Remote sensing assessment of spectral characteristics of mangrove forests in Cuba: a methodological approach

Abstract

The uses of remote sensing information for study of Cuban mangroves have been limited. Generalized methods focused on field data provide local perspectives with restricted temporal amplitude, not enough to fully describe forest distribute over thousands of squared kilometers of coastal zones with high spatial and temporal variability. In current paper we describe and apply the method to assess spectral characteristics and describe it spatial variability of Cuban mangrove forest in Cuba, from Landsat 8 imagery in 2017, as baseline for future studies. With processed imagery we create nation-wide mosaic for ten spectral vegetation indexes. Spatial variability was statistically sampled from 11 584 points to characterize values distribution among regions, coastal zones and main wetland systems of Cuba. Indexes were correlated among them, with tree cover and with distances to potential drivers such as sea line, water bodies or rivers and human populations. The two first principal components explained 80% of variance and lead to detection of global differences in scores distribution among regions. Mangrove forest of the four main wetland systems of the country showed specific spectral response patterns according to spectral indexes. Applications of this type of data in the study and monitoring of this important Cuban ecosystem were discussed and its potential described.