Search

The Copernicus Climate Change Service (C3S) provides open and free access to state-of-the-art climate data and tools for use by governments, public authorities, and private entities around the world. It is fully funded by the European Union and implemented by the European Centre for Medium-Range Weather Forecasts (ECMWF) together with public and private entities in Europe and elsewhere. With over 120,000 registered users worldwide, C3S has rapidly become an authoritative climate service in Europe and beyond, delivering quality-assured climate data and information based on the latest science. Established in 2014, C3S became fully operational in 2018 with the launch of its Climate Data Store, a powerful cloud-based infrastructure providing access to a vast range of global and regional information, including climate data records derived from observations, the latest ECMWF reanalyses, seasonal forecast data from multiple providers, and a large collection of climate projections. The system has been designed to be accessible to nonspecialists, offering a uniform interface to all data and documentation as well as a Python-based toolbox that can be used to process and use the data online. C3S publishes European State of the Climate reports annually for policy-makers, as well as monthly and annual summaries that are widely disseminated in the international press. Together with users, C3S develops customized indicators of climate impacts in economic sectors such as energy, water management, agriculture, insurance, health, and urban planning. C3S works closely with national climate service providers, satellite agencies, and other stakeholders on the improvement of its data and services.

The Copernicus Atmosphere Monitoring Service (CAMS), part of the European Union’s Earth observation program Copernicus, entered operations in July 2015. Implemented by the European Centre for Medium-Range Weather Forecasts (ECMWF) as a truly European effort with over 23,500 direct data users and well over 200 million end users worldwide as of March 2022, CAMS delivers numerous global and regional information products about air quality, inventory-based emissions and observation-based surface fluxes of greenhouse gases and from biomass burning, solar energy, ozone and UV radiation, and climate forcings. Access to CAMS products is open and free of charge via the Atmosphere Data Store. The CAMS global atmospheric composition analyses, forecasts, and reanalyses build on ECMWF’s Integrated Forecasting System (IFS) and exploit over 90 different satellite data streams. The global products are complemented by coherent higher-resolution regional air quality products over Europe derived from multisystem analyses and forecasts. CAMS information products also include policy support such as quantitative impact assessment of short- and long-term pollutant-emission mitigation scenarios, source apportionment information, and annual European air quality assessment reports. Relevant CAMS products are cited and used for instance in IPCC Assessment Reports. Providing dedicated support for users operating smartphone applications, websites, or TV bulletins in Europe and worldwide is also integral to the service. This paper presents key achievements of the CAMS initial phase (2014–21) and outlines some of its new components for the second phase (2021–28), e.g., the new Copernicus anthropogenic CO2 emissions Monitoring and Verification Support capacity that will monitor global anthropogenic emissions of key greenhouse gases.

Framed within the Copernicus Climate Change Service (C3S) of the European Commission, the European Centre for Medium-Range Weather Forecasts (ECMWF) is producing an enhanced global dataset for the land component of the fifth generation of European ReAnalysis (ERA5), hereafter referred to as ERA5-Land. Once completed, the period covered will span from 1950 to the present, with continuous updates to support land monitoring applications. ERA5-Land describes the evolution of the water and energy cycles over land in a consistent manner over the production period, which, among others, could be used to analyse trends and anomalies. This is achieved through global high-resolution numerical integrations of the ECMWF land surface model driven by the downscaled meteorological forcing from the ERA5 climate reanalysis, including an elevation correction for the thermodynamic near-surface state. ERA5-Land shares with ERA5 most of the parameterizations that guarantees the use of the state-of-the-art land surface modelling applied to numerical weather prediction (NWP) models. A main advantage of ERA5-Land compared to ERA5 and the older ERA-Interim is the horizontal resolution, which is enhanced globally to 9 km compared to 31 km (ERA5) or 80 km (ERA-Interim), whereas the temporal resolution is hourly as in ERA5. Evaluation against independent in situ observations and global model or satellite-based reference datasets shows the added value of ERA5-Land in the description of the hydrological cycle, in particular with enhanced soil moisture and lake description, and an overall better agreement of river discharge estimations with available observations. However, ERA5-Land snow depth fields present a mixed performance when compared to those of ERA5, depending on geographical location and altitude. The description of the energy cycle shows comparable results with ERA5. Nevertheless, ERA5-Land reduces the global averaged root mean square error of the skin temperature, taking as reference MODIS data, mainly due to the contribution of coastal points where spatial resolution is important. Since January 2020, the ERA5-Land period available has extended from January 1981 to the near present, with a 2- to 3-month delay with respect to real time. The segment prior to 1981 is in production, aiming for a release of the whole dataset in summer/autumn 2021. The high spatial and temporal resolution of ERA5-Land, its extended period, and the consistency of the fields produced makes it a valuable dataset to support hydrological studies, to initialize NWP and climate models, and to support diverse applications dealing with water resource, land, and environmental management. The full ERA5-Land hourly (Muñoz-Sabater, 2019a) and monthly (Muñoz-Sabater, 2019b) averaged datasets presented in this paper are available through the C3S Climate Data Store at https://doi.org/10.24381/cds.e2161bac and https://doi.org/10.24381/cds.68d2bb30, respectively.

The Observations for Model Intercomparison Project (Obs4MIPs) was initiated in 2010 to facilitate the use of observations in climate model evaluation and research, with a particular target being the Coupled Model Intercomparison Project (CMIP), a major initiative of the World Climate Research Programme (WCRP). To this end, Obs4MIPs (1) targets observed variables that can be compared to CMIP model variables; (2) utilizes dataset formatting specifications and metadata requirements closely aligned with CMIP model output; (3) provides brief technical documentation for each dataset, designed for nonexperts and tailored towards relevance for model evaluation, including information on uncertainty, dataset merits, and limitations; and (4) disseminates the data through the Earth System Grid Federation (ESGF) platforms, making the observations searchable and accessible via the same portals as the model output. Taken together, these characteristics of the organization and structure of obs4MIPs should entice a more diverse community of researchers to engage in the comparison of model output with observations and to contribute to a more comprehensive evaluation of the climate models. At present, the number of obs4MIPs datasets has grown to about 80; many are undergoing updates, with another 20 or so in preparation, and more than 100 are proposed and under consideration. A partial list of current global satellite-based datasets includes humidity and temperature profiles; a wide range of cloud and aerosol observations; ocean surface wind, temperature, height, and sea ice fraction; surface and top-of-atmosphere longwave and shortwave radiation; and ozone (O3), methane (CH4), and carbon dioxide (CO2) products. A partial list of proposed products expected to be useful in analyzing CMIP6 results includes the following: alternative products for the above quantities, additional products for ocean surface flux and chlorophyll products, a number of vegetation products (e.g., FAPAR, LAI, burned area fraction), ice sheet mass and height, carbon monoxide (CO), and nitrogen dioxide (NO2). While most existing obs4MIPs datasets consist of monthly-mean gridded data over the global domain, products with higher time resolution (e.g., daily) and/or regional products are now receiving more attention. Along with an increasing number of datasets, obs4MIPs has implemented a number of capability upgrades including (1) an updated obs4MIPs data specifications document that provides additional search facets and generally improves congruence with CMIP6 specifications for model datasets, (2) a set of six easily understood indicators that help guide users as to a dataset's maturity and suitability for application, and (3) an option to supply supplemental information about a dataset beyond what can be found in the standard metadata. With the maturation of the obs4MIPs framework, the dataset inclusion process, and the dataset formatting guidelines and resources, the scope of the observations being considered is expected to grow to include gridded in situ datasets as well as datasets with a regional focus, and the ultimate intent is to judiciously expand this scope to any observation dataset that has applicability for evaluation of the types of Earth system models used in CMIP.

