Hydrological Drought Monitoring, Forecasting, and Projection System Development in Ethiopia
No Thumbnail Available
Date
2024-05
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
Due to its multifaceted effects and gradual beginning and end dissemination, the concept of drought is very debatable. However, experts divided drought into four categories: meteorological, agricultural, hydrological, and socioeconomic drought, which were characterized according to the lack of precipitation, soil moisture, surface, and subsurface water availability, and the imbalance between supply and demand, respectively. This study is primarily focused on hydrological drought monitoring and forecasting using two hydrological drought indices namely streamflow drought index (SDI) and surface water supply index (SWSI) across eight river basins in Ethiopia. Besides this spatiotemporal variability of meteorological drought in the Abbay river basin was analyzed using the standardized precipitation index (SPI) and reconnaissance drought index (RDI) to compare hydrological and meteorological drought correlation. Meteorological and streamflow data were collected from 50 rainfall stations and 35 streamflow stations from 1973 – 2014.
The result indicates several severe and extreme drought events occurred during the 1980s and 1990s compared to the 2000s and 2010s. The most identified severe drought years are 1975, 1981, 1984, 1986, 1991, 1994, and 2010 whereas 1983, 1984, 2001, 2003, and 2010 were extreme drought years. The spatial analysis shows that the Tekeze, Abbay, and Baro river basins have similar characteristics; Awash and Rift Valley river basins show relatively the same character, and Genale Dawa and Wabishebele river basins have a similar drought trend. However, the Omo Gibe River basin has a unique character in that the severe drought occurred in a different year than other river basins.
The statistical correlation of RDI and SPI, and SPI and SDI, RDI and SDI were found 0.95, 0.87, and 0.83 respectively, at an annual time scale. It implies that both hydrological and meteorological drought indices have an excellent correlation for long-term time scale and it also indicates the possibility of using SPI and RDI indices instead of SDI in areas having streamflow data scarcity. On the other hand, SDI and SWSI have a good relationship in all river basins except the Rift Valley basin. However, the overall result of hydrological drought analysis using SDI is better than SWSI compared to the previous historical drought events.
Climate change-induced hydroclimatic hazards have increased from decade to decade overall in the world. So, projecting drought conditions for the future plays a great role in hydrology. In this regard, in this study, streamflow was forecasted from 2026 to 2099 using an artificial neural network (ANN) model using downscaled precipitation data as input to the model. Recently, ANN has been a suitable forecasting technique in water resource engineering. The future input data was downscaled using the Regional Climate Model (RCM) and the downscaled data have bias corrected using the linear scaling bias correction technique. The ANN model was trained and tested using historically observed precipitation and streamflow data as input and output variables respectively. Then the bias-corrected precipitation data is used to forecast future streamflow. The statistical performance parameters such as the Mean Absolute Percentage Error (MAPE), Root Mean Square Error (RMSE), and Coefficient of Determination (R2) were used to evaluate the performance of the ANN model and the result shows ANN has an acceptable value in humid areas than arid areas to forecast streamflow from precipitation data. Finally, the future hydrological drought condition of the country was investigated using SDI based on forecasted streamflow. The finding shows that the 2030s, 2040s, and 2060s are the expected critical and worth drought years in the country.
The result of this thesis presents the establishment of hydrological drought monitoring, forecasting, and projection skills. The articles presented in this thesis seek to offer novel perspectives on hydrological drought monitoring, forecasting, and projection, emphasizing existing difficulties and opportunities for the generation of valuable information that can help decision-makers and policymakers in the management of water resources. Overall, the historical and future drought trends of Ethiopia indicate the country is frequently hit by severe droughts. So, appropriate drought mitigation measurement is needed. However, the commonly adopted drought mitigation trend in Ethiopia is a reactive approach which is a short-term drought mitigation technique during the drought event that has occurred. However, this approach will never bring a sustainable solution for drought-victim societies in the country. Therefore, it is better to shift into a new paradigm, a proactive approach which is a long–term drought mitigation system. To do this, the national government should actively promote the construction of water conservation infrastructure like dams and reservoirs, afforestation (green legacy), and the development of a national drought policy.
Description
Keywords
hydrological drought, meteorological drought, monitoring, forecasting, regional climate model