Accurate prediction of thunderstorms during the pre-monsoon season (April-June) in India is essential for human activities such as construction, aviation and agriculture. Two objective forecasting methods are developed using data from May and June for 1985–89. The developed methods are tested with independent data sets of the recent years, namely May and June for the years 1994 and 1995. The first method is based on a graphical technique. Fifteen different types of stability index are used in combinations of different pairs. It is found that Showalter index versus Totals total index and Jefferson's modified index versus George index can cluster cases of occurrence of thunderstorms mixed with a few cases of non-occurrence along a zone. The zones are demarcated and further sub-zones are created for clarity. The probability of occurrence/non-occurrence of thunderstorms in each sub-zone is then calculated. The second approach uses a multiple regression method to predict the occurrence/non-occurrence of thunderstorms. A total of 274 potential predictors are subjected to stepwise screening and nine significant predictors are selected to formulate a multiple regression equation that gives the forecast in probabilistic terms. Out of the two methods tested, it is found that the multiple regression method gives consistently better results with developmental as well as independent data sets; it is a potential method for operational use.

Regional air pollution models are usually off-line coupled with numerical weather prediction models. The present study, however, focuses on on-line coupled modelling, for which the air pollution model is an integral part of the meteorological model. To this purpose, simulations of the first European Tracer Experiment (ETEX), which supplies the best suited verification data for dispersion modelling of a passive tracer, have been performed using versions of the Danish Meteorological Institute High Resolution Limited-Area Model (DMI-HIRLAM). Simulation results have been compared with observed values as well as with results from other models.

The operational version of DMI-HIRLAM using well-known basic advection schemes has been tested, showing the need for improvements for air pollution purposes. The fourth-order Bott algorithm for advection has thus been implemented in the semi-Lagrangian version of DMI-HIRLAM. Due to its practically perfect mass conservation properties and its low computational cost, this method turns out to be efficient for simulations of transport of pollutants in the atmosphere. This represents a first effort towards a fully integrated air pollution model, e.g. for forecasting purposes.

Surface observations, satellite and radar imagery and cloud-to-ground lightning data are used in an observational study of two cases that produced severe weather in Catalonia (Spain). The first one occurred on 24 August 1993; a squall line crossed Catalonia from west to east producing heavy rain with rates of up to 100 mm h−1 and hail of 7 cm diameter. The observational information provided is a good tool for monitoring the event and issuing a reasonable nowcast. The second case, which occurred on 31 August 1994, was associated with the development of a tornado (F1 in the Fujita scale) as well as hail of up to 5 cm diameter. In this case the convection was almost stationary and no clear signatures of severe weather can be identified from available satellite and radar imagery.

This paper presents a general review of ensemble forecasting techniques, with a focus on short-range and tropical cyclone predictions. The basic ideas and terminology of ensemble forecasting are introduced, and using four measures to evaluate an ensemble (ensemble mean forecast, consistency, spread versus skill, and inclusiveness), various potential utilities (e.g. dynamical probabilistic forecasts) are illustrated. Since the perturbation methodologies designed for medium-range forecasts of mid-latitude synoptic-scale systems - singular vectors, bred modes, and so on - are already quite mature, they are only briefly described here. The general problems encountered in applying ensemble forecasting techniques to short-range and tropical cyclone forecasts are diagnosed, and some recent studies on these topics reviewed. In general, the perturbation methodologies used for short-range ensembles to date can have a skill comparable to or slightly higher than their corresponding high-resolution control forecast. However, the complicated problem of the relationship between initial condition errors and model deficiencies persists. A similar situation also applies to ensembles designed for tropical cyclone forecasting. An additional difficulty is the different error characteristics encountered in the tropics, mainly the result of the strong convection in the area and the mutual interaction with the ocean. Studies from several research groups use quite different perturbation methodologies, but the results are encouraging. Most of them performed ensemble forecasting of tropical cyclone motion, but extensions to tropical cyclone intensity forecasts are also being developed.

We discuss how weather forecasts can be used in the pricing of weather derivatives and derive results for the most important types of weather index and contract. We show that calculating the expected payoff of linear contracts on linear indices requires only forecasts of the mean temperature over the contract period. Calculating the expected payoff of linear contracts on non-linear indices requires forecasts of both the mean and the distribution of temperatures, but not of the dependence between temperature distributions on different days. Calculating the expected payoff of non-linear contracts requires forecasts of the full multivariate distribution of temperature over the whole contract. For contracts that extend beyond the end of available forecasts, correlations between the forecast and post-forecast periods must be taken into account when estimating this distribution. We present two methods by which this can be achieved, both of which combine information from climatological models of daily temperature with information from probabilistic forecasts.

