Skip to content

VG LSM — Vertical Gravity Gradient

gravtools.models.vg_lsm

Classes for the estimation of vertical gravity gradient by least-squares adjustment of differential relative gravimeter observations.

Copyright (C) 2021 Andreas Hellerschmied andreas.hellerschmied@bev.gv.at

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program 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 General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see https://www.gnu.org/licenses/.

VGLSM

Bases: LSM

VG estimation by least-squares adjustment of non-differential gravimeter observations.

Notes

The estimation of vertical gradients is restricted to th observations of a single survey and one station!

Attributes:

Name Type Description
setup_obs_df Pandas DataFrame

Pandas Dataframes for logging (differential or absolute) setup observations, the related metadata, estimation results and statistics. The columns of the dataframe are defined in self._SETUP_OBS_COLUMNS.

observed_stations list of str

Unique list of names of observed stations, defining the order of stations (station IDs) for all matrices and vectors used in the adjustment scheme.

Source code in gravtools/models/vg_lsm.py
 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
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
class VGLSM(LSM):
    """VG estimation by least-squares adjustment of non-differential gravimeter observations.

    Notes
    -----
    The estimation of vertical gradients is restricted to th observations of a single survey and one station!


    Attributes
    ----------
    setup_obs_df : Pandas DataFrame
        Pandas Dataframes for logging (differential or absolute) setup observations, the related metadata, estimation
        results and statistics. The columns of the dataframe are defined in `self._SETUP_OBS_COLUMNS`.
    observed_stations : list of str
        Unique list of names of observed stations, defining the order of stations (station IDs) for all matrices and
        vectors used in the adjustment scheme.
    """

    # Column names of self.setup_obs_df:
    # - keys: Column names of the pandas dataframe
    # - values: Short description for table headers, etc., in the GUI
    _SETUP_OBS_COLUMNS_DICT = {
        'survey_name': 'Survey',
        'ref_epoch_dt': 'Epoch',
        'obs_id': 'Obs. ID',
        'station_name': 'Station',
        'setup_id': 'Setup ID',
        'g_obs_mugal': 'g [µGal]',
        'sd_g_obs_mugal': 'SD [µGal]',
        'dhf_sensor_m': 'Sensor height [m]',
        'sd_g_obs_est_mugal': 'SD_est [µGal]',
        'v_obs_est_mugal': 'Residuals [µGal]',  # Post fit residuals
        'sd_v_obs_est_mugal': 'SD_v [µGal]',  # SD of post-fit residuals
        'w_obs_est_mugal': 'Std. Residual []',
        'r_obs_est': 'Redundancy []',
        'tau_test_result': 'Outlier Test',
    }
    _SETUP_OBS_COLUMNS = list(_SETUP_OBS_COLUMNS_DICT.keys())

    # Column names of self.drift_pol_df:
    # - keys: Column names of the pandas dataframe
    # - values: Short description for table headers, etc., in the GUI
    _DRIFT_POL_DF_COLUMNS_DICT = {
        'survey_name': 'Survey',
        'degree': 'Degree',
        'coefficient': 'Coefficient',
        'sd_coeff': 'SD',
        'coeff_unit': 'Unit',
        'ref_epoch_t0_dt': 't0',
    }
    _DRIFT_POL_DF_COLUMNS = list(_DRIFT_POL_DF_COLUMNS_DICT.keys())

    # Column names of self.vg_pol_df:
    # - keys: Column names of the pandas dataframe
    # - values: Short description for table headers, etc., in the GUI
    _VG_POL_DF_COLUMNS_DICT = {
        'degree': 'Degree',
        'coefficient': 'Coefficient',
        'sd_coeff': 'SD',
        'coeff_unit': 'Unit',
    }
    _VG_POL_DF_COLUMNS = list(_VG_POL_DF_COLUMNS_DICT.keys())

    def __init__(self, stat_df, setups, comment='', write_log=True):
        """
        Parameters
        ----------
        stat_df : :py:obj:`gravtools.Station.stat_df`
            The station dataframe contains all relevant station data.
        setups : dict of dicts
            The setups dictionary contains all observation data used for the adjustment. The keys of the dictionary
            are the survey names (str). The items are again keys with the following items:

            - ref_epoch_delta_t_h : datetime object
                Reference epoch for the relative reference times in the column `delta_t_h` in the `setup_df` dataframe.
                The reference epoch is determined as the epoch of the first (active) observation in this survey.
            -  ref_epoch_delta_t_campaign_h : datetime object
                Reference epoch for the relative reference times in the column `delta_t_campaign_h` in the `setup_df`
                dataframe. The reference epoch is determined as the epoch of the first (active) observation in the campaign.
            - setup_df : Pandas DataFrame
                Pandas dataframes containing the observation data (see :py:obj:`gravtools.Survey.setup_df`).

        comment : str, optional (default = '')
            Arbitrary comment on the LSM run.
        write_log : bool, optional (default=True)
            Flag that indicates whether log string should be written or not.
        """
        # Call constructor from abstract base class:
        lsm_method = 'VG_LSM_nondiff'
        super().__init__(lsm_method, stat_df=stat_df, setups=setups, comment=comment, write_log=write_log)

    @classmethod
    def from_campaign(cls, campaign, comment='', write_log=True):
        """Constructor that generates and populates the LSM object from a Campaign class object.

        Notes
        -----
        Put checks that are dependent on the LSM method into here! All common checks and preparations are implemented
        in the `from_campaign` method of the parent class `LSM`.

        Parameters
        ----------
        campaign : :py:obj:`gravtools.models.survey.Campaign`
            The campaign object needs to provide setup data for all active surveys and the related station data.
        comment : str, optional (default = '')
            Arbitrary comment on the LSM run.
        write_log : bool, optional (default=True)
            Flag that indicates whether log string should be written or not.

        Returns
        -------
        :py:obj:`.VGLSM`
            Contains all information required for adjusting the campaign.
        """
        # Station data:
        # - Just one station (at different height levels) allowed => Check coordinates!
        if campaign.stations is None:
            raise AssertionError(f'The campaign "{campaign.campaign_name}" does not contain any station data!')
        else:
            if not hasattr(campaign.stations, 'stat_df'):
                raise AssertionError(f'The campaign "{campaign.campaign_name}" has not station dataframe!')
            else:
                if len(campaign.stations.stat_df) == 0:
                    raise AssertionError(f'The campaign "{campaign.campaign_name}" has an empty station dataframe!')
                else:  # Just one station allowed => Check the coordinates (lon, lat)!
                    stat_df_observed = campaign.stations.stat_df.loc[campaign.stations.stat_df.is_observed]
                    if (len(stat_df_observed['lat_deg'].unique()) > 1) or (
                            len(stat_df_observed['long_deg'].unique()) > 1):
                        raise AssertionError(f'This campaign contains observed stations with differing latitudes and/or'
                                             f' longitudes. This is an indicator for observations of more than one '
                                             f'stations which is not allowed.')

        # Survey data:
        # - Just one survey allowed!
        if campaign.surveys is None:
            raise AssertionError(f'The campaign "{campaign.campaign_name}" does not contain any survey data!')
        else:
            if len(campaign.surveys) > 1:
                raise AssertionError(
                    f'The campaign "{campaign.campaign_name}" contains more than one survey! The VG estimation is '
                    f'restricted to just one survey.')

        # Check if the reduced setup observations refer to the sensor height of the instrument:
        # Loop over surveys in campaign (just one...):
        for survey_name, survey in campaign.surveys.items():
            if survey.keep_survey and survey.is_active:
                if survey.red_reference_height_type != 'sensor_height':
                    raise AssertionError(f'In survey {survey_name} the reduced gravity values do not refer to the '
                                         f'sensor height! Change the reference height to "Sensor" and recalculate the '
                                         f'setup observation data.')

        return super().from_campaign(campaign, comment=comment, write_log=True)

    def adjust(self, drift_pol_degree=1,
               vg_polynomial_degree=1,
               vg_polynomial_ref_height_offset_m=0.0,
               sig0_mugal=1,
               confidence_level_chi_test=0.95,
               confidence_level_tau_test=0.95,
               verbose=False
               ):
        """Run the VG adjustment based on non-differential observations.

