#!/usr/bin/env python
# -*- coding: utf-8 -*-
# Copyright (c) 2017-2018 Satpy developers
#
# This file is part of satpy.
#
# satpy 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.
#
# satpy 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
# satpy. If not, see <http://www.gnu.org/licenses/>.
"""Common functionality for SEVIRI L1.5 data readers.
Introduction
------------
*The Spinning Enhanced Visible and InfraRed Imager (SEVIRI) is the primary
instrument on Meteosat Second Generation (MSG) and has the capacity to observe
the Earth in 12 spectral channels.*
*Level 1.5 corresponds to image data that has been corrected for all unwanted
radiometric and geometric effects, has been geolocated using a standardised
projection, and has been calibrated and radiance-linearised.*
(From the EUMETSAT documentation)
Satpy provides the following readers for SEVIRI L1.5 data in different formats:
- Native: :mod:`satpy.readers.seviri_l1b_native`
- HRIT: :mod:`satpy.readers.seviri_l1b_hrit`
- netCDF: :mod:`satpy.readers.seviri_l1b_nc`
Calibration
-----------
This section describes how to control the calibration of SEVIRI L1.5 data.
Calibration to radiance
^^^^^^^^^^^^^^^^^^^^^^^
The SEVIRI L1.5 data readers allow for choosing between two file-internal
calibration coefficients to convert counts to radiances:
- Nominal for all channels (default)
- GSICS where available (IR currently) and nominal for the remaining
channels (VIS & HRV currently)
In order to change the default behaviour, use the ``reader_kwargs`` keyword
argument upon Scene creation::
import satpy
scene = satpy.Scene(filenames=filenames,
reader='seviri_l1b_...',
reader_kwargs={'calib_mode': 'GSICS'})
scene.load(['VIS006', 'IR_108'])
In addition, two other calibration methods are available:
1. It is possible to specify external calibration coefficients for the
conversion from counts to radiances. External coefficients take
precedence over internal coefficients and over the Meirink
coefficients, but you can also mix internal and external coefficients:
If external calibration coefficients are specified for only a subset
of channels, the remaining channels will be calibrated using the
chosen file-internal coefficients (nominal or GSICS). Calibration
coefficients must be specified in [mW m-2 sr-1 (cm-1)-1].
2. The calibration mode ``meirink-2023`` uses coefficients based on an
intercalibration with Aqua-MODIS for the visible channels, as found in
`Inter-calibration of polar imager solar channels using SEVIRI`_
(2013) by J. F. Meirink, R. A. Roebeling, and P. Stammes.
In the following example we use external calibration coefficients for the
``VIS006`` & ``IR_108`` channels, and nominal coefficients for the
remaining channels::
coefs = {'VIS006': {'gain': 0.0236, 'offset': -1.20},
'IR_108': {'gain': 0.2156, 'offset': -10.4}}
scene = satpy.Scene(filenames,
reader='seviri_l1b_...',
reader_kwargs={'ext_calib_coefs': coefs})
scene.load(['VIS006', 'VIS008', 'IR_108', 'IR_120'])
In the next example we use external calibration coefficients for the
``VIS006`` & ``IR_108`` channels, GSICS coefficients where available
(other IR channels) and nominal coefficients for the rest::
coefs = {'VIS006': {'gain': 0.0236, 'offset': -1.20},
'IR_108': {'gain': 0.2156, 'offset': -10.4}}
scene = satpy.Scene(filenames,
reader='seviri_l1b_...',
reader_kwargs={'calib_mode': 'GSICS',
'ext_calib_coefs': coefs})
scene.load(['VIS006', 'VIS008', 'IR_108', 'IR_120'])
In the next example we use the mode ``meirink-2023`` calibration
coefficients for all visible channels and nominal coefficients for the
rest::
scene = satpy.Scene(filenames,
reader='seviri_l1b_...',
reader_kwargs={'calib_mode': 'meirink-2023'})
scene.load(['VIS006', 'VIS008', 'IR_016'])
Calibration to reflectance
^^^^^^^^^^^^^^^^^^^^^^^^^^
When loading solar channels, the SEVIRI L1.5 data readers apply a correction for
the Sun-Earth distance variation throughout the year - as recommended by
the EUMETSAT document
`Conversion from radiances to reflectances for SEVIRI warm channels`_.
In the unlikely situation that this correction is not required, it can be
removed on a per-channel basis using
:func:`satpy.readers.utils.remove_earthsun_distance_correction`.
Masking of bad quality scan lines
---------------------------------
By default bad quality scan lines are masked and replaced with ``np.nan`` for radiance, reflectance and
brightness temperature calibrations based on the quality flags provided by the data (for details on quality
flags see `MSG Level 1.5 Image Data Format Description`_ page 109). To disable masking
``reader_kwargs={'mask_bad_quality_scan_lines': False}`` can be passed to the Scene.
