| Home > Research > Ifsdocs > OBSERVATIONS > |
IFS documentation Front Page
Chapter 2. Observations: Types, variables and error statistics
Chapter 3. CMA creation (MAKEMA)
Chapter 4. The FEEBACK task
Chapter 5. The TOOLS task
Chapter 6. Central-memory array (CMA) structure/format
Chapter 7. BUFR feedback data structure/format
Chapter 8. SIMULATED-observations data structure/format
Chapter 9. NAMELISTS
Chapter 10. Processing of scatterometer data
REFERENCES
Previous Section
1.4 Data assimilation data flow and OBSPROC's role in it
Fig. 1.1 shows data flow in the data assimilation system cycle with the OBSPROCs' role in it. Data assimilation system is invoked by starting up the OBSPROC's task MAKECMA. MAKECMA reads input BUFR data (files), extracts all necessary informations and creates so called Expanded CMA (ECMA) file. Here, it is important to note that for each ECMA report there is also its BUFR counterpart with one to one correspondence created and written out onto a separate so called initial BUFR feedback file. Only the ECMA file is then passed on to the IFS screening run for further processing. During the screening run various observation related informations are stored in the ECMA and after performing all its activities it creates so called Compressed CMA (CCMA) file for the IFS minimization run. Note: the ECMA is still in one to one correspondence with the BUFR but not with the CCMA. Although the CCMA file contains only a subset of the ECMA (both in terms of observation reports and pieces of information) the actual numbers are identical at this stage. While the minimization (analysis) is being carried out the CCMA is being constantly updated with fresh observation related informations. After the minimization has finished the FEEDBACK task is started. The idea here is to feedback (append) all observation related informations gathered in the data assimilation cycle onto the original input BUFR data. These feedback informations are: flags, events, black list informations, departures etc. The actual feedback is effectively passing over informations from the ECMA file to the initial BUFR feedback file, bearing in mind the one to one correspondence between them. However, at this stage there are two CMA files, ECMA and CCMA, with the ECMA not updated for the informations stored during the minimization. Hence, the FEEDBACK task has its sub-task called MATCHUP which purpose is to bring back the relevant CCMA file content into the ECMA. The MATCHUP sub-task is realised by an appropriate call to the OBSORT from the FEEDBACK task. Upon completing the MATCHUP the actual feedback takes place. The FEEDBACK task output data structure/format is the appended initial BUFR feedback file. These feedback BUFR files, after running some diagnostics, are then archived. As it can be noted, the OBSPROC's TOOLS task does not figure out in this data flow.
Next Section
Previous Section
 |
16.04.2002 |
 |
© ECMWF |
