Nonlinear storage constant units

[snorfalorpagus]

The nonlinear storage constant (Cq) has units "volume.time" (i.e. L^3.T).

The current docstring for this is incorrect: https://github.com/jetuk/pycatchmod/blob/d43438d9f8b9013c8241a8eebce81ca6ac0cd2a9/pycatchmod/_catchmod.pyx#L150

The excel version of catchmod stores this parameter with units cumec.days^2/km^2, i.e. L^3.T^-1.T^2.L^-2, which cancels to L.T. Before calculation, this is multiplied by area (L^2) to give the required units. In addition, the value is multiplied by 86.4, which converts m3/s to Ml/d.

We need to decide how to handle this.

[snorfalorpagus]

This is a little tricky to fix because of the way NonLinearStore is independent of SubCatchment. It's not clear where the best place to do the conversion is, especially if it needs to be robust to subsequent changes to both the constant, and the area.

[jetuk]

I think this is because the area is multiplied through between linear and non-linear storage components. Perhaps that multiplication can be done last, and then Cq doesn't need to be multiplied through by area.?