.. DO NOT EDIT. .. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY. .. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE: .. "examples/quickstart_low_level_api.py" .. LINE NUMBERS ARE GIVEN BELOW. .. only:: html .. note:: :class: sphx-glr-download-link-note :ref:`Go to the end ` to download the full example code .. rst-class:: sphx-glr-example-title .. _sphx_glr_examples_quickstart_low_level_api.py: Quickstart guide low-level api ============================== This notebook is designed as quickstart guide for the low-level api of :py:mod:`pylawr`. .. GENERATED FROM PYTHON SOURCE LINES 12-16 Introduction ------------ In the following, we will use the low-level api to process data from the ``HHG`` X-band weather radar. .. GENERATED FROM PYTHON SOURCE LINES 18-26 The following processing steps are: 1. Read in data with pre-defined data handler for data, 2. Define a corresponding grid which represents the position of the data, 3. Remove noise, 4. Filter out clutter, 5. Interpolate missing values, 6. Remap the data to a new grid, and 7. Plot the data. .. GENERATED FROM PYTHON SOURCE LINES 28-31 Import packages --------------- Package including the LawrHandler .. GENERATED FROM PYTHON SOURCE LINES 31-56 .. code-block:: default import pylawr.datahandler # Package including all grids import pylawr.grid # For filters import pylawr.transform # For beamexpansion tag import pylawr.transform.spatial.beamexpansion # For interpolation import pylawr.remap # For plotting import pylawr.plot import pylawr.plot.layer # For filters from wradlib import wradlib # For array operations import numpy as np # For plotting purpose import matplotlib.colors as mpl_colors # For georeferencing within plotting import cartopy.crs as ccrs .. GENERATED FROM PYTHON SOURCE LINES 57-63 1. Read in data --------------- To read in the ascii data, we will use a pre-defined ascii data handler. For examples to read in other file types, please take a look to other examples.This data handler will get an opened file handler such that we have to open the file beforehand. .. GENERATED FROM PYTHON SOURCE LINES 63-67 .. code-block:: default file_path = '../tests/data/lawr_data.txt' opened_file = open(file_path, mode='r') .. GENERATED FROM PYTHON SOURCE LINES 68-71 Different data handlers are available in the ``datahandler`` subpackage of :py:mod:`pylawr`. To read in ascii files generated with Universität Hamburg weather radar, we will use ``LawrHandler`` .. GENERATED FROM PYTHON SOURCE LINES 71-74 .. code-block:: default data_handler = pylawr.datahandler.LawrHandler(opened_file) .. GENERATED FROM PYTHON SOURCE LINES 75-78 The next step is to decode the data. Every data handler has a unified method (``get_reflectivity``) to get the reflectivity as xarray.DataArray in ``dBZ`` from the opened file. .. GENERATED FROM PYTHON SOURCE LINES 78-82 .. code-block:: default refl_raw = data_handler.get_reflectivity() print(refl_raw) .. rst-class:: sphx-glr-script-out .. code-block:: none array([[[-30.5125, 39.5594, 36.9296, ..., 22.7212, 23.1982, 22.7786], [-30.3187, 39.5143, 36.2365, ..., 22.7075, 22.4898, 22.6836], [-30.0702, 39.485 , 35.9452, ..., 22.5073, 22.5032, 22.5006], ..., [-30.3587, 39.4875, 37.1375, ..., 22.6017, 22.9643, 23.0368], [-30.4531, 39.3453, 37.0415, ..., 22.7306, 23.7097, 23.6428], [-30.5855, 39.4076, 37.017 , ..., 22.6635, 24.0049, 23.4478]]], dtype=float32) Coordinates: * time (time) datetime64[ns] 2017-07-20T08:46:30 * azimuth (azimuth) int64 0 1 2 3 4 5 6 7 ... 352 353 354 355 356 357 358 359 * range (range) int64 0 1 2 3 4 5 6 7 8 ... 