of errors in the analytical input data.71 has also been added. Fe(total) is ignored. BALL ET AL. Dissolved oxygen is all that is required for calculation of both the H2O/O2 and H2O2/O2 couples. which performs the mass balance calculations separately for Fe(II) and Fe(III) when they are input separately. the option of using pE from dissolved oxygen for redox speciation has been dropped from the model. the input E^ value la used to distribute all redox speiiles including Fe ^^ and Fe^*. fulfilled. p. (79) haye calculated an equilibrium constant for amorphic silica which best fits their field data. named SPLIT_IRON-.) it.303 (i.1 no membrane) and preserved for ferrous-ferric iron analyses. 63-64).i. Morey _e£ sl.Ax + By (16) I where AP is the activity product. x ahd y are the log activities of the constituent ions In solution. the reactions of the iron section have been extensively rewritten (10) and a protedure.824 CttEMICAL MODEUNC IN AQUEOUS SYSTEMS •4 Jii 36. haa been added.2CO3 + 2.thermodynamic data.. -An E^^ value is calculated from the computed activities of Fe^'+' and Fe^"^ and may. Redox Couples.4 I •I These reactions have been added to the model. As alluded to above. Recent studies (81) show that when the following three condltiijns used to compute the potential of the ferroua-ferrlc couple. E.2F2. so that the relative standard deviation of input data and computed activities are. 1) large voluraes of water inust flow past the electrodes during esnf measurement.^. From Bevington (82. data. The basic equation for the activity product of a solid phase is: logloAP . under this definition. Arrangements have also 'been made to enter and output standard deviations for . if there Is only Fe(total) input. speciation Is done among Fe(II) complexes only. all Fe calculations are bypassed. 8(B) t g. a second equation is required tor x and y as the logio Pf -another number. which in practice is the sum of the logm of the activities of the constituent Ions: a'' • • A^0^ + B^o*^ logioAP X y • (17) Unfortunately.48F' + 4. The H2O/O2 couple is klnetlcally inhibited and is grossly out of equilibrium except at elevated temperacures (80) . J.and no input E^ value. 1974) show that when a flow-through cell is used for emf measurements of geothermal waters. equation 18 becomes X . Jenne. and a and b are weighting constants. p. If only Fe(total) Is input. 1414 (C03)i. and As^'''/As^ . and 3) activities of Fe^ and Fe'^ rather than concentrations must. rather the kno'wn quantity is the standard deviation in the input analytical valuei. A.36Hg. 5Na. but we have defined the propagated standard deviation to be strictly in terms of errors in analytical input data.4 for the fluorocarbonato apatite reaction: Cag. S^"/SO?'.logio(+bu) and equation 19 becomes: 5 c r (18a) x • 2. (15) .W. The log K for this reaction of -2. If either Fe(II) or Fe(III) are present. The conditions are'. ES 2 f . HjO/Oa. respectively. unpub. 1/ln (10)). the platinum electrode potential mirrors decreasing dissolved sulfide and then increasing dissolved oxygen values in the downstream drainage of individual hot springs. The model calculates the redox potential of the couples. equal: y V u u (20) . by user option.SCa^"^ + CSSNa"*" f O.e. the atandard deviation in the logio of the activity of the species is not known. given the requisite concentrations of the couple members. Other studies (R. If a Is set equal to 1/2. E. . NOi/NOj. To accomplish this. if Fe(Il) only Is present. Standard Deviations.a(P0i. input data for total iron. 2) water sainples must be properly filtered (<_ 0. and likewise for Fe(III). From Bevington (81. with solute modeling calculations done using whatever data are input. Therefore. H2O2/O2. Computerized Chemical Model 825 Chlen and Black (78) caiculate a log K of -114. A new procedure named ERRCALC was written to calculate the propagated standard deviation. prop agated standard deviations are now computed for a subset of the solid phase activity products considered in the model. Therefore. for: O-a X X " a Ih^bu) (18) ~ tl9) where u and x are the activity and logio of the activity.303 u u (19a) The standard deviation in the activity of the species Is also unknown.UiMg^"'' + I. Fe^^/Fe^ . be used to distribute other redox species in lieu of an input E^ value. Ball. and A and B are the number of lona of each species making up the solid. Fe(II) and/or Fe(III) are accepted by the model. 60) we construct the following equation to calculate the standard deviation in the logio of the activity product.

