Extended Abstract (968K)JP5.3
Monitoring of IR Clear-sky Radiances over Oceans for SST (MICROS): Near-Real Time, Web-Based Tool for Monitoring CRTM—AVHRR Biases for Improved Cloud Mask and SST Retrievals
XingMing Liang, NOAA/NESDIS, Camp Springs, MD; and A. Ignatov, Y. Kihai, and F. Xu
MICROS, a web-based tool whose name stands for Monitoring of IR Clear-sky Radiances over Oceans for SST (www.star.nesdis.noaa.gov/sod/sst/micros/), was established to monitor the Model minus Observation (M-O) biases in clear-sky brightness temperatures over oceans. The Community Radiative Transfer Model (CRTM) uses Reynolds SST and Global Forecast System upper air data as input. Corresponding observations come from the AVHRR Clear-Sky Processor for Oceans (ACSPO) system developed at NESDIS.
The tool publishes, in near-real time, the daily calculations of M-O biases in AVHRR bands 3B (3.7 μm), 4 (11 μm), and 5 (12 μm) onboard NOAA-16, 17, 18 and MetOp-A. Current regression SSTs derived from brightness temperatures, are also trended. In addition, histograms of the M-O biases are reported along with the respective daily statistics of a number of clear-sky ocean observations (N), Mean, and RMSD. Mean biases are further plotted as a function of retrieval conditions (column water vapor, wind speed, SST, sea-air temperature difference, satellite view zenith angle, and latitude). Mean bias, RMSD, and N are also trended in time, with a choice of a 1, 3, or 7-day moving average. Results for four platforms are overlaid, to facilitate cross-platform consistency analyses.
In addition to its current offerings, MICROS infrastructure also allows for easy incorporation of new data. Physical SSTs, being developed in the ACSPO framework, and NPOESS/VIIRS or GOES-R/ABI data, will be added when they become available, and work is underway to include MSG/SEVIRI data as well.
As of the time of this abstract, approximately two months of data are available through the MICROS page. At night, the M-O biases are slightly positive, leaving a few tenths degree of margin for future inclusion of aerosols, using skin SST (instead of the current Reynolds bulk SST) for the ongoing improvements in the ACSPO cloud mask. The biases are generally stable over time and only weakly depend on the retrieval conditions. The daytime M-O biases in ch3b are much lower and noisier than their nighttime counterparts because the extra-terrestrial solar and scattered radiation is not included in the current CRTM. Similar day-night differences, although of much smaller magnitudes, are also observed in the long-wave ch4 and 5. Data from different platforms are largely consistent. The only exception is NOAA16, whose M-O biases are out of family and show some unexplained trends.
The M-O biases in ACSPO are greatly affected by the delayed availability of weekly Reynolds SST data in ACSPO real-time processing (data sometimes lag behind by one to two weeks). Our immediate plan is thus to replace the 1° weekly Reynolds SST currently used in ACSPO with a 0.25° daily product. Other near-term plans also include working with the CRTM team to reconcile CRTM emissivity model with that accepted in the SST community, include extra-terrestrial and scattered radiation, and improve CRTM efficiency in ACSPO to allow parallel CRTM calculations (including calculation of the Jacobian, which is needed for physical SST retrievals).
Once sufficiently long time series of the M-O bias are accumulated, the remaining cross-platform inconsistencies in the M-O bias (e.g., out-of-family behavior of NOAA-16) should be better understood and corrected as needed. Next, we plan to start exploring physical SST retrievals in ACSPO. In a longer perspective, we plan to test GOCART aerosols in conjunction with CRTM for improved SST. The effect of all new improvements and developments will be evaluated using the MICROS methodology described in this paper.
Joint Poster Session 5, Climate
Wednesday, 14 January 2009, 2:30 PM-4:00 PM, Hall 5
Previous paper Next paper