Calibration processors for AEOLUS A. Dabas
Overview of AEOLUS processors For each height-bin FP A N A AEOLUS reciever FP B N B Rayleigh Brillouin spectrum + Mie peak Fizeau N mie (k) Pieces of information for each observation and each height bin N mie AUX_MRC Mie wind L1B N A AUX_RBC Rayleigh wind L2B N B AUX_CAL Aerosol products L2A
Mie wind: AUX_MRC From N mie (k), the Mie core algorithm estimates an index (position on the CCD). This response has then to be converted into a frequency. This is done with the calibration function characterized in the AUX_MRC product. Internal Response Calibration (IRC) AEOLUS pointed vertically (no wind Doppler shift) 40 observations are made at 40 different frequency steps f = f base + 475MHz 450MHz 25MHz 0MHz +25MHz +475MHz +500MHz Mie Core MR(1) MR(2) MR(19) MR(20) MR(21) MR(39) MR(40) For each frequency step, application of Mie core algorithm to spectra registered in internal reference channel (laser light sent directly to reciever) AUX_MRC IRC performed on a regular basis.
Rayleigh winds: AUX_RBC Rayleigh winds are retrieved from the Rayleigh Response measurements: R = N A N B N A + N B The conversion of R measurements into frequency Doppler shifts Δν D is done with calibration curves. PROBLEM: R does not solely depend Δν D, but also also on the atmosphere N A,B (ρ 1)I aer f + Δν D f 0 + I mol f + Δν D f 0, P, T T A,B f df where ρ = 1 + β aer /β mol is the scattering ratio, P and T the pressure and temperature inside the probed volume. There is a need for as many calibration curves are there are possible combinations of (P,T) in the atmosphere (the impact of aerosols on R can be corrected before the inversion). This need is covered by the AUX_RBC product
AUX_RBC The AUX_RBC is a 3D look-up table that contains the value of R for many sets (Δν D, P, T) paving the volume of possible combinations in the atmosphere. Δν k R i, j, k Inverting a measurement R amounts to making a interpolation in the LUT, the actual pressure and temperature being given by a meteorological forecast and contained in the AUX_MET product. T j P i Pressure The LUT is computed by integrating N A,B Δν D, P, T = I mol f + Δν D f laser, P, T T A,B (f) df where is the Rayleigh Brillouin molecular spectrum simulated with B. Wistchas analytical model. Note: a specific study was conducted on the shape of Rayleigh- Brillouin spectra. NOTA BENE: The procedure assumes T A f and T B f are known B. Witschas, Analytical model for Rayleigh Brillouin line shapes in air, Appl. Opt. 50, 267 270 (2011).
AUX_RBC R Δν D i, P j, T k AUX_RBC
Aerosol products AUX_CAL The L2A processor retrieves optical properties of the atmosphere from N ray = N A + N B and N mie = N mie (k) k N ray = N A + N B N mie = N mie k k L2A processor β aer α aer These quantities are related to the optical properties of the atmosphere through Molecular basckatter in Rayleigh Aerosol basckatter in Rayleigh N ray = ΔTK R ray R 2 C 1 (Δν D, P, T)β mol + C 2 Δν D β aer exp 2 α aer r + α mol r dr N mie = ΔTK mie R 2 C 3 Δν D β aer + C 4 (Δν D, P, T)β mol exp 2 α aer r + α mol r dr 0 R 0 Molecular basckatter in Mie Molecular basckatter in Mie
AUX_CAL Calibration coefficients C 1 to C 4 can easily be computed C 1 Δν D, P, T C 4 Δν D, P, T = I mol( f + Δν D f 0, P, T) T A f + T B f f T Fiz C 2 Δν D T = I C 3 Δν aer f + Δν D f A f + T B f 0 D T Fiz f δ f+δν D f 0 df df Coefficients K ray and K mie are estimated by comparing the signal level actually measured by the lidar in aerosol-free regions of the atmosphere with the signal level predicted by N ray = ΔTK R ray R 2 C 1 (Δν D, P, T)β mol exp 2 α mol r dr N mie = ΔTK mie R 2 C 4 (Δν D, P, T)β mol exp 2 α mol r dr 0 with β mol and α mol computed from the profiles of air density predicted by the meterological model (and provided in the AUX_MET file). β mol [m 1 sr 1 ] 1.38 α mol z 1.16 550 355 4.09 550 P[hPa] 355 1013 4.09 P[hPa] 1013 0 R 288 T[K] 10 6 288 T[K] 10 5
AUX_CAL Retreival of AUX_CAL from data simulated with the E2S The AUX_CAL processor looks for a «clean» region in the atmosphere between 6 and 16km, detects the cloud between 8 and 10km, and computes K ray and K mie with data between 10km and 16km.
AUX_CAL K ray, K mie C 1 Δν D i, P j, T k, C 2 Δν D i C 3 Δν D i, C 4 Δν D i, P j, T k AUX_CAL
FP transmission: AUX_CSR The derivation of both AUX_RBC and AUX_CAL both assumes the spectral transmission characteristics of the Fabry-Perot and Fizeau interferometers are known. In principle, this information is provided by a dedicated operational mode of the lidar: the Instrument Spectral Registration (ISR) Laser light f laser = f base + FSR FSR : 25MHz: 2 2 Aeolus receiver FPA FPB Fizeau T A ISR T B ISR ISR T Fizeau f laser f laser f laser However, the étendue of the laser beam in the Rayleigh channel is different from the étendue of the beam coming from the atmosphere T A f T A ISR f T B f T B ISR f A scheme has been devised to estimate T A,B f from T ISR A,B is based on the follwing model for the étendue effect: f. It Δ 1 (1 + 0.5x) Δ 1 Δ 1 (1 0.5x) T A,B f, Δ, x = T ISR A,B Π Δ,x (f) 0 Δ 2 Δ 2
Rayleigh Response AUX_CSR The parameters Δ and x are determined by comparing a Rayleigh Response curve (R versus Δν D actually observed during an IRC and a prediction of the Rayleigh Response based on T A,B f, Δ, x The prediction of the Rayleigh Response use the same equations as the RBC and relies on the provision of pressure and temperature profiles in the atmosphere by a meteorological model (AUX_MET). 0.5 0.4 0.3 0.2 0.1 0-0.1 IRC_Std_noise_w500_tilt100 RRC ISR PRED CSR PRED -0.2-0.5 0 0.5 Frequency shift (GHz) True étendue parameters: Δ = 500MHz x = 1 Estimated étendue parameters Δ = 654MHz x = 0.8
AUX_CSR T A, T B AUX_CSR
Summary AUX_MRC Mie winds AUX_RBC Rayleigh winds AUX_CSR AUX_CAL Aeorosol products A new AUX_MRC and a new AUX_CSR shall be produced every time an IRC is carried out, that is, once a week.