LAKE LADOGA SURFACE FEATURES ON THE ERS-1 SAR IMAGERY

ABSTRACT

Three series of experiments have been conducted in Lake Ladoga (Russia) with use of ERS-1 Synthetic Aperture Radar (SAR) imagery on July, 1993-1995. To support SAR imagery NOAA AVHRR data and contact measurements have been also collected. The paper reports on the results obtained from analysis of available ERS-1 SAR imagery and subsatellite observations as a step in the investigation the Lake Ladoga with use a spaceborne SAR. The lake surface features visible on the ERS-1 SAR images include internal waves, natural films, associated with summer bloom, variations of the wind speed and interface of two water masses (front). A number of examples of ERS-1 SAR images are presented to illustrate these phenomena. A general conclusion that the ERS SAR is a very powerful tool for studies hydrological phenomena in large lakes is made.

INTRODUCTION

Since long a synthetic aperture radar (SAR) has been used to study ocean phenomena [1]. The ability of SAR to map water surface features has also many potential applications in hydrology. It is evident that large lakes, like the Great Lakes, Lake Ladoga and Lake Onega in Russia are good test sites to study some important hydrological phenomena. Large lakes also exhibit surface signatures of processes of different spacial and temporal scales associated with both lake upper layer and atmospheric boundary layer dynamics. These processes are poor understood yet because there are problems with collection of in situ measurements. With the launch of the ERS-1 acquisition of all-weather SAR images became possible that reduces a lack in data.

Lake Ladoga located NE of St.-Petersburg (Russia) near the Finnish border is the largest freshwater lake in Europe which covers an area of about 18,000 sq.km, has a length of about 190 km, a width of 130 km, and a maximum depth of 230 m (Fig.1). From December to April the lake is covered with ice. The lake is stratified from May to November and has a characteristic lymnological regime due to the presence specific temperature distribution.

It is evident that some hydrophysical phenomena (thermal fronts, upwellings, oil slicks, internal waves etc.) have the same nature both in large lakes and in ocean and seas. It is possible in a number of cases to study these phenomena in large lakes under more controlled conditions.

Fig.1 Morphometry of Lake Ladoga, location of ERS-1 SAR image of July 12, 1993; straight lines are RV tracks

Although results of several field studies of thermohydrophysic processes in Lake Ladoga have been recently summarized [2-4], its surface features still remains to be studied. In the last years some studies were devoted to investigation of a thermal regime using NOAA AVHRR data [5,6].

Recently the ERS-1 SAR has demonstrated significant ability to image a large variety of ocean phenomena as well as phenomena associated with atmospheric boundary layer [1]. Therefore, in 1993 a field program in Lake Ladoga was initiated by NPO Mashinostroenia, Lymnology Institute, Russian Academy of Sciences and Institute of Oceanography, University of Hamburg to study the relationship between radar backscattering and hydrological phenomena occurred in the lake. An idea of the study was to use the specific thermal structure in the lake as a source of different hydrophysical processes and surface roughness variations for a comparison the response of the ERS-1 SAR and contact measurements. Other interest was to study different phenomena recently detected in the large lakes such as internal waves. Since the thermal bar and post-thermal bar situation in Lake Ladoga are associated with rather characteristic surface temperature distribution we expected that combined use of SAR and infrared (IR) data as well as contact measurements can improve our understanding the hydrophysical processes in the lake.

Fig.2 ERS-1 image (100x200 km) of Lake Ladoga acquired on July 12, 1993, at 8:56 UTC. It shows large dark region (slick) over cold water body. © ESA, 1993

Three series of experiments have been conducted on July 1993-1995 with acquisition of ERS-1 SAR imagery and IR imagery from the NOAA-9 satellite. In situ measurements carried out by research vessel (RV) operated by Lymnology Institute were simultaneous with the ERS-1 SAR and made it possible to identify a number of the surface signatures visible on the ERS-1 SAR and relate them to the hydrological phenomena. This paper presents an overview of phenomena visible in the ERS-1 SAR imagery of Lake Ladoga.

a) Cold Water Body

The ERS-1 SAR image in Fig.2 shows large dark region among the bright wind-roughened water surface adjacent to the north eastern shoreline. In an area to south-east of the Isl.Valaam the transition between the bright and the dark areas is very sharp, whereas in other areas it is quite diffuse. Note the wind direction (wind from the NE, speed 3-4 m/s) can also be inferred from the direction of the streaks visible in bright section of the image.

