Sea surface brightness temperature at L-band - IEEE Xplore
Sea Surface Brightness Temperature At L-band: Impact Of Surface Currents A. Camps(1), M. Vall-llossera(1), J. Miranda(1), J. Font(2) (1)
Dept. of Signal Theory and Communications, Universitat Politècnica de Catalunya Campus Nord, D4-016, E-08034 Barcelona, Spain Tel. +34+934016085, Fax. +34+934017232, e-mail: [email protected] (2) Institut de Ciencies del Mar, CMIMA - CSIC Dept. Marine Geology and Physical Oceanography Passeig Maritim, 37-49, E- 08003 Barcelona, Spain
Abstract Drastically different sea states result for a growing, decaying, or a fully-developed sea due to the presence of a wind field [1]. In addition, the sea state is altered by the superposition of oceanic currents on the wave field created by the wind. This study presents a numerical analysis of the change in the sea surface brightness temperature at L-band computed using the SSA model and a modified Kitaigorodskii-Pierson-Moskowitz sea surface spectrum to account for the presence of surface water currents [2]. Their impact on the brightness temperature variations is evaluated and compared to the brightness temperature fluctuations observed from a tower-based field experiment [3,4]. The consequences for salinity retrievals from space using L-band radiometry are discussed.
oceanographic and atmospheric data (fig. 1). LAURA performed incidence angle scans at different azimuth angles. The flat surface brightness temperature contribution (eqn. 1) was subtracted from the calibrated brightness temperatures to obtain ∆TB wind , p .
I. INTRODUCTION Sea surface salinity can be derived from L-band radiometric measurements [5]. At this frequency the sensitivity of the brightness temperature (TB) to the sea surface salinity (SSS) is low: 0.5 K/psu for a sea surface temperature (SST) of 20°C, decreasing to 0.25 K/psu for a SST of 0°C. Since other variables than SSS influence the TB signal (sea surface temperature, surface roughness and foam), the accuracy of the SSS measurement will degrade unless these effects are properly accounted for. The main objective of the ESAsponsored WInd and Salinity Experiment (WISE 2000 and WISE 2001) field experiments [3,4] was the improvement of our understanding of the sea state effects on TB at different incidence angles and polarizations. The brightness temperature sensitivity to the 10 m height wind speed U10 ( ∆TB wind , p ) was computed from the flat surface emissivity model:
TB , p (θ i , SST , SSS ,U10 ) = = TB Fresnel , p (θi , SST , SSS ) + ∆TB wind , p (θ i , U10 ) ,
(1)
where
TB Fresnel , p (θ i , SST , SSS ) = SST ⋅ e p (θ i , SST , SSS )
(2)
is the brightness temperature of a flat sea surface, and e p (θ i , SST , SSS ) = 1 − Γ p (θ i , SST , SSS )
2
(3)
is the emissivity computed from the Fresnel field reflection coefficient at p-polarization using the Klein and Swift model [6]. The UPC L-band AUtomatic RAdiometer (LAURA) was deployed at the Casablanca oil rig, with a number of other instruments and buoys to acquire 0-7803-8742-2/04/$20.00 (C) 2004 IEEE
Fig. 1. Instrument distribution and observation geometry (WISE).
The linear regression of the ∆TB wind , p points vs. U10 at each incidence angle and polarization was obtained (fig. 2), and the slope of this linear regression is the sensitivity of the brightness temperature to wind speed. The scatter of the data points in fig. 2 has been a matter of discussion, and several possible sources have been pointed out including: errors associated to the measurement of the wind speed itself and its conversion from 2.6 (buoy anemometer) and 70 m (tower anemometer) height to reference 10 m height, the presence of swell [1] ( TBwind + swell − TBwind from -0.12 K to +0.25K at V-pol, and from -0.2 K to + 0.1 K at H-pol, at 60∞ incidence angle), the azimuthal dependence induced by wave reflections and its interference with the incident ones (changing the rms slope and the associated amount of foam) [4], and an imperfect correction for the galactic noise (~±0.5 K) due to differences in the different surveys and measurements performed by sky-looking radiometers [7].