Metadata
--------
The SEVIRI L1.5 readers provide the following metadata:
* The ``orbital_parameters`` attribute provides the nominal and actual satellite
position, as well as the projection centre. See the `Metadata` section in
the :doc:`../reading` chapter for more information.
* The ``acq_time`` coordinate provides the mean acquisition time for each
scanline. Use a ``MultiIndex`` to enable selection by acquisition time:
.. code-block:: python
import pandas as pd
mi = pd.MultiIndex.from_arrays([scn['IR_108']['y'].data, scn['IR_108']['acq_time'].data],
names=('y_coord', 'time'))
scn['IR_108']['y'] = mi
scn['IR_108'].sel(time=np.datetime64('2019-03-01T12:06:13.052000000'))
* Raw metadata from the file header can be included by setting the reader
argument ``include_raw_metadata=True`` (HRIT and Native format only). Note
that this comes with a performance penalty of up to 10% if raw metadata from
multiple segments or scans need to be combined. By default, arrays with more
than 100 elements are excluded to limit the performance penalty. This
threshold can be adjusted using the ``mda_max_array_size`` reader keyword
argument:
.. code-block:: python
scene = satpy.Scene(filenames,
reader='seviri_l1b_hrit/native',
reader_kwargs={'include_raw_metadata': True,
'mda_max_array_size': 1000})
References:
- `MSG Level 1.5 Image Data Format Description`_
- `Radiometric Calibration of MSG SEVIRI Level 1.5 Image Data in Equivalent Spectral Blackbody Radiance`_
.. _Conversion from radiances to reflectances for SEVIRI warm channels:
https://www-cdn.eumetsat.int/files/2020-04/pdf_msg_seviri_rad2refl.pdf
.. _MSG Level 1.5 Image Data Format Description:
https://www.eumetsat.int/media/45126
.. _Radiometric Calibration of MSG SEVIRI Level 1.5 Image Data in Equivalent Spectral Blackbody Radiance:
https://www-cdn.eumetsat.int/files/2020-04/pdf_ten_msg_seviri_rad_calib.pdf
.. _Inter-calibration of polar imager solar channels using SEVIRI:
http://dx.doi.org/10.5194/amt-6-2495-2013
"""
from __future__ import annotations
import warnings
from datetime import datetime, timedelta
import dask.array as da
import numpy as np
import pyproj
from numpy.polynomial.chebyshev import Chebyshev
from satpy.readers.eum_base import issue_revision, time_cds_short
from satpy.readers.utils import apply_earthsun_distance_correction
from satpy.utils import get_legacy_chunk_size
CHUNK_SIZE = get_legacy_chunk_size()
PLATFORM_DICT = {
"MET08": "Meteosat-8",
"MET09": "Meteosat-9",
"MET10": "Meteosat-10",
"MET11": "Meteosat-11",
"MSG1": "Meteosat-8",
"MSG2": "Meteosat-9",
"MSG3": "Meteosat-10",
"MSG4": "Meteosat-11",
}
REPEAT_CYCLE_DURATION = 15
C1 = 1.19104273e-5
C2 = 1.43877523
VISIR_NUM_COLUMNS = 3712
VISIR_NUM_LINES = 3712
HRV_NUM_COLUMNS = 11136
HRV_NUM_LINES = 11136
CHANNEL_NAMES = {1: "VIS006",
2: "VIS008",
3: "IR_016",
4: "IR_039",
5: "WV_062",
6: "WV_073",
7: "IR_087",
8: "IR_097",
9: "IR_108",
10: "IR_120",
11: "IR_134",
12: "HRV"}
VIS_CHANNELS = ["HRV", "VIS006", "VIS008", "IR_016"]
# Polynomial coefficients for spectral-effective BT fits
BTFIT = dict()
# [A, B, C]
BTFIT["IR_039"] = [0.0, 1.011751900, -3.550400]
BTFIT["WV_062"] = [0.00001805700, 1.000255533, -1.790930]
BTFIT["WV_073"] = [0.00000231818, 1.000668281, -0.456166]
BTFIT["IR_087"] = [-0.00002332000, 1.011803400, -1.507390]
BTFIT["IR_097"] = [-0.00002055330, 1.009370670, -1.030600]
BTFIT["IR_108"] = [-0.00007392770, 1.032889800, -3.296740]
BTFIT["IR_120"] = [-0.00007009840, 1.031314600, -3.181090]
BTFIT["IR_134"] = [-0.00007293450, 1.030424800, -2.645950]
SATNUM = {321: "8",
322: "9",
323: "10",
324: "11"}
CALIB = dict()
