Coenostate descriptors and spatial dependence in vegetation: derived variables in monitoring forest dynamics and assembly rules.

Long term vegetation monitoring provides valuable information on spatio-temporal patterns in plant communities that could be analysed to detect spatial relationship changes among species and to interpret dynamic tendencies and assembly rules in non-equilibrium phytocoenoses. In studies of this kind, one should take into account recent ecological theories emphasizing the scale dependence of vegetation; in particular, fine-scale spatial patterns of vegetation are important constraints in the genesis and maintenance of diversity. The information theory models of Juhász-Nagy offer an appropriate tool for describing the relationship between diversity and multispecies spatial dependence in vegetation. Diversity (florula diversity) and spatial dependence (associatum) are calculated for a series of increasing plot sizes (spatial scaling). The plot sizes at which the two coenostate descriptors reach the maximum information represent the characteristic scales that should be considered as optimal plot sizes in monitoring data collection. Moreover, this methodology enables us to study non-equilibrium dynamics and assembly rules in a more effective way. Diversity and spatial dependence are related, but the power and direction of this relationship change according to environmental characteristics, vegetation type and successional context. The demonstrated correspondence between dominant pattern-generating mechanisms and the related trajectories in abstract coenostate spaces (florula diversity and associatum maximum values), obtained by exploratory simulation studies, can improve interpretation of dynamic state and vegetation tendencies and can support a better inference about the relative role of different background mechanisms. We present some results obtained using this methodology with field data from the forest of Bialowieza National Park (Poland). In particular, we compared the herb layer spatial patterns of dynamically contiguous regeneration phases of the same phytocoenosis. Sampling was performed by recording the presence of plant species in 10 cm x 10cm contiguous microquadrats arranged in 150 m long circular transects. Field data were analysed with the same information theory methods as the ones applied to simulated data. Results show that assemblages of plant individuals are less diverse and more associated in primary than in regenerating stands, suggesting, in both situations, competitive dominance and disturbance as the main ecological mechanisms. Thus, the method was proven effective in distinguishing slightly different dynamical processes.