Estimating impacts of plantation forestry on plant biodiversity in southern Chile-a spatially explicit modelling approach

摘要

Monitoring the impacts of land-use practices is of particular importance with regard to biodiversity hotspots in developing countries. Here, conserving the high level of unique biodiversity is challenged by limited possibilities for data collection on site. Especially for such scenarios, assisting biodiversity assessments by remote sensing has proven useful. Remote sensing techniques can be applied to interpolate between biodiversity assessments taken in situ. Through this approach, estimates of biodiversity for entire landscapes can be produced, relating land-use intensity to biodiversity conditions. Such maps are a valuable basis for developing biodiversity conservation plans. Several approaches have been published so far to interpolate local biodiversity assessments in remote sensing data. In the following, a new approach is proposed. Instead of inferring biodiversity using environmental variables or the variability of spectral values, a hypothesis-based approach is applied. Empirical knowledge about biodiversity in relation to land-use is formalized and applied as ascription rules for image data. The method is exemplified for a large study site (over 67,000 km(2)) in central Chile, where forest industry heavily impacts plant diversity. The proposed approach yields a coefficient of correlation of 0.73 and produces a convincing estimate of regional biodiversity. The framework is broad enough to be applied to other study sites.