Thanks to the regional climate models set up by the team of modellers from KULeuven, it was possible to establish that these variations could be exacerbated under the influence of changes in atmospheric conditions. Longer rainy seasons reinforce the stratification of the water column, reducing the periods of phytoplankton growth and leading to a natural decrease in productivity at all levels, including fish. “Establishing the extent of atmospheric humidity was one of the major conclusions that these models provided”, the researchers explain. “We overestimated the importance of the wind as a mechanical source causing the waters to mix. It is mainly evaporation, which is expressed by a loss of energy and therefore a loss of heat, that cools the surface waters and encourages them to mix.” And this evaporation is more efficient during the dry season. “In the short term, there’s not a lot we can do to curb these climate changes”, says Jean-Pierre Descy. “On the other hand, this information will allow us to react to reduce the risk of the lake’s decline. For instance, we can think about how to better manage fishing, in a way that takes into account the seasonal evolution of fish stocks. But in order to propose regulated fishing, we need better statistical knowledge. We need to know the real evolution of the fishing yield over the course of the year. This inventory is quite good in Rwanda, but is still embryonic in DRC.”

Generalisation of observations through remote sensing

The traditional approaches for gathering information on the waters’ characteristics such as temperature or the concentration of chlorophyll indicating the quantity of phytoplankton, which can provide clues to the lake’s biological productivity, are relatively classic and costly. During campaigns, a boat periodically takes water samples. This provides specific information at a given moment and often at different depths, an operation that can be repeated in space and time to start statistical and comparative studies. Remote sensing and the analysis of the satellite images can also be used to obtain data that is more representative of the whole lake. Within the framework of the EAGLES project, this type of data was provided by a team of two researchers under the responsibility of Yves Cornet, lecturer in ULg’s Geomatics Unit, following a similar fruitful operation at Lake Tanganyika. “Through satellite imaging, remote sensing allows us to acquire daily data on the surface for a whole system, and it is ideal for highlighting its variability, in support of the traditional sampling methods”, Yves Cornet adds. “As regards Lake Kivu, the information collected was supposed to give us a more precise idea of temperature variations in the surface waters and primary production fluctuations, by observing the concentration of chlorophyll-a.” While the technology was particularly effective at Lake Tanganyika, Lake Kivu unfortunately turned out to be more hostile. “We suffered a series disturbances particular to the region’s physiography and climatology which rendered the majority of our data practically unusable.” Despite its 2,700 km2, the lake is in fact too small for efficient data collection. “No matter where you are on the lake, the shoreline is never far and it reflects a light that interferes with the signal recorded by the sensors. In addition, there was too much cloud cover. In the case of Lake Tanganyika, which is much bigger, despite similar climate conditions, we had more chance of observing cloudless areas. As for Lake Kivu, sometimes we only obtained a single piece of good quality data per week at a few points on the lake, which we then had to compare with the information collected on the ground to calibrate our calculation methods.” A task that was made all the more arduous owing to the lake’s relative ecological poverty. “Kivu is an oligotrophic lake. Even during the dry season, primary production remains low, as does the concentration of chlorophyll-a. Therefore, the signal was often very poor, with significant disturbances, making it very challenging to obtain quality products.” Nevertheless, this satellite data provided relatively precise information on the surface temperature, providing a better understanding of the spatial and temporal behaviour of the lake’s water over more than a decade, making it easier to predict future changes.

Reading the history of the lake through its sediments  

By taking sediment cores and analysing the different components in the sediments, including fossils such as diatoms, it is possible to piece together the lake’s history (paleolimnology). In the East African Great Lakes, which include Lake Kivu, this type of research enables the observation of the ecosystem’s responses to changes that occurred in the past, and thus draw similarities with future scenarios. “We were able to record a whole series of major variations in terms of the richness of nutrients, or the stratification of the waters”, Jean-Pierre Descy explains. “But Lake Kivu has a geochemical particularity which prevented us from dating the samples sufficiently accurately to be able to relate these variations to climate events.” On the other hand, these cores allowed the researchers to test hypotheses on the recent history of the lake, based on questionable samples. “For instance”, Alberto Borges points out, “the introduction of endemic sardines from Lake Tanganyika to Lake Kivu was considered an ecological disaster by many. In particular, these sardines eradicated a species of daphnia observed in a sample taken in the 1950s. We found no trace of this daphnia’s past existence in our sediment cores, and it is likely that it never lived in Lake Kivu. On the other hand, on the basis of old data, we were able to estimate that zooplankton was three times more abundant before the introduction of the sardines.” “Which is normal” Jean-Pierre Descy adds. “The lake was a true paradise for the zooplankton, which didn’t have any predators at the time. Rather than initiating an ecological disaster, the sardines actually re-established a ‘natural’ balance, or at least that observed in other lakes. One piece of evidence that the balance has been restored is that for 30 years, the stock of sardines, estimated by hydroacoustic methods (depth sounder), seems to be constant.