High-resolution Numerical Weather Prediction Research Articles

Imaging gravity waves in lower stratospheric AMSU-A radiances, Part 2: Validation case study

Abstract. Two-dimensional radiance maps from Channel 9 (~60–90 hPa) of the Advanced Microwave Sounding Unit (AMSU-A), acquired over southern Scandinavia on 14 January 2003, show plane-wave-like oscillations with a wavelength λh of ~400–500 km and peak brightness temperature amplitudes of up to 0.9 K. The wave-like pattern is observed in AMSU-A radiances from 8 overpasses of this region by 4 different satellites, revealing a growth in the disturbance amplitude from 00:00 UTC to 12:00 UTC and a change in its horizontal structure between 12:00 UTC and 20:00 UTC. Forecast and hindcast runs for 14 January 2003 using high-resolution global and regional numerical weather prediction (NWP) models generate a lower stratospheric mountain wave over southern Scandinavia with peak 90 hPa temperature amplitudes of ~5–7 K at 12:00 UTC and a similar horizontal wavelength, packet width, phase structure and time evolution to the disturbance observed in AMSU-A radiances. The wave's vertical wavelength is ~12 km. These NWP fields are validated against radiosonde wind and temperature profiles and airborne lidar profiles of temperature and aerosol backscatter ratios acquired from the NASA DC-8 during the second SAGE III Ozone Loss and Validation Experiment (SOLVE II). Both the amplitude and phase of the stratospheric mountain wave in the various NWP fields agree well with localized perturbation features in these suborbital measurements. In particular, we show that this wave formed the type II polar stratospheric clouds measured by the DC-8 lidar. To compare directly with the AMSU-A data, we convert these validated NWP temperature fields into swath-scanned brightness temperatures using three-dimensional Channel 9 weighting functions and the actual AMSU-A scan patterns from each of the 8 overpasses of this region. These NWP-based brightness temperatures contain two-dimensional oscillations due to this resolved stratospheric mountain wave that have an amplitude, wavelength, horizontal structure and time evolution that closely match those observed in the AMSU-A data. These comparisons not only verify gravity wave detection and horizontal imaging capabilities for AMSU-A Channel 9, but provide an absolute validation of the anticipated radiance signals for a given three-dimensional gravity wave, based on the modeling of Eckermann and Wu (2006).

A radar-based verification of precipitation forecast for local convective storms

Local flash flood storms with a rapid hydrological response are a real challenge for quantitative precipitation forecasting (QPF). It is relevant to assess space domains, to which the QPF approaches are applicable. In this paper an attempt is made to evaluate the forecasting capability of a high-resolution numerical weather prediction (NWP) model by means of area-related QPF verification. The results presented concern two local convective events, which occurred in the Czech Republic (CR) on 13 and 15 July 2002 and caused local flash floods. We used the LM COSMO model (Lokall Model of the COSMO consortium) adapted to the horizontal resolution of 2.8 km over a model domain covering the CR. The 18 h forecast of convective precipitation was verified by using radar rainfall totals adjusted to the measured rain gauge data. The grid point-related root mean square error (RMSE) value was calculated over a square around the grid point under the assumption that rainfall values were randomly distributed within the square. The forecast accuracy was characterized by the mean RMSE over the whole verification domain. We attempt to show a dependence of both the RMSE field and the mean RMSE on the square size. The importance of a suitable merger between the radar and rain gauge datasets is demonstrated by a comparison between the verification results obtained with and without the gauge adjustment. The application of verification procedure demonstrates uncertainties in the precipitation forecasts. The model was integrated with initial conditions shifted by 0.5° distances. The four verifications, corresponding to the shifts in the four directions, show differences in the resulting QPF, which depend on the size of verification area and on the direction of the shift.

A Radar Simulator for High-Resolution Nonhydrostatic Models

Abstract A full radar simulator for high-resolution (1–5 km) nonhydrostatic models has been developed within the research nonhydrostatic mesoscale atmospheric (Meso-NH) model. This simulator is made up of building blocks, each of which describes a particular physical process (scattering, beam bending, etc.). For each of these blocks, several formulations have been implemented. For instance, the radar simulator offers the possibility to choose among Rayleigh, Rayleigh–Gans, Mie, or T-matrix scattering methods, and beam bending can be derived from an effective earth radius or can depend on the vertical gradient of the refractive index of air. Moreover, the radar simulator is fully consistent with the microphysical parameterizations used by the atmospheric numerical model. Sensitivity experiments were carried out using different configurations for the simulator. They permitted the specification of an observation operator for assimilation of radar reflectivities by high-resolution nonhydrostatic numerical weather prediction systems, as well as for their validation. A study of the flash flood of 8–9 September 2002 in southeastern France, which was well documented with volumetric data from an S-band radar, serves to illustrate the capabilities of the radar simulator as a validation tool for a mesoscale model.

