Science Focus: Sea Surface Temperature Measurements of the MODIS and AIRS Instruments Onboard of Aqua Satellite

Dongliang Yuan, NASA Goddard Earth Sciences Data and Information Services Center,
Code 610.2.

On May 24, 2002, the Aqua satellite was launched into a Sun synchronous orbit as part of the NASA-centered international Earth Observing System (EOS) to provide observations of the Earth´s oceans, atmosphere, land, ice, snow covers, and vegetation. Onboard Aqua carries several of the most important instruments of the EOS system, the Moderate Resolution Imaging Spectroradiometer (MODIS), the Atmospheric InfraRed Sounder (AIRS), and the Advanced Microwave Sounder Unit (AMSU). Both of MODIS and AIRS measure radiances in the infrared bands so that the surface temperature of the ocean can be derived. The infrared measurements of AIRS and the microwave measurements of AMSU have been combined to generate a sea surface skin temperature product. This article explains the sea surface temperature (SST) products of MODIS and AIRS and gives an introduction to the synchronized use of the two products in scientific researches.

Over the surface of the ocean, there frequently exists a very thin layer called the surface skin layer in remote sensing sciences (Schluessel et al., 1990) (Figure 2). The existence of the surface skin layer can be demonstrated both in theory (Hinzpeter, 1967, 1968) and in observations (Ewing and McAlister, 1960; Saunders, 1967; Clauss et al., 1970; Schluessel et al., 1990) by the need to regulate the long wave radiation and the sensible and latent turbulent heat fluxes across the sea surface. Above and below the thin skin layer, turbulent eddy fluxes enhance heat flux in the ocean and/or atmosphere across the interface. However, the eddy cannot transport heat across the ocean surface by itself. The heat balance in the skin layer must be accomplished by molecular processes, hence the thin skin layer. The actual thickness of the skin layer depends on the local energy flux of the molecular transports, which is usually less than 1 mm thick and can persist at wind speed up to 10 m/s. For stronger winds, the skin layer is destroyed by breaking waves. Observations indicate that the skin layer can re-establish itself within 10 to 12 seconds after the dissipation of the breaking waves (Ewing and McAlister, 1960; Clauss et al., 1970).

The infrared wave can only penetrate water no deeper than ~500 µm. The MODIS instrument measures infrared radiances at around two wavelengths: 3.9 µm and 11 µm. The penetration depths for these two bands are ~100 µm and ~10 µm. The AIRS uses wavelengths from 3.75 µm to 13 µm for surface properties retrievals, which have penetration depths ranging from ~100 µm to ~10 µm. Thus, the radiances measured by the MODIS and AIRS satellite instruments are emitted from within the surface skin layer of the ocean. Figure 1 shows the infrared optical constants of clear water.

Figure 1. Absorption coefficient and penetration depths of infrared waves. Adopted from Wieliczka et al., 1989).

The SST can also be measured by the microwave radiometer. The penetration depth of the microwaves can be an order of magnitude larger than that of the infrared waves. For low-frequency (6-10 GHz) microwaves, the penetration depth can exceed 1 mm. The penetration depth of the microwaves is sensitive to the surface salinity and roughness conditions. For the frequencies used in the AMSU instrument, which span from 23 GHz to 90 GHz, the penetration depth is generally smaller than 1 mm and the surface radiation is heavily influenced by the skin temperature of the surface ocean.

The vertical structure of the skin layer SST can be generally described as in Figure 2. The interface SST, SST_int, is the temperature at the infinitely thin layer at the exact air-sea interface. This temperature cannot be measured using current technology. The skin SST, SST_skin, is the temperature measured by an infrared radiometer at a depth of order of 500 µm depending on the wavelength of the measurement. This temperature is depth (wavelength) dependent, but the differences measured by the infrared radiometers are very small (less than 0.01 K due to the small penetration depth differences). Therefore, the wavelength dependence of SSTskin is usually ignored. The subskin SST, SST_subskin, is representative of the SST at the bottom of the skin temperature layer and is usually the value measured by a low-frequency (6-10 GHz) microwave radiometer. The SST at depth, SST_depth, (traditionally referred to as the bulk SST) represents the temperature of the upper mixed layer produced by the turbulence associated with wind stirring and convective overturning, etc.

