以電腦模擬取樣探討比特立希法與方區法在森林調查上之比較研究

A Comparison of Bitterlich and Quadrate Sampling Methods for the Forest Inventory Using Computer Simulation

本研究主要以北東眼山天然闊葉樹林型大面積全面調查所得之實測資料，應用電腦模擬取樣技術，針對方區法及比特立希法分別以變異度法、信賴區間法及適合度測驗等三種方法進行比較，以供日後植群調查之參考。在變異度法及信賴區間法之比較上，顯示相同的取樣面積下，方區法推估優勢度所需設置之理論樣區數明顯多於推估密度之所需，而比特立希法在取樣上之準確性，受角度尺之比例是否適當所影響。在本研究中，當角度尺比例小於1：40，則產生低估，且低估之情形隨著角度尺之比例愈小更加明顯。藉由適合度測驗發現，比特立希法較小比例角度尺對優勢度之低估，主要來自於對較大直徑級林木產生低估所致；而對密度之低估，則主要係來自於對較小直徑級林木產生低估所致。 The quadrate sampling method and the Bitterlich sampling method were compared by computer simulation, using the tree census data obtained from the broad-leaved forest at Mt. North Tungyen. Based on variability and confident interval, for the quadrate method the number of sampling plots required for estimating dominance was much higher than that for density. For the Bitterlich method the sampling accuracy was affected by appropriateness of angle gauge; underestimation of both dominance and density was resulted when the angles were smaller than 1:40 and the underestimation increased with the decrease in angle. Based on the result of fitness test, the underestimation of dominance was mainly caused by bigger diameter classes of trees, while the underestimation of density was by smaller diameter classes.