DOI: 10.1111/are.

13674

ORIGINAL ARTICLE

Effects of mechanical aeration on evaporation rate and water

temperature in aquaculture ponds

Hisham A Abdelrahman1,2 | Claude E Boyd2

1
Department of Veterinary Hygiene and
Management, Faculty of Veterinary
Medicine, Cairo University, Giza, Egypt
2
School of Fisheries, Aquaculture and
Aquatic Sciences, Auburn University,
Auburn, AL, USA
Correspondence
Hisham A Abdelrahman, Department of
Veterinary Hygiene and Management,
Faculty of Veterinary Medicine, Cairo
University, Giza, Egypt.
Email: Hisham@auburn.edu
Abstract
Water temperature and water loss by evaporation were monitored in control ponds
and in ponds with different rates of aeration (9.2, 18.4, 27.6 and 36.9 kW/ha). The
mean decrease in water temperature at 70-cm depth was greater than that at the sur-
face in aerated ponds than in control ponds. The greater the aeration rate, the cooler
was water, both at the surface and at 70 cm. Evaporation rates were found to
increase with greater aeration rate. Water loss increased by 32%–92% over 24-hr
periods in ponds with one to four 0.37-kW Air-O-Lator aerators, respectively. The
nutrient-enriched control pond was more turbid, had cooler surface and deep water
temperature, and had greater evaporation loss than the control pond without nutri-
ent addition and less turbid water. But, aeration did not increase turbidity. Aeration
can increase water loss from ponds and result in lower water temperature. Although
aeration should not be used excessively in order to conserve water and reduce
production cost, it is essential for many types of feed-based aquaculture.

KEYWORDS
aquaculture, earthen-lined ponds, evaporation rate, mechanical aeration, vertical pump, water
temperature

1 | INTRODUCTION
The application of feed to aquaculture ponds results in a large oxy-
gen demand (Chatvijitkul, Boyd, Davis & McNevin, 2017) that often
exceeds the supply of dissolved oxygen from aquatic plant photo-
synthesis and diffusion of atmospheric oxygen into the water (Boyd
& Tucker, 1998). Water exchange can be used to avoid low dis-
solved oxygen concentration, but it usually is more efficient to apply
mechanical aeration than water exchange in ponds to increase DO
concentration (Hopkins, Hamilton, Sandier, Browdy & Stokes, 1993;
McGee & Boyd, 1983). Mechanical aeration has become a standard
practice for avoiding low dissolved oxygen concentration in feed-
based culture of channel catfish Ictalurus punctatus (Tucker & Harg-
reaves, 2004), marine shrimp such as Litopenaeus vannamei and
Penaeus monodon (Boyd, McNevin, et al., 2017), and several other
species.
Paddlewheel and other types of aerators that splash water into
the air are most commonly used in commercial aquaculture (Boyd &
Tucker, 2014). The splashing action of these aerators increases the
surface area of water exposed to the air and facilitates oxygen trans-
fer rate from air to water (Boyd, 1998). Of course, greater surface
area of water exposed to air would be expected to increase evapora-
tion, because the process occurs at the air–water interface (Harman,
2005).
Phytoplankton blooms in ponds tend to increase the absorption
of solar radiation near the surface, the result being a sequential cool-
ing of the deep layers of the pond and a warming of its surface lay-
ers (Idso & Foster, 1974). Mechanical circulators that do not splash
appreciable water into the air actually increase the temperature of
the entire water column, because of the mixing of higher tempera-
ture surface water with deeper, cooler water (Busch, Flood & Allison,
1978). However, mechanical aerators that function by splashing
water into the air likely lower water temperature, because evapora-
tion is a cooling process as heat from the surroundings is required to
convert liquid water to water vapour (Harman, 2005). Thus, mechan-
ical aeration in ponds may not only increase water loss but it may

