VIIRS/AQUA Satellite Ocean Color

Overview

The OCEAN COLOR (NASA’s ocean color web) supports the collection, processing, calibration, archive and distribution of ocean-related products from a number of missions which are supported within the framework and facilities of the NASA Ocean Data Processing System. This converter support ocean color data from two different missions:

1. AQUA/MODIS (or Moderate Resolution Imaging Spectroradiometer) is a key instrument aboard the Terra (EOS AM) and Aqua (EOS PM). Terra MODIS and Aqua MODIS view the entire Earth’s surface every 2 days.

2. VIIRS (or Visible and Infrared Imager/Radiometer Suite) is a multi-disciplinary instrument that is being flown on the Joint Polar Satellite System. VIIRS is the successor to MODIS for Earth Science data product generation.

Data source

In order to have access to the data, you will need a log in on: https://earthdata.nasa.gov/

Once a username and a password have been created, the ocean color data can be downloaded from: data-access-webpage where a manual file search can be performed or using the provided script: shell_scripts/get_ocdata.sh

Programs and Scripts

The programs and scripts in the observations/obs_converters/ocean_color directory download the data and create DART observation sequence (obs_seq) files. All programs are built in the ocean_color/work directory and can be done by running ./quickbuild.sh In addition to the main converter (i.e., convert_sat_chl), other programs such advance_time and preprocess utilities will be built.

Converter namelist convert_sat_chl_nml: This namelist is added to the rest of DART program namelists in file input.nml. Namelists start with an ampersand ‘&’ and terminate with a slash ‘/’.

&convert_sat_chl_nml
   file_in         = '../data/V2020350.L3m_DAY_SNPP_CHL_chlor_a_4km.nc'
   file_out        = '../data/obs_seq_chl'
   chl_thresh      = 0.03
   subsample_intv  = 1
   special_mask    = .true.
   debug           = .false.
   /

Contents	Type	Description
file_in	character	Name of the input netcdf data file. Example of absolute path: `$DART/observations/obs_converters/ocean_color/data/V2020336.L3m_DAY_SNPP_CHL_4km.nc`
file_out	character	Partial filename for the output file. The date and time are appended to `file_out` to construct a unique filename reflecting the time of the observations in the file.
chl_thresh	real(r8)	When the observed chlorophyll values are small, a threshold value is used for the obs. Example: `chl_thresh = 0.03`
subsample_intv	integer	It is possible to ‘thin’ the observations. `subsample_intv` allows one to take every nth observation.
special_mask	logical	A simple procedure to ignore data in certain areas. Users can the code in convert_sat_chl.f90 to change the mask location
debug	logical	Print extra information during the `convert_sat_chl` execution.

It’s widely known that chlorophyll follows a log-normal distribution. After reading the observed chlorophyll data from the input netcdf file, a LOG10 transformation is applied such that the resulting distribution is normal. The idea behind this step is to prepare the data for the EnKF which assumes both the state and observations to be Gaussian.

The get_ocdata.sh script is placed inside the shell_scripts directory. Technically, this is the only the script that the user needs to run. Prior to running the script, one should edit some parameters such as: date range, the resolution of the data, the instrument, username and password to access the data site, frequency, domain area coordinates, etc. These parameters are hard coded in the script. The tasks that this script do are:

It Downloads ocean color data for the requested period and frequency one file at a time.
It retains the chlorophyll variable for the requested domain and gets rid of unnecessary information in the netcdf data files.
It modifies the netcdf files by adding a time dimension with the necessary attributes.
It runs the obs converter convert_sat_chl for each data file.

So, each data file will have a corresponding obs_seq file. The observation sequence files are named such that the observation time is part of the file name. For example: obs_seq_chl.2020-12-15-00000

This script is NOT intended to be turnkey script; it will certainly needs to be customized for your use. This script will make use of some DART utilities and namelists such as advance_time and input.nml. If an existing observation sequence file of the same output name is found when the converter is run again, it will replace that file.