chriscra/biodiversity_index_variation

This repository houses scripts to accompany Crawford et al. 2021 (Ecological Applications), in which we apply a range of biodiversity indices to an agricultural land-use prioritization model, using Zambia as a case study to investigate how variation in how biodiversity is represented affects the results of land-use prioritization analyses. # Consequences of under-explored variation in biodiversity indices used for land-use prioritization This repository houses scripts with data and analyses for: > Crawford C.L.\*, Estes L.D., Searchinger T.D., and Wilcove, D.S. 2021. Consequences of under-explored variation in biodiversity indices used for land-use prioritization. *Ecological Applications* 31(7): e02396. 10.1002/eap.2396 \*@chriscra, ccrawford@princeton.edu, Robertson Hall, Princeton University, Princeton, NJ We explore how variation in the design of biodiversity indices affects the outcome of land-use prioritization. We use the `agroEcoTradeoff` land-use prioritization model (Estes and Spiegel 2016, `doi.org`), a trade-off model designed to identify areas for agricultural expansion that meet a given production target at the least environmental cost, and apply the model to a case study in Zambia. The `agroEcoTradeoff` model allows users to minimize four constraints -- 1) biodiversity loss, 2) total agricultural area (maximizing yields), 3) carbon loss, and 4) transportation costs. Our analysis focuses on how biodiversity loss is modeled: specifically, we assess agreement between the least biodiverse areas in Zambia as identified by biodiversity indices that vary in their construction. We explore results for a wide range of criteria and methods that biologists and land-use planners have used, including: published composite indices, vertebrate taxonomic groups, metrics of species richness, methods for combining layers, and spatial resolutions. This is a living repository. See the public release and Zenodo archive for the code used at time of publication. ## Components of this Repository This repository has one primary directory housing the scripts used in our analysis. Because the data sources are all either publicly available or available upon request from the appropriate organizations, this repository does not include the raw data itself. The five primary `.Rmd` script …