PROGRESS AND CHALLENGES.

Atmospheric, oceanic, and land reanalyses have become fundamental tools for weather, ocean, hydrology, and climate research. With great progress having been made in the last five years, reanalyses have even become established long-term climate and environmental records. And they continue to evolve with improvements in data assimilation, numerical modeling, observation recovery, and quality control, with newer ideas, projects, and data coming forward.

While reanalysis has typically been carried out for the individual domains of atmosphere, ocean, and land, it is now moving toward coupled systems using Earth system models. Observations are being reprocessed to improve their quality for use in reanalysis data. New applications are being investigated, and the need for climate reanalyses is as strong as ever. At the heart of it all, new investigators are exploring the possibilities for reanalysis data, and developing new concepts in research and applications. The total number of reanalyses is increasing through varying disciplines and motivating goals (e.g., ocean, land, and cryosphere research centers, as well as weather and climate prediction centers), along with the development of new ideas (e.g., families of reanalyses), and new and innovative diagnostics and output data.

With the applications of reanalysis data growing steadily, there is an increasing need for open discussion and comment on the quality of the data. This article presents current progress and broad directions to continue the advancement of reanalysis data. It summarizes the presentations and discussions of the Fourth World Climate Research Programme (WCRP) International Conference on Reanalyses, held in spring 2012. A complete conference report with a listing of all authors and contributors is available online at http://ICR4.org.

While originating in the atmospheric sciences and numerical weather prediction (NWP), the essential methodology for reanalysis data has been adopted in the fields of oceanography and terrestrial ecosystems and hydrology, with emerging research in atmospheric composition, cryosphere, and carbon cycle disciplines. The latest generation of atmospheric reanalysis systems—such as the Modern-Era Retrospective Analysis for Research and Applications (MERRA; see, e.g., Fig. 1), the Climate Forecast System Reanalysis (CFSR), and the European Centre for Medium-Range Weather Forecasts (ECMWF) Interim Re-Analysis (ERA-Interim)—has been evaluated, and many strengths and weaknesses have been identified. CFSR has demonstrated the first attempt at land and ocean coupling with the atmosphere, and the first results are becoming available from the new Japanese 55-yr Reanalysis (JRA-55). There is also much to be learned from the National Oceanic and Atmospheric Administration (NOAA)'s Earth System Research Laboratory (ESRL) Twentieth-Century Reanalysis (20CR) surface-pressure-only ensemble reanalysis. Furthermore, ocean reanalyses are demonstrating that ensembles of multiple reanalysis systems can provide valuable ocean circulation data, air–sea fluxes, and deep-ocean energy transport.

View Full Size

The 1979 President's Day snowstorm depicted from MERRA sea level pressure (mb), surface winds (kt; 1 kt = 0.51 m s⁻¹), and cloud fraction.

Citation: Bulletin of the American Meteorological Society 94, 8; 10.1175/BAMS-D-12-00191.1

• Download Figure
• Download figure as PowerPoint slide

Although there are several reanalysis efforts worldwide at present, the community consensus is that the diversity among them will enable deeper understanding of the reanalysss systems, their strengths and weaknesses, and their representation of the underlying Earth system processes/phenomena. This is then reflected in the producers' plans [notably those of the Japanese Meteorological Agency (JMA) and ECMWF] that lean toward “families” of reanalyses (each system producing various configurations of reanalysis data). There is much to be learned about the observations, data assimilation, modeling, and coupling of the Earth system, but new data systems, efficient computing, and processing of the multitude of reanalysis products are required. There is also a need to encourage the young generation of scientists to pursue the areas of research that comprise a modern reanalysis system.

Observations are the fundamental resource for reanalysis. The importance of observing systems cannot be overstated, especially in the stratosphere and deep ocean, to anchor the reanalysis. In situ observations provide reference datasets for calibration, validation, and bias-correction purposes. Reanalyses would benefit from a greater range of high-quality monitoring products for validation purposes. For example, new precipitation data products from the Global Precipitation Climatology Centre and the Hadley Centre high-resolution climate dataset over land [Hadley Centre Integrated Surface Database (HadISD)] may provide valuable high-quality input data. Data archives such as the International Comprehensive Ocean–Atmosphere Data Set (ICOADS) and the Integrated Global Radiosonde Archive (IGRA) are being continuously populated by newly rescued data. Efforts such as those of the Atmospheric Circulation Reconstructions over the Earth (ACRE) and the International Environmental Data Rescue Organization (IEDRO) are crucial to rescuing and archiving historical data in digital format. The International Surface Temperature Initiative (ISTI) has the potential to become a valuable land data source in the future. Reanalyses can be used to identify and correct datasets, such as those from radiosondes, where the identification of break points in time series of the observations is critical to the success of adjustment methods and subsequent derivation of climate trends. Remote sensing provides useful input data for reanalyses, primarily for the last three decades, and older imagery might also be exploited using ad hoc processing. However, satellite data present some unique challenges, requiring intercalibration, reprocessing, and bias corrections.

Assessing robust observational and model error covariances, preferably varying over time, is complex and expensive. While many producing and research agencies have developed and investigated bias-correction methods, it should be stressed that both models and data contain biases. Preliminary results indicate a potential benefit of coupling the ocean and atmosphere domains for improved forecasts and reanalyses. Data assimilation is also helpful in designing observing systems and in identifying erroneous data, but data assimilation should be consistent with the processes it aims to resolve, requiring appropriate model development for that purpose. Given the discontinuous nature of the observational record, data assimilation techniques will be the primary way to develop more temporally continuous reanalysis output data. Data assimilation methods continue to improve, but they have more challenges ahead, such as the amelioration of shocks associated with changes to the observing system, better characterization and reduction of model bias, and the development of uncertainty estimates for reanalyses.

Reanalyses are natural integrative tools, yet coupling the components of the Earth system in reanalyses remains a challenge. Integrating the components of the Earth system in a reanalysis framework exposes the complexity of an observing and modeling system approach. For example, direct and indirect (cloud albedo) aerosol negative radiative forcing will provide feedback on the other analyzed components. Empirical optical depth retrieval and variable transformation are some of the techniques being used to that effect. Another example includes forward proxy modeling approaches that use ensemble-mean increments to modify single members, which are able to decrease the computational burden of reanalyses and improve overall skill. The high-resolution (30 km) Arctic System Reanalysis (ASR)-Interim shows superior skill to ERA-Interim on many parameters; a new release at 10 km is expected in 2013.

There is a move toward using reanalyses for monitoring some aspects of the climate. While this is a valid objective, there are still some considerable limitations regarding long-term observations and monitoring to be addressed. For example, temporal homogeneity across the entry and drop out of various observing systems [e.g., Advanced Transfer for the Television and Infrared Observation Satellite (TIROS) Operational Vertical Sounder entry in 1997] affects climate, trends, and water and energy budgets, especially over the oceans and Arctic regions. Used with caution, reanalyses are highly valuable as long-term records, and it is recognized that some level of review may be useful to provide context for their future use as monitoring products.

FUTURE DIRECTIONS.

While progress has been very good across major aspects of reanalyses, significant limitations persist. Four broad directions to continue the advancement of reanalysis data are listed below.

• 1) Quantitative uncertainty estimation—Reanalyses are based on observations, and can include the errors of observations and the assimilating system. It is recommended that reanalysis data be made available in a common framework to facilitate the analysis of their strengths and weaknesses. The notion of families of reanalyses will likewise expose the impact of assimilating observations on the analyses. Ensemble methods can also provide quantitative uncertainty estimates. Last, passing observations and innovations through to an easily accessible data format can promote deeper investigation of the use of observations in the reanalyses.

• 2) Qualitative uncertainty estimation—Reanalysis users often ask which reanalysis is best for a given topic. As newer reanalyses come along, the answer may not be widely known—if at all. In this regard, the community of users and developers must collaborate, given the diversity of reanalysis applications. Therefore, sharing reanalysis knowledge and research in a timely manner, among researchers and developers, is a critical need to allow subsequent exploitation by the climate community. The website Reanalysis Intercomparison and Observations (http://reanalyses.org) has provided an initial effort along these lines, but more participation is encouraged. In addition, the Climate Data Guide (http://climatedataguide.ucar.edu) provides informed commentary on reanalysis and other datasets. Likely, even more lines of communications are required. Ultimately, it is incumbent on the researchers to assess the multitude of reanalyses objectively. New data systems are required that allow for more efficient cross comparisons among the various reanalyses [such as those used for Atmospheric Model Intercomparison Project (AMIP) and Coupled Model Intercomparison Project (CMIP) studies] and the Earth System Grid (ESG), which is designed to enable access to such large-scale datasets.

• 3) Earth system coupling—The natural course of reanalysis development is toward longer datasets with coupled Earth system components that will ultimately contribute to improved coupled predictions. The use of more varied observations (e.g., aerosols) will reinforce the physical representation of the Earth system processes in the reanalysis systems. There is a need to develop independent and innovative modeling, coupling, and data assimilation methods to represent the Earth system throughout the time span of the observational record. More interdisciplinary collaborations in system development and observational research will begin to address this need.

• 4) Reanalyses, observations, and stewardship—While the observational records have been greatly improved since the first reanalyses through research, reprocessing, and homogenizations, research and improvements continue their development. Reprocessing and intercalibrations of observed records are critical to improve the quality and consistency of reanalyses. In situ and satellite data need to be found, rescued, and archived into suitable formats to extend the reanalysis record back in time. Reanalysis systems for the atmosphere, ocean, cryosphere, land, and coupled Earth system are needed that maximize use of the observations as far back as each instrumental record will allow. It is important for the observational data and reanalysis developers to maintain communication, so the latest data are used in reanalyses, and also that the output of reanalyses may contribute to the understanding of observations. Such an endeavor should be coordinated at an international level.

CONCLUSIONS AND PERSPECTIVES.

Reanalyses will most likely increase in number and complexity in the coming years. Incorporating reanalyses in improved data systems, such as ESG, the Coordinated Regional Climate Downscaling Experiment (CORDEX), and reanalyses results in support of the Intergovernmental Panel on Climate Change assessment, would also facilitate the comparisons among reanalyses and independent observations, and would shed more light on the quality and variability among reanalyses. The conference findings support deeper use of reanalyses and also improving their capability through international coordination, more and better input observations, and thorough intercomparisons. Sustained and focused support for reanalysis research by the funding agencies will ensure greater progress in this budding field, which has great potential in demonstrating the complementary power of observations and models to offer science-based information for decision makers in addressing the challenges and opportunities associated with weather, climate, and ultimately environmental services. The World Climate Research Programme (WCRP) should continue to facilitate international coordination of reanalyses activities and promote greater use of reanalysis products beyond scientific research and development.