Spatial priorities for climate-change refugia and connectivity for British Columbia (Version 1.1)
Bibliographic record
Abstract
Overview This dataset provides spatial priorities for climate-informed conservation and restoration planning across British Columbia. The products integrate multiple indicators of macrorefugia and microrefugia using the Core Area Zonation (CAZ) algorithm in Zonation. Macrorefugia represent relatively large areas with the potential to maintain suitable climatic conditions or species habitat as climate changes. Microrefugia represent locations where fine-scale topographic, hydrologic, ecological, or disturbance-related conditions may locally buffer ecosystems from regional climate change. The resulting spatial prioritizations are intended to support climate-informed land stewardship and protected-area planning. They can be used to identify areas with high potential to support biodiversity persistence under climate change, evaluate existing conservation networks, and distinguish areas where conservation versus restoration may be appropriate. Analyses were conducted at 1-km spatial resolution in the NAD 1983 BC Albers projection. Ranked Zonation priorities range from 0 (lowest priority) to 1 (highest priority). Land relationship acknowledgement We respectfully acknowledge that we live and work across diverse unceded territories and treaty lands and pay our respects to the First Nations, Inuit and Métis ancestors of these places. We honour our connections to these lands and waters and reaffirm our relationships with one another. Methods 1. Macrorefugia Macrorefugia priorities incorporated both climate-based and species-specific indicators. Climate-type macrorefugia were represented by forward and backward climate velocity for an ensemble of eight CMIP6 global climate models under SSP2-4.5. Climate-velocity values were transformed so that higher values represented greater refugia potential. Species-specific macrorefugia represented areas with relatively high potential to retain suitable climatic habitat through time. Separate Zonation scenarios were developed for: rare species in British Columbia (833 Red-listed, Blue-listed, and SARA Schedule 1 species), grouped by major taxonomic group; 142 climate-sensitive bird species; and 10 dominant tree species in British Columbia. Species-specific macrorefugia inputs were based on indices ranging from 0 to 1, with higher values indicating greater macrorefugia potential. Taxon-specific Zonation outputs were subsequently combined with forward and backward climate-type velocity in an overall macrorefugia CAZ scenario. Inputs were equally weighted. Separate analyses were conducted for the 2050s (2041–2070) and 2080s (2071–2100) under SSP2-4.5. 2. Microrefugia A separate CAZ scenario combined six indicators of microrefugia: topodiversity; cool slopes, glaciers, and wetlands; topographically mediated fire refugia; thermal refugia; drought refugia; and old-growth forest. Inputs were transformed to a common interpretation in which higher values indicated greater microrefugia potential. Binary indicators were assigned values of 0 or 1, while continuous or categorical products were rescaled or reclassified as described in Table 1. The six microrefugia indicators were equally weighted in the CAZ analysis. Microrefugia were treated as temporally static and therefore used for both future periods. 3. Combined macrorefugia and microrefugia priorities The macrorefugia and microrefugia Zonation outputs were combined in a final CAZ scenario, with equal weighting of the two components. Separate combined prioritizations were produced for the 2050s and 2080s. 4. Conservation and restoration priorities A post hoc human-footprint mask derived from the British Columbia cumulative-effects dataset was applied to the combined Zonation outputs. Areas of low human footprint were classified as candidates for area-based conservation, whereas areas of high human footprint were classified as candidates for restoration or other management actions. These products should therefore be interpreted as relative spatial priorities under the assumptions and inputs of the prioritization framework, rather than as prescriptive designations of areas that should necessarily be protected or restored. Table 1. Spatial inputs used to identify macrorefugia and microrefugia Refugia class Input Resolution Processing for Zonation Source Macrorefugia Forward and backward climate velocity, 8-GCM ensemble, CMIP6, SSP2-4.5 1 km Values were rescaled by adding a constant and taking the inverse so that higher values represented greater refugia potential. Carroll (2023); Carroll et al. (2015) Macrorefugia Bird macrorefugia based on backward velocity (142 species), 3-GCM ensemble, CMIP5, RCP 4.5 1 km Existing 0–100 macrorefugia-index values were used. Raymundo et al. (2025); Bateman et al. (2020) Macrorefugia Rare-species macrorefugia based on forward and backward velocity (833 species), 3-GCM ensemble, CMIP6, SSP2-4.5 1 km Existing 0–100 macrorefugia-index values were used. Stolar et al. (2026); Raymundo et