In the Road Condition Model one wants to be able to predict the transition between different winter road conditions. One such example is the transition from loose snow all across the driving lane to loose snow except in wheel tracks, where the condition is changed to ice or bare road. Each vehicle passing as long as there is loose snow across the whole driving lane will remove a certain amount of loose snow from the developing wheel tracks. The amount of removed loose snow is assumed to vary between different types of vehicles. The Road Weather Information System, RWIS, delivers information on the amount of loose snow falling on the road. The traffic flow is measured by inductive loops. The amount of loose snow in wheel tracks is measured intermittent. Then one has information of traffic flow and the amount of removed loose snow from wheel tracks (determined as the difference between the amount af loose snow falling on the road and the amount remaining in the wheel tracks). Trough regression analysis one can determine the statistical relation between traffic flow and amount of removed loose snow. However, one problem is that on some occasions there is no loose snow at all remaining in the wheel tracks at the time of observation. In that case one can assume that the observed amount of traffic flow would have removed more snow if the snow fall would have been heavier. Those observations are considered censurered since the actual amount of removed loose snow exceeds the observed.