Aquaculture Research. 2018;1–9. wileyonlinelibrary.com/journal/are © 2018 John Wiley & Sons Ltd | 1
2 |
also result in a lower water temperature than would occur without
aeration.
There apparently have been no published studies on the effect
of mechanical aeration with surface aerators on evaporation rate or
water temperature in aquaculture ponds. Therefore, the present
study was conducted to evaluate the influence of aeration on evapo-
ration rate in ponds, and to determine the influence of evaporation
on pond water temperature.
2 | MATERIALS AND METHODS
The experiment was performed at the Auburn University E. W. Shell
Fisheries Center (SFC), Auburn, AL (USA) (32° 390 0 0.5″ N, 85° 290 0
6.9″ W). Six, 0.04-ha research earthen ponds located side by side,
with the same dimensions, containing water inlet and outlet control
structures, and supplied by the same water source were selected.
The average depth of each pond was 0.85 m. The ponds had vertical
concrete side walls to assure that water surface area was the same
until water levels fall at least 30 cm below the top of overflow struc-
tures. The study was conducted between early June and early Octo-
ber in 2013.
One pond had no aeration and served as the control (C), while
the other five ponds were treated with nitrogen and phosphorus fer-
tilizer twice per month at the rate of 9 kg N plus 9 kg P2O5/ha to
promote phytoplankton blooms. Aeration was not applied to one
nutrient-enriched pond to provide a nutrient-enriched control (NEC),
and either one, two, three or four 0.37-kW Air-O-Lator aerators
(Air-O-Lator Corporation, Kansas City, MO, USA) were installed in
each of the other ponds. These vertical pump aerators splashed
water into the air (Figure 1).
The aeration rates were equivalent to 9.2, 18.4, 27.6 and
36.9 kW/ha. Aerators were alternatively operated for 24 hr (ON)
and then turned off for 24 hr (OFF). In total, aerators were ON for
25 days and OFF for 25 days. Ten grass carp Ctenopharyngodon
idella (mean body weight and standard deviation 10.32 g  2.64)
were stocked into each of the six ponds for aquatic weed control.
F I G U R E 1 A 0.37-kW Air-O-lator aerator showing water being
splashed into the air during operation
ABDELRAHMAN AND BOYD
Water samples were dipped once per day from approximately
10 cm depth in each pond with 0.5-L plastic bottles, carried to the
laboratory for immediate analysis. Turbidity was measured using a
VWR Model 66120–200 turbidity meter (VWR International, Radnor,
PA, USA). Chlorophyll a (Chl a) concentration (membrane filtration,
acetone extraction and spectroscopy) was measured weekly as
described by Boyd and Tucker (1992).
Water temperature at 70-cm depth (hereafter called “T70″) in
each pond was monitored at 1-hr intervals with a Model 64K HOBO
Pendantâ Temperature/Light Data Logger (UA-002-64, HOBO
wareâ; Janesville, WI, USA). The logger was attached to the top of a
brick using a Zip-Tie. Each brick was suspended from both ends by
70-cm long ropes, and each rope was connected to a float. The
floats were attached by ropes to wooden stakes on opposite sides
of the pond. This maintained the brick and thermometer in the area
deeper than 70 cm.
Surface water temperature (hereafter called “TS”) in each
pond and air temperature were monitored at 1-hr intervals with
Model 64K HOBO Pendantâ Temperature/Alarm Data Loggers
(UA-001-64, HOBOwareâ). These data loggers were tied to the
bottom surface of one of the two floats that suspended the
brick at the pond surface. Two data loggers for monitoring air
temperature were mounted at a height of 3 m under two open
sheds—roof but no sides—located at about 200 m from the
most distant pond. The data loggers were installed the day
before starting the experiment and used until the end of the
experiment. At the end of the study, software provided by the
manufacturer (HOBOware Pro 3.7.1) was used to download data
into a laptop computer. Data loggers were connected to the
computer using HOBO Pendant Optic USB Base Station and
Coupler (part # BASE-U-1).
A standard National Weather Service Class-A type evaporation
pan (Forestry Supply, Jackson, MS, USA) was used to measure evap-
oration. The pan was mounted on a level, wooden platform above a
grassy surface. The pan was not shaded by trees, buildings or other
tall objects, and weeds and grass around it were mowed regularly.
The pan was filled to within 6 cm of the top with city water. Water
loss from the pan was measured by the aid of a stilling well and
hook gauge (Forestry Supply). The pan was cleaned as necessary to
keep it free from algae and surface films that might alter the rate of
evaporation. A standard US Weather Bureau type rain gauge (For-
estry Supply) was installed on a level wooden platform over a grassy
surface near the evaporation pan. It was monitored and depth of
any rainfall was recorded.
A stilling well consisting of a 10-cm diameter polyvinyl chloride
pipe was installed in each pond close to the edge. A 1-cm diameter
hole was drilled into the side of the pipe to maintain the same water
level inside the pipe as outside in the pond. Water levels in the still-
ing wells also were measured with the hook gauge.
Means of daily air and water temperatures were calculated by
averaging all 1-hr readings from the loggers each day. Air temper-
atures were calculated as hourly average of temperature recorded
by the two data loggers mounted in the sheds. The amount of
ABDELRAHMAN AND BOYD
water loss per kilowatt of aeration per hectare was calculated.
The differences in turbidity between aeration periods when aera-
tors were turned on and those when turned off were calculated.
The cost that would have been accrued had the water been
pumped into the ponds was calculated as described by Yoo and
Boyd (1994).
Statistical analyses
Data were analysed using means, standard deviations, t tests and
simple linear regression. The Shapiro–Wilk test was utilized for nor-
mality analysis of the variables. A Levene’s test (Levene, 1960) was
conducted to check the homogeneity of variances between ponds.
Analysis of variance (ANOVA) was used for data with a normal dis-
tribution, and if there were significant differences, the Tukey’s Stu-
dentized Range (HSD) test was used for post hoc analysis.
Otherwise, the non-parametric Kruskal–Wallis test was used for data
not normally distributed, and Wilcoxon signed ranks test was used
to calculate the difference between samples in cases showing differ-
ences with the Kruskal–Wallis test.
To account for variations in water temperature during ON peri-
ods that could be attributed to variations in water temperature dur-
ing OFF periods, two models of one-way analysis of covariance
(ANCOVA) were used separately to assess the effects of aeration
rate on TS, and T70. Least-squares adjusted means were used to
compare the true effects of the treatments, unbiased by differences
in water temperature during OFF periods.
Statistical significance was set at p < .05, and all data were pre-
sented as the mean  standard error (SE). For conducting t tests,
the protocol for calculating the t-value took into account whether
variances were homogeneous as outlined by Steel and Torrie (1980).
Analyses were performed with SASâ version 9.4 statistical software
(SAS Institute Inc., Cary, NC, USA).
Calculation of evaporation
The water-level changes in a pond and in the class-A evaporation
pan were measured daily:
D Pond or D Pan ¼ WL2  ðWL1 þ PÞ
where Δ Pond or Δ Pan = change in water level during a 24 hr per-
iod (mm/day); WL1 = initial water level (mm); WL2 = water level
after 24 hr (mm); P = rainfall during 24 hr period (mm).
Pond evaporation was measured as:
Eoff = Δ Panoff 9 monthly pan coefficient
Eon = Δ Panon 9 monthly pan coefficient
where Eoff = estimated pond evaporation during the aeration OFF
period (mm/day); Eon = estimated pond evaporation during the aera-
tion ON period (mm/day); Δ Panoff = water-level change during the
| 3
aeration OFF period (mm/day); Δ Panon = water-level change during
the aeration ON period (mm/day); monthly pan coefficient previ-
ously measured for converting class-A pan evaporation to actual