Using Radar Wind Observations to Improve Mesoscale Numerical Weather Prediction

Abstract A high-resolution radar data assimilation system is presented for high-resolution numerical weather prediction models. The system is under development at the Naval Research Laboratory for the Navy’s Coupled Ocean–Atmosphere Mesoscale Prediction System. A variational approach is used to retrieve three-dimensional dynamical fields of atmospheric conditions from multiple-Doppler radar observations of radial velocity within a limited area. The methodology is described along with a preliminary evaluation of the impact of assimilated radar data on model forecasts using a case study of a squall line that occurred along the east coast of the United States on 9 May 2003. Results from the experiments show a significant impact from the assimilated radar radial velocity data on the model forecast of not just dynamical but also hydrological fields at all model levels for the duration of the storm. A verification system has also been developed to assess the radar data assimilation impact, and the results show improvements in the three-dimensional wind forecasts but relatively small changes in the prediction of storm locations. This study highlights the need to develop a continuous radar data assimilation system to maximize the impact of the data.

Partitioning of snowy and rainy precipitation in a case of a north Adriatic frontal passage

Abstract. A case of snow fall in the plains of the Northern Italian region Veneto is presented from a forecasters' perspective. Contrasting forecast guidance came from the ECMWF global model and the limited area model LAMI. The former showed a marked warm-moist Sirocco flow coming from the Adriatic Sea onto the coast at all levels, the latter discerned a distinct cold air flow from the north-east along the foothills of the Alps. The integrated observing network of the Centro Meteorologico di Teolo ARPA Veneto revealed this cold-air structure and helped the forecaster in the choice of the forecast and underpin the snowfall alert to the road authorities. It is argued that this feature is a crucial element for the occurrence of snowfall over the Veneto plains, and that the high-resolution numerical weather prediction model was essential in describing this mesoscale feature. The nature of the north-easterly flow is thought to be a combination of a Bora like flow and a barrier jet induced by flow blocking by the Alps.

On the Vertical Structure of Modeled and Observed Deep Convective Storms: Insights for Precipitation Retrieval and Microphysical Parameterization

Abstract An understanding of the vertical structure of clouds is important for remote sensing of precipitation from space and for the parameterization of cloud microphysics in numerical weather prediction (NWP) models. The representation of cloud hydrometeor profiles in high-resolution NWP models has direct applications in inversion-type precipitation retrieval algorithms [e.g., the Goddard profiling (GPROF) algorithm used for retrieval of precipitation from passive microwave sensors] and in quantitative precipitation forecasting. This study seeks to understand how the vertical structure of hydrometeors (liquid and frozen water droplets in a cloud) produced by high-resolution NWP models with explicit microphysics, henceforth referred to as cloud-resolving models (CRMs), compares to observations. Although direct observations of 3D hydrometeor fields are not available, comparisons of modeled and observed radar echoes can provide some insight into the vertical structure of hydrometeors and, in turn, into the ability of CRMs to produce precipitation structures that are consistent with observations. Significant differences are revealed between the vertical structure of observed and modeled clouds of a severe midlatitude storm over Texas for which the surface precipitation is reasonably well captured. Possible reasons for this discrepancy are presented, and the need for future research is highlighted.

MEETING SUMMARIES

WHAT: Scientists from Korea, Japan, and the United States discuss recent developments in the parameterizations of physical processes in nextgeneration, high-resolution numerical weather prediction models WHEN: 17–18 May 2004 WHERE: Yonsei University, Seoul, Korea he Second International Workshop on NextGeneration Numerical Weather Prediction (NWP) Models1 met to discuss the impact of recent developments in modeling for next-generation, high-resolution NWP models, and to exchange ideas for improving the prediction of high-impact weather. In 1999, the Laboratory for Atmospheric Modeling Research (LAMOR) of Yonsei University (YSU) embarked on a national project developing a next-generation NWP model focusing on the parameterization of physical processes in high-resolution models (see information online at http://lamor.yonsei.ac.kr). The ultimate goal of the project is in line with that of the Weather Research and Forecast (WRF) model initiative (see information online at http://wrf-model. org), both of which are to develop a state-of-the-art mesoscale model that is suitable for grid spacing in the range of 1–30 km. The director of LAMOR, Professor Tae-Young Lee, told participants that the focus of this workshop was to discuss the progress in our understanding of the physical parameterizations and the outcome of collaboration between the LAMOR and U.S. groups for the past three years. Major outcomes of collaborations include the Yonsei University PBL (YSUPBL) and WRF single-moment microphysics (WSM) schemes implemented in the WRF model.

