Maximum entropy at intermediate anthropogenic dynamics

Landscape entropy represents heterogeneity within a landscape (Renyi, 1961; Bogaert et al., 2005). Previous researches found increasing values of entropy by studying a limited number of zones increasingly affected by anthropogenic effect (O’Neill et al., 1988; Bogaert et al., 2005). The present research aims to generalise the relationship between anthropogenic effect and landscape entropy, with a further goal of linking these concepts to overall biodiversity.
16 study zones from classified LANDSAT TM scenes and Africover maps presenting
different anthropogenic effect intensities have been used (Djibu et al., 2008; Bamba et al., 2010; Barima et al., 2010; Munyemba, 2010; Vranken et al., 2011). Simpson's H diversity index based on class number has been used for entropy measurement (Renyi, 1961; Pielou, 1975; Shannon and Weaver, 1963). 1-H has been retained to get the relationship between the index and diversity directly proportional. For anthropogenic effect measurement, O’Neill’s U disturbance index, quotient between anthropogenic and natural areas, has been used (O’Neill et al., 1988). Log H has been retained to linearise the exponential relationship between the index and anthropogenic effect. Correlation between those variables has then
been analysed.
The scatter plot of the 16 study zones shows a Gaussian curve (Fig. 1) presenting maximal landscape entropy at intermediate anthropogenic effect. This distribution has been modelled by a second order polynomial regression with determination coefficient and significance.
This phenomenon has been interpreted in terms of spatial transformation processes (Bogaert et al., 2004) and linked to the habitat heterogeneity hypothesis (Tews, 2004), as well as the intermediate disturbance hypothesis (Connell, 1978; Lindenmayer and Brugman, 2005).