pond evaporation at the SFC (Boyd, 1985).
Pond seepage was calculated as:
S ¼ DPondoff  Eoff
where S = pond seepage during 24 hr (mm/day); Δ Pondoff = change
in water level during the aeration OFF period (mm/day).
Total pond evaporation was estimated as shown below:
TE ¼ DPondon  S
where TE = total pond evaporation during the aeration ON period
(mm/day).
Evaporation resulting from aeration, as estimated from water-
level changes, was determined by the following equations:
EAWLðpanÞ ¼ TE  Eon
where EAWL(pan) = evaporation resulting from aeration for 24 hr
(mm/day). It was estimated from water-level change in the pond. In
this method of estimation, seepage was calculated based on ΔPanoff.
EAWLðonoffÞ ¼ DPondon  DPondoff
where EAWL(on-off) = evaporation resulting from aeration for 24 hr
(mm/day). In this method of estimation, seepage was considered to
be constant during aeration ON and aeration OFF periods.
Evaporation resulting from aeration as calculated based on heat
loss from ponds was determined with the following equation:
EAt ¼ ðDÞðT1  T2 Þ=½540 þ ð100  T1 Þ
where EAt = evaporation resulting from aeration for 24 hr (mm/day);
D = average depth of pond (mm); T1 = average water tempera-
ture = [(TS + T70)/2] of NEC pond (°C); T2 = average water tempera-
ture of aerated pond (°C); 540 = latent heat of vaporization of water
is 540 calories/g.
Calculation of pumping cost
Water for ponds at SFC was supplied by gravity flow, but in many
cases, water must be pumped into ponds. The cost of pumping
water to compensate water loss caused by aeration was determined
with the following equation (Yoo & Boyd, 1994):
PC ¼ QHSET=0:102 ep em
where PC = pumping cost ($); Q = discharge (m3/s); H = pumping
head (m); S = specific gravity (S = 1 for water); ep = pump efficiency;
em = motor efficiency; E = cost of electricity (kW/hr); T = pumping
time (hr).
4 |
3 | RESULTS AND DISCUSSION
Rainfall during the study was sufficient to maintain water levels and
provide a constant pond water surface area. Water levels were
never great enough to cause overflow, and water loss from ponds
resulted solely from evaporation and seepage.
Air temperature and pan evaporation
Means, maximum and minimum air temperature were calculated dur-
ing the ON and the OFF periods for aerators over each 24-hr period.
The means  SE for air temperature over a 24-hr period were
25.94°C  0.16 (N = 896) during the ON period for aerators, and
25.65°C  0.15 (N = 910) during the OFF period for aerators. Air
temperature ranged from 19.04 to 37.77°C (N = 904) during daytime
(12-hr period started at 7 a.m.), and from 16.33 to 28.76°C
(N = 902) during night-time (12-hr period started at 7 p.m.).
There was no difference between ON and OFF periods for aera-
tors with respect to measured daily air temperature (t = 0.05;
p = .9639; N = 50). The mean  SE for pan evaporation was
5.86 mm/day  0.51 during the ON period for aerators, and
5.42 mm/day  0.59 during the OFF period for aerators. There was
no difference between ON and OFF periods for aerators for pan
evaporation measurements (t = 0.55; p = .5816; N = 50).
Evaporation pans are widely used to measure evaporation. They
determine potential evaporation by providing a constant free water
surface (Brutsaert, 1982). Air temperature is closely correlated with
water temperature in ponds (Prapaiwong & Boyd, 2012). To deter-
mine the effect of aeration on water temperature and evaporation
rate, it is important for weather conditions to be comparable
between ON and OFF periods. Weather conditions obviously dif-
fered from day to day, but over the 50-day period, differences in air
temperature and pan evaporation did not result between ON and
OFF days for the aerators.
Water temperature
Means  SE for water temperature—at the surface and 70-cm
depth—in each pond were calculated (Table 1) for daytime and
night-time during each 24-hr period. Although all ponds of the
study were at the same location, and filled with water from the
same source, there were differences in water temperature among
these ponds while aerators were OFF. This variation might result
from differences in turbidity. To compare the true effects of differ-
ent aeration rates, unbiased by differences in initial water tempera-
ture during the OFF period, the treatment means should be
adjusted to the same initial water temperature during the OFF per-
iod. Two ANCOVA models were used to assess the effect of a
categorical independent variable (ponds with different aeration
rates) on a numerical-dependent variable (water temperature during
ON periods) while controlling for a numerical covariate (water tem-
perature during OFF periods). The first model was for TS, while the
other one was for T70.
ABDELRAHMAN AND BOYD
For the TS model, TS during aerator OFF periods had no signifi-
cant effect (F = 0.08; p = .7801) on the relationship between TS dur-
ing ON periods and different aeration rates. Therefore, the covariate
was removed from the model. Variances of TS during ON periods
were homogeneous among ponds (Levene’s F = 1.72; p = .1262),
while data were not normally distributed (Shapiro–Wilk W = 0.99,
p = .0065). Thus, the non-parametric Kruskal–Wallis test was per-
formed. The TS during ON periods was different for ponds with dif-
ferent numbers of Air-O-Lator units (Kruskal–Wallis test:
v2ð5Þ = 1512.49; p < .0001; N = 5376). Results of Tukey–Kramer-
adjusted post hoc comparisons (Table 1) showed that, during ON
periods, ponds with higher aeration rate had cooler TS. The C pond
had greater TS than the NEC pond (t = 4.45287; p = .0001). Both
control ponds—NEC and C—had higher TS than all aerated ponds.
For the T70 model, T70 during OFF periods had a significant
effect (F = 19.86; p < .0001) on the relationship between T70 during
ON periods and different aeration rates. Therefore, including the
covariate—T70 during OFF—in this ANCOVA model was essential.
The homogeneity of slopes assumption of the standard ANCOVA
was satisfied, which means slopes of regression lines for pond were
equal. Distribution of data was not normal (Kolmogorov–Smirnov
D = 0.0176, p = .0065). The T70 during ON periods was different for
ponds with different numbers of Air-O-Lator units (Kruskal–Wallis
test: v2ð5Þ = 1669.84; p < .0001; N = 5376). Results of Tukey–
Kramer-adjusted post hoc comparisons (Table 1) showed that both
control ponds had greater T70 than all aerated ponds. During
ON periods, ponds with higher aeration rate had lower T70.
Means of differences in water temperature over a 24-hr period
between each pond and the NEC pond during the ON period were
calculated. The mean  SE for the decrease in the TS caused by
having this aerator turned on for 24 hr was 0.92°C  0.09 for an
aeration rate of 9.2 kW/ha, 1.72°C  0.12 for an aeration rate of
18.4 kW/ha, 1.99°C  0.13 for an aeration rate of 27.6 kW/ha
and 2.84°C  0.13 for 36.9 kW/ha of aeration. The mean  SE of
this decrease in T70 was 1.26°C  0.07 for an aeration rate of
9.2 kW/ha, 2.20°C  0.12 for an aeration rate of 18.4 kW/ha,
2.52°C  0.12 for an aeration rate of 27.6 kW/ha and
3.42°C  0.16 for 36.9 kW/ha of aeration.
Four simple linear regression models were calculated (Figure 2)
to reveal the relationship between aeration rate (kW/ha) and the
mean decrease in water temperature in aerated ponds related to
that in the NEC. During daytime (Figure 2a), the aeration rate had a
significant linear effect on the mean decrease in TS (R2 = .9596;
p = .0207) and mean decrease in T70 (R2 = .9501; p = .0258). During
night-time (Figure 2b), the aeration rate had a significant linear effect
on the mean decrease in TS (R2 = .951; p = .0256) and mean
decrease in T70 (R2 = .9654; p = .0181). There was a clear effect of
increasing the number of Air-O-Lator units on decreasing either the
surface or bottom water temperature during either daytime or night-
time. In all aerated ponds, the decrease in T70 because of aeration
was greater than that in TS (t = 2.85; p = .0047; N = 200).
Shallow aquaculture ponds stratify thermally during daytime and
destratify during night-time (Boyd & Tucker, 1998). The unaerated,
ABDELRAHMAN AND BOYD | 5

