Not sure altogether if we are speaking about the wine or biocide. Wine has a density of 1,005 so mixing weight (mg) with volume (l) is a foul play that gives acceptable results.
Hopefully, we agree that ppm is a unit-less multiplier. Per-cent is 1/100, per-million is 1/10^6.
The AMM above says PPM (by volume) thus I would expect a proper calculation in volume : volume units. To adress the density issue (fuel tank gauges in KG, dosage in l or ml -
I only assume here), the calculation guidelines in a
pilot-proof world would need to take this form:
Using your numbers of 7000 L fuel to achieve 100 ppm by volume, 0.7 L of agent needs to be added. And that is 735 g of the matter. If you say the recipe on the can does actually instruct to mix mili- or kilo- grammes on one side and mili- or full liters on the other (either way) - I'll take your word for it. But calling that ppm is gobbledygook.
One more thing I inserted into that pilot-proof example. Realistic values. Which provides the reader with an outlook on what his result should look like. Compare that with the AMM original text suggesting adding 10000 kg / lbs ontop 10000 kg/lbs of fuel already there. Pardon me? There is no tank on the machine that'd hold 10.000 kg of fuel nor 20000 lbs by a wide margin!
Massive case study on how many things can be unhelpful in a technical writing. Unhelpful actually does mean incorrect and wrong. Human Factors - biting the bottoms since well before the Wright brothers.
With Occam's razor, my gut feeling is that the decimal denominator "," "." actually got mixed (same happened in posts #56 and #57 here) and something got multiplied by 1000 rather than divided. Who knows, happy everyone survived to tell the tale.