Publications > Technical reports > Technical report (CCRA3-IA)

Chapter 1: Latest Scientific Evidence for Observed and Projected Climate Change

Published:
16 June 2021

Assessment:
CCRA3-IA

Country focus:
UK

Authors

Lead authors: Julia Slingo

Contributing authors: Richard Betts, Jonathan Beverly, Ben Booth, John Caesar, Mat Collins, Hayley Fowler, Fai Fung, Laila Gohar, Jonathan Gregory, Helen Hanlon, Tim Johns, Lizzie Kendon, Jason Lowe, Dann Mitchell, Matthew Palmer, Cath Senior

Additional contributors: Kathryn Brown, Piers Forster, Ed Hawkins, Richard Millar

This chapter should be cited as: Slingo, J. (2021) Latest scientific evidence for observed and projected climate change. In: The third UK Climate Change Risk Assessment Technical Report [Betts, R.A., Haward, A.B. and Pearson, K.V. (eds.)] Prepared for the Climate Change Committee, London

CCRA3-IA Chapter 1: Latest Scientific Evidence for Observed and Projected Climate Change

This publication is available in PDF format at the end of the page >

Key Messages

This chapter considers the latest observations of, and future projections for, the changing climate in the UK and across the globe. In particular, this chapter focuses on new projections of climate change arising from developments in climate modelling since the 2^nd Climate Change Risk Assessment (CCRA2) in 2017. These developments allow for a more comprehensive assessment of future UK and global climate changes, including those which might alter, materially, the range of risks and opportunities in the various sectors of this assessment compared to previous projections.

This chapter uses the term ‘climatic impact drivers’ to refer to changes in aspects of UK and global weather and climate (e.g., precipitation, temperature, etc). This distinguishes projected changes in aspects of the UK’s weather and climate from the hazards and opportunities that they drive (e.g., heavy rainfall changes driving changes in flooding hazard in some instances). This chapter also puts greater emphasis than in CCRA2 on assessing the variability of the UK’s weather and climate, and how this variability might change as the planet warms.

The main conclusions of this chapter are:

Since CCRA2 in 2017, the world has continued to warm with effects on UK and global weather and climate becoming more evident and increasingly attributed to human-induced climate change. Temperature and sea level rise are the clearest signals of a changing climate for the UK. There is a growing body of evidence suggesting increases in mean rainfall, particularly in winter, and down the west side of the UK. The evidence base for whether UK storminess will change remains weak and should be addressed as a priority for future research.
New UK weather and climate records are being set more frequently, with the UK experiencing unprecedented high temperatures and heavy rainfall. Extreme events test resilience and preparedness, and there is increasing evidence that, even today, human influences have changed and continue to change the likelihood of weather and climate extremes. New science since CCRA2 has highlighted that unprecedented extreme weather events are possible even in today’s climate; for example, there is currently a 1% chance every year that monthly winter UK rainfall can be 20-30% higher than the maximum observed.
The UK is projected to experience ongoing increases in temperature until the middle of the 21^st Century under all scenarios for future global climate change, including those approximately consistent with achievement of the goal of the Paris Agreement to limit global warming to well below 2°C. Until the middle of the century, the extent and spatial pattern of UK climate change depends more on regional climate and weather responses to global warming than the level of future global greenhouse gas emissions.
From the 2050s onwards, higher emissions scenarios are projected lead to greater increases in extreme weather and sea level both at UK and global scales. In a high scenario by 2080, 40°C is projected to be exceeded as frequently as 32°C is exceeded today. At the time of this assessment 40°C has not yet been reached in the UK, but could occur with a return time of 3.5 years by the end of the century in a high-warming scenario.
The severity of extremes is projected to increase with global warming. Meteorological, agricultural and hydrological droughts are expected to become more severe with implications for water resource management. An increase in the incidence of high summer daytime temperatures throughout the UK. In the future, Scotland and Northern Ireland could start to see high summer temperatures similar to those of England and Wales currently. All parts of the UK will continue to experience a steady reduction in frost days as global warming increases, although some years will still see similar numbers of frost days and cold-related impacts as in recent years.
Future summers are projected to be even hotter and drier than earlier estimates in CCRA2, for equivalent levels of global warming. Based on the new Met Office models in the UK Climate Projections 2018 (UKCP18), reductions in future rainfall are substantially larger over England, typically double those used in CCRA2. This is due to improved simulations of summer circulation anomalies and their impacts on rainfall, as well as higher temperatures. However, year-to-year variations in summer rainfall indicate that while drier summers are generally more likely across the UK, wetter summers are also possible. Furthermore, despite overall summer drying, with wet days projected to become less frequent, the new kilometer-scale projections suggest that when it does rain, the daily rainfall will be more intense by as much as 20%, relative to coarser models.
We can expect more frequent and more severe extreme daily high temperatures and Urban Heat Island effects even though the mean warming is almost identical. Better representation of the landscape and urban areas in the kilometre-scale model have highlighted that there is a very small chance (less than 0.02%) of exceeding 40°C by 2040, but by 2080 the frequency of exceeding 40°C is similar to the frequency of exceeding 32°C today. Urban heat island intensity will increase during both day and night, but new results have shown greater increases in night-time intensities implying significantly more ‘tropical nights’. During summer, night-time temperatures in the urban areas of 10 major cities will increase significantly, at rates between 0.48 and 0.55^oC per decade during the 21^st century.
New kilometre-scale projections show pronounced shifts to more intense hourly rainfall at the expense of lighter rainfall, compared with coarser models as used in CCRA2. For example, summer hourly extremes of 20mm/hour may occur twice as frequently as previously projected. Furthermore, there is new evidence on the frequency of rainfall exceeding 30mm/hour for some UK cities, showing that such events are twice as likely by 2080.
Winter extreme rainfall is projected to be around 40% more intense compared to CCRA2, with future winters becoming warmer and wetter overall. New kilometre-scale projections from UKCP18 shows that physical processes not resolved in the coarser regional and global models increases projections for future winter rainfall by around 40% relative to previous projections. They also show that daily rainfall intensity is projected to increase by as much as 25% relative to coarser models, particularly in the south-east.
Future winter weather is projected to be dominated by more mobile, cyclonic weather systems than was the case in previous assessments. This will affect the western parts of the UK, in particular, and reinforces the evidence for more substantial increases in daily rainfall with related flooding, as well as a higher incidence of strong winds and waves. The projected shift to more mobile, cyclonic winters may also increase the risk of atmospheric river events that bring large amounts of rainfall and are major contributors to severe flooding and landslides, particularly for the mountainous regions of the UK.
The increase in rainfall intensity and mobility of cyclonic weather patterns arise from improvements in climate modelling. Although the HadGEM3-GC3.05 model used for these components of the UKCP18 projections has a very high climate sensitivity, these results for rainfall intensity and storminess are not related to that process. The UKCP18 probabilistic projections do not use HadGEM3-GC3.05 – instead, they use the same base model as the UKCP09 projections, plus a wider consideration of the large set of multi-model projections in the 5^th Coupled Model Intercomparison Project (CMIP5). The UKCP18 probabilistic projections warm faster than the CMIP5 projections because they include uncertainties in carbon cycle feedbacks.
New estimates of global sea-level rise indicate an additional 5 – 10cm rise by 2100 compared with CCRA2 estimates. These include a contribution from Antarctic ice dynamics. These faster projections of sea level rise are due to improved understanding and modelling of land ice processes, not faster warming due to higher climate sensitivity, because the new sea level projections are being driven by the CMIP5 projections rather than the UKCP18 land projections.
Low-likelihood, high-impacts scenarios are considered in CCRA3. The new regional projections warm faster than those used in CCRA2, partly due to high climate sensitivity in the global model, and partly due to the use of a higher emissions scenario. Therefore the timing of these regional changes should not be considered the most likely outcome but factored into planning as a lower probability high risk future. In the most extreme sea level scenarios, global sea-level rise could reach 2m by 2100 but this is viewed as very unlikely and with low confidence. However, the processes behind ice sheet collapse particularly for Antarctica remain very uncertain and continued monitoring and process studies are vital.

Scientific advances since CCRA2, along with the delivery of UKCP18 and the development of a new generation of climate models, have provided new and important evidence regarding expected changes in the UKs weather and climate. This chapter summarises a significant body of new evidence on projected changes in the UK’s weather and climate that will help improve assessment of future climate risks and opportunities. In particular, the most up-to-date and physically comprehensive projections indicate that future changes may be more extreme than previously projected in CCRA2, especially locally and on daily timescales. In some areas, important knowledge gaps remain, for example, further research is required to test the robustness of projections in changes in storminess.

Introduction

This chapter summarises the latest scientific evidence on current and future climate change, including new advances made since the second climate change risks assessment (CCRA2), which either consolidate previous understanding or bring new insights. These assessments will be placed in the context of global changes, especially where these changes may have a material impact on the UK.

CCRA2 drew on literature assessing future climatic impact drivers for the UK and worldwide based mainly on the United Kingdom Climate Projections 2009 (UKCP09) and the global projections published in the IPCC Fifth Assessment Report (AR5). These projections have also informed much of the literature available for assessment in CCRA3.

Whilst the IPCC Sixth Assessment Report (AR6) has not yet been published, a new set of UK Climate Projections (UKCP18) has been published since CCRA2. These projections draw on significant advances in model development since the earlier UKCP09 and IPCC AR5, as well as research conducted in preparation for the creation of new sets of climate projections in the 6^th Coupled Model Intercomparison Project (CMIP6). CMIP6 underpins the IPCC 6^th Assessment Report, due to be published in July 2021.

This chapter will focus specifically on identifying significant differences in the climate science evidence base from UKCP09 and CCRA2, drawing on the results from UKCP18 in particular. Since CCRA2 there have been important scientific advances across a number of fronts:

Extended observational records through data archaeology and improved global and regional reanalyses, including additional recent observations during a time of global warmth unprecedented in the observational record.
Ongoing detection and attribution of climate change trends in the observational record, and the rapidly developing methodologies for attributing extreme events to human-induced climate change.
New global and UK regional projections based on improved modelling systems with higher resolution and better representation of climatic impact drivers, especially associated with climate variability, weather systems and local extremes.
Greater overall understanding of the climate system, its response to forcing, and the potential for accelerating Earth system feedbacks and abrupt changes.

It is widely recognised that some of the more costly, disruptive and dangerous impacts of climate change will be associated with increased frequency and/or intensity of extreme weather and climate events. The UK’s weather and climate are highly variable because of where the country sits on the globe – at the end of the North Atlantic storm track where cold polar and warm sub-tropical air masses collide, and with maritime influences from the ocean to the west, and continental influences from Europe to the east. These factors make detecting, attributing, predicting and projecting changes in UK’s weather and climate, especially for more extreme events, challenging, but vitally important.

Consequently, this chapter puts greater emphasis on assessing the variability of the UK’s weather and climate, how this variability might change in terms of frequency and/or intensity as the planet warms, and what this means for unprecedented extremes and their impacts. This has been made possible by significant advances in global and regional climate modelling and their applications since UKCP09 and CCRA2, with current climate models now able to capture regional and local weather with greater fidelity. These advances are documented in Annex 1 and include significant increases in global model resolution for both the atmosphere (60km spacing between grid cells, reduced from 150km in CCRA2) and oceans (0.25^ofrom 1^o in CCRA2), improvements in the model physics, increases in regional model resolution from 25km to 12km, and the deployment of a new kilometre-scale UK regional model.

Although the socio-economic impacts of climate change will be felt more keenly through extreme weather and climate events, the natural environment is also susceptible to longer-term trends in the climate, such as shifts in the regular seasonal weather and climate conditions, which affect phenology, habitats and survival of species. These impacts may only come into play, or become serious, when certain meteorological thresholds are exceeded. Consequently, as part of CCRA3, additional analysis of UKCP18 has been carried out to provide information on, firstly, the changes to weather and climate variability and extremes, and secondly, on trends in climatic indices that affect the natural environment.

1.2 Our approach to assessing the new climate science that underpins CCRA3

As with the sectoral chapters and documented in Chapter 2, the approach here is to integrate existing information from the literature base and previous projections, with new projections and other new research. UKCP18 has provided important new evidence and much of the chapter will focus on how this has evolved our understanding of the UK’s future climate. CCRA3 uses all the UKCP18 products to some extent, mainly the probabilistic projections and HadGEM3 global projections, but also the regional and local projections where possible.

1.2.1 Representing uncertainty in the new UKCP18 projections

It is essential that we consider the uncertainty in the climatic impact drivers by looking at the various sources of uncertainty, how they evolve with time through the 21^st century (Figure 1.1 based on Hawkins and Sutton, 2009), and how well the various ensembles of global and regional projections are able to sample the range of possible outcomes. It is also important to understand that different climatic impact drivers have different sources of uncertainty. As Figure 1.1(a) shows, surface temperature change is dominated by natural (internal) variability^[1] for the first 2 decades whilst the uncertainty due to the global emission scenarios remains relatively small. For the latter half of the century, the uncertainty is increasingly dominated by the global emission scenarios. Model uncertainty, associated for example with cloud feedbacks, is important throughout. On the other hand, for precipitation (Figure 1.1(b)), natural variability plays a significant role at all lead times whilst scenario uncertainty barely enters the assessment.

Chart, pie chart

Description automatically generated

The UKCP18 projections include a number of components using different sets of models and different approaches to quantifying or exploring uncertainty (Murphy et al., 2018):

the probabilistic land projections (at UK and global scales) with a range of emissions scenarios.
perturbed-parameter ensembles of global, regional and local projections (at resolutions of 60km, 12km and 2.2km respectively) with a very high emissions scenario
derived projections representing long-term climate states at 2°C and 4°C global warming and a low-emissions scenario
marine projections.

The probabilistic projections are based on the same climate model, HadCM3, as used in UKCP09, with additional information from the CMIP5 multi-model ensemble. These provide a comprehensive quantification of uncertainty, drawing on this very wide base of information including a large number of independent models. The global perturbed-parameter ensemble (PPE) uses the HadGEM3-GC3.05 climate model, and the regional and local ensembles use high-resolution limited area models taking boundary conditions from the global PPE. The derived projections use the HadGEM3-GC3.05 PPE and CMIP5. The marine projections use CMIP5, applying these to new models of sea level rise and other aspects of marine impacts. The use of different sets of models in the various strands of UKCP18 arose from the need to carry out developments in all strands in parallel. The final set of projections are all useful for different aspects of climate change risk assessment and are all used in CCRA3 in various places. Since they use different models and different approaches, use of several strands together requires an awareness of potential inconsistencies. The following chapters in the CCRA3 Technical Report make clear which UKCP18 strands have been used, and also whether other sources of climate projections have been used.

UKCP18 uses emissions scenarios linked to the Representative Concentration Pathways (RCPs), with an important feature being the representation of uncertainties in carbon cycle feedbacks. This contrast with the CMIP5 ensemble, in which all models are driven by the same concentration pathway for CO₂ and other greenhouse gases (see Annex 1 section A1.4 for discussion of the important difference between the RCPs used as Concentration Pathways and Emissions Scenarios).

The probabilistic projections considered four RCP scenarios, ranging from RCP2.6 (which is consistent with extensive mitigation of emissions) through to RCP8.5 (which has future emissions considerably higher than pathways considered consistent with current worldwide energy policies). The intermediate scenarios RCP4.5 and RCP6.0 were also included: these are within the range of possible emissions futures considered consistent with current worldwide policies. In CCRA3, RCP6.0 is used to define the higher climate change scenario used for the risk assessment. Many, but not all, projections with RCP8.5 are considered as low-likelihood, high-impact outcomes and not included in the main assessment. Details of the use of the different emissions scenarios and concentration pathways are given in the Introduction chapter (Betts and Brown, 2021), and Chapter 2 (Watkiss and Betts, 2021).

Comparing the UKCP18 global probabilistic projections at 2081-2100 relative to 1850-1900 with those from CMIP5 for the RCP scenarios, UKCP18 projects ranges of global warming which are systematically higher than those of CMIP5 (Figure 1.2). A major factor contributing to this difference is that the CMIP5 projections were driven by CO₂ concentrations from the standard RCP pathways, whereas UKCP18 used emissions scenarios associated with the RCPs, accounting for uncertainties in climate-carbon cycle feedbacks. For each emissions scenario, UKCP18 therefore represented the response to a range of CO₂ concentration pathways, most of which were higher than the standard pathways (Murphy et al., 2018). Further details are given in Annex 1 section A1.4.

Due to the computational cost of the new high resolution regional simulations for UKCP18 (see Annex 1), the UK regional and local climate change scenarios were only produced using a single emission scenario to allow for the largest possible ensemble size to be utilized, in order to cover a wide range of regional climate responses. The highest emissions scenario, RCP8.5, was chosen so that the widest range of future levels of global warming could be explored, including the most extreme climate changes considered as low-probability, high-impact scenarios. This means that, unlike the probabilistic projections, the HadGEM3 global projections and associated regional and local projections are not available for RCP6.0, RCP4.5 and RCP2.6 which project slower rates of warming than RCP8.5. Nevertheless, for many climate impact drivers, the projected regional changes at particular levels of global warming can be considered to be representative of the same level of global warming reached at a later date with a lower emissions scenario and/or as a result of a lower climate sensitivity.

CCRA3 is framed in terms of trajectories of global warming rather than emissions scenarios. Research commissioned on some of the risks therefore used selected components of the UKCP18 projections representing global warming of approximately 2°C and 4°C at the end of the century. The assessment of other risks draws on literature using other models, projections and scenarios that give approximately 2°C and 4°C global warming by 2100.

1.2.2 Improved regional dynamics for the simulation of UK weather and climate in the new UKCP18 projections

Hazardous weather with the potential to cause harm, such as floods and heatwaves, have always occurred due to weather and climate variability, but their frequency and / or magnitude can be affected by anthropogenic climate change. Although climate change is often quantified in terms of global mean surface temperature (GMST), the impactful changes at local scales depend on complex responses of the climate system to greenhouse gas increases, and the interactions between these responses and the processes of climate variability (Sutton et al., 2015). GMST primarily provides information about the level of aggregated global risks from climate change and is the main metric for efforts to reduce global emissions. On the other hand, understanding and quantifying natural variability and the complex response of the global circulation to anthropogenic warming is essential for projecting climatic impact drivers on regional to local scales.

The UK’s weather and climate, especially precipitation, are dominated by natural variability and will continue to be so. In this regard, the position and variability of the North Atlantic Jetstream is of fundamental importance for determining the weather and climate of the UK, and its future behaviour will define many of our future climatic impact drivers.

The behaviour of the North Atlantic Jetstream is particularly complicated, compared, say, with the Pacific Jet. In winter, it has three preferred positions with each position corresponding to specific weather regimes – to the north of the UK (European Blocking), over the UK (positive North Atlantic Oscillation) and to the south of the UK (negative North Atlantic Oscillation). These regimes describe the tracks of storms and the development of blocking episodes, and essentially determine the frequency and/or intensity of windstorms, atmospheric rivers^[2] and extreme frontal rainfall.

Consequently, a focus of this chapter is on how the North Atlantic Jetstream may behave in the future and what this means for the key weather regimes that define the UK’s winter climate. To do this, it is important that global climate models can capture the three positions of the Jetstream described above, and to do it for the right reasons.

Since CCRA2, the Met Office has made significant advances in global and regional modelling for weather and climate prediction. A new climate model, HadGEM3, has been developed which features significant increases in horizontal and vertical resolution in the atmosphere and ocean, as well as improvements in model physics (Williams et al., 2018). The atmosphere model resolution increased from 150km in the horizontal, as used for CMIP5 and IPCC AR5, to 60km, and from 38 to 85 levels in the vertical. The ocean resolution increased from 1⁰to 0.25⁰ and from 40 to 75 levels. These enhancements in resolution have delivered significant improvements in the structure of weather systems and ocean circulation, giving notable reductions in a number of key systematic model biases (Annex 1 and Figure A1.1 from Murphy et al., 2018).

Of particular relevance to CCRA3, is the improvement in the simulation of the position and variability of the North Atlantic Jetstream in HadGEM3, and hence the weather systems that affect the UK (Figure 1.3). Figure 1.3 (a-c) shows the density of the tracks of winter storms over the North Atlantic for 1981 to 2000 from the observations (ECMWF Reanalyses) and from the ensemble means of the HadGEM3 (GC30.5-PPE) and CMIP5 simulations; Figure 1.3(d) provides a summary of the seasonal errors in simulated track density.

HadGEM3 reproduces observed winter storm track density quite well in general, whereas the CMIP5 models tend to underestimate the maximum south of Greenland and the observed extension to the north-east of Iceland is missing. Of particular relevance to CCRA3, the number of storms tracking across the UK and into Europe is overestimated in the CMIP5 models. Seasonal error statistics (Figure 1.3(d)) typically show a larger spread of errors for the CMIP5 models, and several CMIP5 members score worse than any of the HadGEM3 members, in each of the four seasons. One CMIP5 model shows root mean squared error (RMSE) values considerably larger than the other simulations in winter, spring and autumn, because the storm track is shifted south and is too zonally (east-west) oriented, with too many winter storms moving across the UK and western Europe. Further background information on the Jetstream in HadGEM3 and the CMIP5 models can be found in McSweeney and Bett (2020).

As part of major advances in seasonal to decadal prediction^[3] using HadGEM3 (e.g., Scaife et al., 2014; Smith et al., 2019), there has been significant progress recently in understanding the global drivers of the UK’s climate variability and associated weather regimes. These include the El Nino/Southern Oscillation (ENSO), drivers from the stratosphere such as stratospheric sudden warmings, and the patterns of sea surface temperatures in the North Atlantic. These advances play an important role in assessing how the global effects of climate change on phenomena, such as ENSO, might alter the population of the three positions of the Atlantic Jetstream shown in Figure 1.2 and hence on UK weather regimes. Improvements in predictability give us confidence that HadGEM3 has captured natural climate variability and its drivers more faithfully.

HadGEM3 is part of the comprehensive new set of UK climate projections (UKCP18) and other applications, including global forecasts on timescales of months to a decade ahead. The UKCP18 regional and local projections (see UKCP18 Science Reports – Murphy et al., 2018; Palmer et al., 2018; Kendon et al., 2019), all used HadGEM3, with the boundary conditions^[4] coming from variants of that model. As documented in detail in Annex 1, HadGEM3 brings some important benefits to our assessment of future climatic impact drivers for the UK.

In summary, HadGEM3 has delivered significant improvements in simulating the observed variability of the North Atlantic Jetstream in terms of both its latitudinal position and the temporal frequency of each preferred location. This means that UKCP18 results based on HadGEM3 are likely to provide a more reliable evidence base for changes in the UK’s weather and climate, many of which will depend on how natural climate variability will change and not just on the overall warming.

1.2.3 Climate sensitivity in the new UKCP18 projections

The CMIP5 models already cover a wide range of climate sensitivities. However, the climate sensitivity of new Met Office global model, HadGEM3, submitted to CMIP6 lies outside the upper range of the CMIP 5 models. HadGEM3’s Equilibrium Climate Sensitivity (ECS) of 5.4⁰C from preindustrial levels for a doubling of atmospheric CO₂ concentrations can be compared with an ECS of 4.6⁰C for the earlier version of the Met Office model used in CMIP5. This is higher than in all climate models from previous generations used in the CMIP3 and CMIP5 intercomparison projects (Figure 1.4), but consistent with a sub-set of other recently developed climate models submitted to CMIP6 (Andrews et al., 2019; Forster et al., 2020; Meehl et al., 2020). The causes of the higher ECS in these new climate models are currently being studied in detail, with cloud feedbacks and cloud-aerosol interactions in models with prognostic aerosol schemes seeming to be playing an important role. Zelinka et al. (2020) showed that this increase in ECS was primarily associated with stronger positive cloud feedbacks from decreasing extratropical low cloud coverage and albedo in this sub-set of models compared to the corresponding models in the previous generation.

As increased climate sensitivity increases projected global warming under future emissions scenarios, establishing the plausibility of these higher sensitivity models is imperative given the potential implications for climate risk assessments. In 2015, the World Climate Research Programme (WCRP) commissioned a major international study to explore whether it is possible to constrain the estimates of ECS using the latest evidence including (i) feedback process understanding, (ii) the historical climate record, and (iii) the paleoclimate record. A summary of the results of this comprehensive study (Sherwood et al., 2020) is shown in Figure 1.5.

The most important result is that it is impossible to reconcile sensitivities less than 2°C from these three strands of evidence; indeed, this new study suggests that the “likely range” (>66% probability range) has narrowed to, at most, 2.3°C to 4.5°C – or possibly an even narrower range of 2.6°C to 3.9°C when all lines of evidence are considered. The lower end of this range is increased substantially from the 1.5°C lower bound in IPCC AR5, meaning that scientists are now much more confident that global warming will not be small.

In addition, according to Sherwood et al. (2020), there is up to an 18% chance that ECS is above 4.5°C, but no more than a 5% chance that is it above 5.7°C. So, while HadGEM3 is at the high end of these estimates, its ECS cannot be eliminated by other lines of evidence.

Another way to assess the plausibility of a model’s climate sensitivity is to test its skill in weather forecast mode, as advocated by Rodwell and Palmer (2007). This is not widely attempted because very few climate models are also run as weather forecasting models, with the UK Met Office being unusual in this respect. Williams et al. (2020) describe tests of HadGEM3 in weather forecast mode, in which they investigated the validity of the model changes responsible for increasing the climate sensitivity. The results showed that these model changes improved the short-range weather forecast and reduced the error growth over the first few hours of the forecast. This suggests that the physical processes represented by these model changes may be a more accurate representation of the real world, and so it is not possible to dismiss completely the high ECS of HadGEM3 as a plausible possibility.

ECS is an idealized quantity that reflects the very long-term (150 years plus) response of the Earth System to a constant forcing of double CO₂. Transient Climate Response (TCR; warming at the time of CO₂ doubling in an idealized 1% per year increase in atmospheric concentration scenario) is a better measure of warming over the near- to medium-term and therefore more relevant to climate adaptation.

Figure 1.4 highlights that although the upper end of the ECS has increased substantially in CMIP6 compared to CMIP5, the distribution of TCR has not changed as much. HadGEM3’s TCR (2.6°C) is slightly higher than in some CMIP6 models, but only up from 2.5°C for the corresponding model in CMIP5 (Meehl et al., 2020). Again HadGEM3’s TCR is in the upper part of the CMIP6 range and therefore means a greater rate of global warming compared to that simulated in CMIP5 models, especially through the latter half of the 21^st century; this affects the rate of warming at UK scales as well. This indicates that, for the latter half of the 21^st century, HadGEM3 will generate UK warming at the upper end of the range from previous assessments and will challenge us on how to evolve the risk assessments from CCRA2 to CCRA3.

It would be unwise though, on the basis of this higher sensitivity, to de-emphasise the HadGEM3-related regional and local climate projections from UKCP18. As has already been shown, HadGEM3 is a more skilful model in terms of the mean climate and its variability, especially for the Euro-Atlantic sector and the weather patterns that affect the UK, and consequently may provide a more robust assessment of future changes in high-impact or extreme weather and climate events that are fundamental for UK adaptation and risk assessments. Furthermore, the effects of this high climate sensitivity can be minimised if we consider risks at specific global warming levels rather than time horizons through the 21^st century. Consequently, we have used both approaches in the following assessment of future climate impact drivers. Generally the HadGEM3 temperature trajectory should be interpreted as a high impact, low probability future but any temperature-level analysis should be considered to be unaffected by climate sensitivity. In this chapter we note when conclusions are affected by the model’s high climate sensitivity but generally focus on the robust conclusions that would remain true if the model were to have a lower climate sensitivity.

1.3 Climate change that has already occurred

This section summarises the latest evidence that has been accumulated since CCRA2 on the climate change that has been observed across the UK and globally. Each year the Met Office publishes its annual ‘State of the UK Climate’ (e.g., Kendon et al., 2020), which provides a comprehensive analysis of the latest observational records. Since CCRA2, several UK climate records have been extended further into the past through data archaeology, recovery and digitising of past weather records. UK-wide temperatures now go back to 1880 and rainfall to 1862, both on a 1 km grid (HadUK-Grid; Hollis et al., 2019), and there will be further progress in reconstructing historical weather and climate records for CCRA4. There is also more information on finer timescales (daily and even sub-daily), which is enabling much greater understanding of past extremes.

The UK record can be placed within the context of changes in global and European climate, which are now routinely documented in the annual WMO Statement on the State of the Global Climate (WMO, 2019), the international annual assessments by the American Meteorological Society (Blunden and Arndt 2019) and the European State of the Climate (ESOTC) annual report compiled by the EU Copernicus Climate Change Service (ESOTC, 2019).

1.3.1 Temperature and Heatwaves

The Earth has continued to warm as measured by the GMST (Figure 1.6), with 2020 the warmest or second warmest year on record. Furthermore 2010 – 2019 concludes the warmest ‘cardinal’ decade globally (spanning those years ending 0-9) in records that stretch back to the mid-19th century, with the last 6 years being the warmest six years over the whole observed record. Other components of the climate system also show increasing evidence of anthropogenic warming such as declines in Arctic sea ice and glacier mass, rising sea levels, increases in atmospheric humidity, more warm days and fewer cold days (Blunden and Arndt, 2019).

The same evidence for continued warming is also seen in the UK land temperature record (Figure 1.7), although due to the very cold winter of 2010, the last decade has only been the second warmest of the ‘cardinal’ decades over the last 100 years of UK weather records, slightly behind the 2000s. Based on the Central England Temperature record (the longest instrumental temperature record in the world), the 21^st century has so far been warmer overall than any 20-year period in the previous three centuries. Around the UK, coastal waters continue to warm, at rates very similar to UK land temperatures in Figure 1.7. For the most recent decade coastal waters have been 0.3^oC warmer than the 1981-2010 average and 0.6^oC warmer than the 1961-1990 average.

It is also notable how many of the UK’s record extreme monthly temperatures have been set in the most recent decade (Figure 1.8), and how many more of them are reflecting high, rather than low, temperature extremes, again a consequence of the UK’s warming climate. Furthermore, as reported by Kennedy-Asser et al. (2020), UK summer temperature extremes (expressed as the 95^th percentile) are warming 15-48 % faster than the UK summer mean and >50 % faster than the global mean annual temperature. In July 2019, the UK recorded its highest daily maximum temperature of 38.7^oC in Cambridge.

The continuing warming of the UK’s climate across all seasons and nations is also reflected in other climate metrics that influence, for example, energy demand and agricultural production, including increases in the number of cooling degree days and growing degree days, and declines in the number of heating days and frost days (Kendon M. et al., 2020).

A recent study of heatwaves (McCarthy et al., 2019a), using the new HadUK-Grid daily dataset (Hollis et al., 2019), provides important information on their spatial and temporal characteristics since 1961. In this case, a heatwave is defined using the new metric described in McCarthy et al. (2019a). A UK heatwave is declared when a location records a period of at least three consecutive days with maximum temperatures meeting or exceeding a heatwave temperature threshold, in which the threshold varies by UK county in the range 25–28°C.

Using this new definition, Figure 1.9 shows the frequency of heatwaves across the UK between 1961 and 2018, expressed as the percentage of years in which a heatwave is declared. Across the southern half of the UK, 30–50% of years have experienced at least one heatwave period.

The duration and frequency of individual heatwaves is shown in Figure 1.10(a) for major metropolitan areas of the UK. It shows that durations in excess of 1 week are quite common and those in excess of 2 weeks can account for around 10% of all heatwaves. Figure 1.10(b) shows the timeseries of heatwave duration for the major cities of London and Glasgow. The extreme years of 1976, 1995, 2006 and 2018 are clearly seen, along with a tendency for more heatwaves in London in recent years.

1.3.2 Sea Level Rise

Sea level is an important climatic impact driver for the UK, causing inundation in low-lying coastal areas, exacerbating coastal erosion, increasing tidal locking^[5] in some rivers, and making storm surges more damaging. The latest assessments show that sea level continues to rise (Figure 1.11). Since 1901, UK sea level has risen by 1.4 ± 0.2mm/year when excluding the effect of vertical land movement, in line with the global figure of 1.7 ± 0.2mm/year. This rise is not uniform around the coast due a number of large-scale atmosphere and ocean processes.

Only a limited number of stations are used to construct the long-term record. A more comprehensive reconstruction by Hogarth et al. (2020) for the period 1958-2018, has shown that in recent decades the mean rate of sea level rise may be higher at 2.39 ± 0.27 mm yr⁻¹ and that this rate is accelerating by 0.058 ± 0.030 mm yr⁻². This is in line with the IPCC Special Report on the Ocean and Cryosphere in a Changing Climate (IPCC SROCC, 2019), which reported that the rate of rise in global mean sea level for 2006–2015 of 3.6 mm yr^–1 is unprecedented over the last century.

1.3.3 Sunshine

Temperature and sea level rise provide the most compelling evidence of a changing climate for the UK; changes in other weather and climate metrics are more difficult to detect due to their inherent large natural variability. However, a clear trend is emerging for increasing sunshine hours for all parts of the UK and especially during winter and spring, where the most recent decade has been higher by 11% and 16% respectively, compared with 1961-1990 (Figure 1.12). The run of sunny springs in recent years is particularly notable, with 2020 being the sunniest spring on record for all UK countries in series stretching back to 1929.

1.3.4 Mean Rainfall

Annual mean rainfall (Figure 1.13) is dominated by natural variations, although there are indications of small increases over the UK and its nations since the 1970s, especially for Scotland.

Seasonal UK rainfall has always been dominated by interannual and decadal variability (Figure 1.14). There is some evidence of increased winter rainfall in recent decades and some notable extremes (winter 2014 is the wettest winter in this series and 2016 ranked eighth wettest). Any trend towards drier summers is less evident, with a recent spell of wet summers between 2007 and 2012.

For many impacts requiring adaptation, it is the amplitude of the change relative to the local amplitude of climate variability which is more relevant. A new analysis of the UK rainfall data for 1862-2017 by Hawkins et al. (2020) has recently shed some light on the detection of climate change in rainfall using the HadUK-Grid dataset. By expressing the signal (S) as the % change in rainfall per 1^oC change in the global mean surface temperature, they were able to detect regional variations in rainfall trends associated with global warming across the UK (Figure 1.15). When that trend is removed from the timeseries, the remaining noise (N) due to natural variability can be quantified and the ratio between signal and noise indicates if and where the effects of climate change are beginning to emerge from natural variability.

Figure 1.15(a) shows that the signal (S) for trends in annual mean rainfall associated with global warming is dominated by increases, mainly over the west side of the UK and especially over Scotland. Hawkins et al. (2020) further showed where the signal exceeded the noise (S/N), identifying Scotland and the mountainous regions of western England (Figure 1.15(b)). Where S/N exceeds 1, this can be interpreted as places where the climate is now moving from the ‘familiar’ towards being ‘unusual’, relative to lived experience, using the terminology as defined by Frame et al. (2017).

In all regions, the influence of climate change is to increase rainfall, consistent with fundamental physics that says that warmer air holds more moisture (i.e., for 1^oC rise in temperature, atmospheric moisture content – and by inference regional rainfall – increases by 7%). Figure 1.15(b) clearly identifies Scotland and the mountainous regions of western England where S/N exceeds 1 and therefore where the influence of climate change is already emerging and potentially challenging our resilience.

1.3.5 Rainfall extremes

While robust evidence for trends in the UK’s annual and seasonal rainfall is emerging year on year, we need also to consider whether the same is true for extreme rainfall which can be particularly damaging.

There is a growing body of literature (e.g., Guerreiro et al., 2018) which argues that the effects of global warming on the hydrological cycle are being manifested in changes in the frequency and intensity distributions of daily and sub-daily rainfall, even when averages over longer timescales are stable.

There is some evidence for the increasing occurrence of widespread heavy daily rainfall across the UK in the last few decades (Figure 1.16), such as autumn 2000 and winter 2013/14. Although the record is too short to be conclusive, this is in line with fundamental physics that says that warmer air holds more moisture. In other words, a weather system today would give more rainfall than the same system in 1950. The statistics on local extreme daily rainfall (Figure 1.16), based on individual station records, also suggest an increasing occurrence, although the gauge network is not ideally suited to detect localized events.

In a major review of the latest evidence on the current anthropogenic intensification of short-duration rainfall extremes, Fowler et al. (2020) conclude that:

Heavy rainfall extremes are intensifying with warming at a rate generally consistent with the increase in atmospheric moisture (i.e., 7% per ^oC), for accumulation periods from hours to days.
In some regions, stronger increases in short-duration, sub-daily, extreme rainfall intensities have been identified, up to twice what would be expected from atmospheric moisture increases alone.
Stronger local increases in short-duration extreme rainfall intensities are related to convective cloud feedbacks involving local storm dynamics.
The evidence is unclear whether storm size has increased or decreased with warming; however, increases in rainfall intensity and the spatial footprint of the storm can compound to give significant increases in the total rainfall during an event.
Evidence is emerging that sub-daily rainfall intensification is related to an intensification of local flash flooding. This will have serious implications for flood risk management and requires urgent climate-change adaptation measures.

1.3.6 Storminess and winds

Storms are an important climatic impact driver for the UK and there has been considerable debate on whether storminess is increasing, particularly after the 2013/14 winter, which was the stormiest winter for at least 143 years when cyclone frequency and intensity are considered together (Matthews et al. 2014).

A comprehensive review of the evidence has been presented by Feser et al. (2015) who concluded that there is as yet no clear evidence for increased storminess. This was confirmed by a more recent study by Krueger et al. (2019). Both studies showed that trends in storm activity depend critically on the time period analysed, and that the apparent increase in storminess between 1960 and 1990 is actually part of a longer-term record that reveals multi-decadal variability. In other words, large-scale natural climate variability, such as the North Atlantic Oscillation, dominates the intensity and frequency of UK storms.

Kendon M. et al. (2020) confirmed this result when they analysed strong wind gusts across the UK (Figure 1.17), which showed increases in wind gusts over 40 knots in the latter part of the 20^th century and a decline thereafter. Nevertheless, more research is needed to address this important question for the UK, since major storms can cause widespread damage, from flooding, winds and waves to coastal storm surges.

1.3.7 Summary

The latest observations show that the UK continues to warm, sea levels continue to rise, and an increasing number of climatic impact drivers are beginning to show clearer evidence of a changing climate. Beyond the mean climate there is a growing body of evidence for changes in the frequency and/or intensity of high-impact weather events, such as extreme daily rainfall and heatwaves. New records are being set more frequently, with the UK experiencing unprecedented high temperatures and heavy rainfall. However, the evidence base for changes in storminess is weak and this needs to be addressed with some urgency. The next section will provide the latest evidence on how to interpret these events and the extent to which they can be attributed to climate change.