T A B L E 1 Means  standard errors (SE) for water temperature—at the surface (TS) and 70-cm depth (T70)—in each pond over a 24-hr
period, daytime (12-hr period started at 7 a.m.), and night-time (12-hr period started at 7 p.m.)
Aeration Rate (kW/ha)

Duration Depth Period 0 (C) 0 (NEC) 9.2 18.4 27.6 36.9

24 hr TS ON 29.48a  0.06 29.02b  0.06 28.04c  0.06 27.16d  0.06 26.93d  0.06 26.08e  0.06
OFF 29.48  0.08 29.10  0.07 28.05  0.06 27.65  0.07 27.61  0.07 26.76  0.06
T70 ON 29.46a  0.07 29.43a  0.07 28.1b  0.06 27.09c  0.06 26.8d  0.06 25.89e  0.06
OFF 29.42  0.08 29.32  0.08 29.06  0.09 28.09  0.08 27.89  0.08 27.48  0.08
Daytime TS ON 29.78  0.09 29.19  0.09 28.19  0.09 27.41  0.08 27.17  0.09 26.29  0.09
OFF 29.65  0.11 29.19  0.11 28.00  0.09 27.74  0.10 27.76  0.10 26.75  0.09
T70 ON 29.95  0.10 30.05  0.11 28.44  0.09 27.37  0.08 27.11  0.09 26.27  0.09
OFF 29.72  0.12 29.70  0.12 29.75  0.14 28.58  0.12 28.36  0.12 28.14  0.13
Night-time TS ON 29.17  0.08 28.84  0.08 27.88  0.08 26.91  0.08 26.68  0.08 25.86  0.08
OFF 29.32  0.10 29.00  0.10 28.11  0.09 27.55  0.09 27.45  0.09 26.77  0.08
T70 ON 28.97  0.08 28.8  0.08 27.75  0.08 26.8  0.08 26.49  0.08 25.5  0.08
OFF 29.12  0.10 28.95  0.10 28.38  0.10 27.61  0.09 27.43  0.09 26.82  0.09

Means were tested by Tukey’s Studentized Range (HSD) test; entries indicated by the same letter in a row do not differ at p < .05.

