The objectives of this study emphasised a need for a general description of the limnological character of the lake and the seasonal variation in this character. The selection of characteristics to be examined was based on more or less conventional understandings of factors directly related to fisheries. Owing to limitations in time and facilities, a compromise was, however, made between the need for descriptive data used in interpreting seasonal changes in the lake structure and the need for similar data in inshore regions where most of the fishing operations occur. As data relating to the latter areas were obtained by project biologists at times and locations coinciding with the programme of experimental fishing, the limnological sampling placed emphasis on the former objective.
After an initial survey of the lake in late August 1969 and discussions with the project's fishery biologists, a series of sampling stations covering the middle and lower sections of the lake were chosen. The upper arm has remained more or less riverine in character and was being regularly visited by the fishery team, hence stations were confined to the main and lower basins.
The location of the stations chosen is shown in Fig. 1. All but station 14 were located over the old river bed where maximum local depths could be obtained. In practice, these stations were located during cruises by reference to visual landmarks and echosounder, and the position was only approximately reproduced on subsequent cruises.
Cruises were planned at intervals of six weeks to occur midway between fishery surveys. Actual times were somewhat irregular and adjusted to accommodate other project needs, and the intervals between cruises actually varied from four to eight weeks. Each cruise was assigned a reference number; a list of the cruises with dates and stations occupied is given in Table 1.
The cruises usually required three days for completion, with all work accomplished during daylight hours. Stations were occupied for from one half to two hours as the full series of observations were not made at every station.
From August 1969 to July 1970 all work had to be accomplished with a 16 ft outboard boat. In July 1970 a 31 ft houseboat became available, permitting more flexible operations and allowing water samples to be analysed while the boat was underway.
Equipment and methods used in measurement and analysis were primarily dictated by the store of equipment already on hand at the project. While the results were reasonably satisfactory, some of the equipment was cumbersome for small boat use.
Temperature profiles were obtained at each occupied station with a bathythermograph operated from a hand winch. Considerable difficulty was experienced in maintaining calibration owing to high ambient temperatures during storage, and to uncertainties in field calibration against standard mercury thermometers. Thermal gradients near the surface were generally very steep during the day, while rapid changes during ascent and reading made such comparisons somewhat unreliable. In practice, the temperatures of the 1-metre samples were incorporated into a “quality-control” chart on which the difference between their temperature (thermometer incorporated into the sample chamber) and the temperature indicated by the bathythermograph was plotted successively for all observations for which sample temperatures were available. While there was considerable scatter in these points (Fig. 2) and the trend of drifting calibration appeared almost exponential, a fitted curve through the tank observations was used to correct BT readings between successive cruises and maintain seasonal comparability. Unfortunately the BT was designed for temperatures to 30°C, while surface temperatures at Kainji frequently reached 32°C while surface temperatures at Kainji frequently reached 32°C and occasionally as high as 37°C. Hence temperature gradients in the first 0.5 metres could not be regularly observed.
The steep surface gradients observed led to the adoption of locally constructed sampling bottles (Meyer's type) for use by the fishery programme in obtaining water temperature at fishing stations at a standard depth of 1 metre, thus avoiding the uncertain significance of dip-sample observations at surface temperature.
A vertical series of water samples was taken at most of the stations. A “surface” sample was taken at a depth of one half metre. A “bottom” sample was taken at a distance of 1–2 metres above the bottom (judged from the depth recorder). One or more intermediate levels were chosen for sampling after examination of the BT profile. When a thermocline was present the intermediate sample was taken just above this thermocline, otherwise it was taken at mid-depth. On some occasions a more complicated profile was observed, in which case one or more additional samples were obtained.
These water samples were collected with three litre plastic bottles of the Van Dorn type (Gemware) with rubber-cup closing mechanism and internally mounted thermometers. As three were available, they were usually operated in series with the same winch and wire used for the BT.
Concentration of dissolved oxygen
The dissolved oxygen was determined for each sample using standard Winkler reagents (Azide modification). Thiosulphate solutions were standardised at the laboratory before and after each cruise. Prior to the arrival of the large houseboat, oxygen determinations were made with the Hach “powder pillow” reagents and FAO reagent until supplies ran out. While convenient for fishery work, this latter procedure is expensive and, owing to the small quantities titrated, of reduced precision.
Conductivity
Conductivity was determined for the samples with a Hach conductivity bridge designed for field use. Owing to initial errors in the procedure adopted in using the internal standardization procedure, and the relatively low precision of the instrument, the data obtained are suspected to include errors in the range of 5–10% (average conductivity about 55 micromhos/centimetre. While seasonal trends could be observed, these data were of little use in distinguishing between surface and bottom waters, and showed little correlation with independent determinations of total anionic concentration.
Hydrogen-ion concentration
Hydrogen-ion concentration (pH) was generally determined by colorimetry using narrow range indicators. While a battery-operated (Tokai instruments) pH metre was frequently used, corrections for temperature proved difficult to apply in the field and the colorimetric method was judged more reliable.