This report outlines the structure of and summarizes the recommendations made at the 5th International Conference on Reanalysis (ICR5), attended by 259 participants from 37 countries, in Rome (Italy), on 13-17 November 2017. It first summarizes the conference structure. Then, the key recommendations of ICR5 are given for the five main conference topics: production; observations (data rescue and preparation); data assimilation methods; quality assurance of reanalysis; and applications in science, services, and policymaking. Lastly, five high-level recommendations are proposed to managing agencies on how best to advance the field of reanalyses, which serves tens of thousands of users, via enhanced research, development, and operations.

The ERA-Interim and JRA-55 reanalyses of synoptic data and several conventional analyses of monthly climatological data provide similar estimates of global-mean surface warming since 1979. They broadly agree on the character of interannual variability and the extremity of the 2015/2016 warm spell to which a strong El Nino and low Arctic sea-ice cover contribute. Nevertheless global and regional averages differ on various time-scales due to differences in data coverage and sea-surface temperature analyses; averages from those conventional datasets that infill where they lack direct observations agree better with the averages from the reanalyses. The latest warm event is less extreme when viewed in terms of atmospheric energy, which gives more weight to variability in the Tropics, where the thermal signal has greater vertical penetration and latent energy is a larger factor. Surface warming from 1998 to 2012 is larger than indicated by earlier versions of the conventional datasets used to characterize what the Fifth Assessment Report of the Intergovernmental Panel on Climate Change termed a hiatus in global warming. None of the datasets exhibit net warming over the Antarctic since 1979. Centennial trends from the conventional datasets, HadCRUT4 on the one hand and GISTEMP and NOAAGlobalTemp on the other, differ mainly because sea-surface temperatures differ. Infilling of values where direct observations are lacking is more questionable for the data-sparse earlier decades. Change since the eighteenth century is inevitably more uncertain than change over and after a modern baseline period. The latter is arguably best estimated separately for taking stock of actions to limit climate change, exploiting reanalyses and using satellite data to refine the conventional approach. Nevertheless, early in 2016 the global temperature appears to have first touched or briefly breached a level 1.5 °C above that early in the Industrial Revolution, having touched the 1.0 °C level in 1998 during a previous El Nino.

Skin temperature (Tskin) derived from infrared sensors on board satellites provides a continuous view of Earth’s surface day and night and allows for the monitoring of global temperature changes relevant for climate trends. Tskin from the Infrared Atmospheric Sounding Interferometer (IASI) has not been properly exploited to date to assess its long-term spatio-temporal variability and no current homogenous Tskin record from IASI exists. In this study, we present a fast retrieval method of Tskin based on an artificial neural network from a set of IASI channels selected using the information theory/entropy reduction technique. We compare and validate our IASI Tskin product with that from EUMETSAT Level 2, ECMWF Reanalysis ERA5, SEVIRI land-surface temperature products, as well as ground measurements. Our results show good correlation between the IASI neural network product and the datasets used for validation, with a standard deviation between 1 and 4 °C. This method can be applied to other infrared measurements, and allows for the construction of a robust Tskin dataset, making it suitable for trend analysis.

The ECMWF twentieth century reanalysis (ERA-20C; 1900–2010) assimilates surface pressure and marine wind observations. The reanalysis is single-member, and the background errors are spatiotemporally varying, derived from an ensemble. The atmospheric general circulation model uses the same configuration as the control member of the ERA-20CM ensemble, forced by observationally based analyses of sea surface temperature, sea ice cover, atmospheric composition changes, and solar forcing. The resulting climate trend estimations resemble ERA-20CM for temperature and the water cycle. The ERA-20C water cycle features stable precipitation minus evaporation global averages and no spurious jumps or trends. The assimilation of observations adds realism on synoptic time scales as compared to ERA-20CM in regions that are sufficiently well observed. Comparing to nighttime ship observations, ERA-20C air temperatures are 1 K colder. Generally, the synoptic quality of the product and the agreement in terms of climate indices with other products improve with the availability of observations. The MJO mean amplitude in ERA-20C is larger than in 20CR version 2c throughout the century, and in agreement with other reanalyses such as JRA-55. A novelty in ERA-20C is the availability of observation feedback information. As shown, this information can help assess the product’s quality on selected time scales and regions.

The European Centre for Medium-Range Weather Forecasts (ECMWF) has developed a coupled assimilation system that ingests simultaneously ocean and atmospheric observations in a coupled ocean–atmosphere model. Employing the coupled model constraint in the analysis implies that assimilation of an ocean observation has immediate impact on the atmospheric state estimate, and, conversely, assimilation of an atmospheric observation affects the ocean state. In this context, observing system experiments have been carried out withholding scatterometer surface wind data over the period September–November 2013. Impacts in the coupled assimilation system have been compared to the uncoupled approach used in ECMWF operations where atmospheric and ocean analyses are computed sequentially. The assimilation of scatterometer data has reduced the background surface wind root-mean-square error in the coupled and uncoupled assimilation systems by 3.7% and 2.5%, respectively. It has been found that the ocean temperature in the mixed layer is improved in the coupled system, while the impact is neutral in the uncoupled system. Further investigations have been conducted over a case of a tropical cyclone when strong interactions between atmospheric wind and ocean temperature occur. Cyclone Phailin in the Bay of Bengal has been selected since the conventional observing system has measured surface wind speed and ocean temperature at a high frequency. In this case study, the coupled assimilation system outperforms the uncoupled approach, being able to better use the scatterometer measurements to estimate the cold wake after the cyclone.

