Using local spatial autocorrelation to compare outputs from a forest growth model

Comparing model outputs is a critical precursor to successfully applying models to environmental issues. In this paper, we applied a calibrated physiological model (3PG) and predicted two fundamental forest growth attributes (leaf area index (LAI) and stand volume). As part of this simulation, we systematically changed two key model input parameters (soil water holding capacity and soil fertility rating) and compared the model outputs utilising a method that accounts for local spatial autocorrelation. The use of the Getis statistic (G_i*) provides insights on the spatial ramifications of an aspatial change to model inputs. Specifically, the location of significant G_i* values identified areas where the differences in LAI and stand volume occur and are spatially clustered. When soil water is doubled and soil fertility is unchanged, both LAI and stand volume increase; conversely, when soil water is doubled and soil fertility is halved, both LAI and stand volume decrease. The increase and decrease in these model outputs occurred differentially across the study area, although there is a similar pattern to the location of the significant G_i* values (p = 0.10) in both LAI and stand volume outputs, for each model scenario. Analyzing the local spatial autocorrelation of the differences between model outputs identified those areas that have systematic sensitivity to specific model inputs. This information may then be used to aid in the interpretation of model outputs, or to direct the collection of additional data to refine model predictions.