nutrient-enriched pond (NEC) followed the expected trend; it strati-
fied thermally during daytime with deep water warmer than surface
water, but it destratified during night-time. Mechanical surface aera-
tion resulted in destratification of aerated ponds during daytime. The
greater the aeration rate during either day or night, the cooler were
water at surface and 70-cm depth. All aerated ponds had lower tem-
peratures during day and night and at the surface and at 70-cm
depth than did the NEC pond. Higher rates of aeration resulted in
cooler water at 70-cm depth than surface water.
Our findings are in agreement with those of Talati and Stenstrom
(1990) who concluded that aeration in wastewater treatment basins
is the single most important factor causing heat loss. They found
that evaporation by surface aeration resulted in approximately 50%
of the total heat loss. Results of the present study show that
increased aeration rate also lowers water temperature in aquaculture
ponds; an effect can be positive or negative depending on a farm’s
location, the season and the farmed species. Pond temperature has
been shown to affect growth rates of channel catfish (I. punctatus)
(Andrews, Knight & Murai, 1972; Andrews & Stickney, 1972; Stick-
ney & Andrews, 1971), brook trout (Salvelinus fontinalis) (Baldwin,
1957), Pacific white shrimp (L. vannamei) (Perez-Velazquez, Bray,
Lawrence, Gatlin & Gonzalez-Felix, 2001; Wyban, Walsh & Godin,
1995), tilapia, Oreochromis spp., and freshwater prawn, Macro-
brachium rosenbergii (Green & Popham, 2008). The production poten-
tial would be impaired when the water temperature falls outside the
optimum range for significant periods (Wickins & Lee, 2008). Thus,
during high temperature, aeration could benefit production by lower-
ing temperature towards the optimum range. On the other hand,
when temperature is low, aeration could lower production because
food consumption and conversion are poor when temperatures fall
below 16–18°C (Andrews & Stickney, 1972).
A 1°C decrease in temperature in water containing 1 mg/L total
ammonia-nitrogen at pH 8.0 will decrease the NH3-N concentration
by 0.004 mg/L (Zhou & Boyd, 2015). Also, rates of respiration and
photosynthesis are affected by water temperature (Tucker, 2005).
Different types of aeration systems in wastewater treatment ponds
cause different rates of heat loss (Talati & Stenstrom, 1990); there-
fore, further work is required to study the heat loss caused by differ-
ent aerator types in aquaculture ponds.
Evaporation rate
There were no differences in total evaporation (Table 2) among the
ponds aerated at 9.2 kW/ha, and the non-aerated ponds (NEC and
C). Total evaporation of the pond with an aeration rate of 18.4 kW/
ha was not different from pond NEC, but greater than from pond C.
Total evaporation rates of ponds with aeration rates of 27.6 and
36.9 kW/ha were greater than from the control ponds (v2 = 30.29;
p < .0001). Thus, increasing aeration rate resulted in greater total
pond evaporation.
Pond evaporation resulting from aeration for 24 hr was deter-
mined (Figure 3) by three methods. The EAWL(pan) and EAWL(on-off)
methods for calculating evaporation rate were based on water-level
measurements, while the EAt measurement was based on heat loss
from the ponds. For individual aerated ponds, there were no differ-
ences among the results of the three estimation methods (v2 = 1.09;
p = .5810). The pond with one aerator differed from the pond with
four aerators in evaporation as determined by the EAWL(pan) method
(v2 = 11.69; p = .0085) and the EAWL(on-off) method (v2 = 13.39;
p = .0039). While with EAt, ponds with two and three aerators did
not differ, the other two ponds differed in evaporation rate
(v2 = 60.93; p < .0001).
Estimation of aerator-induced evaporation by heat loss calcula-
tions provided the strongest evidence of greater evaporation in
ponds with higher aeration rates. This may have been because the
EAt estimation was based on the difference in the average water
temperature between each aerated pond and the NEC pond; while
EAWL(pan) and EAWL(on-off) were based on each daily estimate of the
6 | ABDELRAHMAN AND BOYD

(a) T A B L E 2 Means and standard errors (SE) for total pond

evaporation in nutrient-enriched control (NEC) pond, control (C)
pond and ponds that were aerated with Air-O-Lator units at
different aeration rate
Aeration rate
Pond (kW/ha) Total evaporation Mean (mm/day)  SE
y = 1.066 + 0.0733x
2 C 0 4.38  0.58d
R = .9501
p = .0258 NEC 0 4.52  0.45 cd

1 aerator 9.2 5.43  0.62 bcd

y = 0.4541 + 0.0641x 2 aerators 18.4 6.98  0.63abc

2
R = .9596 3 aerators 27.6 7.60  0.65ab
p = .0207
4 aerators 36.9 8.29  0.69a

Means were tested by Tukey’s Studentized Range (HSD) test; entries

indicated by the same letter in a column do not differ at p < .05.

(b)

y = 0.4141 + 0.0762x
2
R = .9654
p = .0181

y = 0.4418 + 0.0679x
2
R = .951
p = .0256
F I G U R E 3 Means and standard errors for pond evaporation
caused by different aeration rates for 24 hr. These values were
calculated by three different methods. The EAWL(pan) and EAWl(on-off)
values were based on water-level measurement, while EAt value was
based on heat loss from the water body

