Operational quality control of daily precipitation using spatio-climatological plausibility testing
Quality control (QC) of meteorological data is fundamental for many weather and climate related applications. The development of an automated QC of precipitation data has proven difficult in the past and many weather services, including MeteoSwiss, still use primarily manual inspection as a QC proce...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Borntraeger
2011-08-01
|
| Series: | Meteorologische Zeitschrift |
| Online Access: | http://dx.doi.org/10.1127/0941-2948/2011/0236 |
| Summary: | Quality control (QC) of meteorological data is fundamental for many weather and climate related applications. The development of an automated QC of precipitation data has proven difficult in the past and many weather services, including MeteoSwiss, still use primarily manual inspection as a QC procedure. Man power limitations and the demand for more objectivity force many weather services to move towards a less labour intensive and more automated QC while meeting the challenge to keep data quality high. Here we present a spatio-climatological approach for daily precipitation that has been implemented operationally at MeteoSwiss in 2011. It combines the information on the spatial distribution of a precipitation event and the climatological information on the interpolation error at every station using different precipitation intensity intervals. Data editor judgement shows that the system is able to detect potential outliers well without creating too many false alarms that need human inspection. 60-80 % of all suspicious values found by the approach have been classified as real errors by the data editor. Small scale convective precipitation challenges the current configuration. The planned introduction of combined ground-radar estimates may help mitigating this problem. Very helpful in the QC process is the automatic redistribution of accumulated several day sums. The presented method is able to reduce manual inspection in operations and the QC of precipitation can be objectified substantially. |
|---|---|
| ISSN: | 0941-2948 |