It seems that poor PWS siting has been a source of frustration to some of us on this forum. When checking weather online, the siting of the reporting station is often unknown. In rural areas where weather stations are few and far between, NWS has been known to view personal weather stations. Once when at the local forecasting office, I asked if they were aware of the fact that the station they were viewing was mounted two feet above a metal roof and surrounded by grain bins. This station was almost a hundred miles from the NWS office, and no one had been out to view the install. Sometimes properly installed stations are so far in the minority that the one that is right can experience data rejection by website quality control (very frustrating to the one who went to the effort to perform a quality install).
I’m wondering about taking a different approach to the issue of quality data by creating a website that allows users to see the install, and learn what measures have been taken to ensure quality data. Users who wish to post station data on this site will be required to post a minimum of five clear photos of the station (one from north, east, south and west, as well as one aerial or satellite view). If sensors are installed at more than one location more photos should be provided. The aerial view should contain measurements to nearby obstructions. Information such as station elevation should include a note of how this information was obtained. Of course, information such as whether the barometer displays MSL or altimeter setting would be required.
History would be available in the form of graphs and tables. Beside each data archive, a space could be provided for the station owner to mention any calibration or sensor cross checking which had been performed. For example, beside “rain: .68” one could write “CoCoRaHS recorded .73”. One could also mention anything that would cast a doubt on sensor accuracy. For example, beside “humidity: 96%” one could mention “SHT-31 is failing, it’s been foggy all morning”. A box could be checked to allow only readings with notes to be viewed, so historical sensor accuracy, and frequency of calibration and cross checking could be considered. If accuracy of a sensor is still in question, the station owner could be contacted directly.
If the station siting is fully disclosed to all who use data, I don’t see any installation as “wrong”. It would actually be very interesting to look at data from a station in the middle of the Walmart parking lot (possibly the best data when questioning my vehicle’s a/c performance). A station on a high hill will of course report different weather than other stations in the area, as well as will a station on an island in the lake.
No data quality control needs to be used. Those viewing a station can look at the install and steps that the owner has taken to ensure data quality, and make their own decision as to whether or not a station is trustworthy.
Perhaps such a website already exists?
Any thoughts?