325 326 327 328 329 330 331 332 Attributes: (12/28) unit: dBZ radartype: LAWR datestr: 170720084630 timezone: UTC ave: 30.0 smpl: 5000000.0 ... ... Elevation: 3 institution: Meteorological Institute, Hamburg, Germany title: LAWR data latitude: 53.56833 longitude: 9.973997 corr_offset: 29.5364 .. GENERATED FROM PYTHON SOURCE LINES 83-88 As we can see, the data handler decodes the data into three dimensions. The time is inferred from ``datestr`` within the metadata of the file. The data handler couldn't decode the azimuth and range from this ascii-file and we have to set them later with a grid. The metadata within the file are set as attributes of this data array. .. GENERATED FROM PYTHON SOURCE LINES 90-94 As last step in this read in procedure, we need to close the opened file. For the opened text file, this doesn't matter but for binary files, like netCDF or HDF5-files, this is a crucial step because else the file will be corrupted. .. GENERATED FROM PYTHON SOURCE LINES 94-97 .. code-block:: default opened_file.close() .. GENERATED FROM PYTHON SOURCE LINES 98-104 2. Define a grid ---------------- The next large step is to define a grid, which is used as georeference of the data. Some data handlers can decode the grid from the metadata of the opened file. Within the ascii-file there is not enough information to decode the grid and we have to define it manually. .. GENERATED FROM PYTHON SOURCE LINES 106-116 Almost every weather radar works in polar coordinates. For this, we can use ``PolarGrid``. The center of this grid describes the origin of this grid. For ``PolarGrid``, this is the position of the radar in (center=(latitude, longitude, height)). We further need to define the beam elevation (beam_ele) in degrees from the earth, number of ranges (nr_ranges), the range resolution in meters (range_res) and the number of azimuth angles (nr_azi). We further can set an offset for the azimuth angles (azi_offset, in degrees) and the ranges (range_offset, in meters) if the radar is rotated or valid only after some meters. .. GENERATED FROM PYTHON SOURCE LINES 116-128 .. code-block:: default center = (53.56796, 9.97451, 95) grid = pylawr.grid.PolarGrid( center=center, beam_ele=3, nr_ranges=333, range_res=60, range_offset=0, nr_azi=360, azi_offset=0 ) .. GENERATED FROM PYTHON SOURCE LINES 129-132 We now defined a grid to the corresponding data. We can now set the coordinates of this grid to our data with ``lawr.set_grid_coordinates(grid)``. .. GENERATED FROM PYTHON SOURCE LINES 132-136 .. code-block:: default refl_raw = refl_raw.lawr.set_grid_coordinates(grid) print(refl_raw) .. rst-class:: sphx-glr-script-out .. code-block:: none array([[[-30.5125, 39.5594, 36.9296, ..., 22.7212, 23.1982, 22.7786], [-30.3187, 39.5143, 36.2365, ..., 22.7075, 22.4898, 22.6836], [-30.0702, 39.485 , 35.9452, ..., 22.5073, 22.5032, 22.5006], ..., [-30.3587, 39.4875, 37.1375, ..., 22.6017, 22.9643, 23.0368], [-30.4531, 39.3453, 37.0415, ..., 22.7306, 23.7097, 23.6428], [-30.5855, 39.4076, 37.017 , ..., 22.6635, 24.0049, 23.4478]]], dtype=float32) Coordinates: * time (time) datetime64[ns] 2017-07-20T08:46:30 * azimuth (azimuth) float64 0.5 1.5 2.5 3.5 4.5 ... 356.5 357.5 358.5 359.5 * range (range) float64 30.0 90.0 150.0 ... 