(Updated Februray 2023)
Introduction
There are several NASA remote sensing aerosol products available for research and public use. This users guide will focus on the set of aerosol products created using the Dark Target (DT) algorithm(s) which is applied to space-based passive imagers.
A passive imager observes radiation that is reflected through and emitted by Earth’s systems. All passive remote sensing aerosol algorithms must contend with the problem of separating the aerosol signal from the rest of the atmosphere (clouds, molecular scattering, gas absorption) and surface signals. The DT algorithm relies on the phenomena that aerosols over a dark surface target will generally brighten the observed scene. Originally developed for use on the Moderate-Resolution Imaging Spectroradiometer (MODIS) on NASA’s Terra (launched 1999) and Aqua (launched 2002) satellites, the DT algorithm takes advantage of MODIS’s 36 spectral bands in the visible (VIS), near infrared (NIR), shortwave infrared (SWIR) and thermal infrared (TIR) that together cover the wide spectral range (0.41 – 14.2 mm). Using similar spectral coverage across 22 bands, DT has been ported to the Visible Infrared Imaging Radiometer Suite (VIIRS) sensors aboard NASA’s Suomi-NPP (SNPP; launched in 2011) and NOAA’s Joint Polar Satellite System (JPSS). Now known as NOAA-20 and NOAA-21, JPSS-1 and -2 were launched in 2017 and 2022, respectively. DT, and a sister algorithm known as Deep Blue (DB, XXXX), have been contributing to a set of standard aerosol products that are maintained by NASA.
Terra, Aqua, SNPP, N20, and N21 are all geosynchronous, Low Earth Orbits (LEO) near the poles. Each imager/satellite combination offers a wide observing swath, thus providing nearly-complete (MODIS) or complete (VIIRS) global coverage each day. Towards the poles, each LEO imager provides more than daily coverage. Nominal pixel resolution (directly underneath orbit) for most visible-wavelength bands is 0.25, 0.5 or 1.0 km (depending on band) for MODIS, and 0.375 or 0.75 km (depending on band) for VIIRS.
The 2015 launch of the Advanced Himawari Imager (AHI) on Japan’s Himawari satellite ushered in the 3rd generation (3G) of multispectral imagers in Geostationary orbit (GEO). The United States added Advanced Baseline Imager (ABI) on GOES-16 (in 2016), on GOES-17 (in 2019), and GOES-18 (in 2022). With 16 wavelength bands, these 3G imagers provide similar spectral coverage as MODIS or VIIRS. While each GEO imager provides only a regional view from a fixed vantage point over the equator, it observes multiple times each hour, every hour, every day. With nominal pixel resolution (0.5, 1.0 or 2.0 km, depending on band) comparable to MODIS or VIIRS, these new GEO imagers replace global, once-per-day with regional, 144-per-day observations.
Although not as capable spatially or spectrally, 3G imagers in GEO can support a version of the DT algorithm. Aerosol properties retrieved from ABI and AHI could be added to the existing MODIS and VIIRS data to learn about diurnal and fast aerosol changes. Thus in 2017, our team proposed to NASA’s Making Earth System Data Records for Use in Research Environments (MEaSUREs)
https://www.earthdata.nasa.gov/esds/competitive-programs/measures/leo-ge... ), and began a 5-year project aimed at Adding High Temporal Resolution to the Global Long-Term Aerosol Data Record, an attempt to create a GEO-LEO dataset that could take advantage of both GEO and LEO. By fine-tuning the DT inputs and outputs to the specifics of any sensor, but keeping a single core of subroutines and procedures, we aim to create a consistent framework for aerosol retrieval.
Note that there are already standard aerosol products from MODIS, VIIRS, and each of the GEO-imagers, including some that use the DT algorithm. For example, products retrieved using DT, along with products derived from the Deep Blue algorithm (DB, XXXX), are grouped in a MODIS standard Level 2 product known as MxD04_L2, where x=O or Y refers to MODIS on Terra or Aqua, respectively. Both DT and DB algorithms are being used for VIIRS, processed as separate AERDT_L2 (for DT) and AERDB_L2 (for DB) products. Creation of these standard products requires that the DT (and DB) algorithms are embedded within a specific “operational” processing environment, which includes toolkits and input/output procedures developed by processing teams. In fact, since MODIS products are created by the MODAPS at Goddard Space Flight Center (XXXX), and VIIRS products are produced by the Atmosphere-SIPS at University of Wisconsin, there are in fact two different flavors of operational environments. Products produced under MEaSUREs are separate from NASA’s standard processing, which required our team to develop stand-alone processing and portability across systems. .
This guide will focus on the MEaSUREs version of the DT algorithm, as well as associated Level 2 and Level 3 products. We will not go into extreme detail on the “science” of the DT algorithm (e.g., the equations used for mathematical inversion) because those definitions are already embedded within MODIS Algorithm Theoretical Basis Document (ATBD), which is available in the ATBD section of NASA’s Dark Target website https://darktarget.gsfc.nasa.gov