.. DO NOT EDIT. .. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY. .. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE: .. "auto_examples/rainfarm_downscale.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_auto_examples_rainfarm_downscale.py: Precipitation downscaling with RainFARM ======================================= This example script shows how to use the stochastic downscaling method RainFARM available in pysteps. RainFARM is a downscaling algorithm for rainfall fields developed by Rebora et al. (2006). The method can represent the realistic small-scale variability of the downscaled precipitation field by means of Gaussian random fields. Steps: 1. Read the input precipitation data. 2. Upscale the precipitation field. 3. Downscale the field to its original resolution using RainFARM with defaults. 4. Downscale with smoothing. 5. Downscale with spectral fusion. 6. Downscale with smoothing and spectral fusion. References: Rebora, N., L. Ferraris, J. von Hardenberg, and A. Provenzale, 2006: RainFARM: Rainfall downscaling by a filtered autoregressive model. J. Hydrometeor., 7, 724–738. D D'Onofrio, E Palazzi, J von Hardenberg, A Provenzale, and S Calmanti, 2014: Stochastic rainfall downscaling of climate models. J. Hydrometeorol., 15(2):830–843. .. GENERATED FROM PYTHON SOURCE LINES 30-47 .. code-block:: Python import matplotlib.pyplot as plt import numpy as np import os from pprint import pprint import logging from pysteps import io, rcparams from pysteps.utils import aggregate_fields_space, square_domain, to_rainrate from pysteps.downscaling import rainfarm from pysteps.visualization import plot_precip_field # Configure logging logging.basicConfig( level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s" ) .. GENERATED FROM PYTHON SOURCE LINES 48-53 Read the input data ------------------- As first step, we need to import the precipitation field that we are going to use in this example. .. GENERATED FROM PYTHON SOURCE LINES 53-83 .. code-block:: Python def read_precipitation_data(file_path): """Read and process precipitation data from a file.""" precip, _, metadata = io.import_mch_gif( file_path, product="AQC", unit="mm", accutime=5.0 ) precip, metadata = to_rainrate(precip, metadata) precip, metadata = square_domain(precip, metadata, "crop") return precip, metadata # Import the example radar composite root_path = rcparams.data_sources["mch"]["root_path"] filename = os.path.join(root_path, "20160711", "AQC161932100V_00005.801.gif") # Read and process data precip, metadata = read_precipitation_data(filename) # Nicely print the metadata pprint(metadata) # Plot the original rainfall field plot_precip_field(precip, geodata=metadata) plt.title("Original Rainfall Field") plt.show() # Assign the fill value to all the Nans precip[~np.isfinite(precip)] = metadata["zerovalue"] .. image-sg:: /auto_examples/images/sphx_glr_rainfarm_downscale_001.png :alt: Original Rainfall Field :srcset: /auto_examples/images/sphx_glr_rainfarm_downscale_001.png :class: sphx-glr-single-img .. rst-class:: sphx-glr-script-out .. code-block:: none {'accutime': 5.0, 'cartesian_unit': 'm', 'institution': 'MeteoSwiss', 'orig_domain': (640, 710), 'product': 'AQC', 'projection': '+proj=somerc +lon_0=7.43958333333333 +lat_0=46.9524055555556 ' '+k_0=1 +x_0=600000 +y_0=200000 +ellps=bessel ' '+towgs84=674.374,15.056,405.346,0,0,0,0 +units=m +no_defs', 'square_method': 'crop', 'threshold': np.float64(0.01155375598376629), 'transform': None, 'unit': 'mm/h', 'x1': 290000.0, 'x2': 930000.0, 'xpixelsize': 1000.0, 'y1': -160000.0, 'y2': 480000.0, 'yorigin': 'upper', 'ypixelsize': 1000.0, 'zerovalue': np.float64(0.0), 