Aerosols affect the atmosphere through direct interaction with short-wave and long-wave radiation and the microphysical properties of clouds. In this paper we report in detail on several mechanisms by which the short-term impact of dust on surface radiative fluxes can affect the dust loading of the atmosphere via modification of boundary-layer meteorology. This in turn affects the aerosol radiative forcing itself. Examples of these feedbacks between dust and boundary layer meteorology were observed during a series of dust storms in the Sahara and the eastern Mediterranean in April 2012. These case studies have been analysed using the Monitoring Atmospheric Composition and Climate – Interim Implementation (MACC-II) system. The radiative fluxes in the short-wave and long-wave spectra were both significantly affected by the prognostic aerosol–radiation interaction, which in turn impacted the meteorological simulation. Reduced incoming solar radiation below the aerosol layers led to a decrease in maximum surface temperatures and to a more stable thermal stratification of the lower atmosphere. This in turn forced weaker surface wind speeds and eventually smaller dust emissions. Moreover, we also observed a secondary impact of the aerosol radiative forcing, whereby horizontal gradients of surface temperature were increased at the edge of the dust plume, which led to local increases of surface wind speeds due to the thermal wind effect. The differentiated impact of the aerosol layer on surface pressure also contributed to the increase in surface wind speed and dust production in the same area. Enhanced long-wave radiative fluxes by the dust mass were associated with opposite processes. Less stable thermal stratification at night, brought mainly by higher minimum temperatures at the surface, caused stronger surface winds. At the edge of the dust storm, weaker horizontal temperature and pressure gradients forced lower winds and reduced dust production. Regarding dust emissions, short-wave radiative forcing had a larger impact than long-wave radiative forcing, corroborating several previous studies. For surface temperature, short-wave and long-wave contribution were close in intensity. These feedbacks were amplified when using data assimilation to build the aerosol analysis of the MACC-II global system. This led to an improvement in the short-term forecasts of thermal radiative fluxes and surface temperatures.

Surface skin temperature (Tskin) derived from infrared remote sensors mounted on board satellites provides a continuous observation of Earth's surface and allows the monitoring of global temperature change relevant to climate trends. In this study, we present a fast retrieval method for retrieving Tskin based on an artificial neural network (ANN) from a set of spectral channels selected from the Infrared Atmospheric Sounding Interferometer (IASI) using the information theory/entropy reduction technique. Our IASI Tskin product (i.e. TANN) is evaluated against Tskin from EUMETSAT Level 2 product, ECMWF Reanalysis (ERA5), SEVIRI observations, and ground in situ measurements. Good correlations between IASI TANN and the Tskin from other datasets are shown by their statistic data, such as a mean bias and standard deviation (i.e. [bias, STDE]) of [0.55, 1.86 °C], [0.19, 2.10 °C], [-1.5, 3.56 °C], from EUMETSAT IASI L-2 product, ERA5, and SEVIRI. When compared to ground station data, we found that all datasets did not achieve the needed accuracy at several months of the year, and better results were achieved at nighttime. Therefore, comparison with ground-based measurements should be done with care to achieve the ±2 °C accuracy needed, by choosing, for example, a validation site near the station location. On average, this accuracy is achieved, in particular at night, leading to the ability to construct a robust Tskin dataset suitable for Tskin long-term spatio-temporal variability and trend analysis., SCOPUS: ar.j, info:eu-repo/semantics/published

The Anthropocene represents discontinuity in the tight connection that has always linked survival of human societies to the environment around them. But alongside the undeniable human-induced alteration of the global environment the last couple of centuries have given humans an unprecedented ability to monitor, simulate and predict the climate around us. It is over the course of the last two centuries that technological developments have allowed society to both alter the environment on a global scale and predict how the earth system responds to external and internal changes. Such a set of circumstances calls for a profound rethink of the relationship mankind has with the planet, something that would be impossible without a systematic, and operational production of datasets, tools and information about the past, current and future status of the global climate. Copernicus Climate Change Service aims to fill this gap whilst promoting the transition towards a climate smart society.

To better understand and manage climate risks in climate-sensitive sectors such as agriculture, it is essential to have access to consistent and reliable data and information products. Tailoring these products to the needs of the users they want to serve facilitate informed decision-making and downstream applications. This requires an in-depth understanding of users' needs and the context in which these users operate. Considering the diversity of the economic sectors and their actors it is extremely challenging if not outright impossible to promote the emergence of climate services without empowering a plethora of intermediate users who can act as one of the steps in a potential long knowledge brokers chain that connect the climate data providers and the end-users. In this context, Copernicus Climate Change Service (C3S) has been designed around the Climate Data Store (CDS), a unique entry point to a huge variety of quality-controlled climate data and high-level utilities to process that data to develop user-driven applications. Through the Sectoral Information System, C3S has then developed a series of sector specific applications, which show how the infrastructure can be used to address specific users’ needs. This paper presents the key elements of the CDS and selected cases of sectoral application of C3S in agriculture. Keywords: Copernicus Climate Change Services (C3S), Climate Data Store (CDS), User-driven applications, Agriculture sector

This paper provides an overview of the European Union's Earth Observation Programme, Copernicus, dedicated at monitoring the planet Earth and its environment for the benefit of all European citizens. The Programme offers information services based on satellite Earth Observation, in situ (non-space), and modelling data. The paper focuses on one of the six Copernicus Services, the Copernicus Climate Change Service, which provides comprehensive climate information covering a wide range of components of the Earth-system (atmosphere, land, ocean, sea-ice and carbon), maximising the use of past, current and future earth observations (from in-situ and satellite observing systems) in conjunction with modelling, supercomputing and networking capabilities to produce a consistent, comprehensive and credible description of the past, current and future climate, via the generation of a large number of Essential Climate Variables (ECVs).

International audience; Over the last 30 years, satellite data have been playing a key role for the monitoring of the Earth's atmosphere and climate, with an important French and European contribution. This article summarizes the successive stages of a space mission for Earth's observation, describes the geophysical variables that can be derived from the measurements, and presents the societal benefits associated with these observations.; Depuis une trentaine d'années, les données satellitaires jouent un rôle déterminant pour l'observation de l'atmosphère terrestre et du climat, avec une contribution française et européenne importante. Cet article reprend les différentes étapes de la définition d'une mission spatiale d'observation de la Terre, donne un aperçu des variables qui peuvent être restituées à partir des données mesurées et présente les applications sociétales qui bénéficient de ces observations.

