Calculating how much sugar is in 0.001 SG of a specific volume of beer

I posted this on another thread, but felt it should be its own post. I frequently bottle beers with brettanomyces. Brettanomyces will continue to ferment complex sugars and carbohydrates in the beer slowly over time. It's common practice by commercial breweries (Russian River, Crooked Stave, etc.) to bottle these beers when the beer reaches a certain gravity (say, below 1.007), but they will factor in additional gravity points into their bottle priming sugar so that they do not end up with gushers (or bottle bombs), and because they will be aging the beer for 6 months in the bottle before releasing them.

What I'm looking for is a formula for calculating how many volumes of CO2 will be gained if the beer drops 0.001 specific gravity. Or, more simply, how much corn sugar to subtract from my bottle priming to account for a specific gravity change over time in the bottle. I want to subtract a couple of gravity points worth of corn sugar from my bottle priming sugar calculator when bottling my current batch of beer fermented with brettanomyces. Any help with a formula for calculating this would be greatly appreciated. I've found plenty of bottle priming calculators online, but none of them account for additional gravity change over time.

Thanks in advance if you can point me in the right direction.

As a rough approximation, I plugged into brewmate 5oz corn sugar in 5 gallons = 1.003

One specific gravity 'point' is about 1/4 of a Plato degree and a Plato degree is 1% extract by weight. More precise correspondence between the SG and Plato can be had from the ASBC table or by use of the polynomial Plato = -616.868 +1111.145*SG -630.272*SG*SG +135.997*SG*SG*SG

Thus if you have beer whose True Extract were 2.560 °P (1.010 SG ) and the additional fermentation brought it down to 1.283 °P (1.050 SG) the extract consumed would be 1.27°P corresponding, in this range, to 3.914 'points' per °P. Notice that I referred to true extract. This is because the presence of alcohol influences specific gravity readings. With a little experimentation you should be able to figure out the relationship between decline in true and apparent extract at these levels. Measuring true extract isn't that big a production. Fill a (preferably) volumetric flask with the beer to the mark. Transfer to a beaker. Rinse the flask with a small quantity of DI water and add the rinsings to the beaker. Evaporate to approximately 1/3 the original volume. The evaporation can be done by boiling in which case it is better to use a flask than a beaker. At the conclusion of the evaporation transfer the beer back to the volumetric flask (which has been washed thoroughly and rinsed with DI water). Rinse the evaporation vessel with small amounts of DI water and transfer the washings to the volumetric flask. Make up to the mark with DI water. Mix thoroughly and measure the SG of dealcoholized beer. The true extract is the Plato value of the dealcoholized beer (as calculated from its SG using the formula above) multiplied by the SG of the extract and divided by the SG of the pre-evaporation beer.

Now that you have the change in True extract you find the weight of a liter of beer by multiplying 998.203 grams time the specific gravity of the beer and then multiply that by the change in true extract (divided by 100 as °P are percentages). The result is the number of grams of sugar consumed. Each 2.0665 gram of extract consumed produces, on average, 1 gram of alcohol, 0.9565 grams of CO2 and 0.11 grams of yeast. Note that these are averages and are for normal fermentations. Brett may perform differently but probably not terribly differently. In any case you can use these numbers to calculate the number of grams of CO2 produced per liter of beer. According to the ASBC (MOA Beer -13)

volume CO2 = 5.0607*CO2_%_by_wt*SG

and you can calculate the volumes of CO2 from that.

I've left plenty for you to do but these are the essential facts from which you should be able to figure out what you want to know.