        Parameters
        ----------
        drift_pol_degree : int, optional (default=1)
            Degree of estimated drift polynomial.
        vg_polynomial_degree : int, optional (default=1)
            Degree of the estimated polynomial of the vertical gravity gradient. Valid values are 1, 2 or 3.
        vg_polynomial_ref_height_offset_m : float, optional (default=0.0)
            Vertical offset [m] between the reference height of the control point and the zero-level of the estimated
            VG polynomial. A positive offset describes an offset above the control point.
        sig0_mugal : int, optional (default=1)
            A priori standard deviation of unit weight of observations [µGal] for the stochastic model of the
            least-squares adjustment.
        confidence_level_chi_test : float, optional (default=0.95)
            Confidence level for the goodness-of-fit test.
        confidence_level_tau_test : float, optional (default=0.95)
            Confidence level for the tau test.
        verbose : bool, optional (default=False)
            If True, status messages are printed to the command line, e.g. for debugging and testing

        Notes
        -----
        """

        # Initial checks:
        if (vg_polynomial_degree < 1) or (vg_polynomial_degree > 3):
            raise AssertionError(f'Invalid degree of the VG polynomial (degree = {vg_polynomial_degree})! Valid values '
                                 f'are 1 to 3.')

        # Prepare lists and indices:
        # - Observations and parameters:
        self.observed_stations = []
        survey_names = []
        number_of_observations = 0
        setup_ids = []

        for survey_name, setup_data in self.setups.items():
            setup_df = setup_data['setup_df']
            self.observed_stations = self.observed_stations + setup_df['station_name'].to_list()
            number_of_observations = number_of_observations + len(setup_df)
            survey_names.append(survey_name)
            setup_ids = setup_ids + setup_df['setup_id'].to_list()
        self.observed_stations = misc.unique_ordered_list(
            self.observed_stations)  # Unique list of stations => order of stations in matrices!
        number_of_stations = len(self.observed_stations)
        number_of_surveys = len(self.setups)  # Has to be 1 anyway

        # ### Station data ###
        # Get dataframe with subset of observed stations only:
        filter_tmp = self.stat_df['station_name'].isin(self.observed_stations)
        self.stat_obs_df = self.stat_df.loc[filter_tmp].copy(deep=True)  # All observed stations
        # Drop columns that are not required at VG estimation:
        self.stat_obs_df.drop(columns=['g_mugal', 'sd_g_mugal', 'is_datum'], inplace=True)
        # Calculate statistical data for the sensor height (relevant information for VG estimation):
        tmp_df = setup_df.loc[:, ['station_name', 'dhf_sensor_m']].groupby('station_name').mean().rename(
            columns={"dhf_sensor_m": "dhf_sensor_mean_m"})
        self.stat_obs_df = self.stat_obs_df.merge(tmp_df, left_on='station_name', right_on='station_name', how='left')
        tmp_df = setup_df.loc[:, ['station_name', 'dhf_sensor_m']].groupby('station_name').std().rename(
            columns={"dhf_sensor_m": "dhf_sensor_std_m"})
        self.stat_obs_df = self.stat_obs_df.merge(tmp_df, left_on='station_name', right_on='station_name', how='left')
        tmp_df = setup_df.loc[:, ['station_name', 'dhf_sensor_m']].groupby('station_name').min().rename(
            columns={"dhf_sensor_m": "dhf_sensor_min_m"})
        self.stat_obs_df = self.stat_obs_df.merge(tmp_df, left_on='station_name', right_on='station_name', how='left')
        tmp_df = setup_df.loc[:, ['station_name', 'dhf_sensor_m']].groupby('station_name').max().rename(
            columns={"dhf_sensor_m": "dhf_sensor_max_m"})
        self.stat_obs_df = self.stat_obs_df.merge(tmp_df, left_on='station_name', right_on='station_name', how='left')

        if number_of_surveys > 1:
            raise AssertionError(
                f'Invalid number of surveys ({number_of_surveys})! Only one survey allowed for VG estimation.')
        # Total number of parameters to be estimated:
        # - Drift polynomial coeff.: Polynomial degree * number of surveys
        # - 1 constant instrumental bias per survey
        # - VG polynomial coeff.: Polynomial degree * number of surveys
        number_of_parameters = drift_pol_degree + vg_polynomial_degree + 1

        # Check, if setup IDs are unique:
        self.check_unique_setups(setup_ids)

        if verbose or self.write_log:
            time_now_str = dt.datetime.now(tz=pytz.UTC).strftime('%Y-%m-%d, %H:%M:%S %Z')
            tmp_str = f'#### Adjustment log (VG LGM estimation based on non-differential observations) ####\n'
            tmp_str += f'Processed with GravTools {GRAVTOOLS_VERSION} ({time_now_str})\n'
            tmp_str += f'Comment: {self.comment}\n'
            tmp_str += f'\n'
            tmp_str += f'---- Input data and settings ----\n'
            tmp_str += f'Method: {settings.ADJUSTMENT_METHODS[self.lsm_method]}\n'
            tmp_str += f'Number of stations (i.e. height levels): {number_of_stations}\n'
            tmp_str += f'Number of observations: {number_of_observations}\n'
            tmp_str += f'Number of estimated parameters: {number_of_parameters}\n'
            tmp_str += f'Degree of freedom: {number_of_observations - number_of_parameters}\n'
            tmp_str += f'\n'
            tmp_str += f'Degree of drift polynomial: {drift_pol_degree}\n'
            tmp_str += f'Degree of VG polynomial: {vg_polynomial_degree}\n'
            tmp_str += f'VG polynomial height offset [m]: {vg_polynomial_ref_height_offset_m:4.3f}\n'
            tmp_str += f'A priori std. deviation of unit weight [µGal]: {sig0_mugal}\n'
            tmp_str += f'Confidence level Chi-test: {confidence_level_chi_test:4.2f}\n'
            tmp_str += f'Confidence level Tau-test: {confidence_level_tau_test:4.2f}\n'
            tmp_str += f'\n'
            tmp_str += f'---- Survey infos ----\n'
            tmp_str += self.survey_info_string
            tmp_str += f'\n'
            tmp_str += f'\n'
            if verbose:
                print(tmp_str)
            if self.write_log:
                self.log_str += tmp_str

        # Initialize matrices:
        if verbose or self.write_log:
            tmp_str = f'---- Set up and populate matrices ----\n'
            if verbose:
                print(tmp_str)
            if self.write_log:
                self.log_str += tmp_str
        # => Initialize complete matrices first and then populate them. This is most efficient!
        # - Observation model:
        mat_A = np.zeros([number_of_observations, number_of_parameters])  # Design matrix
        mat_L = np.zeros((number_of_observations, 1))  # Observations
        mat_sig_ll = np.zeros(number_of_observations)  # Variances of obs.