Convective precipitation events in northern Italy during 1996 and 1997 are analysed using two infrared-based geosynchronous satellite rainfall estimation methods to verify the level of applicability of the techniques for operational applications in the area, their quantitative results, and relative performances. The Negri–Adler–Wetzel (NAW) and the convective stratiform technique (CST) are applied to METEOSAT's thermal infrared (IR) data. C-band radar reflectivity fields detail the vertical and horizontal structure of the cloud systems, and radar rainfall data are retrieved. Satellite rain areas are checked against simultaneous radar rainfall retrievals through a contingency analysis procedure. A semi-quantitative analysis is presented. Positive brightness temperature differences between water vapour and thermal IR channels are also examined and related to the storms' development stage and rainrate. Results show that NAW and CST perform reasonably in delimiting rain areas during active convection and care should be used in the initial and final development stage when statistical parameters lose most of their significance. NAW tends to overestimate rainfall while CST approaches more closely radar measurements. Most common errors arise from considering only portions of the storm, contamination from cold non-precipitating cloud, and merging of two or more cloud masses of independent origin. Operational applications, though not completely quantitative, are also possible, including positive values of the difference between water vapour and IR brightness temperature.

A system has been developed and made operational at the Met Office for the real-time diagnosis of soil state and surface hydrology. It is based on the Met Office Surface Exchanges Scheme (MOSES) modified to take account of unresolved soil and topographic heterogeneity when calculating surface runoff by incorporating a Probability Distributed Moisture (PDM) scheme developed by the Centre for Ecology and Hydrology. The implementation of MOSES-PDM in the Met Office's Nimrod nowcasting system is described. High resolution soil characteristics and land cover data, together with Nimrod's analyses of precipitation amount and type, cloud cover and near-surface atmospheric variables are used to drive MOSES-PDM. Hourly values of snowmelt, runoff, net surface radiation, evaporation, potential evaporation, soil temperature, soil moisture and soil moisture deficit are calculated on a 5 km grid.

The motion of thunderstorms in southern Germany was investigated. The thunderstorms were observed by a lightning position system during the summer months of the years 1992–1996. On average every second day thunderstorms were observed somewhere in southern Germany. In general thunderstorms approached from westerly and south-westerly directions. The average speed was 13 m s−1. No significant relation between the occurrence of thunderstorms and the large-scale synoptic pattern described by the Grosswetterlagen (large-scale weather pattern) was found. Thunderstorms were observed during almost all Grosswetterlagen. The reduction to eight weather patterns based on the low-level flow in southern Germany showed that thunderstorms are likely when the flow has a westerly (43%) or easterly direction (20%). Three distinct groups of different lighting patterns could be identified: stationary, moving thunderstorms and thunderstorm lines. The convective available potential energy (CAPE) and the wind shear were retrieved from radio soundings from München and Stuttgart. On average CAPE was 583 J kg−1 for stationary thunderstorms, 701 J kg−1 for moving thunderstorms and 876 J kg−1 for thunderstorm lines. The corresponding average bulk Richardson numbers were 37, 22 and 21. The steering level was found to be at about 6 km m.s.l. However, it should be noted that in most cases the soundings do not completely describe the local environment of thunderstorms, since radio soundings are only available twice a day.

Operational forecasting of mesoscale weather systems in Hong Kong is supported by an automated nowcasting system called SWIRLS (Short-range Warning of Intense Rainstorms in Localised Systems). SWIRLS is based on the extrapolation of radar echoes using the TREC (Tracking Radar Echoes by Correlation) technique. With a suitable choice of pixel array size on the radar reflectivity maps, the TREC vectors derived can be used to monitor and extrapolate echo motion right across the mesoscale spectrum, from individual convective cells, to supercells and clusters, and to groups of rainbands or squall lines.

