Vegetation dynamics and relation with occurrence of fires in the Cerrado Biome
Name: FELIPE GIMENES RODRIGUES SILVA
Publication date: 21/07/2017
Advisor:
Name![]() |
Role |
---|---|
ALEXANDRE ROSA DOS SANTOS | Advisor * |
NILTON CESAR FIEDLER | Co-advisor * |
Examining board:
Name![]() |
Role |
---|---|
ALEXANDRE ROSA DOS SANTOS | Advisor * |
NILTON CESAR FIEDLER | Co advisor * |
RODRIGO SOBREIRA ALEXANDRE | Internal Examiner * |
SAMUEL FERREIRA DA SILVA | External Examiner * |
Summary: SILVA, Felipe Gimenes Rodrigues. Vegetation dynamics and relation with occurrence of fires in the Cerrado Biome. 2017. Dissertation (Masters degree on Forest Science) Federal University of Espírito Santo, Jerônimo Monteiro ES. Advisor: Prof. Dr. Alexandre Rosa dos Santos. Co-advisor: Prof. Dr. Nilton Cesar Fielder.
The Cerrado Biome is threatened by the advance of deforestation and the indiscriminate use of fire. In this scenario, the study of indigenous lands through information obtained from remote sensing helps to evaluate the vegetation and its relation to the occurrence of fires. In this context, the objective was to evaluate the temporal dynamics of the vegetation, related to the precipitation and the occurrence of forest fires in indigenous lands, located in the Cerrado Biome in the state of Mato Grosso, Brazil. Thus, data from the vegetation indices obtained by the Normalized Difference Vegetation Index (NDVI) and generated from the product of the MOD13Q1 provided by the MODIS sensor, instrument of the satellite TERRA. For the analysis of Burn Area throughout the series, the images came from the MCD45A1 product of the same sensor. The images were acquired from 2007 to 2016, and lastly, rainfall images from the Tropical Rainfall Measuring Mission (TRMM) sensor. With the images of NDVI, TRMM and Burn Area, they were standardized in the ArcGis program so that they were worked in the TerrSet software program, the Earth Trends Modeler module was used, making correlation analysis between elements of precipitation, vegetation index and (3, -2, -1, 0, 1, 2, 3), in order to obtain the response time of the dependent variables in relation to the independent variables. With the images of NDVI, TRMM and Burn Area, they were standardized in the ArcGis program so that they were worked on the TerrSet software, using the Earth Trends Modeler module, doing the correlation analysis between elements of precipitation, vegetation index and area (-3, -2, -1, 0, 1, 2, 3), in order to obtain the response time of the dependent variables in relation to the independent ones. In relation to the analyzes of the interannual trends of the time series of vegetation indices, these were carried out using Mann-Kendall monotonic trend methodologies and analyzes of seasonal trends. The images were imported for the TerrSet software, in which the Earth Trends Modeler (ETM) module was also used to process and analyze the vegetation index behavior trends. For the analysis of anomalies of vegetation index, precipitation and burning area, no anomalous periods were found. The correlations obtained were more significant in open field regions with R of 0.84 for NDVI and Precipitation and for NDVI and R burning area of - 0.74. The Monotonic Trend of Mann-Kendall, denote a certain stability of the vegetation index with positive variations in open field regions and in only one place (Areões village) with reduced vegetative vigor. The analysis of seasonal trends identified different vegetation responses, since this biome presents a diversified phytophysiognomy showing the seasonality of vegetation with different phases in amplitudes. In all years, large areas of burning were quantified, totaling approximately one million hectares per year; The variation of the trends, evidences the various phytophysiognomies and their answers regarding the gain and loss of biomass. The best correlation and regression response of NDVI x precipitation was observed according to the type of open field vegetation. The Area of burning tends to increase with the reduction of NDVI, mainly in regions with more open vegetations. Finally, in relation to seasonality, a defined pattern of cycles and vegetation phases was not found.
Keywords: Remote sensing, Precipitation, Forest Fires, Vegetation Index.