1.983e+04 1.989e+04 1.995e+04 Attributes: (12/28) unit: dBZ radartype: LAWR datestr: 170720084630 timezone: UTC ave: 30.0 smpl: 5000000.0 ... ... Elevation: 3 institution: Meteorological Institute, Hamburg, Germany title: LAWR data latitude: 53.56833 longitude: 9.973997 corr_offset: 29.5364 .. GENERATED FROM PYTHON SOURCE LINES 137-141 3. Remove noise --------------- In this package, noise is defined as the 10. percentile of the raw uncorrected reflectivity. .. GENERATED FROM PYTHON SOURCE LINES 143-150 Caused by beam expansion the radar beam losses its intensity through the ranges. In the raw ASCII-data, this effect is already corrected. The data array has a tagging system, where tags can be added and removed from the data array. The next step is to tag the array that it is already corrected. This tag is necessary for the beam expansion filter to determine the direction of the correction. To add a tag to the DataArray, we will use ``lawr.add_tag(TAG_NAME)``. .. GENERATED FROM PYTHON SOURCE LINES 150-156 .. code-block:: default refl_raw.lawr.add_tag( pylawr.transform.spatial.beamexpansion.TAG_BEAM_EXPANSION_CORR ) print(refl_raw) .. rst-class:: sphx-glr-script-out .. code-block:: none array([[[-30.5125, 39.5594, 36.9296, ..., 22.7212, 23.1982, 22.7786], [-30.3187, 39.5143, 36.2365, ..., 22.7075, 22.4898, 22.6836], [-30.0702, 39.485 , 35.9452, ..., 22.5073, 22.5032, 22.5006], ..., [-30.3587, 39.4875, 37.1375, ..., 22.6017, 22.9643, 23.0368], [-30.4531, 39.3453, 37.0415, ..., 22.7306, 23.7097, 23.6428], [-30.5855, 39.4076, 37.017 , ..., 22.6635, 24.0049, 23.4478]]], dtype=float32) Coordinates: * time (time) datetime64[ns] 2017-07-20T08:46:30 * azimuth (azimuth) float64 0.5 1.5 2.5 3.5 4.5 ... 356.5 357.5 358.5 359.5 * range (range) float64 30.0 90.0 150.0 ... 1.983e+04 1.989e+04 1.995e+04 Attributes: (12/29) unit: dBZ radartype: LAWR datestr: 170720084630 timezone: UTC ave: 30.0 smpl: 5000000.0 ... ... institution: Meteorological Institute, Hamburg, Germany title: LAWR data latitude: 53.56833 longitude: 9.973997 corr_offset: 29.5364 tags: beam-expansion-corr .. GENERATED FROM PYTHON SOURCE LINES 157-160 In the output, we can see that the beam expansion tag is added to the DataArray as attribute. These tags can be used to track the processing steps of this DataArray. .. GENERATED FROM PYTHON SOURCE LINES 162-166 The noise is calculated and removed in the raw, uncorrected data and we need to revert the beam expansion correction. For this, we will use the ``BeamExpansion`` filter, which determines the direction of the correction automatically based on set tags. .. GENERATED FROM PYTHON SOURCE LINES 166-169 .. code-block:: default beam_expansion = pylawr.transform.spatial.BeamExpansion() .. GENERATED FROM PYTHON SOURCE LINES 170-173 Every filter within ``pylawr.filter`` has a ``transform`` method, which applies the filter to given DataArray. Normally, the transform method returns a data array as first value. .. GENERATED FROM PYTHON SOURCE LINES 173-177 .. code-block:: default refl_uncorr = beam_expansion.transform(refl_raw.lawr.to_z()) print(refl_uncorr) .. rst-class:: sphx-glr-script-out .. code-block:: none array([[[9.87432690e-07, 1.11546200e+00, 2.19168186e-01, ..., 4.75854912e-07, 5.27898419e-07, 4.76402894e-07], [1.03249357e-06, 1.10393892e+00, 1.86839019e-01, ..., 4.74356230e-07, 4.48446337e-07, 4.66094612e-07], [1.09329535e-06, 1.09651632e+00, 1.74717969e-01, ..., 4.52985833e-07, 4.49832044e-07, 4.46862869e-07], ..., [1.02302777e-06, 1.09714735e+00, 2.29915148e-01, ..., 4.62939832e-07, 5.00219414e-07, 5.05584953e-07], [1.00103091e-06, 1.06180495e+00, 2.24888585e-01, ..., 4.76886085e-07, 5.93882371e-07, 5.81291803e-07], [9.70973504e-07, 1.07714735e+00, 2.23623372e-01, ..., 4.69574589e-07, 6.35653533e-07, 5.55768802e-07]]]) Coordinates: * time (time) datetime64[ns] 2017-07-20T08:46:30 * azimuth (azimuth) float64 0.5 1.5 2.5 3.5 4.5 ... 