Merge pull request #3545 from gem/shakemap

Shakemap to GMFs

openquake/hazardlib/correlation.py

@@ -111,52 +111,38 @@ def __init__(self, vs30_clustering):
         self.cache = {}  # imt -> correlation model
 
     def _get_correlation_matrix(self, sites, imt):
-        """
-        Calculate correlation matrix for a given sites collection.
-
-        Correlation depends on spectral period, Vs 30 clustering behaviour
-        and distance between sites.
+        distances = sites.mesh.get_distance_matrix()
+        return jbcorrelation(distances, imt, self.vs30_clustering)
 
-        Parameters are the same as for
-        :meth:`BaseCorrelationModel.get_lower_triangle_correlation_matrix`.
+    def get_lower_triangle_correlation_matrix(self, sites, imt):
         """
-        distances = sites.mesh.get_distance_matrix()
-        return self._get_correlation_model(distances, imt)
+        See :meth:`BaseCorrelationModel.get_lower_triangle_correlation_matrix`.
+        """
+        return numpy.linalg.cholesky(self._get_correlation_matrix(sites, imt))
 
-    def _get_correlation_model(self, distances, imt):
+
+def jbcorrelation(distances, imt, vs30_clustering):
         """
-        Returns the correlation model for a set of distances, given the
-        appropriate period
+        Returns the Jayaram-Baker correlation model.
 
         :param numpy.ndarray distances:
             Distance matrix
-
-        :param float period:
-            Period of spectral acceleration
+        :param imt:
+            Intensity Measure Type (PGA or SA)
+        :param vs30_clustering:
+            flag
         """
-        if isinstance(imt, SA):
-            period = imt.period
-        else:
-            assert isinstance(imt, PGA), imt
-            period = 0
-
         # formulae are from page 1700
-        if period < 1:
-            if not self.vs30_clustering:
+        if imt.period < 1:
+            if not vs30_clustering:
                 # case 1, eq. (17)
-                b = 8.5 + 17.2 * period
+                b = 8.5 + 17.2 * imt.period
             else:
                 # case 2, eq. (18)
-                b = 40.7 - 15.0 * period
+                b = 40.7 - 15.0 * imt.period
         else:
             # both cases, eq. (19)
-            b = 22.0 + 3.7 * period
+            b = 22.0 + 3.7 * imt.period
 
         # eq. (20)
         return numpy.exp((- 3.0 / b) * distances)
-
-    def get_lower_triangle_correlation_matrix(self, sites, imt):
-        """
-        See :meth:`BaseCorrelationModel.get_lower_triangle_correlation_matrix`.
-        """
-        return numpy.linalg.cholesky(self._get_correlation_matrix(sites, imt))

openquake/hazardlib/imt.py

@@ -106,6 +106,7 @@ class PGA(_IMT):
     Peak ground acceleration during an earthquake measured in units
     of ``g``, times of gravitational acceleration.
     """
+    period = 0.0
 
 
 class PGV(_IMT):

