High Resolution Radiometer Sea Surface Research Articles

Estimation of Secondary Phytoplankton Pigments From Satellite Observations Using Self‐Organizing Maps (SOMs)

AbstractThis study presents a method for estimating secondary phytoplankton pigments from satellite ocean color observations. We first compiled a large training data set composed of 12,000 samples; each sample is composed of 10 in situ phytoplankton high‐performance liquid chromatography (HPLC)‐measured pigment concentrations, GlobColour products of chlorophyll‐a concentration, and remote sensing reflectance (Rrs(λ)) data at different wavelengths, in addition to advanced very high resolution radiometer sea surface temperature measurements. The resulting data set regroups a large variety of encountered situations between 1997 and 2014. The nonlinear relationship between the in situ and satellite components was identified using a self‐organizing map, which is a neural network classifier. As a major result, the self‐organizing map enabled reliable estimations of the concentration of chlorophyll‐a and of nine different pigments from satellite observations. A cross‐validation procedure showed that the estimations were robust for all pigments (R2 > 0.75 and an average root‐mean‐square error = 0.016 mg/m3). A consistent association of several phytoplankton pigments indicating phytoplankton group specific dynamic was shown at a global scale. We also showed the uncertainties for the estimation of each pigment.

Relationship between swordfish swimming behaviour and sea surface temperature in the central Mediterranean Sea during the reproductive period

Swordfish is the target species for harpoon fishery in the Central Mediterranean Sea (southern Tyrrhenian Sea and Strait of Messina). This is a seasonal fishing activity, carried out in daylight during the swordfish's reproductive period. The relationship between sea surface temperature (SST), catches, Catch per Unit Effort (CPUE) and fish swimming behaviour, recorded by direct observation, was investigated between 1999 and 2008. Data from AVHRR-SST (Advanced Very High Resolution Radiometer–Sea Surface Temperature), the meteorological station and the swordfish daily catch were recorded. A low presence of pairs was observed in the 2003 summer season, corresponding to an anomalous sea surface temperature recorded in the study area. After 2003, an early presence of pairs (in May rather than in June) was recorded. ANOVA and Bonferroni correction showed significant differences of CPUE of pairs between May and June, and June and August (F3=8.07, P<0.0001). The relationship between immature and mature adult specimens, single and in pairs, four ranges of SST, seven swimming behaviours and five maturity stages was analysed by the multiple correspondence analysis (MCA). Fish in pairs were mainly observed at temperatures over 24°C, showing a well-defined courtship behaviour, i.e. a female route. At lower temperatures, the dominant behaviour was basking in immature fish and breaching in mature fish. A significant relationship between SST and fish behaviour was detected in both immature and mature adult specimens.

The influence of initial conditions and open boundary conditions on shelf circulation in a 3D ocean-shelf model of the North East Atlantic

Data from climatology (World Ocean Atlas) and two large scale operational ocean models (Forecasting Ocean Assimilation Model (FOAM), UK Met. Office and the Navy Coastal Ocean Model (NCOM), US Naval Research Laboratory) are used to give initial and open boundary conditions for a northeast Atlantic implementation of the Proudman Oceanographic Laboratory Coastal Ocean Model System (POLCOMS). We study the effects of using the different datasets on the temperature fields and the circulation. On the continental shelf, comparisons of POLCOMS output with Advanced Very High Resolution Radiometer sea surface temperature data suggest that the effect of using different ocean model initial and boundary conditions is small and that, after 15 months of model time, the impact of the different initial conditions is negligible. Stronger evidence of influence is seen in the deeper oceanic regions of the domain. Volume fluxes through sections governing flow into and out of the North Sea, through the Irish Sea and along the shelf edge show that the impact of the different boundary conditions is small on the shelf but significant elsewhere. These results are contrasted with the use of climatology to assess the value of these Global Ocean Data Assimilation Experiment ocean model products.