Figure 2. Schematic plot of open ocean surface thermal structures: a) nighttime; b) daytime. SST_int is the temperature at the air-sea interface, SST_skin at about 500 µm, SST_sub-skin at about 1 mm and SST_depth the bulk SST. Figure adapted from Donlon et al. (2004).

Within the surface mixed layer usually of a thickness on the order of 1 m to 100 m, the bulk SST has very small vertical gradient. In fact, the depth of the surface mixed layer is frequently defined as the depth where the temperature drops from the surface bulk SST by a small amount (say 0.5 K or 1 K). Occasionally, vertical gradient of SST is present in the upper mixed layer due to interleaving and overturning processes of high-frequency heating/cooling. But the gradient is quickly eliminated by the turbulence over periods of a few hours to a day. The mean skin temperature is generally several tenths of a degree colder than the mean bulk temperature (Schluessel et al., 1990). The instantaneous bulk-skin temperature differences can be as large as 1.0 K to -1.0 K (Robinson, 1985), depending on the wind and surface flux conditions. For instance, when the long wave radiation from the upper few micrometers of the ocean is upward, the skin temperature is usually cooler than the bulk SST. Latent and sensible heat fluxes can cool the sea surface further if the air is dryer or colder.

Because of the small penetration depths of the infrared and microwave radiation, the satellite instruments from space can only measure the upwelling long wave or microwave radiations from the surface skin layer. However, most oceanographers are interested in the bulk SST in the surface mixed layer because traditionally this is the temperature measured by ships, by drifter buoys, and by moored thermometers, etc., and because the variations of the bulk SST involve large heat exchange, which can impact the earth´s climate. The errors of satellite SST measurements due to the bulk-skin temperature difference can cause significant inaccuracies in global climate studies.

Very few in situ measurements of the surface skin temperature are made on a regular basis, so the MODIS/Aqua SST data have been calibrated primarily by the bulk SST of in situ and ship-board measurements (Smith et al., 1996; Minnett, 1999; Barton, et al., 2003). The calibration is necessary because the atmospheric corrections, which the infrared measurement is sensitive to, involve large uncertainties. Thus the MODIS SST can be regarded as the best representation of the bulk SST based on the space-borne instruments, particularly during high-wind conditions (wind speed at 10 m above the sea level > 6 m/s) (Donlon et al., 2002). Over the tropical oceans, the MODIS algorithms have been calibrated by the Tropical Atmosphere-Ocean (TAO) mooring measurements at a nominal depth of 1 m. However, the use of ship-borne measurements may introduce errors of depth variations and may impact the accuracy of the SST retrievals (Donlon et al, 2002).

The microwave can penetrate clouds and thus the AMSU instrument is advantageous in measuring SST under cloud cover. However, the microwave SST measurements are strongly frequency and surface dependent and can be highly erroneous in the areas of strong precipitation, which make the microwave-only measurements less attractive in many areas of the ocean. The design of the AIRS/AMSU instrument is to achieve effective de-clouding using the high spectral resolution of the AIRS instrument combined with the AMSU field of view (FOV) (Aumann, et al., 2003). The de-clouding makes use of the multiple FOVs of the AIRS within an AMSU footprint and retrieves the surface skin temperature and the vertical temperature profiles of the atmosphere simultaneously in an iterative procedure (Susskind et al, 2003). Because of the intrinsic structure of the vertical temperature profile retrieved, the AIRS/AMSU SST is believed to be close to the skin SST in theory. Thus, the difference between the MODIS SST and the AIRS/AMSU SST contains the bulk-skin SST difference in addition to algorithm errors and instrument inaccuracies and others. Indeed, the AIRS/AMSU SST has been found to be colder than the MODIS SST on global scales (Aumann and Strow, 2003).