openquake/hazardlib/shakemap.py

@@ -0,0 +1,223 @@
+#  -*- coding: utf-8 -*-
+#  vim: tabstop=4 shiftwidth=4 softtabstop=4
+
+#  Copyright (c) 2018, GEM Foundation
+
+#  OpenQuake is free software: you can redistribute it and/or modify it
+#  under the terms of the GNU Affero General Public License as published
+#  by the Free Software Foundation, either version 3 of the License, or
+#  (at your option) any later version.
+
+#  OpenQuake is distributed in the hope that it will be useful,
+#  but WITHOUT ANY WARRANTY; without even the implied warranty of
+#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+#  GNU Affero General Public License for more details.
+
+#  You should have received a copy of the GNU Affero General Public License
+#  along with OpenQuake.  If not, see <http://www.gnu.org/licenses/>.
+from urllib.request import urlopen
+import math
+import numpy
+from scipy.stats import truncnorm
+from scipy import interpolate
+
+from openquake.hazardlib import geo, site, imt, correlation
+from openquake.hazardlib.shakemapconverter import get_shakemap_array
+
+F32 = numpy.float32
+SHAKEMAP_URL = 'http://shakemap.rm.ingv.it/shake/{}/download/{}.xml'
+PCTG = 100  # percent of g, the gravity acceleration
+
+
+def get_sitecol_shakemap(array_or_id, sitecol=None, assoc_dist=None):
+    """
+    :param array_or_id: shakemap ID or full shakemap array
+    :param sitecol: SiteCollection used to reduce the shakemap
+    :param assoc_dist: association distance
+    :returns: a pair (filtered site collection, filtered shakemap)
+    """
+    if isinstance(array_or_id, int):
+        with urlopen(SHAKEMAP_URL.format(array_or_id, 'grid')) as f1, \
+             urlopen(SHAKEMAP_URL.format(array_or_id, 'uncertainty')) as f2:
+            array = get_shakemap_array(f1, f2)
+    else:
+        array = array_or_id
+    if sitecol is None:  # extract the sites from the shakemap
+        return site.SiteCollection.from_shakemap(array), array
+
+    # associate the shakemap to the site collection
+    # TODO: forbid IDL crossing
+    bbox = (array['lon'].min(), array['lat'].min(),
+            array['lon'].max(), array['lat'].max())
+    sitecol = sitecol.within_bb(bbox)
+    data = geo.utils.GeographicObjects(array)
+    dic = data.assoc(sitecol, assoc_dist)
+    sids = sorted(dic)
+    return sitecol.filtered(sids), numpy.array([dic[sid] for sid in sids])
+
+
+# Here is the explanation of USGS for the units they are using:
+# PGA = peak ground acceleration (percent-g)
+# PSA03 = spectral acceleration at 0.3 s period, 5% damping (percent-g)
+# PSA10 = spectral acceleration at 1.0 s period, 5% damping (percent-g)
+# PSA30 = spectral acceleration at 3.0 s period, 5% damping (percent-g)
+# STDPGA = the standard deviation of PGA (natural log of percent-g)
+
+
+def spatial_correlation_array(dmatrix, imts, correl='spatial',
+                              vs30clustered=True):
+    """
+    :param dmatrix: distance matrix of shape (N, N)
+    :param imts: M intensity measure types
+    :param correl: 'no correlation', 'full correlation', 'spatial'
+    :param vs30clustered: flag, True by default
+    :returns: array of shape (M, N, N)
+    """
+    n = len(dmatrix)
+    corr = numpy.zeros((len(imts), n, n))
+    for imti, im in enumerate(imts):
+        if correl == 'no correlation':
+            corr[imti] = numpy.zeros((n, n))
+        if correl == 'full correlation':
+            corr[imti] = numpy.eye(n)
+        elif correl == 'spatial':
+            corr[imti] = correlation.jbcorrelation(dmatrix, im, vs30clustered)
+    return corr
+
+
+def spatial_covariance_array(stddev, corrmatrices):
+    """
+    :param stddev: array of shape (M, N)
+    :param corrmatrices: array of shape (M, N, N)
+    :returns: an array of shape (M, N, N)
+    """
+    # this depends on sPGA, sSa03, sSa10, sSa30
+    M, N = corrmatrices.shape[:2]
+    matrices = []
+    for i, std in enumerate(stddev):
+        covmatrix = numpy.zeros((N, N))
+        for j in range(N):
+            for k in range(N):
+                covmatrix[j, k] = corrmatrices[i, j, k] * std[j] * std[k]
+        matrices.append(covmatrix)
+    return numpy.array(matrices)
+
+
