A method to assess the performance of growth intercept models in British Columbia

An active program of developing growth intercept models in British Columbia is underway. In lieu of a proper validation procedure, we devised a method of assessing the growth intercept models, which we describe and demonstrate for interior spruce. Four sources of error in the site index predictions of the growth intercept models are examined: measurement, sampling, model, and application error. The impact of the measurement errors tend to be intolerably large at young ages, decreasing to acceptable levels as the trees near the index age. Sampling error was small in our example. The tested models were unbiased and have good precision. An analysis of the application error indicated that the growth intercept models may underestimate site index. Key words: Application bias, measurement error, model error, sampling error, spruce, validation