The 2015 Paris Agreement sets out a global action plan to avoid dangerous climate change by limiting global warming to well below 2°C, whilst pursuing efforts to limit warming to 1.5°C.

However, predicting how the climate will change over the next 20-50 years, as well as defining emissions pathways to keep the world on track, requires a better understanding of how several human and natural factors will affect the climate in coming decades. These include how atmospheric aerosols affect the Earth’s radiation budget, and the roles of clouds and oceans in driving climate change.

CONSTRAIN, a consortium of 14 European partners, is investigating these factors, feeding them into climate models to reduce uncertainties in, and create improved climate projections for, the next 20-50 years, on regional as well as global scales.

It is also translating this new scientific understanding into an improved evidence base aimed at providing up-to-date scientific evidence for international climate policy, and supporting decisions on climate mitigation and adaptation.


CONSTRAIN is structured around three science knowledge gaps particularly relevant for climate projections on the 20-50 year timeframe and a fourth knowledge gap on the translation of new science in societal decisions

Knowledge Gap A:  The magnitude and pattern of effective radiative forcing

The objectives under this Knowledge Gap are to reduce uncertainty in key components of Effective Radiative Forcing (ERF), especially from CO2 and aerosol-cloud interactions, and the ERF time history; and to constraining the representation of rapid adjustment processes in climate models by a detailed examination of adjustment mechanisms, their uncertainties, and their role in driving atmospheric circulation changes.

For the first time, we have quantified and broken down ERF uncertainty within CMIP6 models.  We have also set out state-of-the-art understanding of rapid adjustments.

Key publications include:

Knowledge Gap B:  The magnitude of cloud feedbacks and the role of cloud-circulation coupling in determining the pattern of climate change and climate sensitivity

Objectives include improving understanding of cloud feedbacks using a hierarchy of models and numerical experiments focussed on how feedbacks are affected by patterns of surface temperature change and influence climate sensitivity; and reassessing and constraining low-cloud feedbacks using observations from the first field study specifically designed for this purpose and associated modelling activities.

Work to date has culminated in report on climate sensitivity and feedbacks in CMIP models as well as a report on mesoscale organisation of shallow convection.

Further objectives are to improve understanding of circulation responses to warming, and to test hypotheses for factors influencing the response of circulation systems, especially in the tropical and North Atlantic and neighbouring land areas e.g. the Caribbean and in Europe.

Two key papers have emerged:

Knowledge Gap C:  The manner in which ocean variability conditions the response of the climate system to effective radiative forcings on different timescales

The first objective is to determine the processes that control patterns of sea-surface temperature change on decadal to multidecadal timescales, and assess whether these can be constrained for
understanding future change. This is underway but will require more analysis of CMIP6 models.

Related papers include:

A set of plausible sea surface temperature projections – one dominated by the forced response, and one dominated by internal variability – has also been produced.  These have been used as input to developing a climate emulator that will better reflect the effects of climate feedback variations on temperature and rainfall projections.

A further objective is to quantify the impacts of time varying sea-surface temperature patterns on climate feedbacks and atmospheric circulation.  This work is underway and we expect useful publications, but the methods are being adjusted to look at the complementary role of ocean heat uptake – preliminary work has shown that this potentially counters some of the effects of surface temperature change.

A final objective to understand the mechanisms that control uptake of heat and carbon by the oceans on decadal to multidecadal timescales is underway and on target.

Knowledge Gap D:  Translation of insights and uncertainties in variability and forced response on 20-50 year timescales into improved projections and effective adaptation and mitigation policy decisions

The first objective under this Knowledge Gap is to provide analyses on Effective Radiative Forcing (ERF), climate projections, climate feedbacks and climate sensitivity as key science input to the IPCC AR6 process.  This has been achieved: CONSTRAIN produced 64 peer-reviewed publications, with a further 4 accepted, covering the topics set out under this objective as well as many other relevant aspects of climate science in time for the IPCC cut-off.  All of these CONSTRAIN papers are summarised in a briefing note on CONSTRAIN journal publications and their relevance to IPCC AR6.

A further objective aims to characterise and constrain estimates of climate sensitivity and better understand the relationship between different lines of evidence and role in climate projections.  A report has been completed setting out advances made by CONSTRAIN in understanding climate sensitivity. These include analysis of CMIP6 model results that support the finding that there is a substantial role for cloud feedbacks in some of the high sensitivity models. Detailed analysis of some of the key processes driving differences in sensitivity, assessed in some of the models run within the CONSTRAIN project, also highlight the need for more observational constraints of cloud processes.

Key CONSTRAIN publications that have increased understanding of climate sensitivity and the role of clouds in particular include:

Notably, six CONSTRAIN scientists contributed to the landmark Sherwood et al. (2020) assessment of climate sensitivity.  This assessment used different lines of evidence to narrow the ECS range to 2.3 – 4.5°C, providing a constraint on climate sensitivity in the global models, as well as potentially on the processes driving the sensitivity. The understanding gained on drivers of climate sensitivity is providing new avenues to reduce uncertainties in future warming projections.

There is also an objective to provide the science community with robust climate emulator tools to better characterise uncertainty in global and regional climate projections and their variation with emissions pathways. Progress has been made on emulator development, with a set of plausible sea surface temperature projections developed to inform emulator development.

Key CONSTRAIN publications include:

Next steps will clarify how temperatures and rainfall will respond, both globally and regionally, to future greenhouse gas emissions within a complex and changing climate system.

Finally, there are objectives to develop knowledge translation tools targeted at exploring informed mitigation pathways and robust adaptation decisions; and to improve the flow of climate science knowledge gains into stakeholder understanding and decision making, including carbon budgets, UNFCCC negotiations and the general public.

The main knowledge translations tools developed to date are the three CONSTRAIN Zero In reports, launched at the UNFCCC COP and Climate Dialogues events respectively. These summarise the latest understanding on key processes and concepts studied by CONSTRAIN, including the latest climate model results, near-term warming rates, and the remaining global carbon budget. The reports are underpinned by published CONSTRAIN research.  The main findings have been accompanied by infographics aimed at promoting wider understanding of complex climate science.

The Silicone pathway tool is also now openly available, providing Integrated Assessment Modellers with a means of establishing relationships between a commonly modelled emission (e.g. CO2) and a rarer emissions (e.g. N2O), and the ability to “infill” emissions pathways. The emulators described above will form the basis of further knowledge translations tools, particularly in terms of providing robust regional projections of temperature, precipitation, and rainfall to support adaptation and impact studies.