Adaptation of the L-moments method for the regionalization for maximum annual temperatures in Colombia

The present study has developed a regional frequency analysis for Annual Maximum Temperature (AMT) in the hydrographic basins of Colombia. The L-moments methodology was applied for the regionalization. Five stages were considered to apply the methodology: data analysis; the L-Moments estimation; identification of homogeneous regions; fit of probability density functions (pdf) to observed data and estimation of quantile values; and finally, the developing and drawing of regionalized maps. Overall, fifteen homogeneous regions were identified and selected for the regionalization of AMT, which meet specific criteria of homogeneity and discordance. Several pdf for regional frequency analysis were tested in order to select the best probability function. Finally, regionalized temperature maps were generated for several return periods. Using the L-Moments methodology, the regionalization procedure was done using the average of AMT as the key scale parameter. The regionalization procedure ensures, as far as possible, a coherent-basin approach: the boundaries between homogeneous regions were drawn, complying with the catchment borders.