The formulation of robust policies for mitigation of, and adaptation to, climate change requires quantitative understanding of how and why specific changes are unfolding in the Earth System, and what might happen in the future. Quantitative explanation (or “attribution”) of observed changes – through robust process-based detection and attribution – is also fundamental to specification of confidence in climate assessments, predictions and projections. However, the capacity to deliver these capabilities is very immature. The WCRP Lighthouse Activity on Explaining and Predicting Earth System Change (EPESC) is intended to address this gap, with an overarching objective: to design, and take major steps toward delivery of, an integrated capability for quantitative observation, explanation, early warning and prediction of Earth System change on global and regional scales, with a focus on multi- annual to decadal timescales.

The research challenges are organised around 3 themes:

Theme 1: Monitoring and Modelling of Earth System Change
Tighter integration between the global climate observing system and the climate modelling community is necessary to address a number of interrelated challenges. These include (i) understanding and quantifying the uncertainty in key climate metrics, focusing on interannual to decadal climate variations; (ii) providing a quantitative framework for designing or optimizing an observation system suitable for detecting and monitoring interannual to decadal climate variations; and (iii) understanding and overcoming persistent Earth System model and re-analysis biases through the use of comprehensive estimation methods that bring modelling and (re-)analysis closer together and lead to a better usage of the diverse, heterogenous observing networks underlying the Global Climate Observing System (GCOS). The joint consideration of observation and modelling challenges provides a conceptual framework for identifying major gaps and opportunities for progress in both monitoring and observing Earth System variability and change. The benefits of this tighter integration will first be explored by focusing on a number of specific case studies. These case studies will be used to develop a systematic methodology that can be applied subsequently to assess a wider set of events.

Theme 2: Integrated Attribution, Prediction and Projection (including early warning and the potential for abrupt change)
Multi-annual forecasts are now routinely issued on the WMO Lead Centre for Annual to Decadal Climate Prediction website and in the WMO Global Annual to Decadal Climate Update. However, improved understanding and attribution of predicted signals is needed to gain further confidence in the forecasts. The objectives of Theme 2 are (i) to establish and apply attribution methodologies to help explain multi-annual to decadal changes in the climate system and (ii) to design and build an operational capability using these attribution methods. Outputs of this system will be integrated with forecasts and State of the Climate reports issued by WMO.

Theme 3: Assessment of Current and Future Hazards
The goal of Theme 3 is to understand (i.e., explain), quantify and predict (or project) changes on multi-annual-to-decadal timescales in the characteristics and statistics of weather and climate hazards (such as: tropical and extratropical cyclone frequency, intensity and paths; drought duration and severity; floods; heatwaves; cold air outbreaks). Major objectives are: (i) quantifying the current likelihood of specific weather and climate hazards (ii) quantifying changes in weather and climate hazards on multi-annual to decadal timescales; (iii) understanding the processes connecting changes in hazards to natural and anthropogenic drivers of climate variability and change; and (iv) advancing capabilities to predict and project changes in hazards.

Successful delivery of outcomes from the WCRP Explaining and Predicting Earth System Change Lighthouse Activity will require close collaboration with many different groups within WCRP and with key external partners.

EUMETSAT and its network of Satellite Application Facilities have released close to 100 Climate Data Records since starting sustained climate monitoring activities. Reprocessing was most often from the original sensor data, covering long time-series (~40 years) and several generations of instruments. According to the ECV Inventory of the joint CEOS-CGMS Working Group on Climate, the data records produced by EUMETSAT provide inputs to 19 out of 37 ECVs rated observable from space. As Earth-system reanalyses and climate model reconstructions expand to cover more components of the Earth system, they gradually incorporate more information from these historical sensor time-series into ever-higher resolution model-gridded datasets, via direct assimilation and/or via forcing datasets.

In spite of these progresses, gaps remain between the observations on the one side and the models on the other side. This paper presents a method to quantify, detect, and expose these disagreements, such that they can support data providers and model developers in their efforts to improve their respective products. On the observations side, the issues posed by originally inconsistent data processing artefacts and lack of quality control were addressed through climate data reprocessing. A selection of EUMETSAT Fundamental Climate Data Records is considered, from over 40 instruments (past and present), including microwave and infrared sounders and imagers. In order to remain as close as possible to the original instrumental record and its uncertainties, the method considers radiance brightness temperature level-1 data records. Departures from these records are first calculated using radiative transfer models and the latest Copernicus Climate Change Service (C3S) reanalysis ERA5, whose data are accessed via the joint ECMWF-EUMETSAT cloud infrastructure called the European Weather Cloud.

