Mixed-effects models for estimating stand volume by means of small footprint airborne laser scanner data.

Abstract

For this study, hierarchical data sets--in that several sample plots are located within a stand--were analyzed for study sites in the USA and Germany. The German data had an additional hierarchy as the stands are located within four distinct public forests. Fixed-effects models and mixed-effects models with a random intercept on the stand level were fit to each data set. The coefficients varied significantly also between adjacent study sites. The mixed-effects models significantly improved the estimates and especially reduced the bias that was present for numerous stands in the predictions of the fixed-effects models. The RMSE for the German study site was higher (22.5 to 31.3 percent) than for the U.S. study site (16.7 percent). Unlike the American data set, the German data set necessitated the spatial correlation of the data to be considered. A mixed-effects model with random effects on the study site and stand level was fit to the complete German data set. It provided comparable goodness-of-fit statistics for the local mixed-effects models. The study shows the potential of mixed-effects models in this context. It illustrates that use of mixed models could be a good alternative to the common practice of fitting different models for different groups of data.

Citation

Breidenbach, J.; Kublin, E.; McGaughey, R.; Andersen, H.-E.; Reutebuch, S. 2008. Mixed-effects models for estimating stand volume by means of small footprint airborne laser scanner data. The Photogrammetric Journal of Finland. 21(1): 4-15