Invasion by Eragrostis plana Nees in areas of the Brazilian Pampa biome modelled with remotely sensed data and GARP species distribution model

• Autor(a) principal: González, José David Montoya
• Data de Publicação: 2020
• Outros Autores: Fonseca, Eliana Lima da, Perez, Naylor Bastiani
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Institucional da UFRGS
• Texto Completo: http://hdl.handle.net/10183/215478
• topic: Espécies invasoras : Brasil; Imagens SRTM; South African lovegrass; Grasslands; Rangeland; NDVI; MODIS; Landsat

In Brazil, the remnants of Pampa biome represent an area of high environmental fragility due to the expansion of the agricultural frontier and to overgrazing, which promotes conditions for the rapid spread and establishment of invasive species such as Eragrostis plana Nees. The areas most susceptible to invasion by this species are the areas degraded by overgrazing and intensive agriculture, abandoned crop fields, and roadsides. Considering the problems that arise from species invasion in natural areas, and particularly from Eragrostis plana in the Pampa biome, the objective of this study was to use the GARP (Genetic Algorithm Rule-set Production) species distribution model to map, at the local scale, the probability to invasion by this species using as input variables remotely sensed data, as well as, verify the influence of roads maps as input variable at model´s results. The environmental and topographic variables used as input variables were obtained from the spectral images of the MODIS-Terra and OLI-Landsat 8 sensors, from SRTM digital elevation model, and from road maps. The association between GARP species distribution models and remotely sensed data had positive effect in order to modeling plants patterns of invasion at local scale and a greater probability to invasion was found in areas nearest the roads, independent of the use it as the input variable in the model.