Unesco technical papers in marine science 37 Background papers and supporting data on the Practical Salinity Scale 1978 Unesco/ICES/SCOR/IAPSO joint panel on oceanographic tables and standards Ungzsco, 1981No. n 1s 16 7 20 n n 2B Pa UNESCO TECHNICAL PAPERS IN MARINE SCIENCE Numbers 2,3, 5,6.7,9, 10, 12, 13, 15, 16, 18, 19, (English) and 25, are out of stock. For full titles see inside back cover. ‘Numbers 1, 4, 8 and 14 are incorporated in No. 27. ‘An intercomparison of some current meters, orton an experiment 3! WHOI Mooring Site “D", 1624 July 1967 by the Working Group on Continuous Curent Velocity Measurements Sponsored by SCOR, IAPSO ant Unesco Monitoring fein uh ocean: sponsored by SCOR, ACHRR, Unesco, IBP/PM ‘Sixth report ofthe joint panel on oceanographic tables and standards, Kiel 24-26 January 1973: sponsored by Uness, ICES, SCOR, IAPSO {An intercomparzon of some coment meters, teport on an experiment of Retearch Vessel ‘Akademie Kurchatov, March-April 1970, by the ‘Working Group on Curent Velocity Measurenent sponsored by SCOR, TAPSO, Unesco Marine Sclence Teaching atthe University Level. Report of the Unesco Workshop on University CCumicula-Aveilale tn French, Spanish and Arabic Iehthyoplankton. Report ofthe CICAR Iehthyo- plankton Workshop ~Alvo publiched in Spenish [An intercomparison of opens tidal pressure sensors, Report of SCOR Working Group 27: “Tides of the open sea” European subrepional co-operation in oceano- ‘pap. Report of working roup sponsored by the Unesco Seiemifie Co-operation Bureau for Europe and the Divion of Marine Scences {An intercomparison of some curent meters, Il Report onan experiment carted out from the Research Vessel Alanis Il, AugustSeptember 1972, by the Working Group on Continuous ‘Velocity Measurements: sponsored by SCOR, TAPSO and Unesco Seventh report of the joint pane on oceano- ‘maphle tales and sandards, Grenoble, ES seeder 1075 sponte by Une, ICES, SCOR, 1APS( Marine science progamme for the Red Sea: Recommendations of the workshop held in Bremeth.ven, FRG, 2223 October 1974; sponsored by the Devtihe Forschungagemeln- shaftand Unesco 1973 1974 1974 1974 1975 1975 1975 ists 1976 1975 scoR we wor wor wo 10 wo2t wo27 wea wo10 6 n 8 » 30 3 2 2 38 ” 39 Marin science inthe Gulf area - Report of consultative meeting, Pais, 11-14 November 1975 CColeced reports ofthe joint panel on oceano- ‘gaphi tables and standards, 1964-1969, Eighth report of the joint pane on oceano- sgphic ables and standards, Woods Hole, USA, sponsored by Unesco ICES, SCOR, TAPS Committe for the preparation of CLOFETA - Report ofthe frst mecting, Pars, 16:18 janvier 1978 ‘Ninth report of the oint panel on oceanographic ‘ables and standards, Unesco, Paris 11-13 September 1978, (Coastal lagoon survey (1976-1978) CCostl lagoon research, present and future Report and guidelines of mina, Dk University Marine Laboratory, Besufort, NC, USA, August 1978. (Unesco, ABO). To be published in 1981 (Coastal lagoon research, present and future. Proceedings ofa seminar, Duke University, ‘August 1978, (Unesco, ABO). To be ‘ublished in Enalich ond French ‘The carbon budget ofthe oceans. Report of smecting, Paris, 1213 November 1979 Determination of ehlorophyll in seawater. [Report of intereliration tests sponsored by ‘SCOR and carried out by C4. Loreraen and SW. Jeffrey, CSIRO Cronuils, N.S, ‘Australia, September October 1978 Tenth report ofthe joint panel on oceanographic {ables and standards, Sidney, B.C., Canada, 115 September 1980. Sponsored by Uneic, ICES, ‘SCOR, TAPSO. [Background papers and supporting data on the Practical Salinity Seale 1978. [Background papers and supporting dats on the Intemational Equation of State of Seawater 1980, International Ocranographic Tables, Vol. 3. 7 ERS 1982 0.8 HAS 1982 Yeu 1976 1976 1978 1979 1979 1980 1981 1981 1980 1980 1981 1981 1981 1981 SCOR we wo10 wo 10 wos wo10 wo 10 wo10 we10Unesco technical papers in marine science 37 Background papers and supporting data on the Practical Salinity Scale 1978 Unesco/ICES/SCOR/IAPSO joint panel on oceanographic tables and standards Unesco, 1981Published in 1981 by the United Nations Educational, Selenite and Cultural Organization, Place de Fontenoy, 75700 Pat Printed in Unesco’ workshops OUnesco 1982 Printod in France Reproduction authorize, providing that appropriate ‘mention itmade of Unesco Technical Papers in MarinePREFACE This series, the Unesco Technical Papers in Marine Science, is produced by the Unesco Division of Marine Sciences as a means of informing the scientific community of recent developnents in oceanographic research and marine science affairs. Many of the texts published within the series result from research activities of the Scientific Committee on Oceanic Research (SCOR) and are submitted to Unesco for printing following final approval by SCOR of the relevant working group report. Unesco Technical Papers in Marine Science are distributed free of charge to various institutions and governmental authorities. Requests for copies of individual titles or additions to the mailing list should be addressed, on letterhead stationery if possible, to : Division of Marine Sciences, Unesco, Place de Fontenoy, 75700 Paris, France.CONTENTS General Introduction Outline of the work and the laboratories involved = Note on the symbols and units used in the Special Issue on the Practical Salinity Scale 1978 of the IEEE Journal of Oceanic Engineering = Introduction to the Special Issue on the Practical Salinity Scale 1978 - Thomas M. Dauphinee - The Practical Salinity Scale 1978 and Its Antecedents - Edvard Lyn Lewis Determination of the KCl Solution Primary Flectrical Standard - The Effect of Concentration and Temperature on the Conductivity Ratio of Potassium Chloride Solutions to Standard Seavater of Salinity 350/00 (C1 19.3740°/oo) = T. M. Dauphine, J. Ancein, H. P. Klein and M. J. Phillips ~ Determination of the Concentration of Potassium Chloride Solution Having the Sane Electrical Conductivity, at 15°C and Infinite Frequency, as Standard Seavater of Salinity 35.0000°/00 (Chlorinity 19.37394°/oo) - F. Culkin and N. Smith ~ The Concentration of the KCl Solution Whose Conductivity is that of Standard Seawater (352/00) at 15°C = Alain Poisson Measurement of the Conductivity Ratio as a Function of Salinity, Temperature and Pressure = The Electrical Conductivity of Weight Diluted and Concentrated Standard Seawater as a Function of Salinity and Temperature - T. M. Dauphinec, J. Ancsin, H, P. Klein and M. J. Phillips Page* 10 a 13 19 21 27 29 35 7 ‘The pagination of this report itself is at the bottom of the page. The pagination of the original papers is at the top of the page,= Conductivity/Salinity/Temperature Relationship of Diluted and Concentrated Standard Seavater - Alein - Electrical Conductivity of Seawater - A. L. Bradshaw ana K. E, Schleicher = Conductivity Ratio Measurenent of Diluted or Concentrated Standard Scasater snd KCL Solution used to Establish the New Practical Salinity Scale, 1978 ~ Alain Foisson Fitting of the Data ~ The Practical Salinity Scale 1978 : Fitting the Data - R. G. Perkin and E. L. Lewis Conversion of the Existing Data - The Practical Salinity Scale 1978 : Conversion of Existing Data - E. L. Lewis and R. G. Perkin Conclusion = Definition and Algorithm of Practical Salinity Page 20 59 73 105 107 115 7 139 141,GENERAL INTRODUCTION In 1964, a panel of scientists were appointed jointly by Unesco, ICES, SCOR and IAPSC to advise on the establishent of international tables and standards. The first task of this Joint Panel on Oceanographic Tables and Standards (JPOTS) vas the preparation of tables for computing salinity of seavater from deterainations of electrical conductivity. Before these tables could be prepared, it was necessary to redefine salinity in terms of conductivity. After discussions (Unesco 1965), the JPOTS recon- “nended (Unesco 1966,a) a definition of salinity based on determinations of chlorinity and conductivity on samples of natural seavater from all the oceans of the world. In October 1966, the "International Oceanographic Tables" (Unesco 1966,b) for computing salinity from conductivity, based on the above definition, were published. They include a tabulation of this definition at 15°C for salinities from 29 to 42°/oo, along with a correc~ tion table for mezsurements at temperatures other than 15°C, from 10 to a1ec. However, the use of in situ measurements of conductivity for estimating salinity increased rapidly in the early seventies, rendering the "International Oceanographic Tables" unsuited for use in the majority of in situ measurements because the tabies do not go below 10°C, Further- more, a comparison of the conductivities of seven batches of standard sea- water, relative to a KCl solution revealed that the conductivity of sone batches was higher than that calculated from the certified chiorinity (Unesco 1976), This raised the problen of the calibration of the conduc~ tivity salinoneters and CTD probes, as weil as the definition of salinity itself. After discussion, the principle of calibrating standard seawater in electrical conductivity with a potassium chloride solution, vas adopted and the establishnent of a practical salinity scale was recomended by the JPOTS (Unesco 1978). Intensive work was then carried out in different laboratcries with radically different measuring equipment. This resulted in considerable data on which are based the Practical Salinity Scaie 1978, as well as the recommended algorithns for the calculation of practical salinity from the conductivity ratio at all temperatures and pressures over the range of oceanographic interest (Unesco 1979). This was finally adopted by the JPOTS during its meeting in Sidney, B.C., Canada, 1-5 September 1980 (Unesco 1981,a). Whereas the previous salinity scale (Unesco 1966,b) was based on a conductivity-chlorinity relation using natural seawater, the Practical Salinity Scale 1978 is different in that the standard seawater used was diluted by weight with distilled water or evaporated to obtain other salinity values. This procedure was followed to ensure the constancy of composition of this seavater over the salinity range of interest. A precisely specified solution of potassium chloride was chosen as a reproducible electrical con- ductivity standard; an evaluation was then made of the concentration of this solution which yields a conductivity ratio of unity at 15°C with respect to a standard seawater (from the North Atlantic ocean) whose certified chlori- nity was 19,3740°/oo, i.e. its salinity was exactly 35°/oo on the previous salinity scale. By convention, its practical salinity, on the new Practical Salinity Scale 1978, is 35, to ensure continuity at that salinity with the Previous scale.The following is a report on all the work and data on which both the Practical Salinity Scale 1978 and the algoritha for its calculation from in situ measurenents are based. This report consists mainly of papers published in a special issue of the IEEE Journal of Oceanic Engineering and in Deep Sea Research, together with one unpublished paper. The cooperation of the publishers of both Journals in granting permission for reproduction of these papers is greatly appreciated. This report is divided into five sections an¢ a conclusion : = Outline of the work and the laboratories involved. ~ Determination of the KCl solution primary electrical standard. - Measurement of the conductivity ratio as a function of salinity, temperature and pressure, ~ Fitting of the data. = Conversion of the existing data. In the Conclusion, the endorsed definition of the practical salinity scale, as drafted by the JPOTS at its last meeting (September 1980), is reported, together with the algorithm necessary for the calcu- lation of practical salinity from the ccnductivity ratio at all tempera- tures and pressures over the ranges of oceanographic interest. The Practical Salinity Scale 1978 was accepted by ICES in October 1979, by IAPSO in Decenber 1979, by SCOR in September 1980 and by 10C of Unesco in June 1981. The new equations must be used for all values published on and after 1 January 1982. New International Oceanographic Tables, Volume 3 (Unesco 1981,b) have been prepared and will be published shortly as Unesco Technical Papers in Marine Science No. 39. Copies are available, free of charge, upon request from Division of Marine Sciences, Unesco, Place de Fontenoy, 75700 Paris, France,THE PRACTICAL SALINITY, 1978 DEFINITION The practical salinity, syzbol S, of a sample of seavater, is defined in terms of the ratio Kis of the electrical conductivity of the seawater sample at the temperature of 15°C and the pressure of one standard atmosphere, to that of a potassium chloride (KCI) solution, in which the mass fraction of KCl 1s 32.4356 x 1077, at the same temperature and pressure. The Kis value exactly equal to 1 corresponds, by definition, to a practical salinity exactly equal to 35. The practical salinity is defined in terms of the ratio Kis by the following equation i; 2 $= 0,0080 ~ 0.1692 K,,!/? + 25.3851 Ky. 5/2 15 3/2 2 + 14,0941 K,,°/? - 7.0261 K,,” + 2.7081 K, formulated and adopted by the Unesco/ICES/SCOR/TAPSO Joint Panel on Oceanographic Tables and Standards, Sidney, B.C., Canada, 1 to 5 September 1980 and endorsed by the International Association for the Physical Sciences of the Ocean (IAPSO) in December 1979, the International Council for the Exploration of the Sea (ICES) in October 1979, the Scientific Committee on Oceanic Research (SCOR) in September 1980 and the Intergovernnental Oceanographic Commission (10C) of Unesco in June 1981, This equation is valid for a practical salinity $ from 2 to 42.REFE: UNESCO 1965. First Report of the Joint Panel on Oceanographic Tables and Standards, Copenhagen, October 1964. Unesco Tech. Pap. in Mar. Sci., No. 1 (incorporated with Nos. 4, & and 14 in No. 27 ~ Collected Reports of the Joint Panel on Oceancgraphic Tables and Standards, 1964-1969). UNESCO 1966,2. Second Report of the Joint Panel on Oceanographic Tables and Standards, Rome, October 1965. Unesco Tech. Pap. in Mar. Sci., No. 4 (incorporated with Nos. 1, 8 and 14 in No, 27, op. cit.)« UNESCO 1966,b. International Oceanographic Tables. National Institute of Oceanography of Great Britain and Unesco, Paris, p. 118. UNESCO 1976, Seventh Report of the Joint Panel on Oceanographic Tables and Standards, Grenoble, Septenber 1975. Unesco Tech. Pap. in Mar. Sci., No. 24, p. 61. UNESCO 1978. Eighth Report of the Joint Panel on Oceanographic Tables and Standards, Woods Hole, March 1977. Unesco Tech. Pap. in Mar. Sei., No. 28, p. 35. 