and 5) improved set of headings used in the printed results of the solute modeling calculations. For certain expansible layer silicates and two zeolites. data using the method of Laxen (83) has been added to the model as a check on analytical input data.^g2+]i + ANa+ + Air+).303V +B Ow 2. . . . These data rarely contain either sigma values or the data from which they can be calculated. 3 ) has been retained for the major ions. 2) printed table of log K values for which analytical expressions exist. M g .. BALL Err A L . Most of the preexisting reactions which include OH were similarly modified to conform to this convention. (88) presented at this symposium. A mass balance section for the hydrogen sulfide species was added to the anion mass balance calculations when we observed that strong HS complexing of some trace metals sometimes rendered cation mass balance convergence impossible. noted below. Cl and SO14. There are inconsistencies in the model for the calculation of activity products for the "clays.. If the reactions are written in terms of OH . in the analytical input data." .thermochemical data (10). The uncertainties ascribed to the thermodynamic data are labeled SIGMA(T). the usual upper limit for extended Debye-Hiickel calculations. The Davies equation is considered satisfactory to 0. N a .ct^ jal^. Belle Fourche montmorillonite and Aberdeen montmorillonite ( A H + + A^g+ + A]j+). v is the analytical concentration of the species from which the activity of the species was derived and o^ and Oy are the respective standard deviations. All new reactions which include OH are written in terms of H2O and H . Activity coefficients were originally calculated using the extended Debye-Huckel equation and whenever a new complex was added to the program it was necessary to estimate the a parameter. Modifications in the Code The PL/1 language computer code has been extensively altered in the process of building it into WATEQ2. Several errors in the original code were corrected and some major changes. and the Debye-Huckel equation is used to calculate the polysulflde activity coefficients. 3) check on the charge balance computed from input data. 4) revised anion mass balance calculation. errors in the dissociation of water and in its temperature dependence are encountered. following the listing of the input . Substituting from equation 20 to equation 19a we obtain: 2. However. This problem was overcome by substituting the more general Davies equation which has adequate reliability at low ionic strengths and is usually more accurate at high ionic strengths (84). Therefore. and moderate capacity illite. Note that the square root of the divalent cation is used in the sum in keeping with the practice in the ion exchange literature (86).826 C1IEMICL\L MODEUNC IN AQUEOUS SYSTEMS 36. Since acid mine waters can have ionic strengths approaching that of sea water. chlorite. Computcrtzcd C h e m i c a l M o d e l 827 where u is the activity of the species. The sigma values thus computed are printed along with the results of the activity product calculations under the heading SIGMA(A). allowing for. we have resorted to use of twice the uncertainty resulting from rounding of the analytical values as a minimum estimate of the analytical uncertainty.Exchangeable cations are disregarded for the low exchange capacity kaolinite. as well as that of Bassett . Ca. Revision of the calculation of exchangeable cation contribution to the activity product has been delayed pending the pertinent reviews of Kittrick (87).303W (17a) where a„ and w are the analogously obtained values for Oy. are beidellite ([Aj. Other Modifications. Other features of WATEQF (£) which have been incorporated into WATEQ2 include: 1) revised set of analytical expressions for the effect of temperature on the stability constant ( 1 0 ) . for which a parameters have been estimated by Cloke (85). and the solute cations considered as exchangeable cations.5 molal. Sigma values are routinely obtained for analyses performed in our laboratory. K.. clinoptilollte and mordenite (Ana+ + Aj^f). These errors are thus avoided since k y ^ is a directly measured quantity. the uncertainty ascribed to the analytical values used. following the convention of Baes and Mesmer (51). were made to improve program execution and ease of use .1 molal. This also has the advantage of making the propagated standard deviations for different solid phases more comparable than if some are written in terms of OH and some in terms of H2O and K^. the logio "f 'he activity of selected cations is added into the sum of the activity products.303 (19b) Deriving an analogous expression for o in terms of w and substituting it and the right side of l9b into 17 we obtain: ^Log^^AP Ov 2. with aborting of the calculations if the balance is off by >30Z. in fact. These are not distinguished in the model from true sigma values. • — . it is desirable to use a theory for activity coefficients that can reach somewhat above 0. we are utilizing a considerable amount of literature data to evaluate the role of various solid phases as effective solubility controls. . minor alterations are far too numerous to mention here. The extended Debye-Huckel equation with fit parameters (2^. Differing input and calculated specific conductances indicate that one or more errors may exist. halloysite. Specific conductance calculated from input major constituent. faster convergence. The mineral phases treated in this manner.

