Global reference evapotranspiration clustering and its relation to the Köppen-Geiger climate classification
Nikolaos Malamos, Aristotelis Tegos, George Bourantas, Christos Chalvantzis, Demetris Koutsoyiannis
Abstract
This study introduces a global clustering approach for reference evapotranspiration (ETo) based on spectral clustering techniques. By analyzing 3139 stations worldwide, the method segments the globe into 45 unique clusters that reflect shared ETo dynamics. This clustering technique not only enhances the accuracy of ETo estimation but also provides a simplified method for computing ETo in areas where meteorological data are limited. The spatial extent of these clusters allows for the application of pre-calculated parameters, streamlining ETo calculation and reducing the computational complexity typically involved. This approach draws an analogy to the Köppen-Geiger climate classification, aligning ETo clusters with well-established climate zones, thus revealing patterns between evapotranspiration and global climate behavior. Moreover, the study demonstrates that the clusters can be utilized to calibrate other ETo estimation methods, offering a tool for guiding regional model adjustments. The validation of this method, using 30 stations from diverse climates, showed strong correlations and acceptable performance metrics, indicating its applicability for hydrological and engineering tasks. In conclusion, this global ETo clustering provides a robust, accessible framework for water resource management , irrigation planning, and climate adaptation strategies, particularly in data-scarce regions.