Arctic waters and marginal ice zones: 2. An investigation of arctic atmospheric infrared absorption for advanced very high resolution radiometer sea surface temperature estimates

The derivation of sea surface temperatures (SST) from satellite radiometric data is well established in temperate latitudes. Water vapor is typically the greatest clear sky absorber of infrared (IR) energy between the emitting surface and spaceborne sensor, necessitating a corrective term for SST calculation. Algorithms developed for advanced very high resolution radiometers (AVHRR) use the difference in brightness temperatures between Channel 4 (10.3 to 11.3 μm) and Channel 5 (11.5 to 12.5 μm), or T45, to estimate the amount of IR absorption in the atmosphere. While relatively accurate in temperate latitudes, this approach is not applicable to Arctic waters, typically overestimating the SST by 2 to 3 K as a result of high T45 values that are not indicative of IR absorption by water vapor. The high T45 values in the Arctic may be attributable to atmospheric ice crystals. The attenuation of IR energy increases sharply across Channel 4 and 5 for ice crystals, the amount of which is a function of crystal size, shape and orientation. In the development of the Composite Arctic Sea Surface Temperature Algorithm in the North Water polynya (NOW), it was demonstrated that when T45 exceeded a threshold of 2 K the surface temperature could not be estimated owing to the presence of a clear sky absorptive feature. Observations from the NOW study led to the assessment that areas where T45 &gt; 2K were covered by ice fog. This is a significant finding since these regions must be identified to achieve an accurate mapping of the surface temperature.

Seasonal warming of Narragansett Bay and Rhode Island Sound in 1997: Advanced very high resolution radiometer sea surface temperature and in situ measurements

The warming of Narragansett Bay and the offshore waters of Rhode Island Sound (RIS) and Buzzards Bay in the spring and early summer of 1997 was studied using in situ time series data and remotely sensed advanced very high resolution radiometer sea surface temperature (SST) satellite images. High‐resolution SST images of the New England area were expanded to highlight Narragansett Bay and RIS. To validate this procedure, the remotely sensed data were compared to in situ data at the NOAA buoy in Buzzards Bay and at a spar buoy in mid‐Narragansett Bay. The standard error (1.3°C) observed at the buoy in Narragansett Bay was slightly higher than that observed at the buoy in Buzzards Bay (1.0°C). A transect line down Narragansett Bay and into RIS and another across the entrance of Narragansett Bay and Buzzards Bay were extracted from the 47 images. A thermal front was observed at the mouth of the bay with the bay being warmer in the summer and cooler in the winter than the sound. Two areas of cold water were identified in the RIS transect: a cold water plume at the tip of Long Island and a second area near the Elizabeth Islands. We believe that both were caused by vertical mixing. There were three sources of in situ time series data to compare with the SST: (1) a spar buoy with sensors in the surface and bottom waters located near the middle of the Bay, (2) observations from a shore site near the mouth of the Bay, and (3) a National Oceanic and Atmospheric Administration buoy at the mouth of Buzzards Bay. Using the spar buoy data, we were able to calculate the vertical density gradient, and we found that salinity was more important than temperature in controlling the density structure at this site. Time series temperature data from the surface water in Buzzards Bay were almost identical to those observed in the bottom waters of Narragansett Bay, indicating that bottom water in the bay originates as surface water in RIS. Using a cooling event in the surface waters at the end of July, a transport time of ∼4 days was calculated for the offshore surface waters to reach the bottom at the mid‐bay location.

Satellite Derived Sea Surface Temperature Variability Off California During the Upwelling Season

An empirical orthogonal function analysis of an advanced very high resolution radiometer sea surface temperature data set encompassing the spring–summer period during 1991–1993 was used to examine patterns of temperature variability off northern and central California. This analysis decomposed the spatial variability of remotely sensed sea surface temperature in 124 images that covered an offshore distance of 200 km and an alongshore distance of 500 km. The first mode empirical orthogonal function described 65% of the spatial variability in the data set. The spatial amplitude pattern of the first mode, which closely resembled the mean of all the images, revealed a strong cross-shore gradient with much cooler water nearshore and warmer water offshore. These areas of cool waters were associated with known upwelling locales, and the large-scale temperature structure compared favorably to that derived from historical California Cooperative Oceanographic Fisheries Investigation (CalCOFI) climatology. The mode 1 temporal amplitude function, representing the temporal modulation of this spatial pattern, was coupled with changes in the alongshore component of the coastal wind field. Interannual variations in the timing of the spring transition were reflected in this temporal amplitude time series. The second mode empirical orthogonal function described 11% of the variability in the data set. The spatial pattern revealed a north-south structure that appeared to be related to regional spatial variations in the strength of upwelling-favorable wind. The third empirical orthogonal function described only 3% of the data variability and the spatial structure suggested a meandering of the coastal surface flow. The results of this investigation indicate that the empirical orthogonal function method was effective at characterizing from the satellite data the dominant nearshore surface properties of the California Current System related to upwelling and wind, and by extension suggest that it is a suitable approach for remotely investigating other less well known eastern boundary current systems.

