See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/280901443

Comparative transparency, depth of mixing, and stability of

stratification in lakes of Cameroon, West Africa

Article in Limnology and Oceanography · January 1988

CITATIONS READS

149 487

1 author:

George W. Kling
University of Michigan
221 PUBLICATIONS 14,977 CITATIONS

SEE PROFILE

All content following this page was uploaded by George W. Kling on 05 December 2015.

The user has requested enhancement of the downloaded file.

Limnol. Oceanogr., 33(1), 1988, 27-40
© 1988, by the American Society of Limnology and Oceanography, Inc.

Comparative transparency, depth of mixing, and stability of

stratification in lakes of Cameroon, West Africa!
George W. Kling?
Department of Zoology, Duke University, Durham, North Carolina 27706

Abstract
Morphometry, oxygen concentration, temperature, and transparency were studied in 39 natural
lakes in Cameroon, West Africa. Thermal profiles from 31 of the lakes and data from published
studies were used to calculate stability of thermal stratification and evaluate morphological cor-
relates of mixing depth. Twenty-six lakes showed some degree of stratification and 17 had distinct
thermoclines and well-developed, anoxic hypolimnia. Total stability of the water column ranged
from 0 to 5,784 J m-*. The high values were similar to or greater than those of other tropical and
temperate lakes. Lake depth seems to exert a stronger influence on stability than does lake area,
but depth or stability measures alone provide little information about heat distribution or mixing
regime. A strong positive relationship between water transparency and thermocline depth in both
tropical and temperate lakes suggests that reductions in buoyant resistance to vertical mixing,
caused by deeper penetration of solar radiation, are important in establishing mixing depths in
various lakes. Comparisons of persistent thermocline depth in tropical vs. temperate lakes, re-
gardless of size, indicate that mixing depths in tropical lakes are often greater than those in their
temperate counterparts. This difference is caused in part by the narrow ranges of temperature and
smaller absolute density gradients in these tropical lakes, against which the mixed layer is deepened.

Knowledge concerning the formation and summer thermocline depth and lake fetch
maintainence of the upper mixed layer in or area (Gorham and Boyce pers. comm.,;
oceans and lakes has expanded greatly in Patalas 1984).
recent years (Phillips 1977; Imberger 1985). Given these advances in understanding
Entrainment of deeper, denser water into of the structure and energetics of mixed lay-
the base of the epilimnion is the primary ers, it is surprising that in tropical lakes even
mechanism of mixed-layer thickening. The the most fundamental information, such as
turbulent kinetic energy required has been persistent thermocline depth and stability
shown to come from stirring processes such of stratification, is so limited (Lewis 1984).
as wind mixing or penetrative convection In Lake Titicaca heat transfers have been
(Kraus and Turner 1967; Foster 1971; found to dominate diurnal mixing (Vincent
Woods and Barkmann 1986) and from shear et al. 1984), and Henderson-Sellers (1984)
production within boundary layers of the showed that horizontal advection con-
epilimnion (Pollard et al. 1973; Dillon and trolled mixing depths in a subtropical res-
Powell 1979; Price et al. 1986). When cou- ervoir. MacIntyre and Melack (in press) dis-
pled with water transparency and meteo- cussed the frequent, nonseasonal mixing of
rological data, these mechanistic ap- shallow Amazon floodplain lakes, and Wood
proaches can be used to model and predict et al. (1976) presented a climatological study
the vertical transport of heat and momen- of seasonal mixing and stratification in sev-
tum on both diurnal and seasonal time scales eral Ethiopian crater lakes. Finally, Lewis
(Holloway 1980; Woods 1980; Spigel et al. (1973, 1983) suggested that the small den-
1986). Morphological features may also in- sity differences across narrow temperature
fiuence the integrated response of mixed ranges, typically found in tropical lakes, plus
layers to climate, as shown by the positive small Coriolis effects, could explain the deep
correlation in temperate lakes between and variable mixing depths in lakes Lanao
and Valencia. Certainly the basic mecha-
nisms of mixed-layer thickening, i.e. stirring
' Supported by NSF grants BSR 84-00532 to D. A. from above and shear generated at the base
Livingstone and EAR 87-09978 to G.W.K., and by the
French government agencies ORSTOM and CNRS.
of the epilimnion, occur at all latitudes. It
? Present address: The Ecosystems Center, Woods is the relative importance of these processes
Hole, Massachusetts 02543. that we know little about. Regardless of the
27
28 Kling

NGAOUNDERE Ministry of Higher Education and Scientific

13°30" 14°00’
Research (MESRES) and the French over-
b 336 47°30" 1° seas research organization (ORSTOM) for
their cooperation in this research, especially
330 oe {ss CRH-MESRES for use of an inflatable boat
; 40 SS and ORSTOM-Yaounde for use of a vehi-
cle. J. C. Stager, A. Zogning, and Ndoni Paul
-7° assisted in the field. K. Kelts provided the
seismic records. The suggestions of W. M.
Lewis, Jr., D. A. Livingstone, S. MacIntyre,
C. H. Mortimer, W. H. Schlesinger, and an
anonymous reviewer improved the manu-
3°
script.

