Improving forest field inventories by using remote sensing data in novel sampling designs

Abstract

It is becoming more common that auxiliary information from remote sensing is available at the planning stage of a forest field inventory. Recent developments in sampling theory allows the inclusion of such information in the sampling design to obtain better samples and, hence, improve estimates of common forest attributes. We explain the methodology and evaluate the possibility of including data from airborne laser scanning in the sampling design. The novel designs that we use can select samples that are balanced on a set of auxiliary variables and (or) well spread in a set of auxiliary variables. The results from a simulation study with real data indicate that significant improvement is achieved.