However, detailed re-analyses [1] have shown that it is the sea state and not the wind speed who plays a major role in the L-band brightness temperature. In this study we analyze the impact of surface currents in the sea surface height spectrum, and compute the variation of the modeled brightness temperature when currents are present and when they are not. This was a non-negligible effect in WISE, since the direction of the dominant winds was North West, but there were intense storms with winds coming from the East, and then the relative orientation between the wind and the currents varied (fig. 1). 3481
Authorized licensed use limited to: CSIC. Downloaded on April 08,2010 at 07:13:31 UTC from IEEE Xplore. Restrictions apply.
V-pol (red) and H-pol (blue) 0.7
0.5
0.4 0.3
B
∆T /∆U
10
[K/(m/s)]
0.6
0.2 0.1 0
Fig. 2. Derivation of the brightness temperature sensitivity to wind speed: ∆TB wind , h (top) and ∆TB wind , v (center) scatter plots, linear fit (solid line) and percentile 50% (dashed lines) as a function of wind speed for incidence angles from 25° to 65° (left to right). Bottom: derived TB sensitivity to wind speed as a function of polarization and incidence angle (solid line), associated ±1σ error bars, and linear fit (dashed lines). All WISE 2001 data points used.
II. SEA SURFACE SPECTRUM MODELING: EFFECTS OF WIND AND SURFACE’S CURRENTS Many spectra are found in the literature to mathematically describe the sea surface height. On the other hand, several numerical methods exist to compute the scattering coefficients, and from them, the surface’s emissivity using Peake’s relationship [8]. In [9] Vall-llossera et al. performed a comprehensive study of the range of validity of the different methods, comparing the computed sensitivity of the L-band brightness temperature vs. wind speed. They found that the 2-scale model and the small slope approximation (SSA), with the Durden-Vesecky spectrum multiplied by two provided the best agreement to the WISE-derived data (fig. 2). However, it is difficult to modify this spectrum to include the effect of surface’s currents. In [2], Huang et al. derived a closed-form expression for the Kitaigorodskii-Pierson-Moskowitz sea surface height omnidirectional spectrum to include the effect of surface’s currents, shown in eqn. (4) in a slightly modified form: S ( K )) =
a 1 K 3 U current 1 + c
2 0.74 g − , exp 7 4 U current 2 4 K U10 1 + c -3
10
20
30
40
50
60
Incidence angle [deg]
Fig. 3. Brightness temperature sensitivity to wind speed computed for the Kitaigorodskii-Pierson-Moskowitz sea surface height spectrum, U10 from 0 to 5 m/s, and Ucurrent = 0 m/s. As compared to other spectra and measured data, this spectrum overestimates the sensitivity at H-pol and underestimates it at V-pol at large incidence angles.
Figure 4 shows the sea surface height spectrum at two wind speeds U10: a) 5 m/s, and b) 10 m/s, for different values of the surface’s current Ucurrent = +3 m /s (same direction as the wind), 0 m/s (no current), and -3 m/s (direction contrary to the wind). As it can be appreciated, a surface’s current in the same direction as the wind, decreases the amplitude of the spectrum (less roughness), while a surface’s current against the direction of the wind increases it, as experience corroborates intuitively at the river mouths. This effect starts being noticeably for K≥20-30 m-1, corresponding to l≥20-30 cm (l0 = 21 cm at 1.413 MHz). It should be pointed out that since the dispersion relationship is invalid for U current < − c 2 = −1 2 ⋅ g K , the spectrum will have a cutoff at this point. Therefore the higher the sea wave-number K, which depends on the electromagnetic frequency at which the current effects have to be analyzed, the lower the |Ucurrent| that can be modeled. The authors of [2] also warn that the model may not be accurate enough when the surface’s current is against the wind. For this reason, we have limited ourselves to surface current speeds higher than Ucurrent ≥ -0.1 m/s. However, intuitively, if Ucurrent = Uwind at the surface the sea surface should be flat.
(4)
where K is the wavenumber, a=4.05 10 , c is the phase speed, and U10 and Ucurrent are the 10 m height wind and current speed. Equation (4) is very convenient for numerical integration, and despite the trend and the predicted values of the sensitivity to wind speed at low incidence angles is correct, at large incidence angles they are higher than the measured sensitivities by ~0.15 K/(m/s) (figs. 2c and 3).
