AWS/Weatherbug Stations commentary

[peteywheatstraw]

As a weather buff and scientist, I frequently refer to sources of climatic data to get an accurate idea of real time and recent weather conditions and trends in my area and further abroad. In this regard, I have found the information provided by AWS/Weatherbug very useful since it claims 8000 neighborhood stations across the US. I especially like looking up current weather conditions and data from offbeat locations. I've got a customized site list for spots all along the Gulf Coast where I live. One such site, for example, is Cathedral Academy, in the French Quarter of New Orleans. As far as I know, this is the only published w/station that has operated right in the French Quarter. Many sites have live cams, which is really cool, although quality varies.

However, using 'Weatherbug' (which claims to be the represent the third most frequently visited weather site after NOAA/weather.gov and weather.com) does appear to have some drawbacks:

If you use the paid WeatherBug program ($20/yr), be aware that it of course, takes up its share of space, and if it is loaded, produces annoying popups, including your 'home station' of choice upon boot and NWS weather statement prompts (which may or may not be important) which can't be killed until read.

The 'historical data' section for selected stations needs to be improved. Data only goes back one month and is generalized. Basic hourly data only can be retrieved from the past four months, not very long climatologically.

You can't look up climatalogical data for multiple stations at a glance, this must be done one site at a time.

Of the weather stations themselves, herein lies the real issue. While many stations report seemingly accurate real-time information, many suffer from apparent quality control issues.

It is fairly easy to tell if a rain gage is inoperative - if a heavy, widespread rain event occurs in one's area, you know everyone got some appreciable amount of rain. Specific location precip totals can easily be compared against radar estimate and the numerous other rain gages we have these days. A minority of WB stations in my area have what are obviously non-working or non-reporting rain gages. More significantly, many 'work' but apparently under report the totals either slightly or significantly; at least a third of the WB stations in my area appear too conservative. I have not seen any that over report, however.

Temperature bias is another aspect that seems a concern here - Again, while many WB stations consistently report a fairly accurate temperature, a high percentage produce suspect readings. In my area, I'd say about a third or more are not quite right. Commonly, WB stations seem a couple or more degrees off from where they ought to be - usually on the low side. High temperature bias seems to occur less frequently and if it does it's more often with the daily max temps (possibly due to overexposure). Suspected low temp bias appears to be more pronounced on colder, brisk days. I've determined many AWS/WB stations to be off a bit by comparing them repeatedly with nearby NWS stations, agricultural stations, cooperative stations and by interpolation with other WB stations at times when the regional temperature is very consistent, taking into account possible microclimate influence and geographical location.

I am wondering if any one else out there has found similar findings.

[Aslkahuna]

Before one makes a commentary on temperatures, one needs to know the local topography and ground cover as well as the presence to trees. My backyard is consistently cooler at night than the front because of cold air drainage from a nearby wash and I have three thermometers back there. They read close to the official downtown reading from the local Fire Station which is similarly located near a wash. On Post the temperatures are consistently higher than both and even higher than my front yard. OTOH my back yard reads closer to the Post during the Day since it shares the same ground cover while my Front yard with a lawn and trees reads closer to the Fire Station. Mixed terrain locations will show very strong variations over short distances-a bit over a mile away the local Cable TV is in a real pneumonia hollow at night and the areas down by the river have a totally diiferent nighttime climate even dropping below zero at times during the Winter. Another question concerns rainguages. These must be checked frequently as the tipping bucket type has a tendency to read low in heavy rain events-this is why I use my NWS Spotter guage for my rainfall data base. Siting of sensors is also important. I have persoanlly driven around the local area and have seen many of the sites that report on the Weather Underground and by and large the sensors are not properly sited.. First off, the wind sensors are mounted too close to roof line and not 10m above ground. Secondly there's a tendency to have the remaining sensor suite colocated with the wind sensor when they should be mounted separately at the proper height and exposure (temperature sensors should be 2m above ground) putting the temperature sensors near a roof will result in higher highs and lows usually. Temperature sensors are especially sensitive to height since the coldest temperatures at night during the winter occur close to the ground (on Post, the remote mesonet sites had ground level sensors and one at 10m in addition to the standard 2m height and we would record Delta-Ts between those two of as much as 15-20F on very cold nights). Even if the sensors are perfectly calibrated (and most are not-not even ASOS) poor siting will bias the data. My Wind sensor needs to be elevated so I take it's measurements with a grain of salt-especially with southerly winds since I have 10-12m hgh trees nearby to the south.

Steve