B. Petrenko

Clear-Sky Mask for the Advanced Clear-Sky Processor for Oceans

TheAdvancedClearSkyProcessorforOceans(ACSPO)generatesclear-skyproducts,suchasSST,clear-sky radiances, and aerosol, from Advanced Very High Resolution Radiometer(AVHRR)-likemeasurements. The ACSPO clear-sky mask (ACSM) identifies clear-sky pixels within the ACSPO products. This paper describes the ACSM structure and compares the performances of ACSM and its predecessor, Clouds from AVHRR Extended Algorithm (CLAVRx). ACSM essentially employs online clear-sky radiative transfer simulations enabledwithinACSPOwiththeCommunityRadiativeTransferModel(CRTM)inconjunctionwithnumerical weather prediction atmospheric [Global Forecast System (GFS)] and SST [Reynolds daily high-resolution blended SST (DSST)] fields. The baseline ACSM tests verify the accuracy of fitting observed brightness temperatureswithCRTM, checkretrievedSSTfor consistencywithReynolds SST,andidentifyambientcloudiness at the boundaries of cloudy systems. Residual cloud effects are screened out with several tests, adopted from CLAVRx, and with the SST spatial uniformity test designed to minimize misclassification of sharp SST gradients as clouds. Cross-platform and temporal consistencies of retrieved SSTs are maintained by accounting for SST and brightness temperature biases, estimated within ACSPO online and independently from ACSM. The performance of ACSM is characterized in terms of statistics of deviations of retrieved SST from the DSST. ACSMincreasestheamountof‘‘clear’’pixelsby30%to40%andimprovesstatisticsofretrievedSSTcompared with CLAVRx. ACSM is also shown to be capable of producing satisfactory statistics of SST anomalies if the reference SST field for the exact date of observations is unavailable at the time of processing.

AVHRR GAC SST Reanalysis Version 1 (RAN1)

In response to its users’ needs, the National Oceanic and Atmospheric Administration (NOAA) initiated reanalysis (RAN) of the Advanced Very High Resolution Radiometer (AVHRR) Global Area Coverage (GAC; 4 km) sea surface temperature (SST) data employing its Advanced Clear Sky Processor for Oceans (ACSPO) retrieval system. Initially, AVHRR/3 data from five NOAA and two Metop satellites from 2002 to 2015 have been reprocessed. The derived SSTs have been matched up with two reference SSTs—the quality controlled in situ SSTs from the NOAA in situ Quality Monitor (iQuam) and the Canadian Meteorological Centre (CMC) L4 SST analysis—and analyzed in the NOAA SST Quality Monitor (SQUAM) online system. The corresponding clear-sky ocean brightness temperatures (BT) in AVHRR bands 3b, 4 and 5 (centered at 3.7, 11, and 12 µm, respectively) have been compared with the Community Radiative Transfer Model simulations in another NOAA online system, Monitoring of Infrared Clear-sky Radiances over Ocean for SST (MICROS). For some AVHRRs, the time series of “AVHRR minus reference” SSTs and “observed minus model” BTs are unstable and inconsistent, with artifacts in the SSTs and BTs strongly correlated. In the official “Reanalysis version 1” (RAN1), data from only five platforms—two midmorning (NOAA-17 and Metop-A) and three afternoon (NOAA-16, -18 and -19)—were included during the most stable periods of their operations. The stability of the SST time series was further improved using variable regression SST coefficients, similarly to how it was done in the NOAA/NASA Pathfinder version 5.2 (PFV5.2) dataset. For data assimilation applications, especially those blending satellite and in situ SSTs, we recommend bias-correcting the RAN1 SSTs using the newly developed sensor-specific error statistics (SSES), which are reported in the product files. Relative performance of RAN1 and PFV5.2 SSTs is discussed. Work is underway to improve the calibration of AVHRR/3s and extend RAN time series, initially back to the mid-1990s and later to the early 1980s.

Sensor-Specific Error Statistics for SST in the Advanced Clear-Sky Processor for Oceans

AbstractThe formulation of the sensor-specific error statistics (SSES) has been redesigned in the latest implementation of the NOAA Advanced Clear-Sky Processor for Oceans (ACSPO) to enable efficient use of SSES for assimilation of the ACSPO baseline regression SST (BSST) into level 4 (L4) analyses. The SSES algorithm employs segmentation of the SST domain in the space of regressors and derives the segmentation parameter from the statistics of regressors within the global dataset of matchups. For each segment, local regression coefficients and standard deviations (SDs) of BSST minus in situ SST are calculated from the corresponding subset of matchups. The local regression coefficients are used to generate an auxiliary product—piecewise regression (PWR) SST—and SSES biases are estimated as differences between BSST and PWR SST. Correction of SSES biases, which transforms BSST back into PWR SST, reduces the effects of residual cloud; variations in view zenith angle; and, during the daytime, diurnal surface w...