1979. Winch Report of the Joint Panel on Oceanographic Tables and Standards, Paris, Septenber 1978. Unesco Tech. Pap. in Mar. Sei., Ko. 30, p. 32. UNESCO 1981,a. Tenth Report of the Joint Panel on Oceanographic Tables and Standards, Sidney, September 1980. Unesco Tech. Pap. in Mar. Sei., No. 36, p. 25. UNESCO 1981,b. International Oceanographic Tables, Vol. Pap. in Mar. Sci., No. 39 (in press). Unesco TechOUTLINE OF THE WORK AND THE LABORATORIES INVOLVEDNote on the symbols and units used in che Special Issue on the Practicai Salinity Scale 1978 of the IEEE Journal of Oceanic Engineering The papers presented in this report were published before the meeting of the Joint Panel on Oceanographic Tables and Standards, held in Woods Hole, USA, fron 1 to 5 September 1980, dur g which time the definition of the Practical Salinity 1978, S, w s established, and before the IAPSO Generel Assenbly, held at Canberra, Australia, in December 1979. The synbol S(°/oo) has been used in these papers for the practical salinity, instead of the adopted symbol S (Sun Report, 1980). The reader must read $ instead of S(°/oo) and thus ignore the "foo" when they appear in the text regarding the practical salinity. For example, $(°/o0) = 35.238 or $ = 35.238°/oo corresponds to S = 35.238. -1o-Reprinted, with permission, from IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. OF-5, NO. 1, JANUARY 1980 1 Introduction to the Special Issue on the Practical Salinity Scale 1978 ‘THOMAS M. DAUPHINE, soon MEMme. EEE Absract—This paper outlines the work done at the fve laboratories which provided, atthe request of the Jolt Panel on Oceanographe Tables ‘and Standards (IPOTS), the baie data forthe new Practica Sellaty Seale 1978. Imeriaboratory agreement of very high order was achieved. The ‘new scale ls ln proces of scceptance by all major oceanographic scities, ‘At whlch tae I will be ofcally recommended forall future reporting of ‘oceanographic data. IHE PAPERS in this issue describe work that has been done very recently aimed at the development ofa tniform repeatable Practical Salinity Seale, based on electrical conduc- tity. measurements, for use by all oceanographers, both in the laboratory and in the field. ‘The work was undertaken atthe request of the UNESCO- SCORACESIAPSO, Joint Pane! on Oceanographic Tables ané Standards (JPOTS) in several different laboratories with radi- cally diferent measuring equipment. One ofthe conitions for acceptance of any resuliswas that there should be independent confirmation over most of the range of interest. 'As an aid to both internal and interlaboratory consistency it was decided thatthe measurements would be caried out on recent lots of standard zeawater as supplied by the Seawater Service at the Institute of Ocsanopraphic Sciences, Wormley, UK, weight diluted or concentrated as necessary using dis tilled water forall dutions. Deviations from the behavior of these solutions due to different origins of natural water are expecied to be small and would be treated in the future as minor perturbations of the basic equations in the few cases where corrections were required The experimental work and mathematical analyses on which the Practica! Salinity Seale 1978 is based were caried cout at five different laboratories in four countries. The fol Towing groups were involved. A. L. Bradshaw and K. E. Schleicher, Woods Hole Oceano- graphic Institution (WHOD, Woods Hole, MA. F. Culkin and N. D. Smith, Institute of Oceanographic Sciences (10S), Wormléy, U.K. TT. M. Dauphinee, J. Ancsin, H. P. Klein, and M.J. Phillips, National Research Council (NRC), Ottawa, Ont., Canada. R.G. Perkin and E. L. Lewis, Institute for Ocean Sciences (105-0), Sidney, B.C., Canad. ‘A, Poisson, Laboratoire d’Oceanographie Physique (LOP), Paris, France. “Manuscript ecelved October 1, 1979; revised December 13, 1979. ‘The author is with the National Research Coune of Canada, Ot taws, Onc, Canada KIA OR6, In addition important confirmatory evidence not actually used to derive the scale was obtained from Bradshaw and Schleicher, from Millero, University of Miami (U of M), Miami, FL, and from the eatlier work of Brown and Allentoft (B-A), ‘San Diego, CA. ‘THE MEASUREMENTS The work done to develop the scale may be divided into six sections. 1) Set up a reproducible primary standard of electrical conductivity, against which all future lots of standard sea water and other secondary standards could be measured, and as close as possible to the conductivity of recent batches of standard seawater, so a to maintain continuity with pre- viously reported data. The standard chosen was the conduc- tivity at 15°C of a solution of potassium chloride (KCI) in distiled water having a specified concentration Ky chosen to give a conductivity ratio of 1 to current standard seawaters, (referred to 354,,) at 15°C and normal atmospheric pressure. Determination of the standard concentration Ky and ancillary experiments to facilitate future work were carried out at 10S, NRC, and LOP. The standard seawater samples were supplied by the Standard Seawater Service at IOS and were known to be on the curve of conductivity (C) versus chlorinity (C1) of recent batches, thus ensuring continuity of the scale at the transition date. Salinity was calculated from chlorinity using the accepted ratio S = Cl X 1.806550. It was agreed in advance by JPOTS that Merck “Suprapur” KCl suitably dried would be used for the KCI reference because it has excellent purity and a welldocumented chemical analysis. ‘The agreement among the thre laboratories turned out to be as good as one could reasonably expect. The final values for Ky (IOS 32.4353, NRC 32.4356, LOP 32.4358 gKCI/kg s0- lution have a spread of only 0.5 ma/kg (equivalent to 0.6 ppm S)).The average (rounded to 32.4356 g/kg) has been used by JPOTS in the definition of the scale, as reported by Lewis, in the paper immediately following entitled, “The Practical Salinity Scale 1978 and Its Antecedents.” 2) Determine the relationship between silinity and the ratio Ry,15,0 of the conductivity at salinity $ to the con- ductivity of standard seawater (35¢,, exact) under identical conditions of 15°C and one atmosphere pressure (p (gauge) = 0), '3)Determine the variation with temperature of the ratio Rex of conductivity at salinity S and temperature 1 to con- ductivity of standard seawater at the same temperature (p = 9). (0364-9059/80/0100-0001$00.75 © 1980 IEEE Sine‘The measurements for 2 and 3 sbove were carried out over the whole ocean range at NRC (0 assure ‘that a numberof los were represented, Residual water wat driven off by heating the KCI in a cruzile in ait for a period of time. Orinally the heating was carried out at about [10°C 150°C. and 200°C for up to 24h Subsequent measurements indicated. however. that this was cient and the final measurements were based on ‘in bulk form as received, toa least 260°C fora number for fusing the KCI. The fusion process used is very simple aud, we believe. preferable because itis fast and aves smaller scatter. The KCI was heated with an oxygen gas flame in a platinum crucible unti it melted. It was then poured from the crucible onto a clean polished stainles stel plate about 0.5 em thick on which t froze, without sticking. into numerous lumps of very convenient size and rapidly cooled to room temperature. Sticking was prevented by a thin layer of KCI powder which forms by condensatie on the plate and prevents actual contact between the lumps and the steel. Poisson (private communication) has confirmed that the lumps retain the original purity and that even the powder (which we do not use anyway) is barely contaminated Reshydtation of the desiccated granular KC1 dering cooling, and weighing has been shown by timed series af suc : weighings 10 be completely negligible. Even kee desiccated KCI in an atmosphere of 100-percent RH at 22°C resulted in only 30 ppm mass gain in 24 h, equivalent to an error of only 0.001 gikg in our KCI concanteation calculations. For weighing, a suitable amount of dried KCI was placed con a light stainless steel weighing pan with 2 pourspout (as recommended by Culkin and Smith, private communication) and its mass was determined ftom weighings of the pan plus KCI and pan plus residual KCI, respectively. before and after the KCI lumps were poured into a glass bottle. The appro- priate balance weight and ait buoyancy corrections were rade using 3 KCI density of 1.984 glcm?. The bottle was of hour (0364-9059/80/0100-0017500.75 © 1980 IEEE -2-18 IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. OFS, NO. 1. JANUARY 1980 Fig. “The differences aK = Ke - Ky fot the solutions made from KGL dried at 200°C(-), 260°C(2), and by melting (0). Ke is the concentration calculated itor 2g and (2), Kw is eaeulated from 250"C AND FUSED KCI" te y Re wget Stee SesuygT rie Geka dogs iegin iesh Gch ise bases 35888 wits Ete Bes es Hun ssa ise Jess x Sats Hen Histo +» symbols as in Table I Thre different los of Merck “SUPRAPUR™ Khare represented. TABLE MI [ERRORS IN KCl WEIGHT CALCUALTIONS AFTER AIR DRYING AT APPROXIMATELY 150°C" 20 4 108 a0) 5 tot er 3 a Bt 3 * Symbols ay im Tables Tand I. The samples in each Tot are made from the same stock solution, but each lot is fom different butt of Merck "SUPRAPUR™ KCL TABLETV. ‘THE STANDARD SEAWATER CONDUCTANCE READINGS (R) ON ‘WHICH Ky IS BASED (AT 15°C) AND THE CORRESPONDING. 