Predicting the spatial distribution of Cupressus lusitanica productivity in New Zealand

Data from a nationwide set of permanent sample plots was used to develop a multiple regression model of Cupressus lusitanica site index (mean height of the 100 largest diameter trees per hectare at age 30 years) using independent variables obtained from national extent ancillary maps and interpolated surfaces. Using this model New Zealand productivity surfaces for C. lusitanica site index were constructed.The final model accounted for 82% of the variance in the data using mean minimum air temperature, establishment date (expressed as years since 1900), potential root depth, degree of ground frost in summer and modified vegetation cover classification, with each variable accounting for, respectively, 45, 20, 7, 6 and 4% of the variance in the data. A one-at-a-time validation indicated that the final model was relatively unbiased, and accurate, with the predicted values accounting for 76% of the variance in the actual site index.Partial response curves show a positive linear relationship between site index and mean minimum air temperature and the establishment date, and a negative linear relationship with degree of ground frost in summer. The relationship between site index and potential root depth exhibited an exponential increase from root depths of 0.29 m to a maximum at root depths ≥0.53 m. For vegetation cover classification, site index for stands planted into cropland/grassland/pasture exceeded that of stands planted into mixed scrubland/exotic forest by 1.89 m. Maps illustrating the spatial variation of C. lusitanica site index within New Zealand are provided.