Variações da reflectância e dos índices de vegetação em função dos parâmetros da modelagem topográfica no Parque Estadual do Turvo, RS

Abstract

The process of land occupation and the agricultural expansion in the Rio Grande do Sul State contributed to the reduction of native forest cover as the deciduous subtropical forest. Initially this forest covered most of the northwest of the State and today it is restricted to the Parque Estadual do Turvo (PET) and the Terra Indígena de Guarita. Located in the Derrubadas municipality, the PET is the largest preserved area of deciduous subtropical forest of the State, with great environmental importance and an area for conduction of forest studies based on the use of remote sensing techniques as the vegetation indices. These indices are affected by external components that change the spectral responses of the imaged targets. Among the components that can cause this problem type, the relief can influence on the view-illumination conditions. In this sense, the objective of this study was evaluating the magnitude of the bidirectional reflectance variations and of the vegetation indices as a function of local topographical variations, with the use of high spatial resolution data of the Rapideye sensor. The applied methodology included the acquisition of MOD13Q1 product of the Moderate-resolution Imaging Espectroradiometer (MODIS) sensor, a digital elevation model of Advanced Spaceborne Thermal Emission and Reflection Radiometer - Global Digital Elevation Map version 2 (ASTER GDEM v2) and two scenes of Rapideye sensors. The MODIS time series allowed a temporal evaluation of forest dynamics and allowed to identify the more appropriate dates for the acquisition of Rapideye scenes. Rapideye data were acquired at June 28, 2012 and October 17, 2012. The next step was to perfor the topographic modeling in order to extract the slope, aspect and altimetry. Based on the topographic variables and the Rapideye metadata the shaded relief was calculated for both scenes. By slicing topographic variables, the Rapideye per band reflectance and vegetation indices were evaluated for each class. Based on the Rapideye data, four vegetatioin indices were calculated: Normalized Diference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Red Edge Normalized Difference Vegetation Index (RENDVI) and the Red-edge Chlorophyll Index (CIRed-Edge). Randomly samples were averaged and analyzed through graphs, descriptive statistics and its relationship with forest survey in sevem plots (20x100 m) was performed. The results of field inventory have identified 62 dominant species distributed in three strata and showed the presence of discontinuities within the forest. The vegetation indices have presented the highest values in summer and lower in winter, in agreement to the deciduous behavior of the emergent strata of the forest. The obtained results showed that beyond the seasonal phenological variations, the local illumination conditions associated to PET relief, contribute to explain the spectral response of the the forest. The most important topographic variables that affect the spectral response of Rapideye band were the aspect and slope. Despite both the reflectance data and the vegetation indices were conditioned by the illumination conditions in the two dates, the June image was most affect (more shaded relief due to highest solar zenith angle). The bands that showed higher dependence on slope, aspect and illumination were the red and near infrared. The increased of the altimetry caused a positive alteration in reflectance of the near infrared and the red edge bands and a decrease in reflectance of the visible bands. Vegetation indices that have normalization factors, such as the NDVI and the RENDVI presented lower dependence on the topographical conditions, while the indices as EVI and CIred-edge showed higher sensitivity to illumination conditions and shading.