Satellite oceanography: New results 4 page

Certain parts of the sea, especially shallow and isolated regions and those surrounded by land, are largely characterized by features encompassing enhanced annual and daily amplitudes of air temperature, recurrence and duration of low temperatures, significant horizontal gradients of all parameters of air temperature regime in coastal regions, decrease of cloud cover and air humidity, and development of breeze air circulation.

Cyclonic gyres are present in the open parts of the Bassein in front of the entrance to Kandalakshskiy, Dvinskiy, and Onezhskiy Bays. The residual tidal currents, developing due to non-linear interactions of tidal currents with the

Figure 6.26. RADARSAT SAR imagery of the White Sea and the adjacent south-western part of the Barents Sea taken on 28 February (left panel) and 4 March (right panel), 1998.

bottom relief and the coast, play a special role in forming the circulation patterns in the White Sea.

Commonly, mesoscale anticyclonic eddies generate downwelling, and result in lowering of isoclines of various hydrophysical parameters. In studies of the shelf zone in the Black and White Seas, this phenomenon was qualified as a biohydro- chemical barrier or marginal filter (Lisitsyn, 1994; Sapozhnikov, 1994; Skibinsky, 2001; Shevtchenko et al., 2002). It is believed, that the convergence zone generated in such a way, partially isolates the pelagic part of the sea from the coastal influence. Thus, it seems that the inflow of coastal water, rich in biogenic elements and phytoplankton, plays a significant role in generating patches of increased concentra- tion of phytopigments.

Using AVHRR data, it was possible to outline certain trends in marine large- scale and mesoscale thermal and dynamic processes. A diversity of fronts, upwelling areas, eddies of various genesis, zones of river discharge, and diversity of frontal zones, all described in Chapter 4, were documented as mesoscale processes. The new data allowed more detailed descriptions of water dynamics. In summary, the following new thermo-hydrodynamic features in the White Sea have been revealed:

1 A quasi-permanent front in the Gorlo.

2 Generation of fronts of different natures (tidal, upwelling, river discharge).

3 Persistent upwelling around the Solovetzkiy Archipelago.

4 Development of wind-driven Ekman coastal upwelling.

5 Dynamics of the ice field disintegration. The ice cover breaks up earliest in river estuaries and the upwelling zone near to the Solovetskiy Archipelago, but it persists for the longest period in the area neighboring the Kola Peninsula.

6 Development of coherent mesoscale structures are observed in the vicinity of the fronts associated with Dvinskiy and Onezhskiy Bays.

7 Shipborne measurements accompanied by satellite data can be used for the diagnosis and tuning of numerical model performance (Semenov and Luneva, 1999).

6.3 SATELLITE SAR AND PASSIVE MICROWAVE REMOTE SENSING

6.3.1 Satellite SAR monitoring of ice cover parameters

The NERSC/NIERSC and the Polar Institute of Fishery and Oceanography (PINRO) archive database of satellite SAR data of the White Sea and the adjacent marine and land regions is large and includes the data on sea ice relating to winter seasons of various severity. The results of thematic interpretation of ERS SAR images of the White Sea have been discussed (Melentyev et al., 1997; Melentyev et al., 2004).

This section expands on some examples of remote assessments of sea ice param- eters via exploiting conjointly or separately the data from the SAR and ASAR instrumentation installed on board, respectively, the RADARSAT and ENVISAT

platforms. The advantage of employment of RADARSAT is the possibility of both sub-surface ice sounding (the snow is transparent for radio signals) and attaining at- surface high resolution (20 m) when surveying different types of sea ice coverage; the swath can be as large as 500 x 500 km.

Figures 6.25 and 6.26 display RADARSAT SAR images of the White Sea and the adjacent south-western part of the Barents Sea taken on 28 February and

4 March, 1998, respectively. The validation of these satellite SAR images and retrieval of sea ice cover parameters were conducted using appropriate airborne multispectral data (the validation investigations were performed at three levels: satellite, aircraft, and in situ/ground truth observations).