Improved VIIRS and MODIS SST Imagery

Moderate Resolution Imaging Spectroradiometers (MODIS) and Visible Infrared Imaging Radiometer Suite (VIIRS) radiometers, flown onboard Terra/Aqua and Suomi National Polar-orbiting Partnership (S-NPP)/Joint Polar Satellite System (JPSS) satellites, are capable of providing superior sea surface temperature (SST) imagery. However, the swath data of these multi-detector sensors are subject to several artifacts including bow-tie distortions and striping, and require special pre-processing steps. VIIRS additionally does two irreversible data reduction steps onboard: pixel aggregation (to reduce resolution changes across the swath) and pixel deletion, which complicate both bow-tie correction and destriping. While destriping was addressed elsewhere, this paper describes an algorithm, adopted in the National Oceanic and Atmospheric Administration (NOAA) Advanced Clear-Sky Processor for Oceans (ACSPO) SST system, to minimize the bow-tie artifacts in the SST imagery and facilitate application of the pattern recognition algorithms for improved separation of ocean from cloud and mapping fine SST structure, especially in the dynamic, coastal and high-latitude regions of the ocean. The algorithm is based on a computationally fast re-sampling procedure that ensures a continuity of corresponding latitude and longitude arrays. Potentially, Level 1.5 products may be generated to benefit a wide range of MODIS and VIIRS users in land, ocean, cryosphere, and atmosphere remote sensing.

SST algorithms in ACSPO reanalysis of AVHRR GAC data from 2002-2013

In response to a request from the NOAA Coral Reef Watch Program, NOAA SST Team initiated reprocessing of 4 km resolution GAC data from AVHRRs flown onboard NOAA and MetOp satellites. The objective is to create a longterm Level 2 Advanced Clear-Sky Processor for Oceans (ACSPO) SST product, consistent with NOAA operations. ACSPO-Reanalysis (RAN) is used as input in the NOAA geo-polar blended Level 4 SST and potentially other Level 4 SST products. In the first stage of reprocessing (reanalysis 1, or RAN1), data from NOAA-15, -16, -17, -18, -19, and Metop-A and -B, from 2002-present have been processed with ACSPO v2.20, and matched up with quality controlled in situ data from in situ Quality Monitor (iQuam) version 1. The ~12 years time series of matchups were used to develop and explore the SST retrieval algorithms, with emphasis on minimizing spatial biases in retrieved SSTs, close reproduction of the magnitudes of true SST variations, and maximizing temporal, spatial and inter-platform stability of retrieval metrics. Two types of SST algorithms were considered: conventional SST regressions, and recently developed incremental regressions. The conventional equations were adopted in the EUMETSAT OSI-SAF formulation, which, according to our previous analyses, provide relatively small regional biases and well-balanced combination of precision and sensitivity, in its class. Incremental regression equations were specifically elaborated to automatically correct for model minus observation biases, always present when RTM simulations are employed. Improved temporal stability was achieved by recalculation of SST coefficients from matchups on a daily basis, with a ±45 day window around the current date. This presentation describes the candidate SST algorithms considered for the next round of ACSPO reanalysis, RAN2.

Suppressing the noise in SST retrieved from satellite infrared measurements by smoothing the differential terms in regression equations

Multichannel regression algorithms are widely used in retrievals of sea surface temperature (SST) from infrared brightness temperatures (BTs) observed from satellites. The SST equations typically include terms dependent on the difference between BTs observed in spectral bands with different atmospheric absorption. Such terms do account for variations in the variable atmospheric attenuation, but may introduce additional noise in the retrieved SST due to amplification of the radiometric noise. Some processing systems (e.g., the EUMETSAT OSI-SAF) incorporate noise suppression algorithms, based on spatial smoothing of the differential terms in the SST equations. A similar algorithm is being tested for the potential use in the NOAA Advanced Clear-Sky Processor for Oceans (ACSPO). The ACSPO smoothing algorithm aims to preserve natural variations in SST field, while minimizing distortions in the original SST imagery, at a minimal processing time. This presentation describes the ACSPO smoothing algorithm and results of its evaluation with the SST imagery, and with the in situ matchups for NOAA and Metop AVHRRs, Terra and Aqua MODISs, and SNPP/JPSS VIIRS.

Exploring pattern recognition enhancements to ACSPO clear-sky mask for VIIRS: potential and limitations

Discriminating clear-ocean from cloud in the thermal IR imagery is challenging, especially at night. Thresholds in automated cloud detection algorithms are often set conservatively leading to underestimation of the Sea Surface Temperature (SST) domain. Yet an expert user can visually distinguish the cloud patterns from SST. In this study, available pattern recognition methodologies are discussed and an automated algorithm formulated. Analyses are performed with the SSTs retrieved from the VIIRS sensor onboard S-NPP using the NOAA ACSPO system. Based on the analyses of global data, we have identified low-level spectral and spatial features potentially useful for discriminating cloud from clear-ocean. The algorithm attempts to mimic the visual perception by a human operator such as gradient information, spatial connectivity, and high/low frequency discrimination. It first identifies contiguous areas with similar features, and then makes decision based on the statistics of the whole region, rather than on a per pixel basis. Our initial objective was to automatically identify clear sky regions misclassified by ACSPO as cloud, and improve coverage of dynamic areas of the ocean and coastal zones.

