• Antoine, D., , and Nobileau D. , 2003: Recent increase of Saharan dust transport over the Mediterranean seam as revealed from ocean color satellites SeaWiFS observations. J. Geophys. Res., 111, D12214, doi:10.1029/2005JD006795.

    • Search Google Scholar
    • Export Citation
  • Arbelo, M., , Podesta G. , , Hernandez-Leal P. , , and Diaz J. , 2003: Use of TOMS data to correct the Saharan dust effects on SST retrievals from satellite. Adv. Space Res., 32, 21752180.

    • Search Google Scholar
    • Export Citation
  • Bailey, P., 1995: SADIST-2 v100 products. ER-TN-RAL-AT-2164, 124 pp.

  • Brown, S., , Harris A. , , and Mason I. , 1997: New aerosol robust sea surface temperature algorithm for the Along-Track Scanning Radiometer. J. Geophys. Res., 102, 27 97327 989.

    • Search Google Scholar
    • Export Citation
  • Castro, S., , Emery W. , , and Wick G. , 2004: Skin and bulk sea surface temperature estimates from passive microwave and thermal infrared satellite imagery and their relationships to atmospheric forcing. Proc. Int. Geoscience and Remote Sensing Symp. (IGARSS04), Vol. 2, Anchorage, AK, IEEE, 831–834.

    • Search Google Scholar
    • Export Citation
  • Derrien, M., , and Gleau H. L. , 1999: Cloud classification extracted from AVHRR and GOES imagery. Proc. Meteorological Satellite Data User’s Conf., Copenhagen, Denmark, EUMETSAT, 545–553.

    • Search Google Scholar
    • Export Citation
  • Donlon, C., , and Robinson I. , 1998: Radiometric validation of ERS-1 Along-Track Scanning Radiometer average sea surface temperature in the Atlantic Ocean. J. Atmos. Oceanic Technol., 15, 647660.

    • Search Google Scholar
    • Export Citation
  • Donlon, C., and Coauthors, 2007: The Global Ocean Data Assimilation Experiment high-resolution sea surface temperature pilot project. Bull. Amer. Meteor. Soc., 88, 11971213.

    • Search Google Scholar
    • Export Citation
  • GHRSST-PP International Project Office, 2006: The GHRSST-PP product user guide. GDSv1 revision 1.1, GODAE High Resolution Sea Surface Temperature-Pilot Program International Project Office, Hadley Centre, Met Office, Exeter, United Kingdom, 84 pp. [Available online at http://www.medspiration.org/documents/GHRSST-PP-Product-User-Guide-v1.1.pdf.]

    • Search Google Scholar
    • Export Citation
  • Ginoux, P., , Chin M. , , Tegen I. , , Prospero J. , , Holben B. , , Dubovik O. , , and Lin S. , 2001: Sources and distributions of dust aerosols simulated with the GOCART model. J. Geophys. Res., 106, 20 25520 274.

    • Search Google Scholar
    • Export Citation
  • Grey, W. M. F., , North P. R. J. , , Los S. O. , , and Mitchell R. M. , 2006: Aerosol optical depth and land surface reflectance from multiangle AATSR measurements: Global validation and intersensor comparisons. IEEE Trans. Geosci. Remote Sens., 44, 21842197.

    • Search Google Scholar
    • Export Citation
  • Guerzoni, S., , and Chester R. , 1996: The Impact of Desert Dust Across the Mediterranean. Kluwer Academic Publishers, 387 pp.

  • Husar, R., , Prospero J. , , and Stowe L. , 1997: Characterization of tropospheric aerosols over the oceans with the NOAA Advanced Very High Resolution Radiometer optical thickness operational product. J. Geophys. Res., 102, 16 88916 909.

    • Search Google Scholar
    • Export Citation
  • Ignatov, A., , Sapper J. , , Cox S. , , Laszlo I. , , Nalli N. , , and Kidwell K. , 2004: Operational aerosol observation (AEROBS) from AVHRR/3 on board NOAA-KLM satellites. J. Atmos. Oceanic Technol., 21, 326.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437471.

  • Kishcha, P., , Alpert P. , , Brakan J. , , Kirchner I. , , and Machenhauer B. , 2003: Atmospheric response to Sahran dust deduced from ECMWF ReAnalysis (ERA) temperature increments. Tellus, 55, 901913.

    • Search Google Scholar
    • Export Citation
  • Legrand, M., , N’doum C. , , and Jankowiak I. , 1994: Satellite-derived climatology of the Saharan aerosol. Passive Infrared Remote Sensing of Clouds and Atmosphere II, D. K. Lynch, Ed., International Society for Optical Engineering (SPIE Proceedings, Vol. 2309), 127–135.

    • Search Google Scholar
    • Export Citation
  • Lewellyn-Jones, D., , Edwards M. , , Mutlow C. , , Birks A. , , Barton I. , , and Tait H. , 2001: AATSR: Global-change and surface temperature measurements from Envisat. ESA Bulletin, No. 105, European Space Agency, Paris, France, 11–21.

    • Search Google Scholar
    • Export Citation
  • May, D., , Stowe L. , , Hawkins J. , , and McClain E. , 1992: A correction for Saharan dust effects on satellite sea surface temperature measurements. J. Geophys. Res., 97, 36113619.

    • Search Google Scholar
    • Export Citation
  • McClain, E., , Pichel W. , , and Walton C. , 1985: Comparative performance of AVHRR based multichannel sea surface temperatures. J. Geophys. Res., 90, 11 58711 601.

    • Search Google Scholar
    • Export Citation
  • Meloni, D., , di Sarra A. , , Iorio D. , , and Fiocco G. , 2005: Influence of the vertical profile of Saharan dust on the visibile direct radiative forcing. J. Quant. Spectrosc. Radiat. Transfer, 93, 397413.

    • Search Google Scholar
    • Export Citation
  • Merchant, C., , and Harris A. , 1999: Toward the elimination of bias in satellite retrievals of sea surface temperature: 2. Comparison with in situ measurements. J. Geophys. Res., 104, 23 57923 590.

    • Search Google Scholar
    • Export Citation
  • Mona, L., , Amodeo A. , , Pandolfi M. , , and Pappalardo G. , 2006: Saharan dust intrusions in the Mediterranean area: Three years of Raman lidar measurements. J. Geophys. Res., 111, D16203, doi:10.1029/2005JD006569.

    • Search Google Scholar
    • Export Citation
  • Moulin, C., and Coauthors, 1998: Satellite climatology of African dust transport in the Mediterranean atmosphere. J. Geophys. Res., 103, 13 13713 144.

    • Search Google Scholar
    • Export Citation
  • Nalli, N. R., , and Stowe L. , 2002: Aerosol correction for remotely sensed sea surface temperatures from the National Oceanic and Atmospheric Administration Advanced Very High Resolution Radiometer. J. Geophys. Res., 107, 3172, doi:10.1029/2001JC001162.

    • Search Google Scholar
    • Export Citation
  • Nalli, N. R., , and Reynolds R. , 2006: Sea surface temperature daytime climate analyses derived from aerosol bias-corrected satellite data. J. Climate, 19, 410428.

    • Search Google Scholar
    • Export Citation
  • North, P., , Briggs S. , , Plummer S. , , and Settle J. , 1999: Retrieval of land surface bidirectional reflectance and aerosol opacity from ATSR-2 multi-angle imagery. IEEE Trans. Geosci. Remote Sens., 37, 526537.

    • Search Google Scholar
    • Export Citation
  • Noyes, E., , Minnett P. , , Remedios J. , , Corlett G. , , Good S. , , and Llewellyn-Jones D. T. , 2006: The accuracy of the AATSR sea surface temperature in the Caribbean. Remote Sens. Environ., 101, 3851.

