Análise de lacunaridade de fenômenos agroambientais

Abstract

The phenomena studied in agricultural and environmental sciences are characterized by their natural evolution and are also affected by human activity. They are comprised of a large number of components with non linear interactions on some scales, which produce emergent properties on other scales. These are characteristics of complex systems and can be studied using various concepts developed over the last decades such as fractals, multifractals, self-organized criticality and entropy. In this paper we analyzed the temporal dynamics of climate variables rainfall, temperature and stream flow, and spatial distribution of vegetation fires represented by hot pixels detected by satellites. We used the method of lacunarity that serves to describe the distribution of gaps in a data set. Using lacunarity values for different time scales it was possible to describe the rainfall regime (fragmentation of rainy periods) and classify two precipitation patterns in the state of Sergipe, one formed by stations approaching the Sergipe coast, and other formed by stations that are located in semiarid region. The lacunarity analysis was successful for evaluation of impact of human activity on natural regime of river flow, in the case of rivers Atibaia and Jaguari, caused by the construction of Cantareira system reservoirs. After the construction of the reservoirs lacunarity values of stream flow decreased, indicating a reduction of heterogeneity of flow dynamics due to regulation of water flow. The results of lacunarity analysis of time series of temperature and precipitation in the state of Pernambuco showed that this method can be used to identify the geographic regions with characteristic temporal behavior of climatic variables. The lacunarity analysis for two-dimensional data was applied to spatial distribution of hot pixels detected in Amazonia during the period 2000-2013. In the years with high number of hot pixels the lacunarity values were lower, indicating a more homogeneous spatial distribution compared with periods with low number of hot pixels. The lacunarity method has also shown to be effective in identifying the dry and rainy seasons, by capturing the heterogeneity of spatial distribution of hot pixels. These results may be useful in planning the use of water resources, planning of agricultural activities, developing strategies to reduce the impact of extreme weather events (drought, flood) and natural disasters such as forest fires.