Monitoring spatial tree diversity indices using relascope sampling: Estimators, interactions and limitations

Abstract

With ongoing climate change the monitoring of tree diversity has become very important for avoiding or at least decelerating the loss of biodiversity and for maintaining forest ecosystem resilience. As part of such monitoring, spatial indices of tree diversity which are calculated for individual trees often serve as surrogates for more direct measures of biodiversity. Mainly for its efficiency and ease of application, relascope sampling is a widespread method applied in forest inventory of many countries and thus has often been suggested as a data source for the monitoring of tree diversity. Since the interaction between sampling design and spatial diversity indices is not always clear to data analysts, we reviewed existing estimators and experimentally examined a new one, conducted extensive sampling simulations using different indices and estimators and additionally analysed the data from a large-scale forest inventory in Austria. We found that both forest structure and index algorithm greatly influence the sampling error. The largest source of sampling error was the index variance and contrary to our expectation not so much the bias due to spatial effects. For diversity indices related to distances, it has turned out to be best to apply estimators that include spatial edge correction methods. For all other indices an estimator performed better that included information on both the sample trees and their nearest neighbours, as it much reduced overall index variance. However, if possible the plus-sampling edge correction method should be applied.