Climate variability and change that have been intensifying since the 1970s are accompanied by changes in hydrology and water balance of inland aquatic systems. These changes, however, have not been well documented with regard to small and shallow aquatic systems that are more vulnerable. Changes in temperature, rainfall and wind speed around Lakes Wamala and Kawi (Uganda) were examined to provide insight on how the increasing variability and change in climate affect water balance and lake water levels. Around Lake Wamala, average air temperature has increased by 0.018°C y−1 since 1980. Rainfall increased by 9.01 mm y−1 since 1970 and accounted for 79.5% of the water gain during 2011 and 2013 period. However, the gains were exceeded by losses as a result of evaporation that accounted for >85% of the water loss. Despite the increase in rainfall, the mean lake depth of Lake Wamala decreased by 0.015 m y−1, apparently due to high evaporation rates. Around Lake Kawi, average air temperature has increased by 0.036°C y−1 since 1980. Rainfall, however, has decreased, although it still dominates the water inflows, accounting for 83% of the total water gains. Interviews with local fishermen on Lake Kawi indicated that the lake shoreline has receded by ∼50 m over the last two decades. These results suggested that most of the water in small shallow lakes is gained through direct rainfall, but more is lost through evaporation. Therefore, increase in rainfall around these lakes is no longer sufficient to sustain normal lake water levels and lake surface area as long as temperature and wind speed, which enhance evaporation, continue increasing. This has implications for lake productivity processes and needs to be understood and incorporated in management of fisheries and the catchment areas as climate warming intensifies.
Climate variability and change are some of the major environmental challenges of the 21st Century (IPCC, 2013) that are expected to greatly affect natural resources, including fisheries. Temperatures over Africa have increased by 0.5°C since the 1970s (IPCC, 2007) and are projected to increase faster than the global average during the 21st Century (James and Washington, 2013). This is expected to alter rainfall patterns and wind speed, with corresponding changes in lake water levels and surface area. This can have major impacts, especially on small and shallow lakes and marginal areas of deep lakes and, ultimately, aquatic productivity processes, fisheries, and livelihoods. Warming has already been detected in most large and relatively deep African lakes, including Tanganyika (O’Reilly et al., 2003; Verburg et al., 2003), Victoria (Hecky et al., 2010; Sitoki et al., 2010; Marshall et al., 2013), Albert (Lehman et al., 1998), Malawi (Vollmer et al., 2005), Kivu (Lorke et al., 2004) and Kariba (Mahere, 2012) and few shallow lakes, such as Chad (Famine Early Warning Systems Network, 2012), Chilwa (Foli and Makungwa, 2011). Such long-term and repeated investigations that can be used to quantify impact of climate change on fisheries productivity are still lacking for most small and shallow aquatic ecosystems, which are more vulnerable to impacts of climate variability and change. Lack of such information compromises interpretation of the influence of climate variability and change on fisheries production and management for their resilience.
Here, we examine changes in temperature, rainfall and wind speed around Lakes Wamala and Kawi to provide insight on how the increasing variability and change in climate are likely to influence water balance and lake water levels of small and shallow lakes. This will be useful for future assessment of the impacts of the increasing variability and change in climate on aquatic productivity processes, fish production and, ultimately, livelihoods. We choose Lake Wamala in Central Uganda and Lake Kawi in North Eastern Uganda because these lakes have a history of undergoing periodic changes in lake water levels and surface area. Since there is no abstraction of water from these two shallow lakes for large scale purposes such as irrigation and hydropower generation, we hypothesize that these changes are due to changes in climate variables. IPCC (2007) has predicted that climate warming will be accompanied by increased rainfall over the East African region. Since most of the water in lakes within the East African region is gained through direct rainfall (Bootsma and Hecky, 2003), lake water levels would be expected to remain high even with the warming climate.