Description of the study area

Cameroon lies on the coast of West Africa
between 8° and 16°E long (Fig. 1). Nearly
6°00’ all of its lakes lie on or close to the central
highland backbone of the country. Excep-
18 tions include deltaic lakes at the mouth of
1S,
9°00’ 9°30’ . 120 19138 the Sanaga River and lakes beside the Benue
SOUTHWEST 1o°DO I
10480" River in the north. There are three lake dis-
BAMENDA tricts, each of which was included in the
Fig. 1. Location map. Numbers correspond to sampling program: the Southwest, the Ba-
names of lakes in Table 1. menda, and the Ngaoundere (Fig. 1; Table
1). The lakes are distributed from 3°47’ to
7°26'N lat, and range in elevation from Ossa
dominant process of mixed-layer thicken-
at sea level to Bambili at 2,264 m asl (Table
ing, however, a given energy input applied
1).
to a smaller density gradient will produce a
At least 34 of the 39 study lakes have
deeper mixed layer.
formed in volcanic basins. These lakes oc-
The present study of 39 natural lakes in
cur mainly in the trachytic and Recent ba-
Cameroon, West Africa, was designed to
saltic series, although their precise ages are
test the hypothesis that persistent thermo-
unknown. Many of the lakes are maars, sur-
cline depths are deeper in tropical lakes than
rounded by cliffs and tuffrings of pyroclastic
would be expected in similar temperate
material from 5 to 200 m high (Table 1).
lakes, to examine the morphometric fea-
Monoun and Nyos are two volcanic lakes
tures of these lakes as they relate to overall
of special interest, as they are the sites of
stability of the water column, and to identify
lethal gas releases in 1984 and 1986 (Kling
patterns of light attenuation within and be-
et al. 1987).
tween lakes and test their relationship to
mixing and stratification. The Cameroon
lakes, which were studied during February— Methods
October 1985, September 1986, and May Depths were measured along transects
1987, offer great diversity and are thus ideal with a sounding line or a seismic recorder
for comparative research (Kling 1987a). and transceiver or were taken from Hassert
Previous studies of physical limnology in (1912). The geographic features presented
these lakes have been restricted to the in Table 1 were obtained from maps. A YSI
bathymetric survey of eight lakes (Hassert model 46 telethermometer, readable to
1912) and to observations of temperature 0.05°C, and thermistor probe were used to
in three lakes (Green et al. 1973, 1974; Cor- measure the vertical distribution of tem-
bet et al. 1973). perature. The meter and probe were cali-
I thank many people of the Cameroon brated frequently against a certified ther-
 Lake mixing in Cameroon 29

Table 1. Geographic and morphometric features of Cameroon lakes: median crater rim height above the
lake surface—rim; maximum depth—z,,,,; mean depth—zZ; volume development— 32/z,,,,; relative depth—z,.

Rim Zmax 4

Lake Elev (m asl) Area (ha) (m) 32/Zmax z, (%)

1 Ossa 1 1,300 10 - - - -
2 Debundsha 54 6 40 13.5 - - 4.88
3 S. Debundsha 54 1 10 - - - -
4 Mboandong 143 20 5 - - - -
5 B. Kotto 106 140 30 5.5 4.5 2.45 0.34
6 Barombi Mbo 301 415 100 110 68.7 1.87 4.78
7 Disoni 455 165 100 80 41.2 1.54 5.52
8 Manengouba-F 1,920 22 100 168 61.8 1.10 31.7
9 Manengouba-M 1,900 2 70 92 55.0 1.79 57.6
10 Beme 500 60 60 14.5 - - 1.66
11 Mfouet 1,120 7 70 14 - - 4.69
12 Baleng 1,374 8 70 52 - - 16.3
13. Mfou 1,350 8 200 58 - - 18.2
14 Negop Ghang 1,120 8 70 23 - - 7.21
15 Monoun 1,080 53 20 96 25.8 0.81 11.7
16 Banefo 1,100 6 20 64 - - 23.2
17 L. Monoun 1,100 25 30 49 - - 8.69
18 Petponoun 1,120 30 3 12 - - 1.94
19 Nchout 1,140 13 40 16 - - 3.93
20 Ejagham 100 70 10 17 11.4 2.01 1.80
21 Bambuluwe 2,053 28 50 58 39.3 2.03 9.71
22 Bambili 23264 28 100 4 - - 1.49
23 Oku 2,227 243 80 52 32.0 1.85 2.96
24 Enep 697 50 200 78 - - 9.78
25 Elum 960 50 120 35 - - 4.39
26 Wum 1,177 45 20 124 48.4 1.17 15.5
27 Nyi 1,316 50 50 47 - - 5.89
28 Nyos 1,091 158 70 208 111.7 1.61 14.7
29 Njupi 1,020 30 10 - - - -
30 Gegouba 1,180 20 30 104 - - 20.6
31 Ngaoundaba 1,160 10 30 62 - - 17.4
32 Baledjam 1,249 25 15 13 - - 2.3
33 Tizong 1,160 8 70 48 26.2 1.64 15.0
34 Massot 1,054 5 5 - - - -
35 Mbalang 1,130 50 40 52 30.0 1.73 6.52
36 Dang 1,079 80 5 - - - -
37 Bini 1,079 20 5 - - - -
38 Benakuma 576 154 200 138 - - 9.86
39 Edib 1,280 2 60 12.5 - - 7.83

mometer traceable to the National Bureau in wEinst m~? s~!, Diffuse attenuation coef-
of Standards. Water for oxygen analysis was ficients (k, m~') were taken as the slope es-
collected with a 4-liter Van Dorn sampler, timate of the regression z on In(H,,), where
fixed immediately with Winkler reagents Hi, is the downwelling PAR irradiance at
when brought to the surface, and titrated depth z. Underwater light penetration was
within 6 h with phenylarsine oxide. Elec- also estimated with a 20-cm Secchi disk.
trical conductivity was measured in the field Lake stability (S) was calculated follow-
with a 0.1 scale cell, and the values were ing Idso (1973) and expressed in J m~.
later corrected to 25°C (K35). Water density increases due to chemical
Light attenuation was measured on clear stratification were determined with the
days between 1100 and 1200 hours with a chemistry values in Kling (1987a) and Kling
recently calibrated LiCor 185 quantum me- et al. (1987) and following the method pre-
ter and 2-pi cosine sensor. The photon flux sented in MacIntyre and Melack (1982) and
density of PAR (400-700 nm) was recorded Chen and Millero (1977). The waters of lakes
 30 Kling

