Spatial and Temporal Analysis of Precipitation and Drought Trends Using the Climate Forecast System Reanalysis (CFSR)

Abstract

Determination of spatial and temporal patterns of precipitation and drought remains a challenge in many arid and semi-arid regions. Generally, in these areas, precipitation measurements obtained from field stations are few or unreliable. Reanalysis precipitation data has opened up new large-scale hydrological application in data-sparse or ungauged catchments. This study presents a method based on GIS, statistical tests, and satellite-based precipitation for the analysis of spatial and temporal precipitation and drought trends. As a case study, we used 35 years (1979–2013) of precipitation data from the climate forecast system reanalysis (CFSR) for the Mexican state of Durango. For trend detection, we used the statistical test of Mann–Kendall at 5% standard precipitation index timescale of 12 months (SPI-12) to analyze the spatiotemporal trends of drought, and principal component analysis (PCA) was applied to characterize the spatial patterns. In many of the grid locations analyzed, the precipitation trends were not statistically significant. However, the months within winter and spring, except for June, showed a decreasing trend in precipitation, while the months of July, August, and September showed an increasing trend. This variation in rainfall intensity agrees with the climate pattern reported for this region. The frequency of drought events for each month during the period of analysis was mapped. The incidence of drought events was higher in June and September. Drought events are present all around the state of Durango, but they are more frequent in the northern part, in the Sierra Madre Occidental, Mapimi, and Laguna regions.