Atmospheric water vapor is a key element in the global radiation budget, because of its efficiency as green-house gas. In addition, it is involved in the microphysical processes leading to clouds formation and development. On local scale, water vapor has a principal role because of its influence in the local weather especially for what concerns precipitations, with direct effects on the human activities (e.g. agriculture and tourism) and on human life and security (severe events). Therefore it is important to represent this parameter in models for short term weather forecasts. Moreover, water vapor distribution is characterized by a large spatial and temporal variability, considerably different in troposphere and stratosphere, strongly influenced by both large-scale circulation and localized convection. This high variability makes necessary a long-term comparison between accurate and high resolution water vapor observations and operational forecast models. The EU CloudNET project offers an extended database of water vapor profiles provided by four operational forecast models of ECMWF, MetOffice, MeteoFrance and KNMI. On the other hand, at CNR-IMAA accurate vertical profiles of water vapor mixing ratio are provided with very high resolution in a systematic way since May 2002.
At CNR-IMAA (40°36'N, 15°44'E, 760 m above sea level), a Raman lidar system for atmospheric water vapor mixing ratio measurements is operative since May 2002. The CNR-IMAA Raman lidar system is able to provide water vapor mixing ratio vertical profiles up to the tropopause in night-time and up to 5 km of altitude in daytime conditions, with a typical temporal resolution of 10 minutes [Cornacchia et al., 2004]. It is part of the Envisat's ACVT (Atmospheric and Chemistry Validation Team) and is operative within NDSC (Network for the Detection of Stratospheric Change) [Cuomo et al., 2004; D'Aulerio et al., 2004]. Even if, in principle, Raman lidar measurements of water vapor can be calibrated in absolute terms, the absolute calibration accuracy is limited to about 10% by the uncertainties in the ratio of Raman lidar cross sections of water vapor and nitrogen. A better accuracy is typically obtained calibrating the lidar water vapor mixing ratio profile with independent water vapor measurements. The CNR-IMAA Raman lidar for the water vapor measurements has been calibrated by means of an intensive measurement campaign performed in May-June 2002, by means of contemporary and co-located Vaisala RS80-A radiosondes measurements. Since July 2002, about 200 co-located radiosonde (RS80-A, RS90 and RS92) launches have been used to check the stability of the lidar calibration constant, that remains within 5%. Since February 2004, the calibration of the Raman lidar is continuously checked by comparing integrated precipitable water vapor (IPWV) content obtained with the lidar itself to the IPWV obtained with a multichannel microwave profiler operational at CNR-IMAA [Madonna et al., 2005]. This allows to overall all the problems related to radiosounding water vapor measurements, like dry bias, different investigated air volumes, and high costs. Since July 2002, measurements have been collected for about 270 days, during systematic measurements performed twice a week and special measurement campaigns (e.g. ICARTT field campaign in summer 2004 and LAUNCH-2005 campaign in autumn 2005). Besides the water vapor Raman lidar and the multichannel microwave profiler for T and RH profiles and water vapor and liquid water content measurements, a ceilometer for clouds top and base detection, a multi-wavelength Raman lidar and a sunphotometer for aerosol study are operational at CNR-IMAA. The large variety of instrumentation located at CNR-IMAA and the large quantity of data acquired in a systematic way makes the CNR-IMAA one of the best candidates to perform a long-term check of the four operational forecasts models.
In the database provided within CloudNET by the four operation models (ECMWF, MetOffice, MeteoFrance and KNMI), the atmosphere is split up into a series of grid boxes, typically with horizontal size of 50 km and vertical depth between 0.5 and 1 km. For a correct comparison with observational data, lidar high resolution profiles are reduced into large grid boxes on the base of wind speed in order to take into account the advection time. In this, only high resolution lidar data with a total error less than 50% are considered. For each box, water vapor mixing ratio mean value and standard deviation and mean error are calculated. In addition, the number of points considered within each box and the pdf of values contained in it are calculated and stored for a quantitative comparison with model data. On 1-3 October 2005, a long record of measurements of about 30 hours has been collected at CNR-IMAA. This case has been selected as a first case for the comparison with the models. Even if the models do not capture details in the evolution of water vapor fields, a good agreement is found in term of vertical structure and water vapor content. First results on long term comparison between the 4 models and CNR-IMAA lidar data will be presented, focusing on the capability of the models to capture mean aspects of the water vapor field as well as on the possible discrepancies between observations and models.
Cornacchia C., et al., The IMAA Raman lidar system for water vapor measurements, Reviewed and Revised Papers presented at the 22nd International Laser Radar Conference, ESA SP-561, 107-110, 2004.
D'Aulerio P., et al., Intercomparison of water vapor retrieval between three Raman lidar stations, Reviewed and Revised Papers presented at the 22nd International Laser Radar Conference, ESA SP-561, 923-926, 2004.
Cuomo V., et al., Envisat validation campaign at IMAA-CNR, Proceedings of the Second Validation Workshop on the Atmospheric Chemistry Validation of ENVISAT (ACVE-2) Frascati, Italy, May 2004, vol. SP-532, ESA.
Madonna F., et al., Multichannel microwave radiometer and water vapour Raman lidar: comparisons and synergies , AITinforma - Rivista Italiana di telerilevamento, 35, 115-130, 2006.