As a result of the Global Precipitation Climatology Centre (GPCC) offer, the meeting recommended that The (GPCC) be recognized as the global centre for quality monitoring of precipitation. GPCC should provide the results of their monthly precipitation monitoring to the Lead Centres for surface data monitoring for their use in developing 6-monthly consolidated reports.

The Team considered that verification of precipitation including snow depth in the short- and medium-range would, for the time being, continue to be an area of research and development and should remain under the auspices of WGNE and GCOS/AOPC. It further noted that other surface weather forecast elements such as wind and temperature are tailored for national users by using MOS/PPM procedures to correct for local effects. Their verification requires a high density observation network that is available only at a national level. The Team therefore recommended that all GDPS centres make their results from verification of surface weather elements available to users on their web site.

The Team reviewed the current draft standard verification system (SVS) for long-range forecasts, with a view to adopting from these SVS proposals the relevant aspects of the procedures, which may be used to enhance the verification and forecast surface elements in the short and medium-range. It was noted that for reasons above, this did not appear feasible at this time

1.2 Mr Morrison Mlaki, on behalf of Professor G.O.P. Obasi, Secretary General of WMO, thanked Mr Yamamoto for guidance provided to the meeting in his opening address and JMA for hosting the meeting. He expressed his appreciation to the chairman and members of the Team for their efforts in the work of the Team and the efforts they will devote during the week to assure relevant outcome and deliverables to meet CBS expectations. He noted that one deliverable expected of the Team was reporting on feasibility to add quality control of precipitation to existing monitoring of land surface observations. He further noted that other related issues focus on additional standardised quality monitoring of surface data elements, forecast verification and procedures for exchange of results, mainly in anticipation of requirements for increased monitoring and prediction of severe weather events.

2.1 The Agenda of the meeting, as adopted by the meeting, is given in Appendix I. It agreed on its working arrangements and the list of participants is given in Appendix II.

3.1 Preliminary findings of the quality of amounts of precipitation in synoptic reports were presented. This used an accounting method to derive, if logically possible, 6-, 12- and 24-hour amounts from reported amounts, including 18-hour ones, and possibly to infer any of these amounts, if a whole synoptic report were missing. It was then possible to sum these amounts over a period, such as a month, so that an assessment of the consistency of reporting could be made.

3.2 One principal result was that an ambiguity in distinguishing whether the single digit indicator for amounts of 0.0mm refers to a period of 6 or 12 hours does occur and may only be resolved by a decoder with explicit knowledge of reporting practices. The use of t_R =0 as an indicator for a 12-hour amount requires complex conditional processing, which is still not completely exclusive when used in the accounting method. The rounding of reported amounts above 1mm in the 6-groups in Sections 1 and 3 of the SYNOP report and code figures for those below 1mm, in contrast to intervals of 0.1mm for 24-hour amounts in the 7-group in section 3, require a fairly large criterion of 3mm for comparing computed totals and 24-hour reported amounts and may also lead to their use in the coding of the 7-group.

3.3 These results and the fact that it is impossible to extract or infer a 6-hour amount from many countries in one particular Region, because the Regional reporting practices only require 12-hour amounts, limit redundancy, which is very desirable for any reliable quality control. Various options for improving the redundancy in and quality of reported amounts were described. An increase of the number of reported 6-hour amounts would be the most desirable of the three most common reported amounts, and an associated increase in 24-hour ones, including those of 0.0mm, would also improve redundancy.

3.4 To address shortcomings identified in accounting for correct accumulation of precipitation amounts reported over a 24 hour period, in accordance with the current variety of applications of global procedures and different regional reporting practices, the meeting considered the possibility of standardising global procedures to make the reporting of precipitation for the preceding 6 hours mandatory at all four main synoptic hours 0000, 0600, 1200 and 1800 UTC. This would however entail changes to current global reporting procedures and changes to agreed regional reporting practices in most regions. The meeting therefore considered and recommended other approaches that involve only minor changes in global procedures and regional reporting practices and an insignificant increase in data volume exchanged and which are efficient with respect to correct accounting of accumulated precipitation amounts.

3.5 The meeting noted the following main reasons as to why availability of precipitation for the preceding 6 hours was of importance to data processing and forecasting activities and as to why six-hourly rather than 12-hourly or daily precipitation is a minimum requirement in various fields of meteorology and hydrology :

• For synoptic applications and nowcasting techniques in weather and severe weather watch;

• For the purpose of quality control (QC), redundancy in precipitation reports is necessary. SYNOP reports should include at least 6-hourly and daily precipitation amounts.

• Operational validation of NWP needs six-hourly reports ;

• As the simulation of physical processes in regional and global NWP models is becoming more and more sophisticated and their horizontal resolution is continually increasing, they are now able to predict the diurnal variation and meet the target to predict severe weather, particularly heavy precipitation. Therefore 6-hourly precipitation amounts are a minimum requirement and 3-hourly ones are preferable

• Precipitation is one of the important atmospheric forcings which determine soil moisture. Soil moisture is very important in predicting near surface meteorological parameters as well as in climate prediction. Because variation of soil moisture involves fast processes like evapotranspiration and surface runoff, high frequency precipitation data is an absolute necessity.

• In order to be assimilated in NWP, the precipitation data have to have high frequency and high density.

• For the validation of remote-sensing data, high frequency observational data are necessary. Here also resolving diurnal variation is important.

3.6 Noting and considering the above discussions the team recommends that:

• When zero precipitation was measured for the preceding period, it shall be mandatory and reported as zero together with the relevant reference period (6000t_R). The recommended preferred approach only involves an amendment to delete the specification of i_R = 3 in Code Table 1819 and specify RRR = 000 as zero precipitation in Code Table 3590.

• Members be urged to report 24-hour precipitation amounts in the 7R₂₄R₂₄R₂₄R₂₄ group of Section 3.

• Regional Association II Members be invited to consider aligning their regional reporting practices so as to report precipitation for the preceding 6 and 24 hours in addition to current 12 hours, as appropriate, in accordance with global procedures.

• Members whose national practices deviate from global procedures and regional practices, with consequent negative impact on the precipitation quality monitoring programme, be invited to reconsider their position with a view to ameliorating the problem by conforming to global procedures.

• The list of QC focal points, being a list of quality monitoring contacts in each NMHS, should be updated on a regular basis and made available on the WMO web site and occasionally published (say 6 monthly) in the WMO Monthly Newsletter.

• The need to enhance or establish links between monitoring lead centres and observing system operators, where such links are missing, and to foster wider dissemination of quality control results and fast feedback from observation system operators.

• The need to plan for use of CREX/BUFR for exchange of additional data requirements.

3.7 The Global Precipitation Climatology Centre (GPCC), operated by DWD, Offenbach, is a component of the Global Precipitation Climatology Project (GPCP) within the World Climate Research Program (WCRP) and provides global precipitation data based on observations for use in verification of global climate models, investigation of climate anomalies (ENSO) and determination of the earth’s water balance.

3.8 The activities of the GPCC extend from collection of world-wide precipitation data, quality-control (QC) and correction of conventionally measured data, calculation of gridded area-mean monthly precipitation totals from conventionally measured data for the earth’s surface, error assessment on the grid with respect to systematic measuring errors, stochastic errors and sampling errors in area-mean, combination of raingauge-observed and satellite-based estimates to the dissemination of the gridded products.

3.9 The meeting noted with appreciation the offer of the GPCC to provide the results of global precipitation monitoring to the WMO Regional Lead Centres for Land Surface Data Quality Monitoring. The detailed GPCC monitoring procedures are given in the annex to this paragraph. As a result of the GPCC offer the meeting recommended that:

• The Global Precipitation Climatology Centre (GPCC) be recognized as the global centre for quality monitoring of precipitation. GPCC should provide the results of their monthly precipitation monitoring to the Lead Centres for surface data monitoring for their use in developing 6-monthly consolidated reports. These reports to be distributed in accordance with current CBS Lead Centre monitoring procedures, with the request for feedback from the NMHSs. The lead centres should also provide relevant feedback on remedial action on identified problems by NMHSs to other Lead Centres and the GPCC, as appropriate.