    • Search Google Scholar
    • Export Citation
  • OSISAFPT, 2005: North Atlantic regional sea surface temperature product manual. Version 1.7, EUMETSAT Ocean and Sea Ice Satellite Application Facility, 50 pp. [Available online at http://www.osi-saf.org/biblio/docs/ss1_pum_leo_sst_2_2.pdf.]

    • Search Google Scholar
    • Export Citation
  • Piolle, J., , and Prevost C. , 2006: Match-up database user manual. IFREMER, 47 pp.

  • Prospero, J., 1999: Long-term measurements of the transport of African mineral dust to the southeastern United States: Implications for regional air quality. J. Geophys. Res., 104, 15 91715 927.

    • Search Google Scholar
    • Export Citation
  • Robinson, I., 2004: Measuring the Oceans from Space. The Principles and Methods of Satellite Oceanography. Springer-Praxis, 669 pp.

  • Robinson, I., , and Challenor P. , 2004: Medspiration—System requirements document. Tech. Rep. MED-SOC-RS-001-1, SOC Laboratory for Satellite Oceanography, Southampton Oceanography Centre, 193 pp.

    • Search Google Scholar
    • Export Citation
  • Ruescas, A., 2006: Remote sensing applications in the thermal range to the study of the atmospheric processes in the Western Mediterranean basin (in Spanish). Ph.D. thesis, University Jaume I, Castellón de la Plana, Spain, I 300 pp.

  • Schluessel, P., , Emery W. , , Grassl H. , , and Mammen T. , 1990: On the skin–bulk temperature difference and its impact on satellite remote sensing of sea surface temperature. J. Geophys. Res., 95, 13 34113 356.

    • Search Google Scholar
    • Export Citation
  • Stowe, L., , Ignatov A. , , and Singh R. , 1997: Development, validation and potential enhancements to the second generation operational aerosol product at the National Environmental Satellite, Data and Information Service of the National Oceanic and Atmospheric Administration. J. Geophys. Res., 102, 16 92316 934.

    • Search Google Scholar
    • Export Citation
  • Strong, A., , Kearns E. , , and Gjovig K. , 2000: Sea surface temperature signals from satellites—An update. Geophys. Res. Lett., 27, 16671670.

    • Search Google Scholar
    • Export Citation
  • Thorncroft, C., , and Flocas H. A. , 1999: A case study of Saharan cyclogenesis. Mon. Wea. Rev., 125, 11471165.

  • Torres, O., , Bartia P. , , Herman J. , , Ahmad Z. , , and Gleason J. , 1998: Derivation of aerosol properties from satellite measurements of backscattered ultraviolet radiation: Theoretical basis. J. Geophys. Res., 103, 17 09917 110.

    • Search Google Scholar
    • Export Citation
  • Trigo, I., , Davies T. , , and Bigg G. , 1999: Objective climatology of cyclones in the Mediterranean region. J. Climate, 12, 16851696.

  • Vazquez-Cuervo, J., , and Sumagaysay R. , 2001: A comparison between sea surface temperatures as derived from the European remote sensing Along-Track Scanning Radiometer and the NOAA/NAA AVHRR Oceans Pathfinder dataset. Bull. Amer. Meteor. Soc., 82, 925944.

    • Search Google Scholar
    • Export Citation
  • Walton, C., 1988: Nonlinear multichannel algorithm for estimating SST with AVHRR satellite data. J. Appl. Meteor., 27, 115124.

  • Walton, C., , Pichel W. , , Sapper F. , , and May M. , 1998: The development and operational application of nonlinear algorithms for the measurement of sea surface temperatures with NOAA polar-orbiting environmental satellites. J. Geophys. Res., 103, 27 99928 012.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Mediterranean Sea (left) AVHRR and (right) AATSR total negative day and night data.

  • View in gallery

    Temperature residuals vs AOD, mean and standard deviation, 2006–08 AVHRR sensor: (top) day and night data, and (bottom) night data only (negative errors between −3° and 0°C).

  • View in gallery

    As in Fig. 2, but for AATSR sensor.

  • View in gallery

    Mean monthly values of (top) AOD and (bottom) SST residuals between 2006 and 2008. Dashed and solid lines refer to AATSR and AVHRR data, respectively, during the nighttime.

  • View in gallery

    Temperature profiles from radiosonde data from Cagliari for different days at 0000 UTC.

  • View in gallery

    Representation of the air temperature (°C) at 850 hPa and precipitable water (kg m−2) (left) 19 and (right) 20 Aug 2006.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 93 94 20
PDF Downloads 36 36 12

Examining the Effects of Dust Aerosols on Satellite Sea Surface Temperatures in the Mediterranean Sea Using the Medspiration Matchup Database

View More View Less
  • 1 Global Change Unit, Image Processing Laboratory, University of Valencia, Valencia, Spain
  • | 2 Observation of the Earth and Atmosphere Group, Department of Physics, University of La Laguna, Tenerife, Spain
  • | 3 Global Change Unit, Image Processing Laboratory, University of Valencia, Valencia, Spain
© Get Permissions
Full access

Abstract

Dust aerosol plumes from the Sahara cover the Mediterranean Sea regularly during the summer months (June–August) and occasionally during other seasons. Dust can absorb infrared longwave radiation, thus causing a drop in sea surface temperature (SST) retrievals from satellite. To quantify the magnitude of this absorption and to understand the sources of the biases that might be introduced when trying to validate SST algorithms with in situ bulk temperatures, the effects of the dust absorption are studied using the Medspiration Match-up Database. This database provides in situ and satellite SSTs derived from the Advanced Very High Resolution Radiometer (AVHRR) and the Advanced Along-Track Scanning Radiometer (AATSR), and the difference between skin and bulk measurements is calculated in order to obtain errors or residuals, which are classified by ranges and compared to an aerosol optical thickness product derived from the sensors. The behavior of the residuals is studied and there is clear correspondence between higher residual values and higher aerosol concentrations, though in some cases this relation is not so evident, especially during autumn months. Residuals for this period are then related to other atmospheric effects and layer boundary physical processes by using ancillary data (e.g., soundings).

Corresponding author address: Ana B. Ruescas, Global Change Unit, IPL, P.O. Box 22085, E-46071 Valencia, Spain. E-mail: bruescas@uv.es

Abstract

Dust aerosol plumes from the Sahara cover the Mediterranean Sea regularly during the summer months (June–August) and occasionally during other seasons. Dust can absorb infrared longwave radiation, thus causing a drop in sea surface temperature (SST) retrievals from satellite. To quantify the magnitude of this absorption and to understand the sources of the biases that might be introduced when trying to validate SST algorithms with in situ bulk temperatures, the effects of the dust absorption are studied using the Medspiration Match-up Database. This database provides in situ and satellite SSTs derived from the Advanced Very High Resolution Radiometer (AVHRR) and the Advanced Along-Track Scanning Radiometer (AATSR), and the difference between skin and bulk measurements is calculated in order to obtain errors or residuals, which are classified by ranges and compared to an aerosol optical thickness product derived from the sensors. The behavior of the residuals is studied and there is clear correspondence between higher residual values and higher aerosol concentrations, though in some cases this relation is not so evident, especially during autumn months. Residuals for this period are then related to other atmospheric effects and layer boundary physical processes by using ancillary data (e.g., soundings).