Fig.3 Surface temperature distribution in Lake Ladoga extracted from data of AVHRR aboard the NOAA-9 satellite (12.07.1993, 7:12 UTC)

The NOAA-9 image acquired 1:44 hour before the ERS-1 contains signature of same scale and size. It is in a good quantitative agreement with those seen in the ERS-1 SAR image. Then, as a initial test of suggestion that the cold water region looks dark on the ERS-1 SAR image, a comparison of surface temperature distribution (Fig.3) extracted from data of AVHRR aboard the NOAA-9 satellite and spacial distribution of SAR image brightness have been conducted. A comparison shows that the boundary between the bright and the dark regions on the ERS-1 SAR image coincides with location of 10-isotherm in the western part of the lake and 8-isotherm in the eastern and that the sharp boundary between these two regions in a area southeastly off Isl.Valaam is associated with strong horizontal surface temperature gradient.

Fig.4 Profiles of characteristics along the ship tracks: SAR image intensity (a); bulk temperature measured by ship-based towed thermistor from 5:30 to 15:35 UTC (b); surface temperature extracted from AVHRR data (c); difference between surface and bulk temperatures measured by ship-based IR-radiometer (d).

Fig.4 shows profiles of measured bulk temperature, temperature derived from the AVHRR data and SAR image intensity plotted versus distance, i.e. along ship's tracks (see Fig.1). One can find a good agreement between these three profiles. At the positions A and B the horizontal temperature gradients are the strongest. In these areas the reduction of the NRCS also is the largest (about 18 dB). Also note that here exists an area where the surface temperature is higher then the bulk temperature (Fig.4, plot c).

A close inspection of Fig.3 and Fig.4 shows a relation between data of different sensors. It means that the area occupied by cold water strongly affect on distribution of surface roughness. In the moment of image acquisition the wind speed over the central part of the lake was around threshold for ripples generation (<3 m/s), therefore there are no Bragg waves in the lake center, although surrounding areas was preferred for ripple generation (wind speed 3-5 m/s). The cold lake center and warm land modified the wind field such way that it speed decreased towards the lake center. It is found from another ERS-1 SAR images of July 7 and July 15, 1993 that the cold water region is completely or partially covered by waves under moderate winds (>5 m/s).

The modification of the NRCS from -9 dB in the bright area to -28 dB in the dark can be explained by the wind stress variations induced by changes in atmospheric boundary layer over the transition zone with large temperature gradients. Atmospheric stability determined by air-water temperature difference in this case was changing from near-neutral (<+1C) at the eastern lake shore to stable (+7C) in the center of dark area of the SAR image. A decrease in wind stress due to the change of stratification parameter caused a decrease in surface roughness and, consequently, in the NRCS.

The thermal boundaries with strong horizontal gradients were also detected in the ocean by a SAR due to expression through the surface roughness changes associated with atmospheric boundary layer and wind stress transformation [7-9]. Sharp surface temperature gradients at the fronts result in changes atmospheric stability which, in turn, changes atmospheric turbulence, modify the wind profile and wind stress. Mitnik & Lobanov have studied some cases and shown that correlation of the NRCS and surface temperature usually occurs when wind speed < 7 m/s [8] These effects were also found by comparing simultaneous SAR and IR imagery [9]. Since thermal structure of Lake Ladoga is similar to those in the ocean is expected an appearance similar features in the SAR imagery.

It should be noted that large slick on Fig.2 can't be attributed to oil spill. Observations collected during the experiment haven't shown the presence of oil films on the lake surface in the dark region. The most evidences of different authors show that this area very often looks mirror-smooth under weak winds [2].

Separate important item is increasing of surface tension and kinematic viscosity with decreasing of temperature can influence on generation of capillary waves. The last factor may give contribution to reduction the NRCS at significant temperature contrasts [10,11]. Similar effect is observed in upwelling regions where generation of ripples is also lowered [11]. We can't also exclude the formation under calm conditions and day heating of thermal skin which is typically cooler and denser than water deeper and which can prevent generation of capillary waves in favorable conditions [12].

b) Internal Waves

Some of ERS-1 SAR images acquired over Lake Ladoga show manifestations of internal waves (IW). IW are usually generated in Lake Ladoga in period when stable stratification is formed. This period for Lake continues from the beginning of May to the end of October. Earlier internal waves were contact measured in large lakes [13] and detected recently by a SAR [14]. Surface manifestations of IW in Lake Ladoga have been firstly noted on the S-band SAR image which was acquired in 1989 by Kosmos-1870 satellite. Later IW have also been observed in ALMAZ-1 SAR images [14,15]. They become visible on SAR images because they associated with variable surface current field which modify the surface roughness.