Material and methods
The study was carried out on Lake Wamala in the Lake Victoria basin and Lake Kawi in the Lake Kyoga basin (Figure 1). The highest recorded surface area of Lake Wamala is 250 km2, and the maximum mean depth is 4.5 m (Goulden, 2006). Lake Wamala is surrounded by a wetland dominated by Papyrus (Cyperus papyrus L.), Hippo Grass (Vossia cuspidata Roxb) and Ambatch (Aeschynomene elaphroxylon Guill. and Perr.), with the rest of the catchment mainly comprised of agricultural land. The lake is fed by several small rivers, including Nyanzi, Kabasuma, Mpamujugu and Bimbya and drained by River Kibimba into Lake Victoria via the Katonga wetland. The lake surface area has historically fluctuated between 100 and 250 km2 and its mean depth between 1.5 and 4.5 m (UNEP, 2009). Lake Kawi is about 10 km2 with a mean depth of 3.5 m. The lake is also surrounded by a wetland largely composed of Papyrus. Lake Kawi is drained by River Mpologoma into Lake Kyoga. Both Lakes Wamala and Kawi offer livelihoods to the surrounding communities and, therefore, information for their conservation as climate change intensifies is vital.
Monthly minimum and maximum air temperature and rainfall totals from Mubende weather station, which is nearest to Lake Wamala, and Kamuli weather station, which is nearest to Lake Kawi, were obtained from the Uganda National Meteorological Authority (UNMA). Data on wind speed (0900 h average) from Entebbe and Jinja weather stations that are close to Lakes Wamala and Kawi, respectively, were also obtained from UNMA. Data on monthly water inflows and outflows for Lake Wamala, 2011–2013 were acquired from the Directorate of Water Resources Management (DWRM) in the Ministry of Water and Environment (MWE) while those on mean lake depth were obtained from the National Fisheries Resources Research Institute (NaFIRRI). Interviews with local fishermen around Lake Kawi provided information about the changes in the lake water levels.
The data were used to ascertain: long-term trends in the climate variables; deviations of mean annual temperature and total annual rainfall from the 1981–2010 mean (the most recent 30 year period for calculating climate normal; Ted, 2011); decadal variations in mean temperatures; standardized precipitation indices (SPI); and water balance. The coefficient of variation (CV), defined as relative standard deviation, was calculated to give indication of inter-annual rainfall variability. SPI were calculated as standardized difference between long-term mean (1981–2010) and annual rainfall totals. These values were used to characterize wet and dry years, using a classification described by Tumbo (2007). Trend analysis was done using non-parametric Man-Kendall Tau b test (Mann, 1945; Kendall, 1975). Water balance was computed as the sum of inflows (precipitation + river inflow) minus the sum of outflows (evaporation + river discharge). Water balance was calculated to test whether there was any indication of a change in the hydrological pattern attributable to high temperatures and increasing wind speed. Lake Wamala is located in the Lake Victoria basin and, therefore, the interaction between groundwater and surface water, and its influence on other water fluxes, was assumed insignificant (Krishnamurthy and Ibrahim, 1973).
There was an increase in minimum, maximum and average air temperatures around Lake Wamala since 1980, although only the minimum temperature (∼0.02ºC year−1, p < 0.05) and average temperature (∼0.018ºC year−1, p < 0.05) exhibited significant trends (Figure 2a). Similar trends were observed around Lake Kawi but with minimum, maximum and average air temperatures exhibiting significant trends (p < 0.05; Figure 2b). Decadal analysis showed that median temperature around Wamala increased by 0.94ºC between 1980 and 2012, with ∼70% of the increase occurring between 2000 and 2012 (Figure 2c). Around Lake Kawi, median temperature increased by 1.6ºC between 1970 and 2012 but the rate of increase was higher during the period 1970–1990 than the subsequent decades (Figure 2d). Wind speed around both lakes increased, but only Lake Kawi exhibited a significant trend (Figure 3).
Mean annual rainfall around Wamala was 1180 ± 203 mm and the coefficient of variation was 25.6%. There was a significant increase in rainfall around Lake Wamala since 1970 (Tau = 0.402, p < 0.0001, Figure 4a) but no distinct trend was observed around Lake Kawi (Figure 4b). Results of SPI showed that drought periods around Lake Wamala were clustered between years 1970–1975, 1979–1987, 1992–1994, 2004–2005 and 2008 (Figure 4c) while those in the Lake Kawi region did not show any clear pattern (Figure 4d).