An analysis of the differences thus computed is then conducted using machine learning techniques. Interpretation of the results helps pointing to deficiencies in both observations and models that need addressing if one is intent on improving the agreement between observations and models. For example, a Principal Component Analysis highlights how measurements from geostationary satellites are affected by geographical displacements around the nominal position at the equator. Similarly, regional but persistent differences with ERA5 variability are found in the South Atlantic for the infrared window channels. The approach presented in this paper should be of interest to large-data applications that aim to exploit jointly Earth Observation satellite data and model simulations over long time periods. It also provides a possible extension of the classical evaluation comparisons of model data to observations, as performed in Obs4MIPs at geophysical level, with the additional advantage of a better understanding of observational uncertainty.

The Climate Modelling User Group (CMUG) brings a modelling perspective to ESA’s Climate Change Initiative (CCI) programme. CMUG is now in its third phase and works closely with the CCI+ projects. There are twenty-three CCI+ Essential Climate Variables (ECVs), each has its own project devoted to the production of a long-term, good quality, consistent dataset. These datasets are evaluated by CMUG through their application in climate models and in long time-series climate reanalyses.

CMUG also promotes the use of the ECV datasets by contributing to development of data analysis tools, which allow easy access to, and manipulation of, the data by climate scientists, modellers and researchers. Wider use is also encouraged by facilitating communication between the data producers and the data users through various forums such as the Climate Science Working Group (CSWG). CMUG’s work helps to ensure that the CCI ECV datasets meet the needs of the climate research and climate modelling communities.

Results of the latest CMUG experiments investigating quality and consistency of the CCI ECVs will be presented along with progress in CMUG’s work on tools to facilitate the use of the data by climate modellers, namely the observations for Model Intercomparisons (obs4MIPs) data base and the Earth System Model Evaluation Tool (ESMValTool). The use of CCI ECV data for climate services will also be discussed and an outlook for future work based on the CMUG Foresight report will be given.

Reliable estimates of terrestrial carbon fluxes are crucial for the successful implementation of the Global Stocktake under the Paris Agreement. However, current estimates of terrestrial carbon sources and sinks by process-based and semi-empirical models exhibit large uncertainties [1]. The spread in model-based estimates is partly caused by different parameterizations of soil and vegetation processes and different assumptions regarding the amount of carbon contained in vegetation and soil per unit area (=carbon density). Here we aim at reducing these uncertainties by assimilating a novel global observation-based time series of 21st century carbon densities of global woody vegetation [2] into the semi-empirical bookkeeping model BLUE [3]. Our novel approach allows us to disentangle the 21st century observation-based carbon emissions from the terrestrial vegetation into anthropogenic and environmental contributions, which is achieved by employing two different setups of the model-data integration. The assimilation of observational data on woody vegetation carbon allows us to include all impacts on carbon fluxes related to woody vegetation, including processes that are commonly not considered in model-based approaches (e.g. forest degradation). Further, we identify sources of uncertainty in the bookkeeping model BLUE by comparing the woody vegetation carbon stocks from the observed dataset to the estimates from our data assimilation approach.
The results from our different model setups show that the consideration of the effects of synergies between environmental effects (e.g. increasing atmospheric CO2, wildfires, climate) on vegetation carbon leads to much higher global emissions from land use and (land-use induced) land cover changes (=ELUC) compared to only considering anthropogenic effects on vegetation carbon. This effect is currently excluded from common budgeting of anthropogenic influences on the carbon cycle. The spread in ELUC between estimates based on our assimilation approach and other models is reduced by up to 87% compared to earlier estimates of the bookkeeping model, i.e. multi-model uncertainties are substantially reduced. Despite the improvement in ELUC estimates in our data assimilation approach, our analysis reveals that the land use forcing and/or its implementation in the bookkeeping model contribute strongly to uncertainties in the estimated carbon fluxes.
Considering only carbon fluxes from woody vegetation due to environmental processes (=natural carbon sink), our data assimilation approach shows a much higher interannual variability in the natural carbon sink and a stronger reduction in sink capacity in response to extreme events than estimates based on 13 process-based vegetation models [4].
Our results highlight the potentials of using novel observational datasets as constraints on models to identify the main sources of uncertainty in models and to reduce uncertainties in the estimated carbon fluxes under the Global Stocktake.