1.4 Interpreting the observational evidence on extremes

The previous section has identified a number of climatic impact drivers where climate change can be detected, in other words the observations lie outside the envelope of natural variability. The next question is whether those changes can be attributed to anthropogenic forcing or whether they are due to other drivers such as multi-decadal variability. We are now confident that the trends in UK average surface temperature can be attributed to anthropogenic global warming, but other climatic impact drivers, especially related to precipitation are still challenging.

We know that extreme weather and climate events can be very costly, and it is increasingly important to know whether human influences are affecting their severity and frequency so that we can plan accordingly. Knowing the relative contribution of climate change to these events allows us to assess how our risk envelope is changing and what our near-term adaptive responses should be.

Since CCRA2, the science of extreme event attribution has continued to develop (e.g., Stott et al., 2016, Vautard et al., 2019), including progress towards an operational attribution service (e.g., https://www.worldweatherattribution.org). There are two possible approaches to attribution of extremes. The first is framed around the probabilities of such an event occurring in a world with and without raised concentrations of greenhouse gases due to human activity. It relies on model simulations and is therefore dependent on the skill of the model in capturing these extreme events (e.g., Stott et al., 2004). The second approach is framed around a storyline, which examines the role of the various factors contributing to a specific event as it unfolded, including the anomalous aspects of the meteorology. By analyzing the contribution of the particular weather pattern to the event, it is possible to isolate the potential contribution from climate change.

A good example of these two approaches is seen in the recent analysis of the causes of the UK summer 2018 heatwave (McCarthy et al., 2019b). Based on the probabilistic approach using an ensemble of global climate simulations, they concluded that climate change had increased the likelihood of the summer 2018 heatwave by a factor of 30, assuming of course that the simulations are able to represent the weather patterns that give rise to UK heatwaves with some fidelity.

However, these estimates do not include the specific context of each heatwave and we know that each one is set up differently and framing the question around an event-based storyline approach, enables this to be addressed. McCarthy et al. (2019b) showed that most of the high temperature anomalies during the summer 2018 heatwave can be explained by the prevailing circulation anomalies – in particular, a strongly positive summer NAO that raised the sea surface temperatures of UK coastal waters, as well as by other feedbacks, such as the extremely low levels of soil moisture following an extended dry spell during early summer. Importantly, though, McCarthy et al. also note that these factors alone are not sufficient to explain the intensity of the heatwave, which also has an underlying cause related to the warming UK climate.

This study emphasises the importance of understanding whether the specific circulation patterns that give rise to extreme events will become more prevalent or not under climate change. Other studies of recent extremes have shown, for example, that the extremely cold start to the spring of 2018 is much less likely under global warming, although the circulation pattern that gave rise to it may be slightly more likely (Christidis and Stott, 2020).

The Doncaster Floods of Autumn 2019 have yet to be formally attributed to climate change. The summer and autumn were exceptionally wet and with the jet stream positioned anomalously to the south in October, a series of cyclonic systems brought prolonged and persistent rainfall on top of already saturated soils. Likewise, the severe Welsh and Severn floods of February 2020 were preceded by a very wet winter, with the compounding effects of a series of storms, none of which were exceptional. As with the widespread flooding that occurred in 2013/14 and again in 2015/16, the prevailing weather patterns clearly played the key role (e.g., Christidis and Stott, 2015). Nevertheless, the extreme rainfall totals and the severity of the flooding are consistent with the basic premise that a warming world holds more moisture; in other words, the same weather system 50 years ago would have produced less rainfall than today.

One of the issues in understanding extremes for the recent past and present day is that the observational record does not necessarily capture the full range of possible outcomes for extreme weather in a particular location, even under the current climate, simply because such events are by definition rare. For the UK, this is particularly challenging. The natural volatility of our weather means that the observational record is far too short to characterise extreme events with confidence and provide robust estimates of return periods. At the same time the statistics of observational extremes are non-stationary (i.e., they vary with time) due to low frequency, multi-decadal natural variability and the emerging effects of anthropogenic climate change. Empirical methods for analysing extremes based on the limited observational record, tend to assume stationarity and potentially have limitations.

Since CCRA2, weather and climate models have been increasing in the level of skill and granularity. They are now able to provide exciting opportunities to use model simulations to produce large sets of synthetic but meteorologically plausible extreme weather events (orders of magnitude larger than observational records) for the current (and future) climate; in other words, pseudo-observations that act to fill out the extreme ends of the observed distribution. They enable return periods (e.g., 1 in 100 years) to be estimated more robustly than using empirical methods based on the short observational record, provided of course that it can be shown that the model can represent, statistically, the real world (Thompson et al., 2017). Consequently, they may serve to provide an improved baseline understanding of current likelihood of extreme weather events, which is valuable in and for itself; it informs planning today, strengthening resilience, as well as providing a reference against which to assess future changes. Synthetic event sets can also be used to explore correlated extremes (such as simultaneous bread-basket crop failures e.g., Kent et al., 2017) and clustering of extremes (such as European windstorms e.g., Priestley et al., 2018).

This methodology, known as UNSEEN (UNprecedented Simulation of Extremes with Ensembles) was first introduced in the Government’s National Flood Resilience Review (2016) following the severe flooding during the 2015/16 winter and published by Thompson et al. (2017). Figure 1.18 gives an example of the simulated and observed events sets for monthly precipitation in South East England; it shows that by sampling many more meteorologically plausible conditions, a large number of unprecedented extremes can be identified, due to the natural volatility of weather systems. For example, it suggests that the severity of flooding of the Thames in 2014 should not be unexpected, even under present climate conditions, with even more extreme monthly rainfall totals possible.

The size of these simulated event sets is many times larger than observations alone and allows exceedance probabilities to be calculated with much more confidence. For example, it can be concluded from Figure 1.19 that currently, in any year, there is a 10% chance of an unprecedented month’s rainfall, and that furthermore, there is a 1% risk of receiving 20-30% more rainfall than ever observed before, just from natural variability within today’s climate.

The UNSEEN methodology has also been applied to other extremes, such as UK summer heat waves. McCarthy et al. (2019b) considered the 2018 event in which summer average temperatures were close to +2^oC above the 1981–2010 average for a large swathe of southern and central England and Wales. Using a simulated event set of 4720 samples, McCarthy et al. (2019b) showed that there is an 11% chance in any current year of summer temperatures exceeding those in 2018, and a 1% chance that temperatures anomalies may exceed 1^oC or more above the 2018 values.

UNSEEN complements future projections from UKCP18 by providing a valuable new tool for assessing current and near-term climate risks by providing better estimates of the tails of the observed distribution for the current climate and providing bounds on what is meteorologically plausible in terms of extreme events. However, like the attribution studies documented above, UNSEEN relies on the model’s ability to represent accurately the statistics of the real world as far as that is possible from the limited observational record, and so detailed evaluation of the model is an essential first step.

In summary, it is now possible to attribute some changes in UK weather extremes to climate change, and significant progress has continues to be made in the attribution of extreme weather events since CCRA2. The science remains challenging because of the UK’s highly variable weather and the fact that these events are, by definition, rare. New research has emphasised the importance of understanding the meteorological context of each individual event, such as the prevailing weather patterns and the antecedent conditions, as well as recognising that unprecedented extremes will continue to occur just because of natural variability. It remains the case that recent extremes can be largely explained by the prevailing atmospheric circulation anomalies; however, these circulation changes alone are not necessarily sufficient to explain the intensity of the event, which may also have an underlying contribution from the warming UK climate.

1.5 Future Climate Change

This Section draws on the latest UK Climate Projections – UKCP18 – which provide the most up-to-date and comprehensive assessment of future climatic impact drivers for the UK, along with projections of global climate change consistent with the UK projections. It will highlight key differences from the evidence base used in CCRA2 where appropriate. Studies using other climate projections were also used in CCRA3, in addition to information based on UKCP18 – further details on the integration of information from different sources are given in Chapter 2.

1.5.1 Projected future global warming

Future changes in global average temperature will depend on future human-caused emissions of greenhouse gases and on the response of the global climate system to these emissions (see Chapter 0). Any particular pathway of future global warming could therefore arise from a number of different combinations of future emissions and climate system responses.

UKCP18 provides global probabilistic projections of the percentage likelihood of different levels of global average temperature change resulting from four emissions scenarios. Two of these – the RCP2.6 and RCP6.0 emissions scenarios – are broadly representative of the pathways to approximately 2°C and 4°C above preindustrial levels by 2100, as used here in CCRA3 to frame the risk assessment (see Introduction Chapter: Betts and Brown, 2021). As Figure 1.20 shows, with the RCP6.0 emissions scenario, the projections give approximately a 30% probability of global warming exceeding 4°C, relative to 1850-1900 (an approximation of the climatic conditions of preindustrial levels) by 2100; with the RCP2.6 emissions scenario, there is slightly more than a 50% probability of global warming at 2100 being below 2°C, relative to 1850-1900. These probabilities reflect uncertainties in both transient climate response and the strength of carbon cycle feedbacks. Importantly, the probabilistic projections are not based on the high-sensitivity model HadGEM3 – they are built from perturbed-parameter ensembles of the HadCM3 model (as used in UKCP09), and also include information from the CMIP5 multi-model ensemble.

UKCP18 also includes projections with the high-end RCP8.5 scenario, especially for the regional and local projections (see Section 1.2.1). Some of these simulations project warming below the 95^th percentile of that projected with the RCP6.0 emissions scenario, but most are above that. UKCP18 further includes projections with the RCP4.5 emissions scenario; this gives warming that overlaps the lower and upper parts of the ranges with RCP6.0 and RCP2.6 emissions respectively. Further information on this, and comparison with other global warming projections including those used in the UKCP09 projections, and the CMIP5 projections used in the IPCC 5^th Assessment Report, is given in Annex 1.

1.5.2 Projected changes in the UK’s annual and seasonal average climate

UK climate change depends on emissions scenarios and global-scale climate responses, and also on the nature of regional climate responses to the global-scale forcing. Overall, the UK is projected to experience ongoing increases in temperature until the middle of the 21^st Century under all scenarios examined by UKCP18, including RCP2.6. Until that point, the magnitude of regional climate change depends more on the regional climate responses to a given amount of global average warming, than the difference between the emissions scenarios (Figure 1.21).

Beyond mid-century, the magnitude of UK climate change depends on the path of future emissions as well as on the regional and global climate response (Figure 1.21). The spread in the projected temperatures by 2100 is larger for RCP6.0 than RCP2.6 and reflects, in part, the increasing contribution to uncertainty from carbon cycle feedbacks, which come into play more substantially in the latter half of the century and at higher temperatures. The warming experienced in the UK is projected to be greater in the summer than the winter (Murphy et al., 2018).

As in earlier assessments there are significant differences in the precipitation signal between winter and summer (Figure 1.22). Even for the RCP2.6 pathway, winters are projected to become wetter overall, whereas summers are expected to become drier. The spread of probabilities is broader in summer than winter, although the difference between the two scenarios is small throughout the century.

A key point is that while drier summers are generally more likely across the UK, wetter summers are also possible – for most of the country there is a 10% probability of 20-year average summer precipitation increasing by at least 10%. Drier summers are therefore far from certain in the near term. For winter, however, the lower part of the likelihood range remains constant; at any time in the projections, it is about 10% likely that the 20-year mean precipitation would be more than 20% below that for 1981-2000. So, there is no projected change in the likelihood of dry winter conditions when considering the 20-year mean.

A key question is how the HadGEM3-based projections compare with the UKCP18 probabilistic projections for temperature and precipitation shown in Figures 1.21 and 1.22. A comparison of the projections for 2061-2080 is shown in Figure 1.23. Here the RCP8.5 scenario is used because the projections with HadGEM3 and its related regional simulations were only performed for this pathway.

The new probabilistic projections are given by the black box and whisker diagrams; these include CMIP5 (blue dots) and the new ensemble of HadGEM3 (orange dots). Both CMIP5 and HadGEM3 were used to construct the CCRA3 probabilistic projections. The new 12km regional climate model (RCM) and 2.2km convective permitting model results are shown in the pink and green dots respectively; they are both driven by boundary conditions from the HadGEM3 ensemble (see Annex 1 for more details).

Diagram

Description automatically generated

The introduction of the convective permitting model is particularly important. With kilometre-scale resolution, the convective permitting model not only represents the landscape more accurately, but also captures the fundamental physics of thunderstorms and of embedded convection within weather fronts that is missing in lower resolution models (Kendon et al.,. 2014). Consequently the representation of extreme sub-daily rainfall and other local extremes, such as wind gusts and high temperatures, has been transformed compared with previous CCRAs.

In summer, the regional projections are substantially different from the probabilistic projections which incorporate information from CMIP5. HadGEM3 projects hotter and drier summers, a signal that is carried through into the RCM and the convective permitting model results (Figure 1.23(a)). Future increases in temperature over both Scotland and England are projected to be 1-2^oC greater than the CMIP5 models. This is associated in part with HadGEM3’s higher climate sensitivity and partly with the use of higher CO₂ concentrations in HadGEM3 than CMIP5, because the HadGEM3 projections account for uncertainties in climate-carbon cycle feedbacks (see Annex 1, Figure A1.3 and section A1.4). For England, there is also a very strong signal for much reduced rainfall in summer (Figure 1.23(b)), which itself will also act to elevate summer temperatures. Some of the summer heating may be due to HadGEM3’s higher climate sensitivity and higher CO₂ concentrations, but some is clearly associated with changes in the summer atmospheric circulation in HadGEM3 (see Section 1.5.4).

To provide more regional detail, Figure 1.24 compares future changes in the mean summer rainfall from the RCM and the convective permitting model . They are fairly similar which reflects the importance of the driving boundary conditions from HadGEM3. Figure 1.24 also compares these with the 10^th to 90^th percentile ranges in the probabilistic projections and shows that this range is larger than those in the RCM and the convective permitting model. While most project decreased precipitation, the probabilistic projections include the possibility of small increases in average summer precipitation over larger areas of the country than in the RCM and the convective permitting model ensembles.

Map

Description automatically generated

In winter, as shown in Figure 1.23, the results from the various models lie mostly within the range of the probabilistic projections. However, the convective permitting model results show a shift to bigger increases in average precipitation in winter (Figure 1.23) which is countrywide (Figure 1.25).

Kendon et al. (2020) have linked this increase in winter precipitation in the convective permitting model to the advection over land of convective showers initiated over the sea, a process that lower resolution models do not capture. As in observations, these showers penetrate much further inland in the convective permitting model, bringing more rainfall to the eastern side of the country.

This signal is part of a fundamental difference between the RCM and the the convective permitting model in the nature of winter precipitation. A comparison of hourly precipitation diagnostics with a new observational dataset of hourly precipitation amounts (CEH-GEAR1hr, Lewis et al., 2018) for the current climate (Figure 1.26) shows that the RCM has far too many days when it is raining, and that on those days the rainfall intensity is too low. These biases are largely rectified in the the convective permitting model through better representation of the physical processes, as discussed above, giving a simulation that is much closer to observations (Kendon et al., 2020). As a result, the simulation of mean winter precipitation in the the convective permitting model is superior to the RCM for the current climate in a clean test using observed boundary conditions from the ECMWF Reanalyses (Figure 1.27).

Graphical user interface

Description automatically generated

Kendon et al. (2020) go on to show that, under climate change, the convective permitting model gives both more frequent and more intense rainfall in future over land (Figure 1.28). The frequency increases are considerably larger than in the RCM, with a large part of the convective permitting model’s future increases coming from a higher frequency of convective showers. These showers are most likely triggered over the sea (where the warmer ocean and higher levels of atmospheric moisture favour more triggering of convection), and then advected inland, persisting for longer and potentially further development over land. Consequently, as Kendon et al. (2020) explain, these showers are an important contributor to the higher winter precipitation response in the convective permitting model; changes in the mean precipitation from convective showers contribute about 40% of the overall change in winter over land in the convective permitting model.

These are important results; they suggest that previous CCRAs based on traditional coarser-resolution climate models may underestimate future increases in winter precipitation, especially where wintertime convective showers are a key contributor, since the processes important for the advection and further triggering of showers are only well captured in CPMs. These differences in precipitation frequency and intensity will undoubtedly affect the hydrological response to future rainfall changes, and potentially have implications for river flows, flood risk and water resource management.

Sea level rise is an important climate impact driver for the UK. Since CCRA2, new assessments of the contributions to current and future sea level rise from the major Greenland and Antarctic Ice Sheets have been derived from observations (suggesting accelerating mass loss) and included in the future projections. Additionally, there have been improvements in the methodology for estimating the range of uncertainties in the UK estimates of local sea rise level (Palmer et al., 2018; Palmer et al., 2020).

Figure 1.29 (upper panels) shows the evolution of global sea level rise through the 21^st century in the UKCP18 marine projections, using CMIP5 climate scenarios with three concentration pathways. The contributions to sea level rise from the various components are also shown. As in the observations, the thermal expansion of seawater only contributes around one third of the total, emphasising the importance of understanding the vulnerability of the major ice sheets and how mass loss will evolve.

The new estimates in Figure 1.29 include updated estimates of the contribution from Antarctic ice dynamics, which have led to a substantive change in the projections, especially for the 95^th percentile, indicating an additional 5 – 10cm rise in sea level by 2100 compared with earlier estimates also based on CMIP5 climate projections. However, the processes behind ice sheet collapse particularly for Antarctica remain very uncertain and continued monitoring and process studies are vital. The IPCC Special Report on the Ocean and Cryosphere in a Changing Climate (IPCC, 2019) highlighted that global sea-level rise by 2100 could reach 2m in the most extreme scenarios but viewed this as very unlikely and with low confidence (See Section 1.8 on Earth System Instabilities).

Both the UKCP18 and IPCC AR5 sea-level projections use the CMIP5 climate projections, with UKCP18 using updated methods to project sea level rise resulting from the projected climate changes. It is important to note that the CMIP5 models are driven by the standard RCP projections of concentrations of CO₂ and other greenhouse gases, in contrast with the UKCP18 land projections which are driven by emissions scenarios and account for uncertainties in carbon cycle feedbacks (see Annex 1 A1.4). This means that for any given RCP, the sea level rises in the UKCP18 marine projections are mostly driven by a slower rate of global warming than represented in the UKCP18 land projections (see Figure 1.2).

When subsets of the UKCP18 sea level rise projections consistent with global warming of 2°C and 4°C (± 0.1 °C) in 2100 are extracted, these are found to cover much of the likely range of the projections with the RCP2.6 and RCP8.5 concentration pathways (Figure 1.29 middle panels). Therefore the RCP2.6 and RCP8.5 sea level projections can be considered to be reasonably representative of sea level rise consistent with pathways to global warming of 2°C and 4°C (± 0.1 °C) in 2100. The RCP8.5 central estimate is about 0.1 m higher than that of the sea level projection consistent with 4°C warming in 2100.

The lower panels of Figure 1.29 show the regional variations in median sea level rise around the UK with the familiar pattern of higher changes in the south. These variations occur due to land-based ice and land water mass loadings, changes in the ocean circulation and the ongoing isostatic adjustment to the last glacial maximum.

Beyond rises in mean sea level many coastal impacts are associated with storm surges. The latest evidence suggests that changes in extreme water levels will likely be more driven by changes in mean sea level than changes in surge, although notable changes in surge (+ or -) cannot be ruled out (Palmer et al., 2018).

1.5.3 Moving beyond average climate change: The importance of climate variability on annual timescales

There is far more evidence in UKCP18 regarding the volatility of the UK’s weather and climate going forward, and this provides an improved evidence base for assessing future risks (e.g., Sexton and Harris, 2015). An important new element in UKCP18 is the inclusion of estimates of the envelope of interannual variability; these act to broaden the distribution of probabilities and provide information on changes in the likelihood and intensity of extreme months or seasons going forward.

