Comparison of the spatiotemporal variability of rainfall from four different interpolation methods and impact on the result of GR2M hydrological modeling—case of Bani River in Mali, West Africa

Abstract

The climatic evolution of the Bani river watershed, the main tributary to the upper Niger River, is approached through the spatiotemporal variability of rainfall grids over the 1950–2006 period. The analyses are conducted, and their results compared, using four different methods of spatial interpolation of rainfall fields: the spline, kriging, weighted inverse distance, and nearest neighbor methods. The largest changes are observed for all of these grids, but differences—and in some cases divergent results—appear in the details. The analysis shows a substantial decline in rainfall, particularly marked in the center of the basin, during the 1970–2000 period with respect to the 1950–1969 period, and a slight upturn in the northern part, mainly since the beginning of the 1990s. The rainfall deficit can be attributed to a combination of factors: an earlier and drier end of the rainy season, less precipitation in the middle of the rainy season, more dry days and lower amounts of precipitation on rainy days. Two drought indices—the Effective Drought Index and Standardized Precipitation Index—revealed that the maximum duration of drought events increased most in the central part of the basin. Lastly, to supplement this comparison of methods of spatial interpolation of rainfall fields, the sensitivity of a hydrological model (GR2M) to rainfall data was tested. Given the distribution and density of rain gauge stations available in the Bani watershed, the kriging method is found to yield the best hydrological modeling performance.