of the logioK value (Z) to experimentally determined values which may represent a less soluble or more soltible form of the solid phase than that selected as the "best" value. SIGMA (Thermodynamic). la now compared to-a computed "specific conductance" as a qualityo. calculates temperature-dependent coeffieients for the Debye-HUckel equation. Burchard. BALL ET AL. The Input and output aspects of prograa operation are given. MAJ_EL.. In a supplementary report (10). a warning-message is printed and execution continues Just as through convergence had been reached. and computes and prints the saturation state of each mlrieifal with respect to a thermodynamic equilibrium constant. R. Siebert and . Geological Survey. OUTPHCH generates a card deck of a subset of the calculated activities. prints out all the solute data. Ba. Specific conductance. for their careful reviews of this manuscript. primarily for plottl-ng using a subsequent computer program. M. WATEQ2 consists of a main program and 12 subroutines and is patterned similarly to WATEQF (4^). S. discussed previously. Zn. calculation of redox potential from various couples. changes in the poasible operations with Fc (II. for each reaction considered. polysulfides. Ni. W. As. uses input analytical standard deviations to calculate the propagated standard deviation in the ] f l g of the activity products for a subset of minerals considered. The results of solute and mineral calculations will not appear if the activity or activity product. respectively. calculations. and calculates the equilibrium constants as a function of the Input temperature. _58) and new compilations (51. ' APCALC calculates thermodynamic activity products for the various mineral species considered by WATEQ2.. and II + III). mineral equilibrium calculations has been drastically altered. SIGMA (Analytical) . WATEQ2 (the main program) uses input data to set the bounds of all major arrays and calls most of the other procedures. for many helpful diacussions. as the model evolves. I t i subsequent calculations. ERRCALC. Abstract The computerized aqueous chemical model of Truesdell and Jones (^. or until 40 Iterations have elapsed. F .Potter II. and fluoride solute sections. Z) . an advantage especially when a simple laboratory solution is considered and/or a low-speed remote computer terminal is utilized.B. F. N. Sr and the major cations. WATEQ j has been greatly revised and expanded to Include consideration of ion association and solubility equilibria for several trace metals. AG^ per equivalent cation. and performs mass actionmass balance calculations on Li. calculation of propagated standard deviation. i Ackno-wledgements We are pleased to acknowledge the efforts'Sf V. INTABLE reads the thermodynamic data base and prints the thermodynaraic data and other pertinent information. log AP/K^JL^. Geological Survey. and updating the model'a thermodynamic data base using critically evaluated values (81. solubility equilibria for various metastable and(or) sparingly soluble equilibrium solids. sulfate. and a mass balance section for sulfide solutes. Hi. 50. Kn. As discussed previpusly. does a partial speciation of the major atilons. both of the U. Revisions include expansion and revision of the redox.. has not been calculated. SO?". and all values in base 10 form. This eliminates extraneous non-Information and abort ena the listing considerably. Cu. and to J. S.. PREP reads the analytical data. TR__EL performs these calculations on the minor cations. Pb and Zn. and teats the resulte_againat ingut concentration values for the anions