According to NOAA satellite observations, available for the whole winter season, the ice coverage of the White Sea in the winter of 1997/1998 was more extensive than the long-term average one. In late February, 1998, practically the entire White Sea area was filled with packed drift ice, whose northern limit was in the Barents Sea. With regard to the ice age arrangement, the thin first-year ice was predominant in the composition of ice floes. A large amount of precipitation fell in the second-half of February, and the snow cover of ice floes reached 3 points (on a 5-point scale). In late February and early March, the air temperature dropped below

-20oC (Grechukha et al., 1999); however, the formed thick snow cover restrained the intensity of ice thickening. In ice-free water areas, in polynyas and leads, young ice grew rapidly forming large ice breccia. Under the influence of tidal currents and wind drift, active hummocking of gray and gray-white ice took place, and hummocky barriers 1-2 m high were formed on the edges of thick first-year ice floes. In March, 1998, 12 natural hydrometeorological phenomena occurred, partly due to strong winds and partly to abrupt weather variations (Grechukha et al., 1999). According to the data of aircraft observations accomplished on 6 March, Dvinskiy Bay, the central part of the Basin, and the Gorlo were filled with con- solidated (up to 9-10 points) fields of thin, first-year ice, which produced a real problem for ship navigation. Young ice inclusions in the drift ice massif did not exceed 1-2 tenths. The zones of small floes and ice cakes (mainly gray-white ice) were registered only near the Letniy coast of Dvinskiy Bay (areas of finely broken white ice were also revealed near the Terskiy and Zimniy coasts). The situation, anomalously difficult for navigation, that established in late February remained

the same, on the whole, up to mid-March.

As seen from Figure 6.25, on 28 February, 1998 the drifting ice edge in the south-western part of the Barents Sea was located at 40oE, near the Sviatoy Nos Cape. Further on, it displaced eastward and northward (to 42o-42o30tE). The ice edge location made it possible to qualify both the sea ice conditions established during the 1997/1998 winter as very severe, and the winter season in the western part of the Arctic Basin as extremely severe. The cumulative degrees below zero in the region were in excess of 850-900oC.

The satellite SAR image analysis confirms that the extent/concentration of the White Sea ice cover at the end of February, 1998 exceeded the average long-term values. Practically the entire sea area was covered with ice, whose compactness was 90-100%. Only individual regions of the sea surface remained free from ice, which

could be attributed to the inherent hydrological features and the nature of wind effect upon the sea ice cover. The ice-free area located northwards off the confluence of the Ponoy River, and the ice arrangement at this region, was formed due to the impact of the stable current flowing from the Barents Sea. An insignificant width of this ice-free aquatic area (qualifying as a recurring polynya) in this part of the sea indicates that in the western sector of the Arctic in the 1997/1998 winter, the eastern transfer was predominant. Two vast polynyas, revealed in the SAR image, are due to the local wind impact upon the sea ice cover. One of the polynyas stretches along the Letniy being in close vicinity to it; the other one, located beyond the coastal ice, extends along the Karelskiy coast in a north-south direction. These open water areas originated under the influence of strong southern and south-western winds, respec- tively. The opening of the two polynyas took place 2-3 days before the satellite overpass (Figure 6.25). At the very moment of the SAR overpass, the weather in the region of the White Sea was frosty and highly windy. This conclusion has been drawn on the basis of the analysis of SAR signatures discernible across the polynyas and some other ice-free areas in the northern part of the imagery. The bright signal areas seen in the left-hand part of the polynya, extending along the Letniy coast, demarcate the eastern boundary of propagation of freshened water incoming from Onezhskiy Bay to the south-eastern part of the Bassein and the outer part of Dvinskiy Bay. As is known, the waters of the Onega River determine the degree of freshening of water in Onezhskiy Bay; they also affect the dynamics of ice formation in the southern part of the Bassein. The nature of new ice formation processes in various parts of the White Sea is also determined by the effects of tide-driven currents, which revealed themselves very clearly in high-resolution SAR imagery taken on 4 March, 1998 (Figure 6.26).

Frosty weather conditions established on 28 February prevailed in the region, and this is confirmed by the radio portrait of the aquatic area located north- westward of the boundary of drifting ice. The waters directly adjacent to the masses of drifting ice (the white signal areas) represent the regions of overcooled seawater. The revealed features of the water mass state in this part of the Barents Sea make it possible to suppose that the air temperature at the moment of surveying should be -20-25oC. Thus, the SAR data provide diagnostics of the ocean-atmosphere system state, as well as forecasting possible future developments of in-water processes. The nature of the drift ice distribution and the shape of the ice edge are indicative of the persistent penetration of cold air into the region from the Arctic Basin as well as of the possible direction (westwards, to 38oE) of a further displacement of the sea ice edge. The SAR data reflect the income of transformed water of the Barents Sea to the White Sea. Shifted from the Murmanskiy and Terskiy coasts, the inflow of the aforementioned water, of Arctic Ocean origin (the dark signal signature), displaced towards the center of the Voronka.

