Abundance and nestedness in interaction networks

Nestedness is a useful metric that characterizes the generalist–specialist balance in ecological communities. Although several nestedness indices have been proposed, few have explored how species abundance per se affects their performance and the ability to detect true interaction networks. We here develop a mathematical framework that takes into account abundance in estimates of nestedness. We use an analytical approach to relate abundance and nestedness. In our null model the probability of interaction among species is determined solely as function of their abundances. Assuming a power-law abundance model we analytically find the nestedness index and its coefficient of variability. We find that the sloping abundance distribution of our null model generates more nested structures. On the other hand steeper abundances lead to higher coefficients of variability. Both results suggest that nestedness analysis should be evaluated and explanations sought carefully.