Table 2. Secchi depth and diffuse attenuation coef- Results

ficients (k). N is the number of points used in regres-
sions to calculate k. Values in parentheses are the stan- Morphometric features—The morpho-
dard errors of the estimates of k. metric features for 32 of the 39 lakes are
listed in Table 1 (lake numbers correspond
Lake Secchi (m) N k(m™)
to numbers in Fig. 1). The smallest lake
Barombi Mbo surveyed by boat was Edib (2 ha), and the
14 Feb 85 11.0 41 0.148(0.004) largest was Barombi Mbo (415 ha). Lake
Barombi Mbo
26 Feb 85 - 29 0.163(0.004)
Ossa has the greatest surface area (1,300 ha)
Barombi Mbo of any lake in the country except Lake Chad
27 Feb 85 - 31 0.169(0.004) and the reservoirs Bamendjing and Mba-
Oku 10.9 26 0.178(0.002) kaou; the median area of the 39 study lakes
Disoni 10.0 31 0.158(0.006) is only 28 ha. The deepest is Nyos (208 m),
Benakuma 7.80 - -
Mfou 5.60 - — although it was not surveyed by boat before
Gegouba 5.55 35 0.300(0.007) the 21 August 1986 gas release. Visual evi-
Negop Ghang 5.30 - — dence for annual lake level fluctuations of
Wum 5.15 53 0.305(0.006) 1-2 m was common. Most lakes were closed
Baleng 5.00 - -
Manengouba-F 5.00 45 0.262(0.003)
in the dry season and appeared to have sur-
Banefo 4.50 - - face outflow in the wet season.
Beme 4.50 23 0.353(0.004) The volume development of a lake is the
Ejagham 3.95 69 0.425(0.004) ratio of lake volume to that of a cone of
Tizong 3.95 77 0.343(0.002)
equal area and height and is given by three
Petponoun 3.55 61 0.717(0.010)
Nchout 3.50 - - times the ratio of mean depth (Z) to maxi-
Mbalang 3.45 24 0.267(0.006) mum depth (z,,) (Hutchinson 1957). Values
Ngaoundaba 3.45 53 0.427(0.008) approaching 1.8 exceed those of most lakes
L. Monoun 3.10 — — but are typical for small, deep crater lakes.
Baledjam 2.96 63 0.675(0.007)
Bambuluwe 2.45 51 0.517(0.010)
The value of 2.45 for Lake Kotto is extraor-
Debundsha 2.15 45 0.534(0.006) dinary and results from a shoreline drop-
Monoun 1.25 43 1.88 (0.024) ping steeply to a flat, expansive bottom.
Mfouet Relative depth (z,) is the maximum depth
20 May 87 1.20 _ -
as a percentage of the mean lake diameter
Mfouet
16 May 85 0.05 - - (Hutchinson 1957). The center of a lake is
Edib 0.80 _ _ usually closer to the bottom than to any
B. Kotto 0.75 36 1.70 (0.025) shore. Apparently the Manengouba lakes
have the fourth and fifth largest z, values
known (58 and 32%), after Lake Kauhako
Monoun and Nyos can be considered meta- in the Hawaiian islands (z, = 375%: Maci-
stable and were highly charged with CO). olek 1975), Lake Tritriva in Madagascar
Chemical stability in these two lakes was (z, = 89%: Kling 1987a), and Lake Kany-
calculated ignoring the density contribution angeye in Uganda (z, = 71.5%: Melack
of this gas. This is necessary because mea- 1978).
surements of CO, in situ are not available Transparency —Secchi depths ranged
and because the fundamental fluid dynam- from 5 cm in Lake Mfouet to 11 m in Ba-
ics equations are inadequate to describe rombi Mbo (Table 2). The depth at which
water movement in a gas-charged fluid. For 1% of the surface irradiance of PAR remains
example, the effervescence of CO, as water (Zo01) defines the euphotic zone (Talling
is brought from depth to the surface will 1971). The Zo, corresponding to the un-
produce not only a large increase in bulk usually low Secchi depth of the algal-rich
buoyancy but also an increase in turbulence. Lake Mfouet is unknown; the lowest mea-
Water column stability will thus be reduced sured Z 99, (2.61 m) was in Lake Monoun
independent of external energy inputs. and the highest (29.67 m) was in Barombi
Birge’s (1916) work of the wind was cal- Mbo.
culated after Hutchinson (1957). Attenuation coefficients ranged from
 Lake mixing in Cameroon 31

Density, gcm75 -----

(9961 —«.9962 .996) .9963
t '
'

{ lot ! Barombi Mbo |

rs '
x 27 Feb
‘
20} K
€ ee
- : 02. 04
=
a® 9964 .997) .9977
Qa T

Monoun 4

14 May
2b |
L

4t |
Ol * 03 15 * 25
Attenuation Coefficient, m=! —
Fig. 2. Lagged attenuation coefficient (k,) for thin water layers and total density of water plotted against lake
depth (see text).

0.158 in Lake Disoni to 1.88 in Lake Mo- variations in the underwater light field have
noun (Table 2). Coefficients for specific water been observed in other strongly stratified
layers were calculated by comparing each lakes (Bowling and Tyler 1986).
H, to the irradiance in a layer of water just Temperature—The temperatures of sur-
above it instead of to the irradiance at the face waters ranged from 18.6°C in Oku, at
surface. For example, the irradiance value 2,227 m to 32.5°C in Debundsha at 54 m.
H, in any lake was compared to H, instead Diurnal changes as high as 1.8°C at the sur-
of to Hy in determining k (Fig. 2). These face were recorded in Barombi Mbo, and
values best illustrate variation related to changes of 1°C were common in other lakes.
particulate material and mixing. In several In addition, there are seasonal and annual
lakes there was a sharp discontinuity and changes in surface temperatures. A 9-month
sudden increase in k between the epilimnion thermal record for Barombi Mbo (Kling
and the metalimnion that may be related to 1987) showed that the early-morning sur-
plankton or to an accumulation of particles face temperature dropped steadily from
in strong density gradients (Fig. 2). In Lake 28.65°C in February to 26.65°C in Septem-
Monoun, for example, water samples in- ber, closely following the annual cycle of air
dicated a large concentration of Chaoborus temperature and insolation. The surface
suspended at 2.5 m, but not at 1.5 or 3.5 temperature was 29.5°C in April 1972
m. The attenuation profiles for specific lay- (Green et al. 1973).
ers also reveal a slight increase in k for Ba- Thermal stratification—Of the 31 lakes
rombi Mbo near the top of the metalimnion whose thermal profiles were taken, 26
(Fig. 2). The effect moved from about 22.5 showed some degree of stratification. Sev-
m on 14 February up to 15 m on 26 Feb- enteen of these lakes had distinct thermo-
ruary, in concert with a rising metalimnion, clines and well-developed hypolimnia (Figs.
but by 27 February the disturbance was po- 3-6). All lakes >18 m deep showed pro-
sitioned just above the metalimnion and was nounced stratification. No lakes < 18 m deep
smaller. There was also a second distur- formed hypolimnia, although six had well-
bance at 21 m on this last date. Similar defined epilimnia and thermoclines (Figs.
 32 Kling

TEMPERATURE, °C TEMPERATURE, °C
0 20 22 24es 19 21
aa: 23 r
26 278

10F
E r ra 4
> i
x 20+ 1 4
B
a te
4
4 L 4
Barombi Mbo 1404 4

14 Feb 85 a } Dison 4
30 " Tizong 4a ob Mbalang 4 0930 7
28 May 85 30 May 85 17 Feb 85
r 0830 7 fF 0845 7 80; 760, 1040

40- 7 fF 7
Bo, gi Lee
246 8 246 8 (005466 802
4 6 8
OXYGEN, mg L" OXYGEN, mg L'

