Integración de herramientas geoespaciales y datos multifuente para la evaluación de la dinámica de la cobertura forestal en Centroamérica, con aplicaciones potenciales a otras regiones tropicales

Resumen

A theoretical-methodological framework was developed and applied to elaborate cost-efficient mapping on forest cover and its dynamics for the countries of Central America and its scalability was explored in other tropical regions of greater territorial extent selecting for replicate the methodology, the territories of Democratic Republic of the Congo and the State of Pará in Brazil. The potential provided by the combination of satellite remote sensing and geospatial analysis with GIS was exploited, incorporating multi-source secondary data, to automate classification flows and post-classification adjustments (based on expert criteria) of Landsat images from the 2000s , 2012 and 2017. As result, national maps and statistics on forest cover and its dynamics were generated in the different study areas. From the results, quantitative spatial analysis techniques were applied, to generate for Central America a metric at the municipal level called Focus Priority to Stop Deforestation (PFFD by its acronym in Spanish), from which different priority levels were established to focus actions aimed at reducing deforestation in a regional context. The results obtained indicate that in Central America the percentage of forest area decreased from 44% in 2000 to 38% in 2017. The average deforestation of Central America in the period 2000- 2012 was 197,443 ha/year and that of 2012-2017 of 332,243 ha/year. Deforestation for the entire period analyzed (2000-2017) was 264,843 ha/year. The first monitoring event included in this study showed a deforestation of 167,976 hectares for the year 2018. The greatest intervention emergencies according to the PFFD metric were defined for the municipalities located in the areas of broadleaf forests of the Selva Maya (Guatemala - Belize), the Moskitia (Nicaragua - Honduras), the Indio Maíz Reserve (Nicaragua) and the Darien forests (Panama). The 3 highest priority municipalities to stop deforestation were: San Andres in the Department of Petén, Guatemala; Bluefields in the Autonomous Region of the South Caribbean Coast of Nicaragua (RAAS by its acronym in Spanish) and the municipality of Iriona in the Department of Colon, Honduras. In the State of Pará, Brazil, forest cover decreased from 81.3% in 2000 to 75.3% in 2017. In the Democratic Republic of the Congo (DRC), forests decreased from 59% in 2000 to 52.2% in 2017. In terms of accuracy of the classifications obtained, for tree/non-tree classes, as well as for forest/non-forest classes, the lowest were obtained in Honduras, Guatemala, El Salvador and Nicaragua corresponding to the index scale from Kappa to the category of moderate concordance. The best accuracies were obtained in Belize, Panama and Costa Rica, according to the Kappa index in the category of considerable concordance. With the methodological development implemented in this research, a progress has been made respect to the previous studies implemented in Central America, moving from classifying individual scenes of satellite images and separately, to do so in a regional context. In addition, workflows that reduce the replication time of the methodology have been automated, as evidenced in the replica territories selected in this study