Autocorrelation Scales of the SST Distribution and Water Masses Stratification in the Channel of Sicily

Abstract A 1-yr-long series of Advanced Very High Resolution Radiometer sea surface temperature (SST) images is used to study the SST distribution in 1992 over the Channel of Sicily. Time- and spatial scales of the SST are studied by estimating the SST autocorrelation function directly from satellite images. SST timescales vary from 1 week to 45 days, with lower values in areas of intense upwelling. Spatial scales of the SST show seasonal variations with values from 80 to 300 km, with higher values during the summer. Oscillations with periods from 11 to 35 days modulate the annual variability of the SST spatial scales. The first four internal radii of deformation are estimated on the basis of climatological density profiles. The first internal radius varies from 7 km during the winter to 14 km during the summer. It is shown that, although differences between horizontal scales of the SST and internal radii of deformation exist, the annual variability of the SST spatial scales can be interpreted in terms of...

Cleavage of a Gulf of Mexico loop current eddy by a deep water cyclone

Eddy Triton, an anticyclonic eddy shed by the Loop Current in late June 1991, drifted SW across the central Gulf of Mexico in the first 6 months of 1992, along the “southern” of the three characteristic drift paths described by Vukovich and Crissman [1986] from their analyses of 13 years of advanced very high resolution radiometer sea surface temperature data. An expendable bathythermograph (XBT) and conductivity‐temperature‐depth (CTD) transect of opportunity through Triton at eddy age 7 months in January 1992 found that eddy interior stood 23 dyn. cm higher than periphery; this gradient drove an anticyclonic swirl transport of 9–10 Sv relative to 800 dbar. At eddy age 9–10 months and while this eddy was in deep water near 94°W, it interacted with a mesoscale cyclonic circulation and was cleaved into two parts. The major (greater dynamic centimeters) piece drifted NW to end up in the “eddy graveyard” in the NW corner of the gulf, while the minor piece drifted SW and reached the continental margin of the western gulf off Tuxpan. This southern piece of Eddy Triton then turned north to follow the 2000‐m isobath to about 24°N and later coalesced with what remained of the major fragment. Because Eddy Triton's cleavage took place just before the start of marine mammals (GulfCet) and Louisiana‐Texas physical oceanography (LATEX) field programs, the closely spaced CTD, XBT, and air dropped XBT (AXBT) data that were gathered on the continental margin north of 26°N in support of these programs allow a detailed look at the northern margin of the larger fragment of this eddy. Supporting data from the space‐borne altimeters on ERS 1 and TOPEX/POSEIDON allow us to track both pieces of Eddy Triton in the western Gulf and follow their spin down in dynamic height, coalescence, and ultimate entrainment in January 1993 into another anticyclonic eddy (Eddy U).

The physical structure of cold filaments near Point Arena, California, during June 1987

Shipboard conductivity‐temperature‐depth, acoustic Doppler current profiler, and continuous temperature and salinity observations were made in the coastal transition zone off Point Arena, California (39°N, 124°W), during June 15–28, 1987, to describe the hydrographic structure and velocity fields associated with the cold filaments found there. An adaptive sampling plan was used to measure the properties of these filaments, guided in real time by satellite Advanced Very High Resolution Radiometer sea surface temperature imagery and feedback from the in situ sensors. The primary feature observed was a large, cool (12.0°–13.5°C), salty (32.7–33.0 psu) filament which extended over 200 km offshore from Point Arena and exceeded 500 m depth. This feature was bounded in the offshore direction by a continuous equatorward meander, with offshore velocities (60–87 cm s−1) on the northern edge of the cool filament and onshore velocities (69–92 cm s−1) along the southern edge, and persisted for at least 3 weeks. A second feature was advected into the study area from the north by an anticyclonic eddy offshore and later merged with the Point Arena filament. Smaller (30 km wide by 50–100 km long by 50–100 m deep) very cold (10.0°–12.0°C) high salinity (&gt;33.0 practical salinity units (psu)) features were observed within the Point Arena filament, but persisted for only 6–10 days. The net volume transport of the larger feature was offshore at ∼3 × 106 m3 s−1 and suggests it was fed by an inflow to the region from the north. The smaller features were correlated with bursts of equatorward wind stress on an event by event basis but not with times of large‐scale wind stress convergence. Surface drifters deployed during the experiment closely followed the surface dynamic topography. Some of the drifters followed the path of the offshore meander, while others moved south inshore.