This article reviews past and current reanalysis activities at the European Centre for Medium-Range Weather Forecasts (ECMWF) and describes plans for developing future reanalyses of the coupled climate system. Global reanalyses of the atmosphere, ocean, land surface, and atmospheric composition have played an important role in improving and extending the capabilities of ECMWF's operational forecasting systems. The potential role of reanalysis in support of climate change services in Europe is driving several interesting new developments. These include the production of reanalyses that span a century or more and the implementation of a coupled data assimilation capability suitable for climate reanalysis. Although based largely on ECMWF's achievements, capabilities, and plans, the article serves more generally to provide a review of pertinent issues affecting past and current reanalyses and a discussion of the major challenges in moving to more fully coupled systems.

In 2002, the European Space Agency (ESA) launched the ENVIromental SATellite (ENVISAT), which carried ten instruments to provide continuous observation of Earth's land, atmosphere, oceans and ice caps. During the satellite's lifetime, the European Centre for Medium Range Weather Forecasts (ECMWF) has routinely monitored a variety of products from many of its instruments. A subset of these data has also been assimilated in the ECMWF operational system, along with two of its applications: the reanalysis and the Monitoring Atmospheric Composition and Climate (MACC) systems. This article reviews those activities and summarizes the lessons learnt in monitoring and assimilating data from a research satellite within a numerical weather prediction system. The value of continuous data monitoring and the benefits that a close collaboration between data provider and data user can bring to both parties are highlighted. For observations that were assimilated, impact assessments on the ECMWF products have been performed periodically. Two cases are presented in this article. The first one shows that the assimilation of ocean wave information can reduce the wave-height model random error by up to 8% at the analysis time, with benefits at later forecast times extending to up to 5 days in the Tropics. The second example shows that the assimilation of two ENVISAT ozone products improves the agreement of the ozone analyses with independent ozone observations obtained from sondes and the Microwave Limb Sounder. Finally, the use of ENVISAT reprocessed data is presented, with an emphasis on the importance of data reprocessing and long-term data preservation as key activities to ensure the future usage of these datasets.

ERA-Interim is the latest global atmospheric reanalysis produced by the European Centre for Medium-Range Weather Forecasts (ECMWF). The ERA-Interim project was conducted in part to prepare for a new atmospheric reanalysis to replace ERA-40, which will extend back to the early part of the twentieth century. This article describes the forecast model, data assimilation method, and input datasets used to produce ERA-Interim, and discusses the performance of the system. Special emphasis is placed on various difficulties encountered in the production of ERA-40, including the representation of the hydrological cycle, the quality of the stratospheric circulation, and the consistency in time of the reanalysed fields. We provide evidence for substantial improvements in each of these aspects. We also identify areas where further work is needed and describe opportunities and objectives for future reanalysis projects at ECMWF. Copyright © 2011 Royal Meteorological Society

Singular-vector(SV)-based selective satellite data thinning is applied to the Southern Hemisphere (SH) extratropics to reduce analysis uncertainty and forecast error. For two seasons, the European Centre for Medium-Range Weather Forecasts (ECMWF) four-dimensional variational data assimilation system has been run in five different configurations with different satellite data coverage: two reference experiments used low-density and high-density coverage over the globe; in the SH two SV-based selective thinning experiments used low-density data everywhere apart from targeted regions; and one random-based thinning experiment used low-density data everywhere apart from randomly defined regions. The SV-based target regions have been defined either by daily operational SVs computed for the ECMWF Ensemble Prediction System, or by the previous year's mean seasonal distribution. Results indicate that the impact of the additional data largely depends on the season. Overall, forecast errors grow faster in the SH cold season than in the warm season. In the SH cold season, the general impact of adding data is smaller and the relative difference between the impact of the individual targeting experiments is smaller as well. In the cold season, the data assimilation system failed to extract the meteorological signal carried by the extra satellite data in sensitive regions. In the SH warm season, all experiments with more data produce a statistically more significant and longer-lasting positive impact on forecast skill. In this season, the SV-based targeting experiment performs best and as well as the reference experiment in which the data density is increased globally. Copyright © 2011 Royal Meteorological Society

As part of its Data User Element programme, the European Space Agency funded the GlobMODEL project which aimed at investigating the scientific, technical, and organizational issues associated with the use and exploitation of remotely-sensed observations, particularly from new sounders. A pilot study was performed as a ldquodemonstratorrdquo of the GlobMODEL idea, based on the use of new data, with a strong European heritage, not yet assimilated operationally. Two parallel assimilation experiments were performed, using either total column ozone or ozone profiles retrieved at the Royal Netherlands Meteorological Institute (KNMI) from the Ozone Monitoring Instrument (OMI). In both cases, the impact of assimilating OMI data in addition to the total ozone columns from the SCanning Imaging Absorption spectroMeter for Atmospheric CartograpHY (SCIAMACHY) on the European Centre for Medium Range Weather Forecasts (ECMWF) ozone analyses was assessed by means of independent measurements. We found that the impact of OMI total columns is mainly limited to the region between 20 and 80 hPa, and is particularly important at high latitudes in the Southern hemisphere where the stratospheric ozone transport and chemical depletion are generally difficult to model with accuracy. Furthermore, the assimilation experiments carried out in this work suggest that OMI DOAS (Differential Optical Absorption Spectroscopy) total ozone columns are on average larger than SCIAMACHY total columns by up to 3 DU, while OMI total columns derived from OMI ozone profiles are on average about 8 DU larger than SCIAMACHY total columns. At the same time, the demonstrator brought to light a number of issues related to the assimilation of atmospheric composition profiles, such as the shortcomings arising when the vertical resolution of the instrument is not properly accounted for in the assimilation. The GlobMODEL demonstrator accelerated scientific and operational utilization of new observations and its results prompted ECMWF to start the operational assimilation of OMI total column ozone data.

An observing system experiment, simulating a surface-only observing network representative of the 1930s, is carried out with three- and four-dimensional variational data assimilation systems (3D-VAR and 4D-VAR) and an ensemble-based data assimilation system (EnsDA). It is found that 4D-VAR and EnsDA systems produce analyses of comparable quality and that both are much more accurate than the analyses produced by the 3D-VAR system. The EnsDA system also produces useful estimates of analysis error, which are not directly available from the variational systems.

