Improve surface spawn width estimates

[grinnellm]

Improve width estimates for Surface spawn surveys. We have not kept up with Pool/Bed information on recent spawns and the methods to determine pools/beds are not clear. This issue is related to #9. For example, one approach could be to:

• Use the observed width if present.
• If not, get the width for that location if there was a dive survey at some point.
• If there are no dives at that location then:
  • Calculate the average width of spawns within X km, or
  • Calculate the average width of the closest Y number of spawns.