The varying contrasts of the SAR signatures inside the drift ice mass indicate the diversity of ice types present; they also reveal differences in the parameters of their state (especially within the Bassein, Voronka, Mezenskiy Bay, and the southern part of the Barents Sea). The possibility of a detailed interpretation of sea ice state based on SAR data is a great advantage of the methods of sub-surface radio sounding of

ice. Vast dark areas in the SAR imagery correspond to regions of thin and medium first-year ice (50-70 cm thick). In our opinion, this type of ice was formed during frosty weather conditions and therefore is homogeneous in its thickness and expansion. Regions of rafting and ridging of sea ice were revealed only at the southern end of the drift ice area. The main deformations (a layered structure of rafted ice) corresponded to the stage of gray and gray-white ice. This type of ice is elastic and liable to form more layers under the influence of wind; therefore, external effects upon the ice are small-scale and do not exceed 3-5km or at most 7-8 km.

In the Gorlo, ice-breccia floes of thin first-year and young ice are identified from SAR imagery. These floes are the whelping rookeries of the Greenland seal. However, the floes are smaller in this part of the sea than within the Bassein. According to the surveillance data obtained from on-board the PINRO aircraft on 28 February, 1998, 25.6% of whelping of harp seals occurred on this date. The next SAR image of the White Sea, dated to 4 March, 1998 (Figure 6.26), confirms our conclusions drawn earlier about the state of sea ice. The prevalence of the easterly transfer persists at the beginning of the first 10-day period of March as well. The SAR data, dated to 4 March, complement those for the nature of ice drift, its direction, and dynamics.

As seen in Figure 6.26, in the time interval between 28 February and 4 March, 1998, the mass of the drift ice massif was broken into two parts; however, the general direction of the ice drift remained favorable for the migration of whelps of the Greenland seal. Forced by southerly and south-westerly winds, a substantial part of the ice was displaced and shifted from the Bassein and Gorlo northwards, in the direction of the Greenland seal feeding regions situated in the Barents Sea. Counts made from on-board the PINRO airborne laboratory Arktika revealed that, in spite of high ice concentrations, favorable conditions for the preservation of the number of pups of Greenland seal remained. According to the PINRO data, the develop- ment of whelping excludes any possibility of a seal-related ecological catastrophe because 59% of the newborn generation has already formed by 4 March, 1998.

As has been mentioned above, tidal currents are essential for the ice regime in the White Sea. We have investigated, for the first time, the possibility of revealing tide phenomenon from ERS-1/2 SAR data (Melentyev et al., 1998a). It has been shown that the effect of watering (moistening) of the ice edge by the tidal wave essentially changes the level of the sounding radio signal. The intensity of the contrast of the ice-water boundary constitutes a basis for the SAR method of both monitoring the variability of tidal currents and registering the tidal phase shift. The RADARSAT SAR imagery (horizontal polarization) also makes it possible to remotely assess the extent of tidal influence upon the sea ice state (Figure 6.26). The SAR data of 4 March, 1998 revealed considerable variability in sea ice conditions in various parts of the White Sea caused by tidal effects. In particular, the areas of quasi-continuous discharging (rarefactions) and gathering (aggregations) of ice mentioned above can be clearly seen in Figure 6.26.

The assessment of the tide quantitative parameters from the SAR imagery (4 March, 1998, 14 hr 42 min GMT) revealed their considerable spatial variability in various parts of the White Sea. According to the available computational data

(Tables of .. ., 1941), it appears that, at the time when the SAR imagery was taken, the sea level near the Letniy coast increased: about 3-4 hr passed since the phase of low water in this area. In the region of Morzhovets Island, the sea level decreased: about 1-2 hr passed since the onset of the phase of high water. In Mezenskiy Bay, the sea level had risen, but the SAR image had been taken 2-3 hr prior to the moment when the phase of high water reached its maximum.

Calculations of tide characteristics that were provided for the interpretations of the SAR imagery taken on 27 February, 1998, indicate that during the satellite overpass (at 03 hr 54 min GMT) the tidal effects in the White Sea were less pro- nounced. Their manifestations were less contrasting (Figure 6.25). In the region of the Letniy coast the sea level dropped about 3 hr after the moment of high water. At the moment of the SAR overpass of Morzhovets Island, there was low water. In Mezenskiy Bay, the sea level dropped 2-3 hr prior to the onset of low water.

Thus, based on our experience with ERS, RADARSAT, and ENVISAT SAR monitoring of the White Sea in wintertime, it can be argued that this approach allows not only assessment of a variety of ice cover parameters, but also reveals some hydrodynamic manifestations in the ice cover. Many important issues concern- ing the White Sea winter hydrology still remain inadequately studied. This warrants a wider application of remote sensing tools for studying sea ice dynamics, quasi- stable currents, ice exchange between the Barents and White Seas, as well as the generation of sub-surface structures.

