A comparison of periodic and annual forest surveys

Abstract

Forest surveys of large areas often are conducted on a periodic basis, every 10 yr, for example. However, periodic assessments can be problematic due to fluctuating budgets and survey results that do not reflect forest dynamics between measurements. As a result, many countries are interested in designing surveys with annual measurements, such as the U.S. Forest Health Monitoring (FHM) Program design. The existing periodic design for Switzerland was compared with four annual designs in which an equal number of plots are observed annually. Data for the comparisons were collected annually on 682 plots over 7 successive yr across Switzerland. Three estimation methods were applied to each design: Continuous Forest Inventory, Sampling with Partial Replacement, and Mixed Estimator. These used simple means, regression updating, and inverse weighting of new and old plots, and regression combined with growth-model projections, respectively. For 12 combinations of 3 attributes, the Regional Survey design was the most cost-effective, followed by the National or Regional Update Survey designs. The Periodic Survey and the National Survey (similar to the FHM design) were the least cost-effective. Sampling with Partial Replacement was the most efficient estimator, though the Mixed Estimator would have been the most efficient had less biased growth projection models been used.