Allow frac_P and frac_k to vary?

[bpbond]

@kendalynnm

Here are our eight 'problem' samples:

I tried allowing frac_P and frac_k to vary, i.e. for the optimizer to fit their values within constraints (0.001-0.1 and 0.9-0.99 respectively). This greatly improves the model fit for five of our eight problem samples.

Dashed line is old model (not allowing the fractionation constants to vary) and solid line is the new one:

[bpbond]

So, interesting I thought; let's talk about this next week!

[kendalynnm]

Very interesting, tomorrow let's look at this and the dropped-first-timestep fits. I'm curious how divergent the final P and k values are for the three fit types (variable fractionation, dropped-first-timestep, and original).

[bpbond]

Outdated!