ERS-1/2 SAR Detection of Natural Film on the Ocean Surface

Abstract

Ocean mesoscale phenomena such as eddies and current convergence zones can often be seen in SAR images due to characteristic patterns caused by natural film induced damping of the waves. Such films have also been found to exert a significant effect on air-sea gas exchange, which may be important for the global scale climate system. Satellite synthetic aperture radar (SAR) may prove very useful to quantify the extent of natural film. To investigate the composition of these films and their effect on radar return, samples of the sea surface were taken simultaneously with ERS-1/2 SAR coverage of a fjord (the Korsfjord experiment), and later in coastal ocean areas (COASTWATCH'95). During COASTWATCH'95, simultaneous observations were made with a shipmounted C-band dual-polarized Doppler radar, and surface drifters were deployed to investigate the surface current variations in vicinity of the slicks. The dependence of the existence of slicks on wind speed is also investigated, and an estimate of natural film distribution during the experiment period given.

Keywords: SAR, Natural Film, Doppler Radar

Introduction

The spaceborne SAR, with its high spatial resolution (25m), may contribute significantly in global mapping of natural films at the ocean surface during moderate wind conditions. Damping of the capillary and short gravity waves by the film, which are sensed by radars via the Bragg backscatter, produce dark slick signatures in the ERS SAR imagery. So far, oil slicks have been investigated and are therefore well documented in satellite SAR experiments (Wismann, 1993; Wahl et al, 1996). Moreover, artificial monomolecular films have been studied a number of times with airborne radars (Alpers et al, 1991; Huhnerfuss et al, 1994). In contrast, only one previous study exists in which the chemical components of natural films were sampled simultaneously with satellite SAR acquisition. This experiment was conducted under calm wind conditions in a Norwegian fjord (Espedal et al, 1996) and showed that there exists a chemical difference between slicked and non-slicked areas in an ERS-1 SAR image.

The primary objective of the work reported here, was to extend the documentation of the composition of natural films and their effect on radar return, to include different coastal oceanographic and atmospheric conditions. This may in turn give new valuable insight into the geophysical processes responsible for accumulation of different natural films in coastal areas. Eventually this may also help develop satellite SAR as a tool in mapping the global extent, distribution and variability of natural films, to be used in e.g. climate studies. Knowledge of natural films are also important for SAR oil-spill detection systems, since these films are difficult to distinguish from oil in the SAR imagery.

The COASTWATCH'95 Experiment

Extensive temporal and complete spatial satellite SAR coverage of the experiment region was possible due to the ERS-1/2 tandem mission providing exact repeat coverage on ground with 24-hour separation. There were also crossovers from ascending/descending ERS passes at 11-13-hour separation. A total of 53 SAR scenes (each covering 100 x 100km) from the experiment region were analyzed during the field campaign from 11 September to 1 October 1995. In addition, 18 scenes covering the same region in the months before and after the experiment, have been analyzed in this work.

A C-band (4.92GHz) continuous wave (CW) Doppler scatterometer with VV and HH polarization, was installed to acquire shipborne radar backscatter measurements during COASTWATCH'95 (Dano and Kletzli, 1996). The Doppler scatterometer measures surface roughness (radar cross section; RCS). The surface current is retrieved from the Doppler shift (spectral mean), and hence surface current gradients can be measured at very fine spatial resolution (m). In addition, the spectral width provides a measure of the distribution of the speedes of scatterers (if the return was close to the noise floor, i.e. in a slick, the Doppler width will artificially broaden since white noise theoretically has an infinite bandwidth). The scatterometer was mounted on the bow of the R/V Haakon Mosby, 6.5m above the water level. The "spatial resolution" was therefore approximately 2m in the along track direction and 1.5m in the cross track direction. Measurements were typically made with the ship moving at a speed of 2-3m/s. The radar cross section (RCS) and spectral statistics plots in this work are generally based upon a 35sec running average (equal to a distance of 100m). This will tend to smear out the apparent width of the actual transition regions. Only VV polarization results are included here, enabling a more direct comparison with the ERS SAR data.