4.1 The meeting reviewed the current provisions in the Manual and the Guide. It noted with appreciation the proposal made by Dr M. Ondráš and Dr I. Zahumemský, Slovakia and commended them for the extensive work and recommended adoption of corrections and additions proposed to be included in the current guidance on surface data quality monitoring methods for surface weather elements, as contained in the Guide on the GDPS. The proposed updates are given in the annex to this paragraph. The Team agreed to provide any feedback on the final agreement at the latest by 7 August 2000 following which a supplement to the guide would be issued.

4.2 In the light of its recommendation that the Global Precipitation Climatology Centre (GPCC) be recognized as a global centre for quality of precipitation monitoring, the Team entrusted one of its members to develop general guidance reflecting GPCC procedures for precipitation quality monitoring, with relevant reference to their operational procedure publications, to be included under section 6.3.3.1 of the guide on the GDPS.

5.1 The team noted a precipitation verification project undertaken by WGNE, in which daily precipitation amounts predicted by several operational global models were compared with observed precipitation from dense national raingauge and weather radar networks by three NWP centres (DWD, BMRC and NCEP). The verification results are not made public because they are still considered preliminary and the techniques are still being refined.

5.2 The team also noted a GCOS/AOPC project on the intercomparison of global snow depth analyses. It was shown that there were still marked differences in the three analyses from JMA, CMC and DWD, in particular over the Eurasian continent. The differences were attributed to the in-situ snow depth reports received and utilized at those centres, guess fields and spatial scale of smoothing in the analysis scheme. The results showed that using those analyses as a first guess would be difficult and prediction of snow depth would be more difficult.

5.3 The Team considered that verification of precipitation including snow depth in the short- and medium-range would, for the time being, continue to be an area of research and development and should remain under the auspices of WGNE and GCOS/AOPC. It further noted that other surface weather forecast elements such as wind and temperature are tailored for national users by using MOS/PPM procedures to correct for local effects. Their verification requires a high density observation network that is available only at a national level. The Team therefore recommended that all GDPS centres make their results from verification of surface weather elements available to users on their web site.

5.4 The Team recommended that with respect to verification of precipitation, including snow depth, CBS should consider operational implementation of the activity in the light of research results and future availability of more accurate operational remotely sensed data and increased exchange of in-situ data or if WGNE prematurely discontinues precipitation verification.

5.5 The Team reviewed the current draft standard verification system (SVS) for long-range forecasts, with a view to adopting from these SVS proposals the relevant aspects of the procedures, which may be used to enhance the verification and forecast surface elements in the short and medium-range. It was noted that for reasons above, this did not appear feasible at this time. Some members of the Team expressed reservations on the use of the sparse GSN network for verification in lieu of a denser network and also felt that care should be taken in using precipitation from re-analysis data in verification.

6.1 Methods used by the Lead Centres to exchange current monthly and 6-monthly surface quality monitoring reports will be used for the exchange of precipitation quality monitoring results in co-operation with GPCC.

6.2 The Team reviewed current procedures in the Manual of GDPS and recommended the addition of standards for the monitoring of geopotential height from surface observations. It was also agreed to include general guidance reflecting GPCC procedures for precipitation quality monitoring under section 6.3.3.1 of the guide on the GDPS as indicated in paragraph 4.2 above.

6.3 The Team recommends that the Lead Centres for the other types of observations be invited to develop and complete procedures and formats for the exchange of monitoring results for inclusion in the Manual of the GDPS, in accordance with provisions of Attachment II.8 paragraph 1.1.

7.1 The meeting was closed on Friday 23 June 2000.

Monitoring of Precipitation Data on Global Scale at GPCC

The QC procedures are done in different steps: after a pre-check of recorded locations and station identifiers, individual preliminary checks of the precipitation data are performed. This implies primary data of about 1500 CLIMAT reports and about 5000 SYNOP reports processed at DWD, and about 6000 SYNOP reports from NCEP, Washington, processed at its Climate Prediction Center (CPC). Additional secondary data, such as historical data, regional data and national data from NMHSs are also taken into account. All these data are subject to automatic correction of typical typing and encoding errors using full SYNOP information, gap-filling and statistical and visual QC using quintiles. After these checks all data are merged into the precipitation PDB (Point Data Base) of GPCC.

