Tuesday, August 3, 2010 - 10:10 AM

COS 23-7: CANCELLATION - Decoupling fragmentation per-se effects from habitat loss reveals positive effect of fragmentation on spider biodiversity

Yoni Gavish, Ben-Gurion University of the Negev and Michael L. Rosenzweig, University of Arizona.

Background/Question/Methods

After removing the effect of area loss, fragmentation has been shown to increase, decrease or not affect biodiversity. We explored the effect of fragmentation on biodiversity of spiders in the fragmented landscapes of Southern Judea Lowlands (SJL), Israel. SJL lies in the transition zone between the Mediterranean ecosystem in the north, and the arid Negev desert in the south. Over a gradient of 30 km, mean annual precipitation drops from 450 to 250 mm/y. At the landscape scale, SJL is an agro-ecosystem with small, more natural patches of various areas embedded inside the agricultural matrix. We used a data-set comprising more than 15,000 spider individuals from 200 species or morpho-species. The spiders were sampled in three 4×3.2 km landscapes along the sharp climatic gradient. In each landscape, 12 patches of different areas and a continuous area were sampled. We used two methods to decouple the effect of fragmentation from the habitat loss effect. First, for each landscape, we estimated the number of species in each patch using accumulation curves. We conducted a linear regression of log(species) against log(area) and extrapolated the SPAR to the combined total area of all 12 patches. We estimated the actual number of species in all 12 patches combined and determined whether the value lay above, below or on the line of the extrapolation. Second, we selected all combinations of patches whose total area lay in a small range (±5%) around the area of a focal patch. We did it separately for different fragmentation levels (number of patches whose area is combined), ranging from two to nine patches. We calculated the mean Fisher's alpha for different fragmentation levels and compared the value to the Fisher's alpha of the focal patch.

Results/Conclusions

We found that in all three landscapes, the estimated number of species in all 12 patches lay above the expected species diversity of the SPAR. We also found that in SJL, Fisher's alpha increased with fragmentation level. We suggest that in SJL, the species turnover between patches overcompensates for the reduced number of species in each patch due to area loss.