do ABC on number of patches

[petrelharp]

We would like to get the posterior distribution on four parameters (prior distributions in parentheses):

• P_D(1/2 to 1/64)
• DISPERSAL_SIGMA (0.2 to 2.0)
• POP_SIZE, (40 to 320)
• YEAR_SHAPE (1 to 2)

by doing ABC (maybe using abc) with the observed value being "number of occupied patches"; this is calculated in the SLiM script and output to a log file; our "data" is that the actual number of occupied patches today is "between 100 and 500". (So, maybe that means our data is a noisy observation of 250 patches with a SD of100?)

[petrelharp]

Note: in this round of ABC, we should turn off tree seq recording.