Sorry for being pedantic but some offerings seem to have been submitted without adequate thought of what is really happening in the hive throughout, say, a daily period.
Let's consider the 10 000km per 500g of honey. Apparently no account taken that bees collect nectar and not honey, even though that same distance (or much more) may be flown to effectively collect nothing more than the colony running needs. Both these and other factors would, I am sure, have been used to estimate the distance travelled per jar of honey produced.
The accurate hive weight, at any particular time, will not reflect all, either. In peak foraging time we hear of a super of honey in less than a week. That is, say, 10kg (rounded estimate); now that 10 kg will be associated with a rather different weight of collected nectar, the excess water (over around, but likely less than 20% in honey) may be anywhere from around 5kg up to around 25kg, and is driven off from the hive during the time for filling the super. It would need quite a bit of modelling to estimate the actual weight of bees in the hive to the nearest one kg, and with very stable foraging rates, too.
Imagine weighing the hive in the evening after a superb foraging day and then reweighing a little before a 'late start' the following day. One could easily be out by the odd 10 000 bees.
These statistics have been carefully assessed/estimated by those doing continuous monitoring of hives over a long period of time. A single measurement is not too enlightening as to the actual truth, especially in a fickle climate like the UK has.
Rose Cottage is correct with all the other variables and these clearly demonstrate the inadequacies of doing it yourself. 20k foragers could indicate 40k hive population (50% foragers or as much as 65k (at 30%).
Indeed, the population is not static either. It could easily be changing by well over a thousand daily, perhaps two thousand (or more) at peak change rate (fewer forager attrition but peak emergence from brood).
All this lot demonstrates really, is the futility of counting bees. A guess to the nearest 5 thousand is good, during the summer; the indication of an expanding or contracting hive population is far more useful. After all, most of those bees will be dead in 6 weeks time during the summer months.