24°C READINGS * 7 we en Se eee ee se wee Bao) Beat Biss muses pass os 1 Ret} Sen| cie es 2 mse past) 2 ee a i wm Ted a fey 212189 ODD ty 2 sas Ord Pee Nery sha Pe. he waite ea. o0aTe © Zag 1 atRg = 79.14841, R= 100 represents a cel conductance of 4000 fimho. aR = (R-R average). AS is the equivalent i salinity ta hoIEEE JOURNAL OF OCEANIC ENGINEERING, VOL. OF-5, NO. 1, JANUARY 1980 @ ate ~ Fig. 2. The effect of concentration on the ratio Z; 5/234. (2) Deviations (42) of al measured values ‘rom (3). (0) Measured values of Zy9/Z3a plotted on expanded seale of Za4 for the final sat of measurements (260°C and fused), ‘THE EFFECT OF TEMPERATURE ON Zw, By combining all of our 15°C and 24°C data we have also derived on expression (conveniently linear) for the ratio @isiZradg which allows 2,5 to be calculated with great precision, for use in (2), from similar measurements at 24°C (Zz4) made with a laboratory salinometer. Since the equation do. not require a knowledge of the KCI concentration it was permissible to use all available ratios, including those with less than ideal drying, and a total of 83 data points were represented. The 24°C to 15°C conversion equation fits the data with a standard deviation of @ = 3.5 X 10°§ (1.2 X 107* g/kg) and has the form @) fr, in terms of deviation from the standard value Zy,24 = 0.987550, Zys/z0 = 1.004217 + 8.4968 X 10-224 2, 5/Za4 = 1.012608 + 8.4968 X 10-9(Z,4 —0.987550). Ga) Fig. 2(@) shows a plot of the measured values of Zi s/Z20 veesus (3). Fig. 2(6) is an expanded plot of the points neat Z,5 = 1 for the final group of readings (>250°, and fused salt) ‘which were used to fix the intercept 2,24 = 0.987550 at the “normal” concentration (Zy,15 = 1). In addition to the 24°C data a series of rieasurements ‘wore made alternately on KCI solutions very near the normal concentration and on standard seawater with the variable ‘bath close (£<0.001°C) to 15°, 18°, 21°, 24°, 27°, and 30°C and the reference bath at 15°C. The vatiation of Zy,- could bbe expressed very accurately by a second degree equation in TABLE ‘THE EFFECT OF TEMPERATURE ON Zy.1* Tey is » 2 “4 * aw. © 20s (seanired) — Zy (calculated from (3). AK isthe quilt kl Smt citric. Each 2yy the vege of ars (T- 15°C) 2y,e 1 1.4655 X 1072(1) +9.103 X 107840)? -1250°C for at least 24h were found to be satisfactory. After drying the basin and contents were allowed to cool for 12h in vacuum. 2) The dried KCI was weighed in a stainless steel scoop, fitted with a 3.in long delivery spout. and then transferred to 4 tared 1-1 flask (containing a Lin Teflon coated magnetic stirter) by inserting the spout into the neck of the flask and rotating the scoop. The scoop was immediately reweighed to give the weight of KCI taken (w4). ‘The weight of KCl. conected for buoyancy wy was caleu- lated from x] Manuscript received August 28, 1979; revised December 13, 1979, The authors are with the Insitute of Oceanographic Sciences, Wormley, Godalming, Surey, Eggland. 1+ ( wr = wal 1+ p4(— rewad ea eyes where 4 density of air, calculated from atmospheric pressure, temperature, and humidity. Peet density of KCl (solid) = 1.984 g/ml. bw density of balance weights. 3) An appropriate weight of double distilled water, to give a final concentration in the range 32.42-32.44 gKCi/kg. was, introduced into the tared 1-1 flask containing the KCI. The contents of the flask were then stired to give complete dissolution of the KCI and finally the flask and its contents ‘were reweighed to give the weight of solution Ws. The weight of solution, corrected for buoyancy Wy was calculated from menfinl where Pa density of air 23 density of KCI solution = 1.0171 g/ml for ca 32 gKCI/kg at 25°C pw density of balance weights. ‘Then KCI concentration = wy /W, X 1000 g/kg (in vacuurn). MEASUREMENT OF ELECTRICAL CONDUCTIVITY The instrument which was used for these measurements was designed at 1.0.S. for measuring the absolute electrical conductivity of standard seawater. It comprises 2 rectangular quartz cell of uniform cross section, open at both ends, which fs placed in a silver beaker containing the solution to be measured, with the upper end of the cell above the surface of the liquid, The beaker is contained in an oililled bath, servocontrolled at 15° + 0.005°C, which in turn is surrounded bby another bath controlled at 15° + 0.1°C by means of a mer- ccury contact thermometer. The temperature of the solution is ‘measured. by means of a platinum resistance thermometer. The electrode system consists ofa shiny platinum gauze, fixed across the bottom end of the cell, nd a movable electrode, made from shiny platinum sheet, which travels on a rod passing through the vertical axis of the cell. The movable electrode can be located in two positions, the distance between which is meas ‘ured by comparison with a quartz standard length by means (0364-9059/80/0100.0022500.75 © 1980 IEEE -27-CULKIN AND SMITH: KCi SOLUTION CONCENTRATION DETERMINATION of a Michaelson incerferometer. The whole system is sealed to prevent evaporation during measurements. ‘The beaker containing the cell was filed with the solution (ca 1 kg) which was stirred to achieve temperature equilibrium at 15° + 0.0001°C and, with the movable electrode in the upper position, the ac bridge was balanced. The electrode was then moved the fixed distance to the lower position and the change in resistance was compared with a standard re- sistance to 0.0001 percent, From the resistance thus measured, the cross sectional area, and the distance through which the electrode was moved, the conductivity of the solution was obtained. Measurements were made at 1.5, 3.0, 4.5, 6.0, 9.0, and 10.0 kHz and the results were extrapolated to infinite frequency. RESULTS Standard Seawater Chlorinity Batch No, (loo) Infinite Frequency P59 193770 429202 P60 19.3765 42.9193 P68 19.3718 42.9095 Pn 19.3740 42.9139 Pr 19.3748 429163, PIS 193740 42.9146 P77 19.3755 42.9168 P79 19.3760 429178 P80 19.3782 429219 P82 19.3706 42.9075 Psa 19.3762, 429181 Linear regression, conductivity = 5.325053 + 1.940181 14,9. For chlorinity 19.37394%,., conductivity = 42.9140 mmhorem™?, - 28 - a Ki Solutions KCl Concentration Rys (at Infinite (eKCiKg) Frequency) 32.42894 0.999802 32.43328 0.999937 32.43980 1.000146 32.42909 0.999792 32.4362 0.999955 32.43048 0.999865 32.43304 0.999951 32.43661 1.000051 32.43303 0.999930 32.43684 1.000051 32.4300 0999844 32.43232 0.999902 3243552 0.999983 3243357 0.999944 3243283 0.999927 3243072, 0.999853 3242777 0999776 Conductivity of KCI solution at 15°C and infinite ‘= frequency Conductivity of standard seawater of chlorinity 19,37394%o (at 15°C and infinite frequency). Linear regression equation: KCI concentration = 0.82066 + 31.61455Ry5. Hence for Rys = 1.0, KCl concentration = 32.43522 glke. Ris CONCLUSION ‘A KCl solution containing 32.4352 sKCifkg has the same electrical conductivity, at 15°C and infinite frequency, as standard seawater of salinity 35.0000% (chlorinity 19.37394%0).Reprinted, with permission, from 4 IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. OFS, NO. 1, !ANUARY 1980 The Concentration of the KCI Solution Whose Conductivity is that of Standard Seawater (35°/,)at 15°C ALAIN POISSON Aberact—The concentration ofthe poamsam chloride sltion (KC) ‘whl asthe sume