This paper assesses the value of targeted observations over the North Atlantic Ocean for different meteorological flow regimes. It shows that during tropical cyclone activity and particularly tropical cyclone transition to extratropical characteristics, removing observations in sensitive regions, indicated by singular vectors optimized on the 2-day forecast over Europe, degrades the skill of a given forecast more so than excluding observations in randomly selected regions. The maximum downstream degradation computed in terms of spatially and temporally averaged root-mean-square error of 500 hPa geopotential height is about 13%, a value which is 6 times larger than when removing observations in randomly selected areas. The forecast impact for these selected periods, resulting from degrading the observational coverage in sensitive areas, was similar to the impact found (elsewhere in other weather forecast systems) for the observational targeting campaigns carried out over recent years, and it was larger than the average impact obtained by considering a larger set of cases covering various seasons. Copyright © 2007 Royal Meteorological Society

Adding targeted observations to the array of observations routinely available to numerical weather prediction is generally found to give small impact, although studies have often focused on a very limited number of cases in field experiments. It is expected that observation targeting in sensitive areas may have more impact on forecasts ‘busts’. A forecast ‘bust’ in a given verification region is defined as a doubling of the mean monthly error for the day-two forecast. This does not occur very often in the operational ECMWF system. The primary purpose of this study is to produce three reference data-assimilation experiments: the first assimilation experiment uses all available observations (as they are used in the operational ECMWF analysis system). The second assimilation experiment systematically excludes all observations in the North Pacific oceanic region. The third assimilation experiment systematically excludes all observations in the North Atlantic oceanic region. These three experiments are then used to evaluate the potential of targeting strategies which are described in two separate papers (parts II and III of this study). Results, based on a summer and a winter season, indicate that on average: (1) Oceanic data are important to reduce day-2 forecast errors over the downstream land areas, the downstream impact of observations over the Pacific Ocean being larger than the downstream impact of observations over the Atlantic Ocean. (2) The removal of observations taken over the Pacific Ocean has generally a small impact on medium-range forecasts verified over Europe, but results strongly depend on the data-assimilation system used to assimilate the observations. (3) Forecasts started from 3D-Var analyses are worse than forecasts started from 4D-Var analyses and, given the current observing system, the relative impact of ocean observations is higher if a 3D-Var system is used instead of a 4D-Var system. In other words, 4D-Var is better capable of propagating information from data-rich to data-void areas, and is therefore less vulnerable to data gaps in a system which otherwise includes a large number of satellite observations. Copyright © 2007 Royal Meteorological Society

Data-assimilation experiments have been run in seven different configurations for two seasons to assess the value of observations taken in target regions identified either using singular vectors (SVs) or randomly, and located over the Pacific or the Atlantic Oceans. The value has been measured by the relative short-range forecast error reduction in downstream areas, specifically a North American region for observations taken in the Pacific Ocean, and a European region for observations taken in the Atlantic Ocean. Overall, results have indicated (1) that observations taken in SV-target areas are on average more valuable than observations taken in randomly selected areas, (2) that it is important that the daily set of singular vectors are used to compute the target areas, and (3) that the value of targeted observations depends on the region, the season and the baseline observing system. If the baseline observing system is data-void over the ocean, then the average value of observations taken in SV-target areas is very high. Considering for example winter 2004, SV-targeted observations over the Pacific (Atlantic) reduce the day-2 forecasts error of 500 hPa geopotential height forecasts in the verification region by 27.5% (19.1%), compared to 15.7% (14.9%) for observations taken in random areas. By contrast, if the baseline observing system is data-rich over the ocean, then the average value of observations taken in SV-target areas is rather small. Considering for example winter 2004, it has been estimated that adding SV-targeted observations over the Pacific (Atlantic) would reduce, on average, the day-2 forecasts error in the verification region by 4.0% (2.0%), compared to 0.5% (1.7%) for observations in random areas. These average results have been confirmed by single-case investigations, and by a careful examination of time series of forecast errors. These results indicate that more accurate assimilation systems that can exploit the potential value of localized observations are needed to increase the average return of investments in targeting field experiments. Copyright © 2007 Royal Meteorological Society

Emitted infrared limb radiances from the Michelson Interferometer for Passive Atmospheric Sounding (MIPAS) are for the first time assimilated directly in a global numerical weather prediction system, using a fast radiative transfer model that assumes local horizontal homogeneity. This study reports on the monitoring of the observed MIPAS radiances against model equivalents, and the subsequent assimilation experiments within a 4-dimensional variational data assimilation system. The impact of the limb radiance assimilation on the resulting analyses and forecasts is assessed by considering the fit to other observations assimilated in the system and by comparing the resulting analyses with independent data. The results demonstrate the feasibility of direct assimilation of emitted infrared limb radiances and highlight how information on stratospheric temperature, humidity, and ozone can be extracted from the radiances directly within the assimilation system. The assimilation of MIPAS radiances leads to considerable differences in the mean stratospheric analyses, without a significant degradation of the fit to other observations used in the assimilation. The assimilation of MIPAS limb radiances appears to correct temperature biases with an oscillatory structure above 10 hPa in the analyses, and the assimilation leads to a considerable moistening of the stratosphere, typically by 20–40%. For ozone, the assimilation causes an increase in the tropical ozone maximum in the analysis, and a reduction of ozone over the poles. The changes to the humidity and ozone fields are retained in the subsequent 10-day forecast. The changes introduced in the analyses are overall supported qualitatively and quantitatively by independent retrievals of temperature and humidity; results for ozone are more mixed. Analyses with MIPAS radiances agree better with independent ozone sondes or retrievals over the North Polar region, whereas over the Tropics the changes to the mean ozone analyses are not supported by other observations. The limb radiance assimilation shows considerable sensitivity to the bias correction applied to the limb radiances prior to the assimilation, suggesting that the assimilation would benefit from refinements in the bias correction method. Copyright © 2007 Royal Meteorological Society

Four-dimensional variational (4D-Var) assimilation forecast impact experiments were conducted with CHAMP radio occultation measurements from 1st June to 31st July 2004. The CHAMP bending angles were assimilated with one-dimensional (1D) and two-dimensional (2D) bending angle observation operators, enabling the potential benefits of the latter to be investigated. Two computationally affordable 2D bending angle observation operators were considered. The first can be viewed as a straightforward generalisation of the 1D operator. The second is based on a fourth order Runge–Kutta solution of the differential equations defining the ray-path through the refracting atmosphere. It is demonstrated that the computational cost of the 2D bending angle operators is not prohibitive and that they improve the root-mean-square of the observed minus background (O-B) bending angle departures by around 5% in the lower-troposphere of the northern and southern hemispheres. The O-B distributions in the tropical lower-troposphere are biased low, but the first-guess departure quality control procedure removes ∼2–3% of the observations and this reduces the bias by around 20%. The assimilation of the CHAMP measurements with both the 1D and 2D operators improves the analysis and forecast fit to radiosonde temperature measurements in the southern hemisphere on the 300, 200, 100 and 50 hPa levels, over the day-1 to day-5 forecast range. The forecast scores in the lower-troposphere for geopotential height and relative humidity are generally neutral, but there is a small improvement in the anomaly correlation of the 500 hPa geopotential height field in the southern hemisphere, over the day-1 to day-2 forecast range. However, there are no clear, statistically significant differences between the 1D and 2D experiments for this two-month period. Copyright © 2007 Royal Meteorological Society

The development of an assimilation system for radiance data from the Atmospheric InfraRed Sounder (AIRS) is described, in particular the identification of cloud contamination, bias correction and the characterization of errors in the measured radiances and radiative-transfer model. The results of assimilation experiments are presented. These show that a conservative use of AIRS radiance data (in a system already extensively observed with other satellite data) results in a small, but consistent, improvement in the quality of analyses and forecasts. Larger impacts of AIRS are found in hypothetical experiments that test the use of radiances from only a single sounding instrument. In these, the use of AIRS is found to outperform the use of data either from a single Advanced Microwave Sounding Unit-A (AMSU-A) or from a single High-resolution InfraRed Sounder (HIRS). In this hypothetical context the relative forecast performance of each sensor is found to correlate with the size and vertical scale of increments caused by the assimilation of the radiances. Copyright © 2006 Royal Meteorological Society.