Urban meteorological modelling for nuclear emergency preparedness

The main objectives of the current EU project “Integrated Systems for Forecasting Urban Meteorology, Air Pollution and Population Exposure” (FUMAPEX) are the improvement of meteorological forecasts for urban areas, the connection of numerical weather prediction (NWP) models to urban air pollution and population dose models, the building of improved urban air quality information and forecasting systems, and their application in cities in various European climates. In addition to the forecast of the worst air-pollution episodes in large cities, the potential use of improved weather forecasts for nuclear emergency management in urban areas, in case of hazardous releases from nuclear accidents or terror acts, is considered. Such use of NWP data is tested for the Copenhagen metropolitan area and the Øresund region. The Danish Meteorological Institute (DMI) is running an experimental version of the HIRLAM NWP model over Zealand including the Copenhagen metropolitan area with a horizontal resolution of 1.4 km, thus approaching the city-scale. This involves 1-km resolution physiographic data with implications for the urban surface parameters, e.g. surface fluxes, roughness length and albedo. For the city of Copenhagen, the enhanced high-resolution NWP forecasting will be provided to demonstrate the improved dispersion forecasting capabilities of the Danish nuclear emergency preparedness decision–support system, the Accident Reporting and Guidance Operational System (ARGOS), used by the Danish Emergency Management Agency (DEMA). Recently, ARGOS has been extended with a capability of real-time calculation of regional-scale atmospheric dispersion of radioactive material from accidental releases. This is effectuated through on-line interfacing with the Danish Emergency Response Model of the Atmosphere (DERMA), which is run at DMI. For local-scale modelling of atmospheric dispersion, ARGOS utilises the Local-Scale Model Chain (LSMC), which makes use of high-resolution DMI-HIRLAM NWP model data provided to DEMA by DMI four times a day under operational surveillance and covering Denmark and surroundings. The integration of DERMA in ARGOS is effectuated through automated on-line digital communication and exchange of data. The calculations are carried out in parallel for each NWP model to which DMI has access, thereby providing a mini-ensemble of dispersion forecasts for the emergency management.

総合数値予測システムSYNFOS（シンフォス）のH16年豪雨事例への適用

総合数値予測システムSYNFOS（シンフォス）のH16年豪雨事例への適用

The sting at the end of the tail: Model diagnostics of fine‐scale three‐dimensional structure of the cloud head

AbstractThis is the second in a series of papers on ‘the sting at the end of the tail’—a mesoscale region of potentially damaging surface winds that can occur close to the evaporating tip of a cloud head wrapped around the bent‐back front of an extratropical cyclone. In the first paper the sting phenomenon was identified from a purely observational study of the great storm of October 1987. In the present paper an observationally validated forecast run from a high‐resolution numerical weather prediction model is used to examine the evolving three‐dimensional structure of the sting phenomenon. It is shown that the damaging surface winds in the October 1987 storm were due to a well‐defined mesoscale ‘sting jet’ (SJ), identifiable as a coherent ensemble of trajectories originating in cloudy air at about 650 hPa. There is evidence of multiple mesoscale slantwise circulations, and the SJ appears to form within the descending part of these circulations. Air within the SJ descends to the 900 hPa level over a period of about 4 h, during which time it accelerates from less than 20 to above 45 m s−1 with extremes greater than 50 m s−1. Although the wet‐bulb potential temperature of air in the SJ remains constant during its descent to 900 hPa, evaporation leads to a reduction of up to 5 K or more in dry‐bulb potential temperature in some parts of the jet. The SJ is situated behind the primary cold front, and is distinct from the associated warm‐conveyor‐belt low‐level jet; it is also distinct from the cold‐conveyor‐belt low‐level jet which remains below and behind it as the SJ skirts the bent‐back front. © Crown copyright, 2005.

