pastas.rfunc.HantushWellModel.variance_gain#

static pastas.rfunc.HantushWellModel.variance_gain(A: float, b: float, var_A: float, var_b: float, cov_Ab: float, r: float = 1.0, log_b: bool = True) → float | pastas.typing.ArrayLike#

Calculate variance of the gain from parameters A and b.

Variance of the gain is calculated based on propagation of uncertainty using optimal values, the variances of A and b and the covariance between A and b.

Notes

Estimated variance can be biased for non-linear functions as it uses truncated series expansion.

Parameters:

A (float) – optimal value of parameter A, (e.g. ml.parameters.optimal).
b (float) – optimal value of parameter b, (e.g. ml.parameters.optimal).
var_A (float) – variance of parameter A, can be obtained from the diagonal of the covariance matrix (e.g. ml.solver.pcov).
var_b (float) – variance of parameter b, can be obtained from the diagonal of the covariance matrix (e.g. ml.solver.pcov).
cov_Ab (float) – covariance between A and b, can be obtained from the covariance matrix ( e.g. ml.solver.pcov).
r (float or array_like, optional) – distance(s) between observation well and stress(es), default value is 1.0.
log_b (bool, optional) – indicates if parameter b is log10 transformed. Default is True.

Returns:

var_gain – variance of the gain calculated based on propagation of uncertainty of parameters A and b.

Return type:

float or array_like