Impact of Land Use/cover and Climate Change on Surface Water Resources in Semi-arid Lokok and Lokere Catchments, Uganda

Abstract

Globally, the need to conserve and sustainably manage freshwater resources has become a major goal due to concern over the increasing impact of land use and land cover (LULC) and climate change on water resources especially in arid and semi-arid areas of eastern Africa. In Lokok and Lokere Catchments of the semi-arid Karamoja subregion, Uganda, stream flow has been highly fluctuating, with prolonged hydrological drought. However little is known about the contribution of each of these factors on the stream behavior. To assess the historical and plausible future impacts of land use/cover and climate change on water resources in the Catchments, LULC for 1984, 1994, 2003, and 2013 were established through unsupervised and supervised classification of satellite images. Qualitative information was used to obtain a historical account of LULC in order to identify its drivers. By cross comparing 1994 and 2003 LULC, the automatic multi-perceptron neural network built on Markov chain modeling method along with multi-criteria evaluation strategies, all embedded in the IDRISI Land Change Modeler (LCM) software, was applied to develop three plausible LULC scenarios for the Catchments, for 2030 and 2050, namely Business as Usual (BAU), pro-farming policy scenario (PFP) and pro-livestock policy scenario (PLP). The Model was validated with the 2013 LULC. In addition, spatio-temporal trends and variability in temperature and rainfall time series (1980-2009) were assessed using the ordinary least square (OLS) regression and Cumulative Sum (CUSUM). The standard temperature index (STI) and standard rainfall anomaly (SRA) were used to detect hot and dry years respectively. Furthermore, an ensemble of the four locations (stations) future temperature and rainfall scenarios for three periods (2010–2039, 2040–2069, and 2070-2099 or early, mid and end-century respectively) downscaled using the delta method from twenty of the latest International Panel on Climate Change (IPCC) climate models embedded in the Agricultural Model Intercomparison and Improvement Project (AgMIP). These were compared with 1980-2009 as the baseline period. Using the Soil and Water Assessment Tool (SWAT), the water balance of Lokok and Lokere Catchments was simulated using 2003 LULC and 1980-2009 MERRA climate data for the baseline period. Simulations were also made using 2030 and 2050 LULC, with early-century and mid-century ensemble of climate data respectively, under RCP 4.5 and 8.5. These simulation results were compared with the baseline to obtain a change in water balance components. xvii Results showed a change in LULC, especially in the conversion of woodlands and bushlands into small-scale croplands, with degradation of woodland and bushlands increasing grassland area. The area under grasslands, and the largest, increased from 43.64 in 1984 to 60.05 percent in 2013. Small-scale farming (SSF) steadily rose from 9.67 percent in 1984 to 15.69 percent in 2013, at an annual rate of 2.1 percent. The long agro-pastoralism tradition of the inhabitants is expected to continue in the years 2030 and 2050 as SSF would increase in all LULC scenarios. And increase in crop cultivation would persist to the year 2050 even if policy shifts to promote livestock rearing, pro-farming policies would, in both the 2030 and 2050 modeling periods, result in the reduction of grassland as SSF substantially increases – doubling the 2003 land area in 2050. Catchments’ temperature significantly increased (p < 0.05) in the 1980-2009 period. Minimum temperature (Tmin) increased faster than maximum temperature (Tmax), especially during the rainy seasons. During the dry season, Tmax significantly increased and was more variable than Tmin. The increase in rainfall was lowest in Amuria Station which received the highest rainfall. Total annual rainfall significantly increased only in Kotido and Moroto stations during 1980-2009 period. Variability of both temperature and rainfall was higher in the first decade of analysis than in the third; and positive shifts in temperature trends occurred after 2000. Compared to the baseline, temperature was projected to increase, and change in Tmin would be higher than the change in Tmax. Tmax in the Catchments would change by 0.7oC and 0.8oC in the early century, 1.3 oC and 1.9 oC in the mid-century, and 1.7 oC and 3.3 oC in the end-centuries – for RCP 4.5 and RCP 8.5 scenarios. Tmin would change by 0.9 oC and 1.0 oC, 1.6 oC and 2.1 oC, and 2.0 oC and 3.8 oC – for RCP 4.5 and RCP 8.5 in the early, mid, and end-centuries respectively. Future increase in temperature would be higher in the cooler and wetter months and seasons (March-April-Mary, MAM; June-July-August, JJA) than in the warmer season (December-January-February, DJF) – which shows a temporal variation in change. And while rainfall in the catchments is projected to increase by 10 and 8 percent in early-century, 15 and 16 percent in mid-century, and 20 and 30 percent in end-centuries – for RCP 4.5 and RCP 8.5 respectively – the increase would be higher in the drier periods than in the wetter ones. xviii The simulated water balance of the catchments showed that evapotranspiration (ET) is the major component of the hydrological budget in the catchments, as over 97 percent of the precipitation received is lost through ET. As a result, the values of the other components (surface runoff, lateral flow, return flow, and groundwater recharge) are so small, and their changes in percentage terms would be too large. Under future climate scenarios, the percentage increase in water yield would range from 79.5 percent under early-century RCP 8.5 to 204.7 percent under mid-century RCP 4.5M. However, an increase in water yield would be marginal under change in LULC, ranging from 5.7 percent to 18.4 percent under BAU2030 and 2050 pro-farming scenario where SSF is expected to increase. Water yield is expected to be relatively high under combined future scenario of LULC and climate change. It would range from 193.7 percent under the 2050 pro-livestock LULC and RCP 8.5 mid-century climate to 223.2 percent under the 2050 pro-farming LULC and RCP 8.5 mid-century climate. The study demonstrates that the current changes in small holder farming in the catchments would continue into the mid-century (2050). However, grassland would still be more dominant but could be less supportive to livestock herding due to fragmentation by cropland and restriction to sharing of grazing grounds. The results also show that the present increase in rainfall and temperature could continue. And while change in LULC would result in relative increase in water yield, change in climate would have more substantial increase in the water balance of the Catchments. Whereas the projected change in water yield appears minimal, it could have positive ecosystem, social and economic impacts. Given that mobile herding is more adaptive to climate variability, policies and strategies that improve both crop and livestock production could be more beneficial to the population.