| description abstract | The existence of undetected errors in recorded wind observations may have a biasing influence on a statistical study. In the progress of some studies it has been found necessary to reexamine the data being used. A series of upper-air winds has been checked by using available listings of vertical shear and extreme winds. The developed procedure permits correction for major errors and tolerates the minor (random) errors. The test of data by maximum wind profiles uses the highest and second highest scalar wind speed for each station and checks the data by profile scan. The test of data by vertical wind shear uses a critical value, theoretically derived, exceedance of which marks the data as suspicious. A detailed check of the wind observation verifies this suspicious value or it is corrected. In this program 3.5 percent of the observations1 proved suspicious and 85 percent thereof, that is, 2.9 percent of the observations, required correction. Thus the critical value is highly efficient. The errors were traced and split into clerical errors (1.1 percent), instrumental errors (1.3 percent), and computational errors (0.5 percent), which are quite within reasonable limits. | |
