Batcheler-corrected point distance versus belt transect for sampling habitat and woody vegetation

Abstract

Density estimates for habitat objects are used for a variety of forestry and wildlife management applications, but supporting field work can be costly. We used point simulations and field studies of snags and shrub stems to evaluate and compare the plotless Batcheler-corrected point distance (BCPD) method of data collection with the plot-based belt transect (BT) method for estimating object density (snags and shrub stems). The program Transect Point Density ( http://www.cnfer.on.ca/SEP/ ) was used to calculate density and confidence limits. Relative to the BT method, we found the BCPD method performs well under a range of clustered and random dispersion patterns and point densities, with measurements made to the nearest neighbour of the initial point object allowing correction of bias caused by nonrandom spatial dispersion patterns. The BCPD parameter of spatial dispersion, A1, was correlated with I, the variance-to-mean ratio used to estimate spatial dispersion in plot-based samples. The BCPD method also allows measurement to the third object (next-nearest neighbour) to account for second-order clumping, but we found that this estimate was sensitive to measurement error or technique where dense mats of wetland shrub species Ledum groenlandicum Oeder or Myrica gale L. occurred. We also used BCPD to characterize shrub community structure and found no evident bias, relative to BT data, for estimating relative species composition, richness, and diversity. The BCPD method was cost-effective relative to the BT method, with cost ratios 4–47 times higher for the BT method, and for a given investment allowed sampling of a greater area and diversity of stand conditions. Consequently, the BCPD method will facilitate collection of better data for developing and testing habitat models as well as evaluating the effects and effectiveness of forest management options at both the stand and landscape scales.