Here's my hypothesis, for what it's worth....
In its simplest terms, evolution works by two processes. First,
mutation and
reassortment provide genetic variation. Second,
selection favours the sets of genetic material that are most fit (i.e., will leave more progeny than other sets of genetic material). These processes work with all living things, whether they are influenza viruses or humans.
With respect to influenzas, wild birds of various species are the reservoir for all avian influenzas. However, these viruses tend to be of low pathogenicity (LPAI), causing only mild transient illnesses. Until this year, science had described only one outbreak of a high pathogenic avian influenza (HPAI) in wild birds in the absence of affected poultry populations - an
outbreak of HPAI in common terns in South Africa in the 1960s. Since its emergence in 1996, H5N1 had been isolated from a few wild birds that clearly obtained the virus from nearby infected domestic bird populations. However, these were local spill-over events, and the virus was not sustained.
So how do you get HPAI in poultry from LPAI in wild birds. My hypothesis, and I don't claim that this is original, when HPAI strains appear through mutation and/or reassortment, the conditions do not favour the HPAI strains over endemic LPAI strains (see figure below). In contrast, in the wildlife and poultry trade, the high turnover of susceptibles, poor sanitary conditions, and high stress fuel the fire, so to speak, and HPAI strains become dominant through natural selection. This is why HPAI H5N1 first appeared in the markets of southern China.
However, something has changed this year. HPAI H5N1 seems to be persisting in wild bird populations, at least long enough to get to
Mongolia and perhaps to Eastern Europe. Now this is totally speculation, but perhaps the evolutionary landscape for influenza is such that HPAI is unlikely to evolve in wild birds, but now that it has evolved in poultry and domestic waterfowl, it is capable of persisting in wild birds. It remains to be seen how long it will persist.