Citizen Weather Observer Program (CWOP)
APRSWXNET/Citizen Weather Observing Program (CWOP) uses the APRS-IS to collect weather data from citizens (hams and non-hams) with personal weather stations and send these data to users around the country. In addition to providing different types of displays and archiving capability for weather data, CWOP also provides quality checking for the data and feedback to the weather station operator that can be used to improve the quality of the data and increase the value to users.
Over 3000 data contributors send weather data to APRSWXNET/CWOP via APRS-IS on a regular basis. Most of those are stations from North America, but there are a significant number from other parts of the world. There is a wide variety of APRSWXNET/CWOP data available from citizen weather stations around the world.
 Getting Started
There are 22 different software packages that interface with commercially available weather stations and support CWOP. These programs along with links to forums and the software authors are listed here in the table a few screens down. In fact, this is the best place to start if you want to send data to CWOP. Find a software package that interfaces with your weather station and download a trial version. Make sure that you like the look and feel of the software package and if not, try another one. Once the software is all set up, then get a CW number (hams can use their callsign and don't need a CW number) and use that for the ID. Be sure and check the Frequently Asked Questions to see what values to enter into the software. Also, some very helpful software set-up pages are available along with ways to check if your data packets are getting to APRS-IS and to CWOP.
 Weather Station Siting
An important part of measuring accurate weather data is siting of the weather station. Not all residential locations are ideal sites, but some knowledge of what's important in siting weather measuring equipment can greatly improve your data.
- CWOP Weather Station Siting, Performance and Data Quality Guide; 2.2MB PDF developed by CWOP members
- CWOP Resources Page; a listing of many links to different types of information on measuring the weather
- CWOP Radiation Shields; a discussion of the importance of ventilated radiation shields to improve temperature measurements
When the plotted location on the findU maps has been verified as correct, send an e-mail to the indicated location and your station will be registered in CWOP and your data will flow to users including the National Weather Service. Be sure to use as much resolution as possible in your latitude and longitude so your location will be accurate because good weather data from the wrong location are bad data. The APRS lat/lon format will allow position accuracy of about 20 meters, which is good for CWOP weather data where the goal is locations within 200 meters. APRS is a real-time system, so the computer time at the weather station is not a great concern as the data time is taken as the time of arrival at the findU server as is discussed in the FAQ linked above.
 Data Quality Checking Feedback to Station Operators
A unique benefit to CWOP members is the feedback from quality checking of their weather data. This is an important tool that CWOP members can use to improve the quality of their weather data and increase the value of that data to all other users. The quality checking done on the CWOP data is is explained at MADIS Quality Checking and the links from there. Every day, Philip Gladstone (C0003) acquires the CWOP quality checking data from MADIS and assembles them into individualized text reports that can be e-mailed to CWOP members. He also makes individualized graphical content web pages that can be downloaded. One of these is the "site" page, obtained through the first link (Town/City/Meta) on the state list web pages linked from the first group of the CWOP shortform data access page. Among other things, the "site" page shows the station location that was used in the quality checking algorithms. The procedure is to use a numerical weather model that does an analysis to estimate the pressure, temperature, humidity and wind at the station location and then compare these estimates with the measured values. So, it's very important that this station location be correct and the "site" pages gives the station operator a way to verify the location and elevation.
Under the "Data Status" heading, the "site" page also shows the comparison of the numerical weather model and the measurements. These comparisons can be done over different time intervals by selecting the desired interval. For time intervals of 4 weeks or less, the analysis and the measurements are shown in graphs for the last 24 hours of the interval and the statistics are determined over the whole time interval. For time intervals of 8 weeks or longer, the daily graphs are replaced by graphs of the computed statistics. These charts are designed to show the performance of the measurements versus the analysis over a longer period of time. For the barometer chart, the little circle indicates the average error over the week, while the bar indicates one standard deviation. An ideal chart would be one in which the circles were on the zero line and the bars were short. In practice it doesn't work out this way!
The idea is that if you recalibrate your sensor, or resite your temperature sensor, or change the radiation shield, then, after a few weeks, you should be able to see a qualitative difference. The temperature chart is somewhat more complex as it shows the average daytime error and the average nighttime error as well -- these are the yellow dot (think 'sun') and black dot (think 'dark'). Ideally, these averages should be close to each other.
There is a second type of data quality chart which can be reached in the same way, through the CWOP shortform data access page and clicking on the two letter designation for the state you live in. Then click on your station under "CWOP QC" to get a set of graphs that show mean error on the horizontal axis and standard deviation on the vertical axis. It is important to note that the numerical weather model is not "truth" and that the numerical weather model can be in error just as well as your measurements can. Usually, over the long run, the numerical weather model has small average error while over the same time a problem sensor will have larger average error and it is this difference that allows us to find and fix the problem sensor. In each of the charts, the desired target region is indicated with a box area. If your station is not indicated in one of the charts, that is an indication that the location was outside the graph area and not able to be plotted and indicates large error.
One thing that can cause erronous results for this type of analysis is the existence of "micro climates". For exampe, if your location is in a small valley that cold air can settle in, your temperature measurements may be less than predicted by the model if the model topography does not have sufficient resolution to show the valley. In this case, you live in a "micro climate" that is slightly cooler on the average than the surrounding area. So, be sure and consider all aspects of the weather measurement situation when using the quality checking results.
Many aspects of weather data quality checking are discussed in the CWOP Quality Checking e-mail reflector.