Short Name:

GHRSST NOAA/STAR Himawari-08 AHI L2P Pacific Ocean Region SST v2.70 dataset in GDS2

Himawari-8 (H08) was launched on 7 October 2014 into its nominal position at 140.7-deg E, and declared operational on 7 July 2015. The Advanced Himawari Imager (AHI; largely identical to GOES-R/ABI) is a 16 channel sensor, of which five (3.9, 8.4, 10.3, 11.2, and 12.3 um) are suitable for SST. Accurate sensor calibration, image navigation and (co)registration, high spectral fidelity, and sophisticated pre-processing (geo-rectification, radiance equalization, and mapping) offer vastly enhanced capabilities for SST retrievals, over the heritage GOES-I/P and MTSAT-2 Imagers. From altitude 35,800km, H08/AHI maps SST in a Full Disk (FD) area from 80E-160W and 60S-60N, with spatial resolution 2km at nadir to 15km at view zenith angle 67-deg, with a 10-min temporal sampling. The AHI L2P (swath) SST product is derived at the native sensor resolution using NOAA's Advanced Clear-Sky Processor for Ocean (ACSPO) system. ACSPO processes every 10-min FD data, identifies good quality ocean pixels (Petrenko et al., 2010) and derives SST using the four-band (8.4, 10.3, 11.2 and 12.3um) Non-Linear SST (NLSST) regression algorithm (Petrenko et al., 2014), trained against in situ SSTs from drifting and tropical mooring buoys in the NOAA iQuam system (Xu and Ignatov, 2014). The 10-min data are subsequently collated in time, to produce 1-hr L2P product, with improved coverage, and reduced cloud leakages and image noise. The collated L2P reports SSTs and brightness temperatures (BTs) in clear-sky water pixels (defined as ocean, sea, lake or river), and fill values elsewhere. All pixels with valid SSTs are recommended for use. ACSPO files also include sun-sensor geometry, l2p_flags (day/night, land, ice, twilight, and glint flags), and NCEP wind speed. The L2P is reported in NetCDF4 GHRSST Data Specification version 2 (GDS2) format, 24 granules per day, with a total data volume 0.6GB/day.Pixel earth locations are not reported in the granules, as they remain unchanged from granule to granule. Those can be obtained using a flat lat/lon file or a Python script available at Per GDS2 specifications, two additional Sensor-Specific Error Statistics layers (SSES bias and standard deviation) are reported in each pixel (Petrenko et al., 2016). The H08 AHI SSTs and BTs are continuously validated against in situ data in SQUAM (Dash et al, 2010), and RTM simulation in MICROS (Liang and Ignatov, 2011). A reduced size (0.2GB/day), 0.02-deg equal-angle gridded ACSPO L3C product is available at

Map of Earth