Use of topographic variability for assessing plant diversity in agricultural landscapes

The relationship between plant diversity and topographic variability in agricultural landscapes was investigated, with the aim of determining whether sampling landscape units of 1 km(2) (LUs) across a gradient of topographic variability is more efficient than a random design for assessing the range of biodiversity in climatically and biogeographically homogenous areas called sub-regions. Representative plant species data from the Swiss biodiversity monitoring programme were analyzed covering a broad environmental gradient of four altitudinal belts and seven biogeographic regions. The focus of the study laid on agricultural areas but the whole dataset was as well analyzed to put the agricultural LUs in a general context. Plant species lists of LUs were used to calculate two diversity components: ECOrichness, the number of ecological plant types per LU (as a measure of beta diversity) and AGROrichness, the number of species of conservation importance for agriculture. Mixed regression models were used to analyse the effects of topographic variability on the two plant diversity components, including sub-regions (areas with the same combination of altitudinal belt and biogeographic region) as random factor. These analyses were performed for the whole dataset (419 LUs within 22 sub-regions) and for the focal subset of 187 agricultural LUs within 13 sub-regions. ECOrichness increased significantly with topographic variability for both the general and the agricultural dataset. The partial correlations within the sub-regions revealed consistent trends for the agricultural LUs but some inconsistencies for the whole dataset. For the monitoring of agricultural LUs the sampling along a gradient of topographic variability is therefore suggested as an efficient means for assessing the range of plant species diversity within sub-regions. Compared to other measures of landscape heterogeneity like habitat heterogeneity, sampling LUs along topographic variability is cheap and easily applied and it was demonstrated to work over large environmental gradients. (C) 2011 Elsevier BM. All rights reserved.