Use of the visual range of detection to estimate effective sweep width for land search and rescue based on 10 detection experiments in North America.

OBJECTIVE: Standard-of-practice search management requires that the probability of detection (POD) be determined for each search resource after a task. To calculate the POD, a detection index (W) is obtained by field experiments. Because of the complexities of the land environment, search planners need a way to estimate the value of W without conducting formal experiments. We demonstrate a robust empirical correlation between detection range (Rd) and W, and argue that Rd may reliably be used as a quick field estimate for W. METHODS: We obtained the average maximum detection range (AMDR), Rd, and W values from 10 detection experiments conducted throughout North America. We measured the correlation between Rd and W, and tested whether the apparent relationship between W and Rd was statistically significant. RESULTS: On average we found W ≈ 1.645 × Rd with a strong correlation (R(2) = .827). The high-visibility class had W ≈ 1.773 × Rd (also R(2) = .867), the medium-visibility class had W ≈ 1.556 × Rd (R(2) = .560), and the low-visibility had a correction factor of 1.135 (R(2) = .319) for Rd to W. Using analysis of variance and post hoc testing, only the high- and low-visibility classes were significantly different from each other (P < .01). We also found a high correlation between the AMDR and Rd (R(2) = .9974). CONCLUSIONS: Although additional experiments are required for the medium- and low-visibility search objects and in the dry-domain ecoregion, we suggest search planners use the following correction factors to convert field-measured Rd to an estimate of the effective sweep width (W): high-visibility W = 1.8 × Rd; medium-visibility W = 1.6 × Rd; and low-visibility W = 1.1 × Rd.