A one-dimensional bending-angle observation operator for assimilating GPS radio occultation (RO) measurements has been integrated into the ECMWF four-dimensional variational assimilation (4D-Var) system. We have performed forecast impact experiments with 60 days of CHAMP RO measurements, in addition to the latest set of conventional and satellite data that are assimilated operationally, including radiances from the Atmospheric Infrared Sounder. It is demonstrated that the CHAMP measurements provide extremely good temperature information in the upper troposphere and lower stratosphere. In the southern hemisphere (SH), they produce a clear, statistically significant improvement in the r.m.s. forecast fit to radiosonde measurements over the day 1 to day 5 forecast range at 300, 200, 100 and 50 hPa. An improved r.m.s. fit to radiosondes is also evident at 100 hPa in the tropics. However, the observations degrade the 500 hPa geopotential height (500Z) field in the SH. This appears to be mainly caused by erroneous surface pressure increments in Antarctica. As a result, we have modified the GPS tangent-linear and adjoint routines, prior to the evaluation of the model-level pressures and geopotential heights, in order to remove the sensitivity of the model geopotential height values to the surface pressure. This improves the SH 500Z forecast scores, although a small degradation is still evident at the day 1 and day 2 forecast range. A simple method for estimating the degrees of freedom for signal (DFS) of a large variational assimilation system is noted and applied to estimate the DFS of the CHAMP measurements assimilated during a 12-hour assimilation window. The CHAMP measurements increase the total DFS of the 4D-Var system by ∼4%, and the DFS per CHAMP bending-angle profile is ∼34. Copyright © 2006 Royal Meteorological Society

Aerosols affect the atmosphere through the aerosol-radiation and the aerosol-clouds interactions. In this paper we report on a new mechanism whereby the radiative effect of dust aerosol on surface fluxes acts to increase the dust loading of the atmosphere via modification of boundary-layer stability, thereby acting to enhance the radiative aerosol effect. This positive feedback between dust aerosol and boundary layer stability occurred during a series of dust storms in the Sahara and the Eastern Mediterranean in April 2012, which were studied using the Monitoring Atmospheric Composition and Climate – Interim Implementation (MACC-II) system. The radiative fluxes in the shortwave and long-wave spectra were both significantly affected by the prognostic aerosols-radiation interation, which strongly influenced the meteorological simulation. Reduced incoming solar radiation below the aerosol layers caused a decrease in maximum surface temperatures, and consequently a more stable thermal stratification of the lower atmosphere. The increased thermal stability led to decreased surface wind speeds and therefore to smaller amounts of dust aerosol emissions. Larger downwelling long-wave fluxes were associated with the opposite processes: less stable thermal stratification at night, brought mainly by higher minimum temperatures at the surface, caused stronger surface winds. Overall, the impact by the long-wave radiative forcing was more important than the short-wave contribution. This feedback was amplified when taken into account in the aerosol analysis of the MACC-II global system. It lead to a notable improvement in short term forecast of short and long-wave radiative fluxes, of surface temperature but also of the aerosol burden itself. Forecasts of radiative fluxes in the shortwave and long-wave spectrum were also improved. At a longer range the improvement were less important as the forecast error of the aerosol load increased, thereby highlighting the importance of accurate aerosol representation in the study of aerosol-radiation interaction.

This paper describes the direct assimilation of water vapour (WV) clear-sky radiance (CSR) data from geostationary satellites within the context of the ECMWF four-dimensional variational assimilation (4D-Var) system. The assimilation of Meteosat–7 WV CSR data became operational on 9 April 2002. As 4D-Var includes a time dimension, the high temporal resolution of the geostationary radiance data can be exploited to provide information not only on the upper-tropospheric humidity but also on the upper-tropospheric wind field. The data assimilated have a spatial resolution of approximately 80 km and a time resolution of 1 hour. Extensive pre-operational monitoring of the CSR data has been carried out, showing a systematic warm bias of approximately 2–3 K in the Meteosat WV CSR data compared to radiances simulated from model first-guess fields. The systematic biases between the brightness temperatures derived from the model and the observations are investigated and compared to other satellite radiance data. For assimilation purposes this is accounted for by using a bias correction based on a statistical regression. The monitoring also shows contamination of certain time slots caused by intruding solar stray light and a certain degree of residual cloud contamination present in the CSR. Data quality control is introduced to exclude affected data. Initial single-cycle 3D-Var with first guess at appropriate time and 4D-Var experiments demonstrate both the direct and indirect effects of the Meteosat WV CSR data on the model fields, particularly of humidity and wind. The operational implementation of the assimilation of the Meteosat WV CSR data, including results from pre-operational experiments and forecast impacts, is discussed in a companion paper. Copyright © 2004 Royal Meteorological Society

The direct assimilation of water vapour (WV) clear-sky radiances (CSRs) from geostationary satellites within the ECMWF four-dimensional variational assimilation (4D-Var) became operational on 9 April 2002 with the assimilation of radiances from Meteosat-7. To extend the coverage provided by geostationary radiances, the derivation of a similar CSR product from the Geostationary Operational Environmental Satellites GOES-W and GOES-E was initiated and since 14 January 2003 these data have been operationally assimilated as well. This paper discusses results from the pre-operational impact experiments using Meteosat-7 and the subsequent operational implementation of the WV radiance assimilation. The pre-operational data monitoring of the CSRs shows contamination of certain time slots caused by intruding solar stray light and a certain degree of cloud influence present in the CSR. Data quality control is introduced to exclude affected data. When assimilated, the Meteosat WV CSRs correct the upper-tropospheric humidity field in areas of known model problems. While the analysis draws well to Meteosat data, the fit to other conventional observations does not degrade, and the fit to other satellite observations is noticeably improved. This is visible in statistics for the assimilated HIRS-12 as well as for passive Advanced Microwave Sounding Unit B (AMSUB) radiances, both in the pre-operational experiments and in the operational assimilation cycle. The impact on forecast quality is slightly positive to neutral for different areas of the globe. In some experiments a positive impact on upper-level wind fields (around 200 hPa) is seen, especially in the tropics. A relatively large sensitivity is noted of the mean increments and also forecast scores to the bias correction. Copyright © 2004 Royal Meteorological Society