        # Populate matrices:
        obs_id = -1  # Index of differential observations in vectors mat_L, rows of mat_A and mat_p0
        survey_count = -1  # Survey counter for indexing the drift parameters in the A-matrix
        g_obs_mugal_list = []
        station_name_list = []
        setup_id_list = []
        sd_g_obs_mugal_list = []
        obs_id_list = []
        survey_names_list = []
        ref_epoch_dt_list = []  # Reference epochs (datetime objects) of observations
        dhf_sensor_m_list = []  # Height above the control point for VG estimation

        for survey_name, setup_data in self.setups.items():
            setup_df = setup_data['setup_df']
            survey_count += 1

            # Add offset to the height of the sensor above the control point:
            # setup_df['dhf_sensor_m'] = setup_df['dhf_sensor_m'] + vg_polynomial_ref_height_offset_m
            if (setup_df['sd_g_mugal'] < 0).any():
                raise AssertionError(
                    f'ERROR: SD of observations ("sd_g_mugal") in survey {survey_name} <= 0 are not allowed!')

            for index, row in setup_df.iterrows():
                obs_id += 1  # Increment diff. observation ID
                g_obs_mugal = row['g_mugal']
                sd_g_obs_mugal = row['sd_g_mugal']
                station_name = row['station_name']
                setup_id = row['setup_id']
                delta_t_h = row['delta_t_h']  # [hours]
                ref_epoch_dt = row['epoch_dt']
                dhf_sensor_m = row['dhf_sensor_m'] + vg_polynomial_ref_height_offset_m

                # Populate matrices and vectors:
                mat_L[(obs_id, 0)] = g_obs_mugal
                mat_sig_ll[obs_id] = sd_g_obs_mugal ** 2
                # Partial derivative for drift polynomial including constant instrumental bias (pol. degree = 0):
                for pd_drift_id in range(drift_pol_degree + 1):
                    mat_A[obs_id, pd_drift_id] = \
                        delta_t_h ** pd_drift_id
                # Partial derivative for VG polynomial:
                for pd_vg_id in range(vg_polynomial_degree):
                    mat_A[obs_id, drift_pol_degree + 1 + pd_vg_id] = \
                        dhf_sensor_m ** (pd_vg_id + 1)

                # Log data in DataFrame:
                g_obs_mugal_list.append(g_obs_mugal)
                sd_g_obs_mugal_list.append(sd_g_obs_mugal)
                station_name_list.append(station_name)
                obs_id_list.append(obs_id)
                survey_names_list.append(survey_name)
                setup_id_list.append(setup_id)
                ref_epoch_dt_list.append(ref_epoch_dt)
                dhf_sensor_m_list.append(dhf_sensor_m)

        None_list_placeholder = [None] * len(survey_names_list)
        self.setup_obs_df = pd.DataFrame(list(zip(survey_names_list,
                                                  ref_epoch_dt_list,
                                                  obs_id_list,
                                                  station_name_list,
                                                  setup_id_list,
                                                  g_obs_mugal_list,
                                                  sd_g_obs_mugal_list,
                                                  dhf_sensor_m_list,
                                                  None_list_placeholder,
                                                  None_list_placeholder,
                                                  None_list_placeholder,
                                                  None_list_placeholder,
                                                  None_list_placeholder,
                                                  None_list_placeholder,
                                                  )),
                                         columns=self._SETUP_OBS_COLUMNS)

        # Set up all required matrices:
        mat_sig_ll = np.diag(mat_sig_ll)
        mat_Qll = mat_sig_ll / (sig0_mugal ** 2)
        mat_P = np.linalg.inv(mat_Qll)

        if verbose or self.write_log:
            tmp_str = f'\n'
            tmp_str += f'---- Results and statistics ----\n'
            if verbose:
                print(tmp_str)
            if self.write_log:
                self.log_str += tmp_str

        # Solve equation system:
        mat_N = mat_A.T @ mat_P @ mat_A  # Normal equation matrix
        mat_Qxx = np.linalg.inv(mat_N)  # Co-factor matrix of estimates
        mat_x = mat_Qxx @ (mat_A.T @ mat_P @ mat_L)  # Estimates
        mat_v = (mat_A @ mat_x) - mat_L  # Post-fit residuals
        mat_v = misc.numpy_array_set_zero(mat_v, atol=1e-4)  # ToDo: Check atol!
        mat_Qldld = mat_A @ mat_Qxx @ mat_A.T  # A posteriori Co-factor matrix of adjusted observations
        mat_Qldld = misc.numpy_array_set_zero(mat_Qldld)
        # mat_Qll = np.linalg.inv(mat_P)
        mat_Qvv = mat_Qll - mat_Qldld  # Co-factor matrix of post-fit residuals
        mat_Qvv = misc.numpy_array_set_zero(mat_Qvv)

        # Test: "Gewichtsreziprokenprobe nach Ansermet" (see Skriptum AG1, p. 136, Eq. (6.86))
        u = np.sum(np.diag((mat_P @ mat_Qldld)))  # number of unknown parameters (estimates)
        tmp_diff = np.abs(number_of_parameters - u)
        if np.abs(number_of_parameters - u) > settings.ANSERMET_DIFF_THRESHOLD:
            raise AssertionError(f'"Gewichtsreziprokenprobe nach Ansermet" failed! Difference = {tmp_diff}')
        else:
            if verbose or self.write_log:
                tmp_str = f'# Gewichtsreziprokenprobe nach Ansermet (difference = {tmp_diff}) => Passed!\n'
                if verbose:
                    print(tmp_str)
                if self.write_log:
                    self.log_str += tmp_str

        # Condition of normal equation matrix:
        if verbose or self.write_log:
            cond = np.linalg.cond(mat_N)
            tmp_str = f'Condition of normal equation matix N = {cond:1.3f}\n'
            if verbose:
                print(tmp_str)
            if self.write_log:
                self.log_str += tmp_str

        # A posteriori variance of unit weight s02:
        dof = mat_A.shape[0] - mat_A.shape[1]  # degree of freedom
        par_r = mat_v.T @ mat_P @ mat_v  # = v^T * P * v
        if dof == 0:
            # s02_a_posteriori_mugal2 = par_r[0][0]
            raise AssertionError('Degree of freedom has to be larger than 0!')
        else:
            s02_a_posteriori_mugal2 = par_r[0][0] / dof  # Eq. (20)

        s0_mugal = np.sqrt(s02_a_posteriori_mugal2)  # A posteriori std. deviation of unit weight
        if verbose or self.write_log:
            tmp_str = f'\n'
            tmp_str += f'A posteriori variance (sd) of unit weight: ' \
                       f'{s02_a_posteriori_mugal2:5.3f} ({s0_mugal:5.3f})\n'
            if verbose:
                print(tmp_str)
            if self.write_log:
                self.log_str += tmp_str

        # ### Statistics and tests ###

        # Convert co-factor matrices to covariance matrices (variances in diagonal vector)
        mat_Cvv = s02_a_posteriori_mugal2 * mat_Qvv  # A posteriori Covariance matrix of post-fit residuals
        mat_Cxx = s02_a_posteriori_mugal2 * mat_Qxx  # A posteriori Covariance matrix of estimated paramaters
        mat_Cldld = s02_a_posteriori_mugal2 * mat_Qldld  # A posteriori Covariance matrix of adjusted observations
        # diag_Qxx = np.diag(mat_Qxx)

        # Calculate standard deviations:
        mat_sd_xx = np.sqrt(np.diag(mat_Cxx))  # A posteriori SD of estimates
        mat_sd_ldld = np.sqrt(np.diag(mat_Cldld))  # A posteriori SD of adjusted observations
        # mat_sd_vv = np.sqrt(np.diag(mat_Cvv))  # A posteriori SD of residuals
        # mat_sd_vv = np.sqrt(np.diag(abs(mat_Cvv)))  # !!!! without "abs()" the sqrt operation fails because neg. values may occur!

        # A posteriori SD of residuals:
        # - Check just in case, whether all diagonal elements of the Qvv matrix are positive!
        if (np.diag(mat_Qvv) < 0).any():
            mat_sd_vv = np.sqrt(np.diag(abs(mat_Cvv)))
            if verbose or self.write_log:
                tmp_str = f' - Warning: At least one diagonal element of the Qvv matrix is negative!\n'
                if verbose:
                    print(tmp_str)
                if self.write_log:
                    self.log_str += tmp_str
        else:
            mat_sd_vv = np.sqrt(np.diag(mat_Cvv))

        # creating histogram from residuals
        residual_hist, bin_edges = create_hist(mat_v)  # Calculate histogram

        # goodness-of-fit test
        chi_crit, chi_val, chi_test = global_model_test(confidence_level_chi_test, dof,
                                                        s02_a_posteriori_mugal2, sig0_mugal ** 2)
        if verbose or self.write_log:
            tmp_str = f'\n'
            tmp_str += f'# Goodness-of-fit test results:\n'
            tmp_str += f'# Chi-val  Chi-crt-lower  Chi-crt-upper  Status\n'
            tmp_str += f'  {chi_val:7.3f}  {chi_crit[0]:13.3f}  {chi_crit[1]:13.3f}  {chi_test:s}\n'
            tmp_str += f'\n'
            tmp_str += f'# Histogram of the residuals:\n'
            tmp_str += f'# Lower-edge(µGal)  Upper-edge(µGal)  Frequency\n'
            for loop_1 in range(len(residual_hist)):
                tmp_str += f'  {bin_edges[loop_1]:16.4f}  {bin_edges[loop_1 + 1]:16.4f}  {residual_hist[loop_1]:9.0f}\n'
            if verbose:
                print(tmp_str)
            if self.write_log:
                self.log_str += tmp_str

