Land change and loss of landscape diversity at the Caatinga phytogeographical domain - Analysis of pattern-process relationships with MODIS land cover products (2001-2012)

Tropical dry forests are increasingly threatened by land degradation. For the Caatinga seasonally dry forests in Northeast Brazil, recent studies revealed both woody vegetation loss and (re-)gain. A conclusive analysis of land cover and landscape diversity changes is still lacking. To fill this gap, we surveyed pattern-process relationships for the Caatinga phytogeographical domain on different spatial levels for the years 2001-2012 from MODIS land cover data. By a landscape pattern analysis, a set of five landscape metrics was calculated using FRAGSTATS. The spatio-temporal results revealed a general decrease of fragmentation and diversity. The savanna class which comprises also Caatinga dry forests covered large parts of the landscape and gained considerably, whereas other vegetation classes decreased. The distribution of the low vegetated shrublands and grasslands implied degradation risks, in particular at the Sao Fransisco River valley. Large-scaled agriculture was found in the proximity to water reservoirs in the lowlands, which led to local increases of the landscape's diversity. We quantified both gains and losses of woody vegetation with a notable net gain of 6.7% of the study area. Main changes may be attributed to socio-economic shifts from local to large-scaled agriculture and water infrastructural activities. Overall, our study shows that a multi-temporal, global dataset along with a landscape pattern analysis provides valuable contributions to landscape research and for detecting land cover and landscape diversity changes. (C) 2016 Elsevier Ltd. All rights reserved.