laptop with a globe on screen and science data on screens in the background

UKNCSP Webinar: ‘Beyond UKCP18 – A Community Call for Evidence’ - Q&A

A summary of questions and answers from the Q&A session on 19 July 2024

Questions posed at the webinar and the responses

The slide ‘What might be scientific building blocks?’ in Prof Jason Lowe’s presentation illustrates additional sciences required, like decision theory and ethics, but it seems to miss the issue of understanding human behavioural response under climate risk and its contribution to community resilience. Has thought been put to this yet? 

At this stage, we're not fully scoping out options. We're considering whether we need an update driven by user needs or new capabilities. Understanding human behavioural response is indeed crucial for building resilience and adaptation. Behavioural science needs to be incorporated, but we must decide whether it fits within a UKCP project or other programmes like the UKRI's adaptation initiatives. Thoughts on this are welcome. 

Have you considered making UKCPnext more of an operational service rather than a product updated every 6-10 years? Periodic updates don't align well with a rapidly changing climate. Ongoing services and scientific insights are needed. 

It's relevant to consider different delivery models. We need options for continuous, iterative updates and for major updates. Iterative updates provide data sooner but may introduce inconsistencies, whereas major updates ensure consistency but delay information. For CMIP6 and CMIP7, we should use relevant information from both rather than discarding useful data. 

Adaptation pathways to climate change will need information on impacts/losses/damages at a wide range of temperatures. How does UKCP address this? 

We're releasing new data on global warming levels, providing flexibility with a range of scenarios. User feedback often requests a wider range of global warming levels and emission scenarios, which we're considering for future products to support adaptation pathways. 

How does UKCP take into account the uncertainty added by not considering Greenland/ice sheet melting on weather/extreme events and ocean/sea level projections? 

For regional aspects, we're considering changes in ocean colour and feedback mechanisms. For global models, we include sea ice changes and teleconnections. We aim to improve model accuracy in representing these relationships. For Greenland and ice sheets, this involves sea level rise projections and understanding impacts on the Atlantic circulation. Collaborative efforts with the Environment Agency and ongoing research are crucial here. 

Do we expect to use AI/ML in future offerings to deliver a wider range of scenarios and downscaling of simulations? 

Yes, AI/ML will be important for downscaling, ensemble filling, multivariate bias correction, and user data access. The work currently exploring potential for a ‘ChatUKCP’ is an example of such advancements. 

The Met Office is a key partner in the England-wide 'Climate Ambassadors' scheme. Are data folks and partners aware of the scheme? It could support education settings with climate risk/resilience/adaptation expertise. 

Thanks for bringing this scheme to the attention of partners listening to this webinar. As you say, the Met Office very much supports and is well engaged with this scheme, and we will look into the potential to explore risk/resilience in this space. 

Updated chance of 40°C shows that UKCP18 did not adequately include models underestimating change. How can an upgrade ensure it includes current knowledge of out-of-model-range risks? 

At this stage, we're not yet fully scoping out options for an upgrade. We're considering whether we need an update driven by user needs or science developments. However, we are following the scientific developments in this space so that if an upgrade to UKCP is triggered, we can draw from the best available science on e.g. use of large ensembles to better understand/represent natural variability and the development of high impact, low likelihood scenarios.  Your thoughts on this are very welcome - if you have evidence related to the suitability of the current UKCP18 for impacts and resilience, please do make a submission to the Call for Evidence. 

UNSEEN looks exciting. Your probabilities of 40°C differ from what was reported in July 2022. Will UNSEEN outputs be available in UKCPnext? 

As above, we're not yet fully scoping out options for an upgrade. We're considering whether we need an update driven by user needs or scientific developments. However, we are following the developments in UNSEEN and are keen to explore whether it could be a part of any future UKCP update. Your thoughts on this are very welcome - if you have evidence related to the need for approaches like UNSEEN in adaptation and resilience settings as part of UKCP, please do make a submission to the Call for Evidence. 

UKCP18 focused on GCM outputs. Could it be updated to include other evidence about future climate change, such as potential tipping points and larger-than-expected circulation changes? 

As above, we are not yet at the stage where we are fully scoping options for an upgrade. However, we are aware that some of our users are very interested in ‘HILL’ scenarios and we’re following the science developments in this space and keen to explore whether these types of scenario could be part of any UKCP update. Your thoughts on this are very welcome - if you have evidence that scientific developments in this space that point to a need for an update to UKCP, please do make a submission to the Call for Evidence. 

Aside from climate projections, is work needed on observational/historical datasets, especially for land-sea interfaces? 

Observational data is a key component of UKCP18, and so any update will also consider the needs and opportunities for provision of observational/historical datasets. Your thoughts on this are very welcome - if you have evidence relating to user needs for observational datasets or opportunities emerging from scientific developments of historical datasets for UK climate, please do make a submission to the Call for Evidence. 

Regarding using large ensembles to simulate extreme events, how many runs are considered enough? How to avoid delivering events too unlikely to be decision-relevant? 

The larger the ensemble, the more ‘rare’ extreme events can be captured, and the clearer the picture of the characteristics and more confidently we can assign probabilities of rarer extreme event in the simulation. For different decisions there will be different levels of risk aversion and so the return levels of interest, and hence the ensemble size required to inform that decision will vary. For example, information about 1-in-100 year flood levels and 1-in-500 year droughts are extensively used in flood and water resource management planning respectively. This means we need to pay close attention to the physical plausibility of the very extreme events simulated in large ensembles – for example by looking at the physical characteristics such as the spatial and temporal evolution of the event and the large scale conditions. 

Webinar resources 

Please keep up this conversation  

You can get in touch with us as follows: