Comparison of MM5 integrated water vapor with microwave radiometer, GPS, and radiosonde measurements

Abstract

A large dataset of concurrent integrated precipitable water vapor (IPWV) estimates from ground-based microwave radiometers (MWRs), global positioning system (GPS) ground-receivers, and radiosonde observations (RAOBs) has been collected in five different sites in Central Italy. Both MWRs and GPS have shown a capability of accurate and continuous water vapor monitoring. These data are used to study the seasonal and spatial variability of IPWV. A comparison of these data with the IPWV field produced operationally by the nonhydrostatic Mesoscale Model (MM5), running at the University of L'Aquila/Center of Excellence (CETEMPS) is performed in order to find either model shortcomings and to corroborate the IPWV behavior highlighted by the measurements. Both measurements and model outputs span over a period of about one year allowing for a systematic statistical analysis for all the examined stations. The statistical analysis shows a good agreement between GPS and MWR data, whereas discrepancies are found between RAOBs and the other techniques. The IPWV shows the largest diurnal variability, approximately 3%, during the summer season. An overall good agreement is found between the forecasted and observed IPWV. The related statistical parameters show a very low bias (0.001 cm) with a good correlation coefficient (0.939). On the other hands, the seasonal analyses highlight a few discrepancies, mostly due to the MM5 difficulties in correctly forecasting the diurnal cycle.