Sorry, you need to enable JavaScript to visit this website.

Developing Dark Target for VIIRS

VIIRS and MODIS

This page explains some of the theoretical work that has been done to adapt the Dark Target algorithm to VIIRS data.

Due to differences between the MODIS and VIIRS sensors and their processing systems, it is not possible to simply run the MODIS Dark Target algorithm with VIIRS input.  Native MODIS and VIIRS granules are different sizes, have different spatial resolution and store their geolocation information differently.  In order to facilitate comparison of algorithms applied to the input radiances of the two sensors, a special type of Level 1b equivalent granule was created from each sensor's Level 1b files.  These input files have nominal 1 km spatial resolution.  Because of the 1 km spatial resolution, the MODIS Collection 6 algorithm had to be modified to accept special input radiances for the comparison with VIIRS. The 1 km resolution particularly affected the internal MODIS C6 cloud mask that is based on spatial variability tests applied to 0.5 km resolution data.  Thus, the following comparison between Dark Target algorithms applied to MODIS and VIIRS will be using a MODIS-like algorithm, modified to handle the 1 km input radiances, and not the true operational C6 MODIS algorithm.

Some terminology to assist in understanding the discussion of adapting the C6 algorithm to VIIRS data. 

MODIS-like - the algorithm that resembles the C6 dark target algorithm, but has modifications allowing it to ingest and process the special 1 km resolution input files.  

MODIS-like VIIRS: The MODIS-like algorithm applied to VIIRS input data.

MODIS-like MODIS: The MODIS like algorithm applied to MODIS input data.

Before the dark target algorithm can be adapted to use with the VIIRS sensor we have to take careful note of the differences between the two sensors.

The tables below present some of the more significant differences. Their relevance and handling relative to the MODIS DT retrieval will be discussed in the VIIRS algorithm pages.

Below is a table which compares some basic details of the VIIRS and MODIS-Aqua sensors and their respective satellite platforms.

Attribute Aqua-MODIS VIIRS
MODIS and VIIRS Satellite/Sensor Comparison
Orbit Altitude 705 km 824 km
Equator Crossing Time 13:30 Local Time 13:30 Local Time
Swath Width 2330 km 3040 km
Sensor Zenith Angle Range +/- 64 ° +/- 70 °
Wavelength Bands 36 (20 in solar spectrum) 22 bands (15 in solar spectrum)
Pixel size, nadir1 1.0/0.50/0.25 km 0.75/0.375 km
Pixel size, swath edge2 4.8/2.0 1.6 x 1.6

1The pixel sizes at nadir vary according to the resolution of the individual sensor band. 

2As the view shifts from nadir to the swath edge the geographic area covered by an individual pixel increases.  This is called the "Bowtie effect" and is much more pronounced in MODIS than in VIIRS. See our MODIS algortihm page for a better description.

 

Characteristics of the MODIS and VIIRS Sensors and Products
 

Pixel Aggregation for Product

Product Size at Nadir (in Km)

Blue λ**

Green λ

Red λ

NIR λ

SWIR1.2 λ

SWIR1.6 λ

SWIR2.1 λ

Cirrus Detection λ

IR Bands for Cloud Masking

MODIS (Aqua)  20 x 20 (0.5 km pixels)        10  0.47  0.55 0.65 0.86  1.24  1.63*  2.11     1.38  6.7, 8.6, 11.0, 12.0, 13.0
VIIRS  8 x 8 (0.75 km pixels)          6  0.48  0.55 0.68 0.86 1.24  1.61  2.26     1.37  8.6, 10.8, 12.0

** All bands listed are in μm    *The MODS-Aqua 1.63 band has many bad detectors.