Challenge 15 - Benchmarking surface heat fluxes

[EsperanzaCuartero]

Challenge 15 - Benchmarking surface heat fluxes

Stream 1 - Software Development for Earth Sciences

Goal

Adapt existing software package to include a meaningful benchmarking strategy for surface heat fluxes.

Mentors and skills

• Mentors: Jasper Denissen, David Fairbairn, Christoph Herbert
• Skills required:
  • Background in meteorology
  • Basic statistical knowledge
  • Experience with Python
  • Experience with handling in-situ measurements (preferably Eddy-Covariance measurements)

Note: Only nationals or residents from the ECMWF Member States and Co-operating States are eligible to participate (see Terms and Conditions).

---

Challenge description

Validation of soil moisture and soil temperature is crucial for Numerical Weather Prediction (NWP), as they control surface heat fluxes that directly affect near-surface weather. This can be done with LANDVER, which is a validation package for land surface variables, currently consisting of soil moisture and soil temperature. The tool provides an independent validation of soil moisture and soil temperature data using in situ observations from the international soil moisture network (Dorigo et al., 2011). What is currently missing in this software package is the capability to validate latent and sensible surface heat fluxes against Eddy-Covariance measurements, which can provide useful information about how well ECMWF’s Land-Surface Modelling Component ECLand is able to translate soil moisture stress into surface heat fluxes. Implementing an additional routine into the already existing software package paves the way for a standardized land-surface benchmarking tool for the ECMWF.

Data/system to use

• Surface heat fluxes from the ICOS Carbon Portal which is available in Europe
• Possibility to include other networks and observations as well, with the possibility to go global
• ECMWF model output (.grib)
• LANDVER package

Solution

The LANDVER package will be complemented with routines to extract and pre-process sensible and latent surface heat fluxes from the ICOS Carbon Portal and validate and visualize ECMWF model output in a similar fashion to the already existing routines for soil moisture and soil temperature. Taking this further, the applicant will be granted the freedom to, in discussion with the mentors, go in another direction with further developing the software package (see ideas for implementation).

Ideas for implementation

Current logic that is applied for soil moisture and temperature in the LANDVER package can be adapted for benchmarking surface heat fluxes. We encourage complementing the existing framework with additional scores that could benchmark latent and sensible surface heat fluxes on other temporal signatures (seasonal/yearly) than is currently done for soil moisture and temperature. For example, as surface heat fluxes adhere to the diurnal cycle of radiation, it will be useful to validate their respective diurnal cycles. Some other ideas for implementation are:

• Importance of resolution
• Extending the current network by additional soil temperature stations or snow depth observations (dependent on the ambition of the applicant) without restriction to surface heat fluxes
• Slicing per vegetation/climate types or altitude
• Additional statistical scores/metrics

[jardabezdek]

Hello! 👋 Can you provide the link to LANDVER package, please? 🙏 When I type "LANDVER package" into Google, I cannot find anything relevant. 😿

[JasperMCDenissen]

Hey! We are currently working on making the LANDVER package available outside of ECMWF. At the start of April we will provide the LANDVER package together with experimental data to get going. So for now I am afraid you still have to rely on your imagination 😉.

[JasperMCDenissen]

For some inspiration on the application of the LANDVER package, check out David Fairbairns paper: https://journals.ametsoc.org/view/journals/hydr/20/10/jhm-d-19-0074_1.xml