/ (Hargreaves, 2013). Water losses (m3 ha1day1) caused by different

F I G U R E 2 Relationships between the aeration rate (kW/ha), and
the mean for differences in water temperature—at the surface and
70-cm depth—between aerated ponds and the nutrient-enriched
control pond when aerators were operated continuously during
daytime (a), and night-time (b)
water-level change measured directly by aid of a hook gauge and
stilling well. A total of 48 daily readings—24 hourly records of water
temperature at the water surface and at 70-cm depth—were made
in each pond, while only one daily water-level reading was per-
formed in each pond. The estimation method which was based on
more readings—EAt—would likely have resulted in better estimates.
Aeration rates in earthen-lined aquaculture ponds range from 5
to 30 kW/ha, and even greater rates may sometimes be used in
plastic-lined ponds for biofloc-technology (Boyd & Tucker, 2014).
Some intensive tilapia systems are aerated with 75–112 kW/ha
aeration rates were calculated based on heat loss from aerated
ponds (Table 3). The evaporation loss increased linearly with greater
amount of aeration (R2 = .969; p = .015; Figure 4). The regression
equation was ^y = 7.0866 + 0.967x. Evaporation rate increased 32%
and 92% when ponds were aerated for 24 hr with one and four Air-
O-Lator aerators respectively (Table 3).
The costs of well water pumping to compensate the water loss
caused by continuous aeration for 100 days were $3.2/ha and $9.3/
ha for the aeration rates of 9.2 kW/ha and 36.9 kW/ha, respec-
tively, assuming a 5-m pumping head (Table 3). This cost increased
linearly with greater aeration rates, with R2 of .969. Of course, cost
also would be higher when pumping head is greater. Commercial
ponds usually are aerated with larger aeration units than used in this
research. Oxygen transfer rates for surface aerators increase with
the amount of water splashed into the air (Boyd, 1998; Boyd &
Tucker, 1998). Larger aerators have oxygen transfer rates about
ABDELRAHMAN AND BOYD | 7

T A B L E 3 Evaporation caused by aeration (mm/day) for 24 hr by twice those of the units used in this study (Boyd & Ahmad, 1987).
different aeration rates (kW/ha), volume of water loss caused by Thus, in commercial aquaculture, the water loss per unit of aeration
aeration (m3 ha1 day1), percentage of the increase in evaporation is likely greater than reported here.
caused by aeration (%) and the well water pumping cost ($/ha) to
The water depth in ponds does not affect the evaporation rate
compensate the water loss caused by aeration for 24 hr
because evaporation occurs at the water surface (Yoo & Boyd,
Aeration Aeration
1994). The percentage of evaporation loss per unit of water would
Parameter Unit rate (kW/ha) for 24 hr
be less for a deeper pond than for a shallower pond, but it is unclear
Mean  SEa mm/day 9.2 1.52  0.10
how this relationship would apply in aquaculture ponds.
18.4 2.73  0.15
Pond evaporation represented 66.2% of the total water loss from
27.6 3.14  0.17
a channel catfish pond in Alabama (Brown, Boyd & Chappell, 2012);
36.9 4.35  0.22
this percentage could be even higher in arid areas. Increasing evapo-
Volume of water loss m3 9.2 15.18 ration reduces water volume and can concentrate contaminants. It
ha1day1 18.4 27.27 can also increase water salinity significantly. Evaporation is a crucial
27.6 31.41 factor in the determining the required water supply for fish ponds in
36.9 43.55 both humid and arid climates (Boyd, 1986; Brutsaert, 1982). The
b
Increase of evaporation % 9.2 32.01 amount of water required to maintain water levels in ponds
18.4 57.49 increases in drier climates (Boyd, 1986), and consequently, increases
27.6 66.21 water use and pumping cost per unit of production.

36.9 91.82 Some farmers continue to aerate ponds at times when dissolved
oxygen concentrations are adequate. Restricting aerator use to only
Well water pumping costc $/ha 9.2 0.03
the period when oxygen supplementation is needed can conserve
18.4 0.06
water and lessen energy for both aeration and pumping, thereby
27.6 0.07
lowering production costs. The most efficient manner of balancing
36.9 0.09
aeration use with need is to install automated systems for turning
a
Mean  SE for evaporation caused by aeration. aerators on and off in response to dissolved oxygen concentration
b
Percentage of increase in evaporation as a result of aeration for 24 hr
(Boyd & Tucker, 2014). The benefit of aeration for increasing fish
was calculated as compared with pond evaporation (pan evaporation X
pan coefficient) during ON period. production greatly exceeds the pumping cost to replenish water loss
c to greater evaporation. Producers should not lessen aeration use to
Well water pumping cost was estimated assuming electricity cost is 0.1
$/kWh, discharge is 0.032 m3/s (500 gpm), pumping head is 5 m, pump the point that it results in excessively low dissolved oxygen concen-
efficiency is 0.80 and motor efficiency is 0.80.
tration and fish stress.