0-7803-8742-2/04/$20.00 (C) 2004 IEEE
0
Figure 5 shows the sea surface slope spectrum for the same wind speeds and surface’s current. As it can be appreciated, surface’s currents significantly modify the surface’s slope spectrum, especially at high wave-numbers (shorter wavelengths). Since the brightness temperatures are sensitive to the sea surface slopes [4], the change in the sea surface slope spectrum produces a change in the brightness temperatures.
3482
Authorized licensed use limited to: CSIC. Downloaded on April 08,2010 at 07:13:31 UTC from IEEE Xplore. Restrictions apply.
U10= 5 m/s
U10=5 m/s
0
10 0
10
-1
10 -2
10
-2
1
k S [rad m ]
-4
10
-3
10
2
3
S [rad m ]
10
-6
10
-4
10
-8
10
-5
10
-10
10
-6
-2
-1
10
10
a)
0
10 k [m -1]
1
10
10 -2 10
2
10
-1
0
10
10
a)
2
1
10
10
k [m -1]
U10=10 m/s
U10=10 m/s
0
10 0
10
-1
10 -2
10
-2
1
k S [rad m ]
-4
10
-3
10
2
3
S [rad m ]
10
-6
10
-4
10 -8
10
-5
10
-10
10
-6
-2
-1
10
10
b)
0
10 k [m -1]
1
10
Fig. 4. a) Sea surface height spectra at a) U10 = 5 m/s, and b) U10 = 10 m/s, for Ucurrent = 3 m/s (dashed line), Ucurrent = 0 m/s (solid line), and Ucurrent = -3 m/s (dotted line).
III.
10 -2 10
2
10
b)
1
2
10
10
Fig. 5. a) Sea surface slope spectra at a) U10 = 5 m/s, and b) U10 = 10 m/s, for Ucurrent = 3 m/s (dashed line), Ucurrent = 0 m/s (solid line), and Ucurrent = -3 m/s (dotted line).
V-pol (red) and H-pol (blue); U10 = 5 m/s (dashed line) and 10 m/s (solid line)
0.1
The impact of a surface current in the brightness temperature has been computed using eqn. (4) and the SSA model implemented in [8], and it has been computed as the difference between TB , p (θ i , SST , SSS ,U10 , U current ) and
0-7803-8742-2/04/$20.00 (C) 2004 IEEE
10
k [m -1]
SIMULATION RESULTS
Uc = -0.1 m/s
0
-0.1 -0.2
-0.3 Uc = + 3 m/s
-0.4
10
c
B
10
c
T (θ , U , U ) - T (θ , U , U = 0 m/s)
-0.5
B
TB , p (θ i , SST , SSS , U10 , U current = 0 m / s ) . Figure 6 shows the variation in the brightness temperature due to a surface’s current of Ucurrent = - 0.1 m/s and Ucurrent = +3 m/s, for SSS = 35 psu, SST = 15ºC , and U10 = 5 m/s (dashed line) and U10 = 10 m/s (solid line). As it can be appreciated, the presence of strong currents can induce a variation on the brightness temperatures at nadir up to -0.4 K, and at 60∞ incidence angle up to -0.25 K at V-polarization, and -0.6 K at H-polarization, which is not negligible at all when compared to the highest brightness sensitivity to SSS (0.5 K/psu at SST = 20∞C and 0.25 K/psu at SST = 0∞C). The impact at these two wind speeds is very similar.
0
-1
10
-0.6 -0.7 0
10
20 30 40 Incidence angle [deg]
50
60
Fig. 6. Brightness temperature variation due to a surface current of Ucurrent=- 0.1 m/s and Ucurrent=+3 m/s, for SSS=35 psu, SST=15ºC, and U10=5 m/s (dashed line) and U10=10 m/s (solid line).
3483
Authorized licensed use limited to: CSIC. Downloaded on April 08,2010 at 07:13:31 UTC from IEEE Xplore. Restrictions apply.