# Meteosat 8
CALIB[321] = {"HRV": {"F": 78.7599},
"VIS006": {"F": 65.2296},
"VIS008": {"F": 73.0127},
"IR_016": {"F": 62.3715},
"IR_039": {"VC": 2567.33,
"ALPHA": 0.9956,
"BETA": 3.41},
"WV_062": {"VC": 1598.103,
"ALPHA": 0.9962,
"BETA": 2.218},
"WV_073": {"VC": 1362.081,
"ALPHA": 0.9991,
"BETA": 0.478},
"IR_087": {"VC": 1149.069,
"ALPHA": 0.9996,
"BETA": 0.179},
"IR_097": {"VC": 1034.343,
"ALPHA": 0.9999,
"BETA": 0.06},
"IR_108": {"VC": 930.647,
"ALPHA": 0.9983,
"BETA": 0.625},
"IR_120": {"VC": 839.66,
"ALPHA": 0.9988,
"BETA": 0.397},
"IR_134": {"VC": 752.387,
"ALPHA": 0.9981,
"BETA": 0.578}}
# Meteosat 9
CALIB[322] = {"HRV": {"F": 79.0113},
"VIS006": {"F": 65.2065},
"VIS008": {"F": 73.1869},
"IR_016": {"F": 61.9923},
"IR_039": {"VC": 2568.832,
"ALPHA": 0.9954,
"BETA": 3.438},
"WV_062": {"VC": 1600.548,
"ALPHA": 0.9963,
"BETA": 2.185},
"WV_073": {"VC": 1360.330,
"ALPHA": 0.9991,
"BETA": 0.47},
"IR_087": {"VC": 1148.620,
"ALPHA": 0.9996,
"BETA": 0.179},
"IR_097": {"VC": 1035.289,
"ALPHA": 0.9999,
"BETA": 0.056},
"IR_108": {"VC": 931.7,
"ALPHA": 0.9983,
"BETA": 0.64},
"IR_120": {"VC": 836.445,
"ALPHA": 0.9988,
"BETA": 0.408},
"IR_134": {"VC": 751.792,
"ALPHA": 0.9981,
"BETA": 0.561}}
# Meteosat 10
CALIB[323] = {"HRV": {"F": 78.9416},
"VIS006": {"F": 65.5148},
"VIS008": {"F": 73.1807},
"IR_016": {"F": 62.0208},
"IR_039": {"VC": 2547.771,
"ALPHA": 0.9915,
"BETA": 2.9002},
"WV_062": {"VC": 1595.621,
"ALPHA": 0.9960,
"BETA": 2.0337},
"WV_073": {"VC": 1360.337,
"ALPHA": 0.9991,
"BETA": 0.4340},
"IR_087": {"VC": 1148.130,
"ALPHA": 0.9996,
"BETA": 0.1714},
"IR_097": {"VC": 1034.715,
"ALPHA": 0.9999,
"BETA": 0.0527},
"IR_108": {"VC": 929.842,
"ALPHA": 0.9983,
"BETA": 0.6084},
"IR_120": {"VC": 838.659,
"ALPHA": 0.9988,
"BETA": 0.3882},
"IR_134": {"VC": 750.653,
"ALPHA": 0.9982,
"BETA": 0.5390}}
# Meteosat 11
CALIB[324] = {"HRV": {"F": 79.0035},
"VIS006": {"F": 65.2656},
"VIS008": {"F": 73.1692},
"IR_016": {"F": 61.9416},
"IR_039": {"VC": 2555.280,
"ALPHA": 0.9916,
"BETA": 2.9438},
"WV_062": {"VC": 1596.080,
"ALPHA": 0.9959,
"BETA": 2.0780},
"WV_073": {"VC": 1361.748,
"ALPHA": 0.9990,
"BETA": 0.4929},
"IR_087": {"VC": 1147.433,
"ALPHA": 0.9996,
"BETA": 0.1731},
"IR_097": {"VC": 1034.851,
"ALPHA": 0.9998,
"BETA": 0.0597},
"IR_108": {"VC": 931.122,
"ALPHA": 0.9983,
"BETA": 0.6256},
"IR_120": {"VC": 839.113,
"ALPHA": 0.9988,
"BETA": 0.4002},
"IR_134": {"VC": 748.585,
"ALPHA": 0.9981,
"BETA": 0.5635}}
# Calibration coefficients from Meirink, J.F., R.A. Roebeling and P. Stammes, 2013:
# Inter-calibration of polar imager solar channels using SEVIRI, Atm. Meas. Tech., 6,
# 2495-2508, doi:10.5194/amt-6-2495-2013
#
# The coeffients in the 2023 entry have been obtained from the webpage
# https://msgcpp.knmi.nl/solar-channel-calibration.html on 2023-10-11.
#
# The coefficients are stored in pairs of A, B (see function `get_meirink_slope`) where the
# units of A are µW m-2 sr-1 (cm-1)-1 and those of B are µW m-2 sr-1 (cm-1)-1 (86400 s)-1
#
# To obtain the slope for the calibration, one should use the routine get_seviri_meirink_slope
# Epoch for the MEIRINK re-calibration
MEIRINK_EPOCH = datetime(2000, 1, 1)