356.5 357.5 358.5 359.5 * range (range) float64 30.0 90.0 150.0 ... 1.983e+04 1.989e+04 1.995e+04 Attributes: (12/29) unit: dBZ radartype: LAWR datestr: 170720084630 timezone: UTC ave: 30.0 smpl: 5000000.0 ... ... institution: Meteorological Institute, Hamburg, Germany title: LAWR data latitude: 53.56833 longitude: 9.973997 corr_offset: 29.5364 tags: beam-expansion-uncorr .. GENERATED FROM PYTHON SOURCE LINES 178-185 After, we uncorrected the beam expansion, we can remove noise from our radar field. The noise is removed with a ``pylawr.filter.NoiseRemover``. A noise remover needs an appropriate noise level. For this, the noise remover can be fitted. In current implementation, the noise level is a running median of the last ten determined noise levels. For this single test case, the noise level equals the determined noise level from given reflectivity. .. GENERATED FROM PYTHON SOURCE LINES 185-191 .. code-block:: default noise_remover = pylawr.transform.temporal.NoiseRemover() print('Before fit: {0:.3E}'.format(noise_remover.noiselevel)) noise_remover.fit(refl_uncorr) print('After fit: {0:.3E}'.format(noise_remover.noiselevel)) .. rst-class:: sphx-glr-script-out .. code-block:: none Before fit: 8.000E-07 After fit: 4.607E-07 .. GENERATED FROM PYTHON SOURCE LINES 192-196 After, we fitted the noise remover, we can use it to remove the noise from given reflectivity field. After we removed the noise, we need to recorrect the beam expansion. Finally, we translate our radar field into dbZ with `field.lawr.to_dbz()`. .. GENERATED FROM PYTHON SOURCE LINES 196-200 .. code-block:: default refl_wo_noise_uncorr = noise_remover.transform(refl_uncorr) refl_wo_noise = beam_expansion.transform(refl_wo_noise_uncorr).lawr.to_dbz() .. GENERATED FROM PYTHON SOURCE LINES 201-210 4. Filter out clutter --------------------- After we removed noise from our radar field, we can filter out clutter. For this we use techniques from computer vision to identify measured points, where the values and/or gradients are jumping. These points are removed by the TDBZ and SPIN filter and replaced with a NaN value. To further filter out clutter, we also use another clutter filter from `wradlib` called "Gabella filter". This should show how a clutter filter outside of pylawr can be used to filter clutter. .. GENERATED FROM PYTHON SOURCE LINES 212-215 In our package, clutter filters create a map, which will be translated to a binary clutter map via thresholding. The created cluttermaps can be combined such that different combinations of cluttermaps are possible. .. GENERATED FROM PYTHON SOURCE LINES 217-219 As first step, we need to initialize the clutter filters, here the TDBZ and SPIN filter. .. GENERATED FROM PYTHON SOURCE LINES 219-223 .. code-block:: default tdbz_filter = pylawr.transform.filter.TDBZFilter() spin_filter = pylawr.transform.filter.SPINFilter() .. GENERATED FROM PYTHON SOURCE LINES 224-226 Every clutter filter has `create_cluttermap` as method, which creates a `ClutterMap` object, where the the cluttermap is saved as array. .. GENERATED FROM PYTHON SOURCE LINES 226-230 .. code-block:: default tdbz_cmap = tdbz_filter.create_cluttermap(refl_wo_noise) spin_cmap = spin_filter.create_cluttermap(refl_wo_noise) .. GENERATED FROM PYTHON SOURCE LINES 231-238 As previously written, we further want to use a clutter filter, which is not within our package. For this example, we use the "Gabella filter" from `wradlib`. This filter returns a boolean clutter map, which is then converted into a `ClutterMap` object. For this, the clutter map needs to