2013 Microwave Ocean Remote Sensing Training School Sea Surface Temperature Jordi Isern-Fontanet Institut Català de Ciències del Clima Motivation • SST is an important indicator of the state of a climate system • It is also a key variable for establishing atmosphere and ocean boundary conditions • SST affects atmosphere-ocean coupling, which includes water circulation and heat and gas exchanges. • SST can be used to derive ocean currents. Highly accurate SST data are required for climate monitoring, for atmospheric and oceanic simulations and operational applications. 2 of 38 Large scale patterns of SST 3 of 38 ENSO • Coupled ocean-atmosphere phenomenon. • Peridicity: 2 to 7 years. • Recent episodes: 1997-1998, 2002-2003, 2006-2007, ... 4 of 38 Mesoscale variability of SST: microwave observations 5 of 38 Submesoscale variability of SST: infrared observations 6 of 38 Ocean time and scales 7 of 38 The thermocline structure of the ocean • Mixed Layer Depth has a strong seasonal signature ◦ • SST is representative of different depths depending on the period of the year. The vertical structure also varies with latitude. 8 of 38 Spatial variability of the Mixed Layer Isern-Fontanet et al. JGR 2008 9 of 38 The diurnal thermocline Robinson 2004 Price et al. JGR 1986 10 of 38 Foundation temperature • It is defined as the temperature at the first time of the day when the heat gain from the solar radiation absorption exceeds the heat loss at the sea surface 11 of 38 The thermal skin layer 12 of 38 Heat fluxes 30 < QS < 260 Wm−2 −60 < QIR − QG < −30 Wm−2 −42 < QH < −2 Wm−2 −130 < QL < −10 Wm−2 13 of 38 Remote sensing of the oceans Robinson 2004 14 of 38 Blackbody radiation • Spectral emitance at wavelength λ emitted by body in thermodynamic equilibrium at temperature T (Planck’s law) 2hc2 1 Mλ(T ) = 5 λ exp hc − 1 k λT (1) B ◦ Planck constant h = 6.62606957(29) × 10−34Js ◦ Boltzmann constant kB = 1.3806488(13) × 10−23JK−1 ◦ speed of light c = 299, 792, 458ms−1 Robinson 2009 15 of 38 Importance of clouds Chelton and Wentz, BAMS, 2005 16 of 38 Approximations for the microwave band • Spectral emitance at frequency f emitted by body in thermodynamic equilibrium at temperature T (Planck’s law) 2hf 3 1 B̂f (T ) = 2 c exp hf − 1 k T (2) B • At microwave frequencies hf exp kB T • hf ≈1+ kB T (3) Rayleigh-Jeans approcimation: 2kB f 2T B̂f (T ) ≈ c2 (4) The emitted radiance is directly proportional to the temperature of the emitting surface. 17 of 38 Emissivity • Seawater is not a black-body and Planck’s law has to be corrected introducing the emissivity ε 2kB f 2T Bf (T ) = εB̂f (T ) ≈ ε c2 (5) • The temperature retrieved without taking into account ε is known as Brightness Temperature TB • It is necessary to know the emissivity to retrieve the temperature of the emitter, which depends on the viewing angle θ, the polarization H, V and the dielectric constant for sea water e(f ) εH,V = 1 − ρ2H,V (6) p cos θ − e − sin2 θ p = cos θ + e − sin2 θ (7) p e cos θ − e − sin2 θ p ρV = e cos θ + e − sin2 θ (8) ρH Robinson 2004 18 of 38 Dependence with wind • • The roughness of the sea surface creates fluctuations in the incidence angle within the footprint of the radiometer. ◦ The resulting brightness depends on the sea surface slope statistics. ◦ Since this is controlled by wind stress, it contains information about winds. ◦ An effective emissivity ε̂ can be introduced to take roughness into account. The presence of foam also modifies the emissivity. ◦ The presence of foeam also depends on wind stress. 19 of 38 The dielectric constant of sea water e • The emissivity not only depends on the viewing angle but also with e. es − e∞ σ e(f ) = e∞ + −i , 1 + i2πf τr 2πf ε0 (9) valid for f < 85GHz. ◦ ε0 is the permittivity of free space ◦ e∞ is the dielectric constant at very high frequency ◦ es is are the dielectric constant at zero frequency ◦ τr is the relaxation time ◦ • σ is the ionic conductivity Depend on T and S The consequent dependence of ε on T implies that the relation between microwave brightness and T is not linear. 20 of 38 Penetration depth • The penetration depth in the microwave range is given by √ 2πf e d(f ) = c • Consequently, it depends on T and S • Penetration depth for IR (λ = 12µm) Hasoda, JO, 2010 (10) dIR ∼ 4µm • Penetration depth for MW (f = 5GHz) dM W ∼ 5mm • Thickness of the skin-layer dT ∼ 100µm Robinson 2004 21 of 38 Distinguishing different types of SST • SSTint: hypothetical temperature at the exact air-sea interface ◦ • SSTskin: temperature within the conductive diffusion-dominated sublayer ∼ 10 − 20µm ◦ • https://www.ghrsst.org/ Well approximated to the measurement by a MW radiometer (6-11 GHz) SSTdepth: temperature measurements beneath the SSTsubskin ◦ • Measured by an IR radiometer SSTsubskin: temperature at the base of the conductive laminar sub-layer of the ocean surface ◦ • No practical use. Wide variety of platforms and sensors and distinct from those obtained using remote sensing techniques SSTfnd: temperature free of diurnal temperature variability 22 of 38 Sensitivity to SST • The sensitivity of microwave radiance to SST is necessary to set the mounted channel at the most sensitive frequency. dTB dT Hasoda, JO, 2010 (11) • Maximum sensitivity within the range 4–10 GHz. • The higher frequency channels of microwave radiometers (10 GHz) are less sensitive to SST change in the low SST range. • The optimal frequency range was 4–6 GHz for all-range SST estimation. 