condoctvey at IC at P79 standard sexwlercor- ected 10 35.0009 hasbeen evaluate. The variation of the condoctty rio of KCl soitions to standard serwater (1%) has been measured ‘between 148 nd 152°C for KCl obtions whos concentration varies rom RwBek. INTRODUCTION }OR SEVERAL YEARS, we have been studying—as a mem- ber of the Joint Panel—the problem of the definition of salinity and standardization of standard seawater in electrical fot this purpose, it has been proposed several times to use a solution of KCI [1]-[4] The solution recently proposed by the Joint Panel on Oceanographic Tables of Standards (IPOTS), explained in detail elsewhere is close tothe solution that we had proposed: fon one hand, standard seawater will now be standardized in electrical conductivity at 15°C relatively to a KCI solution and, on the other hand, salinity will also be defined relatively 10 this KCL solution. Thus the aim of this work was to measure the concentration of the KCl solution which, at 15°C has the same electrical conductivity as standard seawater with a 35% salinity. We describe here the method we have used to determine this concentration and the results we have obtained. This work has also been performed independ- ‘ently in two other laboratories with different methods [7], (8) ‘The measurement technique and the intermediate calcula- tions, briefly presented here are given in more detail in a UNESCO technical paper (9). APPARATUS ‘The method we have used consists of measuring the re- tistance R and the temperature £ of a definite volume of a KCL solution or of standard seawater. The specific conduc- tance of the KCI solution is determined by x(KCI) = o/ RKC) QM where a is the cell constant at the temperature of the meas- urement and R(KCI) the measured resistance of the KC soluticn. In tie same way, for standard seawater X(6w) = g/R(w). @ Manuscript received October 3, 1979: revised December 13, 1979. ‘This work was mpported by the CNRS, ERA. 27 ‘The autior is with Laboratoire de Physique et Chimie Marine, Uni versité Pie et Marie Curie, Tour 24, 4, Place Jussieu, 75230 Paris Codex 05, Franc, Since these measurements are performed in the same cell, and in she same conditions, the cell constant remains the same for the two sets of measurements. To have X(KC)) = x(ew) it is sufficient to determine that RKC) = (ew). @) The results of these measurements are presented in re- sistance units (2), as obtained directly from the experiment. Resistance was measured with a Jonestype bridge (Leeds and Northrup, type 4666), temperature with a thermometric bridge (ASL, type H7) and a platinum thermometer (Tinley, type 5187 SA), and the volume ofthe solution was defined by a Jones-type cell (Beckman, type 61) with a cell constant equivalent to 120 ¢m™!. The temperature was measured with an accuracy of £1.107?°C and a precision of 0.2 X 107 The resistance given in the tables is the pure ohmic resistance of the solution. In particular. the influence of the frequency woof the current through the solution was eliminated by extrapolating AR, a I/wto infinite frequency. A previous study had shown that at high frequencies (8000, 10 000 Hz), the Falkenhagen effect is not negligible. The measurements which were used for the extrapolation were therefore per- formed at 2000, 2500, 3000, and 4000 Hz. The temperature contections of the Jones-type bridge, the elimination of the resistances of the connecting wires, and the extrapolation to infinite frequency were performed for each measurement. METHOD We first measured the resistance of a number of ampoules of standard seawater from lot P79. That lot was choten because it ison the basic line x(sw) = /(C1%o) of the latest batches of standard seawater prepared at Wormley (Culkin, private communication). We then measured the resistances of a number of KCI solution, close ro the resistance of standard seawater found above and of known concentrations 4s calculated from the weight of KCI and ofthe distiled water used to prepare them. DETERMINATION OF THE CELL RESISTANCE FOR SEAWATER OF EXACTLY 35%), 8 AND 15°C Since the measurements were not made at exactly 15°C and 35% 0S, we first determined by measurements presented elsewhere [10], the coefficients AR(sw)/Ar and AR(sw)/AS close to the resistance of standard seawater at 15°C, The re- (0364-9059/80/0100.0024800.75 © 1980 IEEE oyePOISSON: Kal SOLUTION CONDUCTIVITY TABLEI RESISTANCE OF STANDARD SEAWATER P79 CORRECTED AT IC AND 38%/o9 Comecied Renance a 2789836 2789832 3789.882 Fy 2789794 2789.384 2789.864 3789.538 2789.898 289.870 2789817 27895861 Ampoule Number 2 3 4 $ 6 7 a 3 0 1 sistances of the batch P79 samples, comrected to 15°C and 35% oS ate given in Table 1 To convert the 193760%— chlorinity indicated on the ampoule into salinity, we used the recommended formala [5] S= 1.80655 C. Therefore, salinity adopted for standard seawater P79 was 35.0037%0. ‘The mean value of the resistance of standard seawater P19 at 15.0000°C, corrected to 35.0009) 9§ was 2789.85 220.0322. DETERMINATION OF THE CONCENTRATION OF THE KC SOLUTION HAVING A RESISTANCE OF 2789.855 & AT 15°C Preparation of the KCl Solutions ‘All solutions whose concentration must be known with ‘reat accuracy were prepared with water double distilled on ‘quartz by accelerated evaporation and with Merck “Suprapur™ ka. ‘The distilled water was set in equilibrium with the air of the laboratory before preparing the solutions. In 2 previous study, the KCl used had been purified by recrystallizing it three times, ovenvdried at sbout 100°C, then fused in a platinium crucible. This study has shown that Merck “Suprapur” KCI has sufficient purity for the work carried out here. The data are thus relative to solutions directly ‘prepared with Merck “Suprapur" KCL, without additional purification. One set was prepared with KCl oven-dried at 150°C in vacuum, and 4 second set with fused KCI. For all solutions, corrections were made for evaporation in the flask where the KCI solution was prepared, and for air buoyancy on KCl and on the solution. The weighs in vacuum (vac) were calculated with the equation = rpn/7.76) Je = m(I + 11.8 X 107% 1 = @rpn/p) @ where dypy is the density of air at the time of weighing, at - 30 25 temperature 1, pressure p, and humidity A. p is the density of the body to be weighed and m the value read on the Mettler balance. For those calculations we used PKC (solid) = 1.984 grem? » platinium 1.45 gem? p distilled water =p SM.O.W The density of KCI solutions is given in the Handbook of Chemistry (46th edition). Corrected to the temperature of the ‘weighings (23°C), it can be written as » (solution KCI) = (0.9976 + 6.43 X 10~* C) grcm? ‘where C is the concentration in g(KCI)/kg of solution. The fused KCl was weighed with a Metter balance type MSGD (20 g max to 1 ug) in 2 platinum crucible of precisely known mass; the crucible was then transferred into a 11 screw plugged flask and the total weight determined on a Mettler balance type BSC 1000 (1000 g max to 02 mg). When introducing the distiled water, only a small amount of| fused KCl dissolves. Therefore, a correction was made for the sir displaced by the added disled water. The solutions obtained from dried KCI were prepared in the same way. In this case, however, KCI isin the form of smaller crystals rather than lumps, and thus whe introducing distilled water, all of the KCl dissolves. Therefore, the cor rection was made for the air displaced by a corresponding quantity of KCI solution. The Variation ofthe Resistance of KCl Solutions Around 15°C and 32.43 pkg-! A previous study had enabled us to determine that 2 solution of approximately 32.44 g of KCi/kg of solution has the same electrical conductance 25 standard seawater with a 359%, salinity at 18°C, Therefore, we first measured the variations of resistance of KC) solutions around 