About the comparisons

Background

There are an increasing number of UK-based hydrological simulations and related data to understand how floods and droughts may change in the future. A range of organisations, such as the Environment Agency and water utility companies, need to access, explore, and use these datasets to aid their planning.

We are working with representatives from those organisations to better understand how to present these datasets to meet their needs.

This work started in 2022 when it became clear that colleagues in the Environment Agency were overwhelmed by the Enhanced Future Flows and Groundwater (eFLaG) data release. The eFLaG dataset provided nationally consistent hydrological projections for the UK up to 2090 of river flows, groundwater levels, and potential recharge. A first version of this site was created to help users navigate the eFLaG dataset.

Recent developments

Since the release of the eFLaG data, there have been several hydrological simulations and related datasets created. These include the July 2025 UKCP18 Hydrology Intercomparison Hackathon, and the UK Hydro-MIP.

The UKCP18 Hydrology Intercomparison Hackathon coordinated an intercomparison of model simulations driven with the latest regional climate model projections from the UK Climate Projections project (UKCP18). The aims were to enable a better understanding of the robustness of future river flow projections, identify uncertainties, and build a stronger, shared evidence base to support climate resilience planning.

The UK Hydro-MIP was a community model intercomparison which compared daily streamflow simulations across Great Britain from a range of machine learning, hydrological and land-surface models. The aims were to improve understanding of our current modelling capability, understand the strengths and weaknesses of different model structures, and help quantify modelling uncertainties.

UKCP18 Hydrology Intercomparison Hackathon

About

In July 2025, a three-day UK climate and hydrology-focused hackathon was held in collaboration with UKCEH and the HydroJULES programme.

Across the UK, there are many hydrological simulations to understand how floods and droughts may change in the future based on UKCP18 climate projections. But to date, no coordinated intercomparison of these simulations exists. The outputs from this event enable the comparison of these model simulations, to better understand the robustness of future river flow projections, identify uncertainties, and build a stronger, shared evidence base to support climate resilience planning.

View the GitHub code generated during the Hackathon [feel free to edit this link]

Data

Metrics calculated for 30-year transient windows and the dimensions were:

  • 12 regional climate models
  • 71 window start dates
  • 7 seasons (annual, DJF, MAM, JJA, SON, Apr-Sep, Oct-Mar)
  • 47 metrics
  • 14 hydrological models
  • 887 gauging stations (note that not all gauging stations have data for all models)
  • 4 countries (tied to gauging station, used to aggregate up to constituent countries of the UK)

There is also an additional NetCDF file, which is a subset of the one defined above where, instead of transient windows, it has data for nine global warming levels.

UK Hydro-MIP

About

The UK Hydro-MIP is a community-led Model Intercomparison Project (MIP), seeking to further hydrological understanding by comparing streamflow simulations over Great Britain from a range of machine learning, hydrological and land-surface models.

The purpose of the MIP was to produce modelled river flow outputs spanning a historical 40-year period (20 years of calibration, 20 years of evaluation) to help explore and summarise the current strengths of hydrological models used by the UK community. The MIP includes a diverse selection of models; crucially they all used consistent driving data to generate outputs for over 628 catchments.

The UK Hydro-MIP represents a diverse range of model types, including:

  • machine learning models (LSTM)
  • lumped hydrological models (EXP-HYDRO, HBV, GR6J, PDM, RRMT)
  • HRU-based hydrological models (DECIPHeR, Pycatchmod)
  • gridded hydrological models (Grid-to-Grid, HBV-TYN, SHETRAN, Wflow_sbm)
  • land surface models (JULES)

The full UK Hydro-MIP covers 673 catchments across Great Britain.

Read more about the UK Hydro-MIP

Data

Driving data

Gridded meteorological drivers from, or derived from, the CHESS-met dataset were used with precipitation from the HadUK-Grid dataset.

Models

Placeholder: A sentence or two about each model and a link to project page or GitHub repo.

Output data

Daily mean timeseries of river flow in m3s-1 from each model for the 628 core catchments (and up to 673 catchments) for 1980 to 2019 inclusive.