Corresponding author address: Ana B. Ruescas, Global Change Unit, IPL, P.O. Box 22085, E-46071 Valencia, Spain. E-mail: bruescas@uv.es

1. Introduction

High accuracy in the retrieval of sea surface temperatures (SSTs) from satellite imagery is compulsory when studying upper-layer ocean dynamics related to climate applications, weather prediction, and other environmental studies. Global average biases of these temperature measurements are considered to be of good quality when the errors do not surpass 0.3°C as a minimum requirement, and 0.1°C as a desirable goal (Walton et al. 1998). In some ocean regions, high loads of tropospheric aerosols from volcanic eruptions, burning biomass, or dust aerosols from desert areas can introduce large negative biases in the SST thermal infrared estimations (Vazquez-Cuervo and Sumagaysay 2001). For example, the influence of volcanic aerosols was evident in depressed Advanced Very High Resolution Radiometer (AVHRR) derived SSTs during most of 1982–83 and in 1991–92 due to the eruptions of the El Chichón and Mount Pinatubo volcanoes, respectively (Strong et al. 2000). Large quantities of mineral dust are lifted, and are vertically and horizontally transported by winds to areas far from their origins (Torres et al. 1998). Dust plumes from sources in North Africa are the most prominent and persistent, and cover the largest ocean areas (Husar et al. 1997; Prospero 1999). May et al. (1992) proposed a method of correcting for Saharan dust effects on satellite-derived SSTs. This methodology linked SST errors in the North Atlantic Ocean to an estimate of aerosol optical depth derived from the AVHRR. Similar methods have been proposed on a global scale by Nalli and Stowe (2002) and Nalli and Reynolds (2006), in which empirical data are produced by merging the Pathfinder Atmospheres (PATMOS) gridded AVHRR data with in situ buoy matchups. In the box-and-whiskers plots presented by Nalli and Reynolds (2006), a distribution of the dust residuals grouped by months is shown for the PATMOS Multichannel SST (MCSST) dataset using daytime data only, revealing clear seasonal trends that will be discussed later in this paper.

One of the most frequently used thermal infrared sensors from which SSTs can be extracted is the AVHRR operating on board National Oceanic and Atmospheric Administration (NOAA) satellites since 1981. This instrument incorporates five channels (six when considering that channel 3 modifies its spectrum from daytime to nighttime), three of which are in the infrared part of the spectrum. These channels are used to extract the temperature measurements by applying multiband techniques developed initially by McClain et al. (1985) and refined by Walton (1988), who finally proposed a nonlinear SST algorithm (NLSST) that is still in operation.

Another important remote sensor used to extract global SSTs is the Advanced Along-Track Scanning Radiometer (AATSR) on board the European Environmental Satellite (ENVISAT). This is the third in a series of instruments, after ATSR-1 and ATSR-2 [on board the European Remote Sensing Satellite-1 (ERS-1) and -2 (ERS-2), respectively]. SSTs derived from this sensor are highly accurate, less than ±0.3°C (Noyes et al. 2006), and are used for climate change and numerical weather prediction (NWP). AATSR has a dual view angle, nadir and forward (55°), and as a result the same point on Earth can be viewed through two different atmospheric pathlengths. Additionally, AATSR has 7 channels, 3 of which are in the thermal range of the spectrum (at 3.7, 11, and 12 μm), and are widely used to retrieve SSTs. AATSR products are aerosol robust due to the dual-angle geometry (Merchant and Harris 1999).

SST measurements retrieved through remote sensing correspond to a layer at a depth of approximately 10–20 μm, which means they represent measurements almost in direct contact with the atmosphere. These SSTs are not the same as temperatures measured in situ by buoys and ships using very different kinds of instruments (thermistors, thermosalinographs, etc.) at depths ranging from decimeters to meters. These kinds of measurements are technically called bulk SSTs (Schluessel et al. 1990).

A strong bias detected in SST satellite retrievals is due to dust intrusions in the atmosphere. Desert dust plumes are easy to detect in satellite images in the solar spectrum. They can absorb radiation in the violet, blue, and green portions of the spectrum (400–500 nm), which is why remote sensing of aerosol properties over the oceans is performed using the red and near-infrared bands, since the clear-column contribution to the top-of-the-atmosphere radiance is negligible at these wavelengths (Moulin et al. 1998). The desert dust plumes also affect thermal IR channel measurements of satellite sensors. During the daytime, aerosol presence causes a decrease in the shortwave radiative flux at the surface, resulting in a temperature drop and a longwave emission decrease at the surface. In addition, the decrease of emitted radiance from the surface is further reduced during its transfer through the dust layer (Castro et al. 2004; Legrand et al. 1994).

The unique contribution of this work is to identify and evaluate the bias in satellite SST retrievals caused mainly by large quantities of absorbing atmospheric aerosols in the Mediterranean Sea. This paper also provides some explanations for anomalies observed in these biases. To this end, the Global Ocean Data Assimilation Experiment (GODAE) High Resolution SST Pilot Project (GHRSST) Match-up Database of collocated satellite and in situ SSTs are used (Donlon et al. 2007), a service of Medspiration, the European node to the GHRSST-PP system, a project funded by the European Space Agency (ESA). The aerosol content of the troposphere was identified by means of the aerosol optical depth (AOD) included in the aforementioned Medspiration Match-up Database (MMDB; Piolle and Prevost 2006). The behavior patterns of SSTs from AVHRR and AATSR data and their corresponding in situ measurements are studied by comparing the results obtained from the satellite sensors and the buoys through calculations of the difference between the two products (hereafter called residuals) and relating them to the presence or absence of aerosols provided by the MMDB data in the lower atmosphere.

This paper is organized as follows. Section 2 describes the study area and section 3 presents the Medspiration Match-up Database and the aerosol products. In section 4, the results obtained and the comparisons between AVHRR and AATSR residual behavior dependence on aerosol optical depth are presented. Results also include an analysis of the seasonal behavior of the residuals and a discussion of the anomalies observed. Finally, section 5 presents our conclusions and final remarks.

2. The Mediterranean Sea study area

The geographic limits of the Mediterranean Sea fall within 25° and 50°N and 15°W and 45°E. This area is unique because of its structure and geometry, as an oceanographic basin of small dimensions (4000 km from east to west with 46 000 km of coastline) yet with a complex climatology characterized by intense atmosphere–sea–land interactions. The average water temperature of the Mediterranean Sea is relatively warmer than that of the Atlantic, though there is significant variation throughout the year, which we refer to here as large thermal amplitude. The maximum absolute temperature value occurs during summer (>30°C) due to strong and continuous solar radiation and a lack of precipitation. The minimum value occurs in spring (11°–12°C) when waters become cooler from the increased cloudiness and the lack of radiation, which results from fewer hours of sun exposure during the winter months, in addition to the contribution made by water discharged from European rivers and the Nile.

The Mediterranean Sea is bordered on its southern and eastern shores by arid desert regions such as the Sahara and Middle Eastern deserts. These deserts are a source of lithosphere aerosol, which is transported largely in the form of dust pulses. This often occurs in summer and mainly originates in the central part of the desert. The transport is related to the position of the subtropical anticyclone over the northeast Atlantic and the presence of polar air masses over both the Atlantic Ocean and the European continent. The prevailing conditions determining dust transport toward the Mediterranean are 1) a depression over Spain (50% of the cases), 2) a depression over North Africa (25%), and 3) an anticyclone over the Mediterranean basin (Guerzoni and Chester 1996). Differences are observed between the western and eastern parts of the basin. In the western part, the transport usually occurs at a deeper atmospheric layer for a longer period of time than in the eastern areas. A large number of Saharan dust events are recorded in the May–July period, which means significant seasonal activity of the dust transport.

Long-term observation of the desert dust layer in the Mediterranean region using Raman lidar instruments [from the European Aerosol Research Lidar Network to Establish an Aerosol Climatology (EARLINET) project] performed by the Istituto di Metodologie per l’Analisi Ambientale (IMAA) showed that it extends between 2.5 and 5.9 km above sea level (Mona et al. 2006). This wide-altitude range is due to the proximity to the source region and to the strong convective regimes that develop over the desert. After a few days, it arrives to the Mediterranean area, and the largest aerosol load descends to the 0–2.5-km layer (Meloni et al. 2005). This altitude is the top height of the local planetary boundary layer (between 2 and 2.6 km at night). During springtime, the altitude of the center of mass reaches its maximum (3.8 km), slowly decreasing in autumn to 3 km. The same seasonal behavior pattern is observed for the altitude of the base of the aerosol layer: the top reaches a maximum at the beginning of the summer (because of high convective conditions) and drops to two minima, one at the end of the spring and the other in autumn.

