CO 2 mixing ratio retrievals from JPL airborne Laser Absorption Spectrometer flight campaigns in

Transcription

1 CO 2 mixing ratio retrievals from JPL airborne Laser Absorption Spectrometer flight campaigns in Robert Menzies, Gary Spiers, Joseph Jacob Jet Propulsion Laboratory California Institute of Technology CLRC XVI Long Beach, California June, 2011

2 Laser Absorption Spectrometer (LAS) Concept Transmit and receive with near nadir-pointing of on- and off-line wavelength channels Ground surface backscatter (land and sea) provides return signal Measure integrated path differential absorption (IPDA) Heterodyne detection at 2-μm wavelength High sensitivity: quantum-noise limited; relatively better at longer wavelengths; Detected signal spectrum provides ancillary information (spectral purity, stability) Use additional sensor data (T, P, RH, GPS, altimetry, line parameters) and forward model to retrieve weighted CO 2 column mixing ratio. Instrument development: NASA ESTO; Integration on DC-8: NASA AITT; Data analysis: NASA ASCENDS

3 JPL CO2 LAS Airborne Operations. Flight Campaigns: California: engineering checkout flights were conducted over the Mojave desert and the Pacific Ocean during summer Virginia; 5 flights were conducted in Virginia in October 2007 under the joint LaRC/ITT and JPL campaign. California: El Mirage, April, Oklahoma: 4 flights near the ARM SGP site, July/August, 2009 joint campaign with LaRC/ITT and GSFC airborne instruments. NASA DC-8 flights over 5 sites (California, Pacific, Oklahoma), July, 2010.

4 JPL Airborne CO2 LAS Parameters Parameter CO 2 line center wavelength JPL LAS ON wavelength JPL LAS OFF wavelength Laser output power Transmit/Receive Telescope apertures Receiver FOV (diffraction limited) Photomixer type Receiver heterodyne frequency window Signal Digitization Value cm cm cm mw 10 cm diameter 60 μrad InGaAs MHz 14 bits / 50 MHz

5 JPL Airborne CO2 LAS Transceiver Optics Opto-mechanical structural and thermal design/construction sufficiently robust for hands-off operation in a turbo-prop powered, unpressurized aircraft. (JPL/CTI collaboration) LAS transceiver with optical bench horizontal, telescope side up, base plate in background. LAS transceiver in hermetically sealed enclosure, windows mounted in base plate

6 Two flight campaigns in 2009 El Mirage Dry Lake, Mojave April flight days 3 altitudes each flight day: 2kft, 6kft, 8kft sequence; Concentrated analysis on 6-8 kft layer; 6 overpasses at each altitude Surface: flat, relatively uniform reflectance along ground tracks Atmosphere: dry, moderate temp., well-mixed; negligible change from day to day including CO2 column profile Oklahoma (ARM SGP vicinity) July/Aug flights days 4 altitudes each flight day: 10kft, 9kft,7kft,8kft sequence 2 overpasses at each altitude Surface: varied surface cover (grass, trees, bare soil, water), large reflectance variations Atmosphere: ~40% RH, moderate temp.; layered, inversion at 2-3 km; large day to day variation in CO2 column profile below ~3 km

7 Data Collection Location El Mirage Dry Lake Bed 34 o N, W, 2800 asl 96 km 8 km

8 Oklahoma Ground Tracks over ARM SGP Site 321 m 304 m 304 m 287 m 290 m 298 m 56.1 km

9 DC flight campaign: 5 sites Overpasses at 9 different altitudes between 8kft and 40 kft ; Super spiral in addition; 5 sites: CA Central Valley (July 8, 2010) Atmosphere mostly clear, mild surface topography, large reflectance dynamic range; Mojave/Needles (July 9) Scattered clouds; significant topographical variability, relatively uniform reflectance along ground tracks; Railroad Valley, Nevada (July 12) a favorite satellite sensor calibration surface (e.g. GOSAT) Pacific Ocean off coast of Calif./Baja Calif. border (July 14) Atmosphere mostly clear, no topographical variability! Oklahoma DOE Southern Great Plains site (July 18) Concentrated on 3 sites to date: Central Valley, Mojave/Needles, Pacific Ocean

10 Airborne CO2LAS Methodology CO2LAS measurements produce ln(ratio) or DAOD values for various atmospheric layers defined by the aircraft flight altitudes and/or the surface LBLRTM forward model transforms ln(ratio) values into [CO2] column (# CO2 molecules per unit area in the measurement column)

11 Signal Processing First Steps Heterodyne Intermediate Frequency signals appear in MHz window based on ac ground speed and offnadir point-ahead angle; Digitization rate: 50 MHz; (1) FFT operation (each successive 16K FFT: 320 μs time slice); (2) periodogram formation; (3) periodogram summing Data collection volume is ~0.5 TB/hour Periodograms from DC-8 Mojave Desert overflight, July 9, 2010; T INT = 40 ms

12 The ln(ratio) On-line and Off-line return signals are given by I R on = IT on.η on.r.exp(-2.n CO2 σ on.l) and 2.n CO2 σ off.l) I R off = IT off.η off.r.exp(- where the subscripts on and off denote the online and offline channels respectively, I R and I T are the receive and transmit powers, η on and η off are the instrument response functions for each channel, R is the target reflectivity, n CO2 is the CO 2 number density, σ is the absorption cross section, l is the path length between the instrument and target. We can normalize the detected return power on each channel, divide and take the natural log to obtain the ln(ratio) : ln(i N on /IN off ) = - 2.n CO2 l(σ on σ off ), where the N superscript denotes the normalized power. Summary: The ln(ratio) is twice the Differential Absorption Optical Depth (DAOD) due to the absorbing gas molecules in the atmospheric path.

13 IPDA Measurement of CO 2 (vertical path) Differential Absorption Optical Depth measured through a thin layer of thickness δz at height Z in the atmosphere: δ(daod) Z = n CO2 (z) w(z) [(σ on σ off )] SURFACE δz, where w(z) is the normalized weighting function; Weighting Function vertical weighting depends on the specific displacement of online frequency from line center; w(z) depends on atmospheric T, p, RH as well. Airborne CO2LAS measured DAOD is related to the number of CO2 molecules (per unit area) in the path through a layer defined by upper and lower flight altitudes (at z 1, z 2 boundaries) or the surface. Airborne CO2LAS Measurement precision, repeatability, stability will be discussed in terms of relative DAOD units: i.e. DAOD