CHIRPS Precipitation, Copernicus DEM, (Ag)ERA5 Meteorological Data, GEOS-5 Meteorological Data, IMERG Precipitation, Landsat satellites, MODIS sensors, MSG satellites, Sentinel-2 satellites, VIIRS sensors and WorldCover Land Cover

VIIRS

The VIIRS instrument is the primary source for the Land Surface Temperature (LST) data at all levels and the primary source for optical data at 300m (L1).

Purpose of the data

Land Surface Temperature data is used for the production of the following data component:
- Land Surface Temperature for level 1.
- Input to the thermal sharpening procedure to produce soil moisture stress for level 2 and 3.
- NDVI/fAPAR composites for Level 1
- Surface albedo for Level 1

Approach

The Visible Infrared Imaging Radiometer Suite (VIIRS) instrument is aboard the joint NASA/NOAA Suomi National Polar-orbiting Partnership (Suomi NPP, launched October 2011) and National Oceanic and Atmospheric Administration (NOAA-20, launched November 2017, and NOAA-21, launched 10 November 2022) satellites. VIIRS was developed based on the experiences and lessons from both Terra and Aqua MODIS which were used for previous versions of WaPOR but have reached the end of its design life. There are currently 2 more missions planned: JPSS-3 in 2027 and JPSS-4 in 2032.

VIIRS has 22 channels ranging from 0.41 μm to 12.01 μm. Five of these channels (I1-5) are high resolution image bands (375m) and sixteen (M1-16) serve as moderate-resolution bands.

For the calculation of L1 Albedo and NDVI the VIIRS VNP09 Atmospherically Corrected Surface Reflectance products is used.

NASA provides VIIRS Land Surface Temperature and Emissivity data product based on three thermal infrared bands (M14, M15 and M16) of 750m at nadir (VPN21_L2). As the VIIRS Brightness Temperature layer (band I5) provides a higher resolution (375m) than the Land Surface Temperature and Emissivity data product (750m), it is the preferred input for LST calculations at 300m, 100m and 20m. This layer provides the brightness temperature in Kelvin (K) calculated from top-of-the-atmosphere radiances. The VIIRS Brightness Temperature layer is calculated from VIIRS Calibrated Radiances (VNP02IMG/VJ102IMG) and is available from the Suomi NPP and NOAA satellites.

Furthermore the L1 terrain corrected geolocation product (VNP03IMG and VJ103IMG) and cloud mask product (CLDMSK_L2_VIIRS_SNP) are downloaded.

Top of atmosphere measurements must be corrected for atmospheric and emissivity effects to convert brightness temperature into land surface temperature. The VIIRS 750m LST product is based on the split window technique which eliminates atmospheric effect and applies surface emissivity by using the brightness temperature at the top of atmosphere in two adjacent thermal infrared channels (M15 and M16). The split window technique cannot be applied on the single 375m brightness temperature layer (band I5). JPL has recently developed a NRT LST algorithm for the 375m VIIRS I5 band but does not yet provide a historical archive.

To calculate 375m land surface temperature from VIIRS, we calculate surface emissivity (see below) and simulate the atmospheric effects by estimating the atmospheric water vapour and air temperature profiles using a single channel technique. Jimenez-Munoz et al (2009) revised, updated and extended a single-channel algorithm for theraml infrared band 6 of Landsat 5 TM and added atmospheric transfer functions to allow land surface temperature retrieval from several Landsat thermal infrared sensors. The authors developed a solution where the atmospheric water vapour content is used to estimate the atmospheric parameters using a 2nd degree polynomial fit based on simulations with the radiative transfer model MODTRAN:

The surface emissivity accounts for the fact that the land surface is not a perfect black body (where the surface emissivity would be equal to 1). Surface emissivity may range from 0.9 for bare soils to close to 1 for vegetation. Emissivity is estimated using the simplified Normalised Difference Vegetation Index (NDVI) thresholds method (SNDVI^THM) developed by Sobrino et al (2008).

First the fractional vegetation cover (FVC) is calculated from the NDVI:

The SNDVI^THM method uses NDVI thresholds to distinguish between soil pixels (NDVI < 0.2) and pixels of full vegetation (NDVI > 0.5). For soil pixels it is assumed that the emissivity is equal to soil emissivity, estimated at 0.97. For full vegetation pixels the emissivity is equal to vegetation emissivity, estimated at 0.985. For those pixels composed of soil and vegetation (mixed pixels) the emissivity is calculated using the soil and vegetation emissivity and the fractional vegetation cover FVC:

The VIIRS sensor was designed to extend and improve upon the series of measurements initiated by its predecessors, among which the Moderate Resolution Imaging Spectroradiometer (MODIS) used in version 1 and 2. Active since 2011, VIIRS offers a relatively large archive with historical remote sensing data, and since future missions are already planned, data continuation is also ensured. This makes it a stable and consistent data source for WaPOR data components.

All VIIRS data are freely available from the Level-1 and Atmosphere Archive & Distribution System (LAADS) Distributed Active Archive Center (DAAC) platform, which provides access to global science data products as the VIIRS data.

Challenges

The spatial resolution of the LST product that can be derived from the VIIRS Brightness Temperature data product (375m) does not meet the required spatial resolution of level 2 and level 3, that is 100m and 30m respectively. Therefore, a thermal sharpening technique is needed to provide the LST inputs with the correct resolution and quality.

Alternative sensors

In case of failure of the VIIRS sensors, the following back-ups could be used, listed in order of importance:
‐ Sentinel-3, launched in 2016, with its on-board sensors OLCI and SLSTR, can provide information on land reflectance and land surface temperature in resolutions of 300 and 1000m respectively.
‐ MODIS Land Surface Temperature and Emissivity (LST/E) products provide per-pixel temperature and emissivity values on a daily basis at 1 km spatial resolution. Both Aqua and Terra MODIS have reached the end of its design life but are still operational until 2025-mid 2026 (https://nsidc.org/data/modis).