3. Data and method

a. The Medspiration Match-up Database

In 2000, the international GODAE steering team (IGST) initiated the GHRSST-PP project, which was involved in the development, provision, and application of global SST data products. GHRSST-PP provides a new generation of global high-resolution (<10 km) SST products to the operational scientific community (Donlon et al. 2007). The GHRSST-PP Match-up Database of collocated satellite and in situ SSTs is a service of Medspiration, the project funded by the European Space Agency (ESA). This match-up database is required by GHRSST-PP to perform the quality control of satellite SST datasets, in particular for deriving or verifying static sensor specific error statistics (SSESs) using in situ SSTs. Such observations provide a reliable, independent reference dataset that must be matched in space and time to satellite observations. The MMDB is a multisensor dataset that integrates in situ data stored in the Coriolis system (which is responsible for the collection and archiving of global in situ oceanographic data) with an archive of GHRSST-PP data products, both hosted by the Institut Français de Recherche pour L’Exploitation de la Mer (IFREMER) (GHRSST-PP International Project Office 2006).

In situ and satellite data are collocated on a daily basis within 25 km and 6 h of the satellite overpass as a worst-case scenario. The in situ SST data available in Coriolis and used for the matchups currently include all surface measurements (thermosalinographs on ships and drifting buoys), in addition to data from profiling sensors (Argo floats, XBTs–CTDs–XCTDs from ships and moored buoys). The satellite sources are restricted to products from the European Medspiration project and will be progressively extended to other GHRSST-PP datasets. Between 100 000 and 150 000 matchups are registered each month in the MMDB. All ancillary data attached to level 2 P (L2P) and level 4 (L4) products are available for each satellite matchup in the MMDB (Piolle and Prevost 2006).

In the present work, L2P AVHRR/3 data from the National Oceanic and Atmospheric Administration-17 (NOAA-17) and -18 (NOAA-18) satellites from January 2006 to December 2008 are analyzed (Fig. 1, left). SSTs are calculated for these sensors by the Ocean and Sea Ice Satellite Application Facility (OSISAF) organization. The SST products are available within 2 h after the last satellite data acquisition over a grid of 2-km resolution for the near Atlantic Region (NAR). The NAR products are then delivered in Network Common Data Form (NetCDF) through IFREMER. IFREMER indicates that calculated SSTs are equivalent to in situ SSTs at night (buoy data), but during the day a bias of several degrees kelvin may be found under favorable diurnal heating conditions. The operational cloud masking is based on a multispectral threshold algorithm (Derrien and Gleau 1999) with some refinements specific to the marine conditions (OSISAFPT 2005).

Fig. 1.
Fig. 1.

Mediterranean Sea (left) AVHRR and (right) AATSR total negative day and night data.

Citation: Journal of Atmospheric and Oceanic Technology 28, 5; 10.1175/2010JTECHA1450.1

AATSR SST data from the same period (2006–08) and associated in situ measurements are also analyzed (Fig. 1, right). The calibration on board the AATSR is very accurate and the along-track scanning technique provides two different viewing angles (0° and 55° at nadir and forward, respectively), making this a uniquely sensitive and stable instrument. SSTs are calculated using the 11- and 12-μm channels during the day, and the 11-, 12-, and 3.7-μm channels during the night. This last channel is selected to provide an additional channel at night due to its very high radiometric sensitivity. Two results are obtained for each pixel whenever possible: one using the combined nadir and forward views and the other using the nadir view alone. SSTs are calculated using a predefined set of retrieval coefficients derived from a forward model representing several SSTs and atmospheric states (Lewellyn-Jones et al. 2001). AATSR uses the same processor developed by the Rutherford Appleton University (RAL) called Synthesis of ATSR Data Into Surface Temperature (SADIST; Bailey 1995)) that was used for ATSR-1 and ATSR-2 instruments, but reengineered to be integrated within the wider Envisat payload data segment architecture. The algorithms have been reused to maintain consistency across the three missions.

b. Aerosol products

AOD data used here are included in the MMDB and also matched with the SST records. The AOD derived from the AVHRR/3 uses channel 1 (0.58–0.68 μm). To make the AOD calculations, the upward radiance is scaled in clear weather conditions for oceans and the atmosphere using a simplified form of the radiative transfer equation (Ignatov et al. 2004). All aerosol types are detected, including weakly absorbing aerosols such as industrial sulfates, throughout the entire thickness of the atmospheric column. The first algorithm created for AVHRR AOD (phase 1) was not able to detect nonspherical particles and dust particles with a size of more than a 0.1 μm. There were other assumptions like the scattering model (Mie single-scattering albedo) and an ocean albedo independent of the overpass daytime. The second-generation algorithm has many features in common with the phase 1 algorithm, though it has been improved by lowering the Lambertian surface reflectance. The bidirectional Fresnel reflectance has also been taken into account by introducing a diffuse glint correction. The conclusion after testing with the phase 2 algorithm is that a similar result as would have been observed had the validation been performed with sun photometers is reached (Stowe et al. 1997).

AATSR AOD has the advantage of the multilook approach, making it possible to retrieve aerosols over any kind of surface. The observations made in forward view are more influenced by the atmosphere than in the nadir view due to the longer path of the former. North et al. (1999) developed a simple physical model of light scattering to deal with variations in reflectance between simultaneous multiangular measurements of the ground due to the scattering of the surface, the difference in the atmospheric pathlength, and the scattering phase function. The theory holds that scattering by aerosols increases the diffuse contribution of light at the surface, thus reducing the anisotropy due to the decrease in the contrast between shadowed and sunlit surfaces. The model of surface scattering was applied to AATSR top-of-the-atmosphere (TOA) reflectance data in Grey et al. (2006). They compared the results with Aerosol Robotic Network (AERONET) ground-based sun-photometer data, the aerosol index (AI) derived from the Total Ozone Mapping Spectrometer (TOMS), and the Moderate Resolution Imaging Spectroradiometer (MODIS) and Multiangle Imaging Spectroradiometer (MISR) aerosol and surface products. The overall Pearson correlation coefficient for the sites tested was 0.70. The rms error was 0.16, and Grey et al. found no evidence of systematic error in the estimates of AOD because the mean AOD was 0.27, compared with the mean AERONET AOD of 0.26. In the intersensor comparison, little bias was found between AERONET, AATSR, and MISR.

c. Calculation of residuals

The residuals are the difference between the skin–satellite and the bulk–buoy sea temperatures. Data studied here have been previously filtered based on different requirements, and only a restricted set of residuals between −3° and 0°C is analyzed to avoid high error values associated with external sources like, for example, cloud contamination and poorly functioning sensors (Arbelo et al. 2003). Regarding the AOD data, these are divided into bins adapted to the dataset, with a maximum of 2.0 for the AVHRR dataset and 1.4 for the AATSR dataset due to the marked decrease in the number of cases above those thresholds. For the calculation by AOD bins, the residuals are averaged. The same process is followed for the extraction of monthly residual values. The total numbers of observations for each sensor are 72 217 for AVHRR and 11 001 for AATSR for the 3-yr period. When daytime and nighttime data are separated, there are 50 510 data points during the day and 21 707 data points at nighttime for AVHRR, with 9720 during the daytime and 1281 at nighttime for AATSR.

