, Volume 20, Issue 8, pp 1677-1694
Date: 27 Apr 2011

Geostatistical modelling of regional bird species richness: exploring environmental proxies for conservation purpose

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access


Identifying spatial patterns in species diversity represents an essential task to be accounted for when establishing conservation strategies or monitoring programs. Predicting patterns of species richness by a model-based approach has recently been recognised as a significant component of conservation planning. Finding those environmental predictors which are related to these patterns is crucial since they may represent surrogates of biodiversity, indicating in a fast and cheap way the spatial location of biodiversity hotspots and, consequently, where conservation efforts should be addressed. Predictive models based on classical multiple linear regression or generalised linear models crowded the recent ecological literature. However, very often, problems related with spatial autocorrelation in observed data were not adequately considered. Here, a spatially-explicit data-set on birds presence and distribution across the whole Tuscany region was analysed. Species richness was calculated within 1 × 1 km grid cells and 10 environmental predictors (e.g. altitude, habitat diversity and satellite-derived landscape heterogeneity indices) were included in the analysis. Integrating spatial components of variation with predictive ecological factors, i.e. using geostatistical models, a general model of bird species richness was developed and used to obtain predictive regional maps of bird diversity hotspots. A meaningful subset of environmental predictors, namely habitat productivity, habitat heterogeneity, combined with topographic and geographic information, were included in the final geostatistical model. Conservation strategies based on the predicted hotspots as well as directions for increasing sampling effort efficiency could be extrapolated by the proposed model.