Errors in Wind Fields Derived from Multiple-Doppler Radars: Random Errors and Temporal Errors Associated with Advection and Evolution

T. L. Clark National Center for Atmospheric Research, Boulder, CO

Search for other papers by T. L. Clark in
Current site
Google Scholar
PubMed
Close
,
F. I. Harris National Center for Atmospheric Research, Boulder, CO

Search for other papers by F. I. Harris in
Current site
Google Scholar
PubMed
Close
, and
C. G. Mohr National Center for Atmospheric Research, Boulder, CO

Search for other papers by C. G. Mohr in
Current site
Google Scholar
PubMed
Close
Restricted access

Abstract

Simulations from a time-dependent model of moist convection have been used to assess magnitudes of errors in the estimates of derived wind fields from the synthesis of data from a network of Doppler radars. The two types of errors considered are, first, those due to temporal changes in the scales of deep convection resolved by the model, and second, those due to the random contributions of radial velocity estimates by scales smaller than model resolution (noise). Due to the coarse spatial resolution of the model, much of the assumed noise error is due to spatial scales between the model's resolution (∼1 km) and the Doppler radar sampling scale (∼100 m) and should not be considered in reality as “white” noise with respect to the radar sampling problem. The results presented in this paper must be interpreted only in terms of wind estimates derived by using radar sample volumes comparable to the models resolution. Much higher spatial resolution experiments with the model are necessary to clearly delineate the differences between temporal and noise errors for scales larger than the typical radar sampling volumes.

The temporal errors for the resolved scales of the model using a 3 min scan time were found to be less than those due to noise and in general quite tolerable in magnitude for three or more radars. A dual-Doppler analysis in x, y, z Cartesian space (as opposed to x, y, elevation angle coplane analysis) was considered. In this case the derived errors (in the steady state) were found to be significantly large.

The effects of scan time and number of radars were assessed and two methods of reducing temporal errors were investigated.

Abstract

Simulations from a time-dependent model of moist convection have been used to assess magnitudes of errors in the estimates of derived wind fields from the synthesis of data from a network of Doppler radars. The two types of errors considered are, first, those due to temporal changes in the scales of deep convection resolved by the model, and second, those due to the random contributions of radial velocity estimates by scales smaller than model resolution (noise). Due to the coarse spatial resolution of the model, much of the assumed noise error is due to spatial scales between the model's resolution (∼1 km) and the Doppler radar sampling scale (∼100 m) and should not be considered in reality as “white” noise with respect to the radar sampling problem. The results presented in this paper must be interpreted only in terms of wind estimates derived by using radar sample volumes comparable to the models resolution. Much higher spatial resolution experiments with the model are necessary to clearly delineate the differences between temporal and noise errors for scales larger than the typical radar sampling volumes.

The temporal errors for the resolved scales of the model using a 3 min scan time were found to be less than those due to noise and in general quite tolerable in magnitude for three or more radars. A dual-Doppler analysis in x, y, z Cartesian space (as opposed to x, y, elevation angle coplane analysis) was considered. In this case the derived errors (in the steady state) were found to be significantly large.

The effects of scan time and number of radars were assessed and two methods of reducing temporal errors were investigated.

Save