This study investigates the impact of a new satellite-derived wind product over the polar regions on ECMWF's global four-dimensional variational assimilation system. The winds are derived at the University of Wisconsin— Madison by tracking structures in successive swaths from the polar-orbiting Moderate Resolution Imaging Spectroradiometer (MODIS) on the Terra satellite. The new wind product provides unprecedented coverage of the polar wind field. The data are assimilated subject to cautious quality control at lower levels and over land. The MODIS winds have a positive impact on medium-range global weather forecasts, particularly over the polar regions, but also elsewhere over the Northern Hemisphere. The mean polar wind analysis can be considerably altered as a result of the MODIS winds assimilation. Sensitivity calculations highlight the importance of polar regions for forecasts in the midlatitudes and indicate how MODIS winds can reduce key analysis errors in this region. The MODIS winds are n...

An improved version of the RTTOV (Radiative Transfer for the Television Infrared Observation Satellite (TIROS) Operational Vertical Sounder) fast radiative transfer model used operationally at the European Centre for Medium-Range Weather Forecasts (ECMWF) for the assimilation of Advanced TIROS Operational Vertical Sounder (ATOVS) radiances has been developed. This new model computes radiances for the Atmospheric Infrared Sounder and reproduces line-by-line radiances and Jacobians for the surface-sensing, water vapour and ozone channels of the ATOVS with significantly improved accuracy. The profile-dependent predictors used by the improved model to parametrize the atmospheric optical depths are based on the approach followed by the ECMWF fast radiative transfer model for the Infrared Atmospheric Sounding Interferometer (RTIASI). To improve the accuracy of the fast model in reproducing line-by-line radiances for the Infrared Atmospheric Sounder, modifications have been made to the predictors used in RTIASI by introducing a revised set of predictors for ozone and adding new predictors to model the water vapour continuum type absorption. To eliminate discontinuities in the water vapour Jacobians observed in RTIASI, data are now weighted prior to performing the regression. Copyright © 2004 Royal Meteorological Society

This study examines atmospheric motion vectors derived by tracking features in image sequences from Meteosat for anomalies during the second satellite eclipse period of 2001. During eclipse periods (March–April; September–October), data from geostationary sensors are prone to anomalies caused by solar stray light entering the radiometer around local midnight. As a result, imagery from geostationary satellites can exhibit local anomalies for certain time slots, and these anomalies can change considerably from one time slot to the next. The effect of these anomalies on atmospheric motion vectors is characterized in this study by investigating the temporal consistency of the data and by monitoring the winds against short-term forecasts from the European Centre for Medium-Range Weather Forecasts (ECMWF) global model. Atmospheric motion vectors from the Meteosat water vapor channel exhibit spuriously fast winds for certain time slots around local midnight for extended periods around the satellite's ec...

The Atmospheric Infrared Sounder (AIRS) on board the National Aeronautics and Space Administration (NASA) Earth Observing System (EOS) Aqua satellite provides 2378 channels for each field of view of the instrument. As it is neither feasible nor efficient to assimilate all the channels in a numerical weather-prediction system, a policy of channel selection has to be designed in this context. This paper attempts to assess the optimality of the selection of the AIRS radiance channels that are made available to the scientific community in near real time (hereafter called AIRS NRT) by the National Oceanic and Atmospheric Administration (NOAA) National Environmental Satellite Data and Information Service. This assessment is done by comparing this channel selection with a method preserving the information content of the instrument, the so-called ‘global’ method. It turns out that although the selected channels are different and the information content as measured by the entropy reduction (ER) and the degrees of freedom for signal (DFS) is slightly smaller for the AIRS NRT channel set than for the ‘global’ set, both channel selections give similar results in terms of analysis error for temperature, humidity and ozone. The robustness of the results is then evaluated by varying the range of input parameters to the channel-selection scheme, in particular the atmospheric training dataset on which the channel selection is based, and the background-error covariance matrix. It is found that the performance of the ‘global’ channel selection is sensitive to the training dataset, while the AIRS NRT channel selection remains robust, even, to some extent, for the retrieval of key analysis-error structures. Altogether, the ‘manually selected’ AIRS NRT channels provide a good compromise between robustness and quality. Copyright © 2003 Royal Meteorological Society.

This study investigates and quantifies in detail the spatial correlations of random errors in atmospheric motion vectors (AMVs) derived by tracking structures in imagery from geostationary satellites. A good specification of the observation error is essential to assimilate any kind of observation for numerical weather prediction in a near-optimal way. For AMVs, height assignment, tracking of similar cloud structures, or quality control procedures may introduce spatially correlated errors. The spatial structure of the error correlations is investigated based on a 1-yr dataset of pairs of collocations between AMVs and radiosonde observations. Assuming spatially uncorrelated sonde errors, the spatial AMV error correlations are obtained over dense sonde networks. Results for operational infrared and water vapor wind datasets from Meteosat-5 and -7, Geostationary Operational Environmental Satellite-8 and -10 (GOES-8 and -10), and Geostationary Meteorological Satellite-5 (GMS-5) are presented. Winds fr...

A complete re-analysis of the Fronts and Atlantic Storm-Track EXperiment (5 January–27 February 1997) is performed. It provides a reference description of the life cycle of an ensemble of cyclogeneses over the north Atlantic. Many aspects make this re-analysis noteworthy. In particular, the FASTEX period is characterized by an exceptional increase in the number of dropsoundings or radiosoundings concentrated on the storm track. Another striking point of the re-analysis is the use of a four-dimensional variational (4D-Var) assimilation scheme on such a long period associated with the French Arpege (Action des Recherche Petite Echelle et Grand Echelle) global stretched model, that allows a locally high resolution in the area where dense networks of dropsondes were deployed. A twin re-analysis with the same characteristics but without any FASTEX data is also performed. It shows that the impact of the special FASTEX data is quite noticeable in the resulting analyses and is lower, but still perceptible, in the associated short-range forecasts (6 h–36 h). The impact of FASTEX observations on a particular FASTEX case is also shown. Copyright © 2003 Royal Meteorological Society

