Speed, slope, and synchrony: Empirical insights into SAR searcher behavior.

Abstract

Wilderness search and rescue (SAR) missions are time-critical and terrain-dependent, so planners must quickly allocate resources across complex landscapes. In practice, they rely on expert judgment and experience-based assumptions to coordinate individuals and teams, yet few of these assumptions have been formally validated with field data and modeling. We address this gap by analyzing GPS tracks from 64 SAR incidents, selecting 61 tracks from 13 cases. The tracks are categorized by search tactic: hasty, sweep, or team sweep, with the last divided into six teams of varying size. To quantify how tactic and terrain shape movement, we bootstrapped exponential speed-slope fits, ran Kolmogorov-Smirnov tests, and used a nested ANOVA with random effects. Median uphill and downhill speeds are statistically indistinguishable (0.48 m/s vs. 0.52 m/s; KS p = 0.093), suggesting that slope penalties on pedestrian speed can be modeled symmetrically. Hasty searches are faster than sweeps (0.53 m/s vs. 0.39 m/s; KS p < 10-3), with no interaction between slope and tactic. Team-level analyses using Spearman correlation, time-lagged cross-correlation, and transfer entropy revealed tightly coupled movement and identifiable leaders, with follower reaction lags of only a few seconds. These empirically derived parameters-search-type specific baseline speeds, a single slope coefficient, and realistic coordination bounds-offer practical inputs for SAR coverage calculations and agent-based models. Incorporating values drawn from real field data could refine human mobility assumptions while remaining compatible with the existing SAR operational framework.