Finetune kinematice_wave

[kordejong]

• Try to make the numerical scheme converge in all valid cases. Currently it still doesn't in certain cases (see comment in code)
• Use domain policy to verify all input values are within their respective valid domains.
• Revisit handling of extraction of water, if needed
• Update tests

[saeb-faraji-gargari]

-Just as an idea, the Newton-Raphson method may diverge when lateral_inflow is negative. In this case, I would suggest using discharge_guess as new_discharge (new_discharge = discharge_guess;) since discharge_guess is the solution to the linear discretized version of the kinematic-wave equations.

-The following criterion needs to be satisfied for the convergence of the Newton-Raphson method:
| fq * d2fq| < |dfq|^2

Ref: https://math.stackexchange.com/questions/3136446/condition-for-convergence-of-newton-raphson-method

When lateral_inflow < 0, − _known_terms (-C in Chow's book) becomes positive, may lead to a large positive value for fq. This can violate the convergence criteria of the Newton-Raphson method. This issue can be considered in the future to improve the kinematic-wave operator.