Total ionic concentration and alkalinity
Two one half litre sub-samples were saved from each sample and carried back to the laboratory. One set was used for the determination of alkalinity using a methyl orange endpoint, and for the determination of total anions by the method ion-exchange (method of Mackereth, 1963). As in the determination of conductivity, sensitivity was poor, and little confidence could be placed in comparisons among samples.
The other set was saved for isotope analysis and for reserve.
Turbidity
After watching the changes occurring in the lake over the first year, it became evident that the striking variation in colloidal turbidity of inorganic origin was likely to be the most sensitive variable in distinguishing between water masses in the lake, at least with the equipment at hand. Baring periods of higher turbidity, the percent transmission of blue light was measured on each water sample using the Hach portable colorimeter (No. 4445 filter). Unfortunately the length of the transmission path (approx. 2 cm) was insufficient for use during periods when the Secchi disc readings increased beyond 1 metre. An in situ transmissometer had been ordered for this work, but after a long delay, procurement was abandoned.
With the recruitment of the counterpart limnologist in December 1969, a programme of plankton sampling was initiated and was developed to a semi-quantitative procedure by the following June.
Phytoplankton samples were obtained with separate casts of the water bottles used to obtain water samples, and one half litre sub-samples were returned to the laboratory for counting. Portions of these were then sedimented with Lugol's solution. Details of the methods used will be reported in conjunction with a more complete report of findings by the counterpart limnologist at a later stage in the project.
Zooplankton samples were obtained with cone nets (10 mesh) approximately 20 cm in diameter and 30 cm long, with plastic vials attached by an elastic sleeve at the apex of the net. A frame was evolved which allowed two pairs of these nets to be simultaneously drawn through the water with a one metre distance between pairs and members of a pair 25 cm between centres (i.e., nearly adjacent). These nets, attached to the weighted cable used in other sampling, were lowered to a depth of 5 metres and returned immediately to the surface. While the scheme of multiple nets provided good control of heterogeneity in the distribution of zooplankton, the uncertainties of vertical tows with such short nets leave the results somewhat questionable. Again, a more detailed discussion of the sampling and counting procedures will be left to a later report on the findings of the plankton investigations of the project.
In addition to the specific observations just discussed, a number of other parameters were observed or measured at each station. These included air temperature, relative humidity, wind speed and direction, cloud cover, wave height and Secchi dise transparency. The meteorological data proved to have little value in interpreting limnological events, particularly as such data could be obtained from the meteorological stations at Kainji and Yelwa. Even though the latter data were not representative of conditions over the lake itself, they had the advantage of day-by-day observations, and were less influenced by diurnal fluctuation.
Unfortunately wave heights were only “guesstimated” until November 1970, when a damped wave-pole became available (constructed at the project).
Stationary echo-sounding records for a minimum of three minutes were obtained at each station. In most oases at least ten minutes of record was obtained. These gave indications of presence or absence of fish and scattering layers. Before the arrival of the larger boat, a Mark II Furuno FG-11, a 25 kilo signal, was used with a hand-held transducer. A newer version of the Furuno Mark II, with a 200 kiloHerz signal and expanded depth scale, was permanently mounted in the larger boat soon after it became available.
In order to examine the implications and significance of the routine observations of Secchi disc transparency, light penetration was examined on two occasions with a GM under-water photometer equipped with red, green and blue filters. As scattering by colloidal clay seemed to be the dominant factor in changes of light penetration, several attempts were made to ascertain the physical properties of the colloidal suspensions found in the lake. A much more thorough study is now underway in connexion with a project study of seasonal variation in primary production.
2.2.1 Light Transmission
Light intensity as a function of depth was measured in October 1969 and again in February 1970, when Secchi disc readings of 0.5 m and 2.2 m were obtained respectively. A more thorough study was not undertaken owing to uncertainties in the calibration of the various scales of the meter (the meter scales had to be readjusted on arrival owing, apparently, to moisture problems in the shunts). A rough adjustment was effected by adjusting serially the next more sensitive scale to the appropriate multiple (10x on most ranges) of the reading obtained on the previous scale. This procedure accumulates errors arising from the low precision of the lower part of the scale and non-linearities in the measuring circuit. In spite of these difficulties, a substantial shift in the spectrum of the transmitted light was observed and tentative attenuation coefficients for each filter were calculated. From the sensitivity and transmission curves published for the photocell and the filters respectively, the transmission coefficients were converted to relative energy units from which spectral curves could be drawn. (Vollenweider, ed., 1969).
2.2.2 Properties of the Colloidal System
Data obtained by the previous field observations were supported by similar observations on dilutions of water samples with the Hach colorimeter. The stability of these suspensions was examined by several experiments which will be described briefly in the sections where the results are discussed.
