Aim: Climate change is altering suitable habitat distributions of many species in high latitudes. Fleshy fruit-producing plants (hereafter “berry plants”), important in arctic food webs and as subsistence resources for human communities, may be impacted, but their response to a warming and increasingly variable climate at a landscape scale has not yet been examined.  Here, we identified influential environmental determinants of berry plant distribution and produced predictions on how climate change might shift these distributions.

Location: Bristol Bay and Togiak NRCS Survey Areas, Alaska.

Methods: We built species distribution models using the Random Forests algorithm to identify key characteristics and predict the spatial distribution of habitats suitable for five berry plant species: Vaccinium uliginosum L., Empetrum nigrum L., Rubus chamaemorus L., Vaccinium vitis-idaea L., and Viburnum edule (Michx.) Raf. Then, we used future climate projections (2081-2100; representative concentration pathways 4.5, 6.0, & 8.5) to predict shifts in species’ suitable habitat distributions based on future climate conditions.

Results: The predicted amount and spatial patterns of suitable habitat for the current time period were variable among species, consistent with species’ diverse life history attributes and habitat preferences. Future climate models predicted both positive and negative changes to suitable habitat probability for all species; future binary classification maps predicted net declines in suitable habitat area for all species and climate scenarios tested. Models identified elevation, soil characteristics, and January and July temperatures as important drivers of suitable habitat distributions.

Main conclusions: Our work contributes to understanding the response of important berry plant species to climate change at a landscape scale. Shifting and retracting distributions may alter where communities have access to harvesting areas, suggesting that access to these resources may become restricted in the future. Our prediction maps may help inform climate adaptation planning as communities anticipate shifting access to harvesting locations.

We used presence and absence data for five berry plant species collected by the USDA Natural Resources Conservation Service (NRCS) to build species distribution models. The NRCS data were collected between the years 2006-2013 as part of routine soil surveys conducted throughout the state. These location data were used in tandem with geospatial predictor variables (topography, soils), climate data from Alaska-specific climate models, and the Random Forests algorithm to assess and predict berry plant habitat suitability across the Bristol Bay landscape under current (2006-2013) and future projected (2081-2100; representative concentration pathways 4.5, 6.0, and 8.5) climate conditions.  

See README file for details.