5. Composites

Compositing is the process in Geo2Grid of combining multiple products together to make a new one. Most often this is done to make RGB color images like true_color and false_color. The most common RGB recipes are already configured internally to Geo2Grid, but users can make their own combinations too. The following instructions will go over some basic examples of how to make your own composites.

One type of composite that you may want to make is an image that combines one type of product for the night side of the data and another on the day side. An example of this type of day/night composite can be found in:

$GEO2GRID_HOME/etc/polar2grid/composites/abi.yaml

This abi.yaml file is meant to hold all custom user composites for the ABI instrument. There are separate .yaml configuration files for each supported instrument available in the $GEO2GRID_HOME/etc/polar2grid/composites directory. This abi.yaml file contains the true_color_night composite recipe which combines the visible reflectance daytime true_color composite with the nighttime ABI Channel 14 C14 infrared 11 micron brightness temperatures into one image. The abi.yaml file contents are displayed below for reference:

sensor_name: visir/abi

composites:
  true_color_night:
    compositor: !!python/name:satpy.composites.DayNightCompositor
    prerequisites:
      - true_color
      - C14
    standard_name: day_night_mix

A composite recipe consists of 4 main parts:

1. Name:

   The name of the composite which will be used to request the product on the command line with the -p flag. In this example it is true_color_night. The name for a composite should be unique within a single yaml file or it may be overwritten.

2. Compositor:

   The compositor is a pointer to the python code that does the work of combining the products together. In this case we are using the DayNightCompositor code from the SatPy package. Another common option is the GenericCompositor for joining three bands together in to an RGB.

3. Inputs:

   The prerequisites are the products that are passed as inputs to this compositor. In the case of the day/night compositor the first product listed will be used for day time observations and the second product listed will be used for night time data.

4. Standard Name:

   Used later in Geo2Grid processing to map a composite to a particular enhancement or scaling. For the DayNightCompositor this should almost always be day_night_mix.

Once the composite recipe has been added to the <instrument>.yaml file it will appear in the list of available products when using the --list-products option under the ### Custom User Products heading. It can then be invoked like any other product to geo2grid.sh.

The existing true_color_night composite can be modified directly or used as a template for additional composites. Make sure to change the composite name and what prerequisites are used in the composite. After that the composite can be loaded with your data by using the following command:

$GEO2GRID_HOME/bin/geo2grid.sh -r abi_l1b -w geotiff -p true_color_night -f /path/to/files*.nc

The image created by executing the command on a GOES-16 ABI Full Disk dataset from 12:30 UTC, 12 November 2018 is shown below.

GOES-16 ABI true color day/Channel 14 brightness temperature night composite using input Full Disk observations from 12:30 UTC, 12 November 2018.

It is possible to use the compositor to combine RGBs as well. In the following example, I want to use the day/night compositor to combine the true color RGB for day data and the nighttime microphysics RGB for nighttime data. In this case, I can add the following lines to the abi.yaml file. Make sure to follow the formatting exactly, including the indentations.

true_color_night_microphysics:
  compositor: !!python/name:satpy.composites.DayNightCompositor
  prerequisites:
    - true_color
    - night_microphysics
  standard_name: day_night_mix

Once the .yaml files has been updated, the composite can be generated using the following command:

$GEO2GRID_HOME/bin/geo2grid.sh -r abi_l1b -w geotiff -p true_color_night_microphysics -f /path/to/files*.nc

The image created by executing the command on a GOES-16 ABI Full Disk dataset from 12:30 UTC, 12 November 2018 is shown below.

GOES-16 ABI true color RGB day/nighttime microphysics RGB night composite using input Full Disk observations from 12:30 UTC, 12 November 2018.