During the first part of the COASTWATCH'95 experiment (11-19 September), surface slick material was collected using two remotely controlled surface slick samplers, INTERFACE I (length 1.2m) and II (length 2m) (Knulst, 1996). They were both equipped with a rotating hydrophilic teflon drum (Espedal et al, 1996). The INTERFACE II was also fitted with a GPS receiver, temperature probes, an anemometer and a data logger. Bulk water samples were obtained at 20cm depth for comparison. The samples were analyzed for salinity, anions, fatty acids, alcanes, total organic carbon, cations and trace metals. Ultra violet (UV) and visual light (VIS) scans were made on all samples, and pH, specific conductivity and turbidity were determined.

A Slick Case Study

Figure 1: The slick study on 17 September, 1995. (1) the rough zone, (2) the slick zone, and (3) the slightly rougher zone. The current was stronger in the convergence zone (1) than in the slick zone (2). The circles in (a) indicate surface current drifters, and the long arrow gives the R/V Haakon Mosby ship track.

The RCS plot from the Doppler scatterometer (17:50GMT, Figure 2) shows a clear increase as the R/V Haakon Mosby crossed the front perpendicularly from the smoother to the rougher side. A drop of 5dB can be seen when crossing into the slick zone (zone 2, Figure 2). With a ship speed of 2.3m/s, it took approximately 8sec to cross the slick. This implies a maximum slick width of 18.4m, consistent with visual estimates of a 10-15m wide (and 10km long) slick zone (Fig.~\ref{fig:drawing}). Abundant surfactant material was visually observed in this slick zone. The corresponding current plot (Figure 2), indicates a change in the surface current, when crossing the slick, of about 0.25m/s over a distance of about 50m. This implies the existence of a convergent current with a strain rate (convergence) on the order of 0.005s^{-1}. The Doppler radar measured a surface current of 1.1m/s+-0.1m/s (component along ship heading) in zone 1. These radar measurements are supported by visual observations of drifters deployed at the east side of the front (zone 1). The drifters were tracked using a handheld Global Positioning System (GPS), with an estimated accuracy of +-0.2m/s. They were found to drift into the slick zone (northwestward direction) with a velocity of 0.7m/s, and then remained there (see Figure 1).

Figure 2: Doppler radar results from 17 September 1995 (17:50 GMT). The different zones (see Figure 1) are indicated in the radar cross section (RCS) plot.

The chemical analysis of the samples (inside and outside slick, surface and bulk water samples) indicate that the surface microlayer was enriched both in fatty acids and in alkanes, compared to bulk water samples. However, a clear difference was observed between in and out of slick fatty acid and alkane concentrations. The slick sample was enriched 10 times in fatty acids compared to the bulk water, while the out of slick sample was only about 1.5 times enriched compared to the corresponding bulk water. The relative amounts of the different components of available fatty acids in inside and outside slick samples were similar, in spite of these concentration differences. This indicates some connection between in and out of slick material, consistent with the behavior of a converging current front.

The ERS-2 SAR image from September 17, 1995 (10.50GMT), indicates several slicks east (south- and north-east) of the location of in situ sampling (Figure 2). Some of these dark slick areas are manifestation of eddies which may increase the concentration of natural film (Johannessen et al, 1996b). Moreover, evidence of film material extending offshore are seen in vicinity of the fjords. Taking the northwesterly drift speed of 0.7m/s and the wind speed of 3m/s (from west-southwest) into account, some of the film material seen in the SAR image may have reached the in situ sampling location by the time sampling began (7 hours later). Thus, assuming the film material in these eddies is representative also for the in situ sampling location, the backscatter damping of 4.8dB may be used as an estimate for the sampled slick.

Figure 3: Part of the ERS-2 SAR image taken on 17 Sept. 1995 (10:50GMT). In situ measurements were taken at the location of the asterisk (7 hours after SAR imaging). With a northwesterly drift speed of 0.7m/s during 7h, the white half-circle indicates the maximum distance (18km) between the in situ sampling location, and imaged slicks that were able to reach this location before sampling began.

