Assessment of Meteorological and Streamflow Drought Characteristics in Agago Sub-Catchment, Northern Uganda

Abstract

A sound understanding of the characteristics and interactions between meteorological and streamflow drought patterns is vital for drought early warning. This study assessed meteorological and streamflow drought characteristics in Agago sub-catchment, Northern Uganda. The study specifically: i) assessed the spatial-temporal characteristics of meteorological drought; ii) assessed the characteristics of streamflow drought, and iii) assessed the time-lags between meteorological and streamflow drought events in Agago sub-catchment. Archived climate and streamflow data obtained from Princeton University datasets and Directorate of Water Resource Management; Uganda were used. A suite of analytical approaches including non-parametric tests, descriptive statistics and spatial interpolation were utilized to determine the significance of spatial-temporal trend of meteorological drought events and streamflow drought events. A correlation analysis was conducted to determine the relationship between the occurrence of meteorological and streamflow drought. Results of the study revealed that meteorological and streamflow drought mutually occurred in the same years and meteorological drought events were mainly experienced in Abaneka area compared to the other areas in the sub-catchment. The probability of meteorological drought occurrence for a 3 months’ duration was 79.7% at all the four stations and these drought episodes were mainly observed from December to February. The most severe meteorological drought years recorded were 1983, 1998-1999, 2002, 2004, 2006-2007 2011-2012 and 2016. However, streamflow drought events were mainly experienced in December-June and the most severe streamflow drought years were recorded in 1987, 1991, 1998, 2002, 2004, 2007, 2012 and 2016. There was a significant positive relationship between meteorological drought (Standardized Precipitation Index (SPI) and streamflow drought (Streamflow Drought Index (SDI) in the sub-catchment. This relationship was however most pronounced in the SPI-12/SDI-12 compared to SPI-3/SDI-3 and SPI-6/SDI-6. Cross-correlation analysis indicated that SPI and SDI at 3 and 6 months timescale have a zero month time-lag whereas SPI and SDI at 12-month timescale reached a maximum correlation coefficient at one-month (30 days) timelag indicating that occurrence of streamflow drought in a catchment is delayed by at least one month after meteorological drought has occured. Therefore, a one-month lagtime should be used when considering drought early warnnig and management options for drought hazard in the sub-catchment.