TEMPERATURE, °C TEMPERATURE , °C
19 2i 23. 2d o-2-+-20 5
1 ay 1
Tef
por Te a
e 20r EO.
E P € “ef af 4
xE - v §
a. z= 40 ;3
ry]
a a f
wf
3018 Gegouba Ngaoundaba | 60} f Manengouba-+
3 Jun 85 2 Jun 85
L 19 Feb 85 |
1000 7 + 0900 1150
40) 4 tf 4 80;
104i
246
ii. FF Leiria
0246 8
2
Low,
166 2468
OXYGEN, mg L’!
OXYGEN, mg L"
Fig. 3. Temperature and dissolved oxygen plotted
Fig. 4. As Fig. 3.
against depth for the dates and times shown in rep-
resentative lakes. Stippled area represents O, satura-
tion > 100% with respect to the surface. Thermal and
oxygen data for lakes 11, 12, 13, 14, 16, 17, 19, 24, ing or sensible heat loss could explain the
38, and 39 (see Table 1) are given elsewhere (Kling slight surface instability of Beme, Ejagham,
1987a). Oku, and Tizong (Figs. 3-5), while very
heavy rains the evening before the second
5, 6). Lake Kotto also was not stably strat- profile was taken in Ejagham may have con-
ified in March or April 1972 (Green et al. tributed to its inverse stratification. The diel
1973, 1974). Five lakes had very striking variation in Ejagham is apparent from the
metalimnia at depths between 8 and 12 m profile taken in late afternoon compared to
(temperature gradients of 2.0-3.5°C: Figs. that of the following morning. Superficial
3, 5, and 6). Much deeper metalimnia (28— diurnal stratification was clearly illustrated
36 m) were found in Barombi Mbo and Di- in Manengouba, Ejagham, and Oku (Figs.
soni (Fig. 4). Green et al. (1973) reported 4, 5), and is common in lakes exposed to
that Barombi Mbo had a metalimnion intense solar heating (see Wetzel 1983; Vin-
around 18 m and was strongly stratified in cent et al. 1984). The unusual profile for
April 1972. The greatest temperature range Monoun was distinguished by a maximum
was in Lake Monoun, where the tempera- temperature at the surface, a minimum of
ture decreased from 26.70°C at the surface 20.20°C at 10 m, and an increase to 21.80°C
to 20.40°C at 8 m, which included a single at 74 m (Fig. 5). Such dichothermic profiles
drop of 2.5°C between 3.5 and 4 m. have been observed in other tropical (Bea-
Inverse stratification at the surface was dle 1981) and temperate (Hutchinson 1957)
observed in several lakes. Evaporative cool- lakes.
 Lake mixing in Cameroon 33

TEMPERATURE, °C riod of surface warming in a time of low

24 25
wind stress during which a second ther-
T
mocline develops above the first. The clear-
120) est example is in Disoni, where a thermo-
cline at 29 m formed over a deeper and
g 77 Monoun
m

207 44or
sharper thermocline at 33.5 m (Fig. 4). Pet-
OEPTH,

1100 ponoun, Bambuluwe, and Tizong also

“See

Wum 7 7 4
showed evidence of a second thermocline,
30 4 May 85
and two of the lakes of highest elevation,
PAvSeeceece,

1030
Oku (2,227 m) and Manengouba (1,920 m),
407 480" exhibited intense vertical microstructure
124
OXYGEN, mg L! with many narrow, homogeneous temper-
ature layers superimposed on the typical
TEMPERATURE, °C
stratification curve (Fig. 4). These layering
30 3t 28 29
Oo
08003 phenomena commonly occur in oceans and
| ¢/ 1 \
temperate lakes (Simpson and Woods 1970;
Ejagham
Dillon et al. 1975), although strong second-
ary thermoclines may be more persistent in
Kotto
tropical lakes (Lewis 1983).
m

16 Feb85 +
0930
Oxygen — Maximal oxygen concentra-
DEPTH,

tions were measured in the metalimnia of

Ngaoundaba, Baledjam, and Tizong, and
in the upper mixed layer of all other lakes.
Debundsha Oxygen saturation of surface water varied
10 Mar 85 4 from 66% in Ejagham to 121% in Disoni,
1130
and oxygen saturation in excess of 100%
occurred throughout the epilimnion in sev-
OXYGEN, mg L"! eral lakes (Figs. 3-6). Metalimnetic oxy-
Fig. 5. As Fig. 3. clines were well defined, and anoxic or near-
ly anoxic conditions were found in the
Multiple thermoclines found in some lakes hypolimnia of all stratified lakes. In the
were the most common deviation from the shallow Petponoun and Baledjam, water be-
classical thermal profiles found in Gegouba, came anoxic below the bottom of the me-
Ngaoundaba, and Mbalang. The formation talimnion at 10 m, while in lakes >50 m
of multiple thermoclines occurs when a rel- deep the anoxic boundary ranged from 15
atively cool epilimnion is subjected to a pe- to 50 m (Figs. 3-6). The anoxia beginning

TEMPERATURE , °C

0 18 19 20 2l o-—— 24 ro 26

22.24 | opt ft
E ~ 20f 47
5 i
FP
<= 3 L B 4
ke . 3S 6 ‘
a b 02 _-- 4 /
Wi Bambuluwe cect T° tf

40; 3 May 85 r i Pet ]

a oF Baledjam | Qt! maces 1
E lay 4
i2t ‘coon | tooo
60 wad kd 12! popper

2 8 0246 8 1357

OXYGEN, mg L!
Fig. 6. As Fig. 3.
 34 Kling

Table 3. Stability (S), work of the wind (B), and total work (G) in J m-? calculated for Cameroon lakes on
the dates shown in Figs. 3-6 (except Nyos). Area is in km? and z,,,, is in m. Values for comparison were calculated
from the raw data of Ruttner (1931), or taken from Hutchinson (1957), Johnson et al. (1978), or Lewis (1983).