        # Redundanzanteile (redundancy components):
        # - Measure for the outlier detection effectiveness
        # - AG II, pp. 66-71
        mat_r = np.diag(mat_Qvv @ mat_P)

        # Standardized residuals used as test statistics for outlier detection:
        # - AG II, p. 66
        # - Taking care of zero-elements in the mat_sd_vv vector to prevent division by zero errors!
        mat_w = np.zeros(len(mat_v))
        tmp_filter = ~np.isclose(mat_v[:, 0], 0.0)
        mat_w[tmp_filter] = mat_v[tmp_filter, 0] / mat_sd_vv[tmp_filter]

        # Tau test for outlier detection:
        alpha_tau = 1 - confidence_level_tau_test
        tau_test_result, tau_critical_value = tau_test(mat_w=mat_w, dof=dof, alpha=alpha_tau, mat_r=mat_r)
        number_of_outliers = tau_test_result.count("failed")

        if verbose or self.write_log:
            tmp_str = f'\n'
            tmp_str += f'# Tau-test results:\n'
            tmp_str += f'Critical value (for testing w): {tau_critical_value:1.3f}\n'
            tmp_str += f' - Number of detected outliers: {number_of_outliers}\n'
            tmp_str += f' - Number low redundancy component: {tau_test_result.count("r too small")}\n'
            tmp_str += f'\n'
            if verbose:
                print(tmp_str)
            if self.write_log:
                self.log_str += tmp_str

        # #### Store results ####
        drift_pol_coeff = mat_x[:drift_pol_degree + 1, 0]
        drift_pol_coeff_sd = mat_sd_xx[:drift_pol_degree + 1]
        vg_pol_coeff = mat_x[drift_pol_degree + 1:, 0]
        vg_pol_coeff_sd = mat_sd_xx[drift_pol_degree + 1:]
        sd_g_obs_est_mugal = mat_sd_ldld
        v_obs_est_mugal = mat_v
        sd_v_obs_est_mugal = mat_sd_vv
        w_obs_est_mugal = mat_w
        r_obs_est = mat_r
        tau_test_result_obs = tau_test_result

        # Drift parameters:
        survey_name_list = []
        degree_list = []
        coefficient_list = []
        sd_coeff_list = []
        coeff_unit_list = []
        ref_epoch_t0_dt_list = []
        tmp_idx = 0
        x_estimate_drift_coeff_names = []
        for survey_name, setup_data in self.setups.items():
            for degree in range(drift_pol_degree + 1):
                survey_name_list.append(survey_name)
                degree_list.append(degree)  # starts with 1
                coefficient_list.append(drift_pol_coeff[tmp_idx])  # * (3600**(degree + 1))  # [µGal/h]
                sd_coeff_list.append(drift_pol_coeff_sd[tmp_idx])
                ref_epoch_t0_dt_list.append(setup_data['ref_epoch_delta_t_h'])
                if degree == 0:
                    coeff_unit_list.append(f'µGal')
                else:
                    coeff_unit_list.append(f'µGal/h^{degree}')
                tmp_idx += 1
                x_estimate_drift_coeff_names.append(f'drift-{degree}')
        self.drift_pol_df = pd.DataFrame(list(zip(survey_name_list,
                                                  degree_list,
                                                  coefficient_list,
                                                  sd_coeff_list,
                                                  coeff_unit_list,
                                                  ref_epoch_t0_dt_list)),
                                         columns=self._DRIFT_POL_DF_COLUMNS)

        # VG parameters:
        degree_list = []
        coefficient_list = []
        sd_coeff_list = []
        coeff_unit_list = []
        tmp_idx = 0
        x_estimate_vg_coeff_names = []
        for degree in range(1, vg_polynomial_degree + 1):
            degree_list.append(degree)  # starts with 1
            coefficient_list.append(vg_pol_coeff[tmp_idx])  # * (3600**(degree + 1))  # [µGal/h]
            sd_coeff_list.append(vg_pol_coeff_sd[tmp_idx])
            if degree == 0:
                coeff_unit_list.append(f'µGal')
            else:
                coeff_unit_list.append(f'µGal/m^{degree}')
            tmp_idx += 1
            x_estimate_vg_coeff_names.append(f'vg-{degree}')
        self.vg_pol_df = pd.DataFrame(list(zip(degree_list,
                                               coefficient_list,
                                               sd_coeff_list,
                                               coeff_unit_list)),
                                      columns=self._VG_POL_DF_COLUMNS)

        # Observation-related results:
        for idx, v_obs_mugal in enumerate(v_obs_est_mugal):
            filter_tmp = self.setup_obs_df['obs_id'] == idx
            self.setup_obs_df.loc[filter_tmp, 'v_obs_est_mugal'] = v_obs_mugal
            self.setup_obs_df.loc[filter_tmp, 'sd_v_obs_est_mugal'] = sd_v_obs_est_mugal[idx]
            self.setup_obs_df.loc[filter_tmp, 'sd_g_obs_est_mugal'] = sd_g_obs_est_mugal[idx]
            self.setup_obs_df.loc[filter_tmp, 'w_obs_est_mugal'] = w_obs_est_mugal[idx]  # standardized residuals
            self.setup_obs_df.loc[filter_tmp, 'r_obs_est'] = r_obs_est[idx]  # redundancy components
            self.setup_obs_df.loc[filter_tmp, 'tau_test_result'] = tau_test_result_obs[idx]  # str

        # Print results to terminal and to log string:
        if verbose or self.write_log:
            tmp_str = f'\n'
            tmp_str += f' - Station data:\n'
            tmp_str += self.stat_obs_df[['station_name', 'dhf_sensor_mean_m', 'dhf_sensor_std_m',
                                         'dhf_sensor_min_m', 'dhf_sensor_max_m']].to_string(index=False,
                                                                      float_format=lambda x: '{:.4f}'.format(x))
            tmp_str += f'\n\n'
            tmp_str += f' - Drift polynomial coefficients:\n'
            tmp_str += self.drift_pol_df.to_string(index=False, float_format=lambda x: '{:.6f}'.format(x))
            tmp_str += f'\n\n'
            tmp_str += f' - VG polynomial coefficients:\n'
            tmp_str += self.vg_pol_df.to_string(index=False, float_format=lambda x: '{:.6f}'.format(x))
            tmp_str += f'\n\n'
            tmp_str += f' - Observations:\n'
            for survey_name in survey_names:
                filter_tmp = self.setup_obs_df['survey_name'] == survey_name
                tmp_str += f'   - Survey: {survey_name}\n'
            tmp_str += self.setup_obs_df.loc[filter_tmp, ['station_name', 'g_obs_mugal',
                                                          'sd_g_obs_mugal', 'sd_g_obs_est_mugal',
                                                          'v_obs_est_mugal']].to_string(index=False,
                                                                                        float_format=lambda
                                                                                            x: '{:.1f}'.format(x))
            if verbose:
                print(tmp_str)
            if self.write_log:
                self.log_str += tmp_str

        # Save data/infos to object for later use:
        self.drift_polynomial_degree = drift_pol_degree
        self.sig0_a_priori = sig0_mugal
        self.confidence_level_chi_test = confidence_level_chi_test
        self.confidence_level_tau_test = confidence_level_chi_test
        self.number_of_stations = number_of_stations
        self.number_of_estimates = number_of_parameters
        self.degree_of_freedom = dof
        self.s02_a_posteriori = s02_a_posteriori_mugal2
        self.Cxx = mat_Cxx
        self.x_estimate_names = x_estimate_drift_coeff_names + x_estimate_vg_coeff_names
        self.global_model_test_status = chi_test
        self.number_of_outliers = number_of_outliers
        self.drift_ref_epoch_type = 'survey'  # Not relevant anyway, because only ONE survey allowed in the campaign!
        self.vg_polynomial_ref_height_offset_m = vg_polynomial_ref_height_offset_m
        self.vg_polynomial_degree = vg_polynomial_degree
        self.mat_A = mat_A
        self.mat_x = mat_x