When interannual variability within any long-term average is included, the spread in the various climatic impact drivers is increased significantly, particularly for variables, such as precipitation, that are highly variable in space and time (Figure 1.30). As Sexton and Harris (2015) note, the spread associated with natural variability is larger for near term climate change, when the forced changes and their inherent uncertainties are smaller. As Figure 1.30 demonstrates, individual summers have high likelihoods of being either wetter or drier than the 20-year average would suggest.

By including the interannual variability in UKCP18 it is possible to look at specific extreme years from the past and to estimate the probability of experiencing such events in the future. Figure 1.31 shows how the probability of experiencing a hot summer like 1976 or 2018 changes with time through the 21^st century for the RCP8.5 pathway. By the middle of the century the probability of a summer as warm or warmer than 1976/2018 has a projected probability of around 50%, while by the end of the century the probability is greater than 90%; in other words, summers like 1976 and 2018 could become commonplace.

In summary, UKCP18 has delivered a new capability to look beyond changes in the mean climate to consider shifts in the envelope of interannual variability. This provides a much richer evidence base for assessing future climatic impact drivers and provides important links between what we experience today and what we will experience in the future.

1.5.4 Moving beyond average climate change: The importance of weather regimes under climate change.

As discussed in Section 1.2, it is important to look beyond simple metrics of average climate change to consider changes in prevailing weather patterns. As outlined in Section 1.2 and documented in Annex 1, HadGEM3 is a more skilful model in representing the North Atlantic Jetstream, its variability and related weather regimes, which may have a strong influence on the UK’s future weather and related impacts. For example, Senior et al. (2016) investigated the impact of increasing resolution in HadGEM3 on climate change simulations and noted that although long-term averages at continental and large scales were largely unchanged, there were important regional impacts, including a greater increase in the frequency of the most intense winter storms at the higher resolution used in UKCP18.

In winter, the UK’s weather is dominated by the phase of the North Atlantic Oscillation (NAO), which gives rise to two distinct Weather Types (WT), describing either blocked, colder and drier winters (WT 1) or warmer, wetter and windier winters (WT 2; Figures 1.32(a) and (b)). One of the striking results from UKCP18 is the difference between the CMIP5 and HadGEM3 ensembles in the population of these two distinct weather regimes through the 21^st century (Figures 1.32(c) and (d)).

A picture containing timeline

Description automatically generated

Both ensembles capture the mean and year-to-year variability of the observations reasonably well, bearing in mind the short observational record and multi-decadal fluctuations in the NAO. However, whereas CMIP5 suggests no change in the population of these weather types, HadGEM3 shows clear trends towards WT 2 and away from WT 1. This implies that future winter weather may be dominated by more mobile, cyclonic weather systems, with fewer blocked winters. This will affect the western parts of the UK, in particular, and may contribute to more substantial increases in daily precipitation with related flooding, as well as a higher incidence of strong winds and waves. As well as the severity of flooding, its special extent is also higher under the types of strong westerly airflow that are expected to become more common (Wilby and Quinn, 2013).

The projected shift to more mobile, cyclonic winters may also increase the risk of atmospheric river events that bring large amounts of precipitation and are major contributors to severe flooding, particularly for the mountainous regions of the UK (e.g., Lavers et al., 2011). Storm Desmond in 2015 was a notable example, with the classic streamer of moist air being drawn from the Caribbean, carried by a strong Jetstream, and impinging on the mountains of Lake District.

Matthews et al. (2018) argue that the long-term warming of the North Atlantic may have increased the chance of such an atmospheric river event. Similarly, a recent study by Payne et al. (2020), shows that climate change is likely to increase the frequency of atmospheric river events which, when combined with increased atmospheric moisture due to warming, may make them even more severe, increasing the risk of serious flooding and landslides.

This is important new evidence. Even though average annual trends suggest minimal differences in mean winter precipitation between CMIP5 and HadGEM3, changes in the population of these weather regimes may lead to different impacts associated with storms and heavy rainfall, for example.

As in winter, UK summers are also affected by the population of specific weather regimes. Extreme summer heat, as experienced in 2018, was due mainly to a strongly positive summer NAO (SNAO; e.g. Folland et al., 2009), which is characterised by a high-pressure anomaly over and to the north-east of the UK. The positive phase of the SNAO corresponds to anomalous easterly winds, which bring warm air from continental Europe, as well as more local solar radiation and surface sensible heating. These effects reinforce the temperature response over the UK, resulting in the SNAO being a very important control on UK summer heat.

In addition to influencing summer heat, the SNAO is also an important driver of summer rainfall anomalies. Folland et al. (2009) showed that there is a strong, negative correlation between the SNAO and England-Wales precipitation, with a positive SNAO favouring low summer rainfall. Studies of the future behavior of the SNAO have shown a tendency for more positive phases under a warming climate, which would indicate an increased prevalence of high temperatures and drought in the future. However, there are large discrepancies between the change in the SNAO and its projection on to UK precipitation, between the CMIP5 models and HadGEM3 (Figure 1.33). Whereas some of the CMIP5 models suggest increased rainfall with a shift to more positive SNAO conditions, HadGEM3 shows a clear signal for reduced rainfall, consistent with the observed relationship between the SNAO and rainfall.

The strong reduction in precipitation in HadGEM3 in response to changes in the circulation, may therefore be a contributing factor in generating more extreme high temperatures in future summers compared with previous projections. Although more research is needed to understand the physical mechanisms behind these changes in the SNAO and its projection on to UK precipitation, it is important that CCRA3 considers the possible impacts of these more extreme scenarios.

Heatwaves and drought are not the only hazards related to the projected change in the summer climate. Wildfires are increasingly likely and potentially more severe (Arnell et al. 2021). Poor air quality, linked for example to surface ozone, is exacerbated by high temperatures (Doherty et al., 2013) and by the prevalence of continental air masses which occur during summer blocking episodes (Royal Society, 2008).

The clear message emerging from these studies is that atmospheric circulation patterns (i.e., weather regimes on timescales of days to weeks and modes of climate variability on timescales from months to decades) remain the dominant influence on UK climate impact drivers, now and into the future. Projections of UK climate change therefore need to be underpinned by climate models that have the capability to reproduce these atmospheric patterns, their spatial and temporal characteristics, if they are to tell us how these may change in a warming world.

This is only now beginning to be realised with the latest generation of high-resolution global climate models although many issues still remain. UKCP18 is the first set of UK climate projections to exploit this new generation of high-resolution projections that can begin to answer the question of what the UK’s weather will be like in the future and what this means for our future risks and opportunities.

1.5.5 Moving beyond average climate change: Daily climatic impact drivers and extreme events

Changes in daily weather and related short-timescale extremes are a key component of many climatic impact drivers. They are typically localized and hence challenging for global and even regional models, with their coarse granularity, to provide reliable estimates. Through the innovative use of the 2.2km convection permitting model UKCP18 has delivered some important new insights on local climatic impact drivers, not just confined to extremes, but including the impacts of representing convection on winter mean precipitation changes (see Section 1.5.2).

So far, the focus of the convection permitting model results has been on high temperatures and extreme rainfall, but other information (for example on wind extremes, hail and lightning) will gradually emerge as the convection permitting model projections are analyzed further.

1.5.5.1 Extreme Temperatures

The UK record temperature of 38.7°C set in Cambridge in July 2019 has raised the question of whether exceeding 40 °C is possible. A frequency analysis of exceedances for specific high temperature thresholds in the RCM and convection permitting model ensembles, versus the NCIC observations, is documented in Table 1.1. The results are expressed as the number of counts of threshold exceedances for all points (based on 12km grid) and for all days in a 20-year period, across the UK (upper table) and just for London (lower table). Table 1.1 shows the median value and also the range across the model ensembles; these indicate quite a broad spread in the models which, for the current climate (1981-2000), generally encompasses the observations. It is worth noting that counts of 4871 and 184 for the 32^oC threshold in the present-day simulation of the convection permitting model, for all-UK and London respectively, correspond to only 0.04% and 0.21% of values. Therefore exceedances of these thresholds are very much in the tail of the temperature distribution and hence subject to quite a bit of noise, especially with only a small model ensemble.

Table 1.1 Counts of exceeding certain daytime temperatures for 20-year periods from NCIC observations, the RCM and convection permitting model, for all gridpoints over the UK (1614 points; upper table) and London (12 points; lower table) over all days for the current climate (1981-2000) and for future projections (2021-2040; 2061-2080) using the RCP8.5 scenario. All data have been regridded to 12km RCM grid. Thus, the total number of days is 7200 days for the models (360-day calendar) and 7305 days for the observations, and the maximum number of counts is therefore 11,620,800 (11,790,270 for the observations) for all UK land points, and 86,400 (87,660) for London. The numbers show the median count (in bold), and the low to high estimate (in brackets, corresponding to 2^nd lowest to 2^nd highest member) counts across the ensemble for the convection permitting model and RCM, compared to NCIC observations, for the 20 years of each timeslice. Source: Met Office.
ALL UK
Threshold (^oC)	32	35	38	40
OBS (1981-2000)	2119	108	0	0
CPM (1981-2000)	4871 (1999,8032)	883 (37,1265)	1 (0,208)	0 (0,0)
RCM (1981-2000)	1199 (139,10584)	38 (2,984)	1 (0,7)	0 (0,2)
CPM (2021 – 2040)	26798 (18985,46989)	6435 (3752,15122)	1(416,3736)	232 (22,966)
RCM (2021 – 2040)	14079 (4926,43211)	2936 (732,10964)	347 (1,1868)	21 (0,126)
CPM (2061 – 2080)	182842 (115971,257960)	64343 (32330,83721)	16850 (6317,22595)	5998 (1955,10052)
RCM (2061 – 2080)	137297 (42261,290801)	39434 (7923,96155)	9223 (1461,21335)	2868 (584,7155)
LONDON
Threshold (^oC)	32	35	38	40
OBS (1981-2000)	86	8	0	0
CPM (1981-2000)	184 (104,320)	40 (1,50)	0 (0,10)	0 (0,0)
RCM (1981-2000)	67 (24,230)	6 (0,38)	0 (0,1)	0 (0,0)
CPM (2021 – 2040)	944 (694,1476)	204 (144,486)	51 (22,100)	11 (0,34)
RCM (2021 – 2040)	428 (182,1014)	83 (24,259)	14 (1,49)	1 (0,9)
CPM (2061 – 2080)	4949 (3326,6412)	1754 (1022,2267)	476 (227,713)	141 (38,269)
RCM (2061 – 2080)	3259 (1480,5428)	891 (355,1717)	150 (65,362)	41 (19,124)

The results in Table 1.1 show a clear signal of increasing frequencies of high temperature exceedances through the 21^st century, in accordance with the overall warming. The CPM systematically records higher frequencies than the RCM, even though the mean warming is almost identical (see Figure 1.22). However, the convection permitting model provides a better representation of urban processes and heat-island effects, due to both the high spatial resolution and also the use of the 2-tile MORUSES urban scheme. Importantly, both models show that there is a very small chance of exceeding 40°C by 2040, but that by 2080 the frequency of exceeding 40°C is similar to the frequency of exceeding 32°C today with the RCP8.5 concentration pathway^[6]. Furthermore, the median likelihood of exceeding 40°C by 2080 is three times higher in London than across the whole of the UK (0.16% versus 0.05%).

The potential for the UK to experience daytime temperatures above 40°C has also been demonstrated in an independent study by Christidis et al. (2020) using a different methodology. They addressed the question of high temperature exceedances by using observations to relate local extremes to UK-wide mean extremes, and then applying the resulting relationships to 16 CMIP5 global model projections in a risk-based attribution methodology. This enables them to distinguish between high temperatures due solely to natural variability and those that have a contribution from anthropogenic warming, both for the present day and for the late 21^st century (Figure 1.34).

Application

Description automatically generated

For the present day, the incidence of high temperatures is dominated by natural variability but later in the century human influence dominates. Across the whole of the UK the likelihood, locally, of exceeding 30°C, and even 35°C, increases with time. By 2100 many areas in the north are likely to exceed 30 °C at least once per decade. In the south-east temperatures above 35°C become increasingly common, and temperatures exceeding 40°C also become possible. Summers that experience days above 40°C somewhere in the UK have a return time of 100-300 years at present. This is projected to decrease to 3.5 years by 2100 with the RCP8.5 concentration pathway, which is consistent with a scenario of 4°C global warming at the end of the century (see Chapter 2: Watkiss and Betts, 2021).

Using the UKCP18 RCM projections, Lo et al. (2020) have estimated the 1981–2079 trends in summer urban and rural near-surface air temperatures and in urban heat island (UHI) intensities during day and at night in the 10 most populous built-up areas in England. There are larger upward trends in daytime than nighttime temperature for both urban and rural areas (Figure 1.35), where rural areas are defined as those contiguous to the city.

Their results also show a signal of an urban cool island effect during the day but an urban heat island effect during the night (Figure 1.36). For example, by 2080, London’s ensemble-mean summer nighttime UHI intensity is projected to increase to 2.1^oC, whereas its daytime UHI intensity is projected to decrease slightly to 0.8^oC. These summer daytime urban cool islands are likely to be the result of a phase delay in the increase in upward sensible heat flux in the urban areas during the day because of their large thermal inertia (e.g., Bohnenstengal et al., 2014). This means that cities absorb heat during the day and release it at night. The increased intensity of the UHI at night has implications for the frequency of tropical nights in cities with associated heat stress and health implications.

1.5.5.2 Extreme Precipitation

The average precipitation is a combination of the frequency of rainfall (often termed wet days) and the intensity of the rainfall when it is raining. Even when the average precipitation is unchanged, there can be shifts between frequency and intensity that have important hydrological consequences. Flooding, water resource management and agriculture are all sensitive to the frequency of rain days and how intense the rain is when it falls.

The convection permitting model has proved to be particularly effective in capturing the frequency versus intensity of precipitation (e.g., Kendon et al., 2019) and for simulating extreme daily and sub-daily precipitation (e.g., Kendon et al., 2014). This is because the convection permitting model is able to resolve much more of the physics of rain-bearing systems, as well as resolving the local landscape, especially mountainous terrain, with much greater fidelity than the RCM and global models.

The mean signal of wetter winters is a combination of more wet days (Figure 1.37 top row), as well as an increase in the intensity of rainfall (Figure 1.37 bottom row), which is projected to increase by as much as 25%, particularly in the south-east. The same analysis for summer (Figure 1.38 shows that despite overall summer drying, with wet days projected to become less frequent, the convection permitting model projections nevertheless suggest that when it does rain, the rainfall will be more intense.

Sub-daily and hourly precipitation rates are important drivers of flash flooding events. Figure 1.39 shows future changes in the shape of the hourly precipitation distributions from the RCM and convection permitting model. There is clear evidence of a shift to more intense hourly rainfall at the expense of lighter rainfall in all seasons, in both models, and for all ensemble members. In autumn and winter, changes in the shape of the wet value distribution are very similar between the convection permitting model and RCM. The greatest difference in changes between the models is seen in summer, where there is a much greater increase in the fractional contribution from high intensities in the convection permitting model. This is consistent with Kendon et al. (2014) who showed that in the convection permitting model summer convective storms are strengthened by local dynamical feedbacks.

Chart, line chart

Description automatically generated

Overall, as Figures 1.37 to 1.39 show, there are large changes in the frequency and intensity of daily and hourly precipitation. The hydrological implications of these major shifts in precipitation characteristics for both winter and summer could be profound.

Urban flash flooding is an increasing problem as heavy, sub-daily rainfall events become more frequent. Using the convection permitting model ensemble as an event set, the frequency of precipitation exceeding 30mm/hour has been analysed for some UK cities for present-day (1981-2000) and future (2061-2080) periods with the RCP8.5 emissions scenario (Table 1.2). This methodology, which follows UNSEEN described in Section 1.4, enables more robust estimates of return times for rare events. The results show that currently the return period for such events is typically around 10 years, but the return period decreases to around 5 years by 2080.

Table 1.2 The changing frequency of precipitation exceeding 30mm/hour for 4 cities around the UK expressed as the number of instances per year based on the 12-member ensemble with the convection permitting model. Source: Met Office
Location	Present Day (1/year)	Future (1/year)	Frequency Change
Greater London	0.095	0.139	x 1.5
Edinburgh	0.054	0.099	x 1.8
Belfast	0.136	0.233	x 1.7
Cardiff	0.092	0.191	x 2.1

1.5.6 Summary of Evidence of Future Changes

The new generation of Met Office models that underpin UKCP18 has generated some significant changes to the previous evidence used in CCRA2, such as increased winter rainfall and summer drying, as well as new evidence of physically plausible changes in local extremes on daily and sub-daily timescales. The framing of future climatic impact drivers in terms of changes in weather and climate regimes has been a new feature of the analysis and has emphasized the importance of model skill in representing these regimes.

The main results can be summarized as follows:

The overall message of future winters becoming warmer and wetter still prevails. However, the changes in rainfall are likely to be more extreme than was anticipated in previous CCRAs. The mean signal of wetter winters is a combination of more wet days, as well as an increase in the intensity of daily rainfall, which is projected to increase by as much as 25%, particularly in the south-east. This increase in rainfall intensity potentially implies an enhanced risk of surface water flooding.
Future winter weather may be dominated by more mobile, cyclonic weather systems, with fewer blocked winters than was the case in previous assessments. This will affect the western parts of the UK, in particular, and may contribute to more substantial increases in daily precipitation with related flooding, as well as a higher incidence of strong winds and waves. The projected shift to more mobile, cyclonic winters may also increase the risk of atmospheric river events that bring large amounts of precipitation and are major contributors to severe flooding, particularly for the mountainous regions of the UK.
Future summers are projected to be even hotter and drier than earlier estimates. Reductions in rainfall are substantially larger over England, typically double those used in CCRA2. This can be attributed to improved simulations of summer circulation anomalies as well as higher temperatures. Despite overall summer drying, with wet days projected to become less frequent, the new kilometer-scale projections suggest that when it does rain, the rainfall will be more intense by as much as 20%.
Better representation of the landscape and urban areas have highlighted more frequent and more severe extreme daily high temperatures and urban heat island effects. There is a very small chance of exceeding 40°C by 2040, but by 2080 on a pathway to 4°C global warming at the end of the century, the frequency of exceeding 40°C is similar to the frequency of exceeding 32°C today. Also, night-time urban heat island effects are expected to be more intense, leading to more ‘tropical nights’ in major cities.
Sub-daily and hourly precipitation rates show pronounced shifts to more intense hourly rainfall at the expense of lighter rainfall, than in previous assessments. This has serious hydrological consequences including for flash flooding events.

1.6 An alternative view: UK climate change for specific levels of global warming

As well as assessing climatic impact drivers for specific time horizons, it can also be helpful to assess them at specific levels of global warming, e.g., 2°C above the pre-industrial state. International climate policy discussions and agreements under the UNFCCC, such as the Paris Agreement, currently frame global goals in terms of levels of global warming to be avoided. Moreover, when comparing multiple climate projections from different models with different emissions scenarios, different climate sensitivities and GHG concentrations, the use of specific time horizons can make it difficult to disentangle new evidence on climatic impact drivers from the diversity of warming rates. The use of global warming levels (GWLs) allows a systematic comparison of different sources of evidence on impacts, if the relevant climate quantities scale linearly with global mean warming, as is generally the case for many climate metrics. RCMs suggest that extreme weather metrics may also scale linearly with global mean warming, but whether this is the case for convection permitting models, and especially extreme precipitation, is still to be explored.