Lake Ss B G Area Zmax

Cameroon lakes
Barombi Mbo 5,784 9,294 15,078 4.15 110
Wum 3,068 662 3,730 0.45 124
Manengouba-F 2,115 910 3,025 0.22 168
Disoni 1,925 2,223 4,148 1.65 80
Mbalang 1,627 949 2,576 0.50 52
Tizong 1,291 999 2,290 0.08 48
Bambuluwe 847 355 1,202 0.28 58
Oku 320 434 7154 2.43 52
Monoun 1,550 - - 0.53 96
Nyos 6,870 - - 1.58 208
Other tropical and temperate lakes
Atitlan 21,084 3,669 24,753 137 341
Ranau 6,599 4,728 11,327 126 229
Pakis 4,846 1,489 6,335 0.40 156
Minindau 4,816 5,991 10,807 97.9 169
Lamongan 442 253 695 0.30 28.5
Amatilan 407 737 1,144 8.20 34
Valencia (1978) 345 3,095 3,440 350 39
Giija 172 939 1,111 44.3 26
Pyramid (1914) 10,057 3,977 14,034 532 104
Mendota 504 1,186 1,690 39.2 26
Lunzer Untersee 382 525 907 0.68 34
Lawrence 204 295 499 0.05 13
Mirror 132 276 408 0.15 11
Findley 130 172 302 0.114 24

at 5 m in Lake Monoun was unusual among to the surface was great in many lakes,
the Cameroon lakes. Green et al. (1973) re- chemical stratification contributed insignif-
ported anoxia below 19 m in a study of icantly to overall stability. For example, in
Barombi Mbo (April 1972), which contrasts Barombi Mbo the density changes due to
with the presence of oxygen to about 40 m dissolved ions increased S by <1%. The
in 1985. Although thermally stratified, De- only exceptions to this generality are Mo-
bundsha and Oku had oxygenated hypolim- noun and Nyos, although, as explained ear-
nia (Figs. 4, 5). Oxygen was also present lier, their S values must be considered ap-
throughout the water column in Ejagham, proximate. Chemical stability in Monoun
Beme, Debundsha, and Kotto, but concen- (460 J m’) was roughly half that of thermal
trations were reduced at greater depths in stability (1,090 J m~’). There is evidence
the latter three lakes (Fig. 5). that the 21 August gas release from Nyos
Stability and work of the wind—Stability was associated with substantial mixing in
(S) ranged from 0 in the unstratified lakes the water column (Kling et al. 1987), and
to 5,784 J m~? in Barombi Mbo (Table 3). just after the event the lake was nearly iso-
The value of S for Barombi Mbo probably thermal and estimated stability was due en-
represents its maximum annual value, while tirely to chemical stratification. The Cam-
the stabilities shown in Table 3 for the other eroon lakes can be grouped as minimally
Cameroon lakes may not, because they are stable (Kotto and Ejagham), slightly stable
computed from single profiles. Slight to (Bambuluwe and Oku), moderately stable
moderate stratification of inorganic ions was (Manengouba-F, Disoni, Mbalang, and Ti-
found in all lakes with anoxic hypolimnia zong), strongly stable (Barombi Mbo and
(Table 4). Even though the proportional in- Wum), and metastable due to high CO, con-
crease in hypolimnetic conductivity relative centrations (Monoun and Nyos). The scar-
 Lake mixing in Cameroon 35

Table 4. Conductivity (K,;) for lakes with anoxic Table 5. Regression equations for stability (S), work
or nearly anoxic hypolimnia. of the wind (B), thermocline depth (z,), maximum depth
(Zmax)> and associated morphometric features including
Kos (vS cm™') area (A) and fetch (F). Values in parentheses below
Lake Surface Bottom each parameter are the standard errors of those esti-
mates.
Edib 13.2 39.4
Debundsha 14.5 14.2 N Pr
Bambuluwe 15.0 50.8
Petponoun 29.0 36.8 [log S] = 0.416 + 1.48[log Za] 220.852
Disoni 33.0 71.2 (0.250) (0.138)
Manengouba-F 35.0 43.8 [log B] = —0.17 + 0.45flog z] 9 0.77
Wum 44.8 64.8 (0.26) (0.09)
Barombi Mbo 49.1 80.2 [z] = 9.94 [F}° 300 52 0.83
Benakuma 73.0 306 (0.959) (0.019)
Manengouba-M 96.0 327 [log z,] = 0.996 + 0.168[log 4] 520.715
Baledjam 112 136 (0.026) (0.015)
Monoun 112 2,190
Nchout 113 120
Enep 114 179 Hutchinson (1957) and Wetzel (1983)
Ngaoundaba 142 206 stressed the apparent strong dependence of
Mbalang 167 194 Sand G on lake area. This dependence may
Gegouba 168 185 not apply to tropical lakes. Most Cameroon
Baleng 213 255
Banefo 221 297 lakes, for example, have higher stabilities
L. Monoun 226 406 than Lake Valencia (Table 3), which is al-
Tizong 235 258 most two orders of magnitude larger than
Mfou 243 272 Barombi Mbo. In addition, the relationship
Negop Ghang 273 287 is not found in the range of areas spanned
Mfouet 357 446
by the Cameroon lakes, as Kotto and Oku
are two of the largest lakes and yet are least
stable. Given similar heat incomes, the more
city of similar data makes comparisons dif- sheltered crater lakes would be expected to
ficult, but these stabilities are similar to or have lower rather than higher values of S
higher than those calculated for other trop- compared to lakes similar or larger in size
ical and temperate lakes. that are more exposed to the wind. It is
Birge’s (1916) work of the wind (a mis- obvious from the data in Table 3 that max-
nomer, because heat transport is effected by imum lake depth is related more strongly
turbulence from convective mixing as well to S than is lake area (Fig. 7; Table 5).
as from wind stress) is computed over the Mixing depth—Morphometric controls on
time between minimum and maximum heat the extent of mixing, as measured by epi-
content, which is known only for Barombi limnion depth, thermocline depth, or depth
Mbo. Minimum heat contents were esti- to anoxia, have been noted in temperate
mated assuming homiothermy at the lowest, (Patalas 1984; Gorham and Boyce pers.
hypolimnetic temperature of each lake, and comm.) and tropical (Ruttner 1931; Melack
the heat content at the time the profiles were 1978) lakes. A persistent thermocline is the
taken was used instead of the maximum most common measure of mixing depth,
heat content. Thus the values of B are although a thermocline can be buried in a
underestimates, but they do provide a lower metalimnion of any thickness, and thus
bound. As might be expected, B is signifi- bears an uncertain relationship to the depth
cantly related to thermocline depth (Table of the mixed layer. Modeling of temperate
5). The sum of S and B is the total work lakes has established that summer ther-
(G) required to heat the lake from its min- mocline depth is predictable and that it in-
imal temperature to the observed mean creases as a function of lake fetch or area
temperature without stratification (Hutch- (Patalas 1984; Gorham and Boyce pers.
inson 1957). The grouping of lakes accord- comm.). This relationship is strong for tem-
ing to values of B or G was similar to group- perate lakes up to about 30 km in fetch, but
ing by S (Table 3). in the Laurentian Great Lakes observed
36 Kling