IW are usually generated on the shallow thermocline typically located on the depth 5-10 m in the central deepest part of the lake, propagate in the southern directions and have typically wavelengths ranged from 100 to 400 m. Calculations with use of two-layer model have gave the propagation speed about 10-16 cm/s. Maximum amplitude of IW measured in Lake Ladoga is 10 m [16]. Since we have no special measurements of IW characteristics during the ERS-1 overflights, we give above the values obtained from analysis of a number of SAR images containing IW manifestations and contact measurements [15-17].

Fig.5 shows IW packet as it seen on the ERS-1 SAR image taken over the central part of Lake Ladoga on July 12, 1993. The packet consists of three dark bands with slowly decreasing wavelength, coincident with the smooth regions, on a nearly uniform background. The width of the first dark band on the SAR image is about 350 m. The IW propagated with speed about 15 cm/s to the south direction. The generation mechanism of the IW in large lakes is now under discussion [15,17].

Fig.5 Fragment of the ERS-1 SAR image of July 12, 1993 containing manifestations of internal waves

Note, that ocean tidal IW are predominant among other types of IW which appear in the ocean at the shelf edge or in condition with sharp depth changes. This type of IW are most frequently visible in SAR imagery but another types are infrequent and their origin is not so clear.

Contact measurements carried out in Lake Ladoga during the LADEX-80 experiment shown that generation of high frequency IW in the lake resulted from shear unstability of subsurface currents associated with strong current gradients [16]. Periodic current strengthens and reductions result from large-scale wind stress variations. Energy of wind redistributes between wind waves, wind-induced currents and seiches. It is shown that all mentioned processes can lead, in turn, to internal wave generation in the lake but with different spacial and temporal scales. Generation of IW in the lake according to [16] is highly correlated with synoptic activity, i.e. the maximum activity occurs after sharp atmospheric fronts passes. Cyclones or fronts, however, are not the single factor affecting the intensity of IW activity. Another possibility is discussed is that IW may originate due to instability of the zone with sharp temperature gradients. This interface is strongly unstable, and its energy is approximately an order of magnitude greater than the energy of the average movement [3]. Moreover, another important factor of IW generation may be seiches produced by atmospheric forcing on the lake surface.

c) Atmospheric Boundary Layer Features

A number of ERS-1 SAR images of Lake Ladoga show a pronounced features associated with atmospheric boundary layer transformation over the lake. The image shown in Fig.6 exhibits three regions of different SAR brightness (or different wind speed) which associated with wind speed increase.

Fig.6 ERS-1 image (100x100 km) of Lake Ladoga acquired on July 16, 1995, at 19:45 UTC, showing the three regions of different wind speed. © ESA, 1995

Typically, a water surface appears black in ERS-1 SAR imagery if wind speed does not exceed 3 m/s. In the case of higher wind speeds over water that is dark the presence of surface film is probable. However, at wind speed exceeding 6-7 m/s slicks don't appear on SAR images. Therefore slicks formed by natural films are usually observed at the wind speed between 3 and 7 m/s.

Fig.7 Image intensity scan through the three regions (from the top to the bottom of the image on Fig.6). On the left vertical axis the NRCS is plotted and on the right vertical axis the wind speed calculated from the CMOD4 model

Profile of the NRCS through these three regions visible in the Fig.6 is plotted in Fig.7. The NRCS changes from -28 dB in the dark region through -13 dB in slick-covered region to about -7 dB in bright region. The NRCS results in the wind speed between 2 and 4 m/s according to the CMOD4 model and wind speed is about 2 m/s in slick-covered region. The meteorological station on the south shore reports the wind speed of 5 m/s. The threshold wind speed estimated by Donelan & Pierson for C-band is 3 m/s [18]. The result suggests that the ERS-1 SAR responds in specific way to surface roughness in conditions of weak winds.

It is also well-known when wind is blowing from land to the water it's speed changes due to a number of effects. Similar features on a number of ERS-1 SAR images can be originated from increase of wind blowing from land to the lake due to decrease of surface roughness and turbulent exchange [19].

