T-Square Sampling of Spatial Tree Diversity Indices

Tree diversity indices play a crucial role in monitoring biodiversity in forest stands which is even more important with ongoing climate change. Distance sampling methods are ideal for spatial diversity indices based on the nearest neighbour principle. For this purpose, the tree nearest to the nodes of a systematic sampling grid are chosen and the k nearest neighbours of this tree are determined to quantify spatial tree diversity indices. This conventional variant of the distance method, however, can be biased, because the selection of trees closest to a sample point usually favours isolated individuals. A possibility to correct this bias is T-square sampling (Besag and Gleaves, 1973). As part of this method the tree nearest to a sample point is only selected, if the angle formed by the sample point, the first nearest tree and the tree nearest to the first nearest tree is larger than 90^o. This method has been widely used in ecology but never with spatial diversity indices. Using sampling simulations, RMSE and bias of T-square and conventional distance sampling are compared.

Literature

Besag, J. and Gleaves, J. T., 1973. On the detection of spatial pattern in plant communities. Bulletin of the International Statistical Institute 45, 153-158.

Krebs, C. J., 1999. Ecological methodology. 2^nd edition. Addison Wesley Longman, Menlo Park.

Contact: Prof. Arne Pommerening (arne.pommerening@slu.se)

Tsquare.pdf