Análise da biomassa acima do solo do mangal a partir de imagens de radar do sensor ALOS-2/PALSAR-2 na Baía de Maputo

Abstract

Mangrove are unique coastal ecosystems composed by trees and shrubs adapted to salinity that grow along the coastal zone, between low and high tides and along riverbanks. The principal cause of mangrove degradation in Mozambique is the total or partial extraction of trees for use as domestic fuelwood and timber. In Mozambique, do not exist research of multiple linear regression using radar sensors. Hence, the need to evaluate radar images from ALOS-2/PALSAR-2 sensor in mapping aboveground biomass in the mangrove forest on the Maputo River. Biophysical parameters (tree height, aboveground diameter, volume, and biomass) that in 115 grids in 2021 March and April. The correlation between the variables and the backscattering coefficient of the HH and HV polarizations of the L-band estimated through biophysical parameters. The aboveground biomass estimated through allometric equations specific for each species. Backscattering coefficients of the HH and HV polarizations of ALOS-2/PALSAR-2 used to build multiple linear regression model. The results indicate the presence of four (4) mangrove species in the Maputo river: A Avicennia marina with 51 085 individual trees, Rhizophora mucronata with 3657, Ceriops tagal with 191 e a Bruguiera gymnorhiza with 697. The correlation between the biophysical parameters and the backscatter coefficient was weak low, with the Person coefficient of 0.151 corresponding to 15.1% for the backscatter in the HH polarization and -0.010 for the backscatter coefficient in the HV polarization. The model shows an R 2 of 0.051 corresponding to 5.1% with an average of 528172.15 Mg/ha. However, these results indicate that the multiple linear regression relationship and the variables derived from radar polarizations have limited capacity to accurately predict aboveground biomass, resulting in the construction of models with a low coefficient of determination. Despite these results, the potential of images ALOS-2/PALSAR-2 to estimate biomass cannot be neglected.