Because of the use of the surrounding FOVs, the de-clouding of AIRS/AMSU SST has been achieved at the sacrifice of the spatial resolution. Currently, the MODIS swath resolution at nadir is 1 km in comparison with the AIRS/AMSU resolution at swath nadir of 40.5 km. Figures 3 and 4 show examples of the MODIS (product MYD28L2) and AIRS SST (product AIRX2RET) nighttime swaths over the Japan area. The black areas in Figure 3 over the ocean indicate cloud coverage in the MODIS SST data. The missing data in Figure 3 indicate bad retrievals of the AIRS/AMSU SST data. It is noticed that the MODIS data has separate SST retrievals at 11 µm and 4 µm while the AIRS/AMSU use the radiances from 13 µm to 3.75 µm for surface property retrievals. We have only compared the MODIS 11 µm SST with the AIRS/AMSU SST in this article.

Figure 3. MODIS SST image on April 11, 2004.

In Figures 3 and 4, the warm SST fronts south of Japan indicate the Kuroshio fronts. North of Japan, in the Sea of Japan, the SST fronts of the Tsushima Current are visible in both images. Direct comparison of the two images shows that the MODIS SST image contains much more cloud coverage but much higher spatial resolution than the AIRS/AMSU image. Along the confluence zone of the Kuroshio and the Oyashio at the 36° N east of Japan, the MODIS image indicates clear sky with detailed structures of the SST variations revealed while the AIRS/AMSU image indicates bad retrievals of the SST pixels along the front. The AIRS/AMSU data are only retrieved within the latitudinal range of 40° S to 40° N at present while the MODIS SST covers areas at much higher latitudes. Because of the coarse spatial resolution, the AIRS/AMSU SST data in coastal waters are not retrieved well. Despite the differences, the basic features of the Kuroshio and the Tsushima Current are observed well by both instruments. For example, the meander of the Kuroshio just east of Japan is captured by both images.

Figure 4. AIRS SST images on April 11, 2004.

MODIS algorithms for SST retrieval sometimes fail at the front where large SST gradient is indicated by the observations. The black points along the eastern front of the Oyashio are an example of it. However, these SST pixels are kept in the MODIS SST swath data. If a user wants to use them, they can simply relax the quality filtering criterion.

Because of the bulk-skin SST difference and of the different retrieval algorithms, there are discrepancies between the MODIS and the AIRS/AMSU measurements of the SST at the same location. Note that both of the MODIS and AIRS/AMSU instruments are onboard of Aqua satellite so that they scan the same point of the earth's surface at exactly the same time. A few examples of the comparisons are made in Table 1. The comparisons are made at AIRS/AMSU SST pixels of good quality by averaging the MODIS SST pixels of good quality within the AIRS/AMSU pixel.

Table 1. Comparison of AIRS and MODIS SST

The values in Table 1 indicate that the AIRS/AMSU SST is generally colder than the MODIS SST, but the difference is within ±1.0 K, consistent with the bulk-skin SST difference and with the uncertainty of the satellite SST retrievals.

Although the AIRS/AMSU SST is generally colder than the MODIS SST on global scales (Aumann and Strow, 2003), the difference between the two SSTs can vary significantly over space. Figure 5 shows a comparison of the two SST measurements along a section of 30° N south of Japan. According to the MODIS image, the area is clear of clouds. The comparison shows that the MODIS SST is colder than the AIRS/AMSU SST along this section.

Figure 5. AIRS/AMSU and MODIS SST along 30° N south of Japan. The unit of SST is °C.

