Wednesday, August 4, 2010

PS 54-78: Effects of fragment metrics on woody species diversity in semiarid-oak forests of the Upper Mixtec region, Oaxaca, Mexico

Ma. de los Remedios Aguilar S and Rafael F. del Castillo. Centro Interdisciplinario de Investigación para el Desarrollo Integral Regional_U. Oaxaca, Instituto Politécnico Nacional

Background/Question/Methods The upper Mixtec Region in Oaxaca, southern Mexico, has one of highest deforestation rates in the world, resulting in highly fragmented landscape with 200 - 280 Mg h y-1 of soil lost by erosion.  Nevertheless, the effects of fragmentation on forest richness and structure are unknown.  We analyzed the relationship between fragment metrics and diversity in semiarid-oak forests in this region.  Twelve fragments distributed in three municipalities were selected using 2005 Spot satellite images.  In each fragment we established an average of 36 randomly distributed sampling plots, 18 of them (102.7 m2 each) were for tree sampling, and 18 (12.56 m2 each) were for shrubs.  A total area of 2.5 ha (532 sampling points) was surveyed.  All woody plants ≥ 2.5 cm dbh and ≥ 2 m height were included in tree sampling, whereas those < 2.5 dbh or < 2 m height were included in the shrub sampling plots.  Species accumulation curves randomized 50 times and diversity indices were calculated with EstimateS 7.5.2 software.  The relationship between diversity indices (Jack2, α-Fisher, Shannon-Wiener, and Simpson) with fragment metrics (fragment area, perimeter, and perimeter/area ratio) were analyzed using general linear mixed effects models.

Results/Conclusions A total of 11,029 plants distributed in 155 species (52 trees, 103 shrubs) were recorded.  Fragment metrics ranged from 12 to 3,206 ha (area); 5–328 Km (perimeter); and 70–385 m ha-1 (perimeter/area ratio).  In general, the same relationship with fragment metrics and diversity indices were detected for both trees and shrubs.  This relationship was positive logarithmic for fragment area and perimeter, and negative logarithmic for perimeter area ratio.  Shannon and Simpson indices were less sensitive to fragment metric changes than Jack2 and Fisher alpha, in particular for shrubs.  Overall, small changes in diversity were observed for larger fragments, but in fragments under 40 ha or 30 km perimeter, small fragment reductions resulted in high tree diversity declinations.  Shrubs required higher reductions in fragment size than trees before drastic changes in diversity could be detected.  These results can be used as guides for oak forest management and conservation in the Upper Mixtec region.