The Ontario-Quebec VHF windprofiler radar network

The O-QNet / DATA


March 6, 2004. A team of researchers within Ontario and Quebec has recently been funded by the Canadian Foundation for Innovation (CFI) to construct a multi-million dollar network of windprofiler radars within those two provinces. The project was originally awarded to a team led by Prof. W. Hocking of the University of Western Ontario in 2002, but was then returned to the granting body at the special request of the UWO administration. The project was then re-allocated to York University, where Prof. Peter Taylor is now the principal investigator. In this article, we will describe the importance of these radars, and discuss some of their general features of operation.

The participants in this project are: Peter Taylor (Principal Investigator) and Gary Klaassen (York University), Wayne Hocking, Bob Sica, Gordon McBean, Horia Hangan, John Barron, Bob Mercer (all from the University of Western Ontario), and Isztar Zawadzki and Frederic Fabry (McGill University). The Meteorological Research Branch (MRB) of MSC plans to collaborate on applications Research and Development projects related to assimilation of the windprofiler data into NWP models, model validation, and severe weather studies. (Contacts: Jim Abraham, Director, MRB; Dave Steenbergen, Chief, Data Assimilation and Satellite Meteorology Division; Stewart Cober, Chief, Cloud Physics Research Division)

As currently proposed, the network will have radars at the sites indicated in this MAP. However, small changes could occur due to logistic reasons as the project unfolds. The project is due for completion in 2007/2008.

Here are some photographs of some VHF windprofiler radars. The new radars will all look a bit like the ones in the photograph of the McGill radar below, which was completed in mid-2002 by a Canadian company (Mardoc Inc.), and has been running successfully ever since.
1. The Clovar VHF windprofiler radar north of London, Ont.
2. The VHF windprofiler radar at Resolute Bay, Nunavut, Canada.
3. The VHF windprofiler radar at McGill University, Montreal, Quebec.

You can see the results from the McGill profiler updated hourly at The McGill WindTtracker web site

About Wind Profilers

Radar windprofilers are instruments which measure wind speeds and other related parameters within the atmosphere over a relatively deep region. Depending on the frequency and power used, this height region could cover typically 400 meters altitude to 10 or 15 km altitude, and even as high as 25km for the most powerful radars. Unlike the more widely known weather radars, which primarily produce horizontal maps of precipitation and its motion, (and which are sometimes called Doppler radars, a terminology we will avoid in this text because windprofilers also measure Doppler winds), windprofilers use near-vertical radar beams. While weather radars can measure atmospheric motion to ranges of 100-200 km, these measurements are restricted to areas with precipitation or occasionally clear-air targets in the lowest 1-2 km of the atmosphere. When wind measurements are possible, often only the radial wind component can be directly measured. Windprofiler radars, on the other hand, are designed to measure the wind vectors, and other pertinent atmospheric parameters, immediatly above the site. Their specialties include their height coverages, their ability to determine both vertical and horozontal wind vectors, and their ability to monitor motions with high temporal resolution from heights close to the ground and into the upper troposphere and stratosphere during all conditions, including clean-air. They can also measure turbulence strengths. They have also been called MST (Meso-Strato-Troposphere) radars because of their ability to measure winds in each of the mesosphere, stratosphere, and troposphere. The most powerful such radars can make useful measurements up to about 25 km altitude, and also between 60 and 90 km altitude. In the troposphere and stratosphere, the inhomogenities responsible for the radar scatter can be generated by turbulent mixing of existing temperature and humidity gradients, and in the mesosphere they are generated primarily by the mixing of existing large-scale electron density gradients. A recent review of these radars can be found in Hocking, Radio Science, vol. 32, pp2241-2270, (1997).

The Physics103 web site contains an extensive discussion about the general principles of weather radars, including windprofilers. This discussion is pitched at a moderately elementary level.

Here is another useful link which describes something about the operation of windprofiler radars in a general way. This describes the NOAA network. However, our network will use different frequencies - we will use the 40-60 MHz band, whereas the NOAA network operates at frequencies around 400MHz. We have chosen to use the lower frequencies because they are less susceptible to contamination by birds, insects and precipitation.

The attached figures show some typical data obtained with the CLOVAR windprofiler radar, situated near London, Ont. The radar can measure winds from about 600 metres up to 8 km or so above the ground. It can also measure vertical wind speeds, as also shown in some of the graphs below. The new radars are anticipated to have about 30 times better detectability than the Clovar radar, so that we expect to receive useful winds up to about 14-15 km in altitude.

1. Upper level winds, Oct. 13, 2001, Clovar Radar
2. McGill WindTtracker winds, September 21-23, 2002
3. Lower level winds,May 4-5, 2001
4. Sample vertical winds

Project Specifics and Rationale.

The technology behind these radars is now relatively mature, and they have been used in experimental modes for the last 15-20 years. However, their capabilities are best realized when used as networks, and especially when used in conjunction with numerical weather prediction (NWP) computer models. While attempts have been made to integrate them into such NWP schemes, their potential is still largely untapped. One reason for this is that NWP models are not quite yet capable of fully utilizing the large amounts of data available from these radars. This capability will only occur with the development of new four-dimensional veriational analysis computer models running on computers with massive data-handling capabilities, and it is only now that such machines are becoming avaliable. Most major weather centers, including the Meterorological Service of Canada,are working to develop such schemes.

