Differential colonization causing non-random forest plant community structure in a fragmented agricultural landscape

We investigated forest plant community structure of recent forest patches in a highly fragmented agricultural landscape. A nestedness analysis has been performed at the community level and at the individual species level. We tested the hypothesis that plant species composition showed a nested structure and whether this was generated by isolation and differential colonization. Alternative hypotheses formulated in the past, such as nested habitats and patch area dependent species relaxation, are verified. Isolation measures were calculated between clearly defined source and target patches. At the community level target patches are defined as forest patches < 35 yr old while older patches are considered as source patches. At the individual species level we distinguished between actually occupied and potentially suitable target patches based on a habitat space model and only occupied source patches are taken into account. All results point out that the nested community pattern in recent forest patches is generated primarily by isolation and differential colonization. These findings are confirmed at the species level, since the degree of nestedness of the individual species is highly correlated with their isolation sensitivity. Logistic regression analyses prove that many forest plant species are sensitive to isolation. The probability of occurrence of the isolation sensitive species drops almost to zero, when the nearest source patch is situated further than 200 m. Probabilities of colonization increase significantly when source patches are situated adjacent to target patches.