eWaterCycle: A high resolution global hydrological model (original) (raw)

Last updatedOctober 11, 2025

Abstract

In 2013, the eWaterCycle project was started, which has the ambitious goal to run a high resolution global hydrological model. Starting point was the PCR-GLOBWB built by Utrecht University. The software behind this model will partially be re-engineered in order to enable to run it in a High Performance Computing (HPC) environment. The aim is to have a spatial resolution of 1km x 1km. The idea is also to run the model in real-time and forecasting mode, using data assimilation. An on-demand hydraulic model will be available for detailed flow and flood forecasting in support of navigation and disaster management.

Key takeaways

AI

1. The eWaterCycle project aims for a 1km x 1km global hydrological model.
2. Model re-engineering enables operation in High Performance Computing (HPC) environments.
3. Memory management is critical due to hydrological models' high memory demands.
4. An alternative to the Ensemble Kalman Filter (enKF) was developed to manage memory usage.
5. The presentation includes a 5km x 5km simulation and discusses financial sustainability.

Loading Preview

Sorry, preview is currently unavailable. You can download the paper by clicking the button above.

FAQs

AI

What parallelization strategies were employed in the eWaterCycle model?add

The eWaterCycle project implemented watershed logic as a primary parallelization strategy to facilitate model analysis.

How does the memory requirement of hydrological models compare to meteorological models?add

Hydrological models exhibit greater memory intensity than meteorological models due to their localized parameterization, complicating operations.

What alternative was developed to mitigate enKF's excessive memory demands?add

An alternative data assimilation approach was created to produce equivalent results without incurring the memory burden of a standard Ensemble Kalman Filter.

What was the resolution of the recent simulations in the eWaterCycle project?add

A recent simulation achieved a high resolution of 5km x 5km, enhancing hydrological analysis accuracy.

What challenges does high-performance computing present for hydrological modeling?add

High-performance computing faces challenges such as significant memory demands and the necessity for efficient memory management to avoid information swapping.