Frost (WP13)

1. Which applications/models are planned to get into operation in Phase 2

R-written script computing frost risk based on expected 2-m minimum air temperature, timing within the growing season and possibly other factors (point 2)

2. Status: ready to go, under development

The first version has been already developed (https://github.com/destination-earth-digital-twins/DE330_WP44-Frost-CHMI); further development is planned in the Phase 2:

• running the risk scheme for multiple events and regions
• evaluate with frost damage data (availability?), redefine parameterizations if needed
• evaluate risk level for multiple crop species (apple, apricot, wine,…) – using available literature about frost damage thresholds
• determine triggering thresholds within global NWP (expected minimum air temperature, cloudiness)

3. Targeted platform (HPC needed?, EWC?)

• LUMI as we are the most familiar with the system, HPC needed for DEODE itself
• the script is relatively less computationally intense, one node at LUMI should be enough

4. Input needs, where to access to NWP data

Predicted 2-m air temperature, surface temperature or cloudiness (distinguish between advection-driven and radiation-driven events), degree-day sums (preferentially accumulated from IFS HRES runs from the start of the year), frost duration -> hourly forecasted air temperature. Data preferentially in netcdf or grib2 format.

5. Expected output, we need to sort out the format management

• maps with frost risk, possibly for multiple crop species (in netcdf or grib2 format)

6. What is the timeline? Milestones events?

• porting on LUMI, testing with DEODE output – 09/2024
• validation for multiple events/regions following data availability – 01/2025
• model redefinition following the validation runs – 05/2025
• design and implementation of triggering system – 09/2025
• full implementation into DEODE framework – 01/2026

7. PoC for sector-specific working groups

WP9

8. PoC for 3rd line support

Michael Matějka, Vojtěch Vlach