Inclusión de relaciones interespecíficas usando Maxent.

Morales-Ortiz, José

2015

Descargar documento

Instituto Tecnológico de Costa Rica. Escuela de Ingeniería en Computación.

The species' distribution models are normally feed by occurrences, geo referenced data points where a specie has been seen, and climatic information such as precipitation, solar radiation, temperature, among others. The species geographic distribution not only depends on those variables, there are other factors such as the relationship between the species or the geographic barriers, as ridges and lakes. The present research try to improve the quality of the species' distribution models by including the relationship between the species. In order to achieve that three bird species and the relationship with the plants which serve them as food were defined. The selected species are the lapa verde (Ara ambiguus) and its relationship with the almendro amarillo (Dipteryx panamensis), the pájaro campana (Procnias tricarunculatus) and its relationship with the Lauracea family and finally the quetzal (Pharomachrus mocinno) and its relationship with the Lauracea family as well. In order to include the relationship between species into the species' distribution models a methodology in order to build a layer of "range of influence" were defined; with this new layer a set of experiments were defined, and those experiments verifies that by including the relationships between the species into the species' distribution models improve their precision. In order to measure the improvements the AUC curve were calculated for each model, in that way a numeric comparison were done.

Proyecto de Graduación (Maestría en Computación con énfasis en Ciencias de la Computación). Instituto Tecnológico de Costa Rica. Escuela de Ingeniería en Computación, 2015.

Sistemas de información geográfica; Distribución; Especies; Variabilidad climática; Aves; Plantas;