Decision support systems are needed by natural resource managers to guide decisions regarding how to best protect species of concern while considering other land use needs. However, the effectiveness of such systems is currently constrained by both the quality of data describing the known locations of species of concern and the need for methods that can objectively evaluate the tradeoffs between protection and other needs. An objective decision support system that is based on high-quality data is especially needed on Department of Defense (DoD) lands that contain a diverse mix of land uses, including provision of habitat critical to the persistence of rare, endangered, or otherwise sensitive species. The optimal mix of land uses within a landscape would confer protection to critical habitat while minimizing the costs of restricting activities critical to the DoD's mission. The primary objective of this research is to develop a data-driven, spatially-explicit, decision support system that can be used by multiple, collaborating jurisdictions to identify the mixes of land use that optimally protect species of concern on DoD and surrounding lands. eDNA survey data will be a primary source of input to the decision support system, and its use represents a significant step forward in conservation management due to its increased ability to accurately detect the presence of species. A secondary objective is to determine if species occupancy and species distribution models calibrated with eDNA data can predict occurrences accurately enough to serve as input to the decision support system.

The research team will test the utility of a decision support system that takes advantage of (1) high-quality species occurrence data obtained from eDNA surveys, (2) models that can predict the habitat suitability of different locations within a landscape, and (3) state-of-the-science software that can objectively identify optimal conservation solutions. Researchers will conduct eDNA surveys for six aquatic vertebrate species of concern that occur in and around military installations in coastal California. The eDNA surveys are much more sensitive than traditional survey methods in detecting species occurrences. Researchers will use the survey results coupled with environmental data to build both species occurrence models and species distribution models to predict the habitat suitability of in-surveyed locations. Species occurrence models use replicated samples to produce unbiased estimates of detection probabilities and then adjust predictions of occurrence based on those probabilities. However, they are data intensive and can be difficult to fit. Species distribution models can often be more easily fit, but they typically assume perfect detection, which compromises their accuracy. Researchers will specifically test the hypothesis that when calibrated with highly sensitive eDNA data, species distribution models will perform as well or better than species occupancy models. Researchers will then use Marxan software to identify the mixes of land use that are optimal in both protecting the target species and minimizing costs. Marxan analyses will be based on four scenarios: (1) directly measured occupancy (eDNA survey) with current landscape settings, (2) predicted occupancy (occupancy and species distribution models) with current landscape settings, and scenarios (1) and (2) with estimated future risks of wildfire, additional land use, and climate change.

The project will produce a state-of-the-science decision support system that takes advantage of the enhanced level of species detection provided by eDNA survey methods. The comparison of species distribution and occupancy models will inform the researchers if species distribution models, when calibrated with eDNA data, are sufficiently accurate to use in conservation planning. If the tests of the eDNA-driven decision support system are successful, this approach should be easily transferable to other regions. As part of project deliverables, the research team will provide training so the approach can be readily applied to other regions in the future. This type of decision support system can provide a common tool that multiple, collaborating jurisdictions can use in conservation planning.