There are a couple reasons why this could occur. If a queue is spilling through an upstream intersection, this could cause it. Assume that a vehicle is just over the upstream stopbar. Since it has begun to cross the stop-bar, it will be measured with the downstream intersection. Generally, this is not a huge impact.
Second, the 95th percentile queue is never really observed. The average queue is first determined. The average queue is the average of two minute maximums. The 95th percentile then is a statistical calculation on the average. So, if the average queue was near the link length, the 95th could be much higher (1.65 standard deviations more).
Is the 95th percentile queue calculated in SimTraffic as well as Synchro? It was my understanding that all queue measurements in SimTraffic were observed during the simulation.
The 95th Percentile Queue in SimTraffic is calculated, and may never really be observed. The 95th Percentile Q is a calculation on the Average Queue. In SimTraffic, a maximum queue is measured every 2 minutes. These 2 minute maximums are then averaged to get the Average Queue. The maximum queue is simply the maximum that is every observed in the simulation.