Simulations of deep convection in the Mediterranean area using 3DVAR of conventional and non-conventional data

Abstract. In autumn deep convection in the Mediterranean region is a common phenomenon. The local events characterized by deep convection are still a difficult task even for high resolution numerical weather prediction. Three flood cases, produced by convection either embedded in a large scale system or locally developed, occurring in Italy, are presented. All these case were not correctly forecasted: Sardinia (Cagliari, 13 November 1999); Calabria (Soverato, 7 September 2000) and Sicily (Catania, 16 September 2003). The first case occurred during the Mesoscale Alpine Programme (MAP) campaign, therefore a lot of data are available; for the second one only data from SSM/I and local rain-gauge are available; the third one occurred during the operational experimentation of the TOUGH project. The last one was not well predicted even using the operational assimilation of ground based GPS. To improve the forecast of these cases the assimilation of several data is tested. The variational assimilation performed using 3DVAR of GPS, SSM/I and surface and upper air data is applied to improve the Initial Conditions of the Sicily case. The Sardinia case is improved using either GPS and surface data, whereas for the Soverato case only ZTD is assimilated. The experiments are performed using the MM5 model from Pennsylvania State University/National Center for Atmospheric Research (PSU/NCAR); the model is initialized using the new Initial Conditions produced by the variational assimilation of conventional and non conventional data. The results show that the assimilation of the retrieved quantities does produces large improvement in the precipitation forecast. Large sensitivity to the assimilation of surface data and brightness temperature from SSM/I is found.

A Numerical Modeling Study of the Weather in East Antarctica and the Surrounding Southern Ocean

Abstract The requirement for more detailed and accurate weather forecasts in Antarctica has steadily grown over the past decade, to the point where more sophisticated high-resolution numerical weather prediction (NWP) systems need to be employed to provide the mesoscale prediction services demanded. The following paper discusses the performance of the Antarctic Limited-Area Prediction System (ALAPS) in providing weather forecasting support at high latitudes. Verification results based on traditional statistical techniques, including the S1 skill score and bias and root-mean-square errors, are presented. They suggest that the high-resolution gridpoint model provides no improvement over the global model used for the nesting fields. However, it is argued that the traditional statistical techniques fail to capture the advantages of the high-resolution gridpoint model, and three case studies are presented to highlight the ALAPS model performance characteristics. The first case study is an example of the model ...

An Objective Technique for Verifying Sea Breezes in High-Resolution Numerical Weather Prediction Models

An ongoing challenge in mesoscale numerical weather prediction (NWP) is to determine the ideal method for verifying the performance of high-resolution, detailed forecasts. Traditional objective techniques that evaluate NWP model performance based on point error statistics may not be positively correlated with the value of forecast information for certain applications of mesoscale NWP, and subjective evaluation techniques are often costly and time consuming. As a result, objective event-based verification methodologies are required in order to determine the added value of high-resolution NWP models. This paper presents a new objective technique to verify predictions of the sea-breeze phenomenon over east-central Florida by the Regional Atmospheric Modeling System (RAMS) NWP model. The contour error map (CEM) technique identifies sea-breeze transition times in objectively analyzed grids of observed and forecast wind, verifies the forecast sea-breeze transition times against the observed times, and computes the mean post-sea-breeze wind direction and wind speed to compare the observed and forecast winds behind the sea-breeze front. The CEM technique improves upon traditional objective verification techniques and previously used subjective verification methodologies because it is automated, accounts for both spatial and temporal variations, correctly identifies and verifies the sea-breeze transition times, and provides verification contour maps and simple statistical parameters for easy interpretation. The CEM algorithm details are presented and validated against independent meteorological assessments of the sea-breeze transition times and results from a previously published subjective evaluation.

Verified combination of numerical weather- and avalanche-prediction models at Kootenay Pass, British Columbia, Canada

AbstractThe combination of numerical weather prediction (NWP) and snow avalanche forecasting has been performed using the output of two weather models run at the University of British Columbia, Canada, and a local numerical avalanche-forecasting model developed for Kootenay Pass (McClung and Tweedy, 1994). The main motivations for this work are. (1) to extend the lead time of avalanche forecasts by using NWP forecasts of meteorological variables as input to statistical avalanche-threat models (instead of the traditional method of using current/past observed meteorological variables as input); and (2) to create another tool to help avalanche forecasters in their daily decision-making by making true forecasts instead of “nowcasts”. Therefore, verified weather-forecast model output was used as input for the local avalanche-forecasting model at Kootenay Pass. The resulting 24 hour avalanche forecast was compared to observed avalanche occurrences and to the 12 hour avalanche forecast with current weather observations. As a result, the avalanche-model output for the test runs with numerically predicted weather data is comparable in accuracy to the runs with observed weather data. The results also suggest that avalanches may be predicted statistically for 24 hours into the future when high-resolution NWP is used as input, weather- and avalanche-forecast errors taken into account during operational use.