4. Results

If the difference between the approximate skin SST measurements by satellite and the bulk SST measurements by buoys is calculated on the days when aerosol concentration has been detected, residuals must be negative because bulk SSTs will be warmer than the skin SSTs affected by the layer of aerosols above it. Dust aerosol presence and biases in SST differences are closely related in many cases. Table 1 shows an increase in the SST residuals responding to the increase in the aerosol concentration (AOD). AVHRR has a mean maximum of −2.0°C with a range of 1.9–2.0 AOD. AATSR shows mean values around −0.5°C with a maximum of −0.93°C, with AOD values between 1.3 and 1.4. The result of SST dependence on aerosols organized by ranges for the two sensors and using all data available (within −3° to 0°C) are shown in the top panels of Figs. 2 and 3. In both graphs, an increase can be observed in the residuals, which is related to an increase in the aerosol content, as shown in Table 1. The AVHRR database registered higher values of both aerosol concentration and difference errors, as expected. There is then a clear increase in the residuals with an increase in the aerosol concentration. In the AATSR graph, all that can be observed is an increase in the residuals greater than −0.5°C, with aerosol concentration values greater than 1.2. This is the main difference between the two datasets, because residuals in AVHRR are greater than −0.5°C, even in the lower AOD ranges.

Table 1.

Negative residuals (−3° to 0°C) for ranges of aerosols against AVHRR and AATSR sensors.

Table 1.
Fig. 2.
Fig. 2.

Temperature residuals vs AOD, mean and standard deviation, 2006–08 AVHRR sensor: (top) day and night data, and (bottom) night data only (negative errors between −3° and 0°C).

Citation: Journal of Atmospheric and Oceanic Technology 28, 5; 10.1175/2010JTECHA1450.1

Fig. 3.
Fig. 3.

As in Fig. 2, but for AATSR sensor.

Citation: Journal of Atmospheric and Oceanic Technology 28, 5; 10.1175/2010JTECHA1450.1

In the shallower layers of the sea, skin and bulk SSTs are affected by two kinds of thermoclines: one seasonal, during the summer, found below 20 m, and another diurnal, caused by solar heating under clear-sky conditions and more frequent during summer, especially in the Mediterranean area. Because the shortwave solar radiation is progressively absorbed, the water temperature rises more near the surface, causing a strong gradient of temperature (several degrees kelvin) in the top meter or so. If there is wind mixing, the effect is weakened because the energy is transmitted downward. The diurnal thermocline disappears at night, and the surface layer, now colder, sinks and promotes gravitational mixing. Nighttime bulk measurements might be free of the influence of the diurnal thermocline and therefore more appropriate for studying the skin–bulk relationship at any time of the year. We then decided to analyze data from the nighttime only. In the MMDB, an AOD measurement nearest in space and time to the input pixel SST value is used in each case, allowing us to proceed with the comparisons at night. If no AOD measurement is available, an AOD value derived from an NWP system or aerosol forecast nearest in space and time to the SST measurement is used instead (Robinson and Challenor 2004).

The total data at night for AVHRR (Fig. 2, bottom) show a decrease in the residuals with aerosol range > 1.2. Furthermore, the thermal amplitude (difference in residuals between day and night) reaches 0.8°C in the last AOD range (1.9–2.0; see Table 1). This bias is indicative of the effects of the diurnal heating during the day and the kind of errors that might arise in the SST extraction, even masking any other source of errors. AATSR aerosol dependence is shown in Fig. 3 (bottom), and, due in part to the small number of cases, no clear trend can be observed, in particular, during the nighttime. This is strongly related to the processor used to extract the SST, which corrects the effects of tropospheric aerosols. It is important to note that AATSR instruments properly return SST measurements for the skin of the ocean, while the AVHRR data processing scheme uses a regression method with buoy measurements, which introduces a bias toward a bulk temperature in that dataset (Lewellyn-Jones et al. 2001).

a. Temporal behavior

The seasonal and interannual patterns of behavior of aerosol concentrations and residuals for the 3-yr period of study are analyzed using nighttime data. Table 2 shows the mean values of AOD and residuals by month. Unfortunately, there are no data for the summer period for the AATSR sensor within this threshold due to the strict requirements chosen when analyzing the data and the efficiency of the AATSR algorithm and filtering process (Noyes et al. 2006). For the AVHRR AOD, the values are higher from April to July, with the maximum value in June. The highest residual value occurs in April, followed by May and, surprisingly, by December. Residual values in AVHRR are in general higher than AATSR residuals during the whole year with the exception of April, when unfortunately there are only 18 records and they are probably not sufficiently representative of the behavior for this month. The lack of a clear seasonal pattern of behavior in the residuals contrasts with the results shown by Nalli and Reynolds (2006) for AVHRR and mentioned in the introduction, where there was a marked depression of the global summer values due to the effects of the dust aerosols.

Table 2.

Mean values of matchup residuals and aerosol products by months (°C).

Table 2.

Figure 4 shows mean values of AOD and residuals per month for AVHRR and AATSR. AVHRR data show three peaks in SST residual values: one in June 2006 (−0.99°C), one in April 2007 (−1.36°C), and the last in May 2008 (−1.25°C). These peaks are preceded by a constant increase in the residuals throughout the year, and they are clearly related to the aerosol concentration increase: 0.47 in June 2006, 0.28 in April 2007, and 0.77 in May 2008. Another two peaks in residuals are not related to AOD mean values: in July 2007 there is a value of −0.93°C with an AOD of 0.18, and in October 2008 there is a residual value of −1.03°C with an AOD value of only 0.11. The low residual values observed in June 2008 are remarkable, especially after the high values reached in the preceding month. We can conclude that the fit between the two trends is quite good on the AVHRR, with the exception of the last part of 2008. There are some gaps in the data for November and December 2007 and December 2008. Antoine and Nobileau (2003) did a 7-yr study on the presence and behavior of dust aerosols over the Mediterranean Sea using Sea-viewing Wide Field-of-view Sensor (SeaWiFS) data (1998–2004). They confirmed this marked seasonality, with minimum values in winter in the western basin (<0.125) and in the eastern part (<0.15). They observed an increase in the AOD during springtime in the southern region of the eastern Mediterranean (<35°N) and to a lesser extent in the western part. During the summer the maximum is in the southwest (0.25) whereas the southeastern area has a clearer atmosphere (<0.15). Antoine and Nobileau (2003) also studied the interannual variability, with the spring and autumn having the largest interannual changes. The variability from year to year is said to be characterized by dramatic changes, resulting in highly contrasted situations. A general increase in the dust occurrence during the 7-yr period is pointed out in their work as well.

Fig. 4.
Fig. 4.

Mean monthly values of (top) AOD and (bottom) SST residuals between 2006 and 2008. Dashed and solid lines refer to AATSR and AVHRR data, respectively, during the nighttime.

Citation: Journal of Atmospheric and Oceanic Technology 28, 5; 10.1175/2010JTECHA1450.1

Trends in AATSR are not so easily observable due to the lack of data during summer. There is no clear relationship between the two variables, but it can be observed that during autumn, when there is a slight increase in the AOD index, there is also some response to SST residuals. In any case, early validation results for ATSR-1 made by Donlon and Robinson (1998) suggested that SST retrievals were affected by different kinds of aerosols. Consequently, stratospheric aerosol-robust coefficients were developed by Brown et al. (1997) and Merchant and Harris (1999). Corrections were applied to ATSR-2 and AATSR. The following results from algorithm validations indicated a clear advantage of alongtrack scanning over traditional single-view SST retrievals of the AVHRR.

We can conclude that even when there is a direct relationship between the increase in aerosols and the increase in the residuals during part of the spring and early summer, this fact is not clearly observable in the Mediterranean Sea during part of the autumn and early winter. In this last case when the AOD is at its minimum, the residual values are still relatively high. There must therefore be other factors that modify the expected pattern of behavior of the relationship between the two variables.

b. Discussion

What is happening in the Mediterranean Sea during autumn? In Fig. 4, a progressive increase in the SST residuals from September to December 2006 can be observed while AOD values are kept low. The same happens in late 2008. There are no data for November and December 2007, but in September and October of that year this pattern of behavior appears again.

