Droppings and densities: can sprainting intensity predict otter abundance

Abstract

Reliable estimates of population abundance are fundamental to understand population dynamics and inform conservation management. However, obtaining robust abundance estimates for rare and elusive species remains challenging because rigorous population estimation methods are often not feasible on large scales, require long survey periods, resources and technical expertise. Consequently, wildlife monitoring efforts frequently rely on population indices derived from indirect evidence of the species of interest or proportion of sites occupied. Although the relationship between these indices and population abundance is seldom evaluated. This uncertainty is particularly relevant for otters, which have primarily been studied using their indirect signs which were developed to monitor population status. Smooth-coated otters (Lutrogale perspicillata), a semi-aquatic carnivore for which ecological studies have traditionally relied on indirect signs to study their habitat and diet preferences. Despite its threatened status, there is very little quantitative information on population sizes and trends. This study aimed to estimate the population status of smooth-coated otters inhabiting three Himalayan foothill rivers using multiple complementary approaches and to evaluate whether spraint encounter rates could provide a reliable indication of relative abundance. Camera trap distance sampling, N-mixture modelling based on direct sightings were used to estimate otter populations in each river. Sign surveys were conducted to obtain relative abundance indices such as encounter rates or occupancy probability from each river. By incorporating the use of camera traps, insights into their activity pattern was analysed in these rivers which are outside a protected area that vary in levels of anthropogenic disturbance. The N-mixture model produced highly uncertain abundance estimates because of low detection probability and high variability in count data, indicating that direct-sighting approaches are unlikely to provide reliable abundance estimates for smooth-coated otters under similar field conditions. Camera trap distance sampling generated substantially larger datasets and yielded more precise estimates of detection probability and density, demonstrating greater potential for estimating otter populations. Nevertheless, the number of detections remained insufficient to estimate abundance independently for each river, highlighting the need for increased sampling effort and more rigorous approaches, such as non-invasive genetic approaches or telemetry information, to obtain robust population estimates. Single-season occupancy modelling was of limited value for comparing relative abundance because otters occupied nearly all sampling units within the study area, providing no variation in occupancy estimates between rivers. Similarly, indirect sign surveys for spraint sites failed to show a significant relationship with camera trap encounter rates or estimated population status. Although Khoh exhibited substantially higher camera trap encounter rates than Kolhu, spraint encounter rates showed the opposite pattern, suggesting that sprainting intensity alone is not a reliable proxy for smooth-coated otter abundance in these river systems. These findings support evidence that social behaviour, communal latrine use and habitat-specific detectability can weaken the relationship between spraint site encounter rates and population size. Activity pattern analyses revealed that smooth-coated otters were primarily diurnal and crepuscular. However, otters in the more disturbed Khoh River exhibited a pronounced shift towards crepuscular and nocturnal activity compared with those in Kolhu. This demonstrates the behavioural adaptability of smooth-coated otters to human disturbance in their habitat. Therefore, this study provides one of the first quantitative assessments of smooth-coated otter populations using multiple abundance estimation frameworks in the Himalayan foothills. Although constrained by sample size and limited to two rivers, the study establishes important baseline estimates and demonstrates both the potential and limitations of current monitoring approaches which have been used for other taxa. The findings suggest that camera trap distance sampling currently represents the most promising non-invasive method for estimating smooth-coated otter populations, whereas spraint encounter rates should be interpreted cautiously and not used as a standalone indicator of abundance. Future studies incorporating larger spatial scales, increased sampling effort and rigorous abundance estimation methods will be essential for validating these findings and improving conservation monitoring of this threatened freshwater carnivore.