MEIRINK_COEFS: dict[str, dict[int, dict[str, tuple[float, float]]]] = {}
MEIRINK_COEFS["2023"] = {}
# Meteosat-8
MEIRINK_COEFS["2023"][321] = {"VIS006": (24.346, 0.3739),
"VIS008": (30.989, 0.3111),
"IR_016": (22.869, 0.0065)
}
# Meteosat-9
MEIRINK_COEFS["2023"][322] = {"VIS006": (21.026, 0.2556),
"VIS008": (26.875, 0.1835),
"IR_016": (21.394, 0.0498)
}
# Meteosat-10
MEIRINK_COEFS["2023"][323] = {"VIS006": (19.829, 0.5856),
"VIS008": (25.284, 0.6787),
"IR_016": (23.066, -0.0286)
}
# Meteosat-11
MEIRINK_COEFS["2023"][324] = {"VIS006": (20.515, 0.3600),
"VIS008": (25.803, 0.4844),
"IR_016": (22.354, -0.0187)
}
[docs]
def get_meirink_slope(meirink_coefs, acquisition_time):
"""Compute the slope for the visible channel calibration according to Meirink 2013.
S = A + B * 1.e-3* Day
S is here in µW m-2 sr-1 (cm-1)-1
EUMETSAT calibration is given in mW m-2 sr-1 (cm-1)-1, so an extra factor of 1/1000 must
be applied.
"""
A = meirink_coefs[0]
B = meirink_coefs[1]
delta_t = (acquisition_time - MEIRINK_EPOCH).total_seconds()
S = A + B * delta_t / (3600*24) / 1000.
return S/1000
[docs]
def should_apply_meirink(calib_mode, channel_name):
"""Decide whether to use the Meirink calibration coefficients."""
return "MEIRINK" in calib_mode and channel_name in ["VIS006", "VIS008", "IR_016"]
[docs]
class MeirinkCalibrationHandler:
"""Re-calibration of the SEVIRI visible channels slope (see Meirink 2013)."""
def __init__(self, calib_mode):
"""Initialize the calibration handler."""
self.coefs = MEIRINK_COEFS[calib_mode.split("-")[1]]
[docs]
def get_slope(self, platform, channel, time):
"""Return the slope using the provided calibration coefficients."""
coefs = self.coefs[platform][channel]
return get_meirink_slope(coefs, time)
[docs]
def get_cds_time(days, msecs):
"""Compute timestamp given the days since epoch and milliseconds of the day.
1958-01-01 00:00 is interpreted as fill value and will be replaced by NaT (Not a Time).
Args:
days (int, either scalar or numpy.ndarray):
Days since 1958-01-01
msecs (int, either scalar or numpy.ndarray):
Milliseconds of the day
Returns:
numpy.datetime64: Timestamp(s)
"""
if np.isscalar(days):
days = np.array([days], dtype="int64")
msecs = np.array([msecs], dtype="int64")
# use nanosecond precision to silence warning from XArray
nsecs = 1000000 * msecs.astype("timedelta64[ns]")
time = np.datetime64("1958-01-01").astype("datetime64[ms]") + \
days.astype("timedelta64[D]") + nsecs
time[time == np.datetime64("1958-01-01 00:00")] = np.datetime64("NaT")
if len(time) == 1:
return time[0]
return time
[docs]
def add_scanline_acq_time(dataset, acq_time):
"""Add scanline acquisition time to the given dataset."""
dataset.coords["acq_time"] = ("y", acq_time)
dataset.coords["acq_time"].attrs[
"long_name"] = "Mean scanline acquisition time"
[docs]
def dec10216(inbuf):
"""Decode 10 bits data into 16 bits words.
