1. Evolving Capabilities and Expectations for the GPCP Precipitation Products George J. Huffman(1), Robert F. Adler (2), David T. Bolvin (1,3), Eric J. Nelkin (1,3) (1) NASA/Goddard Space Flight Center (2) University of Maryland/ESSIC (3) Science Systems and Applications, Inc. 1.Heritage 2.Successes 3.Lessons Learned 4.Goals for GPCP V.3 5.GPCP V.3 Products and Inputs 6.GPCP V.3 Output Data Fields 7.GPCP V.3 Observation-Model Products 8.Final Comments

2. 1.Heritage Global Precipitation Climatology Project (GPCP) founded in mid-1990’s use then-new satellite passive microwave instruments to estimate “global” precip develop a 2.5° climatology characterize El Niño precip prior development work in Adler precip group quickly led to an operational data set Subsequent development work made the product fully global on three different time scales different algorithms, but 1DD, Pentad forced to sum to the monthly Time thin arrows denote heritage GPCP V1 SGMAGPI GPCP V2,2.1,2.2 SG GPCP V1,1.1,1.2 1DD GPCP V1,1.1,1.2 Pentad

3. 2.Success (1/2) Yes, the long-term climatology validates well intelligent-guess fill-ins are no longer necessary highest latitudes still have issues

4. 2.Success (2/2) Yes, we (quantitatively) see the annual cycle and ENSO events variations in the land and ocean time series largely offset each other note: interannual variation the small residue of large differences

5. 3.Lessons Learned (1/2) Community now recognizes two styles of precipitation data sets High Resolution Precipitation Product (HRPP) focus on “best” instantaneous answer TMPA, IMERG, CMORPH, PERSIANN, GSMaP, … Climate Data Record (CDR) focus on homogeneity GPCP follows CDR standards GPCP Version 2 now in the NOAA CDR program Users do analyses we hadn’t expected extremes analysis schemes in current GPCP don’t focus on preserving extremes intermediate products originally released to show processing steps, but getting used for science

6. 3.Lessons Learned (2/2) Users want new things precipitation phase (liquid/solid) current state of the art is a diagnostic using reanalysis T and R.H. file formats heritage binary format seen as unfriendly (as is HDF!) non-expert users tend to use off-the-shelf applications may of these have highly restrictive set of formats for inputting data NetCDF, WorldTIFF, KMZ, KML, text, CSV, shapefile-average, … file services area/parameter subsetting, data rods, WMS, GDS, THREDDS, … “rapid” availability of data all these raise the bar for introducing a “modern” data set archive centers have mostly taken the role of providing formats and services

7. 4.Goals for GPCP V.3 (1/2) In Version 3 we want to take advantage of: advanced merger schemes improved Level 2 (swath) retrievals and Level 3 (grid) products GPROF PERSIANN IMERG TRMM Composite Climatology finer time-space resolution stronger quality control and validation next-generation uncertainty estimates but still maintain: consistency among the various products CDR standards Datasets should be compatible with other water and energy cycle variables prepared by GDAP members

8. 4.Goals for GPCP V.3 (2/2) GPCP Version 3 was funded in 2013 by the NASA MEaSUREs program. The team consists of: PI:George J. Huffman (NASA) Co-Is:Robert F. Adler (U. Maryland), KuoLin Hsu (U. Calif. Irvine), Mathew Sapiano (U. Maryland), Pingping Xie (NOAA) Collaborators:Andreas Becker (DWD), Michael Bosilovich (NASA), Richard Cullather (U. Maryland), Kenneth Knapp (NOAA), Christian Kummerow (Colorado State U.), Udo Schneider (DWD), Joel Susskind (NASA)

9. 5.GPCP V.3 Products and Inputs Product scale and coverageInput data 6 a.m./p.m. GPROF-SSMI/SSMIS August 1987–present Monthly Susskind TOVS/AIRS August 1987–present GridSat-B1 1982–2010 (to be extended) Monthly 1979–presentCPC 4-km Merged IR 2011–present 0.5° globalOPI 1979–present GPCC Full Gauge Analysis 1979–2010 (to be extended) GPCC Monitoring Gauge Analysis 2011–present TRMM Composite Climatology nnHIRS reanalysis data (T, RH, sfc p) Pentad 1979–presentCMAP Pentad Precipitation 1979–present 0.5° globalGPCP Daily 1982–present GPCP Monthly 1982–present Daily 1982–present6 a.m./p.m. GPROF-SSMI/SSMIS August 1987–present 0.5° 60°N-S 1982–1996Daily Susskind TOVS/AIRS October 1996–present 0.5° global 1996–presentGridsat-B1 1982-2010 (to be extended) CPC 4-km Merged IR 2011–present GPCP 3-hr 1998–present 3-hr 1998–presentGPCP Monthly 1998–present 0.1° 60°N-S (to be extended)GPM IMERG 1998–present

