An article via American Bee Journal's 'Extra' emails presents the treatment-free slow-collapse scenario, as modeled in a new computer simulation program, BEEHAVE.
"The first results of the model show that colonies infested with a common parasitic mite (varroa) can be much more vulnerable to food shortages. Effects within the first year can be subtle and might be missed by beekeepers during routine management. But the model shows that these effects build up over subsequent years leading to eventual failure of the colony, if it was not given an effective varroa treatment.
"BEEHAVE can also be used to investigate potential consequences of pesticide applications. For example, the BEEHAVE model can simulate the impact of increased loss of foragers. The results show that colonies may be more resilient to this forager loss than previously thought in the short-term, but effects may accumulate over years, especially when colonies are also limited by food supply.
"BEEHAVE simulations show that good food sources close to the hive will make a real difference to the colony and that lack of forage over extended periods leaves them vulnerable to other environmental factors. Addressing forage availability is critical to maintaining healthy hives and colonies over the long term."
I doubt that their model includes data from long-term treatment-free hives, where the microbial ecology is already well established, so there's reason to doubt their projections. But as I've been trying to ask here, what practical measures can beekeepers use to build up that ecology during the 2-3 year span of time that, according to simulations, treated packages and hives might be collapsing without further treatment? Having good forage available nearby seems to be one important factor.
(The link to ABJ's page for articles emailed in March doesn't yet list the one about Beehave, but it will soon. A webpage for the project is at http://www.beehave-model.net .)