What is the number of influenza strains occurring at a given time?

Question

My question was initiated by reading on mock-up influenza vaccines. I understand that the manufacturer pre-prepares a certain vaccine and tries to get it tested and ready before it is actually needed, by taking a novel virus strain that has infected a few people, and could potentially cause a pandemic.

My questions:

• How many different influenza strains are there at any given time at one geographical spot (say, one US state? - narrow down at will). I always thought there would be just one novel strain appearing each year, explaining why people get sick time and again of the 'same' virus. However, the mock-up virus vaccine production seems to imply there are various strains occurring at once?

• Do influenza strains vary on an annual basis, or is it a continuous process (like evolution), and is a new strain re-named after it becomes potentially dangerous (like speciation?) - i.e., what is an influenza strain? It seems, given the vaccine preps, that various strains occur at once, and strains survive for extended periods? Any hard evidence for these indirect assumptions?

And as a side -

• Do vaccine manufacturers prepare a whole bunch of mock-up influenza vaccines on a yearly basis just awaiting to identify a dangerous strain and start manufacturing it once people get sick? Or are mock-up vaccines combined into a cocktail mix containing many strains and are these cocktails re-vamped annually? Or is a yearly shot necessary because immunity only lasts a year?
  ^{I might have to re-post this last question at Health.SE as a dedicated question, but since related to above I prefer to leave it here for now. Grab me in chat if need be!}

Answer 1

Antigenic drift in the coding portions of the genes that involve antibody binding sites accounts for the seasonal variability of flu strains, i.e. they mutated enough that the antigens require new antibodies, hence a new strain. The process is as you said constant, like evolution, and in an analysis of pandemic H1N1 the authors are able to note 1 - 18.1 mutations per genome per replication cycle (ref). Because of this, flu vaccines tend to have more A strains than B strains due to the greater mutability of influenza A, at approx. 30% more mutations per site per infectious cycle (ref). At current It will take me time to research the remaining questions.

Interestingly enough, however, and I've never looked at this but CDC offers weekly influenza reports that detail by region what they're finding (CDC):

Source: CDC's so-called FluView

Beyond that, I'm pretty sure they take this weekly data, see what strain is on the rise in hospitals, and manufacture the vaccine to mirror the "popular" cocktail of strains.