1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

# http://pyrocko.org - GPLv3 

# 

# The Pyrocko Developers, 21st Century 

# ---|P------/S----------~Lg---------- 

from __future__ import absolute_import 

 

import os.path as op 

import math 

from collections import defaultdict 

 

import numpy as num 

 

from pyrocko import util, config 

from pyrocko import orthodrome as od 

from pyrocko.guts_array import Array 

from pyrocko.guts import Object, String 

from .util import get_download_callback 

 

 

PI = math.pi 

 

 

class Plate(Object): 

 

name = String.T( 

help='Name of the tectonic plate.') 

points = Array.T( 

dtype=num.float, shape=(None, 2), 

help='Points on the plate.') 

 

def max_interpoint_distance(self): 

p = od.latlon_to_xyz(self.points) 

return math.sqrt(num.max(num.sum( 

(p[num.newaxis, :, :] - p[:, num.newaxis, :])**2, axis=2))) 

 

def contains_point(self, point): 

return od.contains_point(self.points, point) 

 

def contains_points(self, points): 

return od.contains_points(self.points, points) 

 

 

class Boundary(Object): 

 

name1 = String.T() 

name2 = String.T() 

kind = String.T() 

points = Array.T(dtype=num.float, shape=(None, 2)) 

cpoints = Array.T(dtype=num.float, shape=(None, 2)) 

itypes = Array.T(dtype=num.int, shape=(None)) 

 

def split_types(self, groups=None): 

xyz = od.latlon_to_xyz(self.points) 

xyzmid = (xyz[1:] + xyz[:-1, :]) * 0.5 

cxyz = od.latlon_to_xyz(self.cpoints) 

d = od.distances3d(xyzmid[num.newaxis, :, :], cxyz[:, num.newaxis, :]) 

idmin = num.argmin(d, axis=0) 

itypes = self.itypes[idmin] 

 

if groups is None: 

groupmap = num.arange(len(self._index_to_type)) 

else: 

d = {} 

for igroup, group in enumerate(groups): 

for name in group: 

d[name] = igroup 

 

groupmap = num.array( 

[d[name] for name in self._index_to_type], 

dtype=num.int) 

 

iswitch = num.concatenate( 

([0], 

num.where(groupmap[itypes[1:]] != groupmap[itypes[:-1]])[0]+1, 

[itypes.size])) 

 

results = [] 

for ii in range(iswitch.size-1): 

if groups is not None: 

tt = [self._index_to_type[ityp] for ityp in num.unique( 

itypes[iswitch[ii]:iswitch[ii+1]])] 

else: 

tt = self._index_to_type[itypes[iswitch[ii]]] 

 

results.append((tt, self.points[iswitch[ii]:iswitch[ii+1]+1])) 

 

return results 

 

 

class Dataset(object): 

 

def __init__(self, name, data_dir, citation): 

self.name = name 

self.citation = citation 

self.data_dir = data_dir 

 

def fpath(self, filename): 

return op.join(self.data_dir, filename) 

 

def download_file( 

self, url, fpath, username=None, password=None, 

status_callback=None): 

 

util.download_file( 

url, fpath, username, password, status_callback=status_callback) 

 

 

class PlatesDataset(Dataset): 

pass 

 

 

class PeterBird2003(PlatesDataset): 

'''An updated digital model of plate boundaries.''' 

__citation = ''' 

Bird, Peter. "An updated digital model of plate boundaries." Geochemistry, 

Geophysics, Geosystems 4.3 (2003). 

''' 

 

def __init__( 

self, 

name='PeterBird2003', 

data_dir=None, 

raw_data_url=('https://mirror.pyrocko.org/peterbird.name/oldFTP/PB2002/%s')): # noqa 

 

if data_dir is None: 

data_dir = op.join(config.config().tectonics_dir, name) 

 

PlatesDataset.__init__( 

self, 

name, 

data_dir=data_dir, 

citation=self.__citation) 

 

self.raw_data_url = raw_data_url 

 

self.filenames = [ 

'2001GC000252_readme.txt', 

'PB2002_boundaries.dig.txt', 

'PB2002_orogens.dig.txt', 

'PB2002_plates.dig.txt', 

'PB2002_poles.dat.txt', 

'PB2002_steps.dat.txt'] 

 

self._full_names = None 

 

def full_name(self, name): 

if not self._full_names: 

fn = util.data_file(op.join('tectonics', 'bird2003_plates.txt')) 

 

with open(fn, 'r') as f: 

self._full_names = dict( 

line.strip().split(None, 1) for line in f) 

 

return self._full_names[name] 

 

def download_if_needed(self): 

for fn in self.filenames: 

fpath = self.fpath(fn) 

if not op.exists(fpath): 

self.download_file( 

self.raw_data_url % fn, fpath, 

status_callback=get_download_callback( 

'Downloading Bird 2003 plate database...')) 