Also interesting wave pattern is in the north-eastern part of the ERS-1 SAR (is shown in Fig.8) which appears to be atmospheric gravity waves generated behind mountain islands of the Valaam archipelago. They are typically lee waves and also are associated with varying wind stress which modulate surface roughness.

d) Features Associated with Biological Activity

It is well now known that increased biological activity of plankton can change conditions on the water-air interface causing strong backscatter signal variations in the seas and lakes. The SAR image on Fig.6 contains surface slicks which usually concentrate near the shores, bays and river mouths but under favorable conditions may occupy the whole of the lake surface. The variability of thermal structure in Lake Ladoga determines the special inhomogeneity of hydrophysical, hydrochemical and, in turn, hydrobiological characteristics. The spacial distribution of chlorophyll-A on the water surface shows close correspondence to the spacial distribution of surface temperature [3]. The slicks visible on the ERS-1 SAR images appear to be associated with summer bloom and to be caused by natural films. They appear as dark strips or spots on the SAR images with characteristic width of hundred meters and can exist in wind speed range 3-5 m/s. But using only monofrequent SAR it is very difficult to determine what kind of oil films either natural or anthropogenic they are.

Due to a SAR was found another type of surface signatures connected with algae activity in large lakes. Lake surface manifestations of algae accumulations have been firstly noted on the S-band Kosmos-1870 SAR image acquired over Lake Onega in summer 1989 [20]. The conducted contact measurements made it possible to interpret this SAR image containing different size patches of non wind-induced roughness. This roughness was attributed to high algae concentration and appears as bright spots in the SAR imagery. Fig.8 shows local patches of strong backscatter signal with sharp boundaries in the ERS-1 SAR image where some patches are brighter than surrounding water surface covered by wind-induced short waves. It may be hypothesized that these patches are similar nature that mentioned above, but, unfortunately, there are no strong experimental evidences. Most probably finger-like signatures visible in southern part of Fig.6 have also similar origin.

CONCLUSION

The ERS-1 SAR imagery collected over Lake Ladoga from 1993 to 1995 shows various signatures on the lake surface. It is documented that the ERS-1 SAR can provide observations of mesoscale surface features in large lakes including thermal boundaries, internal waves, algae patches, wind inhomogeneities and oil slicks. But many of lake surface signatures visible in the SAR imagery aren't simply to interpret because they are frequently manifestations of complex phenomena implicating different physical mechanisms with different spacial and temporal scales. An interpretation of ERS-1 SAR imagery can be significantly improved by combined use of a SAR and IR data.

Under favorable wind conditions (wind speed 3-4m/s) location of zone with maximum temperature gradients is determined by the line with jump of the SAR image brightness or the NRCS. The dark area on the SAR image is related to cold water body which is located in the deepest part of the lake. In presented case there is a correlation between profiles of temperature measured in situ and the SAR image intensity. But there is no simple relation between temperature distribution and the NRCS.

The radar pictures of Lake Ladoga strongly depend on subsurface winds. It is also shown that the most of the phenomena became visible in the ERS-1 SAR imagery due to the wind stress fluctuations. Typically they appear as surface features under winds of 2-5 m/s. Unstable wind field over the lake modifies the lake surface roughness and thus, the radar backscatter. The transformation of the wind field over the lake as well as intensification of synoptic activity leads to atmospheric forcing of a number of hydrological processes which, in turn, appear on the lake surface as different signatures.

Of special importance is the observation of internal waves in large lakes. The ERS-1 SAR imagery has also provided an experimental evidence of the existence of internal waves in Lake Ladoga.

In despite of several speculations about the origin of a number phenomena in the Lake Ladoga are made, it is evident that a satisfactory explanation of the responsible generation mechanisms has yet to be presented. Special attention should be paid to processes on the air-water interface which determine the SAR imaging mechanism in conditions of weak winds (2-3 m/s). So, a general conclusion can be made: although the ERS SAR is a very powerful tool for hydrological studies of large lakes, additional studies are needed to improve our understanding of above mentioned phenomena.

Fig.8 ERS-1 image (100x200 km) of Lake Ladoga acquired on July 7, 1994, at 19:40 UTC. It shows the patches with sharp boundaries and manifestations of microscale atmospheric lee waves generated by wind blowing over mountain Islands of Valaam. © ESA, 1994

ACKNOWLEDGMENTS

The authors gratefully acknowledges to the ESA for providing the ERS-1 data. Special thanks to A.Michelsen of Hamburg University, V.Zaitsev and S.Egorkin of NPO Mashinostroenia for preparing the figures.