__init__(stat_df, setups, comment='', write_log=True)

Parameters:

Name Type Description Default
stat_df :py:obj:`gravtools.Station.stat_df`

The station dataframe contains all relevant station data.

required
setups dict of dicts

The setups dictionary contains all observation data used for the adjustment. The keys of the dictionary are the survey names (str). The items are again keys with the following items:

  • ref_epoch_delta_t_h : datetime object Reference epoch for the relative reference times in the column delta_t_h in the setup_df dataframe. The reference epoch is determined as the epoch of the first (active) observation in this survey.
  • ref_epoch_delta_t_campaign_h : datetime object Reference epoch for the relative reference times in the column delta_t_campaign_h in the setup_df dataframe. The reference epoch is determined as the epoch of the first (active) observation in the campaign.
  • setup_df : Pandas DataFrame Pandas dataframes containing the observation data (see :py:obj:gravtools.Survey.setup_df).
required
comment (str, optional(default=''))

Arbitrary comment on the LSM run.

''
write_log (bool, optional(default=True))

Flag that indicates whether log string should be written or not.

True
Source code in gravtools/models/vg_lsm.py
 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
def __init__(self, stat_df, setups, comment='', write_log=True):
    """
    Parameters
    ----------
    stat_df : :py:obj:`gravtools.Station.stat_df`
        The station dataframe contains all relevant station data.
    setups : dict of dicts
        The setups dictionary contains all observation data used for the adjustment. The keys of the dictionary
        are the survey names (str). The items are again keys with the following items:

        - ref_epoch_delta_t_h : datetime object
            Reference epoch for the relative reference times in the column `delta_t_h` in the `setup_df` dataframe.
            The reference epoch is determined as the epoch of the first (active) observation in this survey.
        -  ref_epoch_delta_t_campaign_h : datetime object
            Reference epoch for the relative reference times in the column `delta_t_campaign_h` in the `setup_df`
            dataframe. The reference epoch is determined as the epoch of the first (active) observation in the campaign.
        - setup_df : Pandas DataFrame
            Pandas dataframes containing the observation data (see :py:obj:`gravtools.Survey.setup_df`).

    comment : str, optional (default = '')
        Arbitrary comment on the LSM run.
    write_log : bool, optional (default=True)
        Flag that indicates whether log string should be written or not.
    """
    # Call constructor from abstract base class:
    lsm_method = 'VG_LSM_nondiff'
    super().__init__(lsm_method, stat_df=stat_df, setups=setups, comment=comment, write_log=write_log)

adjust(drift_pol_degree=1, vg_polynomial_degree=1, vg_polynomial_ref_height_offset_m=0.0, sig0_mugal=1, confidence_level_chi_test=0.95, confidence_level_tau_test=0.95, verbose=False)

Run the VG adjustment based on non-differential observations.

Parameters:

Name Type Description Default
drift_pol_degree (int, optional(default=1))

Degree of estimated drift polynomial.

1
vg_polynomial_degree (int, optional(default=1))

Degree of the estimated polynomial of the vertical gravity gradient. Valid values are 1, 2 or 3.

1
vg_polynomial_ref_height_offset_m (float, optional(default=0.0))

Vertical offset [m] between the reference height of the control point and the zero-level of the estimated VG polynomial. A positive offset describes an offset above the control point.

0.0
sig0_mugal (int, optional(default=1))

A priori standard deviation of unit weight of observations [µGal] for the stochastic model of the least-squares adjustment.

1
confidence_level_chi_test (float, optional(default=0.95))

Confidence level for the goodness-of-fit test.

0.95
confidence_level_tau_test (float, optional(default=0.95))

Confidence level for the tau test.

0.95
verbose (bool, optional(default=False))

If True, status messages are printed to the command line, e.g. for debugging and testing

False
Notes
Source code in gravtools/models/vg_lsm.py
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
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
def adjust(self, drift_pol_degree=1,
           vg_polynomial_degree=1,
           vg_polynomial_ref_height_offset_m=0.0,
           sig0_mugal=1,
           confidence_level_chi_test=0.95,
           confidence_level_tau_test=0.95,
           verbose=False
           ):
    """Run the VG adjustment based on non-differential observations.

    Parameters
    ----------
    drift_pol_degree : int, optional (default=1)
        Degree of estimated drift polynomial.
    vg_polynomial_degree : int, optional (default=1)
        Degree of the estimated polynomial of the vertical gravity gradient. Valid values are 1, 2 or 3.
    vg_polynomial_ref_height_offset_m : float, optional (default=0.0)
        Vertical offset [m] between the reference height of the control point and the zero-level of the estimated
        VG polynomial. A positive offset describes an offset above the control point.
    sig0_mugal : int, optional (default=1)
        A priori standard deviation of unit weight of observations [µGal] for the stochastic model of the
        least-squares adjustment.
    confidence_level_chi_test : float, optional (default=0.95)
        Confidence level for the goodness-of-fit test.
    confidence_level_tau_test : float, optional (default=0.95)
        Confidence level for the tau test.
    verbose : bool, optional (default=False)
        If True, status messages are printed to the command line, e.g. for debugging and testing

    Notes
    -----
    """

    # Initial checks:
    if (vg_polynomial_degree < 1) or (vg_polynomial_degree > 3):
        raise AssertionError(f'Invalid degree of the VG polynomial (degree = {vg_polynomial_degree})! Valid values '
                             f'are 1 to 3.')

    # Prepare lists and indices:
    # - Observations and parameters:
    self.observed_stations = []
    survey_names = []
    number_of_observations = 0
    setup_ids = []

    for survey_name, setup_data in self.setups.items():
        setup_df = setup_data['setup_df']
        self.observed_stations = self.observed_stations + setup_df['station_name'].to_list()
        number_of_observations = number_of_observations + len(setup_df)
        survey_names.append(survey_name)
        setup_ids = setup_ids + setup_df['setup_id'].to_list()
    self.observed_stations = misc.unique_ordered_list(
        self.observed_stations)  # Unique list of stations => order of stations in matrices!
    number_of_stations = len(self.observed_stations)
    number_of_surveys = len(self.setups)  # Has to be 1 anyway

    # ### Station data ###
    # Get dataframe with subset of observed stations only:
    filter_tmp = self.stat_df['station_name'].isin(self.observed_stations)
    self.stat_obs_df = self.stat_df.loc[filter_tmp].copy(deep=True)  # All observed stations
    # Drop columns that are not required at VG estimation:
    self.stat_obs_df.drop(columns=['g_mugal', 'sd_g_mugal', 'is_datum'], inplace=True)
    # Calculate statistical data for the sensor height (relevant information for VG estimation):
    tmp_df = setup_df.loc[:, ['station_name', 'dhf_sensor_m']].groupby('station_name').mean().rename(
        columns={"dhf_sensor_m": "dhf_sensor_mean_m"})
    self.stat_obs_df = self.stat_obs_df.merge(tmp_df, left_on='station_name', right_on='station_name', how='left')
    tmp_df = setup_df.loc[:, ['station_name', 'dhf_sensor_m']].groupby('station_name').std().rename(
        columns={"dhf_sensor_m": "dhf_sensor_std_m"})
    self.stat_obs_df = self.stat_obs_df.merge(tmp_df, left_on='station_name', right_on='station_name', how='left')
    tmp_df = setup_df.loc[:, ['station_name', 'dhf_sensor_m']].groupby('station_name').min().rename(
        columns={"dhf_sensor_m": "dhf_sensor_min_m"})
    self.stat_obs_df = self.stat_obs_df.merge(tmp_df, left_on='station_name', right_on='station_name', how='left')
    tmp_df = setup_df.loc[:, ['station_name', 'dhf_sensor_m']].groupby('station_name').max().rename(
        columns={"dhf_sensor_m": "dhf_sensor_max_m"})
    self.stat_obs_df = self.stat_obs_df.merge(tmp_df, left_on='station_name', right_on='station_name', how='left')

    if number_of_surveys > 1:
        raise AssertionError(
            f'Invalid number of surveys ({number_of_surveys})! Only one survey allowed for VG estimation.')
    # Total number of parameters to be estimated:
    # - Drift polynomial coeff.: Polynomial degree * number of surveys
    # - 1 constant instrumental bias per survey
    # - VG polynomial coeff.: Polynomial degree * number of surveys
    number_of_parameters = drift_pol_degree + vg_polynomial_degree + 1

