1. Global Precipitation Climatology Centre (GPCC)
Report from the
Global Precipitation Climatology Centre (GPCC)
25 Years,
Report to AOPC-XX, Zurich, March 2015
Doc Agenda item 7.4
Andreas Becker
Head, Precipitation Monitoring Unit & GPCC
Department of Hydrometeorology
Deutscher Wetterdienst (DWD), Offenbach am Main, Germany
AOPC-Meeting XX GPCC Report

2. Outline Background of Global Precipitation Climatology Centre (GPCC)
Data base, storage and data policy
Status Quo, monthly archive and products
Status Quo, daily archive and products
Products released and published through ESSD/DOI references
GPCC’s follow ups on AOPC-XIX recommendations
GPCC’s requests to AOPC
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

3. Background of GPCC
analysis of precipitation on the basis of in-situ data for the land-surface
established at the beginning of 1989 at Deutscher Wetterdienst (DWD) on invitation by WMO → 26 year of experience with precipitation
Contributing to GEWEX (Global Energy and Water Exchanges Project) and GCOS (Global Climate Observing System)
Many users world wide, analyses used in IPCC-AR5
Data sources: SYNOP, CLIMAT, SYNOP from CPC, ECA&D,CRU, FAO, GHCN, national meteorological services, regional data collections
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

4. KU42/GPCC Staff
KU42 Unit GPCC Team: 1 Head 3 Scientists, 1 Proj. Scientist 1 Programmer, 3 TAs RADOLAN Team 2 Scientists 2 Proj. Scientists 1 Programmer 2 TAs
From left to right:
M.Sc. Elke Rustemeier, Dr. Markus Ziese, Heinz-Jörn Müller, Dipl. Met. Udo Schneider, Jan-Nicolas Breidenbach, Astrid Heller, Cathleen Sterker, Mario Hafer, Dr. Tanja Winterrath, Peter Stender, Dipl. Met. Elmar Weigl, Dipl. Met. Anja Meyer-Christoffer, Peter Finger, Dr. Andreas Becker, M.Sc. Thomas Junghänel, M.Sc. Anna Schmitt
A. Becker, Beobachtung und Bewertung extremer Niederschlagsereignisse. Quo Vadis DWD? ; Besuch PIK-Potsdam, 3. Februar 2015
Beitrag RADOLAN von KU 42 A beim KU-Workshop Tagesanalysen am 4

5. GPCC data sources Near real-time regularly exchanged data
via the WMO Global Main Telecommunication Network
hourly/daily weather reports (SYNOP)
 up to 7,500 stations
monthly climatological totals (CLIMAT)
up to 2,900 stations
up to 8,800 unique stations
Non real-time:
Additional data received from ca. 190 countries
Historical data collections (CRU, FAO, GHCN, Nicolson)
International project data (GEWEX and other)
plus:
> 95,000 stations in total
Global Precipitation Climatology Centre, AOPC XIX – April Agenda Item 5.3

6. WMO Data Request for the GPCC
GPCC acquisition of additional data is supported by WMO -> GPCC benefits from its operational implementation in a national meteorological service (DWD)
which is closely connected under umbrella of WMO to the Met Services of the World,
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

7. Steps of quality control at GPCC
GTS-Data (SYNOP reports):
check of precipitation amounts against the weather information
consistency check of reports overlapping in time
(fill gaps without precipitation, if weather group indicates no precipitation)
Delivered data
Check station location within country and over land
Coding of missing values (sometimes 0 is used)
Data from existing station or new one (check of meta data)
precipitation data checked against background statistics for the station or, if not available, for the corresponding 2.5° grid (since 2009)
data being flagged as questionable (below the 5% or above the 95% percentile), checked manually
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

8. QC: Errors typically detected
Stations are sometimes located in the ocean or outside of the boundaries of the country
Unusual annual cycle or extreme outliers of monthly precipitation
Temporal shifts in the data
Factor*10 errors
Typing or coding errors
Errors in the conversion of inch, mm etc. (mostly with historical data)
Incorrect flagging of missing precipitation observations (might be misinterpreted as „0“)
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

9. User requirements Timeliness (for drought monitoring)
Features of gridded precipitation data as required by the users:
Timeliness (for drought monitoring)
High resolution (for regional structures in global maps)
High accuracy (for verification of model results)
Homogeneity (for climate change and variability analysis)
All of these requirements cannot be met by one single gridded data set
==>
A portfolio of different analysis products has been designed and optimized with respect to the application purposes
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

10. Data base for different monthly GPCC products
Break in 1986 because of initial start of data acquisition almost homogenized
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

11. GPCC Climatological data base
Station data base of GPCC‘s Precipitation Climatology V.2011 (V.2015) as basis for anomaly analyses ca. 67,200 (75,000)
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