Water turbidity
There was no difference in turbidity between the ON and OFF peri-
ods (t = 0.15; p = .8822). In addition, the change in turbidity during
ON days than that in OFF days was not different among ponds with
different numbers of Air-O-Lator units and the control ponds
(F = 0.22; p = .9521). Thus, the type of aerators used in this study
does not increase water turbidity significantly.
y = 7.0866 + 0.967x During the study, the means  SE for turbidity of the C and
2
R = .969 NEC ponds were 0.99 NTU  0.04 and 2.21 NTU  0.18 respec-
p = .015 tively. The NEC pond was more turbid than the C pond (t = 6.5;
p < .0001). Mean  SE for chlorophyll a of the pond C and the NEC
pond were 3.49 lg/L  0.32 and 14.79 lg/L  1.11 respectively
(t = 9.8; p < .0001). Fertilization resulted in greater phytoplankton
abundance that resulted in the increased turbidity. Evaporation
caused by the turbidity increase was calculated based on the tem-
perature difference between the NEC and C ponds. Evaporation
/ rates in the NEC pond were more than those in the C pond by
1.34  0.09 mm 12 hr1 during daytime and 1.25  0.10 mm
F I G U R E 4 Relationship between the aeration rate (kW/ha) and 12 hr1 at night. Greater turbidity as a result of fertilization
the mean volume of water loss (m3 ha1day1) caused by aeration
increased evaporation rates.
using 0.37-kW Air-O-Lator units for 24
8 |
The practical use of aerators in aquaculture is based on applying
enough aeration to avoid dissolved oxygen concentration from
declining below an acceptable level of about 3–4 mg/L for most spe-
cies (Boyd, Torrans & Tucker, 2017). Although the effect of aeration
on evaporation is usually not considered by aquaculturists. However,
the findings of this study suggest that aeration does increase the
evaporation rate. There are some steps that can be taken to lessen
the evaporation rate in aerated ponds. Diffuser aerators or aerators
that operate on the venturi principle aerate by releasing bubbles into
the water and do not cause as much disturbance of the water sur-
face to increase evaporation rate. But, these aerators usually are less
effective than paddlewheel and vertical pump aerators in ponds
(Boyd, Torrans & Tucker, 2017). Nevertheless, aeration typically is
not required or required at a lesser rate in daytime than at night in
ponds. Water could be conserved by aerating ponds only when the
dissolved oxygen concentration is low enough to negatively impact
fish, shrimp or other cultured species. The most reliable means of
controlling aerator operation to periods of undesirable low dissolved
oxygen concentration is to use an automated system that turns
aerators on and off in response to concentration monitored by an
oxygen sensor that communicates with the aerator (Jescovitch, Boyd
& Whitis, 2017; McGraw, Teichert-Coddington, Rouse & Boyd,
2001).
In conclusion, increased aeration rate increases the evaporation
rate, which in turn, lowers water temperature. A nutrient-enriched
pond evaporates more than does an unfertilized pond. Air-O-Lator
units do not increase water turbidity. Economic benefits from
reduced pumping costs may be obtained by restricting aeration to
periods of need, particularly in regions where the water is scarce.
There are many kinds of aerators and commercial ponds are usu-
ally aerated with larger units than used in this research; therefore,
further work is required to study the heat loss and water loss caused
by different aerator types.
ACKNOWLEDGMENTS
Thanks are extended to the fellow students in the water quality lab-
oratory at Auburn University. We also thank the management team
of the E. W. Shell Fisheries Center, North Auburn research station
for their support. Hisham Abdelrahman also extends his appreciation
to the Egyptian Cultural and Educational Bureau (ECEB) in Washing-
ton, DC for financial aid to support his study abroad.
ORCID
Hisham A Abdelrahman http://orcid.org/0000-0003-1076-5536
REFERENCES
Andrews, J. W., Knight, L. H., & Murai, T. (1972). Temperature require-
ments for high density rearing of channel catfish from fingerling to
market size. The Progressive Fish-Culturist, 34, 240–241. https://doi.
org/10.1577/1548-8659(1972)34[240:TRFHDR]2.0.CO;2
ABDELRAHMAN AND BOYD
Andrews, J. W., & Stickney, R. R. (1972). Interactions of feeding rates
and environmental temperature on growth, food conversion, and
body composition of channel catfish. Transactions of the American
Fisheries Society, 101, 94–99. https://doi.org/10.1577/1548-8659
(1972)101&lt;94:IOFRAE&gt;2.0.CO;2
Baldwin, N. (1957). Food consumption and growth of brook trout at dif-
ferent temperatures. Transactions of the American Fisheries Society,
86, 323–328. https://doi.org/10.1577/1548-8659(1956)86[323:FCA
GOB]2.0.CO;2
Boyd, C. E. (1985). Pond evaporation. Transactions of the American Fish-
eries Society, 114, 299–303. https://doi.org/10.1577/1548-8659
(1985)114&lt;299:PE&gt;2.0.CO;2
Boyd, C. E. (1986). Influence of evaporation excess on water require-
ments for fish farming. In Proceedings of the conference on climate and
water management—A critical era (pp. 62–64). Boston, MA: Meteoro-
logical Society.
Boyd, C. E. (1998). Pond water aeration systems. Aquacultural Engineering,
18, 9–40. https://doi.org/10.1016/S0144-8609(98)00019-3
Boyd, C. E., & Ahmad, T. (1987) Evaluation of aerators for channel catfish
farming. Bulletin 584, Alabama Agricultural Experiment Station, Auburn
University, Alabama, 52.
Boyd, C. E., McNevin, A. A., Racine, P., Tinh, H. Q., Minh, H. N., Viriya-
tum, R., . . . Engle, C. (2017). Resource use assessment of shrimp,
Litopenaeus vannamei and Penaeus monodon, production in Thailand
and Vietnam. Journal of the World Aquaculture Society, 48, 201–226.
https://doi.org/10.1111/jwas.12394