In summary, the MODIS SST and the AIRS/AMSU SST are equivalent measurements to the accuracies of the bulk-skin SST difference and of the satellite retrieval. The MODIS SST data have batter spatial resolution and high-latitude retrievals but more cloud coverage. In comparison, the AIRS/AMSU SST data have less cloud coverage but much coarser spatial resolution and fewer coastal SST retrievals. AIRS/AMSU SST data indicate bad retrieval near the Kuroshio and Oyashio confluence zone under clear sky. Both SST datasets have captured the major oceanography phenomena, like the Kuroshio meander and the Tsushima Current front, etc.

References:

1. Aumann, H.H., et al., AIRS/AMSU/HSB on the Aqua mission: Design, science objectives, data products, and processing systems, IEEE transactions on geoscience and remote sensing, 41, No.2, 253-264, 2003.
2. Aumann, H.H., and L. Strow, Sea surface temperature measurements with the Atmospheric Infrared Sounder (AIRS), Proceedings of IGARSS 2003 meeting, Toulouse, France, 2003.
3. Barton, I. J., P. J. Minnett, C. J. Donlon, S. J. Hook, A. T. Jessup, K. A. Maillet and T. J. Nightingale, 2003: The Miami2001 infrared radiometer calibration and inter-comparison: Ship comparisons. J. Atm. Ocean. Tech., (submitted).
4. Clauss, E., H. Hinzpeter, J. Mueller-Glewe, Messungen der Temperaturstruktur im wasser an der Grenzflache Ozean-Atmosphare, Meteor Forschungsergeb., Reihe B, 5, 90-94, 1970.
5. Donlon, C.J., P.J. Minnett, C.L. Gentemann, T.J., Nightingale, I.J. Barton, B. Ward, and M.J. Murry, 2001: Toward Improved Validation of Satellite Sea Surface Skin Temperature Measurements for Climate Research, J. Clim., Vol. 15, No. 4, pp. 353-369.
6. Ewing, G., and E.D. McAlister, On the thermal boundary layers of the ocean, Science, 131, 1374-1376, 1960.
7. Hinzpeter, H., Der Tagesgang Der Wasseroberflachentemperatur in der Nahe des Aquators, Meteor Forschungsergeb., Reihe B, 1, 41-44, 1967.
8. Hinzpeter, H., Tagesperiodische Anderungen des oberflachennahen Temperaturfeldes uber dem Meer als Folge von Strahlungsquellen und senken, Kiel Meeresforsch., 24, 1-13, 1968.
9. Minnett P.J., and R. H. Evans, Comparison of Satellite Retrieved Sea Surface Temperature with the Marine Atmospheric Emitted Radiance Interferometer. MUBEX Workshop, Kyoto, Japan. November 1999.
10. Robinson, I.S., Satellite Oceanography, 455 pp., Ellis Horwood, Chichester, England, 1985.
11. Saunders, P., The temperature at the ocean-air interface, J. Atmos. Sci., 24, 269-273, 1967.
12. Schluessel, P., W.J. Emery, H. Grassl, and T. Mammen, On the bulk-skin temperature difference and its impact on satellite remote sensing of sea surface temperature, J. Geophys. Res., 95, C8, 13 341-13 356, 1990.
13. Smith, W. L., R.O. Knuteson, H.E. Revercomb, W. Feltz, H.B. Howell, W.P. Menzel, N. Nalli, O.B. Brown, J. Brown, P. Minnett, and W. McKeown, Observations of the infrared radiative properties of the ocean--Implications for the measurement of sea surface temperature via satellite remote sensing. Bulletin of the American Meteorological Society, 77(1), 41-51, 1996.
14. Susskind, J., C.D. Barnet, and J.M. Blaisdell, Retrieval of atmospheric and surface parameters from AIRS/AMSU/HSB data in the presence of clouds. IEEE Trans. Geosci. Remote Sensing, 14, 390-409, 2003.
15. Wieliczka, D. M., S. Weng, and M. R. Querry, Wedge shaped cell for highly absorbent liquids: infrared optical constants of water, Appl. Opt., 28, 1714-1719, 1989.