You can fairly easily and accurately estimate the volume of existing reservoirs from seismic surveys. But you then need to make assumptions about how much of it you can extract and what effect new extraction technolgies would have.
Then looking for new fields is a bit more specultive, you can do surveys which show there are similair features under the atlantic as under the North Sea and then say - if they contain the same amount of oil and are 10x bigger then there is 10x as much oil as in the North Sea. Do the same for oil sands etc..
Then there is the question of consumption / extraction technology / extraction cost.
If consumption goes up we have less oil, but then demand goes up so it is worthwhile expoiting more expensive sources such as oil sands. As price goes up it becomes worth investing in new technology to both find oil and extract it - so supply goes up.
Then there's politics. If you are a green campainger you pick the most conservative estimates of available reserves and assume the worst estimates of growth in demand to get 15years. If you are a car maker you pick the opposites.
The oil industry is notorious for overstating it's known reserves - partly this is just human nature, everybody on the exploration teams wants to give good news to his manager, the layers of management all want to give good news to their superiors - so reserves get exagerated. Then the board just flat out lies to the investors.