【正文】 C. In order to better estimate the future atmospheric state ensemble models give better guidance than a single model run. Combined quantities like ensemble mean, spread and others allow probabilities to be estimated which can be used for advanced planning. Here output of the KENDA ensemble model from the German Meteorological Service, DWD, can be used in future to provide this probability information.However, in order to mitigate the impact of wintry weather conditions on airport operations more efficiently, the focus should be laid on shortterm forecasting (termed “nowcasting”) these conditions. This prises the onset, duration and type of precipitation as rain, snow, freezing rain, or fog. DLR is developing a nowcasting system that provides users in aviation with 0 to 2 hour forecasts of these winter weather conditions[8]. We argue that a similar system imbedded in the process of information sharing for collaborative decision making would also be beneficial for operations on road networks. 2 WINTER WEATHER OBJECTSA certain winter weather phenomenon, like . freezing precipitation, can be thought of a certain volume of air within which this phenomenon can be observed. Various observations are suited for describing one or the other attribute of that phenomenon, as . the surface temperature, the precipitation type. With no doubt the actual weather phenomenon can be determined more precisely when data from various sensors are bined [7]. It is therefore advisable to think of such volumes as weather objects with certain inherent attributes. For our purposes, a winter weather object (WWO) in a certain limited area, . an airport or a dense motorway network, can be defined through the following parameters:? a vertical column of air consisting of several layers? issued time? valid time? next update time? layer description, .: Snow: upper and lower boundary with intensity: light, moderate, severe Rain: upper and lower boundary with intensity: light, moderate, severe Freezing rain: upper and lower boundary Freezing drizzle: upper and lower boundary? surface conditions? trends, . intensity increasing, change to melting, etc.Figure 1 sketches how weather parameters from various sources are bined by data fusion to a winter weather object, WWO (yellow cylinder), with different attributes in different layers. SYNOP and automatic sensors (as from SWIS) allow determining surface conditions, in this example rain with temperature above zero. The temperature/humidity sounding can be provided from a numerical weather forecast model, aircraft measured data (AMDAR), or constructed from both depending on data availability. Radars observe the precipitation height and may also be able to determine the hydrometeors within the cloud (polarimetric capability and related algorithms). ADWICE – the Advanced Diagnosis and Warning System for Icing Environments – [6,4] uses the information of reported weather at the ground together with the soundings of temperature and humidity and radar measurements to determine the icing threat to aircraft in flight. ADWICE is now further expanded to diagnose and predict snow and icing conditions at the surface, too, see following Section. Taken together, the derived analysis can be pacted into the WWO which is shown schematically as a yellow cylinder on the right of Figure 1. It is obvious that the object can have several different hazard layers in the vertical. For the given case there would be a near surface layer with temperatures above freezing up to height H1 which contains rain drops, a second layer from H1 to H2 which contains supercooled droplets with corresponding icing threat, and a precipitating cloud layer on top. For nowcasting icing amp。2Hydrometeorological Innovate Solutions ., Barcelona, SpainEmail: ABSTRACTRecent developments are reported on techniques to determine the onset, duration, amount and type of precipitation as well as the snow and icing conditions at the surface. The algorithms, still under development, will be used to forecast the weather in short to medium lead times, . for the next 30 minutes up to a few hours (“nowcasting”). An algorithm aims at detecting potential areas of snow fall by bining reflectivity data of precipitation and surface temperature data from a numerical model as well as surface stations in high spatial resolution. Another approach bines profiling measurements with numerical weather forecast products. (., meteo data measured by aircraft and polarimetric radar data)Keywords: type and amount of precipitation, nowcast, anticipating the weather 1 INTRODUCTIONWeather phenomena contribute to congestions, accidents and delays in all traffic modes. The road traffic in particular is derogate by adverse weather like snow, ice, fog, rain, strong wind and wind gusts. Increasing traffic makes transportation even more vulnerable to adverse weather conditions. Today stakeholders and participants in transportation (be it airborne or groundbased) most of the time only react on adverse weather when the disruption has already happened or is just about to happen. Future road management systems should proactively anticipate disruptive weather elements and their time scales of minutes to days well in advance to avoid or to mitigate the impact upon the traffic flow. But “weather” is not a technical problem that can be simply solved. Predicting the weather is a difficult and plex task and only possible within certain limits. It is therefore necessary to observe and forecast the changing state of the atmosphere as precisely and as rapidly as possible. Moreover, measures are required that translate “weather” to “impact” and minimise those impacts on traffic flow and its management. To inform traffic participants and traffic management centres in due time on (expected) adverse conditions, tailored and accurate meteorological information is required on short notice. This information