smash.Model.set_rr_initial_states#

Model.set_rr_initial_states(key, value)[source]#

Set the values of a Model rainfall-runoff initial state.

This method performs an in-place operation on the Model object.

Parameters:
keystr

The name of the rainfall-runoff initial state.

valuefloat or numpy.ndarray

The value(s) to set to the rainfall-runoff initial state. If the value is a numpy.ndarray, its shape must be broadcastable into the rainfall-runoff initial state shape.

See also

Model.get_rr_initial_states

Get the values of a Model rainfall-runoff initial state.

Model.rr_initial_states

Model rainfall-runoff initial states.

Examples

>>> from smash.factory import load_dataset
>>> setup, mesh = load_dataset("cance")
>>> model = smash.Model(setup, mesh)

Set a specific value to a rainfall-runoff initial state grid

>>> model.set_rr_initial_states("hp", 0.22)

Access its value

>>> model.get_rr_initial_states("hp")
array([[0.22, 0.22, 0.22, 0.22, 0.22, 0.22, 0.22, 0.22, 0.22, 0.22, 0.22,
        0.22, 0.22, 0.22, 0.22, 0.22, 0.22, 0.22, 0.22, 0.22, 0.22, 0.22,
        ...
        0.22, 0.22, 0.22, 0.22, 0.22, 0.22]], dtype=float32)

Set a grid with a shape equivalent to the rainfall-runoff initial state filled with random values between 0 and 1

Get the rainfall-runoff initial state shape

>>> shape = model.get_rr_initial_states("hp").shape
>>> shape
(28, 28)

Generate the random grid

>>> np.random.seed(99)
>>> random_arr = np.random.rand(*shape)
>>> random_arr
array([[6.72278559e-01, 4.88078399e-01, 8.25495174e-01, 3.14463876e-02,
        8.08049963e-01, 5.65617420e-01, 2.97622499e-01, 4.66957205e-02,
        ...
        8.83966213e-01, 3.73980927e-01, 2.98742432e-01, 8.37281270e-01]])

Set to the rainfall-runoff initial state the random grid

>>> model.set_rr_initial_states("hp", random_arr)
>>> model.get_rr_initial_states("hp")
array([[6.72278559e-01, 4.88078399e-01, 8.25495174e-01, 3.14463876e-02,
        8.08049963e-01, 5.65617420e-01, 2.97622499e-01, 4.66957205e-02,
        ...
        8.83966213e-01, 3.73980927e-01, 2.98742432e-01, 8.37281270e-01]],
      dtype=float32)

Note

This method is equivalent to directly slicing the rr_initial_states.values array (as shown below) and change the values but is simpler and safer to use

Access the rainfall-runoff initial state keys

>>> model.rr_initial_states.keys
array(['hi', 'hp', 'ht', 'hlr'], dtype='<U3')

Get the index of the rainfall-runoff initial state 'hp'

>>> ind = np.argwhere(model.rr_initial_states.keys == "hp").item()
>>> ind
1

Slice the rr_initial_states.values array on the last axis and change its values

>>> model.rr_initial_states.values[..., ind] = 0.56
>>> model.rr_initial_states.values[..., ind]
array([[0.56, 0.56, 0.56, 0.56, 0.56, 0.56, 0.56, 0.56, 0.56, 0.56, 0.56,
        0.56, 0.56, 0.56, 0.56, 0.56, 0.56, 0.56, 0.56, 0.56, 0.56, 0.56,
        ...
        0.56, 0.56, 0.56, 0.56, 0.56, 0.56]], dtype=float32)

Warning

In that case, there’s no problem to set the value 0.56 to the rainfall-runoff initial state 'hp', but each rainfall-runoff initial state has a feasibility domain, and that outside this domain, the model cannot run. For example, the feasibility domain of 'hp' is \(]0, 1[\).

Trying to set a value greater than 1 to the rainfall-runoff initial state 'hp' without the setter

>>> model.rr_initial_states.values[..., ind] = 21
>>> model.rr_initial_states.values[..., ind]
array([[21., 21., 21., 21., 21., 21., 21., 21., 21., 21., 21.,
        21., 21., 21., 21., 21., 21., 21., 21., 21., 21., 21.,
        ...
        21., 21., 21., 21., 21., 21.]], dtype=float32)

No particular problem doing this but trying with the setter

>>> model.set_rr_initial_states("hp", 21)
ValueError: Invalid value for model rr_initial_state 'hp'. rr_initial_state domain [21, 21] is not
included in the feasible domain ]0, 1[

Finally, trying to run the Model with a value greater than 1 set to the rainfall-runoff initial state 'hp' leads to the same error

>>> model.forward_run()
ValueError: Invalid value for model rr_initial_state 'hp'. rr_initial_state domain [21, 21] is not
included in the feasible domain ]0, 1[
On this page
Show Source