The final QC is performed in two steps: the automatic part comprises the selection of the ‘best value’ for each station according to the pre-check, calculation of gridded area-means for selected data (first guess) and correction of ‘bad data’ (typographical errors) if possible. The next steps are the check and correction of positional errors, selection of a ‘best value’ of all available sources for suspect stations, check of consistency in space and against climatology and finally a crosscheck of data of different sources.

The visual check of suspect data comprises the consideration of topography in the vicinity of suspect stations using a workstation-based visualization system and a final classification of suspect data as correct or wrong and correctable or uncorrectable.

The most frequently detected problems are errors of position and/or elevation in the station database, unknown or incorrect time interval indicators t_R, inconsistently interlocked precipitation data, for example the 6-hour precipitation amount larger than the 12-hour one, precipitation amount of ‘0 mm’ instead of ‘not available’, format errors within FM12 code, contradictory SYNOP and CLIMAT reports and any kind of typographical errors.

The final product after QC procedures is the calculation of gridded area-means based on the corrected data-set.

The reasons for this extensive multi-stage full quality-control procedure, including manual intervention and visual revision, of monthly precipitation are:

Annex to paragraph 4.1

Proposed updates to the Guide on the GDPS

Quality Control Procedures

Table 6.2.1
Fixed limits for parameters in coded form (in case of indicator i_x = 1 - 4)
(The parameters are considered to be erroneous outside the limits.
The denotations are the same as those used in the Manual on Codes, WMO-No. 306)

6.3.2.1.1 CONSISTENCY CHECKS FOR SURFACE DATA

(in case of indicator i_x = 1 - 4)

The different parameters in SYNOP reports are checked against each other. In the description below, the suggested checking algorithms have been divided into areas where the physical parameters are closely connected:

(a) Wind dd/ff

The wind information is considered to be erroneous in the following cases:

dd = 00 and ff ¹ 00;
dd ¹ 00 and ff = 00;
dd = 99 and ff = 00 or ff ³ 05ms^-1;

(b) Visibility VV and weather ww

The values for visibility and weather are considered suspect when:

(c) Visibility VV and cloud information

The values for visibility and cloud cover are considered suspect when:

(d) Cloud information

The values for cloud cover are considered erroneous when:

N < N_h;

1 £ N_h £ 8 and {C_L = 0 and C_M = 0 or C_L = /};
{N_h = 8 and 1 £ C_L £ 9} and {C_M ³ 0 or C_H ³ 0};
N_h = 8 and C_L = 0 and C_H ³ 0;
N_h = 0 and {C_H = / or C_H = 0};
0 < N < 9 and N_h = C_L = C_M = C_H = 0;
N = 0 and h ¹ 9;
N = 9 and h ¹ /;
N = 9 and 0 £ N_h £ 8;
N = 9 and N_s ¹ 9;
C_L = 0 and C_H = 0 and N_h ¹ N;
C_M = 0 and C_H = 0 and N_h ¹ N;
C_L > 0 and {N_h = 0 or N_h = /};
C_M > 0 and {N_h = 0 or N_h = /};
C_L = 0 and C_M = /;
C_M = 0 and C_H = /;
C_M = / and C_{H ¹} /;
C_L = 0 and C ³ 6;
C_L > 0 and C < 6 in the 1^st group 8N_sCh_sh_s;
C_L = 3, 9 and C ¹ 9 in all groups 8N_sCh_sh_s;
C_L ¹ 3, 9 and C = 9;
{C_M = / or C_M = 0} and 3 £ C _£ 5;
{C_M = 1 or C_M = 2} and C = 3;
{3 £ C_M £ 6 or C_M = 8} and 4 £ C £ 5;
1 £ C_H £ 4 and C = 2;
7 £ C_H £ 8 and C = 0;
0 £ C_H £ 8 and C = 1;

C_H = 7 and N ¹ 8;
N_s = 9 and C ¹ /;
N_s = 9 and N ¹ 9;
N_s > N;
1 £ N_s £ 8 and C = /;

