Development of models to predict <i>Pinus radiata</i> productivity throughout New Zealand

Development of spatial surfaces describing variation in productivity across broad landscapes at a fine resolution would be of considerable use to forest managers as decision support tools to optimize productivity. In New Zealand, the two most widely used indices to quantify productivity of Pinus radiata D. Don are Site Index and 300 Index. Using an extensive national data set comprising a comprehensive set of national extent maps, multiple regression models and spatial surfaces of these indices for P. radiata were constructed. The final models accounted for 64% and 53%, respectively, of the variance in Site Index and 300 Index. For Site Index, variables included in the final model in order of importance were mean annual air temperature, fractional mean annual available root-zone water storage, mean annual windspeed, length and slope factor, categories describing Land Environments of New Zealand (LENZ), and major soil parent material. The variables included in the final model of 300 Index in order of importance included the degree of ground frost during autumn, fractional mean annual available root-zone water storage, categories describing LENZ, vegetation classification, foliar nitrogen, taxonomic soil order, and major soil parent material. These results highlight the utility of thematic spatial layers as driving variables in the development of productivity models.