IV. CONCLUSIONS
REFERENCES
Figure 7 shows the current speed and direction measured from the Aanderaa Instruments Current velocity sensor that hang from the torch of the platform at 2 meters below sea level during WISE 2000. Most of the time the current was parallel to the coast (ϕ=200∞ from South, clockwise) and reached peak values of ~0.5 m/s. This means that in WISE 2000, the brightness temperature variation associated to the presence of the current is estimated to be < 0.1 K at Hpolarization, and < 0.07 K at V-polarization. This error source by itself may account for a salinity retrieval error of 1 psu/0.5 K . 0.1 K ª 0.2 psu, unless the retrieval algorithm leaves the wind as a free parameter and an effective wind speed derived (Ueff) accounts for the current-induced sea surface roughness [4]. In other locations where different water masses meet, for example at river estuaries etc., or where there are strong currents, this effect should be even more noticeably. 0.7
Current Speed [m/s]
0.6
0.5 0.4 0.3 0.2 0.1
0 330
340
350
360 5 DoY [2000-2001]
15
25
340
350
5 360 DoY [2000-2001]
15
25
a) 400
Current direction [deg]
350 300 250 200
150
100 50
0 330
b)
Fig. 7. a) Current speed and b) direction during the WISE 2000 field experiment.
ACKNOWLEDGEMENTS This work has been supported by grants MCYT TIC200204451-C02-01 and Plan Nacional del Espacio ESP200211648-E of the Spanish Government. 0-7803-8742-2/04/$20.00 (C) 2004 IEEE
[1] J. Miranda, M. Vall-llossera, A. Camps, N. Duffo, J. Etcheto, “Sea State Effect on the Sea Surface Emissivity at L-Band,” IEEE Transactions on Geoscience and Remote Sensing Vol. 41, No 10, pp. 2307-2315, October 2003. [2] N. E. Huang, D. T. Chen, C. C. Tung, and J. R. Smith, “Interactions between Steady Non-Uniform Currents and Gravity Waves with Applications for Current Measurements,” Journal of Physical Oceanography Oceanography, Vol. 2, pp. 420-431, October 1972. [3] A. Camps, J. Font, J. Etchetto, V. Caselles, A. Weill, I. Corbella, M. Vall-llossera, N. Duffo, F. Torres, R. Villarino, L. Enrique, A. Julià, C. Gabarró, J. Boutin, E. Rubio, S.C. Reising, P. Wursteisen, M. Berger,y M. Martín-Neira, “Sea Surface Emissivity Observations at L-band: First Results of the Wind and Salinity Experiment WISE-2000,” IEEE Transactions on Geoscience and Remote Sensing, Vol. GRS-40, No 10, pp. 2117-2130, October 2002 [4] A. Camps, J. Font, M. Vall-llossera, C. Gabarró, I. Corbella, N. Duffo, F. Torres, S. Blanch, A. Aguasca, R. Villarino, L. Enrique, J. Miranda, J. Arenas, A. Julià, J. Etcheto, V. Caselles, A. Weill, J. Boutin, S. Contardo, R. Niclós, R. Rivas, S.C.Reising, P. Wursteisen, M. Berger, and M. Martín-Neira, “The WISE 2000 and 2001 campaigns in support of the SMOS Mission: Sea Surface L-Band Brightness Temperature Observations, And Their Application to Multi-Angular Salinity Retrieval,” IEEE Transactions on Geoscience and Remote Sensing, Vol 42 (4), pp. 804-823, April 2004. [5] C.T Swift and R.E. McIntosh, "Considerations fro microwave remote sensing of Ocena-Surface salinity", IEEE Transactions on Geoscience Electronics, Vol. GE21, No. 4, pp. 480- 491, Oct. 1983 [6] L. A. Klein and Swift, C.T. 'An Improved Model for the Dielectric Constant of Sea Water at Microwave Frequencies', IEEE Journal of Oceanic Engineering., OE-Vol 2, No 1, pp 104-111, 1977 [7] J. Etcheto, E. Dinnat, J. Boutin, A. Camps, J. Miller, S. Contardo, J. Wesson, J. Font, D.Long, “Wind speed effect on L-band brightness temperature inferred from EuroSTARRS and WISE 2001 field experiments,” IEEE Trans. on Geoscience and Remote Sensing (in press). [8] W.H. Peake, “Interaction of Electromagnetic Waves with some Natural Surfaces”, IRE Trans. on Antennas and Propagation (special supplement), Vol. 7, pp. S324-S329, 1959. [9] M. Vall-llossera, J. Miranda, A. Camps, R. Villarino, “Sea Surface Emissivity Modelling at L-Band: An InterComparison Study,” ESA SP-525, pp. 143-153, May 2003.
3484
Authorized licensed use limited to: CSIC. Downloaded on April 08,2010 at 07:13:31 UTC from IEEE Xplore. Restrictions apply.
Dept. of Signal Theory and Communications, Universitat Politècnica de Catalunya Campus Nord, D4-016, E-08034 Barcelona, Spain Tel. +34+934016085, Fax. +34+934017232, e-mail: [email protected] (2) Institut de Ciencies del Mar, CMIMA - CSIC Dept. Marine Geology and Physical Oceanography Passeig Maritim, 37-49, E- 08003 Barcelona, Spain
Abstract Drastically different sea states result for a growing, decaying, or a fully-developed sea due to the presence of a wind field [1]. In addition, the sea state is altered by the superposition of oceanic currents on the wave field created by the wind. This study presents a numerical analysis of the change in the sea surface brightness temperature at L-band computed using the SSA model and a modified Kitaigorodskii-Pierson-Moskowitz sea surface spectrum to account for the presence of surface water currents [2]. Their impact on the brightness temperature variations is evaluated and compared to the brightness temperature fluctuations observed from a tower-based field experiment [3,4]. The consequences for salinity retrievals from space using L-band radiometry are discussed.
oceanographic and atmospheric data (fig. 1). LAURA performed incidence angle scans at different azimuth angles. The flat surface brightness temperature contribution (eqn. 1) was subtracted from the calibrated brightness temperatures to obtain ∆TB wind , p .
I. INTRODUCTION Sea surface salinity can be derived from L-band radiometric measurements [5]. At this frequency the sensitivity of the brightness temperature (TB) to the sea surface salinity (SSS) is low: 0.5 K/psu for a sea surface temperature (SST) of 20°C, decreasing to 0.25 K/psu for a SST of 0°C. Since other variables than SSS influence the TB signal (sea surface temperature, surface roughness and foam), the accuracy of the SSS measurement will degrade unless these effects are properly accounted for. The main objective of the ESAsponsored WInd and Salinity Experiment (WISE 2000 and WISE 2001) field experiments [3,4] was the improvement of our understanding of the sea state effects on TB at different incidence angles and polarizations. The brightness temperature sensitivity to the 10 m height wind speed U10 ( ∆TB wind , p ) was computed from the flat surface emissivity model:
TB , p (θ i , SST , SSS ,U10 ) = = TB Fresnel , p (θi , SST , SSS ) + ∆TB wind , p (θ i , U10 ) ,
(1)
where
TB Fresnel , p (θ i , SST , SSS ) = SST ⋅ e p (θ i , SST , SSS )
(2)
is the brightness temperature of a flat sea surface, and e p (θ i , SST , SSS ) = 1 − Γ p (θ i , SST , SSS )
2
(3)
is the emissivity computed from the Fresnel field reflection coefficient at p-polarization using the Klein and Swift model [6]. The UPC L-band AUtomatic RAdiometer (LAURA) was deployed at the Casablanca oil rig, with a number of other instruments and buoys to acquire 0-7803-8742-2/04/$20.00 (C) 2004 IEEE
Fig. 1. Instrument distribution and observation geometry (WISE).
The linear regression of the ∆TB wind , p points vs. U10 at each incidence angle and polarization was obtained (fig. 2), and the slope of this linear regression is the sensitivity of the brightness temperature to wind speed. The scatter of the data points in fig. 2 has been a matter of discussion, and several possible sources have been pointed out including: errors associated to the measurement of the wind speed itself and its conversion from 2.6 (buoy anemometer) and 70 m (tower anemometer) height to reference 10 m height, the presence of swell [1] ( TBwind + swell − TBwind from -0.12 K to +0.25K at V-pol, and from -0.2 K to + 0.1 K at H-pol, at 60∞ incidence angle), the azimuthal dependence induced by wave reflections and its interference with the incident ones (changing the rms slope and the associated amount of foam) [4], and an imperfect correction for the galactic noise (~±0.5 K) due to differences in the different surveys and measurements performed by sky-looking radiometers [7].
However, detailed re-analyses [1] have shown that it is the sea state and not the wind speed who plays a major role in the L-band brightness temperature. In this study we analyze the impact of surface currents in the sea surface height spectrum, and compute the variation of the modeled brightness temperature when currents are present and when they are not. This was a non-negligible effect in WISE, since the direction of the dominant winds was North West, but there were intense storms with winds coming from the East, and then the relative orientation between the wind and the currents varied (fig. 1). 3481
Authorized licensed use limited to: CSIC. Downloaded on April 08,2010 at 07:13:31 UTC from IEEE Xplore. Restrictions apply.
