Introducing L-Correlation for Climate Network Construction: Application to droughts analysis in the Iberian Peninsula

The theory of complex networks, particularly climate networks (CN), is frequently used in the analysis of climate data at different scales, serving both to investigate climate dynamics and to understand extreme phenomena and their temporal evolution. The classic methodology for constructing CN relies on the use of correlation between pairs of nodes in the network, to determine the existence of a given link. The resulting network structure can yield valuable insights about the underlying physical phenomenon. The state-of-the-art reveals multiple methods for constructing CN, most of which are based on correlation or cross-correlation functions, which are L2-norms more sensitive to outliers. In this work, the use of L-correlation as a basis for the construction of CN is proposed. Based on this approach, it is possible to take advantage of the main benefits of the L-moments theory, including its multivariate extension, such as the availability of unbiased estimators and robust performance in the context of extreme events. The specific case study tackled focuses on analyzing different drought phenomena in the Iberian Peninsula using precipitation data in the Reanalysis period. The results obtained have been contrasted and validated through a comparative analysis based on traditional CN methods. The conducted experiments suggest an active desertification process in this region, consistent with the state-of-the-art findings in hydrological process characterization studies. Future research could aim to enhance the interpretability of results derived from CN constructed using higher-order L-comoments, thereby facilitating the application of this method to additional case studies.