The probabilistic projections combine information from several collections of computer models with observations using advanced statistical methods.
Update to UKCP Probabilistic Projections
The Probabilistic Projections were updated in August 2022 and there may be some small changes to the figures within documentation on this webpage. Please refer to the project news webpage update on 4 August 2022 for the latest documentation.”
Using the probabilistic projections
The UKCP18 probabilistic projections address the continued requirement for a product exploring uncertainties. They ensure that scientific evidence used in risk assessments, and other adaptation activities, is as robust as possible. They provide support for future risk assessments and context for applications reliant on output from projection systems containing more limited sampling of uncertainties (such as the global and regional projections). Examples of potential use include:
- Carrying out a robust risk assessment for your system.
- Constructing weather files for input into your building simulation models (see demonstration project on thermal performance of buildings).
- Overlaying the probabilistic projections on response surfaces that show the vulnerabilities of specific systems or sectors to climate change (see demonstration project on water resources and drought planning).
- Downscaling for impacts studies.
Section 4.1 of UKCP18 Guidance: How to use the Land Projections [1.8MB] has links to these examples and further detail on when to use the probabilistic projections.
You can find key results from the probabilistic projections here.
Method and advances since UKCP09
The approach taken to generate probabilistic projections is based on the use of multiple variations of a specific climate model to simulate a wide range of different climate outcomes. This is known as a climate model ensemble. A statistical emulator then estimates climate outcomes for a much greater number of climate model variants based on known sensitivities within the model. These outcomes are weighted by using knowledge of observations from the real world. Probabilistic projections are available for several alternative future pathways of emissions, including RCP2.6, RCP4.5, RCP6.0 and RCP8.5 (see guidance on representative concentration pathways for more information).
The significant advances of UKCP18 over previous probabilistic scenarios provided in UKCP09 are:
- Inclusion of natural inter-annual variability, enabling users to place climate change in the context of natural variability.
- Inclusion of climate models from the 2014 assessment report from the Intergovernmental Panel on Climate Change (see 5th Assessment Report).
- A more comprehensive sampling of Earth System modelling uncertainty (i.e. climate models that simulate the interactions of atmosphere, ocean, land, ice, as well as the biosphere).
- Use of more up-to-date observations to constrain the projections.
- Improvements to the methodological approach.
Further supporting information can be found in Guidance and Science Reports and include:
- Factsheets that summarise the key findings for the available variables across all the projections.
- A summary of the results, methods, evaluation and limitations of the probabilistic projections in Section 2.2. of the UKCP18 Science Overview report.
- Detailed description of the methodology and scientific results in Section 2 of the UKCP18 Land Projections report.
- UKCP18 Marine and Derived Projections reports.
The full dataset is available from the UKCP18 User Interface and CEDA Data Catalogue (note that this requires familiarity with handling large datasets).