Pollution model suggestion

[Ijuin]

Since it has been rumored that the pollution system in Civ4 will be different from what we have known previously, I would like to make a suggestion for a new pollution model.

Instead of having individual tiles get polluted and lose production, each city will have a "pollution index" based on its industrial production and its population. The city's total food/shield/commerce production would then be reduced by a percentage in proportion to this pollution index, in a manner similar to how corruption works. For example, a highly polluting city in the Industrial age (after the availability of production-enhancing buildings but before the availability of the later pollution-enhancing buildings) could lose thirty to forty percent of its food, shields, and gold to pollution.

The pollution index of a city could be reduced by building anti-pollution buildings, or by paying a per-turn fee in gold to "clean up" the pollution (which would be much more expensive than the maintenance costs of the anti-pollution buildings--it should cost at least one gold for every unit of food/shield/commerce recovered by the pollution reduction).

[Dauphin]

Pollution could do with being split into two simple categories. Local and global pollution.

Local:
I don't think local pollution should effect production and commerce per se. I think it should be factored in through population unhappiness and health risks. More pollution leads to unhappier citizens and higher risk of population dying. Pollution is not cumulative, and if no pollution is produced in the recent past then its effects become negated.

Global:
As the global pollution increases to high levels tiles change - desertification, coastal grassland to marshland etc.. It should be gradual and minor changes and not a sudden and 'catastrophic' event. Effects should be cumulative - such as a step based approach where a tile change occurs after every 1,000 pollution smoke stacks accumulate. As its global the location of the tile change is not dependent on where the polluting is occuring.

[Lawrence of Arabia]

and the model would have to be exponential, not constant.

f(x)= c^(x/10) - 1 from x = 0 to x = inf
c is a constant
c>or= 10
x = level of pollution
x axis = amount of pollution
y axis = amount lost to pollution

the cleanup rate can be the area of the function, multiplied by some constant K (a fraction which is that rate at which the cleanup is going at)

f(x) = 20^(0.1x) - x
F(x) = (20/1.1)^(1.1x) -x^2/2

amount of time for total cleanup would be
K(20/1.1)^(1.1x) - x^2/2

if no new pollution was created in that time.
however, if you want it to be more complex (and more lifelike) K can also be an exponential functino (not a constant) with decreasing returns (ie the closer to zero the pollution level is, the more input (recycling plants, 'waste management' (like enterntainers etc)) will be needed to make it go down one level.

EDIT: maybe change f(x) to c^(x/10) - 2, because a little pollution will not upset production or cause wastage.