In 2000: 76 mil. -> (0,76) the measured amount at time zero
In 2010: 87 mil. -> (10,87) the measured amount ten years later
Using these 2 points, find the slope... m = 11/10
Plug the slope and either point (I chose the first) into the point-slope form of the equation of a line:
y - 76 = 11/10(x - 0) --> y = 11x/10 + 76
For 2050: How many million barrels? --> (50, ?) We have x but need y. Use the equation we've just derived:
y = 11(50)/10 + 76 = 131
So, with the same usage every year [that's what makes/keeps it linear], in 2050 the usage would be
131 million barrels.
I don't know what "less linear" means, but to be non-linear would each year usage is changing by a different amount.
To "flatten" I guess, would mean to not change from one year to the next [slope = 0].
Modeling this would mean just that--not change usage amount from one year to the next.