COj". R.f-analysls screening technique. and log AP/K„^jj. P0( . and As.the six anions. SUMS performs the anion mass 1 I action-mass balance calculations. SOLUTES performs computations not related to the mass balance calculations. Rb. If convergence is not reached in 40 iterations. M. 2$^. III. which was initially carried-only data. specific conductance. Plummer and. Cd. Computerized. Ag. iron. MAJ_EL calculates the activity coefficients and. and performs necessary logarithm conversions for use .. Ag. SET initializes values of Individual species for the iterative mass action-masa balance calculations. RATIO calculates and prints mole ratios calculated from analytical molality and log activity ratioa. boron. aetivity products and Input. euch as E^. and tests for charge balance of the input data. The listing of the results of the. Cs. along with the thermodynamic data base. to . Cl and S^~. whose help in many aspects of this work is especially appreciated. Cd. SIGMA(A) la the propagated standard deviation in the analytical valuea and SIGMA(T) Is the standard deviation in the thermodynamic data. concentrations for subsequent use with plotting programs. Information now printed for each species for which an activity product is calculated includes log AP/K. and prints the reaults of each set of iterative calculations. with the deletion of iG^.the model itself or to adjunct calculations.1% of the input concentrations is achieved for. Arrays which miist be increased in alze when species are added are now automatically adjustable in dimension merely by supplying appropriate Input data. U.828 C H E M I C A L MODEUNG IN AQUEOUS SYSTEMS •I 36. However. Chemical Model 829 and to broaden its usefulness. The logioKniin and logio^niax values have been changed from + 5%. converts concentrations to the required units. some of these carried-only data become input to. Fb. pC^and pCHi. on the first Iteration only.. Cu. such as analytical expressions for effect of temperature on selected equilibrium constants. Some data are entered into the model as "car'ried-only" data. Mn. TR_EL and SUMS are executed repetitively until mase balance to within 0. Jones. PHASES prints the results of the activity product and error calculations.

1974. Survey J. Spec. 29. G. 1978. 22. R. H. N. C. M.. 76-13. Amer. "Modem Electrochemistry. (1964). (NTIS PB-215 899) (1973).. 21. G. . Inorganic speciation of silver in natural waters . ed.. T. 11. M." 478 p. Geol. Y. 754 p. 78-116 (in press). E. ^ Klein. H.] Amer.. and Jones. Ivan. Survey Prof. Union Trans. [abst. D. B. 1970-1971. Pluramer. Geol. Christ. and Vasll'ev." 328 p. ed. Ball. W. 2695-2698 (1950). E. J. Naumov. Acta. a computer program for calculating chemical equilibria of natural waters: U. 4^. 237-240 (1975). 1: Chem. Weast. Truesdell. New York. 69.830 aiEMlCAL MODEUNG IN AQUEOUS SYSTEMS 36. 592-600 (1930). A. F. p. C . Jenkins. D. (Frankfurt) 19. 23. and Davies. The thermodynamics of aqueous hydrofluoric acid solution. Soc. LiSOi^.. 16. Dlseociation of Na2S0^ from ultrasonicabsorption reduction in MgS0i. M. Faraday Soc.. C. J. J. 26. "The Thermochemistry of Transition Metal Complexes. Pytkovicz. 362 (1971). Jones. No.. 14. H. Broene. and Reddy. HSO14. U. C. No. Nigrini. Calorlmetrically determined Log K. J. and Jones. E. 47-53 (1969). NaSOi^ ion pairs In aqueous solutions at pressures up to 2000 atm. Inorganic Complexes. Soc. Sci. Girvin. 66. Cosmochem. 