'zr_a': 316.0, 'zr_b': 1.5} /home/docs/checkouts/readthedocs.org/user_builds/pysteps/envs/v1.13.0/lib/python3.10/site-packages/pysteps/visualization/utils.py:439: UserWarning: cartopy package is required for the get_geogrid function but it is not installed. Ignoring geographical information. warnings.warn( .. GENERATED FROM PYTHON SOURCE LINES 84-91 Upscale the field ----------------- To test our downscaling method, we first need to upscale the original field to a lower resolution. This is only for demo purposes, as we need to artificially create a lower resolution field to apply our downscaling method. We are going to use a factor of 16 x. .. GENERATED FROM PYTHON SOURCE LINES 91-111 .. code-block:: Python def upscale_field(precip, metadata, scale_factor): """Upscale the precipitation field by a given scale factor.""" upscaled_resolution = metadata["xpixelsize"] * scale_factor precip_lr, metadata_lr = aggregate_fields_space( precip, metadata, upscaled_resolution ) return precip_lr, metadata_lr scale_factor = 16 precip_lr, metadata_lr = upscale_field(precip, metadata, scale_factor) # Plot the upscaled rainfall field plt.figure() plot_precip_field(precip_lr, geodata=metadata_lr) plt.title("Upscaled Rainfall Field") plt.show() .. image-sg:: /auto_examples/images/sphx_glr_rainfarm_downscale_002.png :alt: Upscaled Rainfall Field :srcset: /auto_examples/images/sphx_glr_rainfarm_downscale_002.png :class: sphx-glr-single-img .. rst-class:: sphx-glr-script-out .. code-block:: none /home/docs/checkouts/readthedocs.org/user_builds/pysteps/envs/v1.13.0/lib/python3.10/site-packages/pysteps/visualization/utils.py:439: UserWarning: cartopy package is required for the get_geogrid function but it is not installed. Ignoring geographical information. warnings.warn( .. GENERATED FROM PYTHON SOURCE LINES 112-116 Downscale the field ------------------- We can now use RainFARM to downscale the precipitation field. .. GENERATED FROM PYTHON SOURCE LINES 116-126 .. code-block:: Python # Basic downscaling precip_hr = rainfarm.downscale(precip_lr, ds_factor=scale_factor) # Plot the downscaled rainfall field plt.figure() plot_precip_field(precip_hr, geodata=metadata) plt.title("Downscaled Rainfall Field") plt.show() .. image-sg:: /auto_examples/images/sphx_glr_rainfarm_downscale_003.png :alt: Downscaled Rainfall Field :srcset: /auto_examples/images/sphx_glr_rainfarm_downscale_003.png :class: sphx-glr-single-img .. rst-class:: sphx-glr-script-out .. code-block:: none /home/docs/checkouts/readthedocs.org/user_builds/pysteps/envs/v1.13.0/lib/python3.10/site-packages/pysteps/visualization/utils.py:439: UserWarning: cartopy package is required for the get_geogrid function but it is not installed. Ignoring geographical information. warnings.warn( .. GENERATED FROM PYTHON SOURCE LINES 127-131 Downscale with smoothing ------------------------ Add smoothing with a Gaussian kernel during the downscaling process. .. GENERATED FROM PYTHON SOURCE LINES 131-142 .. code-block:: Python precip_hr_smooth = rainfarm.downscale( precip_lr, ds_factor=scale_factor, kernel_type="gaussian" ) # Plot the downscaled rainfall field with smoothing plt.figure() plot_precip_field(precip_hr_smooth, geodata=metadata) plt.title("Downscaled Rainfall Field with Gaussian Smoothing") plt.show() .. image-sg:: /auto_examples/images/sphx_glr_rainfarm_downscale_004.png :alt: Downscaled Rainfall Field with Gaussian Smoothing :srcset: /auto_examples/images/sphx_glr_rainfarm_downscale_004.png :class: sphx-glr-single-img .. rst-class:: sphx-glr-script-out .. code-block:: none /home/docs/checkouts/readthedocs.org/user_builds/pysteps/envs/v1.13.0/lib/python3.10/site-packages/pysteps/visualization/utils.py:439: UserWarning: cartopy package is required for the get_geogrid function but it is not installed. Ignoring geographical information. warnings.warn( .. GENERATED FROM PYTHON SOURCE LINES 143-147 Downscale with spectral fusion ------------------------------ Apply spectral merging as described in D'Onofrio et al. (2014). .. GENERATED FROM PYTHON SOURCE LINES 147-158 .. code-block:: Python precip_hr_fusion = rainfarm.downscale( precip_lr, ds_factor=scale_factor, spectral_fusion=True ) # Plot the downscaled rainfall field with spectral fusion plt.figure() plot_precip_field(precip_hr_fusion, geodata=metadata) plt.title("Downscaled Rainfall Field with Spectral Fusion") plt.show() .. image-sg:: /auto_examples/images/sphx_glr_rainfarm_downscale_005.png :alt: Downscaled Rainfall Field with Spectral Fusion :srcset: /auto_examples/images/sphx_glr_rainfarm_downscale_005.png :class: sphx-glr-single-img .. rst-class:: sphx-glr-script-out .. code-block:: none /home/docs/checkouts/readthedocs.org/user_builds/pysteps/envs/v1.13.0/lib/python3.10/site-packages/pysteps/visualization/utils.py:439: UserWarning: cartopy package is required for the get_geogrid function but it is not installed. Ignoring geographical information. warnings.warn( .. GENERATED FROM PYTHON SOURCE LINES 159-164 Combined Downscale with smoothing and spectral fusion ----------------------------------------------------- Apply both smoothing with a Gaussian kernel and spectral fusion during the downscaling process to observe the combined effect. .. GENERATED FROM PYTHON SOURCE LINES 164-175 .. code-block:: Python precip_hr_combined = rainfarm.downscale( precip_lr, ds_factor=scale_factor, kernel_type="gaussian", spectral_fusion=True ) # Plot the downscaled rainfall field with smoothing and spectral fusion plt.figure() plot_precip_field(precip_hr_combined, geodata=metadata) plt.title("Downscaled Rainfall Field with Gaussian Smoothing and Spectral Fusion") plt.show() .. image-sg:: /auto_examples/images/sphx_glr_rainfarm_downscale_006.png :alt: Downscaled Rainfall Field with Gaussian Smoothing and Spectral Fusion :srcset: /auto_examples/images/sphx_glr_rainfarm_downscale_006.png :class: sphx-glr-single-img .. rst-class:: sphx-glr-script-out .. code-block:: none /home/docs/checkouts/readthedocs.org/user_builds/pysteps/envs/v1.13.0/lib/python3.10/site-packages/pysteps/visualization/utils.py:439: UserWarning: cartopy package is required for the get_geogrid function but it is not installed. Ignoring geographical information. warnings.warn( .. GENERATED FROM PYTHON SOURCE LINES 176-183 Remarks ------- Currently, the pysteps implementation of RainFARM only covers spatial downscaling. That is, it can improve the spatial resolution of a rainfall field. However, unlike the original algorithm from Rebora et al. (2006), it cannot downscale the temporal dimension. .. GENERATED FROM PYTHON SOURCE LINES 183-185 .. code-block:: Python # sphinx_gallery_thumbnail_number = 2 .. rst-class:: sphx-glr-timing **Total running time of the script:** (0 minutes 1.578 seconds) .. _sphx_glr_download_auto_examples_rainfarm_downscale.py: .. only:: html .. container:: sphx-glr-footer sphx-glr-footer-example .. container:: sphx-glr-download sphx-glr-download-jupyter :download:`Download Jupyter notebook: rainfarm_downscale.ipynb ` .. container:: sphx-glr-download sphx-glr-download-python :download:`Download Python source code: rainfarm_downscale.py ` .. container:: sphx-glr-download sphx-glr-download-zip :download:`Download zipped: rainfarm_downscale.zip ` .. only:: html .. rst-class:: sphx-glr-signature `Gallery generated by Sphinx-Gallery `_