::
/*
* pack 4 10-bit words in 5 bytes into 4 16-bit words
*
* 0 1 2 3 4 5
* 01234567890123456789012345678901234567890
* 0 1 2 3 4
*/
ip = &in_buffer[i];
op = &out_buffer[j];
op[0] = ip[0]*4 + ip[1]/64;
op[1] = (ip[1] & 0x3F)*16 + ip[2]/16;
op[2] = (ip[2] & 0x0F)*64 + ip[3]/4;
op[3] = (ip[3] & 0x03)*256 +ip[4];
"""
arr10 = inbuf.astype(np.uint16)
arr16_len = int(len(arr10) * 4 / 5)
arr10_len = int((arr16_len * 5) / 4)
arr10 = arr10[:arr10_len] # adjust size
# dask is slow with indexing
arr10_0 = arr10[::5]
arr10_1 = arr10[1::5]
arr10_2 = arr10[2::5]
arr10_3 = arr10[3::5]
arr10_4 = arr10[4::5]
arr16_0 = (arr10_0 << 2) + (arr10_1 >> 6)
arr16_1 = ((arr10_1 & 63) << 4) + (arr10_2 >> 4)
arr16_2 = ((arr10_2 & 15) << 6) + (arr10_3 >> 2)
arr16_3 = ((arr10_3 & 3) << 8) + arr10_4
arr16 = np.stack([arr16_0, arr16_1, arr16_2, arr16_3], axis=-1).ravel()
return arr16
mpef_product_header = MpefProductHeader().get()
[docs]
class SEVIRICalibrationAlgorithm:
"""SEVIRI calibration algorithms."""
def __init__(self, platform_id, scan_time):
"""Initialize the calibration algorithm."""
self._platform_id = platform_id
self._scan_time = scan_time
[docs]
def convert_to_radiance(self, data, gain, offset):
"""Calibrate to radiance."""
data = data.where(data > 0)
return (data * gain + offset).clip(0.0, None)
[docs]
def _erads2bt(self, data, channel_name):
"""Convert effective radiance to brightness temperature."""
cal_info = CALIB[self._platform_id][channel_name]
alpha = cal_info["ALPHA"]
beta = cal_info["BETA"]
wavenumber = CALIB[self._platform_id][channel_name]["VC"]
return (self._tl15(data, wavenumber) - beta) / alpha
[docs]
def ir_calibrate(self, data, channel_name, cal_type):
"""Calibrate to brightness temperature."""
if cal_type == 1:
# spectral radiances
return self._srads2bt(data, channel_name)
elif cal_type == 2:
# effective radiances
return self._erads2bt(data, channel_name)
else:
raise NotImplementedError("Unknown calibration type")
[docs]
def _srads2bt(self, data, channel_name):
"""Convert spectral radiance to brightness temperature."""
a__, b__, c__ = BTFIT[channel_name]
wavenumber = CALIB[self._platform_id][channel_name]["VC"]
temp = self._tl15(data, wavenumber)
return a__ * temp * temp + b__ * temp + c__
[docs]
def _tl15(self, data, wavenumber):
"""Compute the L15 temperature."""
return ((C2 * wavenumber) /
np.log((1.0 / data) * C1 * wavenumber ** 3 + 1.0))
[docs]
def vis_calibrate(self, data, solar_irradiance):
"""Calibrate to reflectance.
This uses the method described in Conversion from radiances to
reflectances for SEVIRI warm channels: https://www-cdn.eumetsat.int/files/2020-04/pdf_msg_seviri_rad2refl.pdf
"""
reflectance = np.pi * data * 100.0 / solar_irradiance
return apply_earthsun_distance_correction(reflectance, self._scan_time)
[docs]
class SEVIRICalibrationHandler:
"""Calibration handler for SEVIRI HRIT-, native- and netCDF-formats.
Handles selection of calibration coefficients and calls the appropriate
calibration algorithm.
"""
def __init__(self, platform_id, channel_name, coefs, calib_mode, scan_time):
"""Initialize the calibration handler."""
self._platform_id = platform_id
self._channel_name = channel_name
self._coefs = coefs
self._calib_mode = calib_mode.upper()
self._scan_time = scan_time
self._algo = SEVIRICalibrationAlgorithm(
platform_id=self._platform_id,
scan_time=self._scan_time
)
valid_modes = ("NOMINAL", "GSICS", "MEIRINK-2023")
if self._calib_mode not in valid_modes:
raise ValueError(
"Invalid calibration mode: {}. Choose one of {}".format(
self._calib_mode, valid_modes)
)
[docs]
def calibrate(self, data, calibration):
"""Calibrate the given data."""
if calibration == "counts":
res = data
elif calibration in ["radiance", "reflectance",
"brightness_temperature"]:
gain, offset = self.get_gain_offset()
res = self._algo.convert_to_radiance(
data.astype(np.float32), gain, offset
)
else:
raise ValueError(
"Invalid calibration {} for channel {}".format(
calibration, self._channel_name
)
)
if calibration == "reflectance":
solar_irradiance = CALIB[self._platform_id][self._channel_name]["F"]
res = self._algo.vis_calibrate(res, solar_irradiance)
elif calibration == "brightness_temperature":
res = self._algo.ir_calibrate(
res, self._channel_name, self._coefs["radiance_type"]
)
return res
[docs]
def get_gain_offset(self):
"""Get gain & offset for calibration from counts to radiance.