On the basis of TREC, quantitative precipitation forecast (QPF) algorithms have been developed to produce high resolution forecast rainfall distribution maps over the local area. These maps provide useful objective guidance for forecasters to assess the likely rain scenario in the next few hours and to facilitate decision-making in operating the Rainstorm Warning System. This review, however, will focus on other aspects of TREC applications and some side benefits for operational mesoscale forecasting in Hong Kong. First, extension of the TREC technique to group tracking of echoes (GTrack) have enabled forecasters to make qualitative educated guesses of the likelihood of prolonged heavy rain or the potential of enhanced storm development. Secondly, extension of the QPF algorithms to a forecast range of 3 hours, supplemented by real-time accumulated rainfall data from a high density raingauge network, has enabled forecasters to provide reliable advice and guidance to geotechnical engineers for landslip risk assessment. Thirdly, rain-related applications aside, TREC vectors are also ingested into a data analysis system LAPS (Local Analysis and Prediction System) adapted from Forecast Systems Laboratory (FSL) of NOAA. Through the assimilation of TREC vectors and other wind observations, a three-dimensional wind structure of tropical cyclones can be generated in near real-time for forecasters' reference in assessing landfall impact.

A forecasting system was set up to infer the impact of mountain waves for producing mesoscale temperature anomalies in the stratosphere. It was successfully applied during a recent aircraft campaign which operated out of Rovaniemi, northern Finland, between 18 December 1996 and 15 January 1997. The forecast information led to the successful detection of mountain-induced polar stratospheric clouds on 9 January 1997. For this event mesoscale forecasts, hindcast runs including grid nesting and comparisons with lidar observations and radiosonde measurements are presented.

Many northern countries use a road weather information system (RWIS) with a network of stations to monitor winter road conditions. Present station locations were selected after field investigations of micro- and local-climate conditions (e.g. using thermal mapping). This paper describes an approach to optimally locate and equip the stations in order to best identify conditions hazardous to road transport. This is achieved using multiple regression analysis of observed data and correlation with location meta-data. A geographical information system (GIS) is used to develop quantitative and objective descriptions of station locations by using knowledge of local and regional climate variations. Road climate is described using a slipperiness classification, in which weather situations are classified into ten types of slipperiness from the meteorological variables collected at RWIS stations. The relationships between quantified locations and data on road slipperiness in southern Sweden during one winter are analysed. The results show that the spatial patterns for different types of slipperiness are significantly related to local parameters. The three most prevalent types are analysed in detail: snowfall on a frozen road surface, hoarfrost and low visibility, and strong formation of hoarfrost.

The performance of three automatic depression tracking schemes developed by Terry & Atlas (1996), Murray & Simmonds (1991a) and Konig et al. (1993) when applied over one month are assessed. The schemes respectively identify depressions by (a) locating the innermost closed contour in a PMSL field, (b) finding maxima in the curvature in a bi-cubic spline fitted to the PMSL data and (c) performing a grid point search to identify a minimum in a PMSL field. The largest number of depressions was found by the Murray and Simmonds scheme, with the Atlas and Terry scheme finding the least. The Murray and Simmonds scheme also found the largest number of tracks, with the other two having comparable numbers of tracks. Two possible explanations for the differences in the number of tracks are considered. Firstly, the case where one or more centres identified by one scheme as corresponding to a single track are not found by the other schemes. Secondly, the case where a single track found by one scheme is split into two or more tracks by the other schemes. All three techniques had a similar latitude of cyclogenesis, although the Atlas and Terry and Konig et al. schemes found more lows at high latitudes as a result of using data on a latitude/longitude grid. The longest mean track length was found with the Konig et al. scheme. A comparison of the PMSL fields with satellite imagery shows that the major NWP centres have trouble producing reliable analyses around the Antarctic because of the lack of data.

A method for classifying different types of slipperiness on roads in Sweden is described. Using this method it is possible to survey road conditions in different areas and between different years to optimise winter road maintenance. Winter road maintenance in Sweden is generally undertaken by the national road administration to improve winter-time road conditions, thereby keeping up the traffic flow and decreasing the accident rate. As a number of different types of slipperiness may develop on roads in winter, each due to a specific set of meteorological variables, maintenance work can be a complicated task. With the proposed classification method it becomes easier for the winter maintenance personnel to analyse information on road conditions and survey the distribution of road slipperiness in a region. The classification is performed with an expert system using meteorological data from the Swedish Road Weather Information System. The road condition is classified as good or as one out of ten different types of slipperiness on roads. Road conditions during three different winter periods are analysed. The results show that variations in climate produce substantial differences in annual road condition characteristics. The output from the expert system classifying road slipperiness is compared with recorded winter road maintenance reports. Maintenance action took place on 49% of all occasions when road conditions were classified as slippery.