6.3.2 SAR studies of the White Sea ice cover as a habitat of ice-associated marine mammals

During the last two decades, satellite remote sensing of the White Sea in the visible and IR has become an integral and indispensable part of different application studies: ice reconnaissance, and fishing and sea-hunting activities in the White Sea as well as in the immense marine areas of the North Atlantic and Arctic Basins (Chernook et al., 1998a). However, determination of ice cover parameters indepen- dently from either accompanying atmospheric conditions or the subsurface structure of sea ice have until recently been beyond the feasibility of multispectral remote sensors. The situation has essentially changed with the advent in the late-1980s/ early-1990s of the era of airborne and spaceborne SAR systems (Complex .. ., 1991; Kondratyev et al., 1992 and 1995; Melentyev et al., 1997). Satellite SAR sensing, for which snow cover is transparent, allows not only investigations of the state of the air-ice interface, but also insight to the variability of ice cover thickness. Satellite SAR data permit mapping of the surface at a spatial resolution of about 10-20 m, though even such high spatial resolution is insufficient for spotting individual animals on the ice cover.

For the first time, the idea of employing the approach of satellite SAR tracking of migration routs of Greenland seals and other ice-associated marine mammals, making use of ice as an abiotic factor of their ecology, has been suggested by Melentyev et al. (1997, 1998a). Greenland seal and harp seal populations have been chosen because these species are the most abundant and at the same time

important in terms of the economy, ecology, and nature protection. A very important fact for the remote sensing of harp seals is that only a certain type of sea ice, namely, drift ice, is used by Greenland seals for whelping and molting. This warranted the application of satellite SAR data for revealing the zones of locations of large communities of Greenland seals through the remotely assessed character- istics of drift ice. Later our approaches were extended to the application of satellite SAR technologies to diagnose the ice seal as a system and investigate the beha- vioral pattern of some representatives of phocid species in the Arctic and sub-Arctic (Melentyev, 2003a, b).

The first field experiment, aimed at comprehensive satellite SAR studies of ice and the behavioral ecology of the Greenland seal during the winter season, was organized within the framework of the Icewatch experiment - international sub- satellite investigations, which took place in the western sector of the Russian Arctic in the winter season of 1995/1996 (Johannessen et al., 1996; Melentyev et al., 1997; Melentyev et al., 1998b). Within the scope of this project, comprehensive SAR surveys of the Kara, Barents, Pechora, and White Seas were accomplished. Dedicated observations of the state of sea ice used by the sea mammals during whelping and molting periods within the SAR image were performed from on board the nuclear powered icebreaker Taymyr. Under this project, a suite of com- parative evaluations of the ice state in different parts of the White Sea have been performed, and the regional specific features of the ERS SAR signatures of drift and fast ice have been established. The above studies discovered significant contrasts between the reflectivity and emissivity of different types of ice in the microwave spectral region. The established contrasts were further linked with the mechanical, electrical and radio-physical properties of drift ice, which the populations of Greenland seals in the White Sea choose for whelping (Figure 6.27).

In the initial stage of our investigations we used airborne multispectral SAR and ERS-1/2 SAR images. Later, RADARSAT and ENVISAT SAR images were used to track the Greenland seal migration routes. The following issues were looked at:

• classification of various environmental objects and natural media in the White Sea region based on satellite SAR images;

• satellite SAR monitoring of water mass distributions and evaluation of the catchment influence (including climatic aspects);

• investigation of the pack ice dynamics and ice-water exchange between the White and Barents Seas (using SAR ice signatures as a tracer of the wintertime hydrologic processes);

• radar and microwave mapping of ice cover parameters and determination of specific features of migration routes of the Greenland seal (ice as an abiotic factor of behavior ecology); and

• combined SAR and passive microwave studies of sea ice parameters as a habitat of various species of ice-associated marine mammals (harp seal, ringed seal, gray seal, polar bear, etc.).

Studies aimed at the interpretation and validation of satellite SAR data were

Figure 6.27. ERS-2 SAR image of the White Sea and neighboring regions taken on 12 March, 1997; (inset) aerial photography of the whelping rookery, situated in the Gorlo. A represents the zone of population location of Greenland seals for whelping (dark satellite radar signatures); B is the area of ice floes and small ice floes, as well as brash ice (light and gray signal background), these zones are inconvenient for whelping, the presence of seals in these areas is unlikely, so this region can be recommended for navigation during the whelping period; C is the ice-free wind-roughened sea surface appearing as bright areas on the SAR image; and D is the terrestrial area.

performed from on board the PINRO airborne laboratory Arktika, fitted with a package of multispectral instruments, including a passive microwave radiometer (with a 5.0-cm channel) as well as a multi-channel SAR, an analogue of the Almaz SAR. The PINRO experts have long-term experience of airborne studies of marine mammals promoted by the fishing industry and developed a variety of methodolo- gies and devices for thermal IR, ultraviolet (UV), and photographic surveillance of various types of sea surface. Special attention has been paid to achieving the required

accuracy of airborne counting of the population of harp seals. Seals on the ice surface were counted by analyzing data from airborne IR measurements (Chernook, 1998a).