In this paper, a weak constraint formulation of the digital filter based on the Dolph–Chebyshev window is introduced in a preoperational version of the 4DVAR analysis of Meteo-France. The constraint is imposed only on the analysis increments to damp spurious fast oscillations associated with gravity–inertia waves. In the incremental formulation of 4DVAR, the analysis increments are obtained from a global model at a uniform low resolution with a simplified set of physical parameterizations, while the high-resolution forecast is obtained with a model that uses a variable-resolution grid having a higher resolution over France and the complete set of physical parameterizations. Both models have the same vertical resolution. In a set of preliminary experiments using the same background field and the same set of observations, it is shown that the weak constraint imposed only on the low-resolution increments manages to control efficiently the emergence of fast oscillations in the resulting high-resolution forecast while maintaining a closer fit to the observations than is possible if the digital filter initialization is applied externally on the final analysis increments. It is also shown that this weak constraint does not add any significant computer cost to the 4DVAR analysis. Finally, 4DVAR has been cycled over a period of 2 weeks and the results show that, compared to 3DVAR, the initial dynamical imbalances are significantly less in 4DVAR even if no constraint is imposed at all. However, it has been noted that the innovation statistics show a positive impact when a constraint is applied.

An analysis correction method is presented that uses directly the sensitivity information coming from gradient fields at initial time to perform an alternative forecast. The method consists in imposing, at initial time, a 'sensitive' structure to the perturbation which is first computed by adjoint tools. The tuning of the perturbation amplitude is obtained by reducing the distance to the observations at the locations where the sensitive structure is of significant amplitude. The method is tested on a severe meteorological event which was operationally badly forecast even at very short ranges. Several sets of homogeneous observations are considered and the method proves its efficiency for part of them.

This paper focuses on Intensive Observation Period 17 of the Fronts and Atlantic Storm-Track EXperiment (FASTEX): the FASTEX special soundings obtained during that period are added to the conventional dataset and assimilated with a 4-dimensional variational (4D-Var) scheme under development at Meteo-France, Results show a consistent use of the high density FASTEX ship soundings by the 4D-Var formulation and an improvement brought by this new scheme over 3D-Var when using only the vertical profiles synchronous with synoptic time. The application of the 4D-Var analysis to dropsondes launched by a system-relative flight designed by the Joint Centre for Mesoscale Meteorology also produces a fine-scale description of sub-structures of the mature system.

In the context of the FASTEX (Fronts and Atlantic Storm-Track EXperiment), this paper presents a number of tests and validations of the four-dimensional variational (4D-Var) data assimilation system under development at Meteo-France, over the Intensive Observation Periods IOP3 and IOP17, in particular the impact of adding a simplified physical package within the incremental approach. the impact of using a 6-hour 4D-Var system (instead of a standard optimum interpolation or a 3D-Var scheme) is positive for the description and the forecast of Atlantic cyclones. Moreover, despite the lack of tuning and remaining deficiencies in the parametrization schemes. the analysis and prediction of Atlantic cyclones seem better handled with a 4D-Var system including physical processes in the assimilating model.

A set of physical parameterizations has been developed for inclusion in incremental four-dimensional variational assimilation (4D-Var). The goal for this physical package is that it be simple, regular (for the efficiency of the minimization in 4D-Var), and at the same time realistic enough. The package contains a simplified computation of radiative fluxes, vertical turbulent diffusion, orographic gravity waves, deep convection, and stratiform precipitation fluxes. Its tangent-linear and adjoint counterparts have also been developed. The validations of the simplified physical parameterizations and of the tangent-linear model with those included have been done. The importance of regularization (removing some thresholds in physical parameterizations that can affect the range of validity of the tangent-linear approximation), which arises during validation of the tangent-linear model, is assessed.

Results of four-dimensional variational assimilations, 4D-Var, in cycling mode, over a few two-week assimilation periods are presented. 4D-Var is implemented in its incremental formulation, with a high-resolution model with the full physical parametrization package to compare the atmospheric states with the observations, and a low-resolution model with simplified physics to minimize the cost-function. The comparison of 4D-Var using several assimilation windows (6, 12 and 24 hours) with 3D-Var (the equivalent of 4D-Var with no time-dimension) over a two-week period shows a clear benefit from using 4D-Var over a 6 or 12—hour window compared to the static 3D-Var scheme. It also exhibits some problems with the forecasts started using 4D-Var over a 24-hour window. The poorer performance of 4D-Var over a relatively long assimilation window can be partly explained by the fact that, in these experiments, the tangent-linear and adjoint models used in the minimization are only approximations of the assimilating model (having lower resolution and crude physics). The error these approximations introduce in the time evolution of a perturbation affects the convergence of the incremental 4D-Var, with larger discontinuities in the values of the cost-function when going from low to high resolution for longer assimilation windows. Additional experiments are performed comparing 4D-Var using a 6-hour window with the 3D-Var system. Two additional 2-week periods show a consistent improvement in extratropical forecast scores with the 4D-Var system. The main 4D-Var improvements occur in areas where the 3D-Var errors were the largest. Local improvement can be as large as 35% for the root-mean-square of the 5-day-forecast error, averaged over a two-week period. A comparison of key analysis errors shows that, indeed, 4D-Var using a 6-hour window is able to reduce substantially the amplitude of its fast-growing error components. The overall fit to observations of analyses and short-range forecasts from 3D-Var and 4D-Var is comparable. In active baroclinic areas, the fit of the background to the data is considerably better for the 4D-Var system, resulting in smaller increments. It appears that in these areas (and in particular over the west Atlantic), 4D-Var is able to better use the information contained in the observations. The ability of 4D-Var to extrapolate some aircraft data in the vertical with a baroclinic tilt is illustrated. Problems exist in the tropics and mountainous areas due partly to a lack of physics in the tangent-linear model. Possible improvements to the system (the introduction of more physics; better behaviour of the incremental approach owing to a line search at high resolution) are also discussed.

Satellite microwave radiometers provide measurements of precipitable water content (PWC) over the oceans, with a horizontal resolution of a few tens of kilometres. These data represent the water vapour content integrated over the atmospheric column. We assess the value of this source of information for numerical weather prediction systems. We used the European Centre for Medium-Range Weather Forecasts Integrated Forecasting System with a four-dimensional variational method to assimilate the Special Sensor Microwave/Imager (SSM/I) PWC data over a 24-hour period. Our experiments use an incremental variational method. Comparison with observations is made using a multi-level global primitive-equation T106 spectral model with physical parametrizations. Minimization is performed using a T63 adiabatic dynamics model which includes only simplified physics (horizontal diffusion, a simple surface drag and a vertical diffusion scheme). Comparing the control and assimilation experiments with aircraft and other data shows that the use of PWC data from SSM/I improves the analysis. We also obtain a slight improvement in short-range forecasts of almost all parameters when SSM/I-PWC data are used in the assimilation.