RESUMEN
Multifractal properties of diurnal temperature range (DTR) have been analyzed in this work, using validated data of maximum and minimum temperature from 197 weather stations in Southern Spain (Andalusia region). DTR is a crucial factor to characterize the regional climate, providing more information than the average daily temperature. Apart from climate change studies, one of the most important applications of DTR in Agrometeorology is as an input variable in the solar radiation or reference evapotranspiration estimation models based on the temperature. With the aim of obtaining a detailed information for different time scales, different multifractal approaches have been applied. Different quality control methods such as range/limits or persistence tests were previously applied in order to detect incorrect and anomalous values, being discarded in the subsequent analysis. The DTR scaling of moments has been analyzed and the moment scaling exponent function K(q) has been obtained, finding some differences between weather stations. In addition, multifractal dimension (D1) and multifractal degree (MD) were also estimated, revealing differences at coastal and inland locations that show heterogeneity across the region, including its multifractal nature and its invariance for a range of scales. The nonlinear characterization carried out in this work improves the understanding of DTR as an indicator of climate changes, and it can have a very positive impact on the calibration of regional models for estimating solar radiation or reference evapotranspiration based on the temperature. This multifractal characterization can be used to group stations with similar nonlinear dynamics, regardless of their geographical features, in such a way that more accurate coefficients than conventional ones are used.