80 Tanganyika ,
4.0 Pyramid,

Barombi 9 akis
Mbo~_:
inindau
60 |
log (Stability)

f @Wum 4
@
Mbalange Manengouba J E + Victoria 4
Tizong®
3.0
N40
Amatilan: @0k: 4
4
z0 be crea “ote se
2.05
2.0 25
0 1 . . _
log (Z max) 0 200 400 600
Fig. 7. Relationship between log of calculated sta- FETCH, km
bility (log S) and log of maximum lake depth (log Zax) Fig. 8. Relationship of thermocline depth (Z,, m)
for temperate and tropical lakes. Data references in and maximum lake fetch for tropical lakes. Data from
Table 3; regression equation in Table 5. references given in Table 3, Hutchinson (1957), and
G. Coulter (pers. comm.).

thermocline depths of 15-20 m are much Discussion

less than predicted (40-60 m: Patalas 1984; Plankton maxima or particle accumula-
Gorham and Boyce pers. comm.). tion at mixing interfaces can abruptly change
For a combination of tropical lakes in light attenuation, and although the epilim-
Cameroon, East Africa, Indonesia, and nion is often considered well mixed and ho-
South America, there is a well-defined re- mogeneous, it may not always be so. For
lationship between thermocline depth and example, peak epilimnetic oxygen concen-
lake fetch (Fig. 8; Table 5). Figure 8 also trations were found at the same depths as
shows that the large East African Rift lakes increases of kK in Barombi Mbo, Ngaoun-
have much deeper thermoclines than the daba, and Tizong and were probably caused
Laurentian Great Lakes. The curvilinear by phytoplankton. Such variation may have
nature of this relationship disappears when diurnal and seasonal structure; measure-
the largest lakes are removed, and a power ment of this structure is uncommon, per-
function fits the remaining data poorly. In haps because of the equipment necessary
addition, the residuals around the regres- (see Talling 1981). Examining the spectral
sion line for all lakes show that the param- effect of narrow water layers requires only
eter estimates used for prediction are inef- irradiance-depth profiles, however, and
ficient. Linearizing the relationship by log- these are widely available. Algal growth
transforming the data corrects the problem. models requiring irradiance data as input
On this basis, lake area is a better predictor (Denman and Marra 1986; Powell and
of z, than is fetch or the minimum height Richerson 1985) thus might be refined by
of the crater rim around the lake (Table 5). studying light intensity microstructure in re-
To facilitate comparison, I have super- lation to epilimnetic mixing.
imposed relationships for temperate lakes Direct solar heating of lower water levels
on a plot of thermocline depth in tropical may be of special importance in transparent
lakes of fetch <10 km (Fig. 9). It is apparent lakes where shelter and small size reduce
from Fig. 9 that most thermocline depths the potential contribution of the wind to
are greater in the Cameroon and other trop- heat distribution (Jassby and Powell 1975).
ical lakes. Gorham and Boyce (pers. comm.) For a given solar input under similar wind
proposed that the high relative mixing conditions, a decrease in optical density of
depths in English lakes are caused by lower the water will increase the depth of diurnal
water temperatures (smaller density gradi- stratification (the “trapping depth”’ of Price
ents) and greater wind force, which is con- et al. 1986) and will thus reduce the net
sistent with observations from northern buoyant resistance of the epilimnion to ver-
Canada (Patalas 1984). tical transport of heat and momentum. On
 Lake mixing in Cameroon 37

36 7 7——+
eDisoni
L4r
| Barombi Mboe 1
27 Singkarak |
= (1.0;
L 3
. 18 ¢ e enone
2
wo
|
x 0.6F
NT. ONTARIO
E
ee eee nEesora 7 = L
o
9 ey .. us JAPAN 2
0.2f 1

OT 68 10
FETCH, km ~0.2° 92 04. 1.0
log (secchi)
Fig. 9. Plot of thermocline depth (Z,, m) and max-
imum fetch for tropical and temperate (Gorham and Fig. 10. Log-log plot of mixing depth vs. Secchi
Boyce pers. comm.) lakes. Points are shown only for depth (Zgecch,) for Cameroon and Uganda crater lakes
tropical lakes; actual data points for the temperate lakes (O and solid line; data from this study and Melack
fall very close to the regression lines presented (Patalas 1978) and Japanese lakes (Mand dashed line; data from
1984). Yoshimura, cited by Hutchinson 1957). Mixing depths
are represented by the top of the metalimnion (Zea)
in Japanese lakes and by the thermocline (z,) in tropical
the basis of a stratification criterion incor- lakes; this may account for the different intercepts. The
porating the effects of solar radiation pen- best-fit simple regression equations are: tropical lakes
etrating the water column, Holloway (1980) (z,) = 2.11 + 1.95 (Zeeccm), N = 31, r? = 0.70, SE inter-
showed that the critical wind speed for mix- cept = 1.18, SE slope coefficient = 0.240; temperate
lakes (Zeta) = 0.77 + 0.53 (Zsecch), N = 45, r2 = 0.61,
ing decreases in clearer waters. If this mech- SE intercept = 0.53, SE slope coefficient.= 0.06.
anism operates in lakes and is important
enough to be reflected in seasonal thermo-
cline depth, then a necessary consequence deeply than brown-water lakes—indepen-
is a pronounced relationship between mix- dent of lake fetch. If conditions are such that
ing depth and water transparency. Data from boundary-layer shear dominates turbulent
sheltered lakes in Cameroon, Uganda, and mixing in the metalimnion (e.g. from an
Japan show that a significant proportion of internal seiche, see Spigel et al. 1986), trans-
the variance in mixing depth is explained parency will be less important because the
by light penetration (Fig. 10). For the trop- solar penetration mechanism only affects
ical lakes, partial r? values from stepwise stirring within the epilimnion.
regression indicate that Secchi depth ex- The effects of lake morphometry on water
plains 71% of the variation in thermocline column stability, and thus on the potential
depth, and lake area only explains an ad- for mixing, are not entirely understood. Lake
ditional 2%. Collinearity between the pre- stability is time-integrated, but its measure
dictor variables, Secchi depth and area, is at a single time offers no information on
negligible. This analysis suggests that re- time-dependent processes. Its magnitude
duction in buoyant resistance to vertical increases with the depth of the lake, the den-
mixing, caused by deeper penetration of so- sity gradient across the thermocline, and the
lar energy into the water column, is partly depth of the thermocline to a point defined
responsible for establishing deeper mixing by basin morphometry. In a deep lake, the
depths in clearer lakes. This mechanism may increased mass of dense water below the
be important in other lakes more exposed pycnocline contributes to the observed re-
to the wind or of larger size, because the lationship between S and lake depth. The
effects of greater solar penetration are in- complex interaction of thermocline lower-
dependent of whether turbulence from wind ing and density gradient production is not
mixing or from convection is actually deep- well studied, and stability as a single mea-
ening the mixed layer. For example, Salo- sure may be viewed as a poor indicator of
nen et al. (1984) reported that in southern physical process and heat distribution. It
Finland clear-water lakes circulate more remains a useful three-dimensional index of
38 Kling

