A COMPARISON OF PLOTLESS DENSITY ESTIMATORS USING MONTE-CARLO SIMULATION

We conducted an extensive simulation study to compare the performance of a large group of plotless density estimators (PDEs) to obtain clari fication of their relative performance in a diversity of sampling situ ations. The PDEs studied included well-known ones from the literature plus some extensions and modifications introduced here. The simulation s cover 96 combinations of 6 spatial patterns, 4 sample sizes, and 4 p opulation densities. We made comparisons within classes of similar est imators, and we indicate the best-performing PDEs out of the complete set studied. Over all spatial patterns, the angle-order estimator with measurements to the third-closest individual in each quadrant had the lowest relative root-mean-squared error (RRMSE), followed by the same estimation method with measurements to the second closest individual in each quadrant. Also performing well were the variable area transect , the ordered distance estimator using the third closest individual, a nd an extension of the Kendall-Moran estimator that searches for the s econd nearest neighbor and pools search areas from all sample points. Opinions and recommendations are given as to which PDEs perform well e nough and are practical enough to deserve strong consideration for use in the field.