Evaluation and selection of SST regression algorithms for S-NPP VIIRS

Currently, two global Level 2 SST products are generated at NOAA from S-NPP VIIRS Sensor Data Records with two independent systems, JPSS Interface Data Processing Segment (IDPS) and Advanced Clear Sky Processor for Oceans (ACSPO) using different retrieval algorithms. The two products differently correlate with in situ SST and L4 analyses, and the performance of IDPS SST is suboptimal. In this context, evaluation of existing operational SST algorithms was undertaken to select the optimal algorithm for VIIRS. This paper describes methodology and results of the evaluation. For all tested algorithms, SST accuracy and precision are estimated from matchups of VIIRS brightness temperatures with in situ SST, and sensitivity of retrieved SST to true SST is calculated using the Community Radiative Transfer Model. These three retrieval characteristics are dependent on observational conditions and show significant spatial variability. Therefore, we evaluate the SST algorithms by quantifying favorability of spatial distributions of retrieval characteristics for global SST product. We define for this purpose Quality Retrieval Domain (QRD) as a part of the World Ocean, within which SST accuracy, precision and sensitivity meet predefined specifications on retrieval characteristics. We show that, given a set of specifications, the QRD significantly varies between the algorithms. This makes QRD an informative measure of the algorithms’ performance. Based on QRD estimates for a variety of specifications, we recommend for VIIRS the algorithms developed at the EUMETSAT Ocean and Sea Ice Satellite Application Facility as ones providing the maximum QRD under reasonable specifications on retrieval characteristics.

Exploring new bands in modified multichannel regression SST algorithms for the next-generation infrared sensors at NOAA

Multichannel regression algorithms are widely used to retrieve sea surface temperature (SST) from infrared observations with satellite radiometers. Their theoretical foundations were laid in the 1980s-1990s, during the era of the Advanced Very High Resolution Radiometers which have been flown onboard NOAA satellites since 1981. Consequently, the multi-channel and non-linear SST algorithms employ the bands centered at 3.7, 11 and 12 μm, similar to available in AVHRR. More recent radiometers carry new bands located in the windows near 4 μm, 8.5 μm and 10 μm, which may also be used for SST. Involving these bands in SST retrieval requires modifications to the regression SST equations. The paper describes a general approach to constructing SST regression equations for an arbitrary number of radiometric bands and explores the benefits of using extended sets of bands available with the Visible Infrared Imager Radiometer Suite (VIIRS) flown onboard the Suomi National Polar-orbiting Partnership (SNPP) and to be flown onboard the follow-on Joint Polar Satellite System (JPSS) satellites, J1-J4, to be launched from 2017-2031; Moderate Resolution Imaging Spectroradiometers (MODIS) flown onboard Aqua and Terra satellites; and the Advanced Himawari Imager (AHI) flown onboard the Japanese Himawari-8 satellite (which in turn is a close proxy of the Advanced Baseline Imager (ABI) to be flown onboard the future Geostationary Operational Environmental Satellites – R Series (GOES-R) planned for launch in October 2016.

Near real time SST retrievals from Himawari-8 at NOAA using ACSPO system

Japanese Himawari-8 (H8) satellite was launched on October 7, 2014 and placed into a geostationary orbit at ~ 140.7°E. The Advanced Himawari Imager (AHI) onboard H8 provides full-disk (FD) observations every 10 minutes, in 16 solar reflectance and thermal infrared (IR) bands, with spatial resolution at nadir of 0.5-1 km and 2 km, respectively. The NOAA Advanced Clear-Sky Processor for Ocean (ACSPO) SST system, previously used with several polar-orbiting sensors, was adapted to process the AHI data. The AHI SST product is routinely validated against quality controlled in situ SSTs available from the NOAA in situ SST Quality monitor (iQuam). The product performance is monitored in the NOAA SST Quality Monitor (SQUAM) system. Typical validation statistics show a bias within ±0.2 K and standard deviation of 0.4-0.6 K. The ACSPO H8 SST is also compared with the NOAA heritage SST produced at OSPO from the Multifunctional Transport Satellite (MTSAT-2; renamed Himawari-7, or H7 after launch) and with another H8 SST produced by JAXA (Japan Aerospace Exploration Agency). This paper describes the ACSPO AHI SST processing and results of validation and comparisons. Work is underway to generate a reduced volume ACSPO AHI SST product L2C (collated in time; e.g., 1-hr instead of current 10-min) and/or L3C (additionally gridded in space). ACSPO AHI processing chain will be applied to the data of the Advanced Baseline Imager (ABI), which will be flown onboard the next generation US geostationary satellite, GOES-R, scheduled for launch in October 2016.