Scale-dependent competition at the stand level assessed from crown areas

The detection and quantification of competition at the stand level is important in forest management because competition reduces growth and increases the risk of mortality. This is of interest for timber production where efficient tools of forest inventory are increasingly demanded. Especially modern planning of thinning based on aerial or satellite images requires a deeper and spatially explicit understanding of the growth dynamics of tree crowns relative to the dynamics of stems. Past studies have evaluated competition in forests with scale-dependent correlation functions applied to tree-size attributes (continuous marks) such as diameter at breast height (DBH) or tree height. Despite the fast reaction of foliage to changes of neighborhood density in the canopy, horizontal crown extent has apparently not been used in such competition analyses of marked point patterns. Here we investigated in a spatially explicit approach the formation of crown-size patterns under neighborhood competition. We also compared how mutual growth reduction in reaction to competition differs between crown extent and stem diameters. This response of tree-size attributes to competition was analyzed with the scale-dependent mark-correlation function (MCF) applied to the marks 'DBH', 'crown area' of all live trees, and 'upper crown area' of overstory trees. These analyses were conducted for two deciduous and two coniferous forests in central Germany. Unlike stem diameters, crown areas were very sensitive for the detection of competition in stands. In relation to 'crown area' of all trees in a plot, this sensitivity was greater when only the mark 'upper crown area' of overstory trees was analyzed because both the strength and the spatial range of negative size correlation increased. Upper crown areas showed a finite range of negative interaction of about 6m. These results demonstrate that (1) the 'functional growing space' of large and light-exposed canopy trees is highly suitable to detect competition and (2) the spatial range and strength of competition can be assessed and differentiated between stands in a spatially explicit manner. Our application may have practical value for monitoring competition based on remotely sensed forest inventory because upper crown areas as seen by the 'bird-eye's view' were most sensitive for detecting competition in stands.