Improving Precision of Forestry Estimation

Presented by:
Sam Olson, Olek Wojcik, & Paul Nguyen (Reed College)
Abstract:

The National Forest Inventory and Analysis (FIA) Program of the United States Forest Service collects and analyzes data on many important forest attributes in the United States. In producing estimations, the current FIA procedure is to utilize post-stratification (PS) estimation, and the Interior West (a region of the United States) uses this estimation technique by using forest and non-forest stratification. We compare the use of various estimation techniques in the Interior West, specifically the Horvitz Thompson, Post-Stratification, and GREG estimators, and introduce and compare a new estimator, GREGORY. Comparing the relative efficiencies of bootstrap standard errors, we find that the more complex estimators generally improve the precision of forest attribute estimates, and to aid in our analysis we created a dashboard to allow for easy comparison.