Quantification of Film Coverage

Figure 4: Outline of the main steps in a slick discrimination algorithm for SAR imagery.

The scatter plot in Figure 5 indicates that for increasing wind speeds, the percentage of film coverage in the SAR scenes expectedly decreases. Many of the 71 investigated SAR scenes have the approximate same % film coverage. The largest film coverages (up to 40%) are found for the lowest wind speeds, 2.5m/s (in these cases large homogeneous areas are classified as caused by low winds, while the surrounding inhomogeneous slick features including eddies are classified as caused by natural film). For 5 to 10m/s wind speed, the natural film coverage has sunken below 5% in all cases. For 12.5m/s and 15m/s, the coverage is below 1%. The backscatter values for slicks classified as natural film varied between -6dB and -26dB (noise floor). This wide range of values indicate a problem in slick quantification. Large areas of less concentrated film may cause less damping, and thus appear in the SAR imagery as dark grey areas (Nilsson and Tildesley, 1995), and not as distinct slicks.

Figure 5: Scatter plot of % film coverage vs. wind speed, based on natural films from 71 ERS-1/2 SAR scenes (many points/values are overlapping).

Figure 6: Scatter plot of backscatter damping vs. wind speed, based on natural films from 71 ERS-1/2 SAR scenes (many points/values are overlapping).

Discussion

The chemical analysis of the samples showed that trace metals were generally enriched in the surface microlayer, especially in the slicked areas. Trace metal enrichment is related to particulate matter present in the films, and possibly associated with humic substances (measured as UV absorbance). Humic substances were also enriched in the surface microlayer. Concentrations of free fatty acids indicate that these are raised inside slicks, but also in one of the outside slick samples. In this outside slick case, the high wind speed (9-10m/s) predicts that much lower values should have been found. The sample is therefore believed to be contaminated (probably by biological fats/fish oil from fishing vessels). Most fatty acids originated from marine organisms (plankton and macro-algae), since there are very few high molecular weight fatty acids available. The latter might have indicated higher plant material from terrestial sources, that could have been introduced from land (e.g. river runoff) via fjords into coastal waters.

Backscatter values for natural film in the 71 ERS-1/2 SAR scenes investigated in this work, varied between -6dB and -26dB (noise floor). The typical drop in backscatter caused by the natural films varied between 2dB and 11dB. The scatter plot in Figure 6 indicates that increasing wind speeds expectedly lead to decreasing damping by natural film. The wide range of damping values, especially for low wind speeds, is probably due to varying film composition and thickness. Moreover, 27 SAR scenes were classified as showing no natural film (no damping). They all had a wind speed of 5m/s or more.

When comparing the COASTWATCH'95 results with the slick sampling experiment carried out in Korsfjord south-west of Bergen, Norway (Espedal et al, 1996), none of the surface microlayer samples seemed heavily impacted by mineral oil products, even though the fjord carried considerable ship traffic. Less, and smaller molecules of fatty acids were found during COASTWATCH'95 compared to the fjord case. The small fatty acid molecules are associated with marine organisms, while the larger come from terrestial sources. Moreover, the investigated slicks from COASTWATCH'95 were generally thinner than those studied in Korsfjord (film thickness is estimated by dividing the sampled volume by the teflon drum area times its number of rotations). This is most likely related to fetch lengths which are shorter in fjords than in coastal areas (Williams et al, 1986; Wheeler, 1975). However, both experiments showed an approximate one order of magnitude increase of the concentration of fatty acids inside compared to outside slicks. For the Korsfjord case, backscatter decreases were found to vary between 6dB and 17dB, which is a slightly higher damping than found in the COASTWATCH'95 cases (2-11dB). This is probably due to film thickness, i.e. the thicker films found in the fjord give higher damping.

Conclusions

Acknowledgement

References

Alpers, W., et al., 1991

The damping of ocean surface waves by monomolecular sea slicks measured by airborne multi--frequency radars during the SAXON--FPN experiment. Proceedings of IGARSS'91, pp. 1987-1990.