Choices for internal coefficients are nominal or GSICS. If no
GSICS coefficients are available for a certain channel, fall back to
nominal coefficients. External coefficients take precedence over
internal coefficients.
"""
coefs = self._coefs["coefs"]
# Select internal coefficients for the given calibration mode
internal_gain = coefs["NOMINAL"]["gain"]
internal_offset = coefs["NOMINAL"]["offset"]
if self._calib_mode == "GSICS":
gsics_gain = coefs["GSICS"]["gain"]
gsics_offset = coefs["GSICS"]["offset"] * gsics_gain
if gsics_gain != 0 and gsics_offset != 0:
# If no GSICS coefficients are available for a certain channel,
# they are set to zero in the file.
internal_gain = gsics_gain
internal_offset = gsics_offset
if should_apply_meirink(self._calib_mode, self._channel_name):
meirink = MeirinkCalibrationHandler(calib_mode=self._calib_mode)
internal_gain = meirink.get_slope(self._platform_id, self._channel_name, self._scan_time)
# Override with external coefficients, if any.
gain = coefs["EXTERNAL"].get("gain", internal_gain)
offset = coefs["EXTERNAL"].get("offset", internal_offset)
return gain, offset
[docs]
def chebyshev(coefs, time, domain):
"""Evaluate a Chebyshev Polynomial.
Args:
coefs (list, np.array): Coefficients defining the polynomial
time (int, float): Time where to evaluate the polynomial
domain (list, tuple): Domain (or time interval) for which the polynomial is defined: [left, right]
Reference: Appendix A in the MSG Level 1.5 Image Data Format Description.
"""
return Chebyshev(coefs, domain=domain)(time) - 0.5 * coefs[0]
[docs]
def chebyshev_3d(coefs, time, domain):
"""Evaluate Chebyshev Polynomials for three dimensions (x, y, z).
Expects the three coefficient sets to be defined in the same domain.
Args:
coefs: (x, y, z) coefficient sets.
time: See :func:`chebyshev`
domain: See :func:`chebyshev`
Returns:
Polynomials evaluated in (x, y, z) dimension.
"""
x_coefs, y_coefs, z_coefs = coefs
x = chebyshev(x_coefs, time, domain)
y = chebyshev(y_coefs, time, domain)
z = chebyshev(z_coefs, time, domain)
return x, y, z
[docs]
class NoValidOrbitParams(Exception):
"""Exception when validOrbitParameters are missing."""
pass
[docs]
class OrbitPolynomial:
"""Polynomial encoding the satellite position.
Satellite position as a function of time is encoded in the coefficients
of an 8th-order Chebyshev polynomial.
"""
def __init__(self, coefs, start_time, end_time):
"""Initialize the polynomial."""
self.coefs = coefs
self.start_time = start_time
self.end_time = end_time
[docs]
def evaluate(self, time):
"""Get satellite position in earth-centered cartesian coordinates.
Args:
time: Timestamp where to evaluate the polynomial
Returns:
Earth-centered cartesian coordinates (x, y, z) in meters
"""
domain = [np.datetime64(self.start_time).astype("int64"),
np.datetime64(self.end_time).astype("int64")]
time = np.datetime64(time).astype("int64")
x, y, z = chebyshev_3d(self.coefs, time, domain)
return x * 1000, y * 1000, z * 1000 # km -> m
def __eq__(self, other):
"""Test equality of two orbit polynomials."""
return (
np.array_equal(self.coefs, np.array(other.coefs)) and
self.start_time == other.start_time and
self.end_time == other.end_time
)
[docs]
def get_satpos(orbit_polynomial, time, semi_major_axis, semi_minor_axis):
"""Get satellite position in geodetic coordinates.
Args:
orbit_polynomial: OrbitPolynomial instance
time: Timestamp where to evaluate the polynomial
semi_major_axis: Semi-major axis of the ellipsoid
semi_minor_axis: Semi-minor axis of the ellipsoid
Returns:
Longitude [deg east], Latitude [deg north] and Altitude [m]
"""
x, y, z = orbit_polynomial.evaluate(time)
geocent = pyproj.CRS(
proj="geocent", a=semi_major_axis, b=semi_minor_axis, units="m"
)
latlong = pyproj.CRS(
proj="latlong", a=semi_major_axis, b=semi_minor_axis, units="m"
)
transformer = pyproj.Transformer.from_crs(geocent, latlong)
lon, lat, alt = transformer.transform(x, y, z)
return lon, lat, alt
[docs]
class OrbitPolynomialFinder:
"""Find orbit polynomial for a given timestamp."""
def __init__(self, orbit_polynomials):
"""Initialize with the given candidates.
