Chiem van Straaten

AS Atmospheric Sciences

The 2022 Outstanding Student and PhD candidate Presentation (OSPP) Award is awarded to Chiem van Straaten for the poster/PICO entitled:

Improving sub-seasonal temperature forecasts by correcting missing teleconnections using ANN-based post-processing (van Straaten, C.; Whan, K.; Coumou, D.; van den Hurk, B.; Schmeits, M.)

Click here to download the poster/PICO file.

Chiem van Straaten is a PhD student at the VU Amsterdam and the Royal Netherlands Meteorological Institute (KNMI). His topic is the long-range forecasting of European weather. Such forecasts require an understanding of drivers of predictable extreme events. Chiem uses tools from Machine Learning and eXplainable AI to quantify these driving mechanisms. He then applies statistical modeling to correct numerical weather forecasts when the mechanisms are incorrectly resolved.

In his presentation at EGU22 he outlined an artificial neural-network architecture that can be used to discover unresolved ‘forecasts of opportunity’. These are circumstances in which the weather remains predictable multiple weeks into the future, but where numerical weather forecasts fail to resolve that predictability. The network learns to restore such opportunities. He further demonstrated that corrections can be interpreted physically, for instance as a restored teleconnection from the tropical Pacific to Europe.

Publication resulting from the award

van Straaten, C., Coumou, D., Whan, K., van den Hurk, B., and Schmeits, M.: Strengthening gradients in the tropical west Pacific connect to European summer temperatures on sub-seasonal timescales, Weather Clim. Dynam., 4, 887–903, https://doi.org/10.5194/wcd-4-887-2023, 2023.