Diagnostics for Dynamics

General informations

Did you ever feel like the tools and datasets you use would surely be useful to others in the institute — but you didn't know to whom exactly? Or you worked on a problem, thinking someone probably solved it before, but didn't know who to talk to? Then Diagnostics for Dynamics (D4D) is for you! In this series of short (<45mins) monthly meetings, we aim to create a space for sharing diagnostic tools and datasets, with a focus on modelling and analysing atmospheric and climate dynamics. Go ahead an Find and suggest topics in this Trello board:).

The aim is to have a low-barrier point of interaction for scientists from different groups to get to know the resources available at IAC, who is working on what, and also find potential for collaboration. Meetings will consist of a short input (20-25 mins) on the tool / dataset of the day, followed and interrupted by questions and comments. The goal is to be as practically relevant as possible.

Next dates:

• Thursday, 11. September, 11:30 (L17) - Topic: Space-time spectral analysis by Zhenghe Xuan and Jacopo Riboldi
• Monday, 13. October, 11:30 (L17) - Topic: Weather query on the servers by Michael Sprenger
• Friday, 7. November, 11:30 (L17) - Topic: TBD
• Monday, 1. December, 11:30 (L17) - Topic: TBD

Topics:

• A Welcome topic for the D4D meetings is everything that is either
  • A diagnostic tool or method (like Lagranto.ftp, but also feature tracking tools, flux or tendency decompositions, spatial or temporal analysis, ...)
  • A dataset (e.g. observational, from simulations or emulations, ...)
  • A model
  • Anything else that is rather a tool or resource that facilitates your research than a scientific result itself
• Basically anything that you think could also be useful to people in other groups working on topics different to yours.
• When in doubt, just suggest the topic and then we'll see whether it fits 😉.
• Also: Tools and presentations don't need to be super polished.

The procedure for suggesting and choosing topics is as follows:

• We use Polybox to collect topics for our D4D meetings.
• If you would like to suggest a topic, enter your topic, a short description and your contact information into this Polybox file.

Regionmask by Mathias Hauser

regionmask is a python package that can be used to create masks of (geo)spatial regions for arbitrary (longitude and latitude) grids. These masks indicate which region a gridpoint belongs to and can then be used to aggregate gridded data over the regions.

• 2025_08_07_regionmask_d4d.pdf

High resolution ICON aquaplanet simulations by Mona Bukenberger

We ran simulations with a gulf-stream-like SST perturbation to otherwise zonally symmetric SSTs and horizontal grid spacing of 2.5, 20, and 80km for three one-(perpetual equinox with hemispherically symmetric forcing)-year ensemble members for each resolution. In this D4D?, I want to talk about the exact setup, the output (3d and 2d, variables, time stepping), how to get access to those high-resolution runs, but also about ideas and collaborations that might use this dataset. Contact: mona.bukenberger@env.ethz.ch

• Presentation with basics on the setup and available data: D4D_highres_APE_data_MB_07_07_25_small.pdf

Moisture budget decomposition by Nora Fahrenbach

This analysis methods can be used to analyse the physical mechanisms that lead to changes in moisture and allows to disentangle the influence of dynamic, thermodynamic and transient eddy impacts on moisture flux (i.e., precipitation minus evapotranspiration). This method can be applied in the tropics and midlatitudes, although different terms will be dominant.

• Presentation of Moisture budget decomposition: Moisture_budget_decomposition_D4D.pdf
• Key reference: Seager et al 2010 https://doi.org/10.1175/2010JCLI3655.1

Low-Frequency Component Analysis by Robb Wills

Low-frequency component analysis is a tool based on EOF analysis for identifying the patterns the evolve on the longest timescales in a dataset (e.g., separating PDO from ENSO, or identifying modes of atmospheric variability). It can also be helpful for separating forced responses from internal variability.

• Presentation of Low Frequency Component Analysis (LFCA): LFC_IAC-D4D.pdf
• Robb's page on, among other things, this: https://iacweb.ethz.ch/staff/rjnglin/code.html
• LFCA Code options:
  • MATLAB and Python code (Wills et al, 2018): https://github.com/rcjwills/lfca
  • "ForceSMIP" repository containing an improved Python LFCA code as well as other methods for forced response estimation: https://github.com/ForceSMIP/Hackathon2023/blob/main/notebooks/ForceSMIP_LFCA_simple.ipynb. The application of LFCA in "ForceSMIP" follows Wills et al. 2020. More information about "ForceSMIP": ForceSMIP webpage.

Lagranto FTP service by Michael Sprenger

The aim of this D4D is to show how kinematic trajectories can be calculated with the Lagranto FTP service, i.e., without the need to install Lagranto locally on a computer. Lagranto is a tool that allows kinematic airmass trajectories to be computed. The driving wind fields are available from different data sources based on the Integrated Forecasting System (IFS) of the European centre for mid-range weather forecasts (www.ecmwf.int). In particular, the operational analysis and deterministic ECMWF forecasts are available. Further, for climatological analysis the ERA5 (1950-2024) can be used. In addition to the paths taken by air parcels, several meteorological fields can be traced along the trajectories, thus allowing for a refined analysis of the meteorological processes taking place along the air parcels’ path.

• Presentation of Lagranto: Presentation_LagrantoFTP.pdf
• Userguide: userguide.pdf

• D4D_highres_APE_data_MB_07_07_25_small.pdf: D4D?_highres_APE_data_MB_07_07_25_small.pdf