Args:
orbit_polynomials: Dictionary of orbit polynomials as found in
SEVIRI L1B files:
.. code-block:: python
{'X': x_polynomials,
'Y': y_polynomials,
'Z': z_polynomials,
'StartTime': polynomials_valid_from,
'EndTime': polynomials_valid_to}
"""
self.orbit_polynomials = orbit_polynomials
# Left/right boundaries of time intervals for which the polynomials are
# valid.
self.valid_from = orbit_polynomials["StartTime"][0, :].astype(
"datetime64[us]")
self.valid_to = orbit_polynomials["EndTime"][0, :].astype(
"datetime64[us]")
[docs]
def get_orbit_polynomial(self, time, max_delta=6):
"""Get orbit polynomial valid for the given time.
Orbit polynomials are only valid for certain time intervals. Find the
polynomial, whose corresponding interval encloses the given timestamp.
If there are multiple enclosing intervals, use the most recent one.
If there is no enclosing interval, find the interval whose centre is
closest to the given timestamp (but not more than ``max_delta`` hours
apart).
Why are there gaps between those intervals? Response from EUM:
A manoeuvre is a discontinuity in the orbit parameters. The flight
dynamic algorithms are not made to interpolate over the time-span of
the manoeuvre; hence we have elements describing the orbit before a
manoeuvre and a new set of elements describing the orbit after the
manoeuvre. The flight dynamic products are created so that there is
an intentional gap at the time of the manoeuvre. Also the two
pre-manoeuvre elements may overlap. But the overlap is not of an
issue as both sets of elements describe the same pre-manoeuvre orbit
(with negligible variations).
"""
time = np.datetime64(time)
try:
match = self._get_enclosing_interval(time)
except ValueError:
warnings.warn(
"No orbit polynomial valid for {}. Using closest "
"match.".format(time),
stacklevel=2
)
match = self._get_closest_interval_within(time, max_delta)
return OrbitPolynomial(
coefs=(
self.orbit_polynomials["X"][match],
self.orbit_polynomials["Y"][match],
self.orbit_polynomials["Z"][match]
),
start_time=self.valid_from[match],
end_time=self.valid_to[match]
)
[docs]
def _get_enclosing_interval(self, time):
"""Find interval enclosing the given timestamp."""
enclosing = np.where(
np.logical_and(
time >= self.valid_from,
time < self.valid_to
)
)[0]
most_recent = np.argmax(self.valid_from[enclosing])
return enclosing[most_recent]
[docs]
def _get_closest_interval_within(self, time, threshold):
"""Find interval closest to the given timestamp within a given distance.
Args:
time: Timestamp of interest
threshold: Maximum distance between timestamp and interval center
Returns:
Index of closest interval
"""
closest_match, distance = self._get_closest_interval(time)
threshold_diff = np.timedelta64(threshold, "h")
if distance < threshold_diff:
return closest_match
raise NoValidOrbitParams(
"Unable to find orbit coefficients valid for {} +/- {}"
"hours".format(time, threshold)
)
[docs]
def _get_closest_interval(self, time):
"""Find interval closest to the given timestamp.
Returns:
Index of closest interval, distance from its center
"""
intervals_centre = self.valid_from + 0.5 * (
self.valid_to - self.valid_from
)
diffs_us = (time - intervals_centre).astype("i8")
closest_match = np.argmin(np.fabs(diffs_us))
distance = abs(intervals_centre[closest_match] - time)
return closest_match, distance
# def calculate_area_extent(center_point, north, east, south, west, we_offset, ns_offset, column_step, line_step):
[docs]
def calculate_area_extent(area_dict):
"""Calculate the area extent seen by a geostationary satellite.
