Acquiring Atmospheric Meteorological Data

The main sources of meteorological data at high altitudes are radiosondes and Aircraft Meteorological Data Relay (AMDAR) data [1]. Mode-select (Mode-S) radars are capable of getting similar readings from aircraft as AMDAR; however, this source of meteorological data has not been used systematically so far. In this article, we analyze the meteorological data obtained by means of Mode-S radars and propose how this data can be used. The first prerequisite for acquiring meteorological data with radars is the necessary aircraft on-board sensors that can measure the meteorological data. Next, the aircraft transponder has to fetch the data from the sensors and report them. Mode-S radar configuration further determines whether the radar will request meteorological data from the aircraft. In our experiments, described later, approximately only 6% of responses from aircraft included meteorological information required by the radar. Still, the amount of data collected by means of Mode-S radars is much larger and much cheaper to obtain in comparison with radiosondes. We expect that in the future this percentage of successful responses from aircraft will grow with the modernization of airliner fleets because newer aircraft are better equipped with meteorological sensors. In this article we analyze the quality of meteorological data collected with Mode-S radars. During a 5-month period we collected meteorological data with a Mode-S radar and compared it with corresponding radiosonde measurements. We developed a method for gathering the data from Mode-S radars and generate atmosphere profiles for wind and temperature. Air traffic control uses upper wind tables generated by numerical weather prediction (NWP) models for flight path calculations. The profiles that we generate from Mode-S radar data can be a relatively good substitute for these tables. A recent study by Strajnar [2] shows that both Mode-S wind and Mode-S temperature are of sufficient accuracy for use in meteorology. De Haan and Stoffelen [3] also used a type of Mode-S data as input to a NWP model to compute 1-hour wind and temperature predictions. Since they could not access the actual Mode-S weather data from the aircraft, they computed Mode-S data indirectly from the flight characteristics of aircraft. Wind, for example, is computed from the difference between airspeed and groundspeed. They reported that when such indirectly derived Mode-S data are used as inputs to the NWP model, in addition to the usual input, the difference between the computed predicted winds and the actual winds is 5% smaller than when the NWP model prediction is computed without Mode-S data. One can assume that by using actual Mode-S weather data that is sent from the aircraft instead of indirectly derived Mode-S data, the predictions would be even more accurate.