See Overview

WP 4 - Reanalysis feedback mechanism to CDR updates


The objectives of this workpackage are to "propose a process to feed back findings on CDR data quality from reanalyses to CDR teams", "propose a process for the provision of ancillary information from reanalyses to CDR teams" and to "propose a process for the reanalysis community to guide subsequent CDR updates". The four mechanisms of feedback from reanalysis to CDR updates observed by ECMWF and satellite agencies over the years (EUMETSATESA) will be turned into four processes in this WP, via four dedicated tasks.

  • Task 4.1 Assessing CDR data quality by comparison with existing reanalysis
  • Task 4.2 Assessing CDR data quality by assimilation into reanalysis
  • Task 4.3 Provision of state-of-the-art reanalysis gridded data for use as ancillary information in CDR generation
  • Task 4.4 Guidance of CDR updates 

Task 4.1 Assessing CDR data quality by comparison with existing reanalysis

CORE-CLIMAX will formalize the process by which existing reanalyses are projected into the space observed by CDR observations. In some cases, this step involves a short integration of an Atmospheric and/or Ocean Circulation Model, a Land Model, or a Hydrology Model. This integration starts from existing reanalysis gridded data, which are the result of historical data and/or other satellite data assimilation. In all cases, observation operators are needed to project the gridded data into CDR space. Consequently, identifying adequate simulating tool to serve as observation operator, if needed, is part of the process.

The project will document how to archive observation minus reanalysis differences, in the observation space, alongside original CDR error characterizations or CDR quality control flags (if provided). Requirements to be taken into account include an open-access database fitted with exploration, data mining, and visualization web-based tools. The ECMWF Observation Feedback Archive being developed by the FP7 ERA-CLIM project shall serve as an example to illustrate the concept.

The team will document how an evaluation stage entails filtering according to any dimension (quality flag, time of day, location, etc…), and visualizing the differences into the time domain (time-series), the spatial domain (map, vertical profile, platform track), or in any other relevant dimension archived in the database (e.g., satellite footprint). The outputs of this subprocess are scientific results (breaks, drifts, and any other observed peculiar features) to be compiled in a report.

Finally, the project will formalize the expected contents of preliminary CDR data quality assessment reports to be submitted to the CDR teams. Such reports could be analysed by the CDR teams for an improvement of their respective CDR before they can be considered for assimilation.

Task 4.2 Assessing CDR data quality by assimilation into reanalysis

This process enables to draw a more quantitative assessment of the CDR quality, after assimilation into reanalysis.

It can only apply to those CDRs that are deemed fit-for-purpose for reanalysis assimilation. The following two conditions must be met. First, the CDR quality has been established and is within reasonable margins as compared to other similar or historical observations and/or previous reanalyses. Second, the reanalysis system can assimilate the information contained in the CDR into an adequate analysis control variable.

CORE-CLIMAX will identify how developments of observation operator tangent linear and adjoints are to be identified.

The project will formalize how the process output is to be enriched to benefit the CDR teams. This will assume an access to an observation feedback archive and visualization of the spatio-temporal differences between the observation and the reanalysis. New materials to be added are the pre- and post-assimilation differences, and estimates of the systematic and random components of the observation errors, produced by the assimilation system. All that shall serve to compile a quantitative report on the assimilated CDR quality. Time-series analysis methods (break detection etc.) can also be applied, knowing that all input data are traceable to the observation feedback archive.

CORE-CLIMAX will propose contents for a post-assimilation CDR data quality assessment report to be fed back to the CDR team and it will take the role of the CDR producer. We will ensure reasonable contents that are meaningful and useful to improve the CDR at its next update. The project will take the example of infrared radiances FCDR that are assimilated into the reanalysis and can be further improved using the assimilation feedback information from the reanalysis.

Task 4.3 Provision of state-of-the-art reanalysis gridded data for use as ancillary information in CDR generation

Under this taks CORE-CLIMAX will explore best practical mechanisms for the CDR teams to acquire gridded reanalysis data. It might turn out that this could be done by open calls for reanalysis gridded data and documentation. Such transparent and documented approach would ensure traceability and trust in the CDR products and error budget studies.

Task 4.4 Guidance of CDR updates

This last task covers the process that ensures that the CDR teams know about the needs of reanalyses in terms of CDRs: essentially which CDR updates are required, and when. The project will investigate how each reanalysis team can best communicate and exchange their needs for CDR or CDR updates. The idea is to make sure that CDR teams (as well as satellite agencies and funding bodies) have up-to-date information on reanalysis needs in terms of CDR (for this information may help attribute priorities to the various CDR update efforts). This information shall feed the digital forum developed in WP1.