 

def get_boundaries(self): 

self.download_if_needed() 

fpath = self.fpath('PB2002_steps.dat.txt') 

 

d = defaultdict(list) 

ntyp = 0 

type_to_index = {} 

index_to_type = [] 

with open(fpath, 'rb') as f: 

data = [] 

for line in f: 

t = line.split() 

s = t[1].lstrip(b':') 

name1 = str(s[0:2].decode('ascii')) 

name2 = str(s[3:5].decode('ascii')) 

kind = s[2] 

 

alon, alat, blon, blat = list(map(float, t[2:6])) 

mlat = (alat + blat) * 0.5 

dlon = ((blon - alon) + 180.) % 360. - 180. 

mlon = alon + dlon * 0.5 

typ = str(t[14].strip(b':*').decode('ascii')) 

 

if typ not in type_to_index: 

ntyp += 1 

type_to_index[typ] = ntyp - 1 

index_to_type.append(typ) 

 

ityp = type_to_index[typ] 

d[name1, kind, name2].append((mlat, mlon, ityp)) 

 

d2 = {} 

for k in d: 

d2[k] = ( 

num.array([l[:2] for l in d[k]], dtype=num.float), 

num.array([l[2] for l in d[k]], dtype=num.int)) 

 

fpath = self.fpath('PB2002_boundaries.dig.txt') 

boundaries = [] 

name1 = '' 

name2 = '' 

kind = '-' 

with open(fpath, 'rb') as f: 

data = [] 

for line in f: 

if line.startswith(b'***'): 

cpoints, itypes = d2[name1, kind, name2] 

boundaries.append(Boundary( 

name1=name1, 

name2=name2, 

kind=kind, 

points=num.array(data, dtype=num.float), 

cpoints=cpoints, 

itypes=itypes)) 

 

boundaries[-1]._type_to_index = type_to_index 

boundaries[-1]._index_to_type = index_to_type 

 

data = [] 

elif line.startswith(b' '): 

data.append(list(map(float, line.split(b',')))[::-1]) 

else: 

s = line.strip() 

name1 = str(s[0:2].decode('ascii')) 

name2 = str(s[3:5].decode('ascii')) 

kind = s[2] 

 

return boundaries 

 

def get_plates(self): 

self.download_if_needed() 

fpath = self.fpath('PB2002_plates.dig.txt') 

plates = [] 

name = '' 

with open(fpath, 'rb') as f: 

data = [] 

for line in f: 

if line.startswith(b'***'): 

plates.append(Plate( 

name=name, 

points=num.array(data, dtype=num.float))) 

 

data = [] 

elif line.startswith(b' '): 

data.append(list(map(float, line.split(b',')))[::-1]) 

else: 

name = str(line.strip().decode('ascii')) 

 

return plates 

 

 

class StrainRateDataset(Dataset): 

pass 

 

 

class GSRM1(StrainRateDataset): 

'''Global Strain Rate Map. An integrated global model of present-day 

plate motions and plate boundary deformation''' 

__citation = '''Kreemer, C., W.E. Holt, and A.J. Haines, "An integrated 

global model of present-day plate motions and plate boundary deformation", 

Geophys. J. Int., 154, 8-34, 2003.''' 

 

def __init__( 

self, 

name='GSRM1.2', 

data_dir=None, 

raw_data_url=('https://mirror.pyrocko.org/gsrm.unavco.org/model' 

'/files/1.2/%s')): 

 

if data_dir is None: 

data_dir = op.join(config.config().tectonics_dir, name) 

 

StrainRateDataset.__init__( 

self, 

name, 

data_dir=data_dir, 

citation=self.__citation) 

 

self.raw_data_url = raw_data_url 

self._full_names = None 

self._names = None 

 

def full_names(self): 

if not self._full_names: 

fn = util.data_file(op.join('tectonics', 'gsrm1_plates.txt')) 

 

with open(fn, 'r') as f: 

self._full_names = dict( 

line.strip().split(None, 1) for line in f) 

 

return self._full_names 

 

def full_name(self, name): 

name = self.plate_alt_names().get(name, name) 

return self.full_names()[name] 

 

def plate_names(self): 

if self._names is None: 

self._names = sorted(self.full_names().keys()) 

 

return self._names 

 

def plate_alt_names(self): 

# 'African Plate' is named 'Nubian Plate' in GSRM1 

return {'AF': 'NU'} 

 

def download_if_needed(self, fn): 

fpath = self.fpath(fn) 

if not op.exists(fpath): 

self.download_file( 

self.raw_data_url % fn, fpath, 

status_callback=get_download_callback( 

'Downloading global strain rate map...')) 

 

def get_velocities(self, reference_name=None, region=None): 

reference_name = self.plate_alt_names().get( 

reference_name, reference_name) 

 

if reference_name is None: 

reference_name = 'NNR' 

 

fn = 'velocity_%s.dat' % reference_name 

 

self.download_if_needed(fn) 

fpath = self.fpath(fn) 

data = [] 

with open(fpath, 'rb') as f: 

for line in f: 

if line.strip().startswith(b'#'): 

continue 

t = line.split() 

data.append(list(map(float, t))) 

 

arr = num.array(data, dtype=num.float) 

 

if region is not None: 

points = arr[:, 1::-1] 

mask = od.points_in_region(points, region) 

arr = arr[mask, :] 

 

lons, lats, veast, vnorth, veast_err, vnorth_err, corr = arr.T 

return lats, lons, vnorth, veast, vnorth_err, veast_err, corr 

 

 

__all__ = ''' 

GSRM1 

PeterBird2003 

Plate'''.split()