Four mechanisms can explain these “anomalies” detected during autumn and occasionally during some summer months. We try here to explain how these different physical mechanisms can bring about a reduction or an increase in the difference between the satellite and the in situ temperatures. On some occasions, the measurement is a result of a real decrease or increase in the differences between the skin and the bulk temperature; in other cases, it is related to modifications of the temperature when passing through the atmosphere and reaching the satellite.

The first mechanism is related to the thermal amplitude (differences in temperature) between the sea and the boundary layer of the atmosphere, which is especially high in altitude during autumn and winter. The Mediterranean Sea is much warmer than the atmosphere and its surrounding land during autumn. Its average temperature is certainly warm, with high thermal inertia, elevated specific heat capacity, high conductivity, and a high rate of evaporation by wind and advection. For instance, near the coast of Castellón (Spain) a mean bulk SST of 19°C was measured between October and December, whereas the air temperature 10 m above the surface of the sea is more than 1°C lower (17.9°C) (Ruescas 2006). This gradient causes a potential convective response, which is the reason for most of the atmospheric perturbations in this area during this time of the year. This may therefore be one reason for the high residuals between the skin and bulk temperatures. Since the skin SST is directly affected by advection and evaporation, the water surface skin gets colder faster than the deep waters underneath, which keep the temperature warmer for a longer period and seem to be less affected by conductivity downward.

The second mechanism is described in detail by Robinson (2004) and based on another physical mechanism that might be controlling the skin layer temperature: the turbulent motion inhibition on the sea surface by gravitational stability. In this respect, heat can be transported downward only by molecular conduction, which is weaker than the normal advection transport in the ocean interior. As a consequence, the skin layer cools until the gradient of the temperature between the two layers is sufficiently steep to drive enough heat flux to the surface. This fact produces an increase in the difference between the skin and the bulk temperatures.

The third mechanism is the wind speed. With wind speed greater than 6 m s−1, the layers of water are mixed together, thus decreasing the difference between skin and bulk temperatures (Robinson 2004), which provides another explanation for decreasing the differences even with the presence of aerosols in the atmosphere.

When wind speed does not appear be the cause behind decreasing residuals, and in particular with regard to summer anomalies, we need to refer to the work published by Kishcha et al. (2003). In this case, the SST measurements taken by the satellite are directly affected by the dust transport in the Mediterranean, whose vertical structure seems to be very different from that studied in the Atlantic Ocean, as do its effects. The atmosphere over North Africa is almost permanently loaded with desert mineral dust during the warm months of the year, that is from April to July (Ginoux et al. 2001). Kishcha et al. (2003) identified two areas of different characteristics: the positive correlation areas (PCAs), with high positive correlation between the European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA) temperature increments and an aerosol index, corresponding to areas where there is a thick dust layer, and the negative correlation areas (NCAs), where there is a relatively small amount of dust and the correlation with the ERA temperature increments is negative. In the areas with the presence of a thick dust layer, solar radiation from above is strongest during the daytime and heating increases with the thickness of the layer (absorption of shortwave radiation). The dust layer might even prove to be so thick as to become impenetrable. If this is the case, the heating of the surface below the dust by solar radiation is negligible and the resulting longwave heating from the surface (absorption of thermal radiation) has no major effect on the dust layer. This keeps the dust layer stable. During the night, the dust layer losses heat gradually by longwave radiation to space, and heat absorption from the sea surface becomes more important. That means that part of what the satellite measures is the longwave radiation coming from the dust layer, and another part from a reheated sea surface.

To check the process just described, the profiles shown in Fig. 5 are made with measurements taken by a sounding station in Cagliari, Sardinia, Italy, for four different days, two in the summer period (June 2006 and July 2008) and two during autumn 2006 (November and December). The sounding station (number 16 560) is located at 39.23°N and 9.05°E at an altitude of 5 m (1020 hPa), very close to the Mediterranean Sea. Sounding data are downloaded from the University of Wyoming sounding database server belonging to the World Meteorological Organization (WMO). Data for this station are processed and filtered because they must fulfill certain requirements, such as being a reasonable geographical distance from the MMDB data (a few kilometers) and having a minimal temporal baseline (both data readings should be taken at the same time). We did not calculate a monthly average of the sounding data because it would cancel out the fluctuations that can be observed in the daily thermal inversion processes. Figure 5 shows the different patterns of behavior on those 4 days of the air temperature from the surface (1020 hPa) up to near the first 2 km of the atmosphere (700 hPa), which corresponds to the altitude of the maximum correlation between the temporal variations of dust presence and changes in air temperatures (Kishcha et al. 2003). During July 2008, a thermal inversion appears between 985 and 955 hPa. In June 2006 this inversion occurs lower, very near the surface, followed by an acute decrease in the temperatures. During summer, temperature increases in the air caused by the dust intrusion could explain the occasionally low differences in the residuals. With relatively high concentrations of aerosol, AOD equal to 0.9 on 12 July 2008 and AOD equal to 0.4 on 11 June 2006, the outcomes observed were: in June 2006 there was an increase in the air temperature in the first 50 m (see also Table 3); and on 12 July 2008 there was a rapid decrease in the first 100 m of altitude (see Table 4)— dropping 0.8°C, to gain them back at 995 hPa (150 m high)— and even surpass the surface temperature (25.6°C) if these data are compared with water temperatures at 1.5-m depth; and at the surface, these latter temperatures were warmer than the air at 5-m elevation, but less warm that along the inversion layer in the lower atmosphere. This means that the difference between SSTs measured by the satellite and the SSTs in situ decreases under the influence of the stability and warming of the dust layer.

Table 3.

Cagliari-Elmas Airport (LIEE, Station #16560) observations at 0000 UTC 11 Jun 2006.

Table 3.
Table 4.

Cagliari (LIEE, Station #16560) observations at 0000 UTC 12 Jul 2008.

Table 4.
Fig. 5.
Fig. 5.

Temperature profiles from radiosonde data from Cagliari for different days at 0000 UTC.

Citation: Journal of Atmospheric and Oceanic Technology 28, 5; 10.1175/2010JTECHA1450.1

Finally, in order to see the synoptic situation of the area during the summer, Fig. 6 shows a typical event in August with conditions for dust intrusions. Figure 6 was extracted by processing 6-h daily data of surface air temperature at the 850-hPa level and precipitable water from the National Centers of Environmental Prediction–National Center of Atmospheric Research reanalysis project (NNR; Kalnay et al. 1996). These data are at 2.5° × 2.5° spatial resolution at global scale, although for this work we gridded them linearly to show the spatial air temperature and precipitable water in the period under study. In August, the synoptic configuration coincided with the case described in Moulin et al. (1998), with a nucleus of higher temperatures located within the western Mediterranean basin, related to the Gulf of Geneva and Sahara low pressure areas (Trigo et al. 1999). These low pressure nuclei amplify the radiative heating profiles due mainly to Saharan dust influences (Thorncroft and Flocas 1999). In the eastern part of the Mediterranean basin (Aegean Sea and Cyprus), there is another area of low pressure that favors the easy transport of desert dust. The synoptic sequence shows that the highest temperature values are probably associated with the dust transport from the western Sahara and the Arabian Peninsula. In addition, the highest values of precipitable water are well correlated with the low pressure areas in the Mediterranean Sea.

Fig. 6.
Fig. 6.

Representation of the air temperature (°C) at 850 hPa and precipitable water (kg m−2) (left) 19 and (right) 20 Aug 2006.