32.44 grkg™? and 15°C asa function of temperature and concentration, ‘Measurement of AR(KCD/At: These measurements were carried out using 2 KCI solution with a concentration close to 32.44 grkg-! prepared from distilled water and KCl dried for about 4 h at 150°C. The ratio AR(KCI)/A¢ has been estimated at -61.02-°C-1, This ratio does not need to be known with a great accuracy, for the measurements of resistance of the KCI solutions to be corrected (Tables II and III) were all carried out at temper- atures very close to 15°C. Therefore, the necessary corrections will be small. ‘Meaturement of AC/AR(KC)): We made three sets of dk- lutions, each from strong concentrations of about 200 g-kg™", prepared without special precautions. The corrections for ‘evaporation were made for the secondary dilutions. Each set of data C = f(R(KCI)) was fitted under the form of 2 polynomial Cm Ag ARE? +AgR +32? ©26 IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. OES, NO. 1, JANUARY 1980 TABLE RESISTANCE OF “DRIED KCl SOLUTION” AT 15°C AND THEIR CONCENTRATION CORRESPONDING TO THE. RESISTANCE OF STAXDARD SEAWATER OF 35¢log AT 15°C Resistance Concentration” Mestured —-Consentration _ Corrected for Tineot Corrected st Measured R= 2789 855.0 Deving tise ke) ay 2h 2795.265 324071 324482 26h 2791 058 324025 324370 sasys 324232 324362 4aays 32eaie 324357 days 32am 324360 ays 32eaa7 324352 TABLET RESISTANCE OF “FUSED KC} SOLUTIONS” AT 15°C AND THEIR CONCENTRATION CORRESPONDING TO THE ‘RESISTANCE OF STANDARD SEAWATER OF 35%og AT 15°C Measured Concentration* Reurtance Concentration ‘Corrected for Solution Meret Conected at Measured RemH9assa Number 15°C (a) wee) 1 791.285 32.4366 2 2791;300, 324365 3 B 3790.67 324360 3790.708 324364 ‘ 8 3790.436 324356 2790:502 324358 s c ag.714 324385 ra9.713 324355 ‘ c 2789388 324350 3799380 324383 7 c 2189.956 320357 2789°940 3243s 8 c 2789.365 324351 2789.60 324386 324355 G0 324387,0 4.7 10 where R = RKC). The three polynomials obtained were then normalized to R = 2789.80 9 for C= 32.436 grks~". Finally, the experimental data, normalized to this point have been fitted to a polynomial of the same form as (5) above whove coefficients are Ag = 521.2965 Ay = 726.6371 Az= 0570524 Ay = -$.772865 X 107 Ag= 227518 x 10-*, From (5) we obtain dC Ay Aa 4 + PAW R+ UR. © a. Since the resistance of standard seawater at 15°C is 2789.88 , the slope at 15°C is given by AC/AR(KCI) = —12.06 X 107? g + kg"! O71. We have used this ratio to correct the experimental data on the KCI solution concentrations given in the following section, ‘The procedure is justified because again these corrections are very small Determination of the Concentration of the KCI Solution Having a Resistance of 2789.855 Nat 15°C Solutions of Dried KCI: The KCt solutions were prepared from KCI dried én an oven in vacuum at 150°C for 24 h (sol 1) or for 4 days (sol 2, 3,4, 8). Tee KCl used came from three different flasks (A,B, C directly delivered by the manufacturer. ‘The resistances measured and the concentrations calculated from the weights of KCI and distilled water are given in Table I In the calculation of the average concentration C and of the standard deviation o, solution 1 which was obviously erratic was not taken into account. Solutions of Fused KC: The solutions were prepared from KC fused in an oven at about 850°C. The data given in Table If: the average concentration C = 32.4357 grkg™! is slightly lower than the concentration obtained with dried KC (= 32.4360 g-kg™1), but the two average concentrations remain within their respective experimental ertors. We can suppose that KCl had not been sufficiently dried, and that there was some earPOISSON: Kal SOLUTION CONDUCTIVITY water left in the sat crystals, which would account for the slight difference between the two sets of data. The average value of all the data (13 independent solutions) is C = 32.4358 grkg™! with a standard deviation of 5 x 10~* ke! ‘THE CONTRIBUTION OF THE SPECIFIC CONDUCTANCE OF DISTILLED WATER TO THE CONDUCTANCE OF THE 32.4358gkg™! KCI SOLUTION The specific conductance of the KCI solution (KCI) is the sum of the specific conductance of K* and C1 ions of KCi,x(K*CI™), and the specific conductance of the distilled water x(w) used to prepare it: x(KC1) = (KCI) + x(u). o Now the conductance of our distilled water can vary from 0.5 X 10-* Q-Frem™! for a water neatly free of gas to 2X 10-6 9-"+cm™! and even more for a distilled water set in equilibrium with air by bubbling. The distilled water we used had previously beén set in equilibrium with the air of the laboratory without bubbling. Its conductance was meas- ured with the method of Justice (private communication): x(w) = 1.0 10767! + om’ Hence, from (1), (2), nd (7) aR(KCD, = 1.0 x 10= RKC) RKC), ® where a 2 120 m= and R(KCI) = 2789.855 2.Consequently R(K*CI~) = 2789920 2 and the conductance of pure water therefore is equivalent to a resistance of -0.065 2, that is to say is equivalent to (from the coefficient AC/AR (KCI) previously determined) + 0.0008 grkg™! in concentra- tion of the KCI sclution. Rxc1 AS A FUNCTION OF C AND t NEAR 3808 AND 15°C. The variation of the ratio Ruci = x(KCe,15.0/ weW)95,15,0 atound 15°C and 32.4 grkg~ !isneeded over an adequate area of temperature and concentration for the standardization of future standard seawater which will bbe made at temperatures and concentrations slightly dif- ferent from 15°C and 32.4358 grkg~!. ‘This ratio can be written Ryey = 1 + Ay (C ~ 32.4358) + Az(C~ 32.4358)? +Ay(t 15) @) where Ay and Ay are constants calculate later S Ka.0350.18.9 4 9%, (10) dt xGW)s5,10 -32- A, had been measured before: Ay = —0.00107°C7!- With polynomial (5) normalized at R = 2789,855 2 for C = 32.4358 grkg™' and the resistance of standard sea- water at 15°C (2789.855 9), the ratio Ryci was measured at various concentrations of the KC! solution.' These data were fitted in a polynomial of the second degree in con- centration C, normalized at Rxci = 1 for C = 32.4358 arkg"! Ayer = 1 + 2.9722 X 1072(C— 32.4358) 8.1 x 1076(C— 32.4388)" any The variation of Ruci around the point of equivalence is then Ryer = 1 +2.9722.10-7(C— 32.4358) —8.1 x 1076(C~ 32.4358)" 1.07 10-3(¢—15). aay This equation can be used for temperatures from 14.8°C to 15.2°C and for KCl solutions with concentrations between 32 grkg™? and 33 grkg™". CONCLUSION In summary, the specific conductance at 15.0000°C of standard seawater P79 corrected to 35.0000 9S is identical to that of a solution of 32.4358 gKCI/kg of solution (in vacuum) whose specific conductance of pure water is 1.0 X 106 2 .em=", Rc near the point of equivalence varies according to (12) above. The value of KCI concentration found in our laboratory has been used with those given by Dauphinee etal. [7} and Culkin er al. {8} to calculate the standard concentration ‘sed inthe “Practical Salinity Scale 1978" [6] by JPOTS. ACKNOWLEDGMENT ‘The author thanks Dr. T. M. Dauphinee for his helpful criticism of this manuscript. REFERENCES 11) A, Poison, “"Measuement of sbiolute elecwial conductiviny of ‘andar senwaer on the bass of KCTes standard." UNESCO Tech. Pap. ‘Mat. Sei ol. 26,p, 61, Annex 3.1976 ——"Proposl te callrute standadseavaterconducivig relative a KCl olton. Preliminary rea,” UNESCO Tech Pap. Ma, Si, vl 28, p35, Asner 3, 1978 AT. M, Dauphiee, C.K. Ross, and F. Culkin “The reli of andard seewater ab an elacical condativty standard.” Oceancl ‘Acta, vl. 120.4, pp. 425-433, 1978, ELL. Lewis and R. 6. Perkin, "Salinity: efi and exiultin.” 