Wind roses help us analyze resources
Submitted by Windy Schor
A wind rose is a statistical tool to analyze wind data, combining speed and direction in a multi-dimensional chart. A wind rose helps predict how well a wind turbine will produce energy in the spot where the data was taken. It is also helpful for aircraft and sea-craft pilots, kite-flying, parachuting, and more.
A wind rose is a histogram, a chart showing bars of the relative amounts of wind in each direction. The wind rose may show only the four cardinal directions, or may split the compass into more sections for the histogram.
Also, for each direction, a bar may be split into colors/patterns to show the relative amounts of time the wind spends at each bin of speeds, in that direction. Some roses discard any winds below a threshold as too low to measure or "calms."
Wind roses for education
Despite the complication of showing statistics in multiple dimensions, I have seen teams of children in the 5th grade easily glean their own inferences from a rose after 15 minutes of explanation and examples. But there is so much one can do with the analysis, that it is certainly a powerful tool for scientists and students alike.
How a wind rose is built
You can access sample wind roses on the internet, but you may better enjoy making your own from raw data. The best would be to gather your own data with an anemometer and direction-vane, then make a rose from that. I have done that in the virtual world "Second Life," which is an online world of hundreds of thousands of users and millions of square kilometers of realistic virtual terrain. In Second Life, the wind blows according to a dynamic, somewhat realistic model, and a wind-rose for a shoreline there looks quite similar to one from real life.
You too can view the data from SL and generate wind roses, excellent statistical tools for analyzing wind resources, at
In real life, The wind usually goes through daily cycles, especially near shorelines. Any periodicity of changes in wind speed/direction should be considered when gathering/using wind data. For example, if you had instantaneous wind speeds at 6pm every day, you would see variance, but the statistics would be biased according to whatever factors might affect the wind every evening, such as onshore winds. Even if you take hourly samples of wind data, if there is a periodicity to the wind that is a multiple of one hour, most obviously 24 hours, it will bias the statistics. One solution is to sample the wind at a random time within each hour. Another solution is to sample the wind more frequently than the frequency of it's variability. The wind doesn't often change drastically in under a second, so a 1-second period would capture most changes in the wind. The problem with this solution is that there are 31557600 seconds a year. It would take about a gigabyte of storage to hold one year of data; not unthinkable, but probably unmanageable. To take a sampling of those data would probably make analysis quicker and easier.
Once you have your samples, you can separate them into bins, one for speeds, and one for directions. A wind rose combines both, so you have two dimensions of bins. For example, if you consider 8 directions (N, NE, E, SE, S, SW, W, and NW) and 10 sets of speed-ranges (0-2 m/s, 2-4 m/s, 4-6 m/s etc.), you'd have 80 bins of data. The precision of your rose is the number of bins; too many bins and you may get lost in the details, too few and the details will be lost themselves.
Many of those bins might have no samples. For example, your samples might never show the wind blowing between 8 and 10 m/s in the NE direction.
Then you must count the number of samples in each bin, and divide that number by the total number of samples to show the portion of time the wind blows at that speed and in that direction. Just the speed bins will tell you most of what you need to know for generating power. In siting wind turbines, direction will be important considering winds from some direction may be more turbulent than others at the site because of hills, trees, and buildings. Also, a line of turbines would best be placed facing the best wind direction.
If you would like to read more about gathering and using wind data, please see the special projects section at EnergyTeachers.org: