Modifying distance methods to improve estimates of historical tree density from General Land Office survey records

ANDERSON, R. C., S. L. JONES, AND R. SWIGART (Department of Biological Sciences-4120, Illinois State University, Normal, IL 61790-4120). Modifying distance methods to improve estimates of historical tree density from General Land Office survey records. J. Torrey Bot. Soc. 133: 449–459. 2006.—Distance sampling methods are widely applied to witness tree data from General Land Office (GLO) survey records to determine historic vegetation. Most researchers apply modifications of the point-centered quarter or random pairs distance methods to witness tree distances as if these data were collected according to the procedures required for the methods. Application of five methods (three modifications of the point-centered quarter method (Q1, Q2, and Q1+Q2) and random pairs with 180u or 203u exclusion angles) to GLO records from the 345,000 ha Shawnee National Forest Purchase Area in southern Illinois, resulted in tree density ranging from 117.8–177.5 trees ha. We present a new procedure, the adjustable exclusion angle method, which empirically estimates the exclusion angle of the random pairs method based on the proportion of trees in quadrants opposite versus adjacent to the quadrant of the first witness tree. For our data, the distributions of witness trees in quadrants was best described as if surveyors used a random pairs method with an exclusion angle of 203u in their selection of witness trees. This method resulted in the highest estimates of tree density. There were significant differences (P , 0.0001) in the square root of the mean area among the five distance methods used. A surveyor bias for selecting trees near the center of the area between the ordinal directions was discovered.