Boyd, C. E., Torrans, E. L., & Tucker, C. S. (2017). Dissolved oxygen and
aeration in ictalurid catfish aquaculture. Journal of the World Aquacul-
ture Society, 49, 7–70.https://doi.org/10.1111/jwas.12469
Boyd, C.E., & Tucker, C.S. (1992). Water quality and pond soil analyses for
aquaculture. Auburn, AL: Alabama Agricultural Experiment Station,
Auburn University.
Boyd, C. E., & Tucker, C. S. (1998). Pond aquaculture water quality man-
agement. Boston, MA: Kluwer Academic Publishers. https://doi.org/
10.1007/978-1-4615-5407-3
Boyd, C., & Tucker, C. (2014). Handbook for aquaculture water quality.
Auburn, AL: Craftmaster Printers.
Brown, T. W., Boyd, C. E., & Chappell, J. A. (2012). Approximate water
and chemical budgets for an experimental, in-pond raceway system.
Journal of the World Aquaculture Society, 43, 526–537. https://doi.
org/10.1111/j.1749-7345.2012.00576.x
Brutsaert, W. (1982). Evaporation into the atmosphere: Theory, history and
applications. Dordrecht, Netherlands: Kluwer Academic. https://doi.
org/10.1007/978-94-017-1497-6
Busch, C., Flood Jr, C. , & Allison, R. (1978). Multiple paddlewheels’ influ-
ence on fish pond temperature and aeration. Transactions of the
American Society of Agricultural Engineers, 21, 1222–1224. https://doi.
org/10.13031/2013.35473
Chatvijitkul, S., Boyd, C. E., Davis, D. A., & McNevin, A. A. (2017). Pollution
potential indicators for feed-based fish and shrimp culture. Aquaculture,
477, 43–49. https://doi.org/10.1016/j.aquaculture.2017.04.034
Green, B., & Popham, T. W. (2008). Probabilities of low nighttime
temperatures during stocking and harvest seasons for inland shrimp
culture. Journal of the World Aquaculture Society, 39, 91–103.
https://doi.org/10.1111/j.1749-7345.2007.00135.x
Hargreaves, J. A. (2013). Biofloc production systems for aquaculture. South-
ern Regional Aquaculture Center (SRAC) Publication No. 4503.
Harman, R. (2005). The Water Cycle: Evaporation Condensation and Ero-
sion. Chicago, IL: Heinemann Library.
Hopkins, J. S., Hamilton, R. D., Sandier, P. A., Browdy, C. L., & Stokes,
A. D. (1993). Effect of water exchange rate on production,
water quality, effluent characteristics and nitrogen budgets of
intensive shrimp ponds. Journal of the World Aquaculture
Society, 24, 304–320. https://doi.org/10.1111/j.1749-7345.1993.tb
00162.x
ABDELRAHMAN AND BOYD
Idso, S. B., & Foster, J. M. (1974). Light and temperature relations in a
small desert pond as influenced by phytoplanktonic density varia-
tions. Water Resources Research, 10, 129–132. https://doi.org/10.
1029/WR010i001p00129
Jescovitch, L. N., Boyd, C. E., & Whitis, G. N. (2017). Effects of mechani-
cal aeration in the waste-treatment cells of split-pond aquaculture
systems on water quality. Aquaculture, 480, 32–41. https://doi.org/
10.1016/j.aquaculture.2017.08.001
Levene, H. (1960). Robust tests for equality of variances. In I. Olkin
(Ed.), Contributions to probability and statistics: Essays in Honor of
Harold Hotelling (pp. 278–292). Stanford, CA: Stanford University
Press.
McGee, M. V., & Boyd, C. E. (1983). Evaluation of the influence of water
exchange in channel catfish ponds. Transactions of the American Fish-
eries Society, 112, 557–560. https://doi.org/10.1577/1548-8659
(1983)112&lt;557:EOTIOW&gt;2.0.CO;2
McGraw, W., Teichert-Coddington, D. R., Rouse, D. B., & Boyd, C. E.
(2001). Higher minimum dissolved oxygen concentrations increase
penaeid shrimp yields in earthen ponds. Aquaculture, 199, 311–321.
https://doi.org/10.1016/S0044-8486(01)00530-0
Perez-Velazquez, M., Bray, W. A., Lawrence, A. L., Gatlin, D. M., &
Gonzalez-Felix, M. L. (2001). Effect of temperature on sperm qual-
ity of captive Litopenaeus vannamei broodstock. Aquaculture, 198,
209–218. https://doi.org/10.1016/S0044-8486(01)00510-5
Prapaiwong, N., & Boyd, C. E. (2012). Water temperature in inland,
low-salinity shrimp ponds in Alabama. Journal of Applied Aquacul-
ture, 24, 334–341. https://doi.org/10.1080/10454438.2012.731
479
Steel, R. G., & Torrie, J. H. (1980). Principle and procedures of statistic: A
biometrical approach, 2nd ed.. New York, NY: McGraw-Hill.
Stickney, R. R., & Andrews, J. W. (1971). Combined effects of dietary
lipids and environmental temperature on growth, metabolism and
| 9
body composition of channel catfish (Ictalurus punctatus). Journal of
Nutrition, 101, 1703–1710. https://doi.org/10.1093/jn/101.12.1703
Talati, S., & Stenstrom, M. (1990). Aeration-basin heat loss. Journal of
Environmental Engineering, 116, 70–86. https://doi.org/10.1061/
(ASCE)0733-9372(1990)116:1(70)
Tucker, C. (2005). Pond aeration. Southern Regional Aquaculture Center
(SRAC), Publication No. 3700.
Tucker, C. S., & Hargreaves, J. A. (2004). Pond water quality. In C. S.
Tucker, & J. A. Hargreaves (Eds.), Biology and Culture of Channel Cat-
fish (pp. 216–278). Amsterdam, The Netherlands: Elsevier.
Wickins, J. F., & Lee, D. O. C. (2008). Crustacean farming: Ranching and
culture. Ames, IA: Blackwell Science.
Wyban, J., Walsh, W. A., & Godin, D. M. (1995). Temperature effects on
growth, feeding rate and feed conversion of the Pacific white shrimp
(Penaeus vannamei). Aquaculture, 138, 267–279. https://doi.org/10.
1016/0044-8486(95)00032-1
Yoo, K., & Boyd, C. (1994). Hydrology and water supply for aquaculture.
New York, NY: Chapman and Hall. https://doi.org/10.1007/978-1-
4615-2640-7
Zhou, L., & Boyd, C. E. (2015). An assessment of total ammonia nitrogen
concentration in Alabama (USA) ictalurid catfish ponds and the possi-
ble risk of ammonia toxicity. Aquaculture, 437, 263–269. https://doi.
org/10.1016/j.aquaculture.2014.12.001
How to cite this article: Abdelrahman HA, Boyd CE. Effects
of mechanical aeration on evaporation rate and water
temperature in aquaculture ponds. Aquac Res. 2018;00:1–9.
https://doi.org/10.1111/are.13674