(e) Cloud information and weather ww

Clouds and weather are considered suspect when:

N ¹ 9 and {ww = 43 or ww = 45 or ww = 47 or ww = 49};
N = 0 and {ww = 03 or 14 £ ww £ 17 or 80 £ ww £ 99};
N < 5 and 50 £ ww £ 59;
N < 3 and {60 £ ww £ 69 or 72 £ ww £ 75 or 77 £ ww £ 79};
N_h = 0 and {50 £ ww £ 75 or 77 £ ww £ 99};
95 £ ww £ 99 and C ¹ 9;

(f) Temperature T and weather ww

Both elements are considered suspect when:

T > +3° C and {56 £ ww £ 57 or 66 £ ww £ 67};
T < -10° C and {56 £ ww £ 57 or 66 £ ww £ 67};
T > +3° C and 48 £ ww £ 49;

(g) Temperature T and dew-point temperature T_d

Both values are considered suspect when:

T - T_d > 5° C and 40 £ ww £ 49;
T_d > T;

(h) Pressure tendency appp

The values of a and ppp are considered erroneous when:

a = / and ppp ³ 0;
a ³ 0 and ppp = /;
a = 4 and ppp > 0;
ppp = 000 and 1 £ a £ 3;
ppp = 000 and 6 £ a £ 8;

(i) Weather ww, W₁, W₂

The values of ww and W₁, W₂ are considered erroneous when:

00 £ ww £ 03 and 0 £ W₁ £ 2;

(j) Past weather W₁ , W₂

The values of W₁, W₂ are considered erroneous when:

W₁ < W₂;
0 £ W₁ £ 2 and W₁ ¹ W₂;

(k) Past weather W₂ and cloud cover N

The values of W₂ and cloud cover are considered suspect when:

W₂ = 0 and 5 £ N £ 8;
W₂ = 2 and 0 £ N £ 4;

(l) Weather ww and wind speed

The values for ww and ff are considered suspect when:

The values of T and T_n , T_x are considered erroneous when:

T < T_n ;
T_x < T;

(n) Precipitation RRR and past weather W₁ , W₂

The values of RRR and W₁ , W₂ are considered erroneous when:

{5 £ W₁ £ 8 or (W₁ = 9 and 5 £ W₂ £ 8)} and RRR = 000;

(o) Indicator i_R and precipitation RRR

The values of i_R and RRR are considered erroneous when:

i_R = 4 and RRR = ///;

(p) Indicator i_x and weather ww, W₁ , W₂

The values of i_x and W₁ , W₂ are considered erroneous when:

(q) Check on station pressure reduction

The paragraph – without any changes

3.1 The criteria for the production of monthly list of suspect stations are as follows:

…………..

3.1.2 List 2: GEOPOTENTIAL HEIGHT

Element: Geopotential height, from surface synoptic observations or derived from station-level pressure, temperature and published station elevations at 0000, 0600, 1200 or 1800 UTC compared to the first guess field of a data assimilation model (usually a six-hour forecast).

Number of observations: at least five for at least one observation time, without distinguishing between observation times.

One or more of the following:

Absolute value of the mean bias 25m
Standard deviation 35m
Percentage gross error 20% (gross error limit: 100m).

3.1.3 PRECIPITATION

General guidance reflecting Global Precipitation Climatology Centre (GPCC) procedures for precipitation quality monitoring are given under section 6.3.3.1 of the guide on the GDPS.

NOTES:

(1) All monitoring centres are asked to conform to the above specified criteria. These monthly lists should be prepared for at least the regional association of the lead centres and, if possible, for other regional associations. Consolidated lists of suspect stations should be produced every six months by the lead centres (January–June and July–December) and forwarded to the WMO Secretariat for further action.

(2) The stations on these consolidated lists should be those appearing on all six monthly lists of the lead centres. Other stations could be added to the consolidated list if the lead centres judges that there is sufficient evidence for their inclusion. Each centre should send its proposed consolidated list to all participating monitoring centres for comment. The final list would then be forwarded to the WMO Secretariat.

APPENDIX I

AGENDA