afstandssampling

Distance sampling is a statistical method used in ecology to estimate the density and abundance of populations while explicitly accounting for imperfect detection that decreases with distance from the observer. It is commonly applied in wildlife surveys and can be implemented with two main designs: line transect sampling, where observers move along lines and record the perpendicular distance to detected animals; and point transect sampling, where observers stand at fixed points and record the radial distance to detections.

Data for distance sampling consist of detected individuals or groups, their measured distances from the observer, and sometimes additional covariates such as group size, species, observer, or environmental conditions. Core assumptions include that detections at distance zero are certain (p=1), distances are measured accurately, detections are independent, animals do not move before detection due to the observer, and transects are placed randomly or systematically to represent the study area.

Analysis proceeds by fitting a detection function that describes the probability of detecting an individual as a function of distance. Common choices include half-normal and hazard-rate forms, sometimes incorporating covariates to explain heterogeneity in detectability. The detection function is used to estimate the effective surveyed area around each transect or point, from which population density is derived. In practice, density is estimated from the observed counts and the estimated detection probability, with the area of influence determined by the geometry of the design (line or point transects). Distances outside a truncation distance are often omitted to improve precision.

Software tools such as Distance and related R packages are standard for distance sampling analyses. Outputs typically include density and abundance estimates, confidence intervals, the fitted detection function, and diagnostic plots. Distance sampling is widely used for birds, marine mammals, and other wildlife where detectability declines with distance.