Citation: Journal of Atmospheric and Oceanic Technology 28, 5; 10.1175/2010JTECHA1450.1

In autumn, the profiles show a thermal inversion in the first 150 m (see Fig. 5), with a visible increase in temperatures on 2 days, 24 November 2006 and 5 December 2006. After this increase, there is a natural decrease in air temperature with altitude. The temperature increase near the surface (<100 m) might be a result of water vapor content brought about by high seasonal evaporation rates, since the water is warmer than the relatively cooler air layer over it, as pointed out before and observed in Tables 5 and 6. Precipitable water amounts perceived in reanalysis data (not shown here) show a maximum over the whole Mediterranean Sea during autumn months, which is related to the general cyclogenesis tracked for this period and the rainfall events associated with it.

Table 5.

Cagliari (LIEE, Station #16560) observations at 0000 UTC 24 Nov 2006.

Table 5.
Table 6.

Cagliari (LIEE, Station #16560) observations at 0000 UTC 5 Dec 2006.

Table 6.

5. Conclusions

Our results showed a general increase in the residuals as the AOD ranges increased as well. The AVHRR database registered higher values of both aerosol concentration and difference errors, as expected. There was then a clear increase in the residuals with an increase in the aerosol concentration. In the AATSR case, we can only observe an increase in the residuals greater than −0.5°C with aerosol concentration values greater than 1.2, with no clear relationship using nighttime data only.

When analyzing the seasonal behavior of the nighttime data, the AVHRR data showed three peaks of SST residual values, one in June 2006 (−0.99°C), one in April 2007 (−1.36°C), and the last in May 2008 (−1.25°C), related to high concentrations of AOD. Another two peaks in the residuals were not related to AOD mean values: in July 2007 there was a value of −0.93°C with an AOD of 0.18, and in October 2008 there was a residual value of −1.03°C with an AOD value of only 0.11. Trends in AATSR were not so easily observable due to the lack of data during the summer. However, there did not seem to be a relationship between the two variables, most likely due to the stratospheric aerosol-robust coefficients developed by Brown et al. (1997) and Merchant and Harris (1999) and applied to the algorithm that generates the SST data.

In an effort to explain the irregular pattern of behavior of the residuals in AVHRR, that is to say, those cases where no dependence on aerosols of the SST residuals could be observed, we related the results to the effects over temperatures of four different physical mechanisms. These “anomalies” lead to an increase in the difference of the two measurements (satellite–buoy) with low aerosol loads, or the opposite effect, a decrease in the difference with high aerosol concentrations. 1) The first mechanism (marked thermal amplitude of the sea due to its high thermal inertia) would explain the first case. The gradient found between the sea and the boundary layer of the atmosphere causes a potential convective response that brings about most of the atmospheric perturbations in this area during autumn. Furthermore, this might be the reason for the high residuals between the skin and bulk temperature, the former being directly affected by advection and evaporation. 2) The second mechanism explains the increased difference due to the skin cool effect for the turbulent motion inhibition of the sea surface. Re-emission of the radiance from the surface is lower, meaning colder temperatures arrive to the sensor, which explains the first case as well. 3) The third cause that might explain the skin–bulk temperature similarities during summer and autumn is the wind speed at the sea surface. The different layers of water are mixed together as a result of strong winds, thus decreasing the difference between the two temperatures. 4) The last explanation indicates that there are areas where a correlation exists between an aerosol intrusion and an increase in the temperature measured by the sensor due to the heat capacity storage of a thick and stable aerosol layer. That is, a reduction in the difference between satellite–buoy temperatures caused by a warm aerosol layer in the summer months.

A logical continuation of this research should be focused on studying the behavior of the residuals by regions in the Mediterranean Sea and, if possible, using daily data. However, the quantity of matchups must increase in the Eastern Mediterranean to achieve this goal. More sounding data located in places near the coast or in the sea are also necessary to have more comparable measurements between the sea surface and air temperatures.

Acknowledgments

We thank the Medspiration project, founded by the ESA, the GHRSST group, and the IFREMER for processing and delivering the Medspiration March-up Database to the scientifc community; and the NOAA/OAR/ESRL PSD for the NCEP reanalysis derived data. Thanks to the Ministry of Education and Science in Spain for Projects CGL2007-66888-C02-01 and CGL2010-22189-C02-01. This work was supported by the Generalitat Valenciana by means of grants for postdoctoral research in Centers of Excellence in the Valencian Community.

REFERENCES

  • Antoine, D., , and Nobileau D. , 2003: Recent increase of Saharan dust transport over the Mediterranean seam as revealed from ocean color satellites SeaWiFS observations. J. Geophys. Res., 111, D12214, doi:10.1029/2005JD006795.

    • Search Google Scholar
    • Export Citation
  • Arbelo, M., , Podesta G. , , Hernandez-Leal P. , , and Diaz J. , 2003: Use of TOMS data to correct the Saharan dust effects on SST retrievals from satellite. Adv. Space Res., 32, 21752180.

    • Search Google Scholar
    • Export Citation
  • Bailey, P., 1995: SADIST-2 v100 products. ER-TN-RAL-AT-2164, 124 pp.

  • Brown, S., , Harris A. , , and Mason I. , 1997: New aerosol robust sea surface temperature algorithm for the Along-Track Scanning Radiometer. J. Geophys. Res., 102, 27 97327 989.

    • Search Google Scholar
    • Export Citation
  • Castro, S., , Emery W. , , and Wick G. , 2004: Skin and bulk sea surface temperature estimates from passive microwave and thermal infrared satellite imagery and their relationships to atmospheric forcing. Proc. Int. Geoscience and Remote Sensing Symp. (IGARSS04), Vol. 2, Anchorage, AK, IEEE, 831–834.

    • Search Google Scholar
    • Export Citation
  • Derrien, M., , and Gleau H. L. , 1999: Cloud classification extracted from AVHRR and GOES imagery. Proc. Meteorological Satellite Data User’s Conf., Copenhagen, Denmark, EUMETSAT, 545–553.

    • Search Google Scholar
    • Export Citation
  • Donlon, C., , and Robinson I. , 1998: Radiometric validation of ERS-1 Along-Track Scanning Radiometer average sea surface temperature in the Atlantic Ocean. J. Atmos. Oceanic Technol., 15, 647660.

    • Search Google Scholar
    • Export Citation
  • Donlon, C., and Coauthors, 2007: The Global Ocean Data Assimilation Experiment high-resolution sea surface temperature pilot project. Bull. Amer. Meteor. Soc., 88, 11971213.

    • Search Google Scholar
    • Export Citation
  • GHRSST-PP International Project Office, 2006: The GHRSST-PP product user guide. GDSv1 revision 1.1, GODAE High Resolution Sea Surface Temperature-Pilot Program International Project Office, Hadley Centre, Met Office, Exeter, United Kingdom, 84 pp. [Available online at http://www.medspiration.org/documents/GHRSST-PP-Product-User-Guide-v1.1.pdf.]

    • Search Google Scholar
    • Export Citation
  • Ginoux, P., , Chin M. , , Tegen I. , , Prospero J. , , Holben B. , , Dubovik O. , , and Lin S. , 2001: Sources and distributions of dust aerosols simulated with the GOCART model. J. Geophys. Res., 106, 20 25520 274.

    • Search Google Scholar
    • Export Citation
  • Grey, W. M. F., , North P. R. J. , , Los S. O. , , and Mitchell R. M. , 2006: Aerosol optical depth and land surface reflectance from multiangle AATSR measurements: Global validation and intersensor comparisons. IEEE Trans. Geosci. Remote Sens., 44, 21842197.

    • Search Google Scholar
    • Export Citation
  • Guerzoni, S., , and Chester R. , 1996: The Impact of Desert Dust Across the Mediterranean. Kluwer Academic Publishers, 387 pp.

  • Husar, R., , Prospero J. , , and Stowe L. , 1997: Characterization of tropospheric aerosols over the oceans with the NOAA Advanced Very High Resolution Radiometer optical thickness operational product. J. Geophys. Res., 102, 16 88916 909.