    # Check, if setup IDs are unique:
    self.check_unique_setups(setup_ids)

    if verbose or self.write_log:
        time_now_str = dt.datetime.now(tz=pytz.UTC).strftime('%Y-%m-%d, %H:%M:%S %Z')
        tmp_str = f'#### Adjustment log (VG LGM estimation based on non-differential observations) ####\n'
        tmp_str += f'Processed with GravTools {GRAVTOOLS_VERSION} ({time_now_str})\n'
        tmp_str += f'Comment: {self.comment}\n'
        tmp_str += f'\n'
        tmp_str += f'---- Input data and settings ----\n'
        tmp_str += f'Method: {settings.ADJUSTMENT_METHODS[self.lsm_method]}\n'
        tmp_str += f'Number of stations (i.e. height levels): {number_of_stations}\n'
        tmp_str += f'Number of observations: {number_of_observations}\n'
        tmp_str += f'Number of estimated parameters: {number_of_parameters}\n'
        tmp_str += f'Degree of freedom: {number_of_observations - number_of_parameters}\n'
        tmp_str += f'\n'
        tmp_str += f'Degree of drift polynomial: {drift_pol_degree}\n'
        tmp_str += f'Degree of VG polynomial: {vg_polynomial_degree}\n'
        tmp_str += f'VG polynomial height offset [m]: {vg_polynomial_ref_height_offset_m:4.3f}\n'
        tmp_str += f'A priori std. deviation of unit weight [µGal]: {sig0_mugal}\n'
        tmp_str += f'Confidence level Chi-test: {confidence_level_chi_test:4.2f}\n'
        tmp_str += f'Confidence level Tau-test: {confidence_level_tau_test:4.2f}\n'
        tmp_str += f'\n'
        tmp_str += f'---- Survey infos ----\n'
        tmp_str += self.survey_info_string
        tmp_str += f'\n'
        tmp_str += f'\n'
        if verbose:
            print(tmp_str)
        if self.write_log:
            self.log_str += tmp_str

    # Initialize matrices:
    if verbose or self.write_log:
        tmp_str = f'---- Set up and populate matrices ----\n'
        if verbose:
            print(tmp_str)
        if self.write_log:
            self.log_str += tmp_str
    # => Initialize complete matrices first and then populate them. This is most efficient!
    # - Observation model:
    mat_A = np.zeros([number_of_observations, number_of_parameters])  # Design matrix
    mat_L = np.zeros((number_of_observations, 1))  # Observations
    mat_sig_ll = np.zeros(number_of_observations)  # Variances of obs.

    # Populate matrices:
    obs_id = -1  # Index of differential observations in vectors mat_L, rows of mat_A and mat_p0
    survey_count = -1  # Survey counter for indexing the drift parameters in the A-matrix
    g_obs_mugal_list = []
    station_name_list = []
    setup_id_list = []
    sd_g_obs_mugal_list = []
    obs_id_list = []
    survey_names_list = []
    ref_epoch_dt_list = []  # Reference epochs (datetime objects) of observations
    dhf_sensor_m_list = []  # Height above the control point for VG estimation

    for survey_name, setup_data in self.setups.items():
        setup_df = setup_data['setup_df']
        survey_count += 1

        # Add offset to the height of the sensor above the control point:
        # setup_df['dhf_sensor_m'] = setup_df['dhf_sensor_m'] + vg_polynomial_ref_height_offset_m
        if (setup_df['sd_g_mugal'] < 0).any():
            raise AssertionError(
                f'ERROR: SD of observations ("sd_g_mugal") in survey {survey_name} <= 0 are not allowed!')

        for index, row in setup_df.iterrows():
            obs_id += 1  # Increment diff. observation ID
            g_obs_mugal = row['g_mugal']
            sd_g_obs_mugal = row['sd_g_mugal']
            station_name = row['station_name']
            setup_id = row['setup_id']
            delta_t_h = row['delta_t_h']  # [hours]
            ref_epoch_dt = row['epoch_dt']
            dhf_sensor_m = row['dhf_sensor_m'] + vg_polynomial_ref_height_offset_m

            # Populate matrices and vectors:
            mat_L[(obs_id, 0)] = g_obs_mugal
            mat_sig_ll[obs_id] = sd_g_obs_mugal ** 2
            # Partial derivative for drift polynomial including constant instrumental bias (pol. degree = 0):
            for pd_drift_id in range(drift_pol_degree + 1):
                mat_A[obs_id, pd_drift_id] = \
                    delta_t_h ** pd_drift_id
            # Partial derivative for VG polynomial:
            for pd_vg_id in range(vg_polynomial_degree):
                mat_A[obs_id, drift_pol_degree + 1 + pd_vg_id] = \
                    dhf_sensor_m ** (pd_vg_id + 1)

            # Log data in DataFrame:
            g_obs_mugal_list.append(g_obs_mugal)
            sd_g_obs_mugal_list.append(sd_g_obs_mugal)
            station_name_list.append(station_name)
            obs_id_list.append(obs_id)
            survey_names_list.append(survey_name)
            setup_id_list.append(setup_id)
            ref_epoch_dt_list.append(ref_epoch_dt)
            dhf_sensor_m_list.append(dhf_sensor_m)

    None_list_placeholder = [None] * len(survey_names_list)
    self.setup_obs_df = pd.DataFrame(list(zip(survey_names_list,
                                              ref_epoch_dt_list,
                                              obs_id_list,
                                              station_name_list,
                                              setup_id_list,
                                              g_obs_mugal_list,
                                              sd_g_obs_mugal_list,
                                              dhf_sensor_m_list,
                                              None_list_placeholder,
                                              None_list_placeholder,
                                              None_list_placeholder,
                                              None_list_placeholder,
                                              None_list_placeholder,
                                              None_list_placeholder,
                                              )),
                                     columns=self._SETUP_OBS_COLUMNS)

    # Set up all required matrices:
    mat_sig_ll = np.diag(mat_sig_ll)
    mat_Qll = mat_sig_ll / (sig0_mugal ** 2)
    mat_P = np.linalg.inv(mat_Qll)

    if verbose or self.write_log:
        tmp_str = f'\n'
        tmp_str += f'---- Results and statistics ----\n'
        if verbose:
            print(tmp_str)
        if self.write_log:
            self.log_str += tmp_str

    # Solve equation system:
    mat_N = mat_A.T @ mat_P @ mat_A  # Normal equation matrix
    mat_Qxx = np.linalg.inv(mat_N)  # Co-factor matrix of estimates
    mat_x = mat_Qxx @ (mat_A.T @ mat_P @ mat_L)  # Estimates
    mat_v = (mat_A @ mat_x) - mat_L  # Post-fit residuals
    mat_v = misc.numpy_array_set_zero(mat_v, atol=1e-4)  # ToDo: Check atol!
    mat_Qldld = mat_A @ mat_Qxx @ mat_A.T  # A posteriori Co-factor matrix of adjusted observations
    mat_Qldld = misc.numpy_array_set_zero(mat_Qldld)
    # mat_Qll = np.linalg.inv(mat_P)
    mat_Qvv = mat_Qll - mat_Qldld  # Co-factor matrix of post-fit residuals
    mat_Qvv = misc.numpy_array_set_zero(mat_Qvv)

