Article from the West Bend News

This article is from this week's issue of the West Bend News.

Last week, we talked about how weather variables in the upper atmosphere are measured, such as temperature, humidity, and pressure. This week, I will talk about how the data is screened, and how wind throughout all levels of the atmosphere is measured.

Usually the radiosonde is sent off by balloon on two separate occasions each and every day. These times are 12 UTC and 00 UTC. For everyone who does not know what UTC means, it stands for Universal Coordinated Time, or Zulu time. It is not the same as military time, as it does not vary with time zone. For instance, if it is 12 z (or UTC) in New York, it is also 12 z in Los Angeles. This Universal Coordinated Time is the same as Greenwich Mean Time (GMT), accept for the fact that it does not change with Daylight Savings Time. If you want to know what the UTC time is right now (depending on your time zone), just subtract five hours from the current military time to obtain the UTC time, if you live in the Eastern Standard Time, while under daylight savings time just subtract four hours. Right now, 8 a.m. Is at 12 z and 8 p.m. is at 00z. The balloon which is sent off is filled with either helium or hydrogen.

During a normal flight, it will take in between 37 and 40 minutes for the balloon to reach its maximum height (usually about 15 kilometers), before bursting and descending through the atmosphere. The weather conditions at the time of lift off will determine whether the balloon will go up or not. For example, usually weather balloons will not be sent up in or near a thunderstorm (unless it is conducted for research purposes), because this will cause noise in the data when sent to the National Centers for Environmental Prediction. Usually, the temperature, humidity and wind profile of a thunderstorm is characterized by by a moist adiabatic lapse rate (temperatures decreasing with height at about 6 degrees Celsius/ kilometer), with 100% relative humidity throughout the entire atmosphere, and chaotic wind patterns due to turbulence, mesocyclones, tornadoes, and other small scale circulations within the thunderstorm. Since the grid-spacing of most models is too large for thunderstorms, if this data was placed into the model's analysis grid, to fit with physical equations that govern model, it would have to generate quite a bit of convective precipitation and latent heat to keep the atmospheric conditions accurate. This can cause quite a problem because the model will either 1.) have to assume that the storm is at least the size of its grid spacing (this will cause problems with global models, such as the GFS, ECMWF, and the GEM), or 2.) it will have to smooth out the latent heating throughout several grid spaces, especially for convective complexes. Latent heating has an enormous impact on our weather (a topic saved for another article), and since model physics are fairly close to the real world (not quite though), they will have to account for this incredible area of heating. Then, the models, when run, will disperse the enormous amount of heating, destabilizing the atmosphere even further, which will generate something called a convective or “QPF” bomb, which is basically a giant land hurricane on the model's grid, as a result of the latent heating, which can completely offset all accuracy of forecasting. This is why the data has to be screened for accuracy. Until the day at which models can actually physically create thunderstorms on their grids, this will be a problem. Other noise problems are also taken care of such as inaccurate data due to broken instruments, or elevation problems. Areas such as the Rocky mountains can not measure variables at levels such as the 850 and 925 millibar levels, so after the data is decoded and sent to the NCEP for use in model grids, the pre-processing that the model does before it runs the data uses the hypsometric equation, which relates atmospheric thickness to the average temperature. Using the atmospheric standard lapse rate, (which is rarely the case in a mountainous and dry region), it calculates data for levels that do not exist in the mountains so that the model can run properly. This why model forecasts of mountain conditions are rarely correct.

I never talked about how radiosondes measure wind speed and direction last week, but now I will. In reality, wind is not actually measured by radiosonde, as they do not mount an anemometer to the side of the radiosonde. What they actually do is use a radar tracking device to measure the angle of elevation, the balloon's displacement from its original position, and calculate the wind vectors from the balloon's movement. By doing this though, it leaves a rather large margin of error for the calculation of wind vectors (wind direction and speed) throughout all levels of the troposphere. When put into the archives, this wind information is put into the format of degrees of direction and wind speed in knots. In meteorology, the scale used for direction is: 0 degrees for a north wind, 90 degrees for an east wind, 180 degrees for a south wind, and 270 degrees for a west wind. This wind information is then plotted on atmospheric soundings, called skew-T charts, and hodographs, on which wind information is the only thing plotted. A hodograph is a chart on which many weather variables can be evaluated, such as total shear, bulk shear, and storm-relative winds. These are very important in determining the severity of storms and their structures.

Next, week we will talk more about improvements in the current technologies that are used in obtaining the weather information (I was trying to get to that this week!).