Source code for pygmm.kempton_stewart_2006

"""Kempton and Stewart (2006, :cite:`kempton2006`) duration model."""

import numpy as np

from . import model

__author__ = ""




class KemptonStewart2006(model.Model):
    """Kempton and Stewart (2006, :cite:`kempton2006`) duration model.

    Parameters
    ----------
    scenario : :class:`pygmm.model.Scenario`
        earthquake scenario

    """

    NAME = "Kempton Stewart (2006)"
    ABBREV = "KS06"

    PARAMS = [
        model.NumericParameter("mag", True, 5, 7.6),
        model.NumericParameter("dist_rup", True, 0, 200),
        model.NumericParameter("v_s30", True, 200, 1000),
    ]



    def __init__(self, scenario):
        super().__init__(scenario)

        # scenario
        s = self._scenario

        # stress drop indices for duration measures
        stress_drop = np.exp(
            [
                6.02,  # D5-75a
                2.79 + 0.82 * (s.mag - 6),  # D5-95a
                5.46,  # D5-75v
                1.53 + 1.34 * (s.mag - 6),
            ]
        )  # D5-95v

        # source duration
        f_0 = self._source_dur(stress_drop)

        # path duration
        f_1 = np.array([0.07, 0.15, 0.10, 0.15]) * s.dist_rup

        # site duration
        f_2 = (
            np.array([0.82, 3.00, 1.40, 3.99])
            + np.array([-0.0013, -0.0041, -0.0022, -0.0062]) * s.v_s30
        )

        # total druation
        self._ln_dur = np.log(f_0 + f_1 + f_2)

        # aleatory standard deviation
        self._std_err = np.array([0.57, 0.47, 0.66, 0.52])


    @property
    def duration(self):
        return KemptonStewart2006._as_recarray(np.exp(self._ln_dur))

    @property
    def std_err(self):
        return KemptonStewart2006._as_recarray(self._std_err)

    @staticmethod
    def _as_recarray(values):
        return np.rec.fromarrays(
            values,
            names=[
                "D_5t75a",
                "D_5t95a",
                "D_5t75v",
                "D_5t95v",
            ],
        )

    def _source_dur(self, stress_drop):
        # seismic moment
        moment = 10 ** (1.5 * self.scenario.mag + 16.05)
        return (stress_drop / moment) ** (-1 / 3) / (4.9e6 * 3.2)