comparison, however, and serves to eval- density (do/dz) over 1 m in the strongly
uate both the potential for hypolimnetic gas stratified Cameroon lakes averaged 0.00023
accumulation and the energy required for g cm™? (N = 24; SE = 0.00003), which is
the upward mixing and release of gas. much less than that found in several ELA
Reduced stability and water column mix- lakes (~0.00090 g cm™?: Schindler 1971).
ing are dependent on appropriate condi- A final factor to consider is the smaller Co-
tions for heat loss. In the tropics, shallow riolis effect at lower latitudes. The magni-
lakes may mix thoroughly each day tude of this effect would be unnoticeable in
(MacIntyre and Melack in press), but the all but the largest lakes (Phillips 1977; cf.
deeper lakes typically experience a period Garwood et al. 1985). In the Laurentian
of minimum stability and deep if not com- Great Lakes, Coriolis effects produce near-
plete mixing during the hemispheric winter shore downwelling and reduce turbulent
(Talling 1969; Lewis 1983). This pattern mixing from internal seiches, which togeth-
may not hold in parts of West Africa, where er limit the comparatively shallow mixing
the period of mixing is coincident with min- depths (Mortimer 1974; Gorham and Boyce
imum air temperatures and insolation dur- pers. comm.). The absence of a strong ro-
ing summer (Kling 1987b). There is no rea- tational component at equatorial latitudes
son to believe that the stratified Cameroon may in turn help explain why the great lakes
lakes do not experience a period of deep of East Africa have such deep mixed layers.
mixing, but the moderate and large stability Detailed measurements on the dynamics of
values found in some lakes suggest that density structure in tropical lakes are need-
yearly mixing may not affect the entire water ed, but it appears that regardless of the dom-
column. inant process of mixed-layer deepening the
Several physical factors may be involved smaller absolute density gradients at trop-
in formation of deeper mixing depths in ical latitudes play a substantial role in for-
tropical lakes. They include the decreased mation of the deeper thermoclines found in
water viscosity at higher temperatures, the many tropical lakes.
intensity of storms or maximum wind
speeds, the smaller density gradients en-
countered in warmer waters, and the re- References
duced Coriolis effects at low latitudes. Kato BEADLE, L. C. 1981. The inland waters of tropical
and Phillips (1969) showed that viscosity Africa, 2nd ed. Longman.
BirceE, E. A. 1916. The work of the wind in warming
can eventually become important to en- a lake. Trans. Wis. Acad. Sci. 18: 341-391.
trainment rates at low Reynolds numbers, BowLiIna, L. C., AND P. A. TYLER. 1986. The un-
although viscosity effects can generally be derwater light-field of lakes with marked physi-
ignored under most conditions (Turner cochemical and biotic diversity in the water col-
umn. J. Plankton Res. 8: 69-77.
1973; Moum and Caldwell 1985). There
CHEN, C-T., AND F. J. MILLERO. 1977. The use and
may be consistent differences in wind power misuse of pure water PVT properties for lake
between specific latitudinal belts, but gen- waters. Nature 266: 707-708.
eralities of storm intensity are difficult to CorseT, S. A., J. GREEN, J. GRIFFITH, AND E. BETNEY.
make, and any effects would be smoothed 1973. Ecological studies on crater lakes in West
Cameroon. Lakes Kotto and Mboandong. J. Zool.
by the cross-sectional nature of the tropical Lond. 170: 309-324.
lake data set. The differences in density gra- DENMAN, K. L., AND J. MARRA. 1986. Modelling of
dients found in tropical and temperate lakes, the time dependent photoadaptation of phyto-
however, are striking. For example, the plankton to fluctuating light, p. 341-359. In J.C.
mean of the maximum density difference J. Nihoul [ed.], Marine interfaces and ecohydrody-
namics. Elsevier.
within the water columns of the tropical DILton, T. M., AND T. M. PoweLi. 1979. Observa-
lakes examined here was 0.00065 g cm™> tions of a surface mixed layer. Deep-Sea Res. 261:
(N = 52; SE= 0.00005), whereas the density 915-932.
difference between 4° and 20°C in a tem- > , AND L. O. Myrup. 1975. Low fre-
quency turbulence and vertical temperature mi-
perate lake during summer stratification is crostructure in Lake Tahoe, California-Nevada.
about three times greater (0.00176 g cm~?). Int. Ver. Theor. Angew. Limnol. Verh. 19: 110-
In addition, the maximum change in water 115.
 Lake mixing in Cameroon 39