23 of 38 Effect of the atmosphere • • The brightness temperature TB as seen from the antenna has three contributions: Robinson 2004 ◦ SST Ts ◦ Cold space Tc = 2.7K ◦ Atmosphere TatmU and TatmD , which depend on the temperature profiel Ta(z) The resulting brightness temperature is given by TB = TatmU + τH εTs + τH R(1 + Ω)(TatmD + τH Tc) (12) where R = 1 − ε is the reflectivity, Ω accounts for the additional reflectivity due sea state and τH is the transmittance. • Transmittance depends on oxygen, water vapor and cloud liquid water. • Rain absorbs and scatters microwave radiation and makes impossible the retrieval of SST. 24 of 38 Attenuation Robinson 2004 • The spectral dependency of the attenuation due to oxygen (- - -), water vapor (—) and cloud water (...) allows to develop corrections based on multispectral measurements 25 of 38 Summary of contributions to TB Robinson 2004 26 of 38 Retrieval of SST: Wentz and Meissner’s algorithm • Proposed by Wentz and Meissner 2000 and updated by Wentz and Meissner 2007 • Based on the Radiative Transfer Model (RTM) ◦ Atmospheric absorption model for water vapor, oxygen and liquid cloud water ◦ Sea surface emissivity model that depends on SST, SSS, sea surface wind speed and direction • The RTM is used to Monte Carlo simulate TB at the Top of the Atmosphere (TOA) • Simulated TB are then used to train a two-stage regressions to retrieve SST. • ◦ First-stage uses a set of regressions trained with global data: good but not optimal ◦ Second-stage uses a set of regressions trained for localized range of environmental conditions around the first-stage retrievals. This algorithm is the basis of the algorithm used by REMSS to retrieve SST from TMI, AMSR-E and WindSat. 27 of 38 AMSR 28 of 38 AMSR-E on October 15 2007 29 of 38 SST measured by TMI, AMSR-E and WindSat on October 15 2007 30 of 38 Spatial resolution • The footprint ∆ is given by cH ∆= Df (13) where D is the antenna diameter and H the satellite altitude. • The spatial resolution of the output will depend on the dominant band for a given product Robinson 2004 31 of 38 Example: hurricane Bonnie 32 of 38 Applications: ocean currents Isern-Fontenet et al. GRL 2006 33 of 38 Main microwave radiometers 1 Sensor acronym Platform(s) Agency Dates Channels Cent. freq. Polarization GHz Main data products TMI TRMM NASA/ JAXA 11/1997- AMSR ADEOS-2 JAXA/ NASA 12/2002-10/2003 AMSR-E Aqua 5/2002-10/2011 AMSR-2 GCOM-W 5/2012- SST Wind speed Water vapor Cloud liquid water Rain rate SST Wind speed Atmos. water vapor Cloud liquid water Rain rate Sea ice WindSat Coriolis 10.7 19.4 21.3 37.0 85.5 6.925 10.65 18.7 23.8 36.5 89.0 50.3/52.81 6.8 10.7 18.7 23.8 37.0 U.S. DoD 1/2003- V, H V, H H V, H V, H V, H V, H V, H V, H V, H V, H V V, H FP FP V, H FP SST Wind speed Wind direction AMSR only 34 of 38 Further reading Papers: • Hasoda, K. (2010). A review of satellite-based microwave observations of sea surface temperatures. Journal of Oceanography, Volume 66, Issue 4, pp 439-473 Web: • http://www.remss.com/ • https://www.ghrsst.org/ 35 of 38 Data sources • • • Remote Sensing Systems: http://www.remss.com/ ◦ AMSR-E, TMI and WindSat ◦ L3 data GCOM-W1 Data Providing Service: https://gcom-w1.jaxa.jp/auth.html ◦ AMSR, AMSR-E, AMSR-2 ◦ L1B, L1R, L2 and L3 data National Snow & Sea Ice Centre: ftp://sidads.colorado.edu/pub/DATASETS/nsidc0302 amsre qtrdeg tbs/ ◦ AMSR-E ◦ L1B and L2 data 36 of 38 Practical: basic examples • Before you start you have to load the environmental variables IDL> @startup • Run example 1: explore the fields available in a AMSR-E L3 file from REMSS IDL> example1 • Run example 2: compare AMSR-E L3 and TMI L3 files from REMSS IDL> example2 • Run example 3: explore averaged AMSR-E L3 files from REMSS IDL> example3 37 of 38 Practical: additional examples • Explore AMSR-E L2 field data/AMSR E L2A BrightnessTemperatures V10 200710150436 D.hdf • Explore AMSR-2 sample data http://suzaku.eorc.jaxa.jp/GCOM W/data/data w sampledata.html • Explore WindSat data and REMSS products and tools ftp://ftp.remss.com/windsat/ 38 of 38