BMI Update may incorrectly assume a 1 hour time step

[hellkite500]

The Update function assumes a conversion of the precip variable from mm/hr to m/hr by simply accounting for the scale difference between mm and m. However, what happens when update_until modifies the dt for whatever reason to a fractional dt. Is this conversion/assumption still valid? A quick look through the model definitions implies that precipitation should be the value for the timestep -- timestep_rainfall_input_m. Just want to double check that there is nothing untoward happening here.

https://github.com/NOAA-OWP/cfe/blob/3617bda7b887ca283a971cc630fd76d890292db1/src/bmi_cfe.c#L970