Global Precipitation Climatology Project GPCP Robert Adler GPCP

Global Precipitation Climatology Project (GPCP) Robert Adler (GPCP Coordinator) U. of Maryland-College Park, USA GPCP VERSION 2. 1 Climatology (1979 -2008) mm/d GPCP data used in > 1500 journal articles

GPCP Global Precipitation Products Version 2. 2 NASA, NOAA, DWD, UMD, GMU, others 1141 journal references 329 • Monthly, 2. 5° Merged Analysis (1979 -present) Adler et al. (2003), J. Hydromet. Huffman et al. (2009) GRL • Pentad, 2. 5° Merged Analysis (1979 -present) Xie et al. (2003) J. Climate • Daily, 1° Merged Analysis (1997 -present) 86 Huffman et al. (2001) J. Hydromet. [although produced using different data sets and algorithms, products are integrated, i. e. they add up] normally produced ~ 3 months after observation time

GPCP Components/People R. Adler (GPCP Coordinator) • Merge Center--Huffman/Adler, NASA Goddard/U. of Maryland (TOVS/AIRS data from Susskind, Goddard) • Gauge Center—Becker, Schneider, German Weather Service, Global Precipitation Climatology Center (GPCC) • Microwave-Land Center--Ferraro, NOAA NESDIS • Microwave-Ocean Center--Chiu, George Mason U. • Geosynchronous Center--Xie, NOAA/NWS/CPC (also does pentad merge)

GPCP Version 2. 2 Initiated August 2011 New version of GPCP V 2 monthly (and therefore other time resolutions) necessitated by failure of last SSMI instrument in “ 6 am/6 pm” orbital time slot (Sep. 2009). This loss required switch to SSMIS on F 17 as calibrating satellite. Good news: about 9 mos. of overlap. About 18 -month delay while we 1) waited for cross-calibrated SSMI/SSMIS Tb data from RSS (Wentz) used for ocean; 2) update of production software to calculate ocean monthly rain from SSMIS Tb’s (new format, GMU/Chiu); 3) adjustment of retrieval parameters to closely reproduce SSMI estimates in overlap period and 4) integration of parts into overall merger system for ocean. Similar problems over land (Ferraro), but less important because of gauge info. over land. Other related problems including SSMIS GPROF update because it is used in GPCP Daily (frequency of precipitation over ocean). Re-started Version 2 production and brought up to Dec 2010. Listed as “provisional” because of remaining issues with SSMIS over land estimates (NOAA/Ferraro).

GPCP Version 2 Product Processing to be Transferred to NOAA Proposal from GPCP “people” accepted by NOAA for converting GPCP product generating software to be run operationally at NOAA/NCDC—project starting NOW. 3 -year effort to clean-up, streamline, test and validate software from various organizations to work in as close to an automated way as possible—current processing is step-by-step with knowledgeable people heavily involved. NCDC will be heavily involved to develop software requirements/standards and approaches. We expect close interactions with NCDC to assure success. Some processing components will remain at home institutions (GPCC gauge analysis, input pentad precipitation analysis from NOAA/CPC). Planning a small GPCP user workshop to make sure result will adequately replace current products. Approach requires modest science support for GPCP scientists for dealing with issues that come up in processing, for evaluation of products and for making software changes to adjust for expected or unexpected changes in input data. This type of science support will hopefully continue even beyond lifetime of proposal.

Dec. -Feb. Columbia U. , etc. 2011

GPCP Monthly Extended Through 2010 Compared with Surface Temperature Analysis Global Tropical Surface Temp. Precip. Global Tropical Time series of monthly-mean (a, c) surface temperature, and (b, d) precipitation averaged over (a, b) global and (c, d) tropical land+ocean. Red solid lines denote the 13 -month-running means. Also shown are the linear changes.

Trend Maps of Surface Temperature, Ocean Water Vapor and Precipitation through 2010 Surface Temp. Precip. Ocean Water Vapor Linear changes in annualmean (a, d) surface temperature (K/decade), (b, e) precipitation (mm day 1/decade), and (c) oceanic columnar water vapor (mm/decade).

GRP-Meeting, 30 Aug. -01 Sept. 2011, Tokyo, Japan Global Precipitation Climatology Centre (GPCC) U. Schneider, A. Becker, A. Meyer-Christoffer, M. Ziese, B. Rudolf Global Precipitation Climatology Centre Deutscher Wetterdienst Department of Hydrometeorology

20% Mean Precipitation Difference Over China From Two Gauge Analyses Deficiency detected in GPCC analyses by B. Adler and J. J. Wang (UMD) – underestimation of rainfall over China in comparison to a CPC analysis (ca. 20%)

GPCP Version 3 • New GPCP Version 3 will have: * Monthly--0. 5° resolution, GPROF microwave algorithm applied to SSMI, SSMIS data as satellite calibrator (1979 present) * Daily— 0. 5° resolution (1998 -present; possibility of extension back in time). [Pentad for whole 1979 -present period] * 3 -hr— 0. 25° (1998 -present) to match with ISCCP and SRB products

GPCP Product Plans (Version 3/Reprocessing) • Continuation of current products (heritage products—at least for 4 -5 years) • Minimization of data set time boundaries, (or other) boundaries • Higher time and space resolutions (~ 3 hr, 25 km) for part of the period (e. g. , 1998 -present) • New passive microwave algorithm (GPROF), new passive microwave data (AMSR, TRMM, AMSU) • Integration of high time/space resolution period with longer period with coarser time/space resolution • Rain/snow discrimination (by temperature) V 3 still in development; production to start ~ April 2012

Readiness and Issues for GPCP Version 3 • Development and testing of Version 3 now re-starting with example products generated, but slowed by problems with Version 2 and readiness of input data sets for Version 3 • Required switch from SSMI to SSMIS also needs to be solved for V 3. Require cross-calibrated Tb’s (L 1 C) and GPROF passive microwave algorithm for SSMI and SSMIS not in final form. • Geo-IR data (B 1) not in final form (but close). Development being done with early versions of B 1 and GPROF for test year of 2007; final versions will likely require adjustments • Questions remain related to 3 -hr input product and daily precipitation product to be used before 1998 • Initiation of production requires ~ 10 year data record minimum; completion of production requires consistent, full data record. Completion dependent on reprocessing of GPROF SSMI/SSMIS.

Example GPCP Version 3 Monthly Analysis January 2007

Example GPCP Version 3 Compared to Version 2. 1 Monthly Analysis Version 3 January 2007 Version 2. 1

Example GPCP Version 3 Daily 1 January 2007

Daily---Example GPCP Version 3 Compared to Version 2. 1 Version 3 Mostly microwave 50°N-50°S [1998 -present] 0. 5° resolution 1 January 2007 Version 2. 1 IR adjusted by microwave 40°N -40°S 1. 0° resolution Polar regions (>60°) the same in both versions

Example GPCP Version 3 3 -hr Product (50 N 50 S) January 1, 2007 00 GMT mm/d

Summary • GPCP products are intensely used by the community (we received many questions/complaints when we fell behind due to SSMIS issue) • NOAA is supporting transfer of GPCP Version 2 set of products for operational processing at NOAA/NCDC • Version 3 development way behind GRP schedule due to delays in input data sets and V 2 problems • Plan is now to start V 3 production in ~April 2012

Extra Slides

GPCP Version 3 Plan (Sep. 18, 2009) Current (heritage) products (monthly, pentad, daily) to continue--at least to compare with new versions, but also to perhaps be used in calibration procedures. (Version 2. 1 has “solved” RSS microwave issues) (will explore impact of using L 1 C data) Tentative suite of Version 3 products (overlapping time resolutions) 3 -hr [00 Z, 03 Z, …. snapshot plus 3 -hr total estimated for 00 -03 Z, 03 -06 Z, etc. ] (0. 25 °) 50 N-50 S (possibly 60° in some areas; incomplete grids); 1998 -present (will test TMPA, CMORPH) Daily [00 Z-00 Z] (0. 5°) global (1998 -present; will use 3 -hr product at low latitudes, current daily (1 dd) product to complete globe; B 1 period [IR], (~1983 -1997) PERSIANN for calibrated IR (50 N-50 S) Pentad (sum of daily), with current pentad technique to complete globe. Sub-monthly products forced to monthly--this is meant to constrain total of 3 -hr or daily estimates which may vary during period Monthly (target: 0. 5°) global; 1979 -present (pre-1987 period will use B 1 supplemented by OPI. (2 possible approaches for SSM/I period: 1 SSM/I and calibrated IR, or multiple SSM/I’s with correction for diurnal--remember 0. 5° spatial resolution)

GPCP Version 3 Plan (cont. ) Use of microwave products: microwave Tb data set (L 1 C and GPROF (V 2008) for conical scanners. Use GPCC Full Data Re-analysis V 4 gauge analysis [lag of 1 -3 years]; GPCC monitoring product for recent period (product updated when full analysis is available) (current undercatch adjustment will be compared to new method being applied to current data). Snow/rain discrimination (will use ancillary sfc. air temp. ) Orography adjustment No global technique available. Our gauge and gauge/sat. analyses will be compared to PRISM analyses in regions where available Error estimates for input data sets, combined data sets, both bias and random will be sought from providers. Monthly already has random errors; bias error may be added AVHRR (non-histogram) data available from ISCCP or NCDC to cover Indian Ocean sector when needed) Grid choice (GRP) ; Ancillary data (OK with ISCCP/SRB, etc. choice)

GPCP Version 3 Plan (cont. ) Initial Test Period 2007 (start with Jan. for mechanics; rest of year after that). All proposed techniques (even if to be applied in pre-1998 period). Practicality, validation, evaluation Schedule ? ?