2. Geopiys Res. ol. $3, pp Abbed78, 1978 'W. S: Woon, AJ. Les, and G. Diesich, "Redefniton of lini." Deep Sea Re, vl. 1, pp. 321-322, 196. E-L- Lewis, “The practical lity vale 1978 andi antcedens," hia By a Co) 5) 6 4 Note that ROKCD is the rerttance of the solution: Rc is the resistance rerio to the standard value.2B ‘ame, pp. 3-8 mt J. Avesn, H.P. Kein and M J. Pilipa, “Te eect of coecenrson onthe condutvi rao of peasom chloride wlons to sear senate fslinty 3949 (Cl. 19.3740) at IS an 20°C, 12d he effect of temperate oo the concoct aio berwee 15 td ‘30°C, tis ase, pp 7-21 (8) F. Cat and N.'D. Saath, “Desemination of te concen of = 33- o (10) IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. OFS, NO. 1, JANUARY 1980 Potasiam chore soon having the tame elcialcnaducvity, 15°C and infinite Pequency, ax andar sere of wity 3g” ht ‘sme, pp. 2-23. (onccnerad standard swaths ae, pp 4150.MEASUREMENT OF THE CONDUCTIVITY RATIO AS A FUNCTION OF SALINITY, TEMPERATURE AND PRESSURE = 35-Reprinted, vith permission, from IEEE JOURNAL OF OCEANIC ENGINEERING, vol. OE-5, no. January 1980, The Electrical Conductivity of Weight Diluted and Concentrated Standard Seawater as a Function of Salinity and Temperature Absrac—Tde ration Ryo of electrical conductivity of seawater samples of precseiy Known salinity to standard seawater at the same temperature have been measured over a wide range of ellis from 02 42%/, und over the fll rane of ocean temperature from ~2t035°C. ‘Th samples with 5 <35/n9 were prepared by acrarate weight dito of ‘mandard sevester with ditled water. High sallnlty samples were ‘prepared by fat evaporation of standard seawater and sabaequent weight ‘lution lato the already determined < 3P ag range. Am equation was derived which erpremes he Sverms 2, eatoaship ‘very precisely from 1-4? qg and at all emperature, La, S* M1 Ran0)* (Rano!) ant where du m 1— 15°C, R= R95 ly the ret tem fe required at LSC. ‘The effect of temperarare oo the electrical conductivity of stndard seawater was also measured. The rai ofthe condoctivity a temperatare 4a the condactiviy at 15°C (C35, fC 5,15) avery accurately exprened ‘bya fourth degree eqation ie, ‘These two equations are sufcent forall salinity determinations at ‘tormal smaspberlepremure. INTRODUCTION 'N CONNECTION with our work on isin measurements of conductivity, temperature, and pressure in the ocean, ‘we have found it necessary to develop facilities for laboratory Manuscript received November 1, 1979; revised Decembet 13, 1979. ‘The authors are withthe National Research Counel of Cansda, Ot tawa, Ont, Canada KIA ORS. calibrations and testing of the equipment. This led to the development of a new and very precise continucas flow system for comparing the conductivty of samples of sea- water with standard seawater at the same temperature, for use in salinity determinations. This in tur ultimately led to production of the Guildline “Autosal” laboratory salinometer ‘which is now in use in almost every major oceanographic standardizing laboratory in the world We then adant-d the basic concept of the salinometer system to allow simultaneous measurements on the sime sample at two different temperatures and began a compre- hensive study of the effect of temperature and salt concentra- tion on the conductivity of seawater samples made from weight diluted or concentrated samples of standard seawater. ‘The measurements were made from below the freezing point to 38°C and from 424,45 to 1%, Swith some measureraents at 15°C carried down to a few ppm S todetermine the shape of the conductivity versus salinity curve at very low salinities, Inaddition, we carried out determinations ofthe concentration of potassium chloride (K) giving a conductivity ratio Rx of I to standard seawater at 15°C, and of the effect of both a few ppt variations around the standard value and temperature ‘between 15° and 30°C which are reported elsewhere in this issue (1). The conductivity measurements were undertaken as part fof an intemational program sponsored by UNESCO-SCOR IAPSOICES (Joint Panel of Experts on Oceanographic Tables and Standards)(2]. The program includes a redef- inition of salinity and development of a standard set of equations and tables for calculation of the salinity of sea- waters both in the laboratory and in the ocean. These wil be recommended for use by all oceanographers after approval by the various sponsoring bodies. The work reported here forms a substantial part of the basis for the new equations (0364-9059/80/0100-0028800.75 © 1980 IEEE orDAUPHINE eral: STANDARD SEAWATER ELECTRICAL CONDUCTIVITY and in fact Jed to the form of the surface pressure equations that have been used. Similar measurements using entirely different equipment were carried out by A. Poisson (Paris, France) [3] and confirmatory work on the effect of tem- perature on the conductivity at various salinities was carried ‘out, using a third type of apparatus, by Bradshaw and Sch- leicher (Woods Hole, U.S.A) [4], who have also repeated with greater accuracy their earlier study [4a] of the effect of pressure on the conductivity. In addition, Culkin and Smith (Wormley, U.K.) [5] and Poisson [6} also carried out deter rinations of the KCI equivalent of standard seawater at 15°C. Ik is extremely gratifying to report that, with few ex- ceptions, wherever comparisons are possible, agreement between the different laboratories using ently different equipment has been better than the equivalent of +1 ppm, almost an order of magnitude better than has been the norm. Except for the effect of temperature near and below 0°C and at salinities below 22,5 the new work is not seriously different from equations fitted to the eatlier work of Brown and Allentoft [7] on which many oceanographic calculations have been based. Because of this solid data base the new equations should bbe accepted with great confidence by the oceangraphic com- munity. It should be pointed out thatthe equations given in this paper represent the best fit to our work only and are not recommended for adoption by anyone. Although the dif- ferences from the new scale are negligible the averaging in of other data has resulted in different constants forthe equations of the oficial version and only the official constants should bbe considered in the interests of completely consistent re- porting of oceanographic data. Practical salinity determinations at zero (surface) pressure axe always made by taking the ratio (R, W) of conductance C of the sample to the conductance of the same cell with Standard seawater (35% 0) either at the same temperature (R,) of at some standard reference temperature, usually 18%, Rr=CrrolCrs.r0, We=Cs1.0/C35.15.0, for most lab mearurements for most field measurements. ‘The above quantities" are related by the function Te CrsnolCas.1s.0 =f wo which expresses the effect of temperature on the conductivity of standard seawater, i, Wem Rete @ Insitu measurements of W also include the effect of pressure and this may be expressed for calculation purposes as @) 4 Subscripts, ¢, and o, refer to salinity. temperature, and pressure (o = 1-atmot, respectively, of the sample Under consideration, one oF ‘other subscript may be omitted when the meaning is obvions - 38 = 29 ‘The basic equation for the calculation of salinity, whatever its form may be, that for the standard temperature of 15°C at zero pressure S=ARis) o Measurements at other than the standard conditions are then usually treated as perturbations of the basic equation. For instance, if the measuring and the reference temperatures are the same but nor 15°C and P = 0 substitution of Ry in Ay alone creates a small error and a correction function f2 is requited, Le., S=fh Rd +hRnn? © If the reference is at 13°C and P = 0 one can substitute from @ Ree Wilh For measurements in situ one substitutes, combining (2) and 8) Re=Wyl(t +0) +1 instead. ‘One possible form of the general expression for S is there- fore © S=fy(WplCl +0) -1d+ fap I+ @) - red) feom which terms may be dropped as appropriate to simplify the caleulations. The present measurements along with those of other workers were undertaken