+def cross_correlation_matrix(imts, corr='cross'):
+    """
+    :param imts: M intensity measure types
+    :param corr: 'no correlation', 'full correlation' or 'cross'
+    """
+    # if there is only PGA this is a 1x1 identity matrix
+    M = len(imts)
+    cross_matrix = numpy.zeros((M, M))
+    for i, im in enumerate(imts):
+        T1 = im.period or 0.05
+
+        for j in range(M):
+            T2 = imts[j].period or 0.05
+            if i == j:
+                cross_matrix[i, j] = 1
+            else:
+                Tmax = max([T1, T2])
+                Tmin = min([T1, T2])
+                II = 1 if Tmin < 0.189 else 0
+                if corr == 'no correlation':
+                    cross_matrix[i, j] = 0
+                if corr == 'full correlation':
+                    cross_matrix[i, j] = 0.99999
+                if corr == 'cross':
+                    cross_matrix[i, j] = 1 - math.cos(math.pi / 2 - (
+                        0.359 + 0.163 * II * math.log(Tmin / 0.189)
+                    ) * math.log(Tmax / Tmin))
+
+    return cross_matrix
+
+
+def amplify_gmfs(imts, vs30s, gmfs):
+    """
+    Amplify the ground shaking depending on the vs30s
+    """
+    n = len(vs30s)
+    for i, im in enumerate(imts):
+        for iloc in range(n):
+            gmfs[i * n + iloc] = amplify_ground_shaking(
+                im.period, vs30s[iloc], gmfs[i * n + iloc])
+    return gmfs
+
+
+def amplify_ground_shaking(T, vs30, gmvs):
+    """
+    :param T: period
+    :param vs30: velocity
+    :param gmvs: ground motion values for the current site
+    """
+    interpolator = interpolate.interp1d(
+        [-1, 0.1, 0.2, 0.3, 0.4, 100],
+        [(760 / vs30)**0.35,
+         (760 / vs30)**0.35,
+         (760 / vs30)**0.25,
+         (760 / vs30)**0.10,
+         (760 / vs30)**-0.05,
+         (760 / vs30)**-0.05],
+    ) if T <= 0.3 else interpolate.interp1d(
+        [-1, 0.1, 0.2, 0.3, 0.4, 100],
+        [(760 / vs30)**0.65,
+         (760 / vs30)**0.65,
+         (760 / vs30)**0.60,
+         (760 / vs30)**0.53,
+         (760 / vs30)**0.45,
+         (760 / vs30)**0.45],
+    )
+    return interpolator(gmvs) * gmvs
+
+
+def cholesky(spatial_cov, cross_corr):
+    """
+    Decompose the spatial covariance and cross correlation matrices
+    """
+    M, N = spatial_cov.shape[:2]
+    L = numpy.array([numpy.linalg.cholesky(spatial_cov[i]) for i in range(M)])
+    LLT = []
+    for i in range(M):
+        row = [numpy.dot(L[i], L[j].T) * cross_corr[i, j] for j in range(M)]
+        for j in range(N):
+            singlerow = numpy.zeros(M * N)
+            for i in range(M):
+                singlerow[i * N:(i + 1) * N] = row[i][j]
+            LLT.append(singlerow)
+    return numpy.linalg.cholesky(numpy.array(LLT))
+
+
+def to_gmfs(shakemap, site_effects, trunclevel, num_gmfs, seed):
+    """
+    :returns: an array of GMFs of shape (N, G) and dtype imt_dt
+    """
+    std = shakemap['std']
+    imts = [imt.from_string(name) for name in std.dtype.names]
+    val = {imt: numpy.log(shakemap['val'][imt] / PCTG) - std[imt] ** 2 / 2.
+           for imt in std.dtype.names}
+    dmatrix = geo.geodetic.distance_matrix(shakemap['lon'], shakemap['lat'])
+    spatial_corr = spatial_correlation_array(dmatrix, imts)
+    stddev = [std[imt] for imt in std.dtype.names]
+    spatial_cov = spatial_covariance_array(stddev, spatial_corr)
+    cross_corr = cross_correlation_matrix(imts)
+    M, N = spatial_corr.shape[:2]
+    mu = numpy.array([numpy.ones(num_gmfs) * val[imt][j]
+                      for imt in std.dtype.names for j in range(N)])
+    L = cholesky(spatial_cov, cross_corr)
+    Z = truncnorm.rvs(-trunclevel, trunclevel, loc=0, scale=1,
+                      size=(M * N, num_gmfs), random_state=seed)
+    gmfs = numpy.exp(numpy.dot(L, Z) + mu)
+    if site_effects:
+        gmfs = amplify_gmfs(imts, shakemap['vs30'], gmfs) * 0.8
+
+    arr = numpy.zeros((N, num_gmfs), std.dtype)
+    for i, im in enumerate(std.dtype.names):
+        arr[im] = gmfs[i * N:(i + 1) * N]
+    return arr
+
+"""
+here is an example for Tanzania:
+https://earthquake.usgs.gov/archive/product/shakemap/us10006nkx/us/1480920466172/download/grid.xml
+"""

