Impact case study database

Improvements in weather forecasting benefit society, both in economic and other terms, facilitating day-to-day planning for the public, agriculture, commerce, utility suppliers, aviation, and transport sectors, and improved preparation and response to extreme events such as storm damage, flooding and drought. Long term climate projections are required for formulation of cost-effective mitigation strategies and climate adaptation. SU research resulted in new global datasets of atmospheric aerosol and surface albedo, and these have been used by prominent meteorological agencies in Europe, Australia, the US and India, leading to improved modelling and forecasts. SU datasets have also been used as benchmarks for the ongoing IPCC 6^th Assessment.

Improvement in weather, climate, and air quality modelling

The Swansea ATSR (SU-ATSR) aerosol dataset (1995-2012) [ R2, R4] has been used by ECMWF to produce the high-resolution atmospheric composition product (CAMS Reanalysis, CAMSRA). The product combines satellite aerosol data with the ECMWF Integrated Forecast Model to produce a continuous spatial and temporal map of aerosol, and to derive the related quantity of atmospheric particulate matter (PM2.5, PM10) needed for air quality. Working with SU under the ESA Climate Change Initiative [ G4], experiments by ECMWF found inclusion of the SU_AATSR data in a version of the IFS model over a 3-month period showed an improved fit with independent observations over Africa, North and South America and Europe compared with the previous use of NASA MODIS data alone, leading to a global error reduction of approximately 30% [ C1] . ECMWF included the full 10-year dataset in the CASMRA reanalysis in 2018 [ C2]. The CAMSRA product serves a wide variety of users, providing expected conditions for regional air quality models, long-range transport of air pollution, radiative forcing for climate and downwelling radiation used to find sites suitable for solar energy generation. For example, in 2020 CAMSRA was used for improved modelling of global air quality and health analysis by the UK Met Office and World Health organisation [ C3].

SU research has led to improvements in the UK Met Office Unified Model (UM). The UM is a numerical model of the land, ocean and atmosphere that is used for weather forecasts across a range of timescales, used for daily forecasting for the UK, as well as the national meteorological agencies of India, Korea, the Philippines, Australia, Republic of South Africa and New Zealand. The same model is also configured to provide climate change projections to inform UK climate planning and contribution to the IPCC AR6. The Swansea background albedo and vegetation albedos [ R7] and the GlobAlbedo dataset [ R6] were included in the UM (GA6.1) in 2014 [ C4] and continued in the current model version (GA7.0) [ C5]. This resulted in consistent improvements in five-day temperature forecasting, with a bias reduction of 20% at the centre of continental land masses in Summer [ C4] . Evaluation by the Met Office shows the radiation changes were responsible for the vast majority of improvements in this field from this upgrade. Out of the radiation changes, the albedo change was responsible for roughly half of the improvement in temperature errors over North America (where biases were particularly large). Specifically, the evaluation by the UK Met office notes: “The impact of this change is to improve the surface energy budget of the model, which specifically improves near-surface temperature errors over continental land in the summer hemisphere” [ C4, p1503]. The albedo changes [ R6, R7] have also been implemented in model configurations publicly available for offline analysis (JULES Global Land GL 7.0/7.1) and for climate modelling by the UK, (HadGEM3-GC3) and have been used for their current submissions of climate forecasts to the IPCC [ C5, C6] .

Impact on satellite dataset providers and use in model verification.

The aerosol dataset has improved the global climate record available to develop and test climate models for IPCC. In 2016, the SU-ATSR aerosol dataset was accepted by the World Climate Research Programme (WCRP), under the Obs4MIPS initiative, to provide quality-assured global datasets for the verification of climate models to be submitted to the IPCC under the Coupled Model Intercomparison Project (CMIP) protocol [ C7, p12]. This dataset forms one of only three global benchmark aerosol datasets, and the only one from European instruments. The SU-ATSR and the new SU Sentinel-3 dataset (2017-2020) were additionally accepted in 2019 by the ECMWF Copernicus Climate Change Service (C3S) intended as a European contribution to climate model verification [ C8]. Under [ G3] SU worked closely with EUMETSAT to develop a satellite aerosol product from Sentinel-3, based on the SU-ATSR algorithm, to be delivered continuously in under 3 hours to meet the needs of operational users in weather and air quality forecast agencies. This resulted in August 2020 in the release of the first Sentinel-3 aerosol product routinely produced in near real time by EUMETSAT allowing access to global users both by internet and direct satellite downlink under the EUMETCAST system [ C9].

The SU-ATSR dataset has been directly used during the period 2016-2020 by national agencies, to provide verification and testing for 14 global models of aerosol optical properties to be submitted to the IPCC 6^th Assessment. The model benchmarking is necessary to provide credibility and bias estimates of climate models. The following national agencies have used the SU-ATSR dataset in this process: Norwegian Meteorological Institute; LSCE/IPSL, CEA-CNRS-UVS (France); NASA Goddard Space Flight Centre (USA); NOAA, Geophysical Fluid Dynamics Laboratory (USA); European Centre for Medium-Range Weather Forecasts (UK); Finnish Meteorological Institute; Royal Netherlands Meteorological Institute; UK Met Office and NASA Goddard Institute for Space Studies (USA). Comparisons are summarised in [ C10, C11]. In these model evaluations [ C11] noted the ATSR series provided the longer record length to 2000, while [ C10] highlighted the unique ability of the SU-ATSR dataset to provide reliable information not only on total aerosol amount but also on aerosol particle size, and how aerosol optical depth varies with wavelength, needed for full evaluation of radiation balance, while other datasets could only provide total aerosol amount at accuracy needed for benchmarking.

Finally, Swansea has enabled wider impact of research by facilitating use of methods, datasets and software tools by other practitioners in remote sensing. The surface reflectance algorithm and software developed by SU [ R4-R6] for the satellite instruments MERIS, (A)ATSR, SPOT-VGT, Sentinel-2, and Sentinel-3 have been made available through the open-source BEAM/SNAP software in collaboration with the companies of Brockmann Consult, ARGANS Ltd, and ACRI-ST [ G2, G6] . For example, this tool forms a fundamental part of the processing for ESA CCI Land Cover projects, allowing accurate atmospheric correction to improve land cover accuracy [ C12, p324], with benefit to a wide range of users of these global land cover products.