To give you a clue how far pollen can reasonably fly.
A study was done a while back by Ernest Small and Tanya Antle in Ottawa, Canada called:
A Preliminary Study of Pollen Dispersal in Cannabis sativa in Relation to Wind Direction
Pollen of Cannabis sativa is disseminated by wind in large amounts and for long distances, and regulations concerning the production of pedigreed seed of industrial hemp, therefore, often call for extremely large isolation distances to prevent unwanted pollination. In Europe and Canada, a standard distance of 5 km is required for the highest classes of hemp seed. This study examines the relative distribution of pollen from an isolated field over the 3-week maximum flowering period, with particular reference to wind direction. The amount of pollen distributed downwind was about six times the amount distributed upwind. In effect, this means that an isolation distance of 5 km on the downwind side is about equivalent to an isolation distance of 0.9 km on the upwind side. In theory, at the experimental site examined, the required isolation area could be reduced by about 58% while still achieving the equivalent of 5 km isolation in all directions. Given that weedy and illegally cultivated plants are widespread, making it difficult to ensure their absence over a distance of 5 km, it seems advisable, when possible, to take advantage of the considerably reduced isolation distance that is necessary on the upwind side. Pollen distribution appeared to follow the expected leptokurtic curve, reducing rapidly with initial distance from the source, but much more slowly with increasing distance. This makes it impossible to guarantee complete absence of potentially contaminating pollen in the field and, for practical purposes, a very low amount of undesired gene flow needs to be tolerated.