Foster, T.D. 1971. A convective model for the diur- of vertical mixing in an Amazon floodplain lake
nal cycle in the upper ocean. J. Geophys. Res. 76: (L. Calado, Brazil). Int. Ver. Theor. Angew. Lim-
666-675. nol. Verh. 23.
Garwoop, R. W., Jr., P. MULLER, AND P. C. GAL- MACIOLEK, J. A. 1975. Limnological ecosystems and
LAGHER. 1985. Wind direction and equilibrium Hawaii’s preservational planning. Int. Ver. Theor.
mixed layer depth in the tropical Pacific ocean. J. Angew. Limnol. Verh. 19: 1461-1467.
Phys. Oceanogr. 15: 1332-1338. MELAcK, J. M. 1978. Morphometric, physical and
GreEEN, J., S. A. CORBET, AND E. BETNEY. 1973. Eco- chemical features of the volcanic crater lakes of
logical studies on crater lakes in West Cameroon: western Uganda. Arch. Hydrobiol. 84: 430-453.
The blood of endemic cichlids in Barombi Mbo Mortimer, C. H. 1974. Lake hydrodynamics. Mitt.
in relation to stratification and their feeding habits. Int. Ver. Theor. Angew. Limnol. 20, p. 124-197.
J. Zool. Lond. 170: 299-308. Moum, J. N., AND D. R. CALDWELL. 1985. Local
; 5, AND . 1974. Ecological studies influences on shear-flow turbulence in the equa-
on crater lakes in West Cameroon. Debundsha torial ocean. Science 230: 315-316.
Lake. J. Zool. Lond. 173: 199-223. PATALAS, K. 1984. Mid-summer mixing depths of
HAsserT, K. 1912. Seenstudien in Nord-Kamerun. lakes of different latitudes. Int. Ver. Theor. Angew.
Z. Ges. Erdk. Berl. 1912: 7-41, 135-144, 203-216. Limnol. Verh. 22: 97-102.
HENDERSON-SELLERS, B. 1984. Development and ap- Puitiirs, O. M. 1977. The dynamics of the upper
plication of “U.S.E.D.”: A hydroclimate lake ocean. Cambridge.
stratification model. Ecol. Model. 21: 233-246. POLLARD, R. T., R. B. RHINES, AND R. O. R. Y.
Ho.ioway, P.E. 1980. A criterion for thermal strat- THompson. 1973. The deepening of the wind
ification in a wind-mixed system. J. Phys. Ocean- mixed layer. Geophys. Fluid Dyn. 4: 381-404.
ogr. 10: 861-869. PowELL, T., AND P. J. RICHERSON. 1985. Temporal
Hurcuinson, G. E. 1957. A treatise on limnology. variation, spatial heterogeneity and competition
V. 1. Wiley. for resources in plankton systems: A theoretical
Ipso, S. B. 1973. On the concept of lake stability. model. Am. Nat. 125: 431-464.
Limnol. Oceanogr. 18: 681-683. Price, J. F., R. A. WELLER, AND R. PINKEL. 1986.
IMBERGER, J. 1985. The diurnal mixed layer. Limnol. Diurnal cycling: Observations and models of the
Oceanogr. 30: 737-770. upper ocean response to diurnal heating, cooling,
JAssBy, A., AND T. POWELL. 1975. Vertical patterns and wind mixing. J. Geophys. Res. 91: 8411-8427.
of eddy diffusion during stratification in Castle RuttTNer, F. 1931. Hydrographische und hydro-
Lake, California. Limnol. Oceanogr. 20: 530-543. chemische Beobachtungen auf Java, Sumatra und
JOHNSON, N. M., J.S. EATON, AND J. E. RIcHEy. 1978. Bali. Arch. Hydrobiol. Suppl. 8, p. 197-454.
Analysis of five North American lake ecosystems. SALONEN, K., L. ARVOLA, AND M. Rask. 1984. Au-
2. Thermal energy and mechanical stability. Int. tumnal and vernal circulation of small forest lakes
Ver. Theor. Angew. Limnol. Verh. 20: 562-567. in southern Finland. Int. Ver. Theor. Angew. Lim-
Kato, H., AND O. M. Puituips. 1969. On the pene- nol. Verh. 22: 103-107.
tration of a turbulent layer into a stratified fluid. SCHINDLER, D.W. 1971. Light, temperature, and oxy-
J. Fluid Mech. 79: 753-768. gen regimes of selected lakes in the Experimental
Kuna, G. W. 1987a. The comparative limnology of Lakes Area, northwestern Ontario. J. Fish. Res.
lakes in Cameroon, West Africa. Ph.D. thesis, Duke Bd. Can. 28:157-169.
Univ. Simpson, J. H., ANDJ.D. Woops. 1970. Temperature
. 19876. Seasonal mixing and catastrophic de- microstructure in a fresh water thermocline. Na-
gassing of tropical lakes, Cameroon, West Africa. ture 226: 832-835.
Science 237: 1022-1024. SPIGEL, R. H., J. IMBERGER, AND K. N. RAYNER. 1986.
, AND OTHERS. 1987. The August 1986 gas di- Modeling the diurnal mixed layer. Limnol. Ocean-
saster at Lake Nyos, Cameroon, West Africa. Sci- ogr. 31: 533-556.
ence 236: 169-175. TALLING, J. F. 1969. The incidence of vertical mixing,
Kraus, E. B., AND J. S. TURNER. 1967. A one-di- and some biological and chemical consequences,
mensional model of the seasonal thermocline, part in tropical African lakes. Int. Ver. Theor. Angew.
2. Tellus 19: 98-105. Limnol. Verh. 17: 998-1012.
Lewis, W. M., JR. 1973. The thermal regime of Lake . 1971. The underwater light climate as a con-
Lanao (Phillipines) and its theoretical implications trolling factor in the production ecology of fresh-
for tropical lakes. Limnol. Oceanogr. 18: 200-217. water phytoplankton. Mitt. Int. Ver. Angew. Lim-
. 1983. Temperature, heat, and mixing in Lake nol. 19, p. 214-243.
Valencia, Venezuela. Limnol. Oceanogr. 28: 273- . 1981. The development of attenuance depth-
286. profiling to follow the changing distribution of
. 1984. A five-year record of temperature, mix- phytoplankton and other particulate material in a
ing, and stability for a tropical lake (Lake Valencia, productive English lake. Arch. Hydrobiol. 93: 1-
Venezuela). Arch. Hydrobiol. 99: 340-346. 20.
MaAcINTYyRE, S., AND J. M. MELACK. 1982. Meromixis TURNER, J.S. 1973. Buoyancy effects in fluids. Cam-
in an equatorial African soda lake. Limnol. Ocean- bridge.
ogr. 27: 595-509. VINCENT, W. F., P. J. NEALE, AND P. J. RICHERSON.
, AND . In press. Frequency and depth 1984. Photoinhibition: Algal responses to bright
View publication stats