Rather than repeating earlier figures, this section will summarise some additional climatic impact drivers at specific global warming thresholds, based on the study by Hanlon et al. (2021a), and the research by Johns (2021), commissioned for CCRA3. These will focus on metrics and threshold exceedances that directly affect the natural environment, agriculture, transport, energy supply, infrastructure and human health. The metrics and threshold exceedances are those used currently for assessing weather impacts; these have been developed by the Natural Hazard Partnership (NHP)^[7] and form part of the National Severe Weather Warning Service (NSWWS)^[8]. They are also part of a designated set of climate variables by the WMO Expert Team on Climate Change Detection and Indices (ETCCDI)^[9]. The definitions of the variables shown in this section are given in Annex 2. This analysis is similar to that reported by Arnell et al. (2020) based on statistical downscaling of the UKCP18 global projections.

Figure 1.40 focuses on cold season impact drivers, specifically frost days, icing days and heating degree days (HDD – see Annex 2 for definition). These are documented for the current climate from the NCIC observations, and from the 21-year averages of regional model climate scenarios centred around specific global warming thresholds – 1.5°C, 2°C, 3°C and 4°C above the average for 1850-1900.

The results show firstly that there is a decline in cold weather metrics even at 1.5^oC warming, compared with the observations. Second, there is very little difference in the various metrics between 1.5^oC and 2^oC, but there are more substantial reductions in cold season impacts at higher warming levels, with icing days almost eliminated for global warming levels of 3^oC and above. Heating Degree Days (HDD)^[10] are typically 50% lower than today at 4^oC global warming.

In summer, the impact metrics revolve around higher temperatures and extended growing seasons. Figure 1.41 shows how the numbers of summer days, tropical nights, Growing Degree Days (GDD)^[11] and Cooling Degree Days (CDD)^[12] increase with global warming (see Annex 2 for definitions). Compared with cold season impacts, there is a greater difference in summer impacts between 1.5^oC and 2^oC global warming. There is a substantial increase in summer days as global warming increases, but outside London, tropical nights only become a serious problem at warming levels approaching 4^oC and tend to be concentrated in the south-east and in urban regions around Manchester. With warmer days and nights, the number of cooling degree days increases quite rapidly even for 1.5^oC and 2^oC global warming, indicating the potential for increased energy demand for air conditioning.

Another approach is to focus on whether the latest results from UKCP18 differ from UKCP09 in ways that are important for assessing climate risks. Johns (2021) documents a comprehensive comparison, and some specific highlights are shown here.

In the context of future renewable energy supplies, changes in near surface winds and in sunshine hours will be important climatic impact drivers, as will changes in peak demand driven for example by clustering of hot days. Figure 1.42 shows projected changes in near surface (10-metre) winds for each season and for each UK nation, from the UKCP09 regional model ensemble (blue), the UKCP18 regional model (pink) and the CMIP5 global model ensemble (green).

Calendar

Description automatically generated

In winter, there is a large spread and no clear signal across the three sources of evidence that would imply robust changes in wind energy supply. In the other seasons there is a consistent, but small, signal of weaker winds, except in summer where the declines are larger, especially in the UKCP18 results. This is consistent with the increased prevalence of summer blocking noted earlier in Section 1.5.4.

The availability of solar energy depends strongly on cloudiness and how this may be affected by changes in the UK’s weather patterns. This is particularly important in summer when the availability of solar energy is at its peak. Figure 1.43 shows a comparison of projected summer changes in the cloudiness (upper panel) and surface shortwave energy (lower panel) for UK nations, from the various sources of evidence available to CCRA3. As well as the UKCP09, UKCP18 and CMIP5 ensembles, Figure 1.43 also shows the new probabilistic global projections (purple), which along with CMIP5 also include HadGEM3 results.

As in previous assessments, there is a decrease in summer cloudiness for all nations, which becomes more marked for higher warming levels. However, UKCP18 shows more substantial reductions in cloudiness for England and Wales, consistent with more dry summer blocking events, as documented in Section 1.5.4. As expected from the greater reductions in cloudiness in UKCP18, the surface shortwave radiation increases for all UK nations, increasing by as much as 30Wm^-2.

Another aspect of electricity supply is its resilience to periods of high demand. As the UK moves to milder winters but hotter summers, some of the peak demand may occur in the summer months associated with prolonged spells of hot weather and an increased demand for air conditioning. One way to assess this risk is to consider the clustering of hot days. Figure 1.44 shows the changes in the maximum number of consecutive hot summer days, defined as days where the maximum temperature exceeds 25^oC. The length of these hot spells increases systematically for all nations as global warming levels increase. For England and Wales hot spells may increase in length by as much as 15-20 days. When these results are compared with the current duration of heatwaves (see Figure 1.10), it suggests that England and Wales may be exposed to very long spells of hot weather.

The results shown in Figures 1.42 to 1.44 may imply that earlier assessments of seasonal changes in the UK’s energy needs are largely unchanged between UKCP09 and UKCP18 except for summer. UKCP18 projects that future summers could be drier and hotter, which could imply increased demand for air conditioning, especially in south-east England (see Cooling Degree Days (CDD) in Figure 1.41).

As the UK warms, drought may become an increasing risk, especially with hotter, drier summers. This has been analysed by Hanlon et al. (2021a) using a simple Drought Severity Index (DSI). The DSI is a rainfall-based drought index expressed in terms of the n-month accumulated precipitation deficit as a percentage of the mean annual rainfall of the location. The DSI has been computed for 3-month, 6-month, 12-month and 36-month periods in order to cover timescales relevant to meteorological, agricultural and hydrological droughts. Meteorological droughts are defined essentially on the basis of short-term rainfall deficiencies, whereas agricultural droughts relate to the gradual depletion of soil water during the growing season, and hydrological droughts are accumulated shortfalls in runoff or aquifer charge over longer periods.

Figure 1.45 shows the DSI for a range of durations from 3 to 36 months across various global warming thresholds, where the drought severity is expressed as the % shortfall in the n-month accumulated precipitation with respect to the climatological annual average for the specific global warming level. The results show that essentially for all drought periods, their severity increases with the warming level. This is particularly the case for the longer period, hydrological droughts where the changes in the accumulated deficits can be substantial. This would imply severe pressures on water resources and the sustainability of agriculture.

In summary, this additional analysis, using the framework of global warming levels rather than time horizons, has demonstrated how climatic impact drivers, directly relevant to specific sectors, are likely to change. It has also highlighted where there are significant differences in the evidence base given by UKCP18 from that used in earlier CCRAs. Key points include:

All parts of the UK will continue to experience a steady reduction in frost days as global warming increases, implying a general trend towards fewer cold weather-related impacts in the long-term average, although some years will still see similar numbers of frost days and cold-related impacts as in recent years.
A significant reduction in the number of icing days across the UK with increases in global warming, with fewer severe cold weather impacts and potentially less transport disruption. Most of the reductions in icing days occur up to a global warming of 3°C with little further reduction from 3 to 4°C.
An increase in the incidence of high summer daytime temperatures throughout the UK. In the future, Scotland and Northern Ireland could start to see high summer temperatures similar to those of England and Wales currently.
A rapid rise in the frequency of ‘summer days’ and ‘tropical nights’. Since most of these events will cluster during the summer months, adaptation to cope with far hotter summers than we are currently experiencing will become important, with cities, especially London, facing the greatest challenge.
A reduction in heating degree days and an increase in cooling degree days is projected for all global warming levels. Over South East England, cooling degree days increase 6-fold for a global warming of 4°C
A clear signal in UKCP18 of decreasing total cloud amount and increasing surface shortwave radiation in summer, especially in southern England, is likely associated with more positive SNAO summers and associated reductions in rainfall and cloudiness. There are also more significant reductions in summer near-surface winds in UKCP18. On the other hand, there is no consistent signal for reductions in windiness in winter.
Meteorological, agricultural and hydrological droughts are expected to become more severe with implications for water resource management.

1.7 Projected climate changes worldwide

The UK is sensitive to climate change beyond its borders and so global projections are helpful for thinking about the international dimensions of climate risk to the UK (such as disrupted food supply chains). Extensive information on projected global patterns of climate change is presented in IPCC Assessment Reports and Special Reports (see https://www.ipcc.ch). While these include comprehensive assessments of uncertainties in regional climate changes, these are typically presented in terms of the average changes from multiple models, along with information on the degree of consensus between models.

For risk assessments, however, it can be useful to provide clear information on ranges of projected changes, hence the presentation style adopted in the UKCP09 and UKCP18 probabilistic projections for the UK (Murphy et al., 2009; Murphy et al., 2018). This can be particularly important for projected changes in precipitation, which can often differ in sign between models or realisations. Since the average change can be substantially smaller in magnitude than for individual realisations, this can fail to provide adequate information on the risk of larger potential changes.

Here, projected global patterns of future changes in selected indices of precipitation are presented from an ensemble based on two different atmosphere models driven by sea surface temperatures from a selection of CMIP5 models, to illustrate the altered character of worldwide precipitation at 2°C and 4°C global warming, showing ranges of outcomes following the style of UKCP09 and UKCP18.

Annual total precipitation is projected to change in all land regions of the world (Figure 1.46). In this ensemble, increased precipitation across the Arctic region is consistently projected by all members at both GWLs. In all other regions, there is no consensus on the sign of the change across the range of outcomes from the driest to the wettest at both GWLs. However, the pattern of changes stays largely the same across GWLs and the magnitude of the changes is generally larger at 4°C global warming compared to 2°C.

Heavy precipitation shows a more consistent increase at 2°C global warming, but uncertainty in the sign of change becomes more widespread at 4°C warming (Figure 1.47). Importantly, the largest projected increases are generally substantially larger than the mean of the projections, underlining the importance of considering the range of projected changes for risk assessments.

The length of dry spells is also projected to change in different ways in different regions worldwide, again with disagreements in the direction of change among the ensemble in many regions (Figure 1.48). At 4°C global warming, there is a consensus across all models on dry spells increasing by 10 days or more across large parts of southern Africa, the Iberian Peninsula and a number of small regions, and a consensus on shorter dry spells in eastern central Asia and some Arctic lands, but in most regions the ensemble projects that dry spells could either increase or decrease. In some regions, including highly populated areas such as the Indian subcontinent, the maximum projected 20-year mean increase is over 40 days per year.

The most extreme high temperatures are projected to occur in regions that are already hot. Human heat stress depend on humidity as well as temperature, and for industry uses is routinely quantified with Wet Bulb Globe Temperature (WBGT) using temperature, humidity and solar radiation. WBGT of 32°C is classified as “extreme risk” of heat stress and is rarely seen in the current global climate. Figure 1.49 shows the percentage of summer days with maximum WBGT above 32°C projected for GWLs of 2°C and 4°C, using the mean of several simulations with a subset of the CMIP5 models. Extreme heat stress conditions are projected for more than 10% of summer days in most tropical regions and many parts of the sub-tropics at 4°C global warming, and many of the more highly-populated areas of the world for more than 40% of summer days.

Engineering drawing

Description automatically generated

1.8 Earth System Instabilities – Potential risks of rapid and/or irreversible changes

Earth System Instabilities (often known as tipping points) describe accelerating, rapid or irreversible changes within the Earth System in response to external forcing. These can involve the physical climate system (e.g. ice sheets), terrestrial carbon cycles and ocean biogeochemistry, they can operate on a range of timescales and can be manifest at global or regional levels. Some are regarded as reversible, but some may persist for centuries or longer. Recent work (e.g. Steffen et al., 2018) has emphasised the links between various Earth System instabilities and considered the risk that self-reinforcing feedbacks, often referred to as tipping cascades, could push the Earth System toward a planetary threshold that, if crossed, could prevent stabilization of the climate at intermediate temperature rises even as human emissions are reduced.

Figure 1.50 provides a useful summary of our current understanding of potential candidates that may exhibit behaviours that could drive the system to more extreme climate states. A report on ‘Effect of Potential Climate Tipping Points on UK Impacts’ has been produced for CCRA3 (Hanlon et al., 2021b).

For the purposes of CCRA3, three classes of Earth system instabilities are considered. Firstly, those that could affect the UK directly through changes in our regional weather and climate without necessarily changing the level of global warming. Secondly, those involving changes in land ice, affecting sea level rise impacts in the UK and worldwide. Thirdly, those related to feedbacks involving carbon or other biogeochemical cycles that could increase the likelihood of higher levels of global radiative forcing, and hence increase the likelihood of large regional climate changes in the UK.

1.8.1 Weakening or collapse of the Atlantic Meridional Overturning Circulation

The UK’s weather and climate are fundamentally controlled by the Atlantic Meridional Overturning Circulation (AMOC), and especially by the warm, returning branch, the Gulf Stream. Although shutdown of the AMOC is considered very unlikely this century, it remains a plausible outcome in the next century (IPCC, 2019). However, significant weakening of the AMOC is considered likely and the latest results (Weijer et al., 2020) suggest that the CMIP6 models project stronger weakening than in IPCC AR5, with a possible AMOC decline between 34% and 45% by 2100. These projected declines are already represented in the UKCP18 results.

Weakening of the AMOC would lead to a cooling effect on the UK’s climate, although not enough to offset anthropogenic warming. However, the effects of AMOC weakening on specific aspects of the UK’s weather and climate may act to exacerbate the trends due to global warming documented in Section 1.5. These include shifts in rainfall patterns (including summer drying), increases in winter storminess, and further rises in sea level due to changes in the ocean density and circulation (e.g., Vellinga and Wood, 2008).

Ritchie et al. (2020) have considered the impacts on UK land use and food production of an extreme scenario in which there is a rapid weakening and AMOC collapse between 2030 and 2050. Figure 1.51 shows the changes in temperature and precipitation through the 21^st century, without and with AMOC collapse, where the climate change signal is taken from UKCP09 regional model projections using the SRES-A1B emission scenario (Nakićenović, N. et al., 2000).

Figure 1.51 provides a striking demonstration of the impacts of AMOC collapse on growing season conditions in the context of ongoing climate change. The impacts are likely to include widespread cessation of arable farming, with losses of agricultural output that are an order of magnitude larger than the impacts of climate change without an AMOC collapse.

1.8.2 Changes in the behaviour of the North Atlantic Jet Stream

As already discussed, the UK’s weather and climate are strongly influenced by the position and strength of the North Atlantic jet stream which in turn can alter the frequency and/or magnitude of high‐impact or extreme weather events. Recent extreme events, involving winter flooding and summer heatwaves, have led to questions around whether the behaviour of the jet stream is changing, with more instances of major meanders north and south, and more stalling of these meanders. Meanders in the jet are often referred to as Rossby or planetary waves and we know that slow moving, amplified Rossby waves favour the occurrence of extreme weather conditions over the UK. For example, extended periods of wet or dry weather, such winter 2013/14 and spring 2020, correspond to a jetstream meander stalling and becoming locked in one position.

It has been postulated that the amplified warming of the Arctic with the associated loss of Arctic sea ice may be weakening the jet stream and hence making it more susceptible to amplified and persistent wave anomalies (e.g., Francis and Vavrus, 2012). However, evidence to support this hypothesis has not yet been forthcoming, but it has raised interesting questions over how the jet stream might respond to warming. We know that there is a link between the strength of the jet stream and its meanders north and south (e.g., Woollings et al., 2018), in which a weaker jet stream favours more meanders, and, relatedly, more blocking events and accompanying high-impact weather. However, Woollings et al. (2018) have shown that North Atlantic jet variability is modulated on multi-decadal time scales, with decades of a strong, steady jet being interspersed with decades of a weak, variable jet.

There is as yet no robust evidence that the North Atlantic Jet Stream is changing. Furthermore, the future behaviour of the North Atlantic Jet stream is not understood, and yet could have profound implications for the UK’s weather and climate. The potential exists for significant changes in its preferred location, in its variability and in its propensity for slow moving, or even stationary, amplified Rossby waves. There is an urgent need to fill this knowledge gap.

1.8.3 Accelerated loss of Antarctic and Greenland ice sheets

Sea level rise has two major contributors – thermal expansion of the oceans as they warm (remembering that the oceans take up around 90% of the additional energy trapped in the planet), and the accumulation of ocean water mass as ice sheets and glaciers melt. Today mass accumulation dominates global sea level rise accounting for about two-thirds of the total. This is primarily due to accelerating loss of mass from the major ice sheets of Greenland and Antarctica.

The main mechanism for Greenland ice melt is changes in surface mass balance, where ice melts faster than snow can accumulate. This mechanism occurs at a steady rate and is not likely to exhibit accelerating or abrupt changes. The IPCC SROCC (2019) estimates that the complete loss of Greenland ice would contribute around 7m to global sea level rise, but this would take more than 1000 years. Sea level rise due to Greenland ice melt during the 21st century would be closer to 10s of centimetres.

The main risk of abrupt change comes from West Antarctica, which is losing ice mass primarily due to ice flow processes but could start losing ice more rapidly from accelerating instability processes. Recent advances have highlighted the potential for collapse of the West Antarctic Ice Sheet and consequent acceleration in the rate of global sea level rise. This is a predominantly marine-based ice sheet, where ice mass input to the ocean is governed primarily by ice flow processes rather than the surface mass balance that dominates for the East Antarctic Ice Sheet.

There are indications that collapse of the West Antarctic Ice Sheet could already be underway, through a positive feedback known as ‘Marine Ice Sheet Instability’ in which the ice sheet separates from its grounding line and floats free where it can melt more rapidly (e.g., Rignot et al., 2014). Recently, a second potential positive feedback on ice loss from West Antarctica has been proposed called ‘Marine Ice Cliff Instability’ (Pollard et al., 2015). This feedback would be triggered by disintegration of the floating ice shelves around Antarctica; wherever these leave behind tall coastal ice cliffs that would be structurally unstable, they may collapse entirely leaving behind further unstable cliffs. This could lead to self-sustaining ice losses and associated global sea level rise of order 1m by 2100 if the feedback were rapid and widespread (e.g. DeConto and Pollard, 2016).

Figure 1.52 shows the impact of Marine Ice Cliff Instability on sea level rise for selected UK cities based on the simulations of DeConto and Pollard (2016) compared with the UKCP18 projections (Palmer et al., 2018), along with the high-end assessments from IPCC (2019). These assessments show that the UK should be prepared for up to 2m sea level rise by 2100 in the event of accelerated Antarctic melting, which suggests a similar upper bound to the UKCP09 H++ scenario (1.9m) for the UK.

1.8.4 Permafrost thawing and additional carbon emissions

Permafrost is a mixture of soil, rocks and ice which remains permanently frozen throughout the year. Carbon stored in the permafrost is relatively inert as temperatures are too cold for much microbial activity to occur. A warming climate can induce environmental changes that accelerate the microbial breakdown of organic carbon and the release of the greenhouse gases, carbon dioxide and methane. Methane is of particular concern because although its lifetime is much shorter, it is a far more potent greenhouse gas than carbon dioxide. The addition of these greenhouse gases to the atmosphere would increase global warming and lead to further thaw, an amplifying process referred to as the “permafrost carbon feedback”.

Should the permafrost thaw, the carbon would not necessarily be released into the atmosphere immediately. The timescales of soil carbon decomposition are much slower than the projected rate of permafrost thaw. In addition, there is likely to be enhanced vegetation growth caused by warmer temperatures with increased CO₂ uptake. Schuur et al. (2015) estimate that for a high emission scenario, carbon release from permafrost is projected to be in the range 37–174 Pg of carbon by 2100, which gives a possible range of additional global warming of 0.13–0.27°C by 2100 and up to 0.42°C by 2300.