be translated into an integer array and need to have three dimensions (time, grid_dim_1, grid_dim_2). After, we converted the clutter map into a `ClutterMap` object, we can use this object as native clutter map. .. GENERATED FROM PYTHON SOURCE LINES 238-247 .. code-block:: default gabella_arr = wradlib.clutter.filter_gabella( refl_wo_noise.values[0], wsize=7, thrsnorain=0., tr1=6., n_p=12, tr2=1.3, rm_nans=False, radial=False, cartesian=False )[np.newaxis, ...] gabella_cmap = pylawr.transform.filter.ClutterMap('Gabella', gabella_arr.astype(int)) .. GENERATED FROM PYTHON SOURCE LINES 248-253 Cluttermaps can be combined via `append()`. This method appends a given cluttermap to existing cluttermaps. To get a clean cluttermap, we initialize an empty `ClutterMap` object, where then the other cluttermaps are appended. The fuzzy threshold can be set to get a fuzzy identification of clutter. .. GENERATED FROM PYTHON SOURCE LINES 253-260 .. code-block:: default cluttermap = pylawr.transform.filter.ClutterMap('ClutterMap', fuzzy_threshold=0.) cluttermap.append(tdbz_cmap) cluttermap.append(spin_cmap) cluttermap.append(gabella_cmap) .. GENERATED FROM PYTHON SOURCE LINES 261-263 After we appended all clutter maps, we can transform a given radar field with this clutter map. This will set all clutter pixels to NaN. .. GENERATED FROM PYTHON SOURCE LINES 263-266 .. code-block:: default refl_filtered = cluttermap.transform(refl_wo_noise) .. GENERATED FROM PYTHON SOURCE LINES 267-274 5. Interpolate missing values ----------------------------- After, we filtered out clutter values, we need to interpolate missing values. To interpolate missing values, we can use an ``Interpolator``, defined in ``pylawr.filter``. This interpolator uses a given remapping algorithm to interpolate missing values. If no remapping algorithm is set, the interpolator uses a single nearest neighbor interpolation. .. GENERATED FROM PYTHON SOURCE LINES 276-279 In this example, we use a nearest neighbor interpolation with five neighbors, which are then averaged with a median. After we defined the algorithm, we initialize an `Interpolator`. .. GENERATED FROM PYTHON SOURCE LINES 279-283 .. code-block:: default int_remapper = pylawr.remap.NearestNeighbor(5, max_dist=500) int_transformer = pylawr.transform.spatial.Interpolator(algorithm=int_remapper) .. GENERATED FROM PYTHON SOURCE LINES 284-289 The interpolator is then used to fill missing values via ``Interpolator.transform()``. This method refits given remapper and interpolates from existing to missing grid points. This interpolated result is then used to fill missing values. This step may take several seconds, because the remapping algorithm is refitted. .. GENERATED FROM PYTHON SOURCE LINES 289-292 .. code-block:: default refl_interpolated = int_transformer.transform(refl_filtered) .. rst-class:: sphx-glr-script-out .. code-block:: none /home/docs/checkouts/readthedocs.org/user_builds/pylawr/conda/stable/lib/python3.11/site-packages/xarray/core/coordinates.py:41: FutureWarning: Updating MultiIndexed coordinate 'grid_cell' would corrupt indices for other variables: ['azimuth', 'range']. This will raise an error in the future. Use `.drop_vars({'range', 'azimuth', 'grid_cell'})` before assigning new coordinate values. self.update({key: value}) /home/docs/checkouts/readthedocs.org/user_builds/pylawr/conda/stable/lib/python3.11/site-packages/pylawr/remap/nearest.py:45: RuntimeWarning: All-NaN slice encountered remapped_data = np.nanmedian(neighbor_values, axis=-1) /home/docs/checkouts/readthedocs.org/user_builds/pylawr/conda/stable/lib/python3.11/site-packages/xarray/core/coordinates.py:41: FutureWarning: Updating MultiIndexed coordinate 'grid_cell' would corrupt indices for other variables: ['grid_lat', 'grid_lon']. This will raise an error in the future. Use `.drop_vars({'grid_lat', 'grid_lon', 'grid_cell'})` before assigning new coordinate values. self.update({key: value}) .. GENERATED FROM PYTHON SOURCE LINES 293-299 6. Remap the data to a new grid ------------------------------- After we interpolated missing values, we can remap our data to a new grid, e.g. a cartesian grid. This is needed to combine different radar fields and for nowcasting purpose. In this example, we will used it to visualize the remapping effects. .. GENERATED FROM PYTHON SOURCE LINES 301-303 As first step for remapping, we need to initialize a remapping object, here nearest neighbour with six averaged neighbors. .. GENERATED FROM PYTHON SOURCE LINES 303-306 .. code-block:: default grid_remapper = pylawr.remap.NearestNeighbor(1, max_dist=300) .. GENERATED FROM PYTHON SOURCE LINES 307-312 This remapping algorithm need to be fitted to translate given radar field with given old grid to a target grid with ``fit()``. As target grid, a cartesian grid is used here. This fit procedure is necessary because the finding of the nearest neighbors is the most time-expensive step within the fitting algorithms. .. GENERATED FROM PYTHON SOURCE LINES 312-316 .. code-block:: default target_grid = pylawr.grid.CartesianGrid(start=-25000) grid_remapper.fit(grid, target_grid) .. GENERATED FROM PYTHON SOURCE LINES 317-319 After we found the nearest neighbors with ``fit()``, we can use ``remap()`` to remap given radar field from old grid to a new target grid. .. GENERATED FROM PYTHON SOURCE LINES 319-323 .. code-block:: default refl_remapped = grid_remapper.remap(refl_interpolated) print(refl_remapped) .. rst-class:: sphx-glr-script-out .. code-block:: none /home/docs/checkouts/readthedocs.org/user_builds/pylawr/conda/stable/lib/python3.11/site-packages/pylawr/remap/nearest.py:45: RuntimeWarning: All-NaN slice encountered remapped_data = np.nanmedian(neighbor_values, axis=-1) array([[[nan, nan, nan, ..., nan, nan, nan], [nan, nan, nan, ..., nan, nan, nan], [nan, nan, nan, ..., nan, nan, nan], ..., [nan, nan, nan, ..., nan, nan, nan], [nan, nan, nan, ..., nan, nan, nan], [nan, nan, nan, ..., nan, nan, nan]]]) Coordinates: * y (y) float64 -2.5e+04 -2.49e+04 -2.48e+04 ... 1.49e+04 1.5e+04 * x (x) float64 -2.5e+04 -2.49e+04 -2.48e+04 ... 1.49e+04 1.5e+04 * time (time) datetime64[ns] 2017-07-20T08:46:30 Attributes: (12/29) unit: dBZ radartype: LAWR datestr: 170720084630 timezone: UTC ave: 30.0 smpl: 5000000.0 ... ... institution: Meteorological Institute, Hamburg, Germany title: LAWR data latitude: 53.56833 longitude: 9.973997 corr_offset: 29.5364 tags: neg-lin-refl-repl-0.0;noise-filtered;beam-expansion... .. GENERATED FROM PYTHON SOURCE LINES 324-326 This remapping is the last processing step within this example. As last step, we will plot this processed data. .. GENERATED FROM PYTHON SOURCE LINES 328-337 7. Plot the data ---------------- Our plotting system is based on an object-oriented approach and matplotlib. As main plotting tool, we introduced a ``Plotter``, which is the main plot and can be split in several subplots. This main plot has a given grid (e.g. 14x14 grid points) and we can define different grid slices, which are translate into subplots. The main plotting logic is hidden into plotting layers. They can be added to a subplot and are then plotted on this specified