Within the USA, a network of profilers has been established by NOAA (National Oceanic and Atmosphereic Administration), using a frequency in the 400-500 MHz band. However, this choise of freqency leads to contamination from other scattering targets like birds and insects, and can on occasion lead to incorrect interpretation of wind motions. Within Europe, another network (CWINDE) has been established, but it is designed on a "contribution basis only". Various researcher organizations there have agreed to contribute data from their own radars to a larger data base, and these radars have a wide variety of operating parameters. Examples from the NOAA site in the USA can be seen at the web site, and from the European network at

Canada at present has very limited capabilities in this area. There are three existing VHF windprofilers which are capable of regularly measuring winds above 2 km altitude; one at London, Ontario, one at Resolute Bay (Nunavut), and one under construction at McGill University. The last one has resulted from a previous successful CFI application (#2594). In addition there are two small UHF-band profilers which can preform studies in the region below about 2-3 km altitude, and a couple of sonar-based instruments.

Project Objectives.

Our overall objective in this proposal is to establish a network of windprofiler radars within Ontario and Quebec, with the intention to use it to demonstrate the usefulness of such a network to weather forecasting and atmospheric science. The network will be closely linked through the internet, with frequent updates of data supply to a central server for access by all relevant researchers. Sites have been found as shown in the map above, and in addition we will be able to employ the existing sites at London and Montreal. Our network design is also arranged so that the existing UHF profilers can contribute to the network in a complementary manner.

Our network will differ in several ways compared to those discussed above.

First, we with to concentrate on the VHF (Very High Frequency) band, using frequencies in the range 40 to 55 MHz. This will largely avoid the problems related to signal contamination due to birds and insects which have affected the NOAA profilers, and will also reduce the contamination which can occur due to precipitation. In the past this frequency band has been avoided, because VHF radars cannot normally measure below about 1.5 km altitude, but recent developments have shown that with the right choice of antennas, and the correct method, measurements as low as 400 m altitude are possible.

The attached photograph shows a loop antenna, which has proven to be very useful for measurements of winds at these lower altitudes.

Secondly, our network will not only produce wind motions (both horizontal and vertical), but will also routinely measure the strengths of turbulence. This feature is not normally implemented in a network of this type. This will be important for studies of atmospheric diffusive transport at these upper levels, and also from the perspective of air traffic safety. This latter issue will be discussed shortly.

Thirdly, our network will be somewhat more tightly clustered than the others (and much more tightly clustered than the existing radiosonde network), and will be sited in a geographically and meterorologically fascinating area, near the Great Lakes of North America. (Tornadoes, for example, are more common here than anywhere else in Canada, and lake breezes have very important effects on local meterorology). For example, see the very interesting work being undertaken in the ELBOW project . Within the USA, weather forecasters have already found the experimental windprofiler network there to be of considerable use, and claim that it has already saved many lives by allowing better anticipation of severe tornadoes e.g. see the site associated with the Oklahama/Southern Kansas Tornado outbreak of May 3, 1999 . (Warning: This is a large .pdf file, 4.9MB in size - do not click this URL if you have a slow computer link). This report states in part: "In the opinion of the Service Assessment Team, without profiler data, SPC forecasters would not have upgraded from moderate to high risk" (in relation to the tornadoes of May 3, 1999). In addition, the report makes the statement "Also, the state of readiness of NWS offices, emergency managers and the media in the severe weather outbreak would not have been as high". Windprofiler data is therefore credited with saving both lives and capital because forecasters were able to recognize the signs of tornadoes with advance warning.

Fourthly, we will work closely with the airlines of Canada, to examine the capabilites of these radars for improvement of airline safety (reduction of encounters with clean air turbulences) and for reduction in fuel costs by utilizing better flight planning strategies based on better knowledge of upper level winds. Encounters with turbulence still represents an important problem for the airline industry, both from the perspective of aircraft damage, and injuries to staff and passengers. Associated with high levels of injury are high insurance premiums, so a better understanding of turbulence can reduce these overheads to the airline industry, resulting (we would hope) in reduced airfares. Recent turbulence encounters have been reported by various news agancies, including CNN and USA TODAY . Research into upper level turbulence is a priority for several organizations, as shown at this site , which discusses a major initiative called SCATCAT to study upper level turbulence.

Fifthly, we will work closely with other organizations such as fire-fighters, forestry officials and local forecasters to make the data accessible for their own particular applications. Fire fighting is one example of possible important applications of these data. Often knowledge about wind motions at heights of 1000 to 1500 metres is inadequate during periods of fire activity, and this makes it hard to determine where the smoke and sparks from the fires is likely to drift. Windprofilers can provide this knowledge if one happens to be located nearby.

Sixthly, we are keen to examine the potential for improvements by incorporating the data from this network into new computer models which take advantage of the latest advances in computer speed, parallel processing and storage capability. These models will cover a variety of grid scales, from eddy-scale simulations (which will take advantage of the high time resolution of these radars) to mesoscale models and beyond. We will especially be looking at the roles of gravity wave processes in these models, and once again the high temporal reolution will be important here.

We will also expand our studies to include topics like troposphere-stratosphere ozone and pollutant transport, in order to investigate upper atmosphere phenomena related to, among other things, global warming. Furthermore, knowledge of the upper level wind field will also be important for longterm studies of long-lived pollution transport and even for tracking radionuclear contaminants which might be produced, for example, in the (hopefully unlikely) event of a nuclear incident.

There are several other interesting aspects to this program, which are perhaps a little less mainstream. One interesting program will be use of an S-band radar near Long Point to study small scale turbulence and at the same time make some studies of bird migration - an interesting and very useful capability of the higher frequency radars. (Birds cannot be seen with the VHF radars).

For discussions about the other research activities of W. Hocking, please go to the The Atmospheric Dynamics Web Site.

Curently available O-QNet sites can be found at:

All sites except McGill

McGill (Montreal)

Wayne Hocking

(Last updated Jan 11, 2011)