Species-driven phases and increasing structure in early-successional plant communities.

Successional phases describe changes in ecological communities that proceed in steps rather than continuously. Despite their importance for the understanding of ecosystem development, there still exists no reliable definition of phases and no quantitative measure of phase transitions. In order to obtain these data, we investigated primary succession in an artificial catchment (6 ha) in eastern Germany over a period of 6 years. The data set consists of records of plant species and their cover values, and initial substrate properties, both from plots in a regular grid (20 m × 20 m) suitable for spatial data analysis. Community assembly was studied by analyses of species co-occurrence and nestedness. Additionally, we correlated lognormal and log series distributions of species abundance to each community. We here introduce a new general method for detection of successional phases based on the degree of transient spatial homogeneity in the study system. Spatially coherent vegetation patterns revealed nonoverlapping partitions within this sequence of primary succession and were characterized as two distinct ecological phases. Patterns of species co-occurrence were increasingly less random, and hence the importance of demographic stochasticity and neutral community assembly decreased during the study period. Our findings highlight the spatial dimension of successional phases and quantify the degree of change between these steps. They are an element for advancing a more reliable terminology of ecological successions.