al. (2026) Macrorefugia Dominant tree-species macrorefugia based on forward and backward velocity (10 species in BC), 13-GCM ensemble, CMIP6, SSP2-4.5 1 km Existing 0–100 macrorefugia-index values were used. Campbell et al. (2025) Microrefugia Topodiversity 1 km Existing categorical values were rescaled so that higher values represented greater microrefugia potential: class 2 = 0, class 3 = 0.5, class 4 = 1. Kehm (2026) Microrefugia Cool slopes, glaciers, and wetlands 1 km Reclassified to binary presence/absence (0/1). Kehm (2023) Microrefugia Fire refugia (Fire_Refugia_Topography.tif) 1 km, resampled from 90 m Existing 0–1 values were used. Kuntzemann et al. (2025a,b) Microrefugia Thermal refugia (lm_slope_LST_Tmax.tif) 1 km Thermal sensitivity was reclassified to binary values (0/1); values <1 were classified as thermal refugia, following the source-data description. Sang et al. (2025a) Microrefugia Drought refugia (drought_pred_12mo.tif) 1 km Positive drought-sensitivity values were set to 0; remaining values were multiplied by −0.124 to rescale them to approximately 0–1, with higher values representing greater refugia potential. Sang et al. (2025b) Microrefugia Old-growth forest (Map 8) 1 km Reclassified to binary presence/absence (0/1). BC Government (2023) Note: Macrorefugia inputs were first combined within relevant taxonomic groups and then integrated with climate-type velocity in the overall macrorefugia scenario. The six microrefugia indicators were combined in a separate, equally weighted CAZ scenario. Macrorefugia and microrefugia outputs were subsequently combined with equal weighting to produce the final 2050s and 2080s prioritizations. Files and directory structure Zonation_macro.7z Contains outputs from the macrorefugia Zonation CAZ analyses. Taxon-specific outputs are provided for the individual taxonomic groups used in the macrorefugia analysis, with separate results for the 2050s and 2080s. Examples: amph_rept_CAZ_2050s.tifamph_rept_CAZ_2080s.tif The archive also contains the combined macrorefugia scenarios: macro_all_CAZ_2050s.tifmacro_all_CAZ_2080s.tif These combine the taxon-specific macrorefugia priorities with forward and backward climate-type velocity. Associated .lyrx files provide ArcGIS layer symbology, and .tfw files provide raster georeferencing information. In addition, a list of bird species is included in the bird macrorefugia component of the Zonation analysis: Birds_species_included_in_Zonation_macrorefugia_scenarios.txt Zonation_micro.7z Contains the combined microrefugia Zonation CAZ output: micro_CAZ.tif This scenario integrates topodiversity; cool slopes, glaciers, and wetlands; fire refugia; thermal refugia; drought refugia; and old-growth forest. Zonation_macro_micro.7z Contains the final combined macrorefugia and microrefugia prioritizations and the products derived after application of the human-footprint mask. macro_micro_CAZ_2050s.tifmacro_micro_CAZ_2080s.tif These represent combined, unmasked macrorefugia and microrefugia priorities. Conservation priorities conservation_priorities_2050s.tifconservation_priorities_2080s.tif These represent combined macrorefugia and microrefugia priorities occurring in areas classified as having low human footprint. Restoration priorities restoration_priorities_2050s.tifrestoration_priorities_2080s.tif These represent combined macrorefugia and microrefugia priorities occurring in areas classified as having high human footprint. Associated .lyrx files provide ArcGIS layer symbology. Zonation_priorities_by_ecoprovince.7z Contains the following root folders: /Macro_scenarios/ /Micro_scenarios/ /Macro_Micro_scenarios/ Within each root folder are .jpg and .tif (GeoTIFF) files of Zonation priorities for the 2050s and 2080s (except for the microrefugia scenarios) using the same inputs as described for the province-wide scenarios above. Example: %taxon_name%_CAZ_2050s_BOP.CAZ_E.jpg %taxon_name%_CAZ_2050s_BOP.CAZ_E.rank.compressed.tif CAZ = Core Area Zonation BOP = Boreal Plains ecoprovince (in this example) rank.compressed.tif = the Zonation ranking for the ecoprovince and scenario "macro_all" refers to the combination of taxonomic groups plus macrorefugia potential based on forward and backward velocity Zonation analyses by ecoprovince were performed using the 'analysis area mask' setting in Zonation. This allows the algorithm to prioritize pixels relative to the rest of the ecoprovince rather than the entire province of British Columbia. Both approaches are useful depending on land stewardship objectives. Please see Table 1 for the list of spatial inputs. Terrestrial ecoprovinces of British Columbia: BOP BOREAL PLAINSCEI CENTRAL INTERIORCOM
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.021 |
| Meta-epidemiology (narrow) | 0.001 | 0.002 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.008 | 0.016 |
| Science and technology studies | 0.004 | 0.001 |
| Scholarly communication | 0.006 | 0.002 |
| Open science | 0.004 | 0.003 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.095 | 0.021 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".