The precision obtained from the poisson methodology may be misleading. There is a lot of variability in traffic data, and peak flow rates can vary significantly from day to day. You may be designing your LTL for peak flow on a Tuesday in February, 2008.

Over the years I have seen many LTLs installed without adequate storage, with the result that queues back into the adjacent through lanes. I have come appreciate the wiggle room provided by using twice the average arrival rate to determine the desirable storage.

If you are confident that your data is representative of the average peak 15 minute flow, that the corridor doesn't have unusual seasonal characteristics, that you have accounted for reasonable background growth to provide adequate service life, then you can make a case for using the poisson methodology for calculating storage length. It may reduce your property acquisition & construction costs, but you should be ready to defend your methodology, especially if the results fall outside the AASHTO guidelines.

Greetings Teeman,

I could not agree with you more about the Poisson Method being misleading as it seems that it's cogency with confidence intervals appears to make it unarguable.  However, I was reviewing a traffic study that utilized this methodology and was calculating the storage lengths utilizing the AASHTO method and the "delta" in values were almost consistently 25% more than Poisson. So conclusively to me I ABSOLUTELY agree that some reasonable "wiggle" room LTL stroage as a buffer wouldn't hurt, especially when compared to the alternative (i.e.,spillback left turns storing in through lanes causing a degradation in throughput capacity and a higher probability of rear end collisions). This doesn't account for the added cost to rework and lengthen turn lanes (provided that there is room to do as such).
 
A thousand-fold thank yous for your response.

Coleagially,

Tbone1