Post-fire succession indices performance in a Mediterranean ecosystem

Abstract

We sought to quantify early species succession (covering a 9 year period) after a natural wildfire in an unmanaged Mediterranean-type ecosystem across three different soil types. We examined different indices that are frequently used in literature as for their sensitivity and ability in tracing vegetation changes during a post-fire succession. Field data from a forest in Sithonia peninsula, northern Greece, during a period of 10 years after a fire event were used as a case study. In order to detect the best model fit of diversity and evenness indices we used pseudovalues generated with the jackknife method. The index of evenness and diversity with the lowest coefficient of variation scores were selected as the ones with the best model fit. In order to select the abundance model with the best model fit we compared the difference of the observed abundances data from the jackknife-predicted ones using the Akaike information criterion. According to our results, in rich and medium soil types, the best diversity ecosystem index from the ones examined was the (inverse) Berger–Parker one, while in poor soil types the Simpson or Brillouin diversity index. According to our results, in rich and medium soil type Mediterranean ecosystems, a good biodiversity ecosystem descriptor is based on how abundant is the most abundant species. In poor soil types where there are fewer species, and the percentage of cover is lower, indices that account for the relative representation of species within the community structure are better biodiversity ecosystem descriptors. All evenness indices examined but the Smith–Wilson were good ecosystem descriptors. The abundance model that showed the best model fit was the log series one. The poor fit of the Smith–Wilson’s index suggests that in species poor ecosystems the number of species is more important than species variance.