I have observed that there are many meteorological stations in the urban area of Girona with questionable data, thermometric and rainfall data. My question is how to know which stations are doubtful and what methodology of stadistics or geography do you propose to separate the stations with reliable meteorological data from those with faulty data?
There are standards to meet in the equipment, location and servicing to meet the research level meteorology station. Most stations include recording instruments such as weighing or tipping bucket raingauges, and the standard non recording raingauge as a backup. The English version has an 8 inch diameter opening, and no obstructions trees, buildings, etc. from a clearing at a 45/degree angle above the raingauge opening. Most include wind dampening apparatus. Thermometers are located in a shaded building with good ventilation. Quality stations will probably include some background information on who collected the data, equipment used, location. It is possible with recording data to have some loss records, but that is why a backup non-recording gauge can help with some data recovery. It would be a challenge to separate the good from the poor without substantial work.
I would like to add, some storm systems are widespread with low variability across the landscape, other systems have very localized cells with great variability across the landscape. Having a vast network of fair station quality may be as good as one or two excellent stations. It is very time consuming to pair individual station data versus Doppler rainfall recordings or rainfall estimates if you have access to that detail. If you learn raingauges are under trees, next to buildings, substandard size or opening not level, these may be especially problematic, sometimes reliable, sometimes not. In GIS, you might color code the various stations and rate them as to quality to help identify the best network for your needs.
As already stated, you want to make sure that are few data gaps (< 20%) of the full record. You also want to make sure that there are no spurious values within the record and that all values seem reasonable (i.e.15 minute temperature readings consistently being 11 C, 11.5 C, 11.5 C, 30 C, 11. 5 C...). The same would hold true for other meteorological variables (precipitation, humidity, wind speed...). Finally, you should verify that a site is using standardized methods to collect the data.
Thank you for your answers. I think that 20 % of meteorological stations of Girona's urban area have data with mistakes.
I have used this tool to homogenize meteorological data: http://www.climatol.eu
Most effective way to prevent the occurrence of systematic errors is to improve observer training and to establish a near real‐rime quality control that allows quick intervention if an error occurs.
Article Identifying, attributing, and overcoming common data quality...
https://www.wmo.int/pages/prog/wcp/wcdmp/hq- gdmfc/documents/QC_QAguidelines-April2019.pdf
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling