Add loading efficiency to xsection plot

[MattBrst]

I would like to add the loading efficiency to the cross section plots.

Example from the benchmark with the following line of code:

ml_leff[0].plots.xsection(params=True, names=False, labels=False, sep="\n")

So basically non-zero loading efficiencies for aquifers are plotted.

[dbrakenhoff]

Good addition, but could you modify it to use the symbol $\beta$, similar to Mark's paper? Looks a bit cleaner than load eff written out, though it is a bit more obscure as to what that means.

We could consider a small list of symbols somewhere to help users understand what's what. Or at least have the loading_efficiency notebook mention $\beta$.

[mbakker7]

And to be one step ahead in the process, the loading efficiency of the leaky layer is on its way. So can we include that right away as well?

[mbakker7]

In the mathematical formulation, we use \beta for the loading efficiency. In the code it is called leffaq and leffll. I don't think that betaaq and betall is better (definitely less readable). But I think the $\beta$ symbol in the xsection plots is better, since the other parameters are also symbols, like $k_h$ and $S_s$.
Once #74 has been merged, we can also add $\beta$ to the leaky layers in this PR.

[MattBrst]

@dbrakenhoff:

• I modified the PR to use the β symbol for both aquifers and leaky layers. This should work even when leffll is not yet defined in the model.
• I had to add the labeling code twice (once in plots.py for the normal models and once in inhom1d.py for the 1D models).

@mbakker7:

• I had to make some changes in aquifer_parameters.py for leffaq, and would guess you will have to do the same for the leffll.

Currently my plot from the benchmark River1D becomes:

And for an 'normal' model with the code below:

import timflow.transient as tft

ml = tft.ModelMaq(
    kaq=50,
    z=[7, 5, 2, 1, 0],
    c=100,
    leffaq=[0.3, 0.1],
    leffll=0.7,
    topboundary="semi",
    tmin=0.1,
    tmax=100,
)

ml.plots.xsection(params=True)

[mbakker7]

Pictures look nice! But I think it should work w/o any changes to aquifer_parameters, I would think. Let's first merge #74, which is almost done (working on some benchmark notebooks) and then see what needs to be done.

[mbakker7]

When Sll=0, Timflow substitutes Sll=1e-20 (as it cannot deal with exactly zero). Maybe this can be done more elegant in the code. But that is what it is for now. In the xsection plot, can you check if Sll=1e-20 and print 0 instead?