LAND COVERS ANALYSES DURING SLASH AND BURN AGRICULTURE BY USING MULTISPECTRAL IMAGERY OBTAINED WITH UNATTENDED AERIAL VEHICLES (UAVs

Abstract

Background. The Milpa is one of the traditional agricultural systems of Yucatan, Mexico; it is implemented by using the agricultural procedure called “Slash & Burn”. Quality and quantity of forest fuel (i.e. biomass) are two of the main factors related to burn severity. Burning affects soil properties related to fertility and crop production. The use of new approaches of remote sensing technologies such as Unattended Aerial Vehicles (UAVs), can allow studying the importance of fire in agriculture to improve productivity in slash & burn agricultural systems. Objective. Analyze the land covers and its influence on the severity of the agricultural burning in a "Milpa" agroecosystem with multi-spectral images acquired by UAVs. Methodology. The study site was located in the municipality of Tzucacab, Yucatan, Mexico. Two plots were selected [10-15 and 20-25 years of fallow]; land cover was characterized before and after slash & burn. Three multispectral sensors [Red, Green, Blue (RGB); Near Infrared (NIR) and; Thermal Infra-Red (TIR)] were mounted on UAVs, to obtained multispectral imagery and generate orthomosaics for later analyses. Results. With the imagery, the Normalized Difference Vegetation Index [NDVI] was calculated and its spectral behavior evaluated. The imagery was used to analyze the fire intensity. On RGB imagery, patterns of areas with greater dry biomass cover associated to high burn severity and, areas with green vegetation or naked soil associated to low burn severity were observed. Land covers with high fuel potential showed low NDVI index values. Implications. The analyses of the multispectral imagery taken by drones allow the quick evaluation of the land covers and the intensity of agricultural fires, with the pertinent adjustments, in the near future this could become a standard methodology to accomplish this kind of evaluations. Conclusions. This approach allowed to analyze the state of land covers to visually assess the quality of fuel and its influence on the intensity of an agricultural fire. The RGB and NIR imagery obtained by UAVs can be a good tool to predict the intensity of an agricultural fire, TIR imagery could be used to find mathematical relations between land covers and fire intensity.