Experiments Concerning Variability among Subjective Analyses

Abstract

The variability of subjective (hand) analyses is explored by examining the results of two synoptic laboratory experiments. Two groups of analysts participated in the first experiment, one consisting of 13 senior meteorology students (1977), the other consisting of six professional synopticians. Each participant analyzed maps of geopotential height and zonal and meridional wind components over the eastern two-thirds of the United States for 1200 GMT 25 June 1968 at 500 mb. Grid-point data were extracted from these analyses and are compared to those obtained from a prior subjective analysis and an objective analysis, both used as standards. Higher order quantities, consisting of geostrophic and ageostrophic wind components, horizontal divergence, relative vorticity, kinetic energy content and generation of kinetic energy, were computed from the grid-point data and also are compared. Methods of comparison include pattern analyses, difference maps and statistical tests. In the second experiment, only one group of analysts, 11 senior meteorology students (1979), participated. The same data were used. Each participant analyzed zonal and meridional wind components as for the first experiment. In addition, individuals analyzed maps for isogons and isotachs from which wind components subsequently were derived. Comparisons. similar to those made in Experiment 1, are made between analyzed and derived components, as well as between kinetic energy, divergence and vorticity values computed from each set of components. The significant findings are as follows: 1) in each experiment, analyses of wind components, regardless of the analysis scheme used (subjective, objective or components derived from wind direction and speed), are in very good agreement; 2) in each experiment, analyses are combined to form composite mean maps which are found to give the best representation of the flow features compared to any of the individual analyses; 3) in Experiment 1, the height and both wind components of the subjective and objective standards differ more from each other than they do from the corresponding composites of subjectively analyzed maps; 4) in Experiment 1. there is reasonably good agreement among analysis techniques for derived quantities, except those that depend on cross-contour flow; 5) average cross-contour flow angles for student composite, professional composite, overall composite, subjective standard and objective standard are 32, 24, 27, 31 and 15°, indicating the tendency of the objective scheme to minimize this variable; and 6) in Experiment 2, as in Experiment 1, kinetic energy and vorticity show good agreement, regardless of analysis scheme, but considerably less agreement is seen for divergence.