Assessing the effectiveness of remotely piloted aircraft to map exposed soil in urban mangroves

Abstract

Although mangroves are ecologically important coastal ecosystems and laws are in place to ensure their protection, anthropogenic activities continue to cause the degradation and/or suppression of mangrove vegetation. Traditional methods to measure and monitor this process, including the use of medium spatial resolution orbital images, are unsuitable for fine-scale environmental degradation and recovery analyses, including the measurement of degraded areas in and around mangroves. Thus, this study aims to analyze the effectiveness of using images from remotely piloted aircraft (RPA) in the mapping of exposed soil areas in mangroves. Imaging with RPA was performed in 22 urban mangroves in Paranaguá, Paraná State, Brazil. Orthomosaics were generated from the collected data and submitted to supervised classification. We then calculated global accuracy and Kappa indices and commission and omission errors. Based on data from the RPA images, the identification of areas of exposed soil on the margins and interior of mangroves was effective since the global accuracy index was higher than 96% for all classified orthomosaics and the Kappa index was above 0.95, indicating excellent classification. The mapping shows different concentrations of exposed soil areas in the analyzed mangroves, enabling us to identify three regional patterns of vegetation degradation. The results can inform municipal planning, including revisions to the Integrated Development Master Plan, Basic Sanitation Plan, and Land Regularization Plan. This information may also be used in studies on the recovery and monitoring of mangrove vegetation.

Keywords:
Remote sensing; Coastal ecosystems; Environmental degradation; Suborbital images; Drone