17. S.. Spec. O'M. 1976. S. Righellato. N. SOLMNEQ: Solutionmineral equilibrium computations: U. Geol. Christensen. C . A. NaSOH. No. The conductances of sodium." NTIS Tech. New York. Chem." 2i8 p. Invest. Paper 498-D. Ohio. "CRC Handbook of Chemistry and Physics" Chemical Rubber Co. V. A. Sol. and DeVries. A. and Jones. The extent of dissociation of salts in water. Pa. 73-002. and Martell. T. Soc. 267. I. Chem. and Kester. A. Phys.. F. 2(2) 233-274 (1974). Report PB214372.. Invest. Sillen. Plenum Press.. Z. New York. Amer. Geochem. Publ. 8. and iS' values for the interaction of sulfate ion with several bland tervalent metal ions. and Martell. R. Kharaka. "Handbook of Thermodynamic Data. Hamed's rule behavior . A. Geol. WATEQF A Fortran IV version of WATEQ. (1971). Trans. A. Nathenson. 2 1 . G. J. Additional and revised thermochemical data and computer code for WATEQ2— A computerized chemical model for trace and major element speciation and mineral equilibria of natural waters.. C . B. Helgeson. p.' 5.. "Critical Stability Constants. 1.. E. Dowder. Amer. 4. E. M. J. 24. Fisher. 88 p. P.. Academic Press. and Truesdell. Smith. Ion association in ferric sulfate and nitrate solutions at low Fe III concentrations. and Monk. 865 p. Chapter 4. J. 41-61. M. J. A. E. A. 729-804 (1969). J. 13. 1960. (Tr. B. J.. 12. Helgeson. L. Unl-bivalent salts. T. Soc. and Barnes. Soc. III. Survey Water Res. 26. 4^. R. B. E. 27. 1970. H. 10. D. H. S. G. 28. F. Jenne.. K. L. Stability of metal complexes in solution. Amer. Eatough. E. L. 1970. Lafon. 51st Ed. U. 25. C.. a computer program for calculating chemital equilibria of natural waters: NTIS-PB2-20464. Ryzhenko. Standard enthalpy_of formation of complex sulfate ions in water.~Paterson). 1289-1293 (1977).. L. and Nordstrom. C.. a computer program for calculating chemical equilibrium of natural waters: U. I . A. D. M. Literature Cited Odum. A. Computerized Chemical Model 831 C. WATEQ. Pergamon Press. Stability constants of metal ion complexes: Chem. 156-167 (1959). and Fox. F. of NaCl-Na2S04 solutions explained by an ion association model. Mattoo. I 20.fresh to marine. Geophys. and Martell. GeOl. 6. R. J.. WATEQ. Survey WTU)-74-001. S. Mechanical revisions include numerous operational Improvements in the computer code. B. 52.'"Environmental Impacts of Nucleating Agents Used in Weather Modification Programs. Hutchinson and Ross. 29 p. 58 p. I. 217-229 (1969).. Chem.'569-592 (1970). B. Fisher. B. (1969). Chemical properties of ground water and their corrosion and encrustation effect on wells. Thermodynamics of hydrothermal systems at elevated temperatures and pressures. J. 267. "Instability Constants of Complex Compounds. and Truesdell. and Mortimer.. H. H. 519-521 (1962). P. Ivan and Clarke. 34. U. unpublished data 1976).. Part II. Chem. Sci.. from Rues. (1973)." 256 p. London. p. W. H. 2. D. J. 25. J. 225-236 (1975). 9. S. B. A.-NaCl solution.. AH"." 1432 p. Res. K. C . NTIS PB 226 722/AS. H. 77 p. and Burchard. Plenum Press. A. L. .. Geol. K. 18. Suppl.. M.. BALL BT AL. "Thermodynaraic calculations. T. W. 61 p. 19. K. A. Brown. Chem. and Bartholomew. Chem. S. Stability constants of metal ion complexes. Ball. P. D. V. Sol." 257 p. F. Sillen.. Truesdell. A. J. 17. A. Publ. R. 3. L . J. The-emergence of ecology as a new integrative discipline: Science 195. and Khodakovsky. Ashcroft. New York. H. A. 1644-1646 (1947). Survey Water-Resour. (1976). Calculation of mass transfer in geochemical processes involving aqueous solutions. F. Izatt. Yatsimirskii. and Malr. H. N. J. potassium and lanthanum sulfates at 25''C. Stroudsberg.. Jenne. Chem. Temperature dependence of sodium sulfate complexing in aqueous solutions. (1973). Bockeris. D. N. Cleveland. 4. Phys. B a m e s . H. 15. Survey Water Res. Austin.. S.

