Verification of Sounding Parameters from the RAP and the RAP-Based and RUC-Based SFCOA Products

This study compared thermodynamic variables from the rapid refresh (RAP) and three different surface objective analysis (SFCOA) models, two based on the RAP, and another based on the rapid update cycle (RUC), to one another and observed soundings from the spring of 2012 to the summer of 2013 from 83 stations around the United States. Surface to 3-km convective available potential energy (CAPE), mixing ratio, precipitable water, surface temperature, and surface dewpoint temperature were examined, in addition to three different measures of CAPE. Statistics were calculated for averages over the entire data set, and statistics for mean difference between forecast and observed values were calculated on a seasonal basis. The tendency for the RAP to underestimate CAPE was evident in both the yearly and seasonal analysis, but the negative bias was more noticeable during the spring and summer months. A tendency for root mean square error to be larger for variables in the spring and summer months and an increased spread between the 10th and 95th percentiles during June, July, and August were observed for the RAP and RAP-based SFCOA products. A tendency for the dewpoint mean difference to be more negative than that of the surface temperature was also observed. This suggests that the RUC-based SFCOA products are likely more accurate when estimating thermodynamic variables in drier environments than the RAP-based SFCOA products.