Numerical prediction of severe convection: comparison with operational forecasts

AbstractThe prediction of severe convection is a major forecasting problem in Australia during the summer months. In particular, severe convection in the Sydney basin frequently produces heavy rain or hail, flash flooding, and destructive winds. Convective activity is a forecasting challenge for the Sydney basin, mainly from October to April. Currently, there is a need for improved numerical model guidance to supplement the official probabilistic convective outlooks, issued by the operational forecasters. In this study we assess the performance of a very high resolution (2 km) numerical weather prediction (NWP) model in terms of how well it performed in providing guidance on heavy rainfall and hail, as well as other mesoscale features such as low level convergence lines. Two cases are described in which the operational forecasts were incorrect on both occasions. Non‐severe thunderstorms were predicted on 1 December 2000 but severe convection occurred. Severe convection was predicted on 8 December 2000, but no convection was reported. In contrast, the numerical model performed well, accurately predicting severe convection on 1 December and no convection on 8 December. These results have encouraged a program aimed at providing an enhanced numerical modelling capability to the operational forecasters for the Sydney basin. Copyright © 2003 Royal Meteorological Society

Advanced flood forecasting in Alpine watersheds by coupling meteorological observations and forecasts with a distributed hydrological model

Flood forecasting may be improved by coupling atmospheric and hydrological models. To investigate the current potential of such an approach in complex mountain watersheds, the authors carried out a number of combined high-resolution one-way driven model experiments to generate runoff hydrographs for seven extreme flood events which occurred in the Lago Maggiore basin between 1993 and 2000. The Alpine Ticino–Verzasca–Maggia basin (2627 km 2) is located directly to the south of the main Alpine ridge embracing a great part of the drainage area of Lago Maggiore. For this basin, the grid-based hydrological catchment model WaSiM-ETH was employed to determine the continuous runoff hydrographs. In the model experiments, two different sets of meteorological input data were used: (1) surface observation data from station measurements and from weather radar, and (2) forecast data from five different high-resolution numerical weather prediction (NWP) models with grid cell sizes between 2 and 14 km. This paper presents and compares selected results of these flood runoff simulations with particular attention to the experimental design of the model coupling. The configuration and initialization of the hydrological model runs are outlined as well as the down-scale techniques which proved to provide an adequate spatial interpolation of the meteorological variables onto the 500 m×500 m grid of the hydrological model. In order to evaluate the various hydrological model results as generated from the different outputs from the five NWP models, some coupled experiments with ‘non-standard’ NWP model outputs have been carried out. In particular, the results of these sensitivity studies point to inherent limits of high-resolution flood runoff predictions in complex mountain terrain.

The modeling and observation of a lee trough event over eastern Tasmania

Due to its complex orography, a lee trough can occur along leeward coasts of Tasmania. Of particular interest to meteorologists and others, is the lee trough that can form along eastern parts of Tasmania. This effect is often observed under broad westerly flow due mainly to a topographic barrier, which lies to the west of the area of interest. The lee trough seems to be most pronounced when the troposphere is stable with an inversion just above the highest parts of the western and central highlands, known as the Western Mountains and Central Plateau, respectively. Blocking of the low-level airflow, coupled with the formation of a windward ridge of high pressure may also contribute to the formation of the lee trough. From time to time, mesoscale thermal lows and pronounced standing wave activity has also been observed to form over parts of the state co-incident with the lee trough. These can further complicate the overall wind pattern over parts of the Tasmania. This study attempts to investigate and verify some aspects of this lee trough by the utilization of a very high resolution numerical weather prediction model, known as HIRES (HIghRESolution).