However, recent research has highlighted the importance of abrupt thawing events that could release far more carbon than a gradual thawing assessed so far (Turetsky et al., 2020). Across the Arctic and Boreal regions, permafrost is collapsing suddenly as pockets of ice within it melt. Instead of a few centimetres of soil thawing each year, several metres of soil can become destabilized within days or weeks. The land can sink and be inundated by swelling lakes and wetlands.

Abrupt thaw would probably occur in up to 20% of the permafrost zone (Olefeldt et al., 2016) but could contribute half of permafrost carbon through collapsing ground, rapid erosion and landslides. Under a high emission scenario Turetsky et al. (2020) estimate that emissions across 2.5 million km² of abrupt thaw could provide a similar climate feedback as gradual thaw emissions from the entire 18 million km² permafrost region. After considering abrupt thaw stabilization, lake drainage and soil carbon uptake by vegetation regrowth, they conclude that models considering only gradual permafrost thaw are substantially underestimating carbon emissions from thawing permafrost.

The impact of permafrost thaw on the UK would be an indirect one, associated with more rapid global warming and subsequent changes to our weather and climate. Also, the release of additional greenhouse gases would reduce the allowable carbon budget to keep within a certain level of global warming and hence is important for mitigation policies. For example, Gasser et al. (2018) compared the carbon budgets and targets of the Paris Agreement with carbon emissions from permafrost in an Earth system model. They concluded that permafrost thaw could use up 10-100% of the allowable carbon budget to stay within 1.5° C and up to 25% of the budget to stay within 2°C. Once emitted these additional carbon emissions would be irreversible for centuries.

1.8.5 Reduced carbon uptake by the biosphere

Land and ocean ecosystems act as natural buffers that limit the increase of CO₂in the atmosphere by absorbing and sequestering nearly half of emitted CO₂. As human emissions have continued to increase, this natural climate change mitigation has so far proportionally kept pace with emissions, with, for example, enhanced vegetation growth from CO₂ fertilisation.

It is expected that the ocean’s ability to take up carbon will continue, although the oceans will become more acidic with consequences for marine organisms (IPCC, 2019). This may not be the case for the land carbon sink where the situation could deteriorate quite rapidly, as deforestation and changing climatic conditions affect the major forests of Amazonia and the northern hemisphere boreal regions.

The Amazon rainforest involves a symbiotic relationship between the trees and the hydrological cycle in which a significant fraction of the rainfall falling on the forest is recycled by the forest in a self-sustaining feedback loop. In recent decades, new forcing factors – deforestation, widespread use of fire to clear vegetation and climate change – have begun to break that loop. There is also evidence that the recent climate of Amazonia has been subject to large oscillations between severe droughts and floods (Yang et al., 2018) that act to destabilise the forest system. The severity of the droughts is part of an emerging picture of an increasingly extended dry season, potentially associated with the warming of the tropical North Atlantic and shifting circulation patterns, but also with deforestation.

All these factors have raised the question of how much would be required to degrade the symbiotic relationship between the forest and the hydrological cycle, to the point that Amazonia is unable to support rain forest ecosystems and lose its role as a robust and important sink of carbon. Lovejoy and Nobre (2018, 2019) have suggested that the negative synergies between deforestation, climate change and the widespread use of fire indicate a tipping point for the Amazon system at 20-25% deforestation. Current estimates of deforestation are around 17% for the whole of Amazonia so the forest system is already close to that suggested tipping point. The loss of forest would lead to substantial losses of biodiversity and carbon with far-reaching ramifications.

The boreal forests of the northern hemisphere are also at risk from climate change. These forests store 30-40% of all land-based carbon in the world, and most of that carbon is found in the soils. High latitude warming is projected to increase dieback and disturbance in boreal forests, with increased prevalence of fires, pests and disease. All these factors could alter the structure, composition and functioning of the boreal forest systems. However, the impact on the climate system is expected to be less profound than for the Amazon rainforest, where most of the carbon is stored in the trees, and deforestation and climate change may lead to an abrupt collapse of the ecosystem.

Regarding UK impacts, the effects of forest loss would be indirect. The loss of forests would reduce the efficiency of terrestrial carbon sinks, leading to increased atmospheric concentrations and accelerated global warming. As with permafrost thaw, allowable carbon budgets to stay within specific levels of warming will be reduced with implications for mitigation.

1.9 Looking ahead to CCRA4

Important advances in climate science evidence have been made leading up to CCRA3 and during its production, particularly through UKCP18 and subsequent releases of further components of the UK Climate Projections. However, in many cases, the full range of evidence, especially from the regional and convection-permitting models, became available while the CCRA3 process was underway and hence in many cases could only be exploited in a limited way in the sectoral risk assessments. This is reflected in the assignment of confidence levels. One example is the use of threshold exceedance metrics with the convention-permitting model: information on these metrics is provided in Annex 2 for potential further use in research to inform CCRA4. Moreover, significant knowledge gaps still exist: key examples are summarised here, with suggestions for further research and development leading up to CCRA4.

1.9.1 Knowledge gaps in the scientific evidence

A theme of this chapter has been the importance of considering the volatility of the UK’s weather and climate and how this may change in the future. Although some basic analysis has been done there needs to be a much greater emphasis on this in the future.

Weather regime analysis is one way forward (e.g. Neal et al., 2016) which is already proving very valuable in operational forecasting and has been explored by De Luca et al. (2019) for interpreting future climate change in the CMIP5 projections. This needs to be repeated with the latest generation of models used in UKCP18 and IPCC AR6. Regime analysis has also been exploited for interpreting the incidence of extreme events which show clear evidence for preferred patterns (e.g. Darwish et al., 2020).

Using weather regimes, insightful diagnostics on frequencies, residence times and transitions can be explored for the current climate and the links with modes of climate variability can be produced. These would form the basis for exploring how the UK’s weather regimes may evolve with global warming and with changes in modes of climate variability. Regime analysis is already used very effectively in weather forecasting for identifying forthcoming risks across a range of sectors and this expertise may therefore be useful for assessing future levels of risk under climate change. Weather regimes may also act as useful vehicles for climate risk communication.

Linked to this, there needs to be a major focus on understanding the behaviour of the North Atlantic Jet Stream and how this will evolve in the future. This is vital for addressing current and future changes in storminess, atmospheric rivers, extreme winds, waves and coastal surges. As discussed in Section 1.8.2, this is a fundamental knowledge gap which needs to be filled before CCRA4.

1.9.2 Storylines and Scenarios

The conventional approach to representing uncertainty is through probabilistic approaches, based on ensembles of climate model simulations. One consequence of this is that the low-likelihood, high-impact events that may pose the greatest risks are difficult to isolate and factor into a risk assessment. An alternative approach is emerging called event-based storylines. Event-based storylines are physically self-consistent unfoldings of past events, or of plausible future events, with an emphasis on plausibility rather than probability (Shepherd et al., 2018). This concept links directly to common practice in disaster risk management using “stress-testing” for emergency preparedness based on events that are conditional on specific (plausible) assumptions about the hazards and possible aspects of exposure and vulnerability of the affected human or ecological system. They are particularly applicable to extreme or unprecedented events whose probability cannot be quantified, but whose impacts could be profound.

There are several reasons why storylines may complement current, probabilistic-based methods:

Improving risk awareness by framing risk in an event-oriented rather than a probabilistic manner, which corresponds more directly to how people perceive and respond to risk.
Strengthening decision-making by allowing one to work backwards from a particular vulnerability or decision point, combining climate change information with other relevant factors to address compound risk and develop appropriate stress tests.
Emphasizing the plausibility rather than probability. This concept links directly to common practices in disaster risk management using “stress-testing” for emergency preparedness based on events that are conditional on specific, but plausible assumptions.
Exploring the boundaries of plausibility, thereby guarding against false precision and surprise. Storylines also offer a powerful way of linking physical with human aspects of climate change.
Exploiting the latest generation of kilometre-scale climate models where the ensemble size may not be sufficient to define probabilities, but where the physical realism is such that plausible, and potentially unprecedented, extremes can be captured.

When co-developed by climate scientists and stakeholders, event-based storylines can provide a useful way of communicating and assessing climate-related risk in a specific decision-making context. Event-based storylines allow for conditional explanations, without full attribution of every causal factor, which is crucial when some aspects of the latter are complex and highly uncertain.

Strategic planning in government and business routinely makes use of scenarios as tools to inform thinking about future possibilities and how to manage them. Thus, scenarios are the obvious tool to describe future climate in ways that are relevant to decision-makers. Here, climate scenarios are defined as a discrete set of physically consistent and self-consistent storylines about the future, under a specified set of assumptions. The impacts and consequences of climate scenarios can be explored in considerable quantitative detail, using metrics that range from meteorological (e.g., rainfall rate) to those that are most decision relevant (e.g., flood level, numbers of people affected, and economic loss).

In summary, the development of CCRA3 has highlighted where significant gaps still exist in the climate science evidence which should be filled before CCRA4. Future options have been proposed which will improve end-to-end risk assessments and the uptake of the latest climate science, as well as enable stress-testing of future adaptation options to extreme or worst-case scenarios.

10. References

Andrews, T., Andrews, M.B., Bodas‐Salcedo, A., Jones, G.S., Kuhlbrodt, T., Manners, J. et al, (2019). Forcings, feedbacks, and climate sensitivity in HadGEM3‐GC3.1 and UKESM1. Journal of Advances in Modeling Earth Systems, 11, 4377–4394. https://doi.org/10.1029/2019MS001866

Arnell, N.W., Kay, A.L., Freeman, A., Rudd, A.C. and Lowe. J.A. (2021). Changing climate risk in the UK: A multi-sectoral analysis using policy-relevant indicators. Climate Risk Management, 31, 100265. https://doi.org/10.1016/j.crm.2020.100265

Betts, R.A., Alfieri, L., Bradshaw, C., Caesar, J., Feyen, L., et al. (2018). Changes in climate extremes, fresh water availability and vulnerability to food insecurity projected at 1.5°C and 2°C global warming with a higher-resolution global climate model. Philos. Trans. R. Soc. A Sci. 376: 20160452 https://doi.org/10.1098/rsta.2016.0452

Betts, R. A. (2020) High-end climate change and adaptation. Science for Adaptation Policy Brief 3. The World Adaptation Science Programme (WASP) Secretariat, UNEP, Nairobi. https://wedocs.unep.org/bitstream/handle/20.500.11822/34437/WASP3.pdf?sequence=1&isAllowed=y

Blunden, J. and Arndt. D.S. (2020) State of the Climate in 2019. Bull. Amer. Meteor. Soc. 101 (8), S1–S429. https://doi.org/10.1175/2020BAMSStateoftheClimate.1

Bohnenstengal, S., Hamilton, I., Davies, M. and Belcher, S.E. (2014). Impact of anthropogenic heat emissions on London’s temperatures. Quart. J. Roy. Meteor. Soc., 140, 687–698. https://doi.org/10.1002/qj.2144 .

Booth, B.B.B., Harris^,G.R., Murphy, J.M., House, J.I., Jones, C.D., Sexton, D., and Sitch, S. (2017) Narrowing the Range of Future Climate Projections Using Historical Observations of Atmospheric CO₂. Journal of Climate, 30, 3039–3053 https://journals.ametsoc.org/doi/pdf/10.1175/JCLI-D-16-0178.1

Christidis, N. and Stott, P.A. (2015). Extreme Rainfall in the United Kingdom During Winter 2013/14: The Role of Atmospheric Circulation and Climate Change. Bull. Amer. Meteor. Soc. 96 (12), S46–S50. https://doi.org/10.1175/BAMS-D-15-00094.1

Christidis N. and Stott, P.A. (2020). The extremely cold start of the spring of 2018 in the United Kingdom. Bulletin of the American Meteorological Society, 101, S23-S28. https://doi.org/10.1175/BAMS-D-19-0084.1

Christidis, N., McCarthy, M. and Stott, P.A. (2020). The increasing likelihood of temperatures above 30 to 40 °C in the United Kingdom. Nature Communications, 11, 3093. https://doi.org/10.1038/s41467-020-16834-0

Darwish, M., Fowler, H.J., Blenkinsop, S., Tye, M.R. (2018). A regional frequency analysis of UK sub-daily extreme precipitation and assessment of their seasonality. International Journal of Climatology, 38, 4758–4776. https://doi.org/10.1002/joc.5694

DeConto, R.M., and Pollard, D. (2016). Contribution of Antarctica to past and future sea level rise. Nature, 531, 591–597. https://doi.org/10.1038/nature17145

De Luca, P., Harpman, C., Wilby, R.L., Hillier, J.K., Franzke, C.L.E. and Leckebusch, G.C. (2019). Past and projected weather pattern persistence with associated multi-hazards in the British Isles. Atmosphere, 10, 577. https://doi.org/10.3390/atmos10100577

Doherty, R.M., Wild, O., Shindell, D.T., Zeng,G., MacKenzie, I.A., Collins, W.J. et al. (2013). Impacts of climate change on surface ozone and intercontinental ozone pollution: A multi-model study. J. Geophys. Res., 118, 3744–3763. https://doi.org/10.1002/jgrd.50266

ESOTC. (2019). European State of the Climate (ESOTC) annual report compiled by the EU Copernicus Climate Change Service. https://climate.copernicus.eu/ESOTC/2019

Feser, F., Barcikowska, M., Krueger, O., Schenk, F., Weisse, R., Xia, L. (2015). Storminess over the North Atlantic and Northwestern Europe – A Review. Q. J. R. Meteorol. Soc., 141, 350-382, January 2015 B. https://doi.org/10.1002/qj.2364

Folland, C.K., Knight, J., Linderholm, H.W., Fereday, D., Ineson, S. and Hurrell, J.W. (2009). The Summer North Atlantic Oscillation: Past, Present, and Future. J. Climate, 22, 1082 – 1103. https://doi.org/10.1175/2008JCLI2459.1

Forster, P.M., Maycock, A.C., McKenna, C.M. and Smith, C.J. (2020). Latest climate models confirm need for urgent mitigation. Nat. Clim. Change, 10, 7–10. https://doi.org/10.1038/s41558-019-0660-0

Fowler, H.J., Lenderink, G., Prein, A.F., Westra, S., Allan, R.P., Ban, N. et al. (2020). Anthropogenic intensification of short-duration rainfall extremes. Nature Reviews Earth and Environment, 2, 107-122. https://doi.org/10.1038/s43017-020-00128-6

Frame, D.J., Joshi, M.M., Hawkins, E., Harrington, L.J. and de Roiste, M. (2017). Population-based emergence of unfamiliar climates. Nature Climate Change, 7, 407–411. https://doi.org/10.1038/nclimate3297

Francis, J.A. and Vavrus, S.J. (2012). Evidence linking Arctic amplification to extreme weather in mid‐latitudes. Geophysical Research Letters, 39(6). https://doi.org/10.1029/2012GL051000

Gasser, T., Kechiar, M., Ciais, P., Burke, E.J., Kleinen, T., Zhu, D. et al. (2018). Path-dependent reductions in CO2 emission budgets caused by permafrost carbon release. Nature Geoscience, 11(11), 830–835. https://doi.org/10.1038/s41561-018-0227-0

Gohar, L., Bernie, D., Good, P. and Lowe, J.A. (2018) UKCP18 Derived Projections of Future Climate over the UK. Met Office https://www.metoffice.gov.uk/pub/data/weather/uk/ukcp18/science-reports/UKCP18-Derived-Projections-of-Future-Climate-over-the-UK.pdf.

Guerreiro, S.B., Fowler, H.J., Barbero, R., Westra, S., Lenderink, G., Blenkinsop, S. et al. (2018a). Detection of continental-scale intensification of hourly rainfall extremes. Nature Climate Change, 8, 803-807. https://doi.org/10.1038/s41558-018-0245-3

Hanlon, H. M., D. Bernie, G. Carigi, and J. A. Lowe, J.A. (2021a) Future Changes to high impact weather in the UK. Climatic Change DOI 10.1007/s10584-021-03100-5

Hanlon, H., Jackson, L., Palmer, M., Burke, E., Eade, R. and Betts, R.A. (2020b). Effect of potential climate tipping points on UK impacts. Prepared in support of the CCRA3 Technical Report. https://www.ukclimaterisk.org/independent-assessment-ccra3/research-supporting-analysis/

Hawkins, E. and Sutton, R. (2009). The Potential to Narrow Uncertainty in Regional Climate Predictions. Bull. Amer. Meteor. Soc. 90 (8), 1095–1108. https://doi.org/10.1175/2009BAMS2607.1

Hawkins, E., Frame, D., Harrington, L., Joshi, M., King, A., Rojas M. et al. (2020). Observed Emergence of the Climate Change Signal: From the Familiar to the Unknown. Geophysical Research Letters, 47(6), e2019GL086259. https://doi.org/10.1029/2019GL086259

HM Government. (2016). National Flood Resilience Review. https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/551137/national-flood-resilience-review.pdf

Hogarth, P., C.W. Hughes, C.W., Williams, S.D.P. and Wilson, C. (2020). Improved and extended tide gauge records for the British Isles leading to more consistent estimates of sea level rise and acceleration since 1958. Progress in Oceanography, 184, 102333. https://doi.org/10.1016/j.pocean.2020.102333

Hollis, D., McCarthy, M., Kendon, M., Legg, T. and Simpson, I. (2019). HadUK‐Grid—A new UK dataset of gridded climate observations. Geoscience Data Journal, Royal Meteorological Society, 6 (2): 151 – 159. https://doi.org/10.1002/gdj3.78

IPCC (2013) Summary for Policymakers. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA https://www.ipcc.ch/report/ar5/wg1/

IPCC (2019) Summary for Policymakers. In: IPCC Special Report on the Ocean and Cryosphere in a Changing Climate [H.-O. Pörtner, D.C. Roberts, V. Masson-Delmotte, P. Zhai, M. Tignor, E. Poloczanska, K. Mintenbeck, A. Alegría, M. Nicolai, A. Okem, J. Petzold, B. Rama, N.M. Weyer (eds.)]. https://www.ipcc.ch/srocc/

Johns, T. (2020). Comparison of climate change and climate impact metrics over the UK for the UKCP09 and UKCP18 RCM-PPE ensembles, within a broader context of climate modelling uncertainty. Prepared in support of the CRA3 Technical Report. https://www.ukclimaterisk.org/independent-assessment-ccra3/research-supporting-analysis/

Kendon, E., Roberts, N., Fowler, H., Roberts, M.J., Chan, S.C. and Senior, C.A. (2014). Heavier summer downpours with climate change revealed by weather forecast resolution model. Nature Climate Change, 4, 570–576. https://doi.org/10.1038/nclimate2258

Kendon, E., Fosser, G., Murphy, J., Chan, S., Clark, R., Harris,G. et al. (2019). UKCP Convection-permitting model projections: Science report. https://www.metoffice.gov.uk/pub/data/weather/uk/ukcp18/science-reports/UKCP-Convection-permitting-model-projections-report.pdf

Kendon, E., Roberts, N.M., Fosser, G., Martin, G.M., Lock, A.P., Murphy, J.M. et al. (2020). Greater Future U.K. Winter Precipitation Increase in New Convection-Permitting Scenarios. J. Climate, 33, 7303–7318. https://doi.org/10.1175/JCLI-D-20-0089.1

Kendon, M., McCarthy, M., Jevrejeva, S., Matthews, A., Sparks, T. and Garforth, J. (2020). State of the UK Climate 2019. Int. J. Climatology, 40 (S1), 1-69. https://doi.org/10.1002/joc.6726

Kennedy-Asser, A.T., Andrews, O., Mitchell, D.M. and Warren, R.F. (2020). Evaluating heat extremes in the UK Climate Projections (UKCP18). Environmental Research Letters, 16 (1), 014039.

https://doi.org/10.1088/1748-9326/abc4ad

Kent, C., Pope, E., Thompson, V., Lewis, K., Scaife, A.A. and Dunstone, N. (2017). Using climate model simulations to assess the current climate risk to maize production. Environ. Res. Lett., 12, 054012. https://doi.org/10.1088/1748-9326/aa6cb9

Krueger, O., Feser, F. and Weisse, R. (2019). Northeast Atlantic Storm Activity and Its Uncertainty from the Late Nineteenth to the Twenty-First Century. J. Climate, 32, 1919-1931. https://doi.org/10.1175/JCLI-D-18-0505.1

Lavers, D.A., Allan, R.P., Wood, E.F., Villarini, G., Brayshaw, D.J. and Wade, A.J. (2011). Winter floods in Britain are connected to atmospheric rivers. Geophysical Research Letters, 38 (23), L23803. https://doi.org/10.1029/2011GL049783 .