openquake/hazardlib/shakemapconverter.py

@@ -24,7 +24,7 @@
 
 F32 = numpy.float32
 SHAKEMAP_FIELDS = set(
-    'LON LAT SVEL PGA PSA03 PSA10 PSA30 STDPGA STDPSA03 STDPSHA10 STDPSA30'
+    'LON LAT SVEL PGA PSA03 PSA10 PSA30 STDPGA STDPSA03 STDPSA10 STDPSA30'
     .split())
 FIELDMAP = {
     'LON': 'lon',

openquake/hazardlib/tests/shakemap/ghorka_grid.xml

@@ -0,0 +1,28 @@
+<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
+<shakemap_grid xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://earthquake.usgs.gov/eqcenter/shakemap" xsi:schemaLocation="http://earthquake.usgs.gov http://earthquake.usgs.gov/eqcenter/shakemap/xml/schemas/shakemap.xsd" event_id="us20002926" shakemap_id="us20002926" shakemap_version="9" code_version="3.5.1440" process_timestamp="2015-07-02T22:50:42Z" shakemap_originator="us" map_status="RELEASED" shakemap_event_type="ACTUAL">
+<event event_id="us20002926" magnitude="7.8" depth="8.22" lat="28.230500" lon="84.731400" event_timestamp="2015-04-25T06:11:25UTC" event_network="us" event_description="NEPAL" />
+<grid_specification lon_min="81.731400" lat_min="25.587500" lon_max="87.731400" lat_max="30.873500" nominal_lon_spacing="0.016667" nominal_lat_spacing="0.016675" nlon="361" nlat="318" />
+<event_specific_uncertainty name="pga" value="0.000000" numsta="" />
+<event_specific_uncertainty name="pgv" value="0.000000" numsta="" />
+<event_specific_uncertainty name="mi" value="0.000000" numsta="" />
+<event_specific_uncertainty name="psa03" value="0.000000" numsta="" />
+<event_specific_uncertainty name="psa10" value="0.000000" numsta="" />
+<event_specific_uncertainty name="psa30" value="0.000000" numsta="" />
+<grid_field index="1" name="LON" units="dd" />
+<grid_field index="2" name="LAT" units="dd" />
+<grid_field index="3" name="PGA" units="pctg" />
+<grid_field index="4" name="PGV" units="cms" />
+<grid_field index="5" name="MMI" units="intensity" />
+<grid_field index="6" name="PSA03" units="pctg" />
+<grid_field index="7" name="PSA10" units="pctg" />
+<grid_field index="8" name="PSA30" units="pctg" />
+<grid_field index="9" name="STDPGA" units="ln(pctg)" />
+<grid_field index="10" name="URAT" units="" />
+<grid_field index="11" name="SVEL" units="ms" />
+<grid_data>
+81.7314 30.8735 0.44 2.21 3.83 1.82 2.8 1.26 0.53 1 400.758
+81.7481 30.8735 0.47 2.45 3.88 1.99 3.09 1.41 0.52 1 352.659
+81.7647 30.8735 0.47 2.4 3.88 1.97 3.04 1.38 0.52 1 363.687
+81.7814 30.8735 0.52 2.78 3.96 2.23 3.51 1.64 0.5 1 301.17
+</grid_data>
+</shakemap_grid>

