Fuzzy logic in the determination of forest fragments for seed collection
Name: TELMA MACHADO DE OLIVEIRA PELUZIO
Type: PhD thesis
Publication date: 17/02/2017
Advisor:
Name |
Role |
|---|---|
| ALEXANDRE ROSA DOS SANTOS | Advisor * |
| NILTON CESAR FIEDLER | Co-advisor * |
| SUSTANIS HORN KUNZ | Co-advisor * |
Examining board:
| Name |
Role |
|---|---|
| ALEXANDRE ROSA DOS SANTOS | Advisor * |
| KARLA MARIA PEDRA DE ABREU | External Examiner * |
| NILTON CESAR FIEDLER | Co advisor * |
| SUSTANIS HORN KUNZ | Co advisor * |
Summary: The Brazilian rain forest is fragmented due to the great exploitation in the process of colonization of the country. For the maintenance and protection of the different ecosystems, a series of legal instruments were adopted, which, when associated to the technological and mathematical tools, especially the Fuzzy logic, allow the adoption of criteria closer to human thought. In the present work, the objective was to select potential forest fragments with a higher degree of conservation for the collection of seeds that meet legal requirements, through the use and association of landscape ecology with Fuzzy logic. The study was carried out in the Itapemirim river basin in the state of Espírito Santo, through the following steps: obtaining the Landsat 8 satellite image; photointerpretation and classification of forest fragments by size class in which: A less than 5 ha, B between 5.1 and 50 ha, C between 50.1 and 300 ha, and D greater than 300 ha; evaluation of the dynamics of the metrics indexes of the area landscape, density and size, shape, border, central area and proximity; application of the Fuzzy logic and their respective Small and Large membership functions, and the Gamma overlay function; selection of potential forest fragments with higher degree of conservation of the seeds for collection; Shannon Wiener diversity, Pielou equability, Bray-Curtis dissimilarity, quality and stem health according to the Brazilian Forestry Society standard, and the Payandeh index to evaluate the pattern of spatial distribution among forest fragments classified as high and low potential for seed collection. 7,515 forest fragments were determined, occupying 19.21% of the study area. Size class A has a smaller territorial area, greater number of fragments and extinction risk due to increased edge. Class D has lower number of fragments, larger area and better condition for seed collection, even with increased edge. Two fragments classified as high (Fragment 1) and low (Fragment 2) were selected, respectively, for the collection of forest seeds. In fragment 1; 1,670 individuals were sampled, corresponding to 172 species, 103 genera and 40 botanical families. In fragment 2; 1,526 individuals, 135 species, 92
genera and 40 botanical families were sampled. Fragment 1 presented lower values for the diversity indexes of Shannon-Wiener, Pielou and Jackknife; Higher basal area and number of individuals per hectare, better quality and stem health among tree individuals and better spatial distribution than fragment 2. Field data analysis allows to confirm that Fuzzy logic was effective in determining potential fragments for collection of forest seeds with higher degree of conservation, based on the determination of the richness of the fragments. Fragment 1 has greater richness, higher basal area, lower diversity, better quality and health of the stem, and better spatial distribution than fragment 2. The methodology can be adapted to other zones and different biomes of the planet.
Keywords: Degraded areas, Nature conservation, Environmental geotechnology, Atlantic forest, Geographic Information Systems.
