May be I miss something, so may answer may affected. In my experience, more than the spatial area, the use of that area affects the trips generation and then, considering the transport supply, the origin_destination flows and path flows. Thus, I would focus more on density of urban activities than spatial area only.

Thank you for your reply Michele. I am quite familiar with OD matrix and trip distribution issue. I understand the importance of use, density, wealth of people, modal split and other atributes of urban areas for traffic generation. Do you mean gravity model should be appropriate mathematical method for determining these linkages? I just wanted to find out if somebody has dealt with the specific issue (spatial area in the context of traffic volumes) or maybe the idea is bad :-)

So, you are looking for a direct link between traffic and area extension. At the best of my knowledge this issue has not been investigated. Traditionally, I would refer to the 4steps modelling approach instead gravity model. For you purposes, gravity models should work only if the population and/or other attributes are uniformely distributed.