V-pol (red) and H-pol (blue) 0.7
0.5
0.4 0.3
B
∆T /∆U
10
[K/(m/s)]
0.6
0.2 0.1 0
Fig. 2. Derivation of the brightness temperature sensitivity to wind speed: ∆TB wind , h (top) and ∆TB wind , v (center) scatter plots, linear fit (solid line) and percentile 50% (dashed lines) as a function of wind speed for incidence angles from 25° to 65° (left to right). Bottom: derived TB sensitivity to wind speed as a function of polarization and incidence angle (solid line), associated ±1σ error bars, and linear fit (dashed lines). All WISE 2001 data points used.
II. SEA SURFACE SPECTRUM MODELING: EFFECTS OF WIND AND SURFACE’S CURRENTS Many spectra are found in the literature to mathematically describe the sea surface height. On the other hand, several numerical methods exist to compute the scattering coefficients, and from them, the surface’s emissivity using Peake’s relationship [8]. In [9] Vall-llossera et al. performed a comprehensive study of the range of validity of the different methods, comparing the computed sensitivity of the L-band brightness temperature vs. wind speed. They found that the 2-scale model and the small slope approximation (SSA), with the Durden-Vesecky spectrum multiplied by two provided the best agreement to the WISE-derived data (fig. 2). However, it is difficult to modify this spectrum to include the effect of surface’s currents. In [2], Huang et al. derived a closed-form expression for the Kitaigorodskii-Pierson-Moskowitz sea surface height omnidirectional spectrum to include the effect of surface’s currents, shown in eqn. (4) in a slightly modified form: S ( K )) =
a 1 K 3 U current 1 + c
2 0.74 g − , exp 7 4 U current 2 4 K U10 1 + c -3
10
20
30
40
50
60
Incidence angle [deg]
Fig. 3. Brightness temperature sensitivity to wind speed computed for the Kitaigorodskii-Pierson-Moskowitz sea surface height spectrum, U10 from 0 to 5 m/s, and Ucurrent = 0 m/s. As compared to other spectra and measured data, this spectrum overestimates the sensitivity at H-pol and underestimates it at V-pol at large incidence angles.
Figure 4 shows the sea surface height spectrum at two wind speeds U10: a) 5 m/s, and b) 10 m/s, for different values of the surface’s current Ucurrent = +3 m /s (same direction as the wind), 0 m/s (no current), and -3 m/s (direction contrary to the wind). As it can be appreciated, a surface’s current in the same direction as the wind, decreases the amplitude of the spectrum (less roughness), while a surface’s current against the direction of the wind increases it, as experience corroborates intuitively at the river mouths. This effect starts being noticeably for K≥20-30 m-1, corresponding to l≥20-30 cm (l0 = 21 cm at 1.413 MHz). It should be pointed out that since the dispersion relationship is invalid for U current < − c 2 = −1 2 ⋅ g K , the spectrum will have a cutoff at this point. Therefore the higher the sea wave-number K, which depends on the electromagnetic frequency at which the current effects have to be analyzed, the lower the |Ucurrent| that can be modeled. The authors of [2] also warn that the model may not be accurate enough when the surface’s current is against the wind. For this reason, we have limited ourselves to surface current speeds higher than Ucurrent ≥ -0.1 m/s. However, intuitively, if Ucurrent = Uwind at the surface the sea surface should be flat.
(4)
where K is the wavenumber, a=4.05 10 , c is the phase speed, and U10 and Ucurrent are the 10 m height wind and current speed. Equation (4) is very convenient for numerical integration, and despite the trend and the predicted values of the sensitivity to wind speed at low incidence angles is correct, at large incidence angles they are higher than the measured sensitivities by ~0.15 K/(m/s) (figs. 2c and 3).
0-7803-8742-2/04/$20.00 (C) 2004 IEEE
0
Figure 5 shows the sea surface slope spectrum for the same wind speeds and surface’s current. As it can be appreciated, surface’s currents significantly modify the surface’s slope spectrum, especially at high wave-numbers (shorter wavelengths). Since the brightness temperatures are sensitive to the sea surface slopes [4], the change in the sea surface slope spectrum produces a change in the brightness temperatures.
