As part of its mission to create a coherent and high-resolution picture of cryo-hydrological changes across key polar and alpine environments, LIQUIDICE has launched an intensive collaborative effort to develop and apply downscaled climate forcing datasets. These datasets are essential for driving local models (T3.3, T3.4) and supporting climate impact studies (WP4) within the project.
To initiate this process, LIQUIDICE convened two key meetings bringing together WP3 and WP4 partners engaged in climate change impact studies at the local scale. These meetings served to define the downscaling strategy based on the specific needs of the community.
First, researchers focused on introducing the methodologies for downscaling climate data to kilometre-scale resolution in the LIQUIDICE study domains: the European Alps, Svalbard, and Greenland. CNR-ISAC presented stochastic and statistical methods for downscaling essential climate variables such as temperature and precipitation. A survey was shared to gather detailed partner input on data needs. Building on the feedback received, a refined overall strategy for downscaling was produced, which addresses six critical aspects:
- Choice of Models to Downscale:
- For the Alps, CORDEX-CMIP6 simulations using RegCM5 at 0.11° resolution will be downscaled. UNESCO-ICTP will provide simulations not yet publicly available.
- For the Arctic, due to the unavailability of CORDEX-CMIP6 data, CMIP6 HighResMIP global simulations (0.25° resolution) will be used. MRI-AGCM3-2-S was identified as the most comprehensive available model. Additional simulations from Task 3.1 will be incorporated post-M24.
- Bias Correction Methods:
Bias correction will be applied prior to downscaling, using reference datasets such as EMO-1, EURO4M-APGD, and E-OBS for the Alps, and the CARRA dataset for the Arctic. Pixel-by-pixel correction will help reduce common model biases. - Future Climate Scenarios:
The project will focus on the SSP3-7.0 scenario from CORDEX and SSP5-8.5 from HighResMIP, following current community protocols. - Time Span for Projections:
Partners have requested the full available periods of CORDEX and HighResMIP simulations for use in local-scale applications. - Downscaling Methods:
- Temperature: Orographic correction using atmospheric lapse rates.
- Precipitation: Stochastic downscaling with orographic considerations.
- Other variables will be downscaled using bilinear interpolation as appropriate.
- Reference Data for Evaluation:
Initial evaluations will use gridded observational datasets. For detailed assessments, local partners are encouraged to use in-situ station data that reflect specific site conditions.
An additional point of discussion was the potential value of downscaled simulations in East Asia, particularly for model evaluation. As a result, the project will explore downscaling over pilot domains in the region.
The final list of data requests is summarized in the document “Milestone 4 Survey Survey of climate data needs from local-scale modellers“ available in the LIQUIDICE online library. The compiled data requests will guide the production of Deliverable D3.2, due by month 24 (December 2026).
Through this collaborative approach, LIQUIDICE is setting a robust foundation for producing high-resolution, bias-corrected climate scenarios that will underpin modelling and impact assessment across diverse cryo-hydrological environments.
For further updates on the progress of WP3 and WP4 activities, stay tuned to the LIQUIDICE website.