REFERENCES

1. W. Alpers. ERS-1 SAR views the ocean: an assessment. Proc. IGARSS'94 Pasadena, California, USA, 8-12 Aug. 1994, p.1951-1953

1. A.I.Tikhomirov. The thermal regime of large lakes, Leningrad: Nauka, 1982, 232 p. (in Russian)
2. M.A.Naumenko. Some aspects of the thermal regime of large lakes: Lake Ladoga and Lake Onega. Water Poll. Res.J.Canada,1994. v.29, Nos.2/3, p.423-439
3. M.A.Naumenko, S.G.Karetnikov, A.I.Tikhomirov. Main features of the thermal regime of Lake Ladoga during ice-free period. Proc. 1st Int. Lake Ladoga Symp. Developments in Hydrobiology, Kluwer Academic Publishers, 1995.
4. M.A.Naumenko and S.G.Karetnikov, Application of IR spaceborne information for investigation of thermal state of Lake Ladoga, Issledovanie Zemli iz Kosmosa (Earth Research from Space), 1993, N 4, p.69-78 (in Russian)
5. J.Malm and L.Jnsson, A study of the thermal bar in Lake Ladoga using water surface temperature data from satellite images, Remote Sens.Environ.,1993, v.44, p.35-46
6. A.Yu.Ivanov, V.P.Nefed'ev, A.V.Smirnov, V.S.Etkin. The analysis of the mesoscale fronts dynamics using data on the ocean remote sensing in microwave range. Izv.Ac.Sc.USSR,FAO,1986, v.22, N 4, p.440-447 (in Russian).
7. L.M.Mitnik and V.B.Lobanov. Reflection of the oceanic fronts on the satellite radar images. In: Oceanography of Asian Marginal Seas, Amsterdam, Elsevier, 1991. Elsevier Oceanography Series v.54, p.85-101
8. R.C.Beal, V.Kudryavtsev, D.Thompson et al. Large and small circulation signatures of the ERS-1 SAR over Gulf Stream. Proc. 2nd ERS-1 Symposium, Hamburg, Germany, 11-14 Oct. 1993, p.547-552
9. W.C.Keller, V.Wismann, W.Alpers. Tower-based measurements of the ocean C-band radar backscattering cross section. J. Geophys. Res., 1989, v.94, N C1, p.924-930.
10. C.-T.Liu, L.M.Mitnik, M.-K.Hsu, Y.Yang Oceanic phenomena northeast of Taiwan from Almaz SAR image Terrestrial, Atmospheric and Oceanic Sciences, 1994, v.5, N 4, p.557-571
11. S.G.Karetnikov and M.A.Naumenko. Combined satellite SAR and IR data application for a determination of skin effects on Lake Ladoga. Proc. EUSAR'96 (European Conference on Synthetic Aperture Radar), Knigswinter, Germany, 26-28 March 1996, p.645-647
12. F.M.Boyce and K.P.B.Thomson, Internal wave trains of finite amplitude in Lake Ontario, Dep.Environ.Inl.Waters Dir.Sci.Series, 1973, N 27
13. A.Ivanov, V. Zaitsev, M. Naumenko, S. Karetnikov. Study of the thermal bar and other hydrological phenomena in Lake Ladoga using ALMAZ-1 and ERS-1 SAR imagery. Proc. IGARSS'95, v.3, Florence, Italy, July 10-14 1995, p.1985-1987
14. K.Ya.Kondrat'ev, M.A.Naumenko, P.A.Shirokov et al. Experience in an application of Almaz-1 SAR data for the large lakes investigations, Doklady Akademii Nauk, 1995, v.304, N 3, p.396-399 (in Russian).
15. N.V.Kochkov. Formation, evolution and dissipation of internal waves in lakes. Ph.D. Thesis, Leningrad, 1989 (in Russian)
16. A.V.Dikinis, A.Yu.Ivanov, I.G.Mal'tseva et al. Interpretation of the ALMAZ-1 SAR imagery of internal waves using hydrometeorological and hydrographic data. Issledovanie Zemli iz Kosmosa (Earth Research from Space), 1996, N 5, p.47-59 (in Russian)
    1. M.A.Donelan and W.J.Pierson. Radar scattering and equilibrium ranges in wind-generated waves with application to scatterometry. J.Geophys. Res., 1987, v.92, p.4971-5029
17. M.P.Timofeev. On the transformation of wind speed over lakes. Proceedings GGO, 1963, v.95, p.33-40 (in Russian)
18. M.A.Naumenko, D.V.Beletsky, V.B.Rumyantsev et al. Investigations of the hydrobiological situation in Lake Onega using joint spaceborne, airborne radars and in situ measurements. Int. J. Rem. Sens., 1994, v.15, N 10, p.2039-2049

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