215-229 (1960). Ellis. E. 1009-1013 (1968). A. Soc. 1977. Washington. Brown. J. J. 105-121. Acid catalysis in hydrofluoric acid buffers. Russi J. G. Hi E.-C. Barthe. P. Geol.". John Wiley and Sons. Paper 650-B. Geology and respurces of fluorine in the United States. 41. A chemical study of some synthetic potassiumhydronlum jarosltes. Chem. Chim. 56-79. Acad. F. E. 1976. 202-210. E. Hydrothermal alteration zones caused by halogen acid solutions. Univ. New York. Deschamps.. J. H. Chem.. Amer. P. Byrne. W.^ Cosmochim. and Barnes. i n Gibb. Jr. Vinyl ether hydrolysis. Chem." 458 .. Survey Prof. Sci. 251^270 (1966). 0. Paper 933_. Chemistry of the reactions between the. Moreno. Nat'l Bur. W.. ed. Bailey. 48. U. Chemical processes in hydrothermal systeos-a review. V. and Jenne.. Patel. A. "Iron speciation and solubility in seawater. Stability of fluoride coraplex with silica and its distribution in natural water systems. Geochim... Stand 74A. Jr. ed. Y. (1968). S o c . Jr. Parker. 53. Bur.. D. 21. Utah. 696-703 (1970). Mineral. Kamada. Stand 75A. Symp. Nat'l. Tech. . ed. arid Malkovsky. J.. D. 33. "PtoceedingB of Symposium on Hydrogeochemistry. 249-252 (1961) . Nazarenko. Acta 41. I Vasll'ev.. and Liang.. H. R. and Gupta. R. M. Nordstrom^ D. T. p. 57. 1544-1546 (1973). Fluorarsenates and their analogues with sulphates. B. The Gibbs free energies of substances in the system Fe-02-H20-C02 at 25°G.-J. Rhode Island. 55." Ph^ D. D. Vanderbotgh. Chem. 59. Yoshlda. Nordstrom. S. R. Res. Equilibria of fluProsilicate solution with special emphasis to the fluoridation of public water suppLies.ioris of multivalent elements and organic reagents. J . R. 77.. S. E. Anal." Amer* Chemi Soc. 822-825 (1973). S. T. J. Res. 100-102 (1969). Wagman. M. Ser. Mineralog. N. 35. Dutt. 1973. and Chiang. Note 270-3. Table for the first thirty-four elements in the standard order of arrangaaient. Mono. A. and Sanemasa. Roberson. Washington. 18.. 43. 54. . L. 5309-5310. "Analytical Methods l i i Oceanography. Chem.. Soe." Amer. M. and Chiang. M." Ph.. Cadek. Amer. Baes. A. of an acid mine drainage systems. R. Spectrophotometric determination of silica at high concentrations usin-g fluoride as a depolymerlzer. 1 atm. J. 37.effect of temperature on the ionization of hydrofluoric acid. Chem. C. Hamer. A. H. E. A. Amer. E. U. Solvent Isotope effects on the ionization of hydrofluoric acid. 239-256 (1978). 264 p. E. Ellis.-. ed. Baumann. 36." v. "Marine Chemistry in the Coastal Environment. Ionization of hydrofluoric acid at 25'C. B. 40. and Wu. Y. 58.. 31. Inorg. : Warren.. p. 39. J..p. D. H. Nucl. R. XVII.. Martin. J.. 34... Shapiro. (1976). Genesle of acid hot spring.. Clarke. L. 162-163 (1951). Ozawa. and Nevskaya. J. Lovering. in Ingerson.. Survey Prof. Chem. S. Catalysis in dilute hydrofluoric acid solution. and Charreton. and Mesmer.. 1. (1968). 51. 11. 44.. 56. Y. S. "Proceedings of Symposium on Hydrogeochemistry. Thesis. in Ingerson. Selected Values of chemical thermodynamic properties. 50. ed. 46.v.. Byrne. Stanford University. Chem. Crosby.*W'' CHEMICAL MODEUNG IN AQUEOUS SYSTEMS 36. M.. B. Langmuir.. H. J.. -Survey J. Kresge. T. B180-B184 (1969). BALL £T AL. 761-768 (1970).. J. 199-202 (1971). 38. Kingston. Determination of stability constant bf hydrogen and aluminum fluorides with a fluoride-selective electrode. The activity coefficients of hydrofluoric acid in water from 0 to 35°C. in Church. T. p.. V. N. Kresge. Thesis. Carollnae Geol.I. Anal Chim. A. 60. C.geothermal waters. Pierre. D. 4300-4304 (1963). D.. 2814-2817 (1972). 