with a view to determining the best form and appropriate constants for the expressions for rand, APPARATUS ‘The measuring system we have used, ilustrated in Figs 1-4,15 essentially the same as that used by Dauphine and. Klein (8} in their measurement of the temperature coef- ficients of standard, synthetic, and Atlantic waters near 35%, 8. The sample water (Fig. 1) is driven by air pressure from an aquarium pump from a sample bottle via a thin Teflon tube. This tube divides to direct the water into two fine bore stainless steel heat exchangers (2.5 m long, 1 mm LD, which are mounted in different temperature controlled baths. From each coll the water passes to one of the two ‘conductance cells, also in the bath, and from the cell, through fan insulated exit tube to the side of the bath where it falls in droplets into a funnel leading to the waste bottle. A pinch clamp on one input line allows balancing of the flows between the cells as the water viscosity changes with change of tempet- ature ‘The cells have four sidearms projecting upwards and each containing a helical Pt1ORh platinized electrode and a Teflon 2 Previously the comection has been made to Ryo to bring it to Rag The reason for the change i iscuned later.30 IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. OFS, NO. 1, JANUARY 1980 , wna) Fig. 2. The component layout inthe bath. (2) Up-down flow separa: to) pals (0 Cook cl. Contac Heat « (0 Heater lamps (@) Control thermistors.(h) Resistance ermomete. = 39-DAUPHINEE er: STANDARD SEAW: 31 Fig. 3. The bath temperatureconttal circuit. Thermistors are Veco, '32A1, 2km at 25°C partly Blackened on one side: INS314~ SmA ‘current regulation diode heaters ae 4-200 incandercent Bulb — aye (oe ' Ae ox > 7 ne ee oe Te | ~ ee | Semen see | NETWORK L ad \ fh i pee eg | corals vee ; for fe eae ‘eae .Po rt 4 Fig. 4. The conductance measuring circuit. Sy ~ Sg are C/MOSFET suitehes (RCA type 4016); amplifiers are AD74N except Ay1.413 (FET input BB3$22 and the curret diver MONO-OP G3). = 40-32 sir escape tube which allows complete filling of the sidearms and flushing of the cel by forcing air down them. To ensure that as much water as possible is driven from the cell on flushing, the cell is mounted at an angle with the exit end the lowest part of the system. This has the added advantage that any bubbles that might be formed at higher temperatures, in the heat exchanger are trapped at the inlet end of the cell, before they get to the measuring zone. The ar escape tubes are electrically isolated in a common header by projecting them vertically downward in a square array about the con- tinuous air flow from a second aquarium pump which keeps. ‘them dry. Flushing is accomplished by pinching ofthe header outlet tube. Leakage to the exit end of the cell at low meas- uring temperatures is prevented by passing the exit tubes ‘through a grease seal in a styrofoam block whose outer surface stays above the condensation temperature. [solation from the waste system results from the separation of the output drop- lets. ‘The bath temperatures are read with a Rosemount platinum resistance thermometer and a Guildline “Dauphinee” direct, reading bridge [9] with interpolation by deflection to0.2 m°C. ‘The baths (Fig, 2) are of 40-and 60-I capacity and stir with a cycle time of a few seconds. They are cooled via copper coll ¢ by cold nitrogen gas evaporated from the liquid by a Veriac controled heater. Actual temperature control is by tubular incandescent lamp heaters f powered in on-off mode by a thermistor controlled supply. The contol circuit, shown in Fig. 3, utlizes two thermistors in parallel (or singly for checking purposes), with some of the ligt energy fed diectly to them so that intemal heating shortens the cycle time to about 1/2 s. Experience has shown that this simple circuit routinely controls the bath temperatures to within=0.2 m°C for many hours and since this is better than the reproducibility, fof the copper thermometer system previously used, the thermometers were dispensed with in favor of frequent direct reference tothe platinum scale, ‘The conductance measuring circuit is shown in Fig. 4. The cell operates on square wave ac curtent generated in the cell by solid-state switching(S, Sz, A2) of apreciselyregulated 1/10 reference voltage #Vq. The cumrent is controlled (At, Aus, Ara) to Keep the square wave component of voltage across the cell potential leads constant, s0 that the current is proportional to the effective conductance of the cell between the potential electrodes. This current is meas- ured by reading the voltage difference across a 250-9 (or 200-9 at high conductances) resistor in series with the cell with an isolating differential amplifier(43, Ag, Ag)- The summing point is held at zero potential by an inverting dif- ference amplifier A, plus a suppression signal (4,)adjustable in 22 steps which is derived from the reference square wave. The output of the difference amplifier is demodulated (Ss, Sa) to de with 2 blocking capacitor in series to eliminate zero offset, and after further dc amplification (4y) to exactly -Visuppression step is read with a precision DVM. Ratios have been chosen so that cell resistance exactly equal to the series reference resistor represents exactly 20 steps of the suppression dial, which is read as 100 and added algebraically to the DVM reading to obtain the conductance =a. IEEE JOURNAL OF OCEANIC ENGINEERING, VOL. OF-5, NO. 1, JANUARY 1980 in arbitrary units. Our cells give a reading of about 85 with standard seawater at 15°C. ‘A simple four-pole selector switch and plug connections can be used to select and interchange the cells since the circuit is effectively independent of lead resistances and thermal emf's. The effect of intetlead capacitance which could cause oscillations is eliminated by use of tape wising with the conductors on either side of the potential leads driven at the same potential (guard). Current offset due to polarization is eliminated to sufficient approximation by a switched (Ss, Se) averaging filter which adtomaticaly adjusts the bias of the drive circuit (49) to keep the de component of cell current small, Shunt current to the heat exchanger is eliminated by holding the potential lead at that end at zero potential (Ajo). We use the inner cell electrodes as current leads because this reduces the driving voltage required of Ay. The extremely low input current of FET amplifies 4, , 412 and regulation of the current to the proper value prevent errors due to changes of potential and current electrode polarization or resistance during the course of one cycle. Balances are normally reproducible from flush to flush to a few parts in 10°, equiv- alent to about £0.1 ppm S. PREPARATION OF SOLUTIONS ‘The salinity of a solution is defined as ” zty where S, x, and y ate the salinity, mass of salt, and mass of pure water, respectively. In order to determine these masses it is necessary to make appropriate corrections for air buoyancy using the relations My 19 My — My where Mo, My are the true masses of the object and the ‘weights of the balance and mg and m,, are the masses of air displaced by object and weights, respectively. We can also write Mo~PaVo =Mus —PeVw where py is the density of air, corrected for barometric pres. sure, temperature, and humidity, and Vp and Vy ate the volumes, of object and weights. Finally Mynite(1-25) +o 1-22 The samples for conductivity measurements were prepared in glass ampoules of about 285-cm? volume made by rounding off one end of a standard seawater bottle and cutting and flame polishing the narrow filling tube at the other end. Each bottle had its own cap made of a short (1-cm) piece of rubber tube with a Plexiglas disk across one end. The caps typically ‘weighed about 0.7 g. The true mass of each ampoule was ®