add precipitation weighting before fitting in prepsources

[courtiol]

There are 2 ways of weighing by precipitation amount:

• before fitting the isoscapes
• after fitting isoscapes per month
  Both methods have pros and cons so we should offer both method.
  So far, none work fully: "before" is not implemented and "after" does not allow for calibration.
  So let us implement the "before" method which is easy and what GNIP does.

[courtiol]

There is now a draft of how to do that in the bookdown.
But this remains very clunky, so I think that we should automate the method 1 in prepsources() and perhaps ditch the method 2 altogether from the bookdown and the examples.

[courtiol]

If you do think this issue matters to you, please add a thumb up!

[courtiol]

Short discussion with David Soto:

• method 1 seems best
• weighing should be done within each year separately (sum(w) = 1 for each year)
• mean isoscape value should represent the mean of the yearly weighted means
• residual variance should represent the variance between the yearly weighted means

This would approximate the isotopic signature of a watershed.

[courtiol]

Now described in the devel version of the bookdown (which will be uploaded soon).