FUZZY LOGIC APPLICATION IN IDENTIFICATION OF FOREST FRAGMENTS WITH POTENTIAL FOR CONSERVATION
Name: CARLOS ROBERTO LIMA THIAGO
Publication date: 30/04/2015
Advisor:
Name |
Role |
|---|---|
| ALEXANDRE ROSA DOS SANTOS | Advisor * |
| JOSÉ EDUARDO MACEDO PEZZOPANE | Co-advisor * |
Examining board:
| Name |
Role |
|---|---|
| ALEXANDRE ROSA DOS SANTOS | Advisor * |
| AUREO BANHOS DOS SANTOS | External Examiner * |
| JÉFERSON LUIZ FERRARI | External Examiner * |
| JOSÉ EDUARDO MACEDO PEZZOPANE | Co advisor * |
| SIDNEY SARA ZANETTI | Internal Alternate * |
Summary: THIAGO, Carlos Roberto Lima. FUZZY LOGIC APPLICATION IN IDENTIFICATION OF FOREST FRAGMENTS WITH POTENTIAL FOR CONSERVATION.Dissertation (Master of Forest Science) - Universidade Federal do Espírito Santo, Jerônimo Monteiro, ES. Advisor: Prof. Dr. Alexandre Rosa dos Santos. Co-advisor: Prof. Dr. José Eduardo Macedo Pezzopane.
The Atlantic Forest biome is notable for its biodiversity, and also as one of the most threatened biomes on the planet. The state of Espírito Santo, which is located in its entirety under the control of the Atlantic Forest biome, has about 12.2% of remaining forests. Forest fragmentation is driven by the uncontrolled activity of land use and land occupation and population growth. The most serious aspects of forest fragmentation process are the edge effect and the loss of biodiversity. The ecological corridors are an alternative to mitigate the effects of forest fragmentation between them allowing gene flow of fauna and flora and the recolonization of degraded areas. In order to join efforts to the protection and management of this ecosystem, this study aimed to show that the Fuzzy logic can be used to identify forest fragments that suggest greater conservation from the perspective of landscape ecology metrics and propose ecological corridors in the catchment area of the river Itapemirim within state boundaries of the Espírito Santo. The mapping of forest fragments was carried out by means of remote sensing techniques using Landsat 8 sensor, available free of charge through the website of the institution United States Geological Survey (USGS) using supervised classification techniques. To calculate the indexes of ecology, the computer application ArcGis 10.2 was used, by extending V-LATE 2.0 public domain. After mapping of fragments verified the accuracy through Kappa (82.03%) and Global accuracy (91.90%), thus validating classification. 11,749 were identified fragments, representing 22% of forest cover in the study area. The fragments were mapped divided into size classes. The fragments (< 5 ha) were found in greater numbers, 8.394, followed by fragments (5.-10 ha), 1.581 (10 - 50 ha), 1.414, (50 -100 ha), 204 (100-300 ha), 105 (> 300 ha), 51. The number of fragments showed an inverse relationship with its contribution in the area. The evaluation of metric indices of forest landscape was employed with area groups of indexes; density and size, shape; proximity and central area, the latter being obtained for different edge effect simulation (40, 60, 100, 120, 140 and 200 m). All indexes showed differences when observing the fragments of size classes, signaling that the fragments of the size classes (100 - 300 ha) and (> 300 ha) have metrics with spatial attributes that suggest a greater degree of conservation. To calculate the spatial potential of forest fragments was used fuzzy logic. The membership functions were employed Fuzzy Small, for the variables, average distance from the nearest neighbor, shape index and fractal dimension, and the membership function Fuzzy Large, for the variables, central area, nearest neighbor area and area fragment. The variables were combined by means of Fuzzy Gamma operator to represent the potential space fragments. Selected and identified 10 fragments with the greatest potential. Then, using the lowest cost approach distance of the potential fragments were connected using fragments of 196 and 95 corridors. Runners totaled 70,879.65 m in length, with average length of 746.10 m, with an average width of 74.61 m. 206 fragments were connected, 24.72% of the total area of the remaining fragments in the catchment area of the river Itapemirim. Factors considered in this study as high cost, or impedance to implementation of the corridor, represented 1.71% of the total connected area.
Keywords: Geographic information systems, remote sensing, forest fragmentation, landscape ecology indexes, multi-criteria analysis.