Asher, W., 1996

The Sea Surface Microlayer and its Effect on Global Air/Sea Gas Transfer. The Sea Surface and Global Change, edited by Duce R., and P. Liss, in press, Cambridge University Press.

Dano, E., and D. Kletzli Jr., 1996

Low cost CW scatterometers for field and laboratory use. Proceedings Natl. Radar Conference, IEEE Press.

Duce, R., et al., 1994

The sea-surface microlayer and its potential role in global change. ESAMP WG 34, XXIV/7, A&M University Press, College Station, TX.

Espedal, H.A., et al., 1994

Norwegian Surface Slick Report. Tech. Rep. No. 81, Nansen Environmental and Remote Sensing Center, Bergen, Norway.

Espedal, H.A., et al., 1996

Satellite detection of natural film on the ocean surface. Geophys. Res. Letters, 23, pp. 3151-3154.

Frew, N., et al., 1990

Impact of Phytoplankton--Generated Surfactants on Air--Sea Gas Exchange. J. Geophys. Res., 95, pp. 3337-3352.

Furnes, G., 1994

Discharges of produced water from production platforms in the North Sea. Report R-064641, Norsk Hydro, bergen, Norway.

Gershey, R., 1983

Characterization of seawater organic matter carried by bubble-generated aerosols. Limnol. Oceanogr., 28, pp. 309-319.

Huhnerfuss, H., et al., 1987

Attenuation of wind waves by monomolecular sea slicks by the Marangoni effect. J. Geophys. Res., 92, pp. 3961-3963.

Huhnerfuss, H., et al., 1994

Classification of sea slicks by multifrequency radar techniques; New chemical insights and their geophysical implications. J. Geophys. Res., 99, pp. 9835-9845.

Johannessen, J.A., et al., 1996b

Coastal ocean fronts and eddies imaged with ERS--1 synthetic aperture radar. J. Geophys. Res., 101, pp. 6651-6667.

Johannessen, O.M., et al., 1996a

COASTWATCH-95: ERS-1/2 SAR Applications of Mesoscale Upper Ocean and Atmospheric Boundary Layer Processes off the Coast of Norway. Proceedings of IGARSS'96, 11, pp. 1158-1161.

Knulst, J.C., 1996

Interfaces in aquatic ecosystems: Implications for transport and impact of anthropogenic compounds. octoral Dissertation, SE-LUNBDS/NBKE-96/1010+136, Lund University, Sweden.

Lucassen-Reynders, E., and J. Lucassen, 1969

Properties of capillary waves. Adv. Colloid Interface Sci., 2, pp. 347-395.

Lyzenga, D.R. and J. Bennett, 1988

Full-spectrum modeling of Synthetic Aperture Radar internal wave signatures. J. Geophys. Res., 93, pp. 12345-12354.

Nilsson, C.S., and P.C. Tildesley, 1995

Imaging of oceanic features by ERS-1 Synthetic Aperture Radar. J. Geophys. Res., 100, pp. 953-967.

Tanis, F.J., et al., 1989

Physics of EOM. Technical Report no. 028, Environmental Research Inst. of Michigan, USA.

Wahl, T., et al., 1996

Radar satellites: A new tool for pollution monitoring in coastal waters Coastal Management, 24, pp. 61-71.

Wheeler, J.R., 1975

Formation and collapse of surface films. Limnol. Oceanogr., 20, pp. 338-342.

Williams, P.M., et al., 1986

Chemical and microbiological studies of sea-surface films in the southern gulf of California and off the west coast of Baja California. Mar. Chem., 19, pp. 17-98.

Wismann, V., 1993

Radar signatures of mineral oil spills measured by an airborne multi-frequency radar and the ERS-1 SAR. Proceedings of IGARSS'93, pp. 940-942.

Keywords: ESA European Space Agency - Agence spatiale europeenne, observation de la terre, earth observation, satellite remote sensing, teledetection, geophysique, altimetrie, radar, chimique atmospherique, geophysics, altimetry, radar, atmospheric chemistry