subplot. .. GENERATED FROM PYTHON SOURCE LINES 339-344 As first step, we will translate our logarithmic reflectivity into a rain rate. For this, we use a z-r relationship with ``zr_convert``. There, the scale (a) and polynomial grade factor (b) can be specified. After this conversion, we set all values below 0.1 mm/h to NaN for plotting purpose. .. GENERATED FROM PYTHON SOURCE LINES 344-348 .. code-block:: default rain_rate = refl_remapped.lawr.to_z().lawr.zr_convert(a=256, b=1.42) rain_rate = rain_rate.where(rain_rate>=0.1, np.nan) .. GENERATED FROM PYTHON SOURCE LINES 349-353 We need to initialize a plotter and we need to specify the grid size for this plotter. In this example, we will use a 14x14 grid. We further can specify the names and grid slices for different subplots (here `map` and `colorbar`). .. GENERATED FROM PYTHON SOURCE LINES 353-369 .. code-block:: default default_gridspec_slices = { 'map': [ slice(None, None), slice(12), ], 'colorbar': [ slice(None, None), slice(12, None), ], } plotter = pylawr.plot.Plotter( grid_size=(14, 14), grid_slices=default_gridspec_slices, backend_name='agg', figsize=(13, 9) ) .. GENERATED FROM PYTHON SOURCE LINES 370-373 Within pylawr, we implemented different colormaps, which are used for our plots. Here, we will use a pre-specified rain colormap and set all NaN values to transparent. We further specify a logarithmic colorscale. .. GENERATED FROM PYTHON SOURCE LINES 373-379 .. code-block:: default cmap = pylawr.plot.available_cmaps['rain'] cmap.set_bad((0, 0, 0, 0)) norm = mpl_colors.LogNorm(vmin=0.1, vmax=200) .. GENERATED FROM PYTHON SOURCE LINES 380-381 In our next steps, we define our plotting layers. .. GENERATED FROM PYTHON SOURCE LINES 383-386 To plot a background map, we can use a `BackgroundLayer`, which will plot an openstreetmap as background (here we will use ``cartopy`` implicit). The zoom level of this openstreetmap layer can be set via `resolution`. .. GENERATED FROM PYTHON SOURCE LINES 386-389 .. code-block:: default # l_bg = pylawr.plot.layer.BackgroundLayer(resolution=10) .. GENERATED FROM PYTHON SOURCE LINES 390-395 As next layer, we define a ``RadarFieldLayer``, which is used to plot a given array and grid on a subplot. Here, we define our rain rate as plotting array with our defined cartesian grid as underlying grid. The previously defined colormap and colorscale are used for this radar layer. .. GENERATED FROM PYTHON SOURCE LINES 395-401 .. code-block:: default l_radar = pylawr.plot.layer.RadarFieldLayer(radar_field=rain_rate, grid=target_grid, zorder=1) l_radar['cmap'] = cmap l_radar['norm'] = norm .. GENERATED FROM PYTHON SOURCE LINES 402-406 As additional information, we want to add a colorbar. For this, we can use ``ColorbarLayer``, we will plot a colorbar for given radar field. The settings of this colorbar are changed after plotting, which will create this colorbar. .. GENERATED FROM PYTHON SOURCE LINES 406-409 .. code-block:: default l_cbar = pylawr.plot.layer.ColorbarLayer(l_radar) .. GENERATED FROM PYTHON SOURCE LINES 410-414 These three layers are added to specified subplots. These subplots are automatically generated based on the names of the slices (here, `map` and `colorbar`). The background and radar field are added to the `map` subplot, while the colorbar is added to `colorbar` subplot. .. GENERATED FROM PYTHON SOURCE LINES 414-419 .. code-block:: default # plotter.add_layer('map', l_bg) plotter.add_layer('map', l_radar) plotter.add_layer('colorbar', l_cbar) .. GENERATED FROM PYTHON SOURCE LINES 420-422 Our map subplot should be