Visualizing 29 Million Wind Measurements in Kansas City

Visualizing 29 Million Wind Measurements in Kansas City

April 16, 2015 6:28 pm Published by

Understory’s network of RTi stations in Kansas City is delivering high fidelity, continuous, and accurate wind speed and gust data. Our sensors are more sensitive and require less maintenance than a typical propeller-type weather station and can collect thousands of measurements every second. This means we know precise statistics for every potentially damaging sustained wind and every gust.

March was a calm month for Kansas City, with only a few blustery days. There were several showers, and one glancing thunderstorm, but no damaging hail was registered by our RTis or reported by other sources, which is unusual for March.

One way to visualize our plethora of wind data is a heat map timeline. The more recent days are on the right; moving to the left is going back in time. Each column represents one hour over this two-week period. Our RTi stations are arranged as rows from the northernmost at the top to southernmost at the bottom. The color of each box represents the maximum wind gust speed for an hour at a particular station. On one end of the color scale, light yellow denotes calm hours, and on the other end, dark red signifies blustery hours. This is a way to visualize more than 8,000 data points synthesized from over 29 million data points.

kc-max-gusts

On most days between March 11th and March 28th, the RTis recorded few gusts above 10 mph. March 14th, 24th, and 26th were windier, with gusts reaching 20 mph. But the strongest gusts occurred during the evening of March 16th, continuing into the morning of March 17th, with a four-hour calm period during the late hours of the 16th. On these days, maximum gusts frequently surpassed 30 mph and the strongest 1-second wind speed measured was 44 mph.

How do we determine the accuracy of our wind measurements?

precipitation-event-report

Our data not only has impressive temporal and spatial resolution, it’s also extremely accurate. As part of our validation we sited our weather station side-by-side a weather station we outfitted with an R.M. Young wind monitor and a Campbell Scientific data logger for some field testing on an accessible flat rooftop.

We had a decently strong Nor’easter come through the Boston area in the last month of 2014. The storm system dropped an estimated 3.1″ in downtown Boston, as shown in our Precipitation Event Report. These reports are a quick way to look at past events with significant precipitation over a 24 hour period.

Now back to the storm as experienced through wind. Let’s take a look at wind speeds over time in the Figure below. Conditions are calm during the early morning hours all the way through 10am. Moving past the mid-morning hours, the storm system arrives and conditions become gusty for several hours.

Our data has so much time resolution that in order to display it here we’ve reduced the size by showing only one out of every 20 seconds. This means some of the highest gusts aren’t plotted; the largest 1-second gust clocked in at 38 mph. Not a major wind event, but enough to be felt on weaker tree branches and vegetation.

 

A graph like this may look like a lot of noise, even to a trained eye. But watch what happens when we zoom in and overlay the data from the weather station sited only 10 feet away.

Both stations are observing real variations in the wind, and the agreement is astounding. Not noise at all. We wouldn’t expect perfect agreement, and we don’t see it, because even though the stations are on the same rooftop, they’re not experiencing identical physical conditions.

Overall the two sets of wind measurements are in excellent agreement. Quantitatively, during times of elevated winds, the mean, median, and max 1-second wind speeds differed between the two stations by 0.4 mph, 0.1 mph, and 2.0 mph, respectively. There are many ways to compare measurements, and if you’re interested in more detail, please ask us!

Back to Kansas City

The accuracy of our hyper-local data networks are very important to us, and that accuracy allows us to provide actionable details about how weather affects properties inside Kansas City. Not only do our weather stations track wind; we have a whole host of measurements — hailstone size and impact angle, temperature, pressure, humidity, and more! We will be describing cool and interesting results from our analyses from our hyper-local weather network in Kansas City (and soon to come, Dallas). Check back here for new content or follow us on twitter.