    • Search Google Scholar
    • Export Citation
  • Ignatov, A., , Sapper J. , , Cox S. , , Laszlo I. , , Nalli N. , , and Kidwell K. , 2004: Operational aerosol observation (AEROBS) from AVHRR/3 on board NOAA-KLM satellites. J. Atmos. Oceanic Technol., 21, 326.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437471.

  • Kishcha, P., , Alpert P. , , Brakan J. , , Kirchner I. , , and Machenhauer B. , 2003: Atmospheric response to Sahran dust deduced from ECMWF ReAnalysis (ERA) temperature increments. Tellus, 55, 901913.

    • Search Google Scholar
    • Export Citation
  • Legrand, M., , N’doum C. , , and Jankowiak I. , 1994: Satellite-derived climatology of the Saharan aerosol. Passive Infrared Remote Sensing of Clouds and Atmosphere II, D. K. Lynch, Ed., International Society for Optical Engineering (SPIE Proceedings, Vol. 2309), 127–135.

    • Search Google Scholar
    • Export Citation
  • Lewellyn-Jones, D., , Edwards M. , , Mutlow C. , , Birks A. , , Barton I. , , and Tait H. , 2001: AATSR: Global-change and surface temperature measurements from Envisat. ESA Bulletin, No. 105, European Space Agency, Paris, France, 11–21.

    • Search Google Scholar
    • Export Citation
  • May, D., , Stowe L. , , Hawkins J. , , and McClain E. , 1992: A correction for Saharan dust effects on satellite sea surface temperature measurements. J. Geophys. Res., 97, 36113619.

    • Search Google Scholar
    • Export Citation
  • McClain, E., , Pichel W. , , and Walton C. , 1985: Comparative performance of AVHRR based multichannel sea surface temperatures. J. Geophys. Res., 90, 11 58711 601.

    • Search Google Scholar
    • Export Citation
  • Meloni, D., , di Sarra A. , , Iorio D. , , and Fiocco G. , 2005: Influence of the vertical profile of Saharan dust on the visibile direct radiative forcing. J. Quant. Spectrosc. Radiat. Transfer, 93, 397413.

    • Search Google Scholar
    • Export Citation
  • Merchant, C., , and Harris A. , 1999: Toward the elimination of bias in satellite retrievals of sea surface temperature: 2. Comparison with in situ measurements. J. Geophys. Res., 104, 23 57923 590.

    • Search Google Scholar
    • Export Citation
  • Mona, L., , Amodeo A. , , Pandolfi M. , , and Pappalardo G. , 2006: Saharan dust intrusions in the Mediterranean area: Three years of Raman lidar measurements. J. Geophys. Res., 111, D16203, doi:10.1029/2005JD006569.

    • Search Google Scholar
    • Export Citation
  • Moulin, C., and Coauthors, 1998: Satellite climatology of African dust transport in the Mediterranean atmosphere. J. Geophys. Res., 103, 13 13713 144.

    • Search Google Scholar
    • Export Citation
  • Nalli, N. R., , and Stowe L. , 2002: Aerosol correction for remotely sensed sea surface temperatures from the National Oceanic and Atmospheric Administration Advanced Very High Resolution Radiometer. J. Geophys. Res., 107, 3172, doi:10.1029/2001JC001162.

    • Search Google Scholar
    • Export Citation
  • Nalli, N. R., , and Reynolds R. , 2006: Sea surface temperature daytime climate analyses derived from aerosol bias-corrected satellite data. J. Climate, 19, 410428.

    • Search Google Scholar
    • Export Citation
  • North, P., , Briggs S. , , Plummer S. , , and Settle J. , 1999: Retrieval of land surface bidirectional reflectance and aerosol opacity from ATSR-2 multi-angle imagery. IEEE Trans. Geosci. Remote Sens., 37, 526537.

    • Search Google Scholar
    • Export Citation
  • Noyes, E., , Minnett P. , , Remedios J. , , Corlett G. , , Good S. , , and Llewellyn-Jones D. T. , 2006: The accuracy of the AATSR sea surface temperature in the Caribbean. Remote Sens. Environ., 101, 3851.

    • Search Google Scholar
    • Export Citation
  • OSISAFPT, 2005: North Atlantic regional sea surface temperature product manual. Version 1.7, EUMETSAT Ocean and Sea Ice Satellite Application Facility, 50 pp. [Available online at http://www.osi-saf.org/biblio/docs/ss1_pum_leo_sst_2_2.pdf.]

    • Search Google Scholar
    • Export Citation
  • Piolle, J., , and Prevost C. , 2006: Match-up database user manual. IFREMER, 47 pp.

  • Prospero, J., 1999: Long-term measurements of the transport of African mineral dust to the southeastern United States: Implications for regional air quality. J. Geophys. Res., 104, 15 91715 927.

    • Search Google Scholar
    • Export Citation
  • Robinson, I., 2004: Measuring the Oceans from Space. The Principles and Methods of Satellite Oceanography. Springer-Praxis, 669 pp.

  • Robinson, I., , and Challenor P. , 2004: Medspiration—System requirements document. Tech. Rep. MED-SOC-RS-001-1, SOC Laboratory for Satellite Oceanography, Southampton Oceanography Centre, 193 pp.

    • Search Google Scholar
    • Export Citation
  • Ruescas, A., 2006: Remote sensing applications in the thermal range to the study of the atmospheric processes in the Western Mediterranean basin (in Spanish). Ph.D. thesis, University Jaume I, Castellón de la Plana, Spain, I 300 pp.

  • Schluessel, P., , Emery W. , , Grassl H. , , and Mammen T. , 1990: On the skin–bulk temperature difference and its impact on satellite remote sensing of sea surface temperature. J. Geophys. Res., 95, 13 34113 356.

    • Search Google Scholar
    • Export Citation
  • Stowe, L., , Ignatov A. , , and Singh R. , 1997: Development, validation and potential enhancements to the second generation operational aerosol product at the National Environmental Satellite, Data and Information Service of the National Oceanic and Atmospheric Administration. J. Geophys. Res., 102, 16 92316 934.

    • Search Google Scholar
    • Export Citation
  • Strong, A., , Kearns E. , , and Gjovig K. , 2000: Sea surface temperature signals from satellites—An update. Geophys. Res. Lett., 27, 16671670.

    • Search Google Scholar
    • Export Citation
  • Thorncroft, C., , and Flocas H. A. , 1999: A case study of Saharan cyclogenesis. Mon. Wea. Rev., 125, 11471165.

  • Torres, O., , Bartia P. , , Herman J. , , Ahmad Z. , , and Gleason J. , 1998: Derivation of aerosol properties from satellite measurements of backscattered ultraviolet radiation: Theoretical basis. J. Geophys. Res., 103, 17 09917 110.

    • Search Google Scholar
    • Export Citation
  • Trigo, I., , Davies T. , , and Bigg G. , 1999: Objective climatology of cyclones in the Mediterranean region. J. Climate, 12, 16851696.

  • Vazquez-Cuervo, J., , and Sumagaysay R. , 2001: A comparison between sea surface temperatures as derived from the European remote sensing Along-Track Scanning Radiometer and the NOAA/NAA AVHRR Oceans Pathfinder dataset. Bull. Amer. Meteor. Soc., 82, 925944.

    • Search Google Scholar
    • Export Citation
  • Walton, C., 1988: Nonlinear multichannel algorithm for estimating SST with AVHRR satellite data. J. Appl. Meteor., 27, 115124.

  • Walton, C., , Pichel W. , , Sapper F. , , and May M. , 1998: The development and operational application of nonlinear algorithms for the measurement of sea surface temperatures with NOAA polar-orbiting environmental satellites. J. Geophys. Res., 103, 27 99928 012.

    • Search Google Scholar
    • Export Citation
Save