For the validation of satellite and aircraft microwave and SAR data, in some cases sampling at sub-satellite footprints was carried out and a description of the ice state was provided. Additionally, a thorough calibration of remote sensors was performed making use of the emission (reflection) of some objects with known emissivity/reflectivity (e.g., a smooth water surface, level fast ice zones, forested areas, concrete surfaces, metal sheets, etc.). The collection of water samples and ice cores in various regions of the White and Barents Seas was aimed at a subsequent study of dielectric properties and scattering in the microwave spectrum of ice of different ages and origins.

For a thematic interpretation of SAR images, synoptic data (e.g., auxiliary data on the air temperature and variability in near-surface wind force provided by the hydro-meteorological stations in Archangelsk, Kandalaksha, and Mezen) are very valuable.

This research on remote tracking of migration routes of whelps of the Greenland seal was supported by the European Space Agency (ESA). ESA provided a series of 12 consecutive ERS SAR images over the White Sea for subsequent thematic inter- pretation (Melentyev et al., 1997). All the SAR images were taken strictly synchron- ously with the airborne seal count. A certain number of SAR images were obtained on a commercial basis. As a result, at present we have an archive of satellite SAR images of the White Sea and adjacent regions for the period 1993-2004 (data from the satellite receiving station in Troms0, Norway). Analysis of the archived SAR images provides a basis from which we can reconstruct the life history of the sea ice. Investigations were focused on various types of drift and fast ice, the physico- mechanical, electrical, and radio-physical properties of which vary during the autumn-winter-spring season sequence.

The aerial photography data were collected in combination with the data provided by a multi-band side-looking radar (SLR). The airborne SLR data were collected at wavelengths in the 0.8-250-cm range (Chernook et al., 1998a; Melentyev and Chernook, 2002). Part of these airborne data have been systematized and prepared as the Atlas of SAR signatures of the White Sea and other marginal Arctic Seas. The Atlas also expands on radar typification of sea ice cover parameters, and its basic properties.

Remote sensing data on the formation of ice cover in the White Sea and its spatial and temporal variability were compared with the data on Ladoga and Onega

- the Great Lakes of north-western Russia - as well as the data on the Saima River and a number of tundra lakes in the northern part of European Russia during periods of freeze-up and ice destruction. These comparisons have been used as benchmark calibration objects. Airborne SAR data on marine, freshwater, freshened and saline ice, as well as the phenomena occurring at the ice edge con- stituted the basis for thematic interpretation of Almaz, ERS, RADARSAT, and ENVISAT SAR images. The appropriate comparative characteristics are given in Table 6.13.

Table 6.13. The comparative characteristics of satellite sensors.

The feasibility of satellite SAR remote sensing is essentially complemented by passive microwave measurements (Figure 6.28, see color section). Passive microwave remote sensing is also immune to weather and sunlight/skylight illumination con- ditions. The methodology we have been developing suggests a combination of airborne microwave surveys that can be used not only for the evaluation of the size and concentration of ice floes, but also for estimation of the total population number residing in whelping rookeries (i.e., for the improvement of the procedure of airborne seal counts within the areas not covered by aerial photography) (Melentyev et al., 2000). The archive with satellite microwave data on sea ice is rather extensive: the calibrated data for the period 1978-2004 from the Defense Meteorological Satellite Program (DMSP) Special Sensor Microwave/Imager (SSM/I) cover the entire period of airborne seal counting performed in recent years by PINRO in the White Sea and adjacent water bodies.

The satellite microwave data (Figure 6.28, see color section) illustrate the scale of displacement of the ice edge during winters of differing severity. Notwithstanding the rather low resolution (10-15km) of these data, they are evidently very useful for airborne ice reconnaissance and counting of harp seals. As seen in Figure 6.28, satellite passive microwave data provide documented evidence of advancement of the outer border of the fast ice zone and the location of the area of drift ice. Passive microwave surveillance also reveals a tendency for ice movement under the influence of winds and currents.

Date: 2016-03-03; view: 660