Args:
area_dict: A dictionary containing the required parameters
center_point: Center point for the projection
north: Northmost row number
east: Eastmost column number
west: Westmost column number
south: Southmost row number
column_step: Pixel resolution in meters in east-west direction
line_step: Pixel resolution in meters in south-north direction
[column_offset: Column offset, defaults to 0 if not given]
[line_offset: Line offset, defaults to 0 if not given]
Returns:
tuple: An area extent for the scene defined by the lower left and
upper right corners
# For Earth model 2 and full disk VISIR, (center_point - west - 0.5 + we_offset) must be -1856.5 .
# See MSG Level 1.5 Image Data Format Description Figure 7 - Alignment and numbering of the non-HRV pixels.
"""
center_point = area_dict["center_point"]
east = area_dict["east"]
west = area_dict["west"]
south = area_dict["south"]
north = area_dict["north"]
column_step = area_dict["column_step"]
line_step = area_dict["line_step"]
column_offset = area_dict.get("column_offset", 0)
line_offset = area_dict.get("line_offset", 0)
ll_c = (center_point - east + 0.5 + column_offset) * column_step
ll_l = (north - center_point + 0.5 + line_offset) * line_step
ur_c = (center_point - west - 0.5 + column_offset) * column_step
ur_l = (south - center_point - 0.5 + line_offset) * line_step
return (ll_c, ll_l, ur_c, ur_l)
[docs]
def create_coef_dict(coefs_nominal, coefs_gsics, radiance_type, ext_coefs):
"""Create coefficient dictionary expected by calibration class."""
return {
"coefs": {
"NOMINAL": {
"gain": coefs_nominal[0],
"offset": coefs_nominal[1],
},
"GSICS": {
"gain": coefs_gsics[0],
"offset": coefs_gsics[1]
},
"EXTERNAL": ext_coefs
},
"radiance_type": radiance_type
}
[docs]
def get_padding_area(shape, dtype):
"""Create a padding area filled with no data."""
if np.issubdtype(dtype, np.floating):
init_value = np.nan
else:
init_value = 0
padding_area = da.full(shape, init_value, dtype=dtype, chunks=CHUNK_SIZE)
return padding_area
[docs]
def pad_data_horizontally(data, final_size, east_bound, west_bound):
"""Pad the data given east and west bounds and the desired size."""
nlines = final_size[0]
if west_bound - east_bound != data.shape[1] - 1:
raise IndexError("East and west bounds do not match data shape")
padding_east = get_padding_area((nlines, east_bound - 1), data.dtype)
padding_west = get_padding_area((nlines, (final_size[1] - west_bound)), data.dtype)
return np.hstack((padding_east, data, padding_west))
[docs]
def pad_data_vertically(data, final_size, south_bound, north_bound):
"""Pad the data given south and north bounds and the desired size."""
ncols = final_size[1]
if north_bound - south_bound != data.shape[0] - 1:
raise IndexError("South and north bounds do not match data shape")
padding_south = get_padding_area((south_bound - 1, ncols), data.dtype)
padding_north = get_padding_area(((final_size[0] - north_bound), ncols), data.dtype)
return np.vstack((padding_south, data, padding_north))
[docs]
def _create_bad_quality_lines_mask(line_validity, line_geometric_quality, line_radiometric_quality):
"""Create bad quality scan lines mask.
For details on quality flags see `MSG Level 1.5 Image Data Format Description`_
page 109.
Args:
line_validity (numpy.ndarray):
Quality flags with shape (nlines,).
line_geometric_quality (numpy.ndarray):
Quality flags with shape (nlines,).
line_radiometric_quality (numpy.ndarray):
Quality flags with shape (nlines,).
Returns:
numpy.ndarray: Indicating if the scan line is bad.
"""
# Based on missing (2) or corrupted (3) data
line_mask = line_validity >= 2
line_mask &= line_validity <= 3
# Do not use (4)
line_mask &= line_radiometric_quality == 4
line_mask &= line_geometric_quality == 4
return line_mask
[docs]
def mask_bad_quality(data, line_validity, line_geometric_quality, line_radiometric_quality):
"""Mask scan lines with bad quality.
Args:
data (xarray.DataArray):
Channel data
line_validity (numpy.ndarray):
Quality flags with shape (nlines,).
line_geometric_quality (numpy.ndarray):
Quality flags with shape (nlines,).
line_radiometric_quality (numpy.ndarray):
Quality flags with shape (nlines,).
Returns:
xarray.DataArray: data with lines flagged as bad converted to np.nan.
"""
line_mask = _create_bad_quality_lines_mask(line_validity, line_geometric_quality, line_radiometric_quality)
line_mask = line_mask[:, np.newaxis]
data = data.where(~line_mask, np.nan).astype(np.float32)
return data
[docs]
def round_nom_time(dt, time_delta):
"""Round a datetime object to a multiple of a timedelta.
dt : datetime.datetime object, default now.
time_delta : timedelta object, we round to a multiple of this, default 1 minute.
adapted for SEVIRI from:
https://stackoverflow.com/questions/3463930/how-to-round-the-minute-of-a-datetime-object-python
"""
seconds = (dt - dt.min).seconds
round_to = time_delta.total_seconds()
rounding = (seconds + round_to / 2) // round_to * round_to
return dt + timedelta(0, rounding - seconds, - dt.microsecond)