Search site

Anglian Water worked example

Objectives: To develop a methodology, using UKCP09, to assess the impact of climate change on reservoir yield. This will be used to evaluate the robustness of the water resource system, supply-demand balance and impact on proposed water resource schemes.

  • How they used UKCP09 dummy data. 

Note: this method was used because the dummy data did not provide Cumulative Distribution Functions.

Reservoir yields were estimated for possible future climate. Two situations were considered: with no restrictions of water use to the customer and where levels of service are met (e.g. hosepipe bans). This required the simulation of the river flows by a rainfall-runoff model and future daily precipitation and evapotranspiration (PET) data for the relevant catchments. Two methodologies were explored to derive these 30-year period climate series.

Change factors from the probabilistic projections (land)

  • Temperature and precipitation information from the CDFs and joint probability plots were used for a general understanding of the climate in the region.
  • Several monthly change factors for future temperature and precipitation were extracted from the sampled data for the time periods, emission scenario, location, and probability levels specified alongside.
  • The change in PET was derived from the change in temperature using a valid method (Oudin et al, 2005).
  • Monthly precipitation and PET factors were applied to the daily recorded raingauge and MORECS PET data respectively to obtain perturbed series necessary for the modeling exercise.

Change factors from the UKCP09 Weather Generator

  • The Weather Generator was run 100 times for the time periods, emission scenario, location and percentiles specified alongside.
  • Daily rainfall and PET series were readily available from the Weather Generator outputs.
  • Future daily precipitation and PET values were used for the modeling exercise.
  • What could they do next?

Results could be used to substantiate changes in levels of service and pricing structure to customers. UKCP09 could be used to inform the Periodic Review 09 and the Water Resources Management Plan.

Estimated yields for the 2020s and 2030s could be used for water resource scheme planning. The simulation of yields for the 2050s and 2080s could be used for a longer term view of potential futures ensuring that policies adopted now do not limit future adaptation measures.

The yields calculated from the two methodologies could be compared to explore the differences in the approaches. Moreover, uncertainty arising from emission scenarios could be analysed for robustness of adaptation.

Results consisting of median reservoir yields (50%ile) with 90% (10%ile and 90%ile) and 95% (5%ile and 95%ile) confidence intervals for all reservoirs would help quantify uncertainty in planning.

Further work could be carried out assessing the implications of future changes to reservoir yield on the levels of service provided to customers.

  • What they learnt about UKCP09

Guidance is required on the correlation between temperature and rainfall .

The largest aggregation of Weather Generator grid squares does not represent a large catchment.

It is still yet to be decided how to deal with multiple runs of the Weather Generator when one daily series is required for the impacts model.

  • Contact details: Dr Catarina Henriques, Water Resources Modeller; Dr Gerry Spraggs, Hydrology Manager, Anglian Water