Lewis, E., Quinn, N., Blenkinsop, S., Fowler, H.J., Freer, J., Tanguy, M. et al. (2018). A rule based quality control method for hourly rainfall data and a 1 km resolution gridded hourly rainfall dataset for Great Britain: CEH-GEAR1hr. Journal of Hydrology, 564, 930-943. https://doi.org/10.1016/j.jhydrol.2018.07.034

Lo, Y.T., Mitchell, D.M., Bohnenstengel, S.I.,Collins, M., Hawkins, E., Hegerl, G.C. et al. (2020). U.K. Climate Projections: Summer Daytime and Nighttime Urban Heat Island Changes in England’s Major Cities. J. Climate, 33, 9015 – 9030. https://doi.org/10.1175/JCLI-D-19-0961.1

Lovejoy, T.E. and Nobre, C. (2018). Amazon Tipping Point. Sci. Adv., 4 (2), eaat2340. https://doi.org/10.1126/sciadv.aat2340

Lovejoy, T.E. and Nobre, C. (2019). Winds of will: Tipping change in the Amazon. Sci. Adv., 5 (12), eaba2949. https://doi.org/10.1126/sciadv.aba2949

Matthews, T., Murphy, C., Wilby, R. and Harrigan, S., (2014). Stormiest winter on record for Ireland and UK. Nature Clim. Change, 4, 738–740. https://doi.org/10.1038/nclimate2336

Matthews, T., Murphy, C., McCarthy, G., Broderick, C. and Wilby, R.L. (2018). Super Storm Desmond: a process-based assessment. Environ. Res. Lett.,13, 014024. https://doi.org/10.1088/1748-9326/aa98c8

McCarthy, M., Armstrong, L. and Armstrong, N. (2019a). A new heatwave definition for the UK. Weather, 74 (11), 382-387. https://doi.org/10.1002/wea.3629

McCarthy, M., Christidis, N., Dunstone, N., Fereday, D., Kay, G., A. Klein-Tank, A. et al. (2019b). Drivers of the UK summer heatwave of 2018. Weather, 74 (11), 390-396. https://doi.org/10.1002/wea.3628

McSweeney, C. and Bett, P. (2020). UKCP European Circulation Indices: Jet Stream Position and Strength. UKCP Factsheet. Met Office. https://www.metoffice.gov.uk/binaries/content/assets/metofficegovuk/pdf/research/ukcp/ukcp18_factsheet_jet_stream.pdf

Meehl, G.A., Senior, C.A., Eyring, V., Flato, G., Lamarque, J-F., Stouffer, R.J. et al. (2020). Context for interpreting equilibrium climate sensitivity and transient climate response from the CMIP6 Earth system models. Science Advances, 6 (26), eaba1981. https://doi.org/10.1126/sciadv.aba1981

Murphy, J., Sexton, D., Jenkins, G., Boorman, P., Booth, B., Brown, K. et al. (2009) UK Climate Projections science report: Climate change projections, Met Office Hadley Centre http://cedadocs.ceda.ac.uk/1320/1/climate_projections_full_report.pdf

Murphy, J.M., Harris, G.R., Sexton, D.M.H., Kendon, E.J., Bett, P.E., Clark, R.T. et al. (2018). UKCP18 Land Report. https://www.metoffice.gov.uk/pub/data/weather/uk/ukcp18/science-reports/UKCP18-Land-report.pdf

Nakićenović, N. et al. (2000) Special Report on Emissions Scenarios. Intergovernmental Panel on Climate Change, Cambridge University Press, UK. https://www.ipcc.ch/report/emissions-scenarios/

Neal, R., Fereday, D., Crocker, R. and Comer, R.E. (2016). A flexible approach to defining weather patterns and their application in weather forecasting over Europe. Meteorol. Appl., 23, 389–400. https://doi.org/10.1002/met.1563

Olefeldt, D., Goswami, S., Grosse, G., Hayes, D., Hugelius, G., Kuhry, P. et al. (2016). Circumpolar distribution and carbon storage of thermokarst landscapes. Nature Communications, 7, 1–11. https://doi.org/10.1038/ncomms13043

Palmer, M., Howard, T., Tinker, J., Lowe, J., Bricheno, L., Calvert, D. et al. (2018). UKCP18 Marine Report. https://www.metoffice.gov.uk/binaries/content/assets/metofficegovuk/pdf/research/ukcp/ukcp18-marine-report-updated.pdf

Palmer, M.D., Gregory, J.M., Bagge, M., Calvert, D., Hagedoorn, J.M., Howard, T. et al. (2020). Exploring the Drivers of Global and Regional Sea-Level Change over the 21st Century and Beyond. Earth’s Future, 8 (9), e2019EF001413. https://doi.org/10.1029/2019EF001413

Payne, A.E., Demory, M-E., Leung, L.R., Ramos, A.M., Shields, C.A., Rutz, J.J. et al. (2020). Responses and impacts of atmospheric rivers to climate change. Nature Reviews, Earth & Environment Reviews, 1, 143-157. https://doi.org/10.1038/s43017-020-0030-5

Pollard, D., Deconto, R.M. and Alley, R.B. (2015). Potential Antarctic Ice Sheet retreat driven by hydrofracturing and ice cliff failure. Earth and Planetary Science Letters, 412, 112–121. https://doi.org/10.1016/j.epsl.2014.12.035

Priestley, M.D.K., Dacre, H.F., Shaffrey, L.C., Hodges, K.I. and Pinto, J.G. (2018). The role of serial European windstorm clustering for extreme seasonal losses as determined from multi-centennial simulations of high-resolution global climate model data. Nat. Hazards Earth Syst. Sci., 18, 2991–3006. https://doi.org/10.5194/nhess-18-2991-2018

Rignot, E., Mouginot, J., Morlighem, M., Seroussi, H. and Scheuchl, B. (2014). Widespread, rapid grounding line retreat of Pine Island, Thwaites, Smith, and Kohler glaciers, West Antarctica, from 1992 to 2011. Geophysical Research Letters, 41, 3502–3509. https://doi.org/10.1002/2014GL060140

Ritchie, P.D., Smith, G.S., Davis, K.J., Fezzi, C., Halleck-Vega, S., Harper, A.B. et al. (2020). Shifts in national land use and food production in Great Britain after a climate tipping point. Nature Food, 1, 76–83. https://doi.org/10.1038/s43016-019-0011-3

Rodwell, M.J. and Palmer, T.N. (2007). Using numerical weather prediction to assess climate models. Quarterly Journal of the Royal Meteorological Society, 133, 129–146. https://doi.org/10.1002/qj.23

Royal Society. (2008). Ground-level ozone in the 21st century: future trends, impacts and policy implications. Available from https://royalsociety.org/topics-policy/publications/2008/ground-level-ozone/

Scaife, A A., Arribas, A., Blockley, E., Brookshaw, A., Clark, R.T., Dunstone, N. et al. (2014). Skillful long-range prediction of European and North American winters. Geophys. Res. Lett., 41, 2514–2519. https://doi.org/10.1002/2014GL059637

Schuur, E.A.G., McGuire, A.D., Schädel, C., Grosse, G., Harden, J.W., Hayes, D.J. et al. (2015). Climate change and the permafrost carbon feedback. Nature, 520, 171–179. https://doi.org/10.1038/nature14338

Senior, C.A., Andrews, T., Burton, C., Chadwick, R., Copsey, D., Graham, T., et al. (2016). Idealized climate change simulations with a high-resolution physical model: HadGEM3-GC2. J. Adv. Model Earth Syst., 8 (2), 813-830. https://doi.org/10.1002/2015MS000614

Sexton, D. and Harris, G. (2015). The importance of including variability in climate change projections used for adaptation. Nature Clim Change, 5, 931–936. https://doi.org/10.1038/nclimate2705

Shepherd, T.G., Boyd, E., Calel, R.A., Chapman, S.C., Dessai, S., Dima-West, I.M. et al. (2018). Storylines: an alternative approach to representing uncertainty in physical aspects of climate change. Climatic Change, 151, 555-571. https://doi.org/10.1007/s10584-018-2317-9

Sherwood, S., Webb, M.J., Annan, J.D., Armour, K.C., Forster, P.M., Hargreaves, J.C., Hegerl, G. et al. (2020). An assessment of Earth’s climate sensitivity using multiple lines of evidence. Rev. Geophysics, 58 (4), e2019RG000678. https://doi.org/10.1029/2019RG000678

Smith, D.M., Eade, R., Scaife, A.A., Caron, L-P., Danabasoglu, G., DelSole, T.M. et al. (2019). Robust skill of decadal climate predictions npj Clim Atmos Sci., 2, 13. https://doi.org/10.1038/s41612-019-0071-y

Steffen, W., Rockström, J., Richardson, K., Lenton, T.M., Folke, C., Liverman, D. et al. (2018). Trajectories of the Earth System in the Anthropocene. PNAS, 115, 8252–8259. https://doi.org/10.1073/pnas.1810141115

Stott, P.A., Stone, D.A. and Allen, M.R. (2004). Human contribution to the European heatwave of 2003. Nature, 432, 610–614. https://doi.org/10.1038/nature03089

Stott, P.A., Christidis, N., Otto, F.E.L., Sun, Y., Vanderlinden, J-P., van Oldenborgh, G.J. et al. (2016). Attribution of extreme weather and climate‐related events. WIREs Climate Change, 7 (1), 23-41. https://doi.org/10.1002/wcc.380

Sutton, R., Suckling, E. and Hawkins, E. (2015). What does global mean temperature tell us about local climate? Philosophical Transactions of the Royal Society A, 373, 20140426. ISSN 1364-503X. https://doi.org/10.1098/rsta.2014.0426

Thompson, V., Dunstone, N.J., Scaife, A.A., Smith, D.M., Slingo, J.M., Brown, S. et al. (2017). High risk of unprecedented UK rainfall in the current climate. Nature Communications, 8, 107. https://doi.org/10.1038/s41467-017-00275-3

Turetsky, M., Abbott, B.W., Jones, M.C., Anthony, K.W., Olefeldt, D., Schuur, E.A.G. et al. (2020). Carbon release through abrupt permafrost thaw. Nature Geoscience, 13, 138-143. https://doi.org/10.1038/s41561-019-0526-0

Vautard, R., Van Oldenborgh, G.J., Otto, F.E.L., Yiou, P., de Vries, H., van Meijgaard, E. et al. (2019). Human influence on European wind storms such as those of January 2018. Earth System Dynamics, 10, 271-286. https://doi.org/10.5194/esd-10-271-2019

Weijer, W., Cheng, W., Garuba, O.A., Hu, A, and Nadiga, B.T. (2020). CMIP6 Models Predict Significant 21st Century Decline of the Atlantic Meridional Overturning Circulation. Geophysical Research Letters 47 (12), e2019GL086075. https://doi.org/10.1029/2019GL086075.

Wilby, R.L. and Quinn, N.W. (2013). Reconstructing multi-decadal variations in fluvial flood risk using atmospheric circulation patterns. Journal of Hydrology, 487, 109-121. https://doi.org/10.1016/j.jhydrol.2013.02.038

Williams, K.D., Copsey, D., Blockley, E.W., Bodas-Salcedo, A., Calvert, D., Comer, R.P. et al. (2018). The Met Ofﬁce Global Coupled model 3.0 and 3.1 (GC3.0 and GC3.1) conﬁgurations. Journal of Advances in Modeling Earth Systems, 10, 357–380. https://doi.org/10.1002/2017MS001115

Williams, K.D., Hewitt, A.J. and Bodas-Salcedo, A. (2020). Use of short-range forecasts to evaluate fast physics processes relevant for climate sensitivity. Journal of Advances in Modeling Earth Systems, 12, e2019MS001986. https://doi.org/10.1029/2019MS001986

WMO. (2019). WMO Statement on the State of the Global Climate in 2019. Available from: https://library.wmo.int/doc_num.php?explnum_id=10211

Woollings, T., Barnes, E., Hoskins, B., Kwon, Y-O., Lee, R.W., Li, C. et al. (2018). Daily to Decadal Modulation of Jet Variability. J. Climate, 31, 1297-1314. https://doi.org/10.1175/JCLI-D-17-0286.1

Wyser, K., Strandberg, G., Caesar, J. and Gohar, L. (2016) Documentation of changes in climate variability and extremes simulated by the HELIX AGCMs at the 3 SWLs and comparison to changes in equivalent SST/SIC low-resolution CMIP5 projections. HELIX project deliverable 3.1 https://helixclimate.eu/wp-content/uploads/2018/04/HELIX-603864-D3.1-Documentation-of-changes-in-climate-variability-and-extremes-simulated-by-the-HELIX-AGCMs-copy.pdf

Yang, Y., Saatchi, S.S., Xu, L., Yu, Y., Choi, S., Phillips, N. et al. (2018). Post-drought decline of the Amazon carbon sink. Nature Communications, 9, 3172. https://doi.org/10.1038/s41467-018-05668-6

Zelinka, M.D., Myers, D., McCoy, T., Po-Chedley, S., Caldwell, P.M., Ceppi, P. et al. (2020). Causes of higher climate sensitivity in CMIP6 models. Geophysical Research Letters, 47, e2019GL085782. https://doi.org/10.1029/2019GL085782

Annex 1. Advances in climate modelling since CCRA2

A1.1 Improvements to climate models

Since CCRA2, significant advances have been made in both global and regional modelling for weather and climate prediction. For climate science serving the UK, a new climate model, HadGEM3, is a key part of a comprehensive new set of UK climate projections (UKCP18) and other applications including global forecasts on timescales of months to a century. HadGEM3 features significant increases in horizontal and vertical resolution in the atmosphere and ocean, as well as improvements in model physics, giving notable reductions in a number of key systematic biases (Williams et al., 2018). The atmosphere model resolution increased from 150km in the horizontal, as used for CMIP5 and IPCC AR5, to 60km, and from 38 to 85 levels in the vertical. The ocean resolution increased from 1⁰to 0.25⁰ and from 40 to 75 levels.

These enhancements in resolution have delivered significant improvements in the structure of synoptic weather systems and ocean circulation, and in subseasonal and seasonal predictability for the UK, including precipitation (e.g., Scaife et al., 2014). This is associated in part with much improved interactions between the stratosphere and troposphere with higher vertical resolution and better representation of the Gulf Stream with higher ocean resolution.

Analysis of the HadGEM3 perturbed parameter ensemble used in UKCP18 showed that, for all members, HadGEM3 out-performs the CMIP5 models across a wide range of climate variables (Figure A1.1 from UKCP18 Land projections: Science Report).

Another advance since CCRA2 has been the creation of a new ensemble of regional climate simulations based on the variants of HadGEM3 used in the perturbed parameter ensemble and driven by boundary conditions from the equivalent global model. So, the regional simulations have also benefited from the improvements in the North Atlantic weather and climate variability in HadGEM3.

The regional model uses a higher resolution of 12km compared with the earlier CORDEX and UKCP09 regional simulations at 25km and gives better representations of the UK landscape and its associated local meteorology, especially in mountainous and coastal areas. Variability in precipitation is higher than in the global model, in better agreement with the observations, and extreme events, such as heavy precipitation and winter cold days, are better represented. Although the ensemble size is still relatively small at only 12 members, the simulations provide meteorologically-consistent scenarios on all timescales from hours to decades and from the local to the national scale, which can be used to drive impacts models. The improved synoptic variability in HadGEM3 and the better representation of extremes in the regional simulations should enable greater stress testing of the UK’s resilience across a range of hazards and their impacts.

Finally, UKCP18 included for the first time, internationally, an ensemble of local simulations at an unprecedented resolution of 2.2km, equivalent to the resolution currently being used for operational ensemble weather forecasting for the UK (see UKCP Convection-Permitting Model Projections: Science Report, 2019). Again, the ensemble is based on HadGEM3 and uses the same global model driving fields as the regional 12km ensemble. The aim of this ensemble is to provide information on local extremes, such as sub-daily rainfall and wind gusts associated with, for example, sting jets. There are numerous examples in operational weather forecasting highlighting the value of this class of simulations, especially for severe weather warnings.

The UKCP18 2.2km ensemble was published in September 2019 and so has only been used to a limited extent in CCRA3. In future, the value of the convection-permitting model (CPM) ensemble can be expected to be increasingly realised for providing detailed spatial and temporal extreme scenarios to stress test the UK’s preparedness and resilience policies at the local level. This was demonstrated for earlier applications of the CPM for the National Flood Resilience Review analysis of the fit-for-purpose of the Extreme Flood Outlines (National Flood Resilience Review 2016).

In summary, climate science and important aspects of the skill of the climate models have advanced significantly since CCRA2. These advances are embodied primarily in the new family of global and regional models, based on HadGEM3, which is an important part of UKCP18. However, the climate sensitivity of HadGEM3 (Andrews et al., 2019) is significantly higher than earlier Hadley Centre global models, such as HadCM3 used in UKCP09 and HadGEM1 used in IPCC AR5. It is also higher than the cohort of CMIP5 models, which was used in CCRA2, but in line with a number of recently developed models being used in CMIP6 for the IPCC AR6. As Andrews et al. (2019) note, none of the model’s forcing and feedback processes are found to be atypical of other models, although the cloud feedback is at the high end.

CCRA3 is framed in terms of trajectories of global warming rather than emissions scenarios. Research commissioned on some of the risks therefore used selected components of the UKCP18 projections representing global warming of approximately 2°C and 4°C by the end of the century. The assessment of other risks draws on literature using other models, projections and scenarios that give approximately 2°C and 4°C global warming by 2100.

A1.2 Methods for climate projections

This section summarises the methods used for the latest projections of future changes in the UK’s climate during the 21^st Century. It draws on the detailed reports produced for the new UK climate projections UKCP18 (Gohar et al., 2018; Kendon et al., 2019; Murphy et al., 2018; Palmer et al., 2018) the emerging literature from the latest global climate models in preparation for IPCC AR6, and analysis by Hanlon et al. (2021a). Comparisons with earlier climate projections are also made, since much of the impacts and risks literature assessed in CCRA3 uses these.

UKCP18 consists of several components which are useful for different aspects of climate change risk assessment.