The mean depth of Lake Wamala decreased by 0.015 m year−1 since 1970 despite rainfall increasing by 9.01 mm year−1 over the same period (Figure 5). However, over time, fluctuations in rainfall were accompanied by changes in mean lake depth. The mean depth of Lake Wamala, for instance, decreased from 4.5 m to 1.5 m between 1977 and 1995, coinciding with the dry years of 1979–1987 and 1992–1994. After the El-Niño of 1997, the mean lake depth increased again to 4.5 m (Figure 5). By 2011, the mean lake depth had dropped to 3.0 m, despite higher than long term average rainfall received around the lake for most of the years since 2006. Although there were no data on lake levels for Lake Kawi, interviews with local fishermen indicated that the lake shoreline had receded by about 50 m since the last two decades.
Data on the inflows and outflows showed that rainfall and evaporation were the most important factors in the water balance of both lakes (Table 1). In Lake Wamala, rainfall contributed 79.6% of the total water input into the lake while evaporation contributed 86.2% of the total outflows, resulting into a negative water balance. Similarly, in Lake Kawi, rainfall contributed about 83%, but this was also exceeded by evaporation, which contributed 99.6% of the water losses from the lake.
Discussion and conclusions
This study was intended to investigate how increasing variability and change in climate might affect hydrology and water balance of small and shallow lakes, and to provide an insight of how this might in turn affect aquatic productivity processes, fish production and, ultimately, livelihoods of dependent communities. The study demonstrated that temperatures around the two lakes had increased by about 0.01–0.03°C annually since 1980. This is consistent with what had been reported globally (IPCC, 2007) and across the tropics (Lewis et al., 2004) for the last three decades of the 20th Century, when climate change intensified. The study also demonstrated an increase in rainfall that was consistent with the general trend for the East African region (IPCC, 2007).
The increase in rainfall around Lakes Wamala and Kawi should have sustained the normal lake water levels since most of the water in aquatic ecosystems within the East African great lakes region is gained through direct rainfall (Bootsma and Hecky, 2003). In lakes that do not experience significant water abstractions for large-scale water uses, such as irrigation and hydroelectricity power generation, lake water levels are expected to remain high as long as the rainfall fluctuates within normal pattern (Sewagudde, 2009). The findings of the present study contradict the above notion, noting that the mean lake depth of Lake Wamala, on average, has decreased since 1970, despite the continuous increase in rainfall (Figure 5), although high mean depth >4 m was observed during high rainfall events and very low mean depth <2 m during drought events. The exception to the latter is the decrease in mean lake depth after 2000, yet most of the years had rainfall above long-term average. This observation indicates that there might have been a change in the hydrological pattern of the lake. Although most of the water in East African Rift Valley lakes is gained through direct rainfall, a similar or even greater quantity is lost through evaporation (Sewagudde, 2009). Thus, the current low depth of Lakes Wamala and Kawi can be attributed to higher evaporation rates (Table 1) associated with increased temperatures and wind speed, especially after 2000 (Figures 2 and 3). These observations imply that although there may be an increase in rainfall associated with climate variability and change, the water gain may be offset by increased evaporation associated with increased temperature and wind speed.
The changes in lake level in relation to rainfall observed for Lakes Wamala and Kawi have also been observed in major lakes within their respective lake basins. For example, the water level of Lakes Victoria and Kyoga increased by about 2 m in 1960 following El-Nino rains (Welcomme, 1970; Ogutu-Ohwayo et al., 2013), but have since then been decreasing and by 2004, lake levels had dropped to pre-1960 levels (Smith, 2011; Ogutu-Ohwayo et al., 2013). The only difference between these lakes and, either Lake Wamala or Lake Kawi, is that because of the large size (especially for Lake Victoria, which is 68,800 km2 and 80 m deep), they may not manifest rapid fluctuation in lake water levels. Lake size is important in buffering evaporation as long as rainfall fluctuates in a normal pattern (Sewagude, 2009). The recession of Lakes Wamala and Kawi, even when rainfall is high, could have therefore been enhanced by their smaller size (i.e. 250 km2 and 10 km2 for Wamala and Kawi, respectively) and their inability to buffer high evaporation rates normally preceded by increased temperature and wind speed. The fact that the rate of water loss from both lakes attributable to evaporation currently exceeds the rate of water gain from direct rainfall (Table 1) suggests that any period of lower than average rainfall will be accompanied by a corresponding decrease in lake depth and lake surface area. This happened between the years 1984 and 1995, when the lake surface area shrank by half and mean depth by three quarters (UNEP, 2009). Given the recent meteorological trends presented both in this study and literature, the current low water levels of these lakes may not reverse in the next few decades if the observed trends in temperature and wind speed continue, or if the increase in rainfall in the region is not sufficiently high to compensate for evaporation losses.
