Modelo de riesgo de fuego para la ecorregión de los llanos colombo-venezolanos
While fire has been part of the natural history of ecosystems, those that go out of control often represent significant threats to public safety, infrastructure, biodiversity and forest resources (Martell, 2007), and are considered one of the most important disturbance factors, especially in tropica...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis de maestría |
| Estado: | Versión aceptada para publicación |
| Fecha de publicación: | 2020 |
| País: | Colombia |
| Institución: | Universidad Nacional de Colombia |
| Repositorio: | Repositorio UN |
| Idioma: | español |
| OAI Identifier: | oai:repositorio.unal.edu.co:unal/79128 |
| Acceso en línea: | https://repositorio.unal.edu.co/handle/unal/79128 |
| Access Level: | acceso abierto |
| Palabra clave: | 570 - Biología Aprendizaje automático Riesgo Peligro Vulnerabilidad Random Forest Machine learning Risk Danger Vulnerability |
| Sumario: | While fire has been part of the natural history of ecosystems, those that go out of control often represent significant threats to public safety, infrastructure, biodiversity and forest resources (Martell, 2007), and are considered one of the most important disturbance factors, especially in tropical and subtropical areas (G. R. Van Der Werf et al., 2010). The ecoregion of the Colombian-Venezuelan plains is characterized by a constant presence of fires (Chacón et al., 2015) caused mainly by human action, derived from land management and preparation practices that include slashing and burning (Armenteras et al., 2005; Leal et al., 2019). This work evaluated fire risk in the region of the colombian-v.00enezuelan plains, based on the probability of occurrence (hazard) and vulnerability at the ecological level. In order to model the probability of occurrence, the automatic learning model Random Forest was implemented, fed with variables associated with topography, climate, vegetation and human presence. To evaluate ecological vulnerability, information regarding biodiversity, conservation and fragmentation in the study area was used. The results of the probability of occurrence model indicate that the most important variable is the NDWI (Normalized Difference Water Index), an index that has been shown to offer better results for estimating the moisture content of living fuel and predicting the risk of fire occurrence in the case of savannah ecosystems (Cheng et al., 2006; Verbesselt et al., 2006, 2007). Finally, both subindices were integrated into a total risk index in order to identify those areas where the occurrence of this type of event is most likely to result in significant ecological damage. The results show that the high and very high probability of occurrence zoning is represented by 544,498 and 499,740 ha respectively, while the high and very high vulnerability zoning is less than 64,500 and 2,298 ha. It was found that in some special management figures, such as El Tuparro National Park (Colombia), Cinaruco Integrated Management District (Colombia) and Cinaruco-Capanaparo National Park (Venezuela), very high probability zoning predominates, which for these areas represents 47.7%, 56.9% and 37.8% of the total area of each park. Evaluating fire risk is a key process within the context of controlling and managing this type of event and represents an important tool for planning on a regional scale. This evaluation allows, among other things, to assess the suitability of landscape protection measures and different types of coverage (Costa et al., 2011), to support planning and protection of forest areas, to take surveillance measures in high-risk areas, to reorganize slash-and-burn practices and to strategically allocate resources to deal with this type of disaster (You et al., 2017). |
|---|