Abstract
The hypothesis that temporal averages of vertical motions over a single radar station are representative of weather systems large enough to be resolved by the radiosonde network is tested using data from the Flatland VHF radar, located in the very flat terrain of central Illinois. Six-hourly means of radar data were compared with four separate estimates of the synoptic or subsynoptic-scale vertical motions computed using the dynamical equations with unsmoothed rawinsonde data and with NMC gridded analyses. Spring and fall cases of large upward and downward vertical motions were selected for study. During the course of this study it was found that contamination of the Doppler radar spectra by heavy or moderate precipitation must be taken into account during analyses of VHF radar data in the troposphere.
The signs of the vertical-motion estimates from the indirect schemes in the extreme cases selected for study here nearly always agree, although the magnitudes often differ by a factor up to about 4. The adiabatic method was found to be unrepresentative due to the large time separation of radiosonde measurements. The 6-b average radar observations usually fall within the envelope of estimates from the various indirect methods. The major source of statistical uncertainty of the temporal means of the vertical motions seen by the radar is the mesoscale structure seen in shorter-period averages and not completely filtered out during averaging. Such structure is not resolved by the radiosonde network data and analyses.