12. Monthly GPCC Products (Becker et al, 2013, ESSD 5, 71-99)
First Guess Product (DOI: /DWD_GPCC/FG_M_100)
based SYNOP data, automated QC, Resolution 1.0 deg
available within 3 to 5 days after the end of each month
Monitoring Product (Version 4, DOI: /DWD_GPCC/MP_M_V4_100)
based on CLIMAT and SYNOP data, enhanced QC, Resolutions 1.0° & 2.5°
available within two month after the analyzed month
Climatology (Version 2011, DOI: /DWD_GPCC/CLIM_M_V2011_025)
based on about stations, Resolutions 0.25°, 0.5°, 1.0° and 2.5°
target reference period 1951 – 2000, stations with at least 10 years of data
background climatology for GPCC products
Full Data Reanalysis (Version 6, /DWD_GPCC/FD_M_V6_050)
uses same stations as Climatology, Resolutions 0.5°, 1.0° and 2.5°
available from 1901 to 2010 (2013 with V7)
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

13. WZN/GPCC Anwendungsbeispiele
Utilization of Full Data Reanalysis Product to identify ENSO and NAO sensitive precipitation regions for the 110yrs period
SOI
NAO
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

14. Utilization of Full Data Reanalysis for ERA_CLIM2 Project
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

15. Steps of quality control at GPCC
GTS-Data (SYNOP reports):
check of precipitation amounts against the weather information
consistency check of reports overlapping in time
(fill gaps without precipitation, if weather group indicates no precipitation)
Delivered data
Check station location within country and over land
Coding of missing values (sometimes 0 is used)
Data from existing station or new one (check of meta data)
precipitation data checked against background statistics for the station or, if not available, for the corresponding 2.5° grid (since 2009)
data being flagged as questionable (below the 5% or above the 95% percentile), checked manually
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

16. QC: Errors typically detected
Stations are sometimes located in the ocean or outside of the boundaries of the country
Unusual annual cycle or extreme outliers of monthly precipitation
Temporal shifts in the data
Factor*10 errors
Typing or coding errors
Errors in the conversion of inch, mm etc. (mostly with historical data)
Incorrect flagging of missing precipitation observations (might be misinterpreted as „0“)
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

17. Absolute Precipitation Anomaly in year 2014
Fig. 5 in „WMO State of the Climate 2014“
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

18. Globally averaged precipitation anomalies (mm yr-1, base period) for three in situ data sets over land.
To appear soon in „BAMS SoC 2014“ d.3
R. S. Vose, K. Hilburn, X. Yin, M. Kruk, and A. Becker
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

19. Additional products available from GPCC
Analyses of daily land-surface precipitation
First Guess Daily: automated QC, based on SYNOP reports; near real-time (within five days)
Bi-decadal ( ) global analysis of daily precipitation utilized to study extremes with ETCCDI package
Additional products:
GPCC drought index: combination of SPI-DWD and SPEI with nearly global coverage
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

20. GPCC‘s First Guess Daily Product
doi: /DWD_GPCC/FG_D_100
Interpolated with ordinary block Kriging
Interpolation of relative values – fraction of daily total in relation to
monthly total (only stations with monthly total could be used, at
least 70% data coverage required)
Released together with First Guess Product
netCDF-files containing total precipitation, standard deviation
regarding Yamamoto (2000), Kriging error and number of stations
per grid
Described in Schamm et al. (2014): “Global Gridded Precipitation
over Land: A description of the new GPCC First Guess Daily product”,
Earth Syst. Sci. Data, 6, 49-60, DOI: /essd
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

21. GPCC First Guess Daily product elements
precipitation
standard deviation
number of stations
Kriging error
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

22. Additional work at GPCC for historic daily data
participating DAPACLIP (global DAily Precipitation Analyses for the validation of medium-range CLImate Predictions)
Daily precipitation analyses for 1988 to 2008 over land
Acquisition of daily data
Developing QC scheme for daily data (including comparison to monthly total)
Loading data into data bank focusing 1988 to present
Testing and optimization of interpolation schemes including error estimation
Test of optimization possibilities of QC → detection of shift by one or two days
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

23. Overview of the daily data collected by GPCC
Daily data loading focusing on 1988 to present, earlier data available at GPCC
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

24. II. Bi-Decadal Global Daily Data Set, 1988-2008
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

25. II. Bi-Decadal Global Daily Data Set, 1988-2008
Example to characterize the bi-decadal DAPACLIP daily data set (Dietzsch et al., 2015) through the precipitation related indices of the ETCCDI package compiled by Klein Tank et al. (2009).Here RX1day (maximum daily precipitation) for the second decade ( ) is shown.
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