These changes will have implications for lake productivity processes, life history of organisms, fish biomass and, ultimately, the livelihoods of fisheries dependent communities (Barange and Perry, 2009; FAO, 2010). There are a number of examples in African small shallow lakes and marginal areas of deep lakes to support this. The shrinking of the surface area Lake Chad by about 90% following a period of drought between 1971 and 1977 was accompanied by a decrease in the number of fish species from 40 to 15 (Leveque, 1995), whereas fish production in Lake Chilwa varies between zero and 24,000 metric tons annually, depending on lake water levels (Allison et al., 2007; Njaya et al., 2011). In Lake Tanganyika, climate warming decreased primary productivity by 20% and fisheries yield by 30% (O’Reilly et al., 2003), although this was later contested by Sarvala et al., (2006). The decline in the catch per unit effort of Kapenta Limnothrissa miodon (Boulenger, 1906) in Lake Kariba between 1986 and 1997 was largely attributed to a 1.9ºC increase in temperature (Ndebele-Murisa et al., 2011). In Uganda, stocks of small pelagic clupeids, such as Rastrineobola argentae (Pellegrin, 1904) in Lake Victoria (Taabu-Munyaho et al., 2014), Neobola bredoi (Poll, 1945) and Brycinus nurse (Rupell, 1832) in Lake Albert (Ministry of Agriculture Animal Industry and Fisheries [MAAIF], 2012) have increased to contribute between 40-80% in the total catches, during the period when climate warming intensified. The changes in lake depth of Lake Wamala were accompanied by changes in catches and life history characteristics of the Nile tilapia, Oreochromis niloticus (Linnaeus, 1758) (Natugonza et al., 2015). These changes affect livelihoods of people who depend on aquatic resources. Studies on the effect of climate change on fisheries productivity of specific aquatic ecosystems are needed to guide management of fisheries for resilience and, to develop adaptation strategies enabling affected communities to sustain their livelihoods.
Additionally, measures are needed to lessen the impact of climate variability and change on aquatic ecosystems due to land use and land cover change within the lake catchment areas. For instance, around Lakes Wamala and Kawi, the wetland cover recedes with lake decreasing lake water levels. Following the recession, the communities around the lakes encroach on riparian zones and, in some cases, cultivate up to the lake shoreline, exposing the lakes to land based contamination (Musinguzi et al., 2015). The international policy on management of lake shores and river banks is based on the principles of coastal area management, which commits states to put in place measures to protect lake shores and river banks. In Uganda, the National policy for the conservation and Management of Wetland resources (GoU, 1995) aims at promoting conservation of wetlands by prohibiting reclaiming, draining and destroying of wetlands. The policy also requires that there should be a protection zone of 100 m from the highest water mark for the rivers and 200 m from the lowest water mark for lakes, but this is hardly enforced. The major challenge with this policy is that the highest and the lowest water marks vary with the lake water levels. This has resulted in encroachment, especially on those shallow lakes like Wamala and Kawi, when the shoreline recedes. This policy requires to be modified to include provisions aimed at demarcating the high and low water marks for specific water bodies, taking into consideration the influence of climate change on the lake shoreline. Additional policy interventions, such as afforestation and reforestation (with autochthonous species, which are not commercially targeted), and efficient use of water will need to be incorporated so that human activities do not exacerbate the impact of climate change on aquatic ecosystems. All these measures should be accompanied with sensitization programs for the riparian populations.
We wish to thank the Director of NaFIRRI for logistical support, UNMA and DWRM for making available their data on climate and water balance, respectively.
This work was done with financial support from The Rockefeller Foundation.