    # Test: "Gewichtsreziprokenprobe nach Ansermet" (see Skriptum AG1, p. 136, Eq. (6.86))
    u = np.sum(np.diag((mat_P @ mat_Qldld)))  # number of unknown parameters (estimates)
    tmp_diff = np.abs(number_of_parameters - u)
    if np.abs(number_of_parameters - u) > settings.ANSERMET_DIFF_THRESHOLD:
        raise AssertionError(f'"Gewichtsreziprokenprobe nach Ansermet" failed! Difference = {tmp_diff}')
    else:
        if verbose or self.write_log:
            tmp_str = f'# Gewichtsreziprokenprobe nach Ansermet (difference = {tmp_diff}) => Passed!\n'
            if verbose:
                print(tmp_str)
            if self.write_log:
                self.log_str += tmp_str

    # Condition of normal equation matrix:
    if verbose or self.write_log:
        cond = np.linalg.cond(mat_N)
        tmp_str = f'Condition of normal equation matix N = {cond:1.3f}\n'
        if verbose:
            print(tmp_str)
        if self.write_log:
            self.log_str += tmp_str

    # A posteriori variance of unit weight s02:
    dof = mat_A.shape[0] - mat_A.shape[1]  # degree of freedom
    par_r = mat_v.T @ mat_P @ mat_v  # = v^T * P * v
    if dof == 0:
        # s02_a_posteriori_mugal2 = par_r[0][0]
        raise AssertionError('Degree of freedom has to be larger than 0!')
    else:
        s02_a_posteriori_mugal2 = par_r[0][0] / dof  # Eq. (20)

    s0_mugal = np.sqrt(s02_a_posteriori_mugal2)  # A posteriori std. deviation of unit weight
    if verbose or self.write_log:
        tmp_str = f'\n'
        tmp_str += f'A posteriori variance (sd) of unit weight: ' \
                   f'{s02_a_posteriori_mugal2:5.3f} ({s0_mugal:5.3f})\n'
        if verbose:
            print(tmp_str)
        if self.write_log:
            self.log_str += tmp_str

    # ### Statistics and tests ###

    # Convert co-factor matrices to covariance matrices (variances in diagonal vector)
    mat_Cvv = s02_a_posteriori_mugal2 * mat_Qvv  # A posteriori Covariance matrix of post-fit residuals
    mat_Cxx = s02_a_posteriori_mugal2 * mat_Qxx  # A posteriori Covariance matrix of estimated paramaters
    mat_Cldld = s02_a_posteriori_mugal2 * mat_Qldld  # A posteriori Covariance matrix of adjusted observations
    # diag_Qxx = np.diag(mat_Qxx)

    # Calculate standard deviations:
    mat_sd_xx = np.sqrt(np.diag(mat_Cxx))  # A posteriori SD of estimates
    mat_sd_ldld = np.sqrt(np.diag(mat_Cldld))  # A posteriori SD of adjusted observations
    # mat_sd_vv = np.sqrt(np.diag(mat_Cvv))  # A posteriori SD of residuals
    # mat_sd_vv = np.sqrt(np.diag(abs(mat_Cvv)))  # !!!! without "abs()" the sqrt operation fails because neg. values may occur!

    # A posteriori SD of residuals:
    # - Check just in case, whether all diagonal elements of the Qvv matrix are positive!
    if (np.diag(mat_Qvv) < 0).any():
        mat_sd_vv = np.sqrt(np.diag(abs(mat_Cvv)))
        if verbose or self.write_log:
            tmp_str = f' - Warning: At least one diagonal element of the Qvv matrix is negative!\n'
            if verbose:
                print(tmp_str)
            if self.write_log:
                self.log_str += tmp_str
    else:
        mat_sd_vv = np.sqrt(np.diag(mat_Cvv))

    # creating histogram from residuals
    residual_hist, bin_edges = create_hist(mat_v)  # Calculate histogram

    # goodness-of-fit test
    chi_crit, chi_val, chi_test = global_model_test(confidence_level_chi_test, dof,
                                                    s02_a_posteriori_mugal2, sig0_mugal ** 2)
    if verbose or self.write_log:
        tmp_str = f'\n'
        tmp_str += f'# Goodness-of-fit test results:\n'
        tmp_str += f'# Chi-val  Chi-crt-lower  Chi-crt-upper  Status\n'
        tmp_str += f'  {chi_val:7.3f}  {chi_crit[0]:13.3f}  {chi_crit[1]:13.3f}  {chi_test:s}\n'
        tmp_str += f'\n'
        tmp_str += f'# Histogram of the residuals:\n'
        tmp_str += f'# Lower-edge(µGal)  Upper-edge(µGal)  Frequency\n'
        for loop_1 in range(len(residual_hist)):
            tmp_str += f'  {bin_edges[loop_1]:16.4f}  {bin_edges[loop_1 + 1]:16.4f}  {residual_hist[loop_1]:9.0f}\n'
        if verbose:
            print(tmp_str)
        if self.write_log:
            self.log_str += tmp_str

    # Redundanzanteile (redundancy components):
    # - Measure for the outlier detection effectiveness
    # - AG II, pp. 66-71
    mat_r = np.diag(mat_Qvv @ mat_P)

    # Standardized residuals used as test statistics for outlier detection:
    # - AG II, p. 66
    # - Taking care of zero-elements in the mat_sd_vv vector to prevent division by zero errors!
    mat_w = np.zeros(len(mat_v))
    tmp_filter = ~np.isclose(mat_v[:, 0], 0.0)
    mat_w[tmp_filter] = mat_v[tmp_filter, 0] / mat_sd_vv[tmp_filter]

    # Tau test for outlier detection:
    alpha_tau = 1 - confidence_level_tau_test
    tau_test_result, tau_critical_value = tau_test(mat_w=mat_w, dof=dof, alpha=alpha_tau, mat_r=mat_r)
    number_of_outliers = tau_test_result.count("failed")

    if verbose or self.write_log:
        tmp_str = f'\n'
        tmp_str += f'# Tau-test results:\n'
        tmp_str += f'Critical value (for testing w): {tau_critical_value:1.3f}\n'
        tmp_str += f' - Number of detected outliers: {number_of_outliers}\n'
        tmp_str += f' - Number low redundancy component: {tau_test_result.count("r too small")}\n'
        tmp_str += f'\n'
        if verbose:
            print(tmp_str)
        if self.write_log:
            self.log_str += tmp_str

    # #### Store results ####
    drift_pol_coeff = mat_x[:drift_pol_degree + 1, 0]
    drift_pol_coeff_sd = mat_sd_xx[:drift_pol_degree + 1]
    vg_pol_coeff = mat_x[drift_pol_degree + 1:, 0]
    vg_pol_coeff_sd = mat_sd_xx[drift_pol_degree + 1:]
    sd_g_obs_est_mugal = mat_sd_ldld
    v_obs_est_mugal = mat_v
    sd_v_obs_est_mugal = mat_sd_vv
    w_obs_est_mugal = mat_w
    r_obs_est = mat_r
    tau_test_result_obs = tau_test_result

    # Drift parameters:
    survey_name_list = []
    degree_list = []
    coefficient_list = []
    sd_coeff_list = []
    coeff_unit_list = []
    ref_epoch_t0_dt_list = []
    tmp_idx = 0
    x_estimate_drift_coeff_names = []
    for survey_name, setup_data in self.setups.items():
        for degree in range(drift_pol_degree + 1):
            survey_name_list.append(survey_name)
            degree_list.append(degree)  # starts with 1
            coefficient_list.append(drift_pol_coeff[tmp_idx])  # * (3600**(degree + 1))  # [µGal/h]
            sd_coeff_list.append(drift_pol_coeff_sd[tmp_idx])
            ref_epoch_t0_dt_list.append(setup_data['ref_epoch_delta_t_h'])
            if degree == 0:
                coeff_unit_list.append(f'µGal')
            else:
                coeff_unit_list.append(f'µGal/h^{degree}')
            tmp_idx += 1
            x_estimate_drift_coeff_names.append(f'drift-{degree}')
    self.drift_pol_df = pd.DataFrame(list(zip(survey_name_list,
                                              degree_list,
                                              coefficient_list,
                                              sd_coeff_list,
                                              coeff_unit_list,
                                              ref_epoch_t0_dt_list)),
                                     columns=self._DRIFT_POL_DF_COLUMNS)

