Climatic space detailed methods:

Materials:

1. Historic mean monthly temperature values for July, August and September (T.historic), taken from 1960-1990 data (1km raster)
2. Historic mean monthly precipitation values for July, August, and September (P.historic), taken from 1960-1990 data (1km raster)
3. Historic mean temperature values for the hottest (Tmax) and coolest (Tmin) quarters in 1960-1990 (1km raster)
4. Historic mean precipitation values from the wettest (Pmax) and dryest (Pmin) quarters in 1960-1990 (1km raster)
5. Projected mean monthly temperature values for July, August, and September (T.future), taken from models for 2025, 2050, and 2099 (6km raster)
6. Projected mean monthly precipitation values for July, August, and September (P.future), taken from models for 2025, 2050, and 2099 (6km raster)
7. Ecosystem map (vector)

Given that the Materials section requires much preparation (raster arithmetic), the steps for analyzing this type of vulnerability are:

• Calculate T.range = Tmax - Tmin
• Calculate P.range = Pmax - Pmin
• Force the projected 6km data to 1km data (1 grid cell now equals 6*6=36 grid cells with the same value
• Perform a Gower Metric (GM) to give different weights to future years (due to a loss of certainty further in the future)
  • GM = 4*(2025 projection) + 2*(2050 projection) + 1*(2099 projection)
  • The Gower Metric gives T.future and P.future
  • Average the GM for all cells within an ecosystem (see diagram)

• For each ecosystem, calculate (T.historic - T.future) / T.range
  • Rank each ecosystem: the largest absolute values indicate the biggest projected change from historic climatic space, so they receive the highest rankings (scale of 1-10)
• For each ecosystem calculate (P.historic - P.future) / P.range
  • Rank each ecosystem on a similar scale of 1-10
• Add the temperature and precipitation ranks to obtain vulnerability in terms of climatic space

Projected differences in temperature and precipitation

More intense colors indicate areas that are projected to experience further deviations from the historic climatic values. While this study did not distinguish between positive or negative changes, it would be an interesting continuation.

Data sources: UNESCO ecosystem map, WORLDCLIM Historical Climate Data, CATHALAC Climate Change Models (2025, 2050, 2099 projections)