Search Results

You are looking at 1 - 7 of 7 items for

  • Author or Editor: L. M. Mcmillin x
  • All content x
Clear All Modify Search
L. M. McMillin and C. Dean

Abstract

To produce atmospheric temperature profiles from measurements of infrared data, it is necessary to obtain infrared radiances for clear areas. Clear radiances are obtained either by identifying spots that are completely clear or by extracting clear values from areas that are partly overcast. Until 10 June 1980, clear radiances were obtained using an algorithm described by Smith and Woolf (1976). At this time, the algorithm was replaced by a new algorithm using current techniques. The new algorithm produced more accurate results which resulted in improved consistency between retrievals from the clear and partly cloudy areas. In addition, the new algorithm produced many partly cloudy retrievals in areas where the former method had produced less accurate retrievals using only microwave channels.

During the study, it was discovered that the crucial assumption of a single layer or even a two- or three-layer cloud is seldom satisfied. The new algorithm uses various tests to identify pairs of spots with the same cloud heights. These tests are an important feature of the new algorithm. The clear radiance procedure is described and the results of the evaluation are discussed.

Full access
M. J. Uddstrom and L. M. McMillin

Abstract

To utilize satellite radiance sounding data, in either explicit or implicit retrieval algorithms, a proper understanding of the noise in the measurements is required. Conventionally, to define the expected accuracy of atmospheric profiles inferred from sounder data, instrument noise equivalent temperature difference (NEΔT) noise specifications have been used to simulate spacecraft data. Here it is demonstrated that NEΔT noise specifications are inappropriate for this purpose. Instead, total system noise estimates should be employed since use of sounding data in any type of physical retrieval algorithm implies application of a radiative transfer model, which in turn must be “calibrated” against in situ and satellite data.

It is demonstrated that the accuracy of atmospheric retrievals inferred from the TIROS Operational Vertical Sounder radiometers is limited by the total system noise rather than NEΔT noise, and that modeled radiance temperatures perturbed by Gaussian total system noise very nearly replicate the accuracy statistics of retrievals computed from satellite measurements. The implications of these results for planned high-resolution infrared sounding instruments are briefly discussed.

Full access
M. J. Uddstrom and L. M. McMillin

Abstract

The National Environmental Satellite Data and Information Service (NESDIS) collocation data archive of satellite and radiosonde measurements is used to investigate errors in in situ radiosonde data and a NESDIS-like radiative transfer forward model. It is shown that the radiative transfer model errors have a strong airmass dependence and that these errors are not primarily due to nonrepresentativeness or radiosonde errors. However, errors in the in situ data do exist. For example, the National Meteorological Center radiosonde radiation adjustment algorithm in use during the period of data collection (1989–90) does not appear to provide adjustments of uniform quality across radiosonde sounding systems. The total system noise appropriate for use in retrieval algorithms is shown to vary from values close to the radiometer noise equivalent temperature difference (NEΔT) specifications for stratospheric channels to several times the NEΔT values for lower-tropospheric channels. Because of the significant discrepancies between measured and modeled radiances for the two most opaque water vapor sounding channels of the TIROS Operational Vertical Sounder, their use in a physical retrieval algorithm is considered problematic. Evidence for errors in NESDIS cloud-cleared (equivalent clear column) radiance temperature estimates is presented.

Full access
L. J. Crone, L. M. Mcmillin, and D. S. Crosby

Abstract

Least squares or regression techniques have been used for many problems in satellite meteorology. Because of the large number of variables and the linear dependence among these variables, colinearity causes significant problems in the application of standard regression techniques. In some of the applications there is prior knowledge about the values of the regression parameters. Since there are errors in the predictor variables as well as the predictand variables, the standard assumptions for ordinary least squares are not valid. In this paper the authors examine several techniques that have been developed to ameliorate the effects of colinearity or to make use of prior information. These include ridge regression, shrinkage estimators, rotated regression, and orthogonal regression. In order to illustrate the techniques and their properties, the authors apply them to two simple examples. These techniques are then applied to a real problem in satellite meteorology: that of estimating theoretical computed brightness temperatures from measured brightness temperatures. It is found that the rotated and the shrinkage estimators make good use of the prior information and help solve the colinearity problem. Ordinary least squares, ridge regression, and orthogonal regression give unsatisfactory results. Theoretical results for the various techniques are given in an appendix.

Full access
L. M. McMillin, D. G. Gray, H. F. Drahos, M. W. Chalfant, and C. S. Novak

Abstract

Root-mean-square differences between satellite and radiosondes for the past three years that TIROS-N has been operational are examined. They show a pronounced annual cycle because the statistics are dominated by the Northern Hemisphere. Differences are smaller in the summer and are larger in the winter, but they reflect a change in the effect of location differences as well as retrieval error. In addition to the annual cycle, there is an increase in retrieval accuracy with time. For the partly cloudy retrievals, the increase approaches 1.3 K for some levels.

Full access
J. A. Schroeder, E. R. Westwater, P. T. May, and L. M. McMillin

Abstract

Temperature profiles from the TIROS-N Operational Vertical Sounder (TOVS) were combined with low-altitude virtual temperature profiles measured by a ground-based 915-MHz/2-kHz radio acoustic sounding system (RASS) at Denver, Colorado. Low-level temperature inversions of more than 20°C in 100 m, as measured by radiosonde, were successfully resolved by RASS, providing critical information that TOVS alone could not. Conversely, TOVS augmented the RASS altitude coverage, which was typically limited to the first 1–2 km above ground, due to strong acoustic attenuation at 2 kHz. (A 405-MHz/0.9-kHz RASS measures to significantly higher altitudes.) The complementary nature of TOVS and RASS is demonstrated with examples chosen from extreme cases that illustrate the capabilities and limitations of the separate and combined systems.

Full access
E. R. Westwater, W. B. Sweezy, L. M. McMillin, and Charles Dean

Abstract

Radiometric soundings from the Wave Propagation Laboratory's ground-based Profiler, the NOAA 6/7 satellites, and the combination of the two, were compared in their ability to derive temperature and moisture profiles. Radiosonde data for the period December 1981-December 1982, taken by the National Weather Service at Stapleton International Airport, Denver, Colorado, were used as “ground-truth” for the comparison; in all, 460 soundings were analyzed. The set of soundings contained 216 clear, 173 partly cloudy and 71 cloudy cases. Comparisons show that Profiler retrievals were more accurate than those of the satellite in the lowest 500 mb of the atmosphere, with the converse being true above that level. The combined temperature retrievals were more accurate, in the rms sense, than either of the separate retrievals at every level from the surface to 10 mb. Below 50 mb, the maximum rms difference of the combined system from radiosondes was 2.7 K; below 300 mb, it was 2.0 K. Geopotential heights and pressure thicknesses were also derived from the combined system with an accuracy approaching that of a radiosonde.

Full access