Publications

da Silva, RG; dos Santos, AR; Peluzio, JBE; Fiedler, NC; Juvanhol, RS; de Souza, KB; Branco, ERF (2021). Vegetation trends in a protected area of the Brazilian Atlantic forest. ECOLOGICAL ENGINEERING, 162, 106180.

Abstract
Among worldwide tropical forest ecosystems, the Atlantic forest of Brazil stands out as a highly fragile area and a biodiversity hotspot, with a high degree of endemism. The Capara?o National Park is an important conservation area for this biome, where recent studies have found new species of animals and plants, including endemic and highly vulnerable ones. We analyzed interannual and seasonal trends in vegetation for Capara?o National Park and its buffer zone, using data from MODIS vegetation indices between the years 2001 and 2019. Our methods included both parametric (linear trend and linear correlation) and non-parametric (Mann-Kendall) analyses of interannual trends, wavelet transform, and seasonal trends analysis. We also studied the correlation of the vegetation index data with the variability of monthly rainfall data from the TRMM satellite. We generated linear regression analyses, with precipitation as the independent variable and Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) as the dependent variable, and repeated this procedure for time lags from 0 to 3 months. In accordance with the findings of the temporal profiles and Mann Kendall trend method identified a prevalence of decreasing trend in the values of vegetation indices in Capara?o National Park, with values higher than 90% of the regions where the result was significant, for the EVI and 80% for the NDVI. For the buffer zone, there is an increasing trend for the NDVI and EVI, with values higher than 50% of the regions where the result was significant. In most of the area, a positive slope was found in the buffer zone, while in the Caparao? National Park the slope was predominantly negative. The r values indicated tendencies predominantly positive for the buffer zone and predominantly negative for the area within the park. 66.9% of the results were positive trends for ndvi in the buffer zone, and 51.7% for EVI, both for r value and slope. In the park area, this trend was negative in 68.9% of the area for ndvi in both methods; 93.8% for the EVI and 93.9% for the NDVI. These differences may be evidence of how vegetation responds to disturbances, once the vegetation in the buffer zone is fragmented and does not have the same protection as vegetation in the park area. The native vegetation in study area shows seasonal similarities, as detected by seasonal trend analysis and wavelets transform. The existence of two annual cycles is a common feature to the biome of Atlantic forest. R values for the correlation of precipitation with vegetation indices indicate high dependence of vegetation on precipitation in the study area. Values in the 0.4 to 0.8 class, represented almost 80% of the buffer zone for the NDVI and 90% for the EVI. For Capara?o National Park, this class accounted for 27% and 38% of the area for the NDVI and EVI, respectively. Vegetation in the protected area was less dependent on rainfall, which can be attributed to the fact that it can better meet its water needs through available water in the soil. In addition, due to anthropogenic pressures and a high degree of forest fragmentation, vegetation in the buffer zone may have less advanced successional stages. Our results collaborate with the range of vegetation studies in the Brazilian Atlantic forest. This methodology can be applied to larger areas around the world to provide information on vegetation trends conditions and influence of climatic variables.

DOI:
10.1016/j.ecoleng.2021.106180

ISSN:
0925-8574