Search Results

You are looking at 1 - 10 of 36 items for

  • Author or Editor: Harry R. Glahn x
  • Refine by Access: All Content x
Clear All Modify Search
HARRY R. GLAHN

Abstract

The regression screening and principal component techniques for developing forecast aids are investigated for their applicability to the objective forecasting of rainfall probabilities. The forecasting of summer rainfall in the Mississippi Delta is the particular problem studied. Subjective forecasts made for the area as well as objective forecasts are verified in terms of reductions of variance and saving over climatology. It is found that many of the forecast equations developed by regression screening and principal component techniques are not stable on test data. The results indicate that subjective screening of predictors is desirable before the regression screening is accomplished. It is found that useful aids can be developed with these linear techniques; at the same time the desirability of an approach that better integrates the physical processes of the atmosphere is indicated.

Full access
Harry R. Glahn

Abstract

There is a popular misconception that the secant form of the Lambert conformal map projection is “better” than the tangent form. It is shown here that the two forms are equivalent; they are different only in the sense that the scale of the map quoted is usually true at the two secant latitudes for the secant projections and at the single tangent latitude for the tangent projection.

Full access
Harry R. Glahn

Abstract

No abstract available.

Full access
Harry R. Glahn

Abstract

No abstract available.

Full access
Harry R. Glahn

Abstract

No abstract available.

Full access
Harry R. Glahn

Abstract

No abstract available.

Full access
Harry R. Glahn

Abstract

Truncation errors of four stencils for evaluating the two-dimensional Laplacian operator are discussed, and response functions are computed. The efficiencies of the stencils, in terms of number of iterations and computer time required, for solving Poisson's equation by relaxation is investigated. The dependence on the relaxation coefficient of the number of iterations required and the error of the solution is shown for a number of examples.

Full access
Harry R. Glahn

Abstract

No abstract available.

Full access
Harry R. Glahn

Abstract

The assumptions inherent in the determination of cloud top heights and areal coverage of clouds from infrared measurements made by satellites are discussed. The problems of interpretation caused by radiometer spatial resolution being of the same order of size as individual cloud elements are studied through mathematical simulation of the viewing process. An analysis of the simulated measurements from simple, specified cloud conditions produces quantitative estimates of the errors of interpretation.

It is found that individual cloud elements of a diameter slightly less than the viewed spot can be very misleading and the height of cloud tops might be judged to be several thousand feet below their true height; tops of larger clouds can be determined more accurately. If the actual height of the tops can be determined, either by the infrared measurements from large cloud masses or by other means, the areal coverage can be estimated rather well.

Full access
Harry R. Glahn

Abstract

Complete automation of public weather forecasts will exist when forecasts are prepared entirely by computer and are disseminated to the final user through associated communications facilities. Preparation of such a forecast must start with observations of various kinds and proceed through the collection and analysis of those observations, the running of numerical and statistical models to arrive at digital forecasts of the pertinent weather elements, and the preparation of the forecast in its final format.

Observing the atmospheric variables used as input to forecasting models is far from automated, although much has been done in this area in recent years. The assembly and analysis of the observations have been largely automated at national centers, although work must continue on incorporating new types of observations and making more efficient use of those already available. Numerical models are being run which do a good job of predicting the future state of certain atmospheric variables. Statistical models are in operation which translate these “atmospheric” forecasts into “weather” forecasts. Automatic preparation of the forecasts in final form has received much less attention than have other aspects of automation, but experimental computer programs do exist for this purpose. And finery, the AFOS (Automation of Field Operations and Services) program of the National Weather Service provides a feasible means for monitoring and communicating the forecasts, if not to the final user—the public—at least to the media which do make the forecasts available to the final user.

This paper describes the automated public weather forecasts issued by the National Weather Service—such as probability, amount and type of precipitation, maximum and minimum temperature, clouds, winds and severe weather—and the progress made on formatting these into a worded message.

Full access