Characterising forest spatial structure through inhomogeneous second order characteristics

Abstract

Point process theory plays a fundamental role in the analysis and modelling of spatial forest patterns. For instance, the Ripley’s K function and its density with respect to the area, i.e. the pair correlation function, have been extensively used to analyse and characterise stationary forest configurations. However, the stationarity condition is not often met in practice when analysing real data. Thus, the development and application of new statistics to measure the degree of inhomogeneity suggests the use of inhomogeneous statistics to describe forest stands. In this paper, we restrict our attention to the inhomogeneous pair correlation function in the context of replicated spatial data. We then analyse the spatial configuration of pure and mixed conifer stands in a case study in Central Catalonia, North-East of Spain. Our results suggest that whilst P. sylvestris tend to be aggregated for short inter-tree distances, P. nigra and P. halepensis keep a minimum inter-event distance between trees. Regarding the mixed stands, trees of distinct species tend to be segregated from each other. Tentative explanations for these results are related with site properties, competition effects, shade tolerance and silviculture practices applied in this forest region.