Entropy analysis and pattern recognition in rainfall data, north Algeria

Abstract

Algerian climate is characterized by the transition between the subtropical climate in the north and the hot Saharan climate in the south. Understanding the spatiotemporal variability of rainfall patterns in such areas has significant implications for water resources management. To account for the spatial variation in the rainfall pattern of north Algeria, Tsallis entropy analysis, pattern recognition, and Precipitation Concentration Index (PCI) have been analyzed over a period of 33 years (1980–2013). The rainfall trend is identified by analyzing the results of the structural break test and Mann-Kendall (MK) trend tests. The Tsallis entropy produced spatial patterns for a better understanding of rainfall characteristics and the results show that entropy values were higher for higher rainfall values. The magnitude of rainfall change indicates that a large amount of rainfall occurs over the northern country boundary compared with the central part of Algeria. The rainfall record shows a structural break in the year 1992 during the selected time period. The MK trend analysis revealed a significant decreasing trend over the central and no trend in north and south of the study area. The PCI indicates a moderate rainfall concentration across the northern border region compared with a significant irregular rainfall distribution over the central region. The Wavelet Coherence Analysis (WCA) between El Nino Modoki (EMI) and Southern Oscillation Index (SOI) events on monthly rainfall data were also investigated to find a possible influence of global climatic indicators on the rainfall events. The results show a significant correlation of EMI and SOI with the rainfall pattern of north Algeria.