The prediction of two cases of severe convection: implications for forecast guidance

Severe convection in the Sydney basin regularly produces destructive winds, heavy rain or flash flooding, and damage from large hail. Such convective activity is a major forecasting challenge for the Sydney basin, especially during the October to April warm season. There presently is a need to provide timely, accurate and reliable numerical guidance to supplement the current probabilistic convective outlooks, issued by the operational forecasters. Initial work has been carried out that examines two cases of severe convection in the Sydney basin. The performance of a very high resolution (2 km) numerical weather prediction (NWP) model is assessed in terms of how well it performed in providing guidance on heavy rainfall and hail, as well as other key mesoscale phenomena such as low level convergence lines. The model results discriminated well between severe convection that actually did occur in the first case (1 December, 2000) and the failure of severe convection to develop in the second case (8 December, 2000). The operational forecasters predicted severe convection to occur in both cases. It is now planned to take the next step of augmenting the quasi-subjective approach of the operational forecasters with the NWP guidance to provide an enhanced capacity to forecast severe convection in the Sydney basin.

Local Data Integration over East-Central Florida Using the ARPS Data Analysis System

The Applied Meteorology Unit has configured the Advanced Regional Prediction System (ARPS) Data Analysis System (ADAS) to support operational short-range weather forecasting over east-central Florida, including the Kennedy Space Center and Cape Canaveral Air Force Station. The ADAS was modified to assimilate nationally and locally available in situ and remotely sensed observational data into a series of high-resolution gridded analyses every 15 min. The goal for running ADAS over east-central Florida is to generate real-time analysis products that may enhance weather nowcasts and short-range (<6 h) forecasts issued by the 45th Weather Squadron (45 WS), the Spaceflight Meteorology Group (SMG), and the National Weather Service (NWS) at Melbourne, Florida (MLB). The locally configured ADAS has the potential to provide added value because it ingests all operationally available data into a single grid analysis at high spatial and temporal resolutions. ADAS-generated grid analyses can provide forecasters with a tool to develop a more comprehensive understanding of evolving fine-scale weather features than could be obtained by individually examining the disparate data sources. The potential utility of this ADAS configuration to operational forecasters is demonstrated through a postanalysis case study of a thunderstorm outflow boundary that postponed an Atlas space launch mission, and a Florida cool-season squall line event. In the Atlas case study, a thunderstorm outflow boundary generated strong winds that exceeded the Atlas vehicle limits. A diagnosis of this event, using analysis products during the decaying phase of a Florida summer thunderstorm, illustrates the potential benefits that may be provided to forecasters supporting space launch and landing operations, and to NWS MLB meteorologists generating short-range forecast products. The evolution of analyzed cloud fields from the squall line event were used to track the areal coverage and tendencies of cloud ceiling and cloud-top heights that impact the evaluation of space operation weather constraints and NWS aviation products. These cases also illustrate how the analyses can provide guidance for nowcasts and short-range forecasts of Florida warm-season convection and fire-weather parameters. In addition, some of the sensitivities of the ADAS analyses to selected observational data sources are discussed. Recently, a real-time version of ADAS was implemented at both SMG and the NWS MLB forecast offices. Future plans of this ADAS configuration include incorporating additional observational datasets and designing visualization products for specific forecast tasks. Finally, the ultimate goal is to use these ADAS analyses to initialize a high-resolution numerical weather prediction model run locally at SMG and the NWS MLB, in order to develop a cycling scheme that preserves fine-scale features such as convective outflow boundaries in short-range numerical forecasts.

Comparison of GPS slant delay measurements to a numerical model: case study of a cold front passage

GPS water vapour measurements are expected to have the ability to fill the lack of moisture information in numerical weather prediction (NWP) and nowcasting. Zenith total delays (ZTD) are obtained by mapping signal delays in the line of sight to the vertical in the processing of the GPS data. The estimated zenith integrated water vapour (IWV) can be extracted from this ZTD when surface pressure and temperature is available. This method is valid under the assumption that the atmosphere is layered and horizontally stratified. When a cold front passes the observation sites, this assumption is not valid. In this paper a case study of such a cold front passage is examined. Slant delays from GPS and the high resolution NWP model High Resolution Limited Area Model (HIRLAM) are derived, and the non-stratified slant wet delay components from GPS are compared with HIRLAM. Furthermore, GPS derived zenith IWV measurements are compared to HIRLAM. Meteosat water vapour channel information together with IWV derived from GPS and HIRLAM are used for interpreting the results. The additional information of slant measurements for use in synoptic forecasting and NWP is briefly discussed.