VARIATION IN FOREST INVENTORY FIELD-MEASUREMENTS

Field crews from the North Central Forest Experiment Station independe ntly measured two forest inventory plots in Michigan's Upper Peninsula ; one plot was measured by eight crews and the other was measured by n ine different crews. For 61 trees, the variation in measurements of di ameter at breast height (DBH), crown ratio, and site index is describe d. For DBH, the distribution of field crew mistakes and the distributi on of measurements without mistakes are described separately. For crow n ratio, the distribution of differences between individual estimates and the most frequently occurring estimate for corresponding trees is described. For site index, the distribution of differences between ind ividual estimates and the mean of plot estimates is described. Coeffic ients of variation were less than 5% for DBH, approximately 73% for cr own ratio, and 13% and 16% for site index for the two plots. The effec ts of variation in measurements on 20-year predictions of basal area a nd cumulative basal area growth were estimated for the two plots using STEMS, TWIGS, and Monte Carlo simulations. Coefficients of variation were 2% and 3% for basal area and 7% and 9% for cumulative basal area growth for the two plots. Variation in site index estimates had the gr eatest effect on variation in the output variables.