10. 6.GPCP V.3 Output Data Fields The standard monthly precipitation estimate will be a satellite-gauge combination The shorter-interval products will be calibrated to the monthly satellite-gauge product Data fields in all products: accumulated precipitation (mm/d) estimated error precipitation phase (percentage probability of liquid) The monthly dataset will also provide: multi-satellite accumulated precipitation (e.g., without month-to-month gauge information; mm/d) gauge relative weighting in the final satellite-gauge product The 3-hourly dataset will also provide: snapshot precipitation rate (expressed in mm/d)

11. 7.GPCP V.3 Observation-Model Products Experience shows that numerical models are more skillful than observational products for cold-season mid- and high-latitude precipitation this is a moving target as both sides improve! In parallel with (not replacement of) the observational products, this project plans to create monthly observation-model products: start work using the NASA Modern-Era Retrospective Analysis for Research and Applications (MERRA) blend the GPCP multi-satellite and MERRA precipitation fields following Sapiano et al. (2008) then combine the gauges with the blended satellite-model field Shorter-interval combinations currently lack a good framework note the recent work by Xie and collaborators on merging daily satellite, gauge, and radar data

12. 8.Final Comments The GPCP precipitation products demonstrated the utility of long-term multi-satellite products and are considered a basic tool in climate work Over 2 decades the landscape has changed new satellite retrievals better models higher user expectations upgrades in dataset formats and archiving GPCP Version 3 is designed to take advantage of these advances still adhere to CDR concepts both observation-only and observation-model products products will be released over the next three years george.j.huffman@nasa.gov

13. 1.Introduction – Success (3/3) Longitude-time Hovmöller diagram for the latitude band 5°N-S, 1997-1998 from daily data neutral, El Niño, La Niño events are distinct longitude time

14. 6.Prototype Monthly Version 3 Compared to Version 2.2 Microwave-adjusted IR extends 60°N-S Microwave-adjusted IR extends 40°N-S TOVS/AIRS dominates 60-90° N and S microwave-only 40-60° N and S 0.5° resolution 2.5° resolution Proto. V.3 Monthly (mm/d) Jun 2012 V.2.1 Monthly (mm/d) Jun 2012

15. 6.Simulated Daily Version 3 Compared to Version 1.1 Microwave-adjusted IR extends 60°N-S Microwave-adjusted IR extends 40°N-S TOVS/AIRS used 60- 90° N and S 0.5° resolution 1° resolution TOVS/AIRS used 40- 90° N and S Sim. V.3 Daily (mm/d) 1 Jan 2007 V.1.1 Daily (mm/d) 1 Jan 2007

16. 1.Introduction – Heritage The Adjusted GPI (early ‘90’s) led to GPCP GPCP concepts were first used in TRMM, then blended with new multi-satellite concepts IMERG adds morphing, Kalman smoother, and neural-network concepts TRMM V4,5 3B42 TRMM 3B42RT Time thin arrows denote heritage TRMM 3B42RT CMORPH IMERG late IMERG final IMERG early GPCP V1 SGMAGPI TRMM V4,5 3B43 GPCP V2,2.1,2.2 SG GPCP V1,1.1,1.2 1DD GPCP V1,1.1,1.2 Pentad V6,7 TRMM 3B42 V6,7 TRMM 3B43 TRMM 3B42RT GPCP V3 PERSIANNPERSIANN-CCS KF-CMORPH

17. 5.Notional Schedule Current-version GPCP products will continue for several years while Version 3 is developed 2014:test versions of observational datasets 2015:beta versions of observational datasets for the passive microwave era test versions of observation-model datasets 2016:formal release of observational datasets beta versions of observation-model datasets 2017:formal release of observation-model datasets potential reprocessing of observation-model datasets