No abstract

Several morphological parameters, and estimates of density and evenness per depth, were analyzed for stands of the subtidal kelp Lessonia trabeculata Villouta et Santelices from 2 zones, central and northern Chile. Two sites in each zone were sampled for 2 yr, and variability patterns were statistically described using site or zone as classification factors. Bathymetric profiles of density showed a strong qualitative and quantitat~ve variation among sites, contrasting with the high evenness in spatial distribution of plants per depth interval, despite intervals of varylng length. Relationships among morphometric parameters changed markedly across spatial scales, producing a high variance in biomass predictions. In most cases, strong site-and zone-specific components were identified, which account for the variability in morphology at both spatial scales. Zone effects may mirror water motion and temperature regimes, among other factors, and site effects may be related to specific habitat configurations. These results show that extrapolation from structural features of single stands for characterizing species may result In misleading inferences. The detection of scale-dependent patterns proved to be a useful starting point for comparative studies of population structure, and necessary before attempting generalizations.

Although networks analysis has moved from static to dynamic, ecological networks are still analyzed as time-aggregated units where time-specific interactions are aggregated into one single network. As a result, several questions arise such as what is the functional form of and how variable is the topology of time-specific versus time-aggregated ecological networks? Furthermore, it is yet unknown to what extent the structure of time-aggregated networks is representative of the dynamics of the community. Here, we compared the topology of time-specific and time-aggregated networks by analyzing a set of intertidal networks containing more than 1,000 interactions, and assessed the spatiotemporal dynamics of their degree distributions. By fitting different distribution models, we found that the out-degree distributions of seasonal and time-aggregated networks were best described by an exponential model while the in-degree distributions were best described by a discrete generalized beta model. The degree distributions of the seasonal networks were highly temporally variable and are significantly different from those of time-aggregated networks. We observed that seasonal degree distributions converged toward time-aggregated network distributions after 1.5 years of sampling. Our results highlight the importance of understanding the dynamics of ecological networks, which can show topological characteristics significantly different from those of time-aggregated networks.