Modelización de datos de geolocalización de especies arbóreas mediante redes complejas. Aplicación al caso andaluz

Network analysis is an essential tool for understanding complex systems through the relationships between their constituent elements. In this project, we map large tree population samples gathered from public data bases into either spatial networks (graphs) of individuals or networks (graphs) of species. In particular, spatial networks are obtained in terms of a distance threshold r: two trees are linked if they are apart less than r. The species network is built from aggregating individuals of the same species and linking two species when at least two or, in general Tp, individuals of each species are closer than r. The topological properties of these networks are studied as a function of r and Tp. Furthermore, this work presents a methodology that brings together various procedures: access to species distribution data, creation of efficient code to form the graphs and to study their properties, graphical or geographical representation of these, etc. which, as a whole, make it possible to carry out an exhaustive analysis of these types of relationships, providing information on both the structure of any tree population and the species coexistence An example is given by the application of this methodology to two Andalusian ecoregions: Andalusian Sierra Morena and Sierra Nevada; They are chosen due to the exceptional coverage of geolocation forest data in this autonomous community, thanks to the work of REDIAM. Despite being a sample of the real population, the high number of trees of the inventory makes the computation of the network properties in a personal computer a big challenge. This project shows that only the design of efficient strategies that combines various computational and mathematical tools can provide a solution.