QUANTIFICATION OF PREDICTIVE SKILL FOR MESOSCALE AND SYNOPTIC-SCALE METEOROLOGICAL FEATURES AS A FUNCTION OF HORIZONTAL GRID RESOLUTION

To quantitatively assess numerical predictive skill for synoptic and m esoscale features as a function of horizontal grid resolution, a serie s of experiments is conducted using the Pennsylvania State University- National Centre for Atmospheric Research Mesoscale Model. For eight ca ses of continental cyclogenesis, 72-h integrations are examined using grids of 160, 80, and 26.7 km. First, we briefly examine error statist ics for synoptic-scale cyclones and anticyclones. Next, a detailed ana lysis of model errors for mesoscale features is presented. A bandpass filtering technique, based on the Barnes objective analysis scheme, is used to separate perturbation quantities associated with the mesoscal e features from the synoptic-scale fields. Error statistics are then c ompiled for various mesoscale features, including the intensity of mes olows, damming ridges, and post frontal troughs, and the thermal gradi ents, propagation speed, and vertical velocity maxima associated with surface cold fronts. Finally, the accuracy of the predicted precipitat ion fields, produced using the Anthes-Kuo cumulus parameterization, is examined. Objective verification reveals that forecast skill does nor improve uniformly for all types of mesoscale features as horizontal g rid resolution is increased, although the general trend is for reduced errors as expected. Improvements do occur on both the 80- and 27-km g rids for all geographically related mesoscale features (such as orogra phic lee troughs). A similar improvement is seen for propagating mesos cale features(such as postfrontal troughs) and synoptic-scale cyclones as the grid length is reduced from 160 to 80km. However,when the grid length is further reduced to 27 km, mean absolute errors and mean pos ition errors actually increase for both classes of features. This grea ter variability in model performance suggests that as grid resolution is enhanced, other factors such as the accuracy of model physics and i nitial conditions become increasingly important. The effect on precipi tation bias and threat scores in these experiments is positive (reduce d errors) when resolution is improved from 160 to 80 km but is general ly insignificant or negative for the 27-km grid. Based on these result s, the Anthes-Kuo convective parameterization used in these experiment s is not recommended for application on grids of about 30 km or less.