205-211 (1971). V. Part III. The.. D. R. J. V. Fluorite solubility equilibria In selected. and Taft. E. Geol. Geol. K. AS. Shawe. 1975. J. 47. 1975. Thermochemistry of acid-base reactions in aqueous hydrofluoric acid solutions. 94. 31. H. T. W. Fla. A. Evaluation of the lanthanum fluoride membrane electrode response in acidic solutions. C. 49. Geol.. Soc. 670-67-3 (1969). "Hydrogeochemical and microbiological factors affecting the heavy metal chemistry. K. F. H . J. D. 3155-3162 (1969). Amer. Acta 53. Bur. and Kozlovskii. Y. 52. R. I. 90. J. D. Redox reactions and solution complexes of iron in marine systems. . Interaction of aluminum and gallium ions with aethylthymol blue. . Appl. and Patel. Sci 258-A. Y. J. Kresge. 99 p. Res. Talanta 15. 19. "The Hydrplysia of Cations. USSR 24(6).. Indian Chem. Ser. Comptes Rendus 232. 2> 357-360 (1974). 1. S. 1-26. and Shepardi A. Stand. Evans. Inorg. Determination of the solubility products of the hydroxide and basic chlorides of lead.. V. Soc.. Computerized Chertiical Model 833 30. Stability of fluorarsenate ion and its comparlBon with fluorphosphate ion. 42... Brown. 32. C. 1973.. 10. M. P. K. D. Nat'l. 175-188 (1977). J. R. •U. Phys. J. Clarke. Transport of fluorine in natural waters and precipitation of fluorite at low temperatures. Occurrence. -4. 147. East Tintic District. I. The measurement of pH in acid fluoride acid solutions and evidence for the existence of (HF)2. 245-252 (1971) . Kester. p. Ghem. 1974. E. and Schuimn. Halow. Jarosite-goethite stabilities at 25'C. B.and implication of sedimentary fluorite In Tampa Bay. and Chiang. Deposita 1. J. J. R. M. Acta Univ.

0. 141 p. M. A. I. J. 64. Stand. Tech. 1969. C.. Soil Sci. The hydrolysis of Iron(III) lon_^and_ the solubility product of Fe(0H)2 7oClo. Tables for the alkaline earth elements (Elements 92 through 97 in the standard order of arrangement): Nat'l Bur.. L. 20. G. 87.. V.3oln 0. S.Solubility. 55. 83. Kittrick. 361368. "Data Reduction and Error Analysis for the Physical Sciences. New York. 103. L... 72. 79. P. Mines Tech. Geophys. 80. and Kinetics. and Kinetics.. Redox equilibria of iron in acid mine waters. Simultaneous evaluation and correlation of thermodynamic data.5 M (Na )Cl" medium." 624 p. D. Ponnamperuma. Oxidation of sulfide ore bodes. and Brandt.. Geochim. Halow. 273. New York. The geologic role of polysulfides . 65. C. 81. Cosmochim. "Ionic Equilibrium. Determination of the Gibbs free energy t G l _„.'Ahalytical Calorimetry. H. R. 71. J. Econ. 234-239 (1976).834 CHEMICAL MODELING IN AQtn»US SYSTEMS 36. (1971). R. Y.Part II. Shomate. S. K." 336 p. and Bates. and Kinetics. Jr. 1964. 739-745 (1973). Geol. ig. 85. W. R. Acta Chem. 62. Kharaka.of jarosite synthesized from a natural solution. J. Helffrich. (this volume). W. K. Solubility. Tomas. Butler. F. Solubility product constant of calcium fluprlde. 525-545 (1976). ed. J. S o c . P. AGf. Kashkal. Stearns. 38.. (1946). G. W. Parker. Motoakl. i a '. A. Proc. 374-382 (1967). J. and Fisher. Speciation. E. W. K. Tables for elements 35 through 53 in the standard order of arrangement. Tomas. I. P. J. R.-M. A. M.S. Bird.. J. 496-498 (1977).. . ' ) Paces." Amer. Wagman. D.. 1968.. L. 63. and Langmuir. Sci.. 429-432 (1974). 1. B. G. Sorption. H. Acta 37.. Paces. J. M. ed. Larson. A.. Chem. Speciation. E. and Lindsay. J. 74.... Speciation.. 84-91 (1973). Macaskill. 