Spatial analysis techniques applied in forest ecology: point pattern analyses

Abstract

Several methods for analyzing uni- and bivariate point patterns, commonly used in ecology, are reviewed in this work. The refined nearest neighbor method, the Ripley’s K function and the SADIE technique were used to analyze simulated and real spatial point patterns. A random pattern, three clumped patterns with radius of 3, 5 and 10 m, and three random patterns with inhibition distances of 2, 4 and 6 m, were simulated to check the efficiency of the methods. Real patterns were measured in two deciduous forests in the Cantabrian lowlands and a coniferous forest in the Cebollera-Urbión Range. All techniques were able to distinguish between random, clumped and regular patterns. Nearest neighbor method and K function displayed inhibition distances in regular patterns, but K function alone detected the clump size. Juveniles showed clumped spatial patterns, while adults displayed a random distribution. For bivariate analyses, some simulated patterns displayed significant evidence of spatial attraction or repulsion. Real patterns indicated spatial repulsion between juveniles and adults, and attraction between differently sized juveniles. The results were interpreted according to the available information on forest ecology. The use of appropriate methods for edge effect correction, and null models alternative to the complete spatial randomness, is also discussed.