Search

Assessing the vulnerability and adaptive capacity of species, communities, and ecosystems is essential for successful conservation. Climate change, however, induces extreme uncertainty in various pathways of assessments, which hampers robust decision-making for conservation. Here, we developed a framework that allows us to quantify the level of acceptable uncertainty as a metric of ecosystem robustness, considering the uncertainty due to climate change. Under the framework, utilizing a key concept from info-gap decision theory, vulnerability is measured as the inverse of maximum acceptable uncertainty to fulfill the minimum required goal for conservation. We applied the framework to 42 natural forest ecosystems and assessed their acceptable uncertainties in terms of maintenance of species richness and forest functional type. Based on best-guess estimate of future temperature in various GCM models and RCP scenarios, and assuming that tree species survival is primarily determined by mean annual temperature, we performed simulations with increasing deviation from the best-guess temperature. Our simulations indicated that the acceptable uncertainty varied greatly among the forest plots, presumably reflecting the distribution of ecological traits and niches among species within the communities. Our framework provides acceptable uncertainty as an operational metric of ecosystem robustness under uncertainty, while incorporating both system properties and socioeconomic conditions. We argue that our framework can enhance social consensus building and decision-making in the face of the extreme uncertainty induced by global climate change. [ABSTRACT FROM AUTHOR]

Ecological niche models (ENMs) are widely used in spatial prioritization for biodiversity conservation (e.g. selecting conservation areas). However, it is unclear whether ENMs are always beneficial for such purposes. We quantified the benefit of using ENMs in conservation prioritization, comparing the numbers of species covered by conservation areas selected on the basis of probabilities estimated by ENMs (ENM approach) and those selected on the basis of raw observation data (raw-data approach), while controlling survey range, survey bias, and target size of conservation area. We evaluated three ENM algorithms (GLM, GAM, and random forests). We used virtual community data generated by simulation for the evaluation. ENM approach was effective when survey bias is strong, survey range is narrow, and target size of conservation area is moderate. The percentage of cases in which the ENM approach outperformed the raw-data approach ranged from 0.0 to 33% (GLM), 31% (GAM), and 75% (random forests) depending on conditions. The number of rare species (< 20 presence records) included in the conservation area based on the ENM approach was less than, or the same as, that of the raw-data approach. The unexpectedly limited cases in which the ENM approach was effective in the present research may depend on the conservation target we used (to cover as many species as possible in conservation area). Our results highlight urgent need for evaluating ENM's effectiveness under other conservation targets for wise use of ENM in conservation prioritization. [ABSTRACT FROM AUTHOR]