Enhancing Historical Climate Record: Data and Observing System Issues

1. Climate Analyses of the Past and Present (CAPP) Establishing fields of the historical climate record and context for today’s observations Panel/WG 4: Data and observing system issues, including changes in observing system Chair: Kevin Trenberth Panel: Data issues: Steve Worley (not available; provided some slides) Dave Easterling Bias corrections and ERA-40 experience Saki Uppala (extra time: ERA40 experience) Bob Kistler (not available) OSEs and more Bob Atlas Other Members: Mike Fiorino, Chris Miller, Chidong Zhang

2. Goals 1. Improve and develop input data for climate analyses • For monitoring low frequency variability, we need long-term stable homogeneous climate data records (of known quality) 2. Improve global estimates of interannual to decadal variability and their uncertainty. Improve the consistency of the record in the face of major changes to the observing system. One or more reanalysis must be targeted at the goal of producing the most consistent time series (rather than the best analysis at any time). • Retain fairly consistent observation sets (e.g., by excluding special obs, short field campaigns) • Use OSEs to determine effects of changes remaining in observing system • Focus on bias-corrected observations • Assess uncertainties in trends etc. • Suggested time periods: post 1950; and post 1979.

3. Datasets for Climate System Studies • Reanalysis is a “data cleansing” process • Gather and develop metadata, and make available • Improve basic observations by assigning or even applying biases: known quality • Be aware that a blacklisted observation may be because of analysis problems (e.g. resolving tropical cyclone) • Should be a continual process, feedback to observing system

4. Observational Archives • Rescue old and operational data • International exchanges • Data Stewardship • Basic Quality Control • Station Library Improvements • Verify station locations and elevations • Is this the source of biases? • Maintain irreplaceable data in perpetuity • Archive data for use • Access • Flag special data (e.g., field program, so that it may be excluded in reanalysis)

5. Findings: 1. The reanalysis “feedback file” is potentially immensely valuable for improving the basic input data by exploiting the use made of the data, its biases relative to the first guess and final analysis, and any QC flags. 2. There is a critical need to provide full tracing of datasets (ID) and versions or generations. 3. There is a need to provide easy access to the feedback file from ERA-40 and NCEP and to the basic observations.

6. Research based on reanalyses FEEDBACK INFORMATION: Observed values OB – FG OB – AN Flags Understanding & Improving Observation datasets: Satellite Conventional Data Assimilation System Development Improved use of Observing Systems in the next reanalysis

7. Filling the possible gaps in time and space with additional data Insert the new data into Data Base Merge of observations and observed values: in the Screening step before minimization Use the merged observations in reanalysis(n+1) Investigate the Feedback information (Ob-Fg, Ob-An) : Statistics and time series per area/ station Statistics of accepted/ rejected data per data source Detailed inventories of the merged dataset and data sources Check/ Modify Station height and position information

8. Finding 4 : There is a need to establish a baseline set of observations Closest to a baseline, but still unsatisfactory, is the GUAN (Global Upper Air Network) of designated radiosonde stations Recommendation: Extract subset of feedback file for all radiosonde stations (emphasis on GUAN), and make available. Validate ERA-40 and Improve GUAN Use the feedback file to adjust the radiosonde record and fill in missing data, with appropriate flags, to generate an improved radiosonde record. Similarly for other in situ components. These steps allow the basic data to then be used independently, for instance, for climate change detection.

9. Issues for trends and low frequency variability: • Model bias • Analysis tends to revert to model climate in absence of data • Real trends • SSTs and radiative gases; • But not total solar irradiance, aerosols, land use change • Perturbations (like Pinatubo) • Changes in observing systems (spurious) While some trends may be captured by the observing system and can be reflected in other quantities through the dynamics, in general the null hypothesis should be that trends and low frequency variability are more likely to be spurious unless proven otherwise.

10. Observing System Changes: • In situ:SSTs (1982), Pibals, Aircraft, etc • Radiosondes • Types, instruments, locations, times, coverage • Satellite data • 1972: VTPR • 1973: some cloud tracked winds • 1979: TOVS (HIRS, MSU, SSU); TOMS, SBUV (ozone) • 1987: SSMI (sfc winds, column water vapour) • 1992: ERS scatterometer • 1998: ATOVS • Satellites vary in number,have finite lifetimes • and are replaced every few years. There is • orbital decay, change in times, platform heating • and instrument degradation • These require bias corrections

11. TOVS/ATOVS for ERA-40 S A T E L L I T E S U P P L I E R NOAA-16

12. Aspects relating to the use of radiances • Problem periods per channel into “blacklist” • Bias tuning necessary • Overlapping periods important • Cloud detection • Contamination by aerosols

13. Finding: There is a need to track system performance viz a viz trends and validate them • Use independent measures and constraints • Global mass of dry air • Surface air temperature over land in selected regions • Dobson ozone measurements • SAGE data (water vapor) • Ocean wave measurements • Alpine summit station data • Field campaign data • Time series of forecast performance measures • “Satellite temperatures” MSU 2 and 4 • GEWEX and SPARC datasets and reports • e.g. ISCCP clouds, etc • Other measurements: • Surface observations, River discharge, Glaciers (?) • Time series of analysis fits to observations • Recommendation: • Document reliability of trends and communicate them • and the results of the studies to users.

14. Finding: OSEs and OSSEs provide effective tools to aid in the interpretation of trends found in model reanalyses, as well as determining the optimal observing system configuration for both weather and climate analysis.

15. Recommendations: • 1. Carry out an ensemble of AMIP-type model simulations • with the available forcings to establish the model climate • and its natural variability. • 2. Carry out selected OSEs with and without major new • observing components • Such as 1973 (VTPR), 1979 (TOVS), 1987 (SSMI) • 3. Carry out a series of OSEs to assess overall gradual • changes in observing system by utilizing results from • a recent year (1998-2003) and degrading the observing • system to match that of: • Late 1950s (include simulated weather ship obs) • Mid 1970s (include simulated VTPR from HIRS) • Mid 1980s (representing TOVS era) • For different seasons

16. Recommendations cont: • 4. Establish a core (base) ongoing activity to carry out OSEs and OSSEs to • better establish the true climate record by determining impacts of changing observing systems and interpreting reality of trends. • help optimize and design observing system for climate • routinely assess subtle changes in the observing system

17. Components of the CAPP for this WG: • Data set development • Research and development, peer reviewed proposals • Operational arm of NOAA/NASA to • update reanalyses (CDAS-activity) • continually pursue OSEs and OSSEs to document impact of continuing observing system changes. • Operations should become proactive on observing system development • Production phases • 1. Post 1979 reanalysis with goal of continuous climate record • 2. Post 1950 reanalysis with same goal. • 3. Post 1900 surface NH oriented $5M/year $3M/year $1M/year enhancement