openquake/hazardlib/tests/shakemap/ghorka_uncertainty.xml

@@ -0,0 +1,37 @@
+<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
+<shakemap_grid xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://earthquake.usgs.gov/eqcenter/shakemap" xsi:schemaLocation="http://earthquake.usgs.gov http://earthquake.usgs.gov/eqcenter/shakemap/xml/schemas/shakemap.xsd" event_id="us20002926" shakemap_id="us20002926" shakemap_version="9" code_version="3.5.1440" process_timestamp="2015-07-02T22:50:42Z" shakemap_originator="us" map_status="RELEASED" shakemap_event_type="ACTUAL">
+<event event_id="us20002926" magnitude="7.8" depth="8.22" lat="28.230500" lon="84.731400" event_timestamp="2015-04-25T06:11:25UTC" event_network="us" event_description="NEPAL" />
+<grid_specification lon_min="81.731400" lat_min="25.587500" lon_max="87.731400" lat_max="30.873500" nominal_lon_spacing="0.016667" nominal_lat_spacing="0.016675" nlon="361" nlat="318" />
+<event_specific_uncertainty name="pga" value="0.000000" numsta="" />
+<event_specific_uncertainty name="pgv" value="0.000000" numsta="" />
+<event_specific_uncertainty name="mi" value="0.000000" numsta="" />
+<event_specific_uncertainty name="psa03" value="0.000000" numsta="" />
+<event_specific_uncertainty name="psa10" value="0.000000" numsta="" />
+<event_specific_uncertainty name="psa30" value="0.000000" numsta="" />
+<grid_field index="1" name="LON" units="dd" />
+<grid_field index="2" name="LAT" units="dd" />
+<grid_field index="3" name="STDPGA" units="ln(pctg)" />
+<grid_field index="4" name="STDPGV" units="ln(cms)" />
+<grid_field index="5" name="STDMMI" units="intensity" />
+<grid_field index="6" name="STDPSA03" units="ln(pctg)" />
+<grid_field index="7" name="STDPSA10" units="ln(pctg)" />
+<grid_field index="8" name="STDPSA30" units="ln(pctg)" />
+<grid_field index="9" name="GMPE_INTER_STDPGA" units="ln(pctg)" />
+<grid_field index="10" name="GMPE_INTRA_STDPGA" units="ln(pctg)" />
+<grid_field index="11" name="GMPE_INTER_STDPGV" units="ln(cms)" />
+<grid_field index="12" name="GMPE_INTRA_STDPGV" units="ln(cms)" />
+<grid_field index="13" name="GMPE_INTER_STDMMI" units="intensity" />
+<grid_field index="14" name="GMPE_INTRA_STDMMI" units="intensity" />
+<grid_field index="15" name="GMPE_INTER_STDPSA03" units="ln(pctg)" />
+<grid_field index="16" name="GMPE_INTRA_STDPSA03" units="ln(pctg)" />
+<grid_field index="17" name="GMPE_INTER_STDPSA10" units="ln(pctg)" />
+<grid_field index="18" name="GMPE_INTRA_STDPSA10" units="ln(pctg)" />
+<grid_field index="19" name="GMPE_INTER_STDPSA30" units="ln(pctg)" />
+<grid_field index="20" name="GMPE_INTRA_STDPSA30" units="ln(pctg)" />
+<grid_data>
+81.7314 30.8735 0.53 0.56 0.72 0.57 0.66 0.73 0.25 0.47 0.24 0.5 0.16 0.7 0.26 0.51 0.33 0.57 0.45 0.58
+81.7481 30.8735 0.52 0.55 0.72 0.55 0.65 0.73 0.24 0.46 0.24 0.5 0.16 0.7 0.24 0.5 0.32 0.57 0.44 0.58
+81.7647 30.8735 0.52 0.55 0.72 0.56 0.66 0.73 0.24 0.46 0.24 0.5 0.16 0.7 0.24 0.5 0.33 0.57 0.44 0.58
+81.7814 30.8735 0.5 0.55 0.72 0.52 0.64 0.73 0.22 0.45 0.24 0.5 0.16 0.7 0.21 0.47 0.31 0.56 0.44 0.58
+</grid_data>
+</shakemap_grid>