Optimizing Reserve Design
Reserve networks are a cornerstone of conservation. Reserves are critical for preventing habitat loss and subsequent biodiversity loss and despite over a century of research developing a robust theoretical framework around it, reserve design remains a persistent and dynamic practical challenge. This is due, in part, to emergent complexity in our ecological understand- ing and, in part, to added complications from social factors such as human population expansion, resource extraction and climate change. These same factors add a level of urgency, further complicating the problem. Thankfully, the sophistication of computational methods has simultaneously evolved with our ecological understanding, and optimization problems are made much easier as a result. With advanced ecological theory and powerful computational tools, the next frontier is how best to synthesize the two and most accurately represent the problem, and which ecological parameters should be included in a highly dynamic optimization problem.
Systematic reserve design exercises start by defining a specific, quantifiable objective. In general terms, this objective consists of defining a set of features to protect (e.g. species), representation targets for each feature (i.e. how much of each feature is to be protected), and planning units (i.e. the areal units that are candidates for protection) with an associated cost metric (typically the area). The goal is then to select the lowest cost set of planning units that simultaneously meets all the representation targets.
In this analysis, I focused on identifying protected areas for a suite of 119 breeding birds in British Columbia. The features used in the prioritizations were the eBird Status and Trends abundance models, and the planning units were the cells of the 3 km raster grid in which these data were provided. The tool used to solve this optimization problem was prioritizr, an R package developed to solve conservation optimization problems.
I also wanted to compare the effectiveness of our current protected areas in British Columbia against an abundance-optimized solution. I downloaded the current protected lands raster from the BC Government Data Portal. I compared the existing protected areas to a prioritization based on abundance data using 15.3% as the protection target to look at how the two overlapped spatially. I also examined how well the existing 15.3% protected areas actually protected the abundance of the suite of breeding birds in BC.
Please see the whole report here.