References

1. Friedlingstein, P., et al., Global Carbon Budget 2020. Earth System Science Data, 2020. 12(4): p. 3269-3340.
2. Xu, L., et al., Changes in global terrestrial live biomass over the 21st century. Sci Adv, 2021. 7(27).
3. Hansis, E., S.J. Davis, and J. Pongratz, Relevance of methodological choices for accounting of land use change carbon fluxes. Global Biogeochemical Cycles, 2015. 29(8): p. 1230-1246.
4. Friedlingstein, P., et al., Global Carbon Budget 2019. Earth System Science Data, 2019. 11(4): p. 1783-1838.

The Copernicus Climate Change Service (C3S) will soon deliver seasonal forecasts of ocean variables, such as ocean heat content, mixed layer depth and sea level. Knowledge of forecast skill is a prerequisite for utilizing forecast information. Assessing the skill of ocean variables from seasonal forecasts has remained elusive due to the lack of ocean verification datasets of sufficient quality and length. Fortunately, recent advances on temporally homogenized ocean observational records means that this is no longer the case. Within the H2020 EuroSea project, we use a subset of observable Essential Ocean/Climate variables (EOVs/ECVs) for verification of two seasonal forecasts systems contributing to the C3S seasonal multi-system product. The EOVs/ECVs are Ocean Heat Content (OHC), SLA and SST. The first is provided by the CMEMS GREP ensemble of ocean reanalyses whilst the other two are come from the ESA-CCI initiative, distributed by C3S. We evaluate the spatial distribution of skill in these variables, with a particular focus on the skill for user-relevant indicators.
A set of observable ocean indicators for monitoring and forecasting has been defined. These indicators target five sectoral applications: i) seasonal forecasts of weather statistics; ii) Climate Variability and Change; ii) Coastal Sea Level Rise; iv) Marine Health and v) Marine Productivity.
Preliminary results show that for most of the indicators the seasonal forecasts of SST clearly beat the persistence forecasts. This is also the case for OHC in many regions, but uncertainties in temperature initial conditions in the upwelling regions limits the assessment of forecast skill. Results also highlight the importance of representing the decadal variability and trends in ocean heat content and sea level in the initial conditions. This is a non-negligible challenge for the ocean data assimilation systems used in the production of ocean initial conditions. The representation of decadal variability and trends is essential for decadal forecasts and climate projections. Therefore, the results from this analysis of seasonal forecasts are also very relevant for these efforts

In recent decades climate variability and change have caused impacts on natural and human systems on all continents. Observations are needed to understand and document these impacts and its causes. Such observations are increasingly based on remote sensing data from satellites which offer global scale and continuous coverage. Only long-term and consistent observations of the Earth system allow us to quantify climate variability and change and their impacts on the natural and human dimension.
Since approximately 20 years, the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT) Satellite Application Facility on Climate Monitoring (CM SAF, https://www.cmsaf.eu) develops capabilities for a sustained generation and provision of Climate Data Records (CDRs) derived from primarily operational meteorological satellites. The product portfolio of the CM SAF comprises long time series of Essential Climate Variables (ECVs) related to the energy and water cycles as defined by the Global Climate Observing System (GCOS). Currently available CM SAF CDRs include, among others, surface and top of the atmosphere radiative fluxes, cloud products, as well as latent heat flux/evaporation, precipitation and freshwater flux over the global ice-free oceans. Upcoming CDR versions will cover the new WMO reference from 1991-2020 and with this partly exceed a temporal coverage of more than 40 years. In order to serve applications with strong timeliness requirements, CM SAF also produces so-called Interim Climate Data Records (ICDRs), which are typically released within a few days of the observations. All products are well-documented, carefully validated and were externally reviewed prior to product release.
After a short introduction to CM SAF this presentation will introduce currently available CDRs and ICDRs from CM SAF. A focus will be on uncertainty characterisation and results from validation. Example applications using CM SAF data records will also be introduced and explained. Finally, the presentation will present an overview of the upcoming new editions of CM SAF CDRs and ICDRs.