Probabilistic projections of climate over UK land which draw together information from perturbed-parameter ensembles with the HadCM3 model and the multi-model ensemble in CMIP5. These give an estimate of the likelihood of particular changes, e.g., long-term average temperature or precipitation change across the UK, from the present day to 2100. The changes at each gridpoint are independent from each other – they are not intended to provide a coherent picture of change across the country, just likelihood of change in individual locations.
A set of global projections at 60km resolution out to 2100 which are spatially coherent, i.e., each individual projection represents climates that could occur simultaneously across the world. Overall, the projected changes across the whole set cover a similar range to those in the probabilistic projections described above, but individual projections are not assigned likelihoods. Within this set, the HadGEM3 model provides a subset of projections. Due to the behaviour of HadGEM3, this subset of projections tends to be at the warmer end of the range of probabilistic projections. The other subset uses projections from other climate models developed by other modelling centres, as part of the 5^th Coupled Model Intercomparison Project (CMIP5). The range of outcomes from the CMIP5 subset covers a similar range to the probabilistic projections.
A set of regional projections of climate over land out to 2100 at 12km resolution, covering the UK. These are driven at their boundaries by output from the global simulations with the HadGEM3 subset of the global projections and provide a more detailed simulation of the climate at the UK scale. Again, because they use HadGEM3 boundary conditions, the projections are at the warmer end of the range of the probabilistic projections.
A further set of local projections of climate over land at 2.2km resolution, for selected time periods during the 21^st Century. These higher-resolution models are designed to provide a more realistic simulation of key meteorological processes, particularly convection, and are particularly for simulating extreme events such as heavy rainfall or high daily maximum temperatures.
A set of projections of long-term sea level rise, storm surges and changes in wave height for UK coastlines
A set of projections derived from the above, representing long-term climate states at 2°C and 4°C global warming and a low-emissions scenario

The regional simulations are driven by HadGEM3 boundary conditions, so the output is influenced not only by HadGEM3’s higher climate sensitivity but also its representation of important weather regimes. These are summarised in Section A1.3 where they are placed in context with the probabilistic projections and global model simulations. Due to the computational cost of the high-resolution regional simulations with HadGEM3, regional climate change scenarios were only produced with one emissions scenario to allow for the largest possible ensemble size to be utilized in order to cover a wide range of regional climate outcomes. The highest emissions scenario, RCP8.5, was chosen so that the widest range of future levels of global warming could be explored, including the most extreme climate changes considered as low-probability, high-impact scenarios. Nevertheless, for many climate impact drivers, the projected regional changes at particular levels of global warming can be considered to be representative of the same level of global warming reached at a later date with a lower emissions scenario and/or as a result of a lower climate sensitivity.

The regional projections basically reflect the driving boundary conditions of HadGEM3. In winter the results mostly lie within the range of the probabilistic projections but indicate slightly wetter and warmer winters than in previous assessments, consistent with the increase in cyclonic weather types in HadGEM3 and its higher climate sensitivity. In summer, however, the regional projections are substantially different and are towards the hotter, drier end of the range of the probabilistic projections. Both Scotland and England are projected to be 1-2^oC warmer than the earlier CMIP5 models, associated in part with HadGEM3’s higher climate sensitivity. For England, there is a very strong signal for much reduced rainfall, which is in line with the link between the phase of the SNAO and summer rainfall noted earlier and therefore not due entirely to the higher climate sensitivity.

A1.3 Comparison of UKCP18 with UKCP09 and other climate projections

Figure A1.2 shows that the trajectories of the 30-year average UK climate are very similar in UKCP18 and UKCP09 when considering a consistent emissions scenario, with the differences at any particular percentile level being much smaller than the 5^th to 95^th percentile spread. This provides reassurance that the evidence for the ongoing trend towards increased likelihood warmer, wetter winters and hotter, drier summers is unchanged. This consistency with UKCP09 on the overall trajectory of long-term average climate change is important because much of the climate risk literature available for CCRA3 uses UKCP09.

CCRA3 commissioned a detailed comparison of these climate metrics based on processed daily data from the regional climate model (RCM) perturbed parameter ensemble components of UKCP18 and UKCP09 (Johns, 2021). In order to place the UKCP results in a broader modelling uncertainty context, results have also been compared to an ensemble of 13 selected CMIP5 global projections, analysed in a similar way using available daily data. The analysis considered future changes in the metrics of interest through the 21^st Century relative to a present-day baseline of 1981-2000, with a focus on changes at specified global warming levels ranging through 1.5, 2, 3 and 4 °C relative to preindustrial.

Reframing the results in terms of 20-year time slices centred on 2, 3 and 4 °C of global mean warming shows that the UKCP18 regional model ensemble actually exhibits relatively lower warming than the UKCP09 regional model ensemble over all UK nations for any given level of global warming (Figure A1.3). However, the UKCP18 results are higher than those based on global CMIP5 model results. Figure A1.3 also shows the bias in the various ensembles for the present day so that the projected climate change can be put in context.

A1.4 Emissions scenarios and concentration pathways in the RCPs

Studies of future climate change impacts and risks using climate model projections use different approaches to the representation of future emissions scenarios. Some, such as UKCP09 and UKCP18, use Earth System Models to calculate a more complete climate system response to a given emissions scenario, including modelling the carbon cycle interactively with the atmosphere and oceans. Others, such as CMIP5, do not model the carbon cycle and instead assign a specific scenario of CO₂ concentrations based on other models. In order to compare or integrate results from these projections correctly, it is important to appreciate the differences between these approaches. Unfortunately, for the RCP scenarios, the same terminology is used in the literature for both emissions scenarios and concentration pathways, and this can often lead to a lack of clarity and poor understanding of the context of research studies of climate change risks.

The RCPs (Representative Concentration Pathways) are defined in terms of levels of radiative forcing at the end of the 21^st Century, which is due to particular level of atmospheric concentrations of greenhouse gases and aerosols. Each RCP has a standard pathway of atmospheric CO₂ concentrations, and it is these concentration pathways that are used as input most climate projections of the CMIP5 generation, including those used in the IPCC 5^th Assessment Report (IPCC, 2013). For example, in RCP8.5, the radiative forcing in 2100 is 8.5 Wm^-2, and the CO₂ concentration at that time in the standard RCP8.5 concentration pathway is 936 parts per million (ppm) (Booth et al., 2017).

However, the same term (RCP) is also applied to specific emissions scenarios that are conventionally associated with these concentration pathways. For example, in the standard RCP8.5 emissions scenario, the cumulative CO₂ emissions from 2020 to 2100 are 6629 Gigatonnes of CO₂ (GtCO₂)^[13]. These emissions scenarios are used in other climate models using a different approach to the main CMIP5 models, taking the emissions as input and then calculating the CO₂ concentrations within the model. The UKCP18 projections uses this approach.

A critical point about the differences between these approaches is that there is uncertainty in the strength of carbon cycle feedbacks in the climate system, with the result that:

Any given emissions scenario can give rise to a wide range of future concentration pathways
Any specific future concentration pathway can arise from a wide range of emissions scenarios

In UKCP18, the probabilistic projections include an exploration of uncertainties in carbon cycle feedbacks, and hence effectively represent a range of CO₂ concentrations compatible with each RCP emissions scenario, not just the standard RCP concentration pathways used in CMIP5. This should be borne in mind when comparing the CMIP5 projections with UKCP18 projections using apparently the same scenario – even though both are labelled “RCP”, the scenarios have been applied differently. The standard RCP8.5 concentration pathway used in the CMIP5 projections is in the lower part of the range of concentrations compatible with the RCP8.5 emissions scenario using a coupled climate-carbon cycle model. (Booth et al., 2017; Murphy et al., 2018; Figure A1.4)

This difference in methodology is an important factor in causing the UKCP18 RCP8.5 emissions-driven probabilistic projections to simulate a much wider range of levels of global warming by 2100 than the CMIP5 RCP8.5 concentrations-driven projections, with the upper end of the UKCP18 range being considerably higher than that of the CMIP5 range (see Figure A1.3). It is also one reason for very rapid projected warming in the UKCP18 60km resolution global projections with the HadGEM3 perturbed-parameter ensemble (the other reason being the high equilibrium climate sensitivity / transient climate response).

An important implication of this is that climate projections driven by RCP concentration pathways could arise from different emissions scenarios from the ones conventionally associated with the RCP pathways used. For example, the RCP8.5 concentration of 936 ppm can be reached by a much lower emissions scenario than the standard RCP8.5 emissions scenario. In the SRES A1B emissions scenario (Nakićenović, N. et al., 2000), cumulative emissions are approximately two-thirds of those in the RCP8.5 emissions scenario, but in a perturbed-parameter ensemble of a coupled climate-carbon cycle model constrained against observed changes in CO₂ concentrations, these emissions lead to a wide range of CO₂ concentrations by 2100 which include the 936 ppm of the standard RCP8.5 pathway (Figure A1.5).

In conclusion: due to substantial uncertainties in translating emissions to concentrations, there is no single one-to-one relationship between an emissions scenario and a concentration pathway. This can lead to confusion over the interpretation of the scenario and the level of climate change impact that it represents. Nevertheless, the use of the same term “RCP” for both emissions scenarios and concentration pathway is common, so for clarity it is helpful to specify whether particular projections used RCP emissions scenarios or RCP concentration pathways.

Annex 2: Threshold Exceedance Metrics

A goal of CCRA3 is to look beyond long-term climatic trends to include volatility of the weather and climate at the regional and local scales. It is the case that some of the more costly, disruptive and dangerous impacts of climate change will be associated with increased frequency and/or intensity of extreme weather and climate events. Furthermore, some of these impacts only come into play, or become very serious, when certain meteorological thresholds are exceeded (Table A2.1).

There is also now the possibility of applying threshold exceedance metrics to the UKCP18 2.2km time-slice scenarios, especially those related to extreme daily and sub-daily rainfall linked to embedded frontal convection and summer thunderstorms, to investigate how the current risk evolves under climate change. Similar to the UNSEEN methodology we can use the 12-member ensemble for 1981-2000 to provide 240 ‘years’ of synthetic observations of the current risk of exceeding certain impact thresholds. This will add to our understanding of the baseline risk from extreme events, which is currently very limited due to the shortness of the observational record. The same analysis can then be applied to the 240 ‘years’ for 2021-2040, which will provide a first assessment of how these metrics may change under near-term climate change. This may provide a valuable tool for CCRA4.

Table A2.1: Impact indices
Index	Threshold	Impact relevance
Frost Days	Daily minimum temperature below 0°C	Cold-weather disruption due to higher than normal chance of ice and snow.
Icing Days	Daily maximum temperature below 0°C	More extreme than frost days, so more severe cold-weather impacts.
Tropical Nights	Daily minimum temperature above 20°C	Health impact due to high night-time temperatures with potential for heat stress. Vulnerable people at increased risk of hospital admission or death.
Summer Days	Daily maximum temperature above 25°C	High daytime temperatures with health impacts for vulnerable people at risk of hospital admission or death. Transport disruption – e.g., track buckling on railways.
Rainfall meeting National Severe Weather Warning Service (NSWWS) criteria	Thresholds, which vary regionally, and is applied as proxy for Low/Medium/High impacts	Informative for estimating periods of increased risk of river flooding (fluvial flooding only). Note these are not appropriate for assessing pluvial flooding (surface water flooding).
Drought Severity Index	This index is not threshold based. Instead, it is calculated with 3-, 6-, 12- and 36-month rainfall deficits	3-month – meteorological/agricultural drought, a short intense dry period leading to low soil moisture levels impacting crop growth. 6-month – longer and higher impact agricultural drought. 12-month – hydrological drought leading to low water supply nationwide. 36-month – longer term hydrological drought leading to low water supply nationwide.
Wind gusts meeting National Severe Weather Warning Service (NSWWS) criteria	Thresholds, which vary regionally, and is applied as proxy for Low/Medium/High impacts	Transport disruption, infrastructure or building damage, bridge collapse and trees falling
Growing Degree Days	Daily Mean temperature above 5.5°C	Energy available for plant growth over a year. Not a measure of season length.
Heating Degree Days	Daily Mean temperature below 15.5°C	Indicator of meteorological contribution to energy demand for heating.
Cooling Degree Days	Daily Mean temperature above 22°C	Indicator of meteorological contribution to energy demand for cooling.

Exceedance metrics are currently used in Met Office National Severe Weather Warning Service (NSWWS), which issues warnings when severe weather has the potential to impact the UK. These warnings are based on a combination of the likelihood of the weather event occurring at any location, and the severity of the impacts if that event happens, where the severity is based on historical links between severe weather and its impacts in different parts of the country. This geographic variation in thresholds reflects the variations in both the exposure of populations and infrastructure, as well as the vulnerability of natural and human systems to these extreme conditions.

Alongside the NSWWS, the Met Office also produces a ‘heat health watch’ service for health professionals, contingency planners and emergency responders for planning purposes. This metric is based on exceedance of daily maximum temperature thresholds for least 3 consecutive days. Again, the thresholds vary geographically, as shown in Figure A2.1 and take account of the urban heat island effects in London.

https://www.metoffice.gov.uk/binaries/content/gallery/metofficegovuk/images/weather/learn-about/weather/uk-heatwave-threshold-temperatures.png

The World Climate Research Programme (WCRP) and World Meteorological Organization (WMO) Expert Team on Climate Change Detection and Indices (ETCCDI) have defined a set of 27 core indices (the ‘ETCCDI’ indices^[14]) which can be derived from land surface observations of daily temperature and precipitation. A selection has been used in the Met Office’s State of the UK Climate reports (Kendon et al., 2018 and McCarthy, 2018) to study observed changes in the UK climate. These can be used in CCRA3 to inform possible changes in the frequency of threshold exceedance, which may have an impact on natural, human and business systems. These differ from the NSWWS metrics which focus on extreme events.

2.2km CPM timeslice scenarios for 1981-2000 and 2021-2040 to provide new estimates of the baseline and near-term climate change risks associated with extreme events at the local scale. For example, based on 3 hourly rainfall, exceedances of 30mm/hour (and potentially 100mm/hour) at any location are currently being studied with respect to surface flooding.

Table A2.2, based on the NSWWS, shows the threshold exceedances that are likely to have a high impact today, and provide a suitable baseline for considering high impact exceedances in the future, where adaptation may be required. The metric refers to any location in the geographical region. Daily rainfall is used to inform forecasters on the possibility of river flooding and these metrics would be applicable to the UKCP18 12km simulations.

Table A2.2 Met Office National Severe Weather Warning Service (NSWWS) thresholds
NSWWS Index based on 12km UKCP18	Geographical Region	Exceedance Metric
Max 10m Wind gust (winter)	South East England	≥ 70 mph
Max 10m Wind gust (winter)	Highlands and Islands	≥ 90 mph
Max 10m Wind gust (winter)	Rest of the country	≥ 80 mph
Max 10m Wind gust (summer)	South East England	≥ 65 mph
Max 10m Wind gust (summer)	Highlands and Islands	≥ 80 mph
Max 10m Wind gust (summer)	Rest of the country	≥ 70mph
24-hour precipitation	England and Wales	≥ 80 mm
24-hour precipitation	Northern Ireland	≥ 80 mm
24-hour precipitation	NW Scotland	≥ 80 mm
24-hour precipitation	SW Scotland	≥ 65 mm
24-hour precipitation	South and East Scotland	≥ 55 mm
24-hour precipitation	NE Scotland	≥ 75 mm
Maximum Temperature	Geographically varying (see Figure)	≥ 25-28^oC for 3 consecutive days

The ETCCDI indices (Table A2.3) identify aspects of our changing climate that influence the functioning of natural ecosystems and aspects of demands on infrastructure. They complement the NSWWS extreme indices.

Table A2.3 World Meteorological Organization Expert Team on Climate Change Detection and Indices (ETCCDI) Indices
ETCCDI Index based on 12km UKCP18	Variable	Threshold	Count
Frost Days	Daily Minimum Temperature	< 0 °C	Days below this threshold
Icing Days	Daily Maximum Temperature	< 0 °C	Days below this threshold
Tropical Nights	Daily Minimum Temperature	> 20 °C	Days above this threshold
Summer Days	Daily Maximum Temperature	> 25 °C	Days above this threshold
Growing Degree Days	Daily Mean Temperature	> 5.5 °C	Degrees above this threshold per day
Heating Degree Days	Daily Mean Temperature	< 15.5 °C	Degrees below this threshold per day
Cooling Degree Days	Daily Mean Temperature	> 22 °C	Degrees above this threshold per day

Footnotes

[1] Natural variability describes the variations in weather and climate that we experience from day-to-day, year-to-year and decade-to-decade. They occur due to internal processes in the climate system associated with the atmospheric (such as weather patterns) and oceanic circulation (such as El Niño). They also include intermittent impacts from explosive volcanic eruptions and associated cooling by aerosols.

[2] Atmospheric rivers are relatively long, narrow regions that form on a strong jet stream and are characterized by intense moisture transport, which, on landfall, produce excessive precipitation that can lead to major flooding (e.g., Lavers et al., 2011). Storms Desmond and Dennis in 2015 and 2020, respectively, are examples of atmospheric river events.

[3] There is an important distinction between prediction and projection. Prediction starts from an observed initial state of the weather and climate system and aims to forecast how the climate system will evolve in the coming days to years (such as in numerical weather prediction and monthly to decadal climate prediction). Projection is simply a simulation of how the climate system might behave in response to changing imposed external forcings such as greenhouse gas and aerosol concentrations.

[4] Boundary conditions refer to the time-varying atmospheric conditions that enter the regional model domain and come from the global climate model in which the regional model is embedded. This means that, to a large extent, the regional model is ‘slave’ to the global model’s simulation of weather and climate variability; this is why the choice of global driving model is so important.

[5] Tidal locking describes the impact of tides on the ability of rivers to drain out to the sea. In very low-lying areas such as the Somerset Levels, the fall on the rivers may be so small that only during the lowest parts of the tidal cycle is the river able to drain to the sea. This was a major factor in the Somerset floods during the 2013/14 winter. This tidal locking will increase as mean sea level rises.

[6] The median of the RCP8.5 projections with the CMIP5 multi-model ensemble is consistent with a scenario of 4°C global warming at the end of the century: see Chapter 2: Watkiss and Betts (2021).

[7] Natural Hazard Partnership (NHP): The NHP provides authoritative and consistent information, research and analysis on natural hazards for the development of more effective policies, communications and services for civil contingencies, governments and the responder community across the UK. It is delivered through a partnership between academia, research organisations, public sector bodies and government departments. See http://naturalhazardspartnership.org.uk/

[8] National Severe Weather Warning Service (NSWWS): The NSWWS is a service provided by the Met Office to warn the public and emergency responders of severe or hazardous weather which has the potential to cause danger to life or widespread disruption. See https://www.metoffice.gov.uk/weather/guides/severe-weather-advice

[9] WMO Expert Team on Climate Change Detection and Indices (ETCCDI): The ETCCDI has the mandate to address the need for the objective measurement and characterization of climate variability and change. The team provides international coordination and collaboration on climate change detection and the indices relevant to climate change detection, and encourages the comparison of modelled data and observations. See: https://www.wcrp-climate.org/etccdi

[10] Heating Degree Days (CDD) are calculated as the product of the number of days below 15.5^oC and the number of degrees below 15.5^oC on each day where the temperature falls below that threshold.

[11] Growing Degree Days (GDD) are calculated as the product of the number of days above 5.5^oC and the number of degrees above 5.5^oC on each day where that threshold is exceeded.

[12] Cooling Degree Days (CDD) are calculated as the product of the number of days above 22^oC and the number of degrees above 22^oC on each day where that threshold is exceeded.

[13] https://tntcat.iiasa.ac.at/RcpDb/dsd?Action=htmlpage&page=compare

[14] http://etccdi.pacificclimate.org/list_27_indices.shtml

Downloads

Technical Report (CCRA3-IA) – Chapter 1

Type: PDF (10 MB)

Open Technical Report (CCRA3-IA) – Chapter 1(PDF)

Share on:

Cookies on this website