Optimizing inventory and yield data collection for forest management planning

This work uses a CostLoss approach to estimate the optimal investment in inventory information for forest planning. A bootstrapping approach is used to simulate the impact of different inventory intensities on the quality of decisions in a linear programming harvest scheduling model. Multiple formulations of the harvest model based on varying inventory intensities are used to calculate the value of a variable labeled Loss that measures the monetary losses resulting from the use of imperfect yield information in the harvest model. The variable Loss and the cost of obtaining the inventory information are used to estimate empirical relationships between their expected value and the amount of inventory information (number of inventory plots and number of experimental plots) used in the harvest models. These two relationships are combined to give an explicit estimate of the expected CostLoss as a function of the inventory intensity variables. This CostLoss relationship is minimized to estimate optimal number of inventory plots and the optimal number of experimental plots. An example is developed with radiata pine information from southern Chile. Results for this example suggest that current practice uses too many experimental plots and too few inventory plots.