SPECIFICATION OF 500-MB. PARAMETERS BY DOWNWARD EXTRAPOLATION

MAURICE E. GRAVES Department of Aerospace Engineering, The University of Michigan, Ann Arbor, Mich.

Search for other papers by MAURICE E. GRAVES in
Current site
Google Scholar
PubMed
Close
and
EDWARD S. EPSTEIN Department of Meteorology and Oceanography, The University of Michigan, Ann Arbor, Mich.

Search for other papers by EDWARD S. EPSTEIN in
Current site
Google Scholar
PubMed
Close
Restricted access

Abstract

With the future utilization of new-type upper tropospheric observations in mind, the estimation of 500-mb. geopotential height from 300-mb. data is accomplished by least squares regression. The regression coefficients form latitudinal patterns which can be expressed by linear relationships in low latitudes and parabolic relationships else-where. From three year's mid-seasonal-month grid data, measures of extrapolation error are obtained over half of the Northern Hemisphere. Verifying tests with radiosonde station data indicate that the error in low latitudes is substantially due to analysis noise in the 500-mb. grid data.

When the same techniques are applied to 200-mb. information, further error studies show considerably less feasibility of extrapolation from that level to 500 mb. However, the temperature at 200 mb. is found to be valuable in predicting the simultaneous temperature at 500 mb.

Abstract

With the future utilization of new-type upper tropospheric observations in mind, the estimation of 500-mb. geopotential height from 300-mb. data is accomplished by least squares regression. The regression coefficients form latitudinal patterns which can be expressed by linear relationships in low latitudes and parabolic relationships else-where. From three year's mid-seasonal-month grid data, measures of extrapolation error are obtained over half of the Northern Hemisphere. Verifying tests with radiosonde station data indicate that the error in low latitudes is substantially due to analysis noise in the 500-mb. grid data.

When the same techniques are applied to 200-mb. information, further error studies show considerably less feasibility of extrapolation from that level to 500 mb. However, the temperature at 200 mb. is found to be valuable in predicting the simultaneous temperature at 500 mb.

Save