3482
Authorized licensed use limited to: CSIC. Downloaded on April 08,2010 at 07:13:31 UTC from IEEE Xplore. Restrictions apply.
U10= 5 m/s
U10=5 m/s
0
10 0
10
-1
10 -2
10
-2
1
k S [rad m ]
-4
10
-3
10
2
3
S [rad m ]
10
-6
10
-4
10
-8
10
-5
10
-10
10
-6
-2
-1
10
10
a)
0
10 k [m -1]
1
10
10 -2 10
2
10
-1
0
10
10
a)
2
1
10
10
k [m -1]
U10=10 m/s
U10=10 m/s
0
10 0
10
-1
10 -2
10
-2
1
k S [rad m ]
-4
10
-3
10
2
3
S [rad m ]
10
-6
10
-4
10 -8
10
-5
10
-10
10
-6
-2
-1
10
10
b)
0
10 k [m -1]
1
10
Fig. 4. a) Sea surface height spectra at a) U10 = 5 m/s, and b) U10 = 10 m/s, for Ucurrent = 3 m/s (dashed line), Ucurrent = 0 m/s (solid line), and Ucurrent = -3 m/s (dotted line).
III.
10 -2 10
2
10
b)
1
2
10
10
Fig. 5. a) Sea surface slope spectra at a) U10 = 5 m/s, and b) U10 = 10 m/s, for Ucurrent = 3 m/s (dashed line), Ucurrent = 0 m/s (solid line), and Ucurrent = -3 m/s (dotted line).
V-pol (red) and H-pol (blue); U10 = 5 m/s (dashed line) and 10 m/s (solid line)
0.1
The impact of a surface current in the brightness temperature has been computed using eqn. (4) and the SSA model implemented in [8], and it has been computed as the difference between TB , p (θ i , SST , SSS ,U10 , U current ) and
0-7803-8742-2/04/$20.00 (C) 2004 IEEE
10
k [m -1]
SIMULATION RESULTS
Uc = -0.1 m/s
0
-0.1 -0.2
-0.3 Uc = + 3 m/s
-0.4
10
c
B
10
c
T (θ , U , U ) - T (θ , U , U = 0 m/s)
-0.5
B
TB , p (θ i , SST , SSS , U10 , U current = 0 m / s ) . Figure 6 shows the variation in the brightness temperature due to a surface’s current of Ucurrent = - 0.1 m/s and Ucurrent = +3 m/s, for SSS = 35 psu, SST = 15ºC , and U10 = 5 m/s (dashed line) and U10 = 10 m/s (solid line). As it can be appreciated, the presence of strong currents can induce a variation on the brightness temperatures at nadir up to -0.4 K, and at 60∞ incidence angle up to -0.25 K at V-polarization, and -0.6 K at H-polarization, which is not negligible at all when compared to the highest brightness sensitivity to SSS (0.5 K/psu at SST = 20∞C and 0.25 K/psu at SST = 0∞C). The impact at these two wind speeds is very similar.
0
-1
10
-0.6 -0.7 0
10
20 30 40 Incidence angle [deg]
50
60
Fig. 6. Brightness temperature variation due to a surface current of Ucurrent=- 0.1 m/s and Ucurrent=+3 m/s, for SSS=35 psu, SST=15ºC, and U10=5 m/s (dashed line) and U10=10 m/s (solid line).
3483
Authorized licensed use limited to: CSIC. Downloaded on April 08,2010 at 07:13:31 UTC from IEEE Xplore. Restrictions apply.
IV. CONCLUSIONS
REFERENCES
Figure 7 shows the current speed and direction measured from the Aanderaa Instruments Current velocity sensor that hang from the torch of the platform at 2 meters below sea level during WISE 2000. Most of the time the current was parallel to the coast (ϕ=200∞ from South, clockwise) and reached peak values of ~0.5 m/s. This means that in WISE 2000, the brightness temperature variation associated to the presence of the current is estimated to be < 0.1 K at Hpolarization, and < 0.07 K at V-polarization. This error source by itself may account for a salinity retrieval error of 1 psu/0.5 K . 0.1 K ª 0.2 psu, unless the retrieval algorithm leaves the wind as a free parameter and an effective wind speed derived (Ueff) accounts for the current-induced sea surface roughness [4]. In other locations where different water masses meet, for example at river estuaries etc., or where there are strong currents, this effect should be even more noticeably. 0.7
Current Speed [m/s]
0.6
0.5 0.4 0.3 0.2 0.1
0 330
340
350
360 5 DoY [2000-2001]
15
25
340
350
5 360 DoY [2000-2001]
15
25
a) 400
Current direction [deg]
350 300 250 200
150
100 50
0 330
b)
Fig. 7. a) Current speed and b) direction during the WISE 2000 field experiment.
ACKNOWLEDGEMENTS This work has been supported by grants MCYT TIC200204451-C02-01 and Plan Nacional del Espacio ESP200211648-E of the Spanish Government. 0-7803-8742-2/04/$20.00 (C) 2004 IEEE
[1] J. Miranda, M. Vall-llossera, A. Camps, N. Duffo, J. Etcheto, “Sea State Effect on the Sea Surface Emissivity at L-Band,” IEEE Transactions on Geoscience and Remote Sensing Vol. 41, No 10, pp. 2307-2315, October 2003. [2] N. E. Huang, D. T. Chen, C. C. Tung, and J. R. Smith, “Interactions between Steady Non-Uniform Currents and Gravity Waves with Applications for Current Measurements,” Journal of Physical Oceanography Oceanography, Vol. 2, pp. 420-431, October 1972. [3] A. Camps, J. Font, J. Etchetto, V. Caselles, A. Weill, I. Corbella, M. Vall-llossera, N. Duffo, F. Torres, R. Villarino, L. Enrique, A. Julià, C. Gabarró, J. Boutin, E. Rubio, S.C. Reising, P. Wursteisen, M. Berger,y M. Martín-Neira, “Sea Surface Emissivity Observations at L-band: First Results of the Wind and Salinity Experiment WISE-2000,” IEEE Transactions on Geoscience and Remote Sensing, Vol. GRS-40, No 10, pp. 2117-2130, October 2002 [4] A. Camps, J. Font, M. Vall-llossera, C. Gabarró, I. Corbella, N. Duffo, F. Torres, S. Blanch, A. Aguasca, R. Villarino, L. Enrique, J. Miranda, J. Arenas, A. Julià, J. Etcheto, V. Caselles, A. Weill, J. Boutin, S. Contardo, R. Niclós, R. Rivas, S.C.Reising, P. Wursteisen, M. Berger, and M. Martín-Neira, “The WISE 2000 and 2001 campaigns in support of the SMOS Mission: Sea Surface L-Band Brightness Temperature Observations, And Their Application to Multi-Angular Salinity Retrieval,” IEEE Transactions on Geoscience and Remote Sensing, Vol 42 (4), pp. 804-823, April 2004. [5] C.T Swift and R.E. McIntosh, "Considerations fro microwave remote sensing of Ocena-Surface salinity", IEEE Transactions on Geoscience Electronics, Vol. GE21, No. 4, pp. 480- 491, Oct. 1983 [6] L. A. Klein and Swift, C.T. 'An Improved Model for the Dielectric Constant of Sea Water at Microwave Frequencies', IEEE Journal of Oceanic Engineering., OE-Vol 2, No 1, pp 104-111, 1977 [7] J. Etcheto, E. Dinnat, J. Boutin, A. Camps, J. Miller, S. Contardo, J. Wesson, J. Font, D.Long, “Wind speed effect on L-band brightness temperature inferred from EuroSTARRS and WISE 2001 field experiments,” IEEE Trans. on Geoscience and Remote Sensing (in press). [8] W.H. Peake, “Interaction of Electromagnetic Waves with some Natural Surfaces”, IRE Trans. on Antennas and Propagation (special supplement), Vol. 7, pp. S324-S329, 1959. [9] M. Vall-llossera, J. Miranda, A. Camps, R. Villarino, “Sea Surface Emissivity Modelling at L-Band: An InterComparison Study,” ESA SP-525, pp. 143-153, May 2003.
3484
Authorized licensed use limited to: CSIC. Downloaded on April 08,2010 at 07:13:31 UTC from IEEE Xplore. Restrictions apply.