    # VG parameters:
    degree_list = []
    coefficient_list = []
    sd_coeff_list = []
    coeff_unit_list = []
    tmp_idx = 0
    x_estimate_vg_coeff_names = []
    for degree in range(1, vg_polynomial_degree + 1):
        degree_list.append(degree)  # starts with 1
        coefficient_list.append(vg_pol_coeff[tmp_idx])  # * (3600**(degree + 1))  # [µGal/h]
        sd_coeff_list.append(vg_pol_coeff_sd[tmp_idx])
        if degree == 0:
            coeff_unit_list.append(f'µGal')
        else:
            coeff_unit_list.append(f'µGal/m^{degree}')
        tmp_idx += 1
        x_estimate_vg_coeff_names.append(f'vg-{degree}')
    self.vg_pol_df = pd.DataFrame(list(zip(degree_list,
                                           coefficient_list,
                                           sd_coeff_list,
                                           coeff_unit_list)),
                                  columns=self._VG_POL_DF_COLUMNS)

    # Observation-related results:
    for idx, v_obs_mugal in enumerate(v_obs_est_mugal):
        filter_tmp = self.setup_obs_df['obs_id'] == idx
        self.setup_obs_df.loc[filter_tmp, 'v_obs_est_mugal'] = v_obs_mugal
        self.setup_obs_df.loc[filter_tmp, 'sd_v_obs_est_mugal'] = sd_v_obs_est_mugal[idx]
        self.setup_obs_df.loc[filter_tmp, 'sd_g_obs_est_mugal'] = sd_g_obs_est_mugal[idx]
        self.setup_obs_df.loc[filter_tmp, 'w_obs_est_mugal'] = w_obs_est_mugal[idx]  # standardized residuals
        self.setup_obs_df.loc[filter_tmp, 'r_obs_est'] = r_obs_est[idx]  # redundancy components
        self.setup_obs_df.loc[filter_tmp, 'tau_test_result'] = tau_test_result_obs[idx]  # str

    # Print results to terminal and to log string:
    if verbose or self.write_log:
        tmp_str = f'\n'
        tmp_str += f' - Station data:\n'
        tmp_str += self.stat_obs_df[['station_name', 'dhf_sensor_mean_m', 'dhf_sensor_std_m',
                                     'dhf_sensor_min_m', 'dhf_sensor_max_m']].to_string(index=False,
                                                                  float_format=lambda x: '{:.4f}'.format(x))
        tmp_str += f'\n\n'
        tmp_str += f' - Drift polynomial coefficients:\n'
        tmp_str += self.drift_pol_df.to_string(index=False, float_format=lambda x: '{:.6f}'.format(x))
        tmp_str += f'\n\n'
        tmp_str += f' - VG polynomial coefficients:\n'
        tmp_str += self.vg_pol_df.to_string(index=False, float_format=lambda x: '{:.6f}'.format(x))
        tmp_str += f'\n\n'
        tmp_str += f' - Observations:\n'
        for survey_name in survey_names:
            filter_tmp = self.setup_obs_df['survey_name'] == survey_name
            tmp_str += f'   - Survey: {survey_name}\n'
        tmp_str += self.setup_obs_df.loc[filter_tmp, ['station_name', 'g_obs_mugal',
                                                      'sd_g_obs_mugal', 'sd_g_obs_est_mugal',
                                                      'v_obs_est_mugal']].to_string(index=False,
                                                                                    float_format=lambda
                                                                                        x: '{:.1f}'.format(x))
        if verbose:
            print(tmp_str)
        if self.write_log:
            self.log_str += tmp_str

    # Save data/infos to object for later use:
    self.drift_polynomial_degree = drift_pol_degree
    self.sig0_a_priori = sig0_mugal
    self.confidence_level_chi_test = confidence_level_chi_test
    self.confidence_level_tau_test = confidence_level_chi_test
    self.number_of_stations = number_of_stations
    self.number_of_estimates = number_of_parameters
    self.degree_of_freedom = dof
    self.s02_a_posteriori = s02_a_posteriori_mugal2
    self.Cxx = mat_Cxx
    self.x_estimate_names = x_estimate_drift_coeff_names + x_estimate_vg_coeff_names
    self.global_model_test_status = chi_test
    self.number_of_outliers = number_of_outliers
    self.drift_ref_epoch_type = 'survey'  # Not relevant anyway, because only ONE survey allowed in the campaign!
    self.vg_polynomial_ref_height_offset_m = vg_polynomial_ref_height_offset_m
    self.vg_polynomial_degree = vg_polynomial_degree
    self.mat_A = mat_A
    self.mat_x = mat_x

from_campaign(campaign, comment='', write_log=True) classmethod

Constructor that generates and populates the LSM object from a Campaign class object.

Notes

Put checks that are dependent on the LSM method into here! All common checks and preparations are implemented in the from_campaign method of the parent class LSM.

Parameters:

Name Type Description Default
campaign :py:obj:`gravtools.models.survey.Campaign`

The campaign object needs to provide setup data for all active surveys and the related station data.

required
comment (str, optional(default=''))

Arbitrary comment on the LSM run.

''
write_log (bool, optional(default=True))

Flag that indicates whether log string should be written or not.

True

Returns:

Type Description
py:obj:`.VGLSM`

Contains all information required for adjusting the campaign.

Source code in gravtools/models/vg_lsm.py
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
@classmethod
def from_campaign(cls, campaign, comment='', write_log=True):
    """Constructor that generates and populates the LSM object from a Campaign class object.

    Notes
    -----
    Put checks that are dependent on the LSM method into here! All common checks and preparations are implemented
    in the `from_campaign` method of the parent class `LSM`.

    Parameters
    ----------
    campaign : :py:obj:`gravtools.models.survey.Campaign`
        The campaign object needs to provide setup data for all active surveys and the related station data.
    comment : str, optional (default = '')
        Arbitrary comment on the LSM run.
    write_log : bool, optional (default=True)
        Flag that indicates whether log string should be written or not.

    Returns
    -------
    :py:obj:`.VGLSM`
        Contains all information required for adjusting the campaign.
    """
    # Station data:
    # - Just one station (at different height levels) allowed => Check coordinates!
    if campaign.stations is None:
        raise AssertionError(f'The campaign "{campaign.campaign_name}" does not contain any station data!')
    else:
        if not hasattr(campaign.stations, 'stat_df'):
            raise AssertionError(f'The campaign "{campaign.campaign_name}" has not station dataframe!')
        else:
            if len(campaign.stations.stat_df) == 0:
                raise AssertionError(f'The campaign "{campaign.campaign_name}" has an empty station dataframe!')
            else:  # Just one station allowed => Check the coordinates (lon, lat)!
                stat_df_observed = campaign.stations.stat_df.loc[campaign.stations.stat_df.is_observed]
                if (len(stat_df_observed['lat_deg'].unique()) > 1) or (
                        len(stat_df_observed['long_deg'].unique()) > 1):
                    raise AssertionError(f'This campaign contains observed stations with differing latitudes and/or'
                                         f' longitudes. This is an indicator for observations of more than one '
                                         f'stations which is not allowed.')

    # Survey data:
    # - Just one survey allowed!
    if campaign.surveys is None:
        raise AssertionError(f'The campaign "{campaign.campaign_name}" does not contain any survey data!')
    else:
        if len(campaign.surveys) > 1:
            raise AssertionError(
                f'The campaign "{campaign.campaign_name}" contains more than one survey! The VG estimation is '
                f'restricted to just one survey.')

    # Check if the reduced setup observations refer to the sensor height of the instrument:
    # Loop over surveys in campaign (just one...):
    for survey_name, survey in campaign.surveys.items():
        if survey.keep_survey and survey.is_active:
            if survey.red_reference_height_type != 'sensor_height':
                raise AssertionError(f'In survey {survey_name} the reduced gravity values do not refer to the '
                                     f'sensor height! Change the reference height to "Sensor" and recalculate the '
                                     f'setup observation data.')

    return super().from_campaign(campaign, comment=comment, write_log=True)