Mammal responses to human recreation depend on landscape context.

Rapid growth in outdoor recreation may have important and varied effects on terrestrial mammal communities. Few studies have investigated factors influencing variation in observed responses of multiple mammal species to recreation. We used data from 155 camera traps, in western Alberta (Canada), and a hierarchical Bayesian community modelling framework to document 15 mammal species responses to recreation, test for differential responses between predators and prey, and evaluate the influence of local context. Factors characterizing context were trail designation (i.e., use by motorized vs non-motorized), management type, forest cover, landscape disturbance, and season. We used three measures to characterize variation in recreation pressure: distance to trail, trail density, and an index of recreation intensity derived from the platform Strava. We found limited evidence for strong or consistent effects of recreation on mammal space use. However, mammal space use was better explained by an interaction between recreation and the influencing factors than by either on their own. The strongest interaction was between trail density and management type; mammals were more likely to avoid sites near a higher density of trails in areas with more restrictive management. We found that responses to recreation varied with the trail designation, although there were not clear or consistent differences between responses to trails designated for motorized vs. non-motorized use. Overall, we found that responses were species- and context-dependent. Limiting the density of trails may be important for reducing negative impacts to mammals within conservation areas. We show that using multiple measures of recreation yields more insight into the varied effects of human disturbances on wildlife. We recommend investigating how different characteristics of recreation (noise, speed, and visibility) influence animal behaviors. Multispecies monitoring and modelling across multiple landscapes that vary in recreation pressure can lead to an adaptive management approach to ensuring outdoor recreation coexistence with wildlife.

Predicting carbon benefits from climate-smart agriculture: High-resolution carbon mapping and uncertainty assessment in El Salvador.

Agroforestry management in smallholder agriculture can provide climate change mitigation and adaptation benefits and has been promoted as 'climate-smart agriculture' (CSA), yet has generally been left out of international and voluntary carbon (C) mitigation agreements. A key reason for this omission is the cost and uncertainty of monitoring C at the farm scale in heterogeneous smallholder landscapes. A largely overlooked alternative is to monitor C at more aggregated scales and develop C contracts with groups of land owners, community organizations or C aggregators working across entire landscapes (e.g., watersheds, communities, municipalities, etc.). In this study we use a 100-km2 agricultural area in El Salvador to demonstrate how high-spatial resolution optical satellite imagery can be used to map aboveground woody biomass (AGWB) C at the landscape scale with very low uncertainty (95% probability of a deviation of less than 1%). Uncertainty of AGWB-C estimates remained low (<5%) for areas as small as 250 ha, despite high uncertainties at the farm and plot scale (34-99%). We estimate that CSA adoption could more than double AGWB-C stocks on agricultural lands in the study area, and that utilizing AGWB-C maps to target denuded areas could increase C gains per unit area by 46%. The potential value of C credits under a plausible adoption scenario would range from $38,270 to $354,000 yr-1 for the study area, or about $13 to $124 ha-1 yr-1, depending on C prices. Considering farm sizes in smallholder landscapes rarely exceed 1-2 ha, relying solely on direct C payments to farmers may not lead to widespread CSA adoption, especially if farm-scale monitoring is required. Instead, landscape-scale approaches to C contracting, supported by satellite-based monitoring methods such as ours, could be a key strategy to reduce costs and uncertainty of C monitoring in heterogeneous smallholder landscapes, thereby incentivizing more widespread CSA adoption.