Cooperative Fish and Wildlife Research Units Program: Florida
Education, Research and Technical Assistance for Managing Our Natural Resources

The line transect distance sampling method provides unbiased estimates of abundance when organisms are distributed randomly or line transects are laid out randomly, sample sizes are large and other assumptions of the method are met; such, however, is rarely the case in real life. We conducted a simulation study to investigate how spatial distribution and density of objects, and total length, layout and number of transects influence bias, precision, and accuracy of estimates of abundance obtained by distance sampling along line transects. Overall, density estimated using the distance sampling method was within 4.9% of the true density, but it varied substantially depending upon spatial distribution of objects. Of the three spatial distribution patterns considered, estimates of density were least biased, and most precise and accurate when objects were distributed randomly; they were most biased, and least precise and accurate when objects followed a clumped distribution. The estimated bias (% difference between true density and estimated density) for clumped, random and uniform distribution was 13.1%, -0.4%, and 2.1%, respectively; precision (% coefficient of variation, CV( D ˆ )) was 13.7%, 9.1%, and 9.2%; and accuracy (root mean-squared error, RMSE) was 27.9%, 7.4%, and 11.7% for clumped, random, and uniform distribution, respectively. Increasing total transect length and using several short transects (as opposed to few long transects) generally reduced bias, and increased accuracy and precision of estimates of abundance. A systematic layout of transects worked as well as, or better than, random layout, except when objects were distributed uniformly in space. This study advances the utility of the line transect method by providing information both on how study design affects accuracy and precision of abundance estimates, and how it can be improved when assumptions of the method are not strictly met based on a priori knowledge of the spatial distribution and presumed density of the target organism through appropriate changes in the study design.