First the regarding the answer to the question of road or rail access - why not both?
My understanding is that logging companies building access roads in this part of Ontario figure $50,000/ mile. The rule of thumb for building railroad track that I am aware of is a $1,000,000 per mile. Neither of these estimates would not include the cost of bridges or dealing with major physical obstacles. Assuming realtively smooth sailing in terms of building track or road to run 200 miles into the ROF one would be looking at a ballpark number of $200,000,000 for rail or $10,000,000 for a gravel road suitable for heavy truck traffic. Looking at the proposed route on the KWG site presentation there appear to be seven bridges involving rivers big enough to show up on the map. Even if you double both numbers to account for bridges and other major obstacles I think the bottom line is that a road is relatively small potatoes compared to the cost of a rail line. Assuming there was enough tonnage to justify a rail link, a road would still be the most economical way to move a lot of supplies in. Having some experience with rail shipments I can assure you rail is not practical for just in time delivery. In studies I have been involved with only 40% of shipments arrived within a 5 day window including the desired delivery date. I am aware of one university based study that demonstrated that in the continental US there was no statistical correlation between miles a shipment moved by rail and the time it took to arrive.
Getting to rail transport you are going to need to move a significant amount of tonnage before rail becomes economical. Typically a mine would own it's own cars figure a 100 tons of payload per car. Probably a good cost estimate would be $100,000 per car - I'm not familiar with ore cars - but this is a typical cost number for a plastic pellet hopper car - and I would guess the costs would be similar. The next key issue would be how fast can you turn your cars. Ideally you would schedule a unit train from the mine to a single steel plant or refinery or even a distribution point to minimize turn time. It really depends on how many end users there are for the material. Concievably you might get your turns down to one to 2 weeks if you have a single end delivery location and ship unit trains from the mine. If you end up breaking up your train and shipping to multiple customers you would be lucky to average better than turning your cars once every 6 weeks. Turn time of course has a huge impact on the number of cars required to move a fixed amount production.
In terms of what do railroads charge for moving freight - the answer is whatever they can get, but if you have a large volume moving the over the tracks you are probably in a position to negotiate a better rate. I believe $30/ton to go a 1000 miles would be about typical for ordinary single car movements in shipper owned cars over railroad owned track but this is not something I am involved with on a day to basis.