Anomalous propagation (anaprop), analogous to the upper mirage in the visual wavelengths, is still a major problem in radar meteorology. This phenomenon assumes particular importance in automatic recognition and estimation of rainfall. Anaprop echoes from terrain features such as hills and coasts often give echoes up to 50-60 dBZ, equivalent to heavy rain or hail in severe thunderstorms. Anaprop echoes from sea waves may be comparable in strength to those from moderate precipitation and also form similar patterns. Based on the evidence that the vertical reflectivity profile of precipitation is quite different from the anaprop profile, two methods for anaprop identification are presented. The method proposed by the Servizio Meteorologico Regionale (SMR, Italy) simply uses the operational scan procedure to discriminate between precipitation and anaprop. At the Swedish Meteorological and Hydrological Institute an ‘ad hoc’ scan strategy has been developed in order to obtain much more detail of the lowest reflectivity profile. A number of statistical parameters have been used to achieve a better discrimination between precipitation, land and sea clutter. A number of case studies, representing different echo intensities and patterns, and including a case of anaprop with embedded precipitation, are presented to assess the impact of these methods.

The pricing of weather derivatives motivates the need to build accurate statistical models of daily temperature variability.Current published models are shown to be inaccurate for locations that show strong seasonality in the probability distribution and autocorrelation structure of temperature anomalies. With respect to the first of these problems, we present a new transform that allows seasonally varying non-normal temperature anomaly distributions to be cast into normal distributions. With respect to the second, we present a new parametric time-series model that captures both the seasonality and the slow decay of the autocorrelation structure of observed temperature anomalies. This model is valid when the seasonality is slowly varying. We also present a simple non-parametric method that is accurate in all cases, including extreme non-normality and rapidly varying seasonality. Application of these new methods in some realistic weather derivative valuation examples shows that they can have a very large impact on the final price when compared to existing methods.

During the last two weeks of March 1980 an intense heat wave developed in central and northern Argentina, with its greatest intensity over the city of Buenos Aires. The sequence of temperature anomalies showed the importance of this heat wave. During the period of maximum intensity the daily minimum temperatures were over 27 °C and the apparent temperatures over 30 °C. It was the most intense heat wave of the last two decades and the only one that has occurred in March. The intensity of the event is shown by the high values of several biometeorological indices that indicate dangerous levels of thermal stress on people. The persistence of an anticyclonic circulation at all tropospheric levels and the extension of the barotropic zone to the south led to the maintenance of high values of temperature and humidity. The advective effect added to the intense solar radiation was the most important cause of the high temperatures.

A whole-farm-level decision model is used to examine the impact of the type of decisions producers make on the value of seasonal climate forecasts. Results suggest that precipitation forecasts directed towards crop mix and applied nitrogen level decisions would have the largest economic value. Further, the results show that the economic value of climate forecasts cannot be obtained by examining only a small set of decision types. Rather, all decision types must be modelled to value seasonal precipitation forecasts correctly. This occurs because in response to seasonal climate forecasts changes in one type of decision may override the need to change other decision types. Finally, forecasts of precipitation during the crop tasselling and grain filling stages may be more valuable than precipitation forecasts for earlier crop growth periods.

The difference between air and road surface temperature in urban and rural areas is an important consideration when modelling the road climate. In this study the effect of the urban heat island in the Stockholm area on road climate is examined. Factors such as distance from the city centre, traffic and topography are analysed in order to assess their impact on the spatial variation of road and air temperature.

Freezing rain in the winter of 1995/96 has drawn attention to the severity of black-ice problems on British roads, additional to the normal increased hazards presented to drivers by any form of precipitation. Disruption to traffic was considerable on several days. Weather radar provides improved nowcasting, for both winter and summer conditions, where available real time directly to highway engineers, but this is not yet generally the case in England and Wales. New developments from the Meteorological Office offer improved possibilities at reasonable costs.

The Champagne vineyards, famous worldwide, are located at the northern limit of viticulture. Consequently, they are very sensitive to spring frosts that can occur after bud burst. These spring frosts occur mainly in radiative atmospheric situations (low speed wind, clear sky) where spatial variations in minimum temperatures are very important. The latter depend on several factors at different scales: from the macroclimatic scale, where the geographical location (latitude/longitude) has a strong influence, to the topoclimatic scale, where environmental and topographical factors play an important role. After a brief recall of the definition of these different factors extracted from a Digital Elevation Model and a land-use database, we propose a spatial interpolation of minimum temperatures in the Champagne vineyards based on multiple regression. From this interpolation, we can map the frost hazard for the whole vineyard from data measured over five spring seasons (1998–2003) and for about 20 weather stations. In this paper, we present this mapping technique and compare it with the data for spring 2003 when particularly strong frosts occurred.