Reliable monitoring of elephant populations in the forests of India: analytical and practical considerations

Abstract

Reliable estimation of elephant population abundance and density assumes great importance in the context of massive threats from illegal hunting and habitat loss. However, available estimates of elephant populations, particularly in Asia, are often unreliable and misleading. We evaluate sources of bias and imprecision in commonly used estimation approaches, and demonstrate that if correctly applied, line transect sampling based on visual detections of elephant clusters can address these issues. We compare our own early transect surveys on foot that relied on purposive line placement, to subsequent surveys in 2011, which employed rigorous survey designs. Estimated elephant density (D^(95%CI^(D^)) in our study sites in India, ranged between 0.25(0.12–0.53) and 3.29(1.74–6.21) elephants/km² in the earlier surveys and between 0.32(0.14–0.75) and 2.24(1.41–3.56) elephants/km² in the 2011 survey. Although coefficients of variation of estimated detection probability (pˆ) and cluster size (E^(S)) were higher at low sample sizes, they dropped to <15% with n > 40 detections. Variance of encounter rate (n/l) was the largest contributor to the variance of density estimates. We recommend that rigorous line transect surveys must ensure: random transect placement with systematic and sufficient spatial replication to ensure adequate spatial coverage; coverage of sufficiently large areas in a short duration to ensure population closure; and investment of adequate effort to ensure reasonable number of detections. Field and analytical protocols presented here can enable reliable estimation of density and abundance of other wildlife species that can be visually detected in forests. They can lead to improved animal monitoring programs that are central to rigorously evaluating the effectiveness of widely employed, expensive conservation interventions meant to counter massive anthropogenic threats facing elephants and other large, diurnal species.