Assessing Characteristics of Drought and Its physical mechanism over Ethiopia using Observational and RegCM4 Driven Indices

This study focuses on the analysis of dryness/wetness over Ethiopia from 1982 to 2009 using Standardized Precipitation Index (SPI) and Palmer Drought Severity Index (PDSI), which are derived from observational and RegCM4 dataset over Ethiopia. Trend empirical orthogonal function (TEOF), regression and wavelet analysis are used to investigate the long term change, frequency and intra-annual variability of drought over Ethiopia. All method analysis of PDSI and SPI are consistent and showed, the southern regions to be more drought-prone than other regions of Ethiopia and statistically significant dry trend are dominated in particular during the recent decade. By contrast, the likelihood of drought occurrence of northern regions is considerably less and dominated by low frequency signals. Both indices indicated that the North and Northwest regions experienced frequent and worst drought conditions centered at the year 1983/1984 and similar conditions centered at the year 2007-2009 over the southern and southwestern regions have been identified. The first two most dominant trend principal components (TPCs) of observational driven PDSI and SPI-12 with corresponding RegCM4 driven TPCs of PDSI and SPI-12 indicates significant correlations (correlation exceeding ~0.8). In addition, the corresponding patterns (TEOFs) of RegCM4 and observational driven drought indices are so similar. Generally, the RegCM4 shows a good performance in simulating the multi-scale spatial and temporal variability of drought occurrence over Ethiopia. The correlation pattern of trend principal components (TPCs) of PDSI and SPI-12 with raw sea surface temperature (SST) indicates that the possible association of the observed variability of drought. The equatorial pacific, the northern and equatorial Atlantic are correlated significantly with first dominant TPC of drought indices, whereas ENSO, Indian and Atlantic Ocean significantly correlated with the second dominant TPC of drought indices. In addition we observe the association between ocean basin and drought indices is highly seasonal dependant with the strongest association in the spring. The physical mechanisms responsible for such linkage in the atmosphere are examined using both ERA-Interim (ERIM) and RegCM4 dataset. Moreover, it was found that droughts occurred in events of strong El Niño years mainly severe and wider areal coverage. During events of negative values of Indian Ocean dipole (-IOD) the southern and west border regions of the country hit by severe drought. The northern and some part of Ethiopia during events of negative anomaly SST of North Atlantic suffered at a significant level by drought conditions. Overall, the common power variance, phase, frequency and confidence of the time-lagged relationships between oceanic basin TPC and drought indices TPCs indicate that oceanic parameters can be used as indicators of drought occurrence in Ethiopia.