a georeferenced subplot. For this, we define a cartopy projection (here a rotated pole). .. GENERATED FROM PYTHON SOURCE LINES 422-425 .. code-block:: default rotated_pole = ccrs.RotatedPole(-170.415, 36.063) .. GENERATED FROM PYTHON SOURCE LINES 426-429 All subplots are saved in the plotter under ``Plotter.subplots`` as dictionary. The projection of the ``map`` subplot is then set to the previously defined projection. .. GENERATED FROM PYTHON SOURCE LINES 429-433 .. code-block:: default map_subplot = plotter.subplots.get('map') map_subplot.projection = rotated_pole .. GENERATED FROM PYTHON SOURCE LINES 434-438 We need to specify the extent of this mapping subplot. We can set the extent as dictionary with latitude and longitude values. If auto_extent of this subplot is set to true, the extent will be extended during plotting to preserve the geometry of this subplot. .. GENERATED FROM PYTHON SOURCE LINES 438-447 .. code-block:: default map_subplot.auto_extent = True map_subplot.extent = dict( lon_min=9.6, lon_max=10.3, lat_min=53.35, lat_max=53.76 ) .. GENERATED FROM PYTHON SOURCE LINES 448-451 The trigger the plotting commands, we have to call ``Plotter.plot()``. This method creates the figure and subplots and plot the layers on created subplots. .. GENERATED FROM PYTHON SOURCE LINES 451-454 .. code-block:: default plotter.plot() .. GENERATED FROM PYTHON SOURCE LINES 455-458 After, we plotted our layers, subplots and plotter, we can manipulate the figure, axes etc. Here, we manipulate the ticks of the colorbar. .. GENERATED FROM PYTHON SOURCE LINES 458-463 .. code-block:: default l_cbar.colorbar.set_ticks([0.1, 0.5, 1, 2, 5, 10, 100]) _ = l_cbar.colorbar.ax.set_yticklabels(['0.1', '0.5', '1', '2', '5', '10', '>100']) .. GENERATED FROM PYTHON SOURCE LINES 464-467 After ``plot()`` was called, the plotter has all methods from the underlying figure. Then, ``Plotter.savefig()`` can be used to save the plotted figure to a specified path. .. GENERATED FROM PYTHON SOURCE LINES 467-471 .. code-block:: default plotter.savefig('/tmp/quickstart_low_level.png') plotter.show() .. image-sg:: /examples/images/sphx_glr_quickstart_low_level_api_001.png :alt: quickstart low level api :srcset: /examples/images/sphx_glr_quickstart_low_level_api_001.png :class: sphx-glr-single-img .. GENERATED FROM PYTHON SOURCE LINES 472-477 We can see the effect of remapping in our saved plot. The remapping decreased the resolution to 100 meters from 60 meters and interpolated values, where no values are available. This is caused by nearest neighbor remapping. A more advanced algorithm, e.g. Kriging, can be used to avoid these artifacts. .. GENERATED FROM PYTHON SOURCE LINES 479-483 After we saved the figure, this quickstart example for the low-level api is finished and showed that different components of :py:mod:`pylawr` can be used to read-in, process and plot radar data. Further examples show how the plotter can be used for comparison plots and how different subpackages can interact. .. rst-class:: sphx-glr-timing **Total running time of the script:** ( 0 minutes 1.415 seconds) .. _sphx_glr_download_examples_quickstart_low_level_api.py: .. only:: html .. container:: sphx-glr-footer sphx-glr-footer-example .. container:: sphx-glr-download sphx-glr-download-python :download:`Download Python source code: quickstart_low_level_api.py ` .. container:: sphx-glr-download sphx-glr-download-jupyter :download:`Download Jupyter notebook: quickstart_low_level_api.ipynb ` .. only:: html .. rst-class:: sphx-glr-signature `Gallery generated by Sphinx-Gallery `_