68.. Soc. Beek. G. R. Phys. J. Computerized Chemical Model 835 61. Cloke. Parker. New York. Rusinov. J. ed. and Anderson. "Chemical Modeling in Aqueous Systems. E. and Evans. F. 195-201. and Huffman." Plenum Press. Studies on the hydrolysis of metal ions. C . Sci. A. Sorption. McGraw-Hill Book Co. 2641-2663 (1973).. J. S o c . L. p. Res.. N. Solubility product variability at constant temperature and pressure. p. Sato. RECEIVED November 16. ' Beidermann. 81. D. Sorption. The solubility of acanthite and covelite in sodium poly. C. and Ball." 1973. G. Geol. A. Blom. J. Note 270-4. Laxen. 69. 73." 547 p.' D. J. 75. Geokhimiya. w. and Roberson. S o c . D. Missouri. and Rowe. Selected values of chemical thermodynamic properties. H. A Mathematic Approach. Amer. 13^6-1388 (1966). Acta 27. and Hepler.. Ion exchange and mineral stability: Are the reactions linked or separate? I j i Jenne. 928-961 (1960). H.298 determination of synthetic jarosite and its sulfate analogues. 76. W. B. Zotov. J. Survey J. 66. and Black. Water Res. "Trace Substances in Environments Health-VI: Univ. Soil Sci. in Hemphill. L. S. 509-517 (1977). J. Salo... 84.. Cosmochim. v. Naylor. 86. 82. Stand.. Free energy of formation of carbonate apatites in some phosphate rocks. W. sulfide solutions. Paper 688'. Bailey. M. P. "Chemical Modeling in Aqueous Systems. ed. . Y. Bassett. D. Geological Survey. Redox equilibria in flooded soils: I. N. H. A. E. 778-784 (1975).. Amer. T. D.. Soil Sci. Evans. Brown.. E. Scand. Kelly. McGraw-Hill. P.. Res. Active mineral surfaces: origin and possible effects on trace elements in natural water systems. Thermodynamic properties of ammonium and potassium alums and related substances with special reference to extraction of alumina from clay and alunite. D. Geochemical environments in terms of Eh and pH. G. Wagman. BALL ET AL. "Ion Exchange. The depositional environments of precambrian banded iron-formations. (1969). B. Borovskaya.. H. and Schumm. Econ. and Chow. U. Amer. Teresita.. B. A specific conductance method for quality control in waters. Soc. D. Note 270-6. L. 1995-2002 (1964). 1299-1319 (1963). 67. Mlronova. The iron hydroxide system. Selected values of chemical thermodynamic properties. Friederich. Eugster. 77. Phys.. 69. Critical review of the equilibrium constants for clay minerals. Disclaimer: The reviews expressed and/ or the products mentioned in this article represent the opinions of the author(s) only and do not necessarily represent the opinions of the U. Chem. Determination of the solubility product of various iron hydroxides and jarosite by the chelation method. Chem.. I j i Jenne. Haas. Solubility. A. 91-94 (1977). Part 57. Steady-state kinetics and equilibrium between ground water and granitic rock: Geochim.. "Chemical Modeling in Aqueous Systems.. Young. Nat'l Bur. F. V.. L. G. M. Chem.. J. Morey. Amer. (this volinne) . A. 104 p. 70.. 1060-1063 (1976). New York. H. S. i n Jenne. R. J. Jenne. and Chou.." Amer. K." Amer. 40.. and Loy. A. Solubility of natural fluorite at 25''C. Tlanco. Vlek. The free energy of formation of magnesium cordierite and phlogopite. V.. and Babozade. S. 5. G. 78.. Chem. Calorimetric measurements dn metal sulfates and their hydrates: electrode potentials and thermodynamic data-for aqueous ions of transition elements. 68. D. (this volume). 88. Columbia. Geol. R. Bur. Geokhimya 5 . Addison-Wesley.. 1144-1168 (1973). 11. Tech. 276. E.. Chlen. 106 p. W. E. E.. 1962. E. U. B. Nordstrom. Th. I.. H. D. Solubility of amorphous Si02 at 25'C. Bevington. Foumier.1978.