26. Bias: Obs-Model (2-5 yrs lead) Bias: Obs-Model (7-10 yrs lead)
Daily resolving observational data sets allow for
model evaluation (Dietzsch et al., EGU )
SDII
Jeweils 2 Grafiken gehören in der Waagerechten zusammen.
— Obere Reihe: Zeigt einen einfachen Bias (Modell minus Beobachtung) des SDII = Simple Daily Intensity Index = Mittlere Niederschlagssumme eines Jahres an Tagen, an denen >1 mm Niederschlag fällt. Das alles nochmal gemittelt für die Model Lead Years 2-5 (links) und 7-10 (rechts) = 2000–2003 und 2005–2008; Verwendet wurde Ensemble Run 1 (bezeichnet mit r1i1p1) des Prototypen, der mit GECCO2 initialisiert worden ist für die Dekade 1999–2008 (vgl. Überschrift der Grafiken).
Modell minus Beobachtung => Rot = Im Modell ist mehr durchschnittlicher Niederschlag; Blau = Im Modell ist weniger durchschnittlicher Niederschlag.
Man sieht, dass es über Land im Modell konsequent zu feucht ist. Ebenso in den trockenen Regionen; z.B. Sc Region im südlichen Atlantik oder um die ITC.
— Untere Reihe: Zeigt den Bias nach selbem Prinzip für CDD = Consecutive Dry Days = Maximale Anzahl aufeinanderfolgender Tage mit Niederschlag <= 1mm
Zu sehen ist vor allem ein markanter Bias im Bereich der pazifischen ITC, die im Modell bzw. im Südzweig extrem trocken simuliert wird, in der Beobachtung so aber überhaupt nicht auftaucht; siehe auch 2. Folie. Weiterhin zu sehen, dass viele trockene Regionen im Modell (leicht) überbetont sind, wobei z.B. die Sc-Region im südl. Atlantik zu feucht simuliert wird.
Bias: Obs-Model (2-5 yrs lead)
Bias: Obs-Model (7-10 yrs lead)
CDD
Courtesy of Felix Dietzsch, CM-SAF, Deutscher Wetterdienst
GPCCs activities in support of WCRP GC on Extremes, Data Reqs WS, Sydney, 25 Feb 2015

27. Models‘ Diffusion Models‘ Bias Model Models‘ basic noise Obs
Daily resolving observational data sets allow for
(decadal) model evaluation (Dietzsch et al., EGU )
Models‘
Diffusion
— Linke Grafik: Zeigt einen doppellogarithmischen Perzentilplot für den schon in Folie 1 verwendeten Prototype-Run sowie einen Historical-Run (d.h. uninitialisiert) des Prototypen für die Dekade 1990–99 (nur dafür gab es einen passenden Historical zum Vergleichen) für die Europa-Domain. Jeder einzelne Punkt entspricht einem ganzzahligen Perzentil aus der Niederschlagsverteilung im Modell und in DAPACLIP. Der Wert der Perzentile ist hier gegeneinander aufgetragen (Ordinate = Modell, Abszisse = DAPACLIP). Für geringe Niederschlagswerte 0.1…1 mm erkennt man ein Grundrauschen des Modells für das Niederschlagssignal, während am oberen Ende das Modell abknickt, d.h. Maxima werden vom Modell quantitativ nicht so gut getroffen. Rot zeigt die entsprechende Normale, die im perfekten Fall zu erwarten gewesen wäre.
— Rechte Grafiken: Zur Illustration des CDD-Bias aus Folie 1 gedacht. Oben: Jährliche CDD DAPACLIP 2005–2008 gemittelt. Unten: Das gleiche für Prototype, wieder für die Dekade 1998 (=1999–2008). Man erkennt im Modell die extrem trocken simulierte pazifische ITC, die in der Beobachtung so nicht existiert. Ebenfalls gut zu sehen die als zu trocken simulierte Sahararegion, während die südafrikanische/südatlantische Sc-Region zu feucht ist (frequentativ).
— Anmerkung: Ich habe dafür jetzt das im Rahmen von Miklip entwickelte Tool zur Berechnung der Indizes verwendet. Allerdings berechnet dieses Tool nur Werte auf maximal jährlicher Basis, was für Indizes auf Basis aufeinanderfolgender Tage (wie z.B. CDD) eigentlich nicht geeignet ist, da gerade Trockenperioden durchaus länger als 1 Jahr andauern können.
Models‘
Bias
Model
Models‘ basic noise
Obs
Courtesy of Felix Dietzsch, CM-SAF, Deutscher Wetterdienst
GPCCs activities in support of WCRP GC on Extremes, Data Reqs WS, Sydney, 25 Feb 2015

28. III. GPCC drought index
GPCC_DI: gridded drought index with nearly global coverage
combination of SPI-DWD and SPEI
precipitation data from GPCC; First Guess Product
monthly mean temperature from CPC
uses mean of SPI-DWD and SPEI, if both can be calculated, otherwise the one which can be computed
parameters derived from Full Data Reanalysis V.6, period
several averaging intervals: 1, 3, 6, 9, 12, 24 and 48 months
using gridded fields, no interpolations → areas with no data possible
analysis from January 2013 until present
provided as netCDF-files
updated 10 to 13 days after each month
DOI: /DWD_GPCC/DI_M_100
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

29. + => III. Example GPCC drought index, January 2013, 1 Month SPI-DWD
SPEI
=>
GPCC_DI
drier than
wetter than
normal
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

30. III. GPCC drought index here for 01.05.-31.01.2015
Sao Paolo
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

31. GPCC comments on AOPC-XVIII (GCOS-169) actions and recommendations:
Referring to Action Item 26, GPCC is grateful for the acknowledgement received.
Referring to Action item 27, GPCC would like to express its gratitude for the continued support of the WMO Secretariat with regard to GPCC’s activities in the field of data acquisition. In the 2nd half of year 2015 GPCC will liaise with GCOS on an optimized data request letter that shall contain better incentives to the data providers and offers additional ways of daily data provision through an XML based data standard.
Referring to Action item 28 GPCC remains open to joint efforts in data acquisition, as demonstrated with ECMWF and KNMI/ECA&D past year with regard to the daily precipitation data originally sent to ECMWF for model verification. The proposed solution to gather the data through ECA&D mechanisms while requesting for co-use by GPCC is a good example for a joint activity. The issue on a too quick re-assignment of stations Identifiers (IDs) from closed stations challenging maintenance of the WMO catalogue remains open. Action from the WMO WWW might be needed to provide sufficient blocks of station IDs.
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

32. GPCC suggestions for AOPC XIX action:
Continue GPCC data acquisition support via WMO Secretariat (e.g. WCP) in case of GPCC communication problems with WMO members; Cooperation with GCOS in 2015 on an updated and optimized data request letter utilized for the communication.
Support alternate ways of joint daily data acquisition in cooperation with scientific projects and with activities of the GEO community. In particular the inclusion of OGC compliant XML standards should be considered constructively due to their enormous potential for data mobilization.
Reliable meta information is key for systematic error correction and correct station association. The latter is challenged in countries of WMO members that have to re-assign station Id’s from closed stations too quickly, challenging the WMO catalogue maintenance. WMO WWW activity might be needed in this context, e.g. to provide for an additional block of station Id’s for those members being short on this;
Consideration of web-services for data and product dissemination should become an appropriate priority level for the WMO Secretariat. DWD and GPCC are ready and open for exchange of requirements and plans.
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

33. GPCC plans for next year
Issuance of the next generation of monthly re-processing products (GPCC Full Data Reanalysis Version 7, GPCC Climatology Version 2014, GPCC Monitoring Product Version 5) in spring For the Monitoring Product the re-processing shall start already in 1986, allowing also for the provision of the systematic error fields (due to snow and wind) for the entire period 1986-today. (So far this was only available since 2007).
Continuation of loading of daily data and development and generation quality assured daily data products. After completion of import from existing distributions (GHCN-daily and ECA&D) a new initiative to enhance the daily data compartment from national contributions (WMO NHMS) will be started in the 2nd half of 2015 through an updated data request letter making better incentives towards data providers with regard to GPCC’s capabilities in backup services and provision of feedback to data providers as for GPCCs results of quality checks.
Continuous enlargement of the GPCC daily database concerning the entire instrumentation periods.
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

34. GPCC plans for next year
GPCC is up to address the data requirements of WCRP GC on Weather and Climate Extremes. Tasks and deliverables have been defined at the Workshop in Sydney, Australia, February 2015 and encompass for the GPCC – inter alia - the identification and combination of unique sources of precipitation data, guidance documents and ultimately the development of improved data sets more ready to study extremes.
Along with the WCRP GC on Weather and Climate Extremes GPCC will continue its efforts towards improved GPCC products derived from historic daily precipitation data information through an elaboration of ETCCDI kinds of extreme indices
Development and release of a new Homogenized Precipitation Re-Analysis (HOMPRA) Product (adjusted for climate variability and trend analyses) covering the time period at spatial resolutions 0.5°, 1.0°, and 2.5°, HOMPRA is planned to become available in a Europe only version in course of year 2015 and in a global version later on. (Delayed from 2013).
Technically GPCC will add an import pre-processor for the OGC compliant XML standard Water ML 2.0 to its data base import software in order to explore data sources and providers beyond the WMO NMHSs
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4

35. Thank you for your attention ! http://gpcc.dwd.de
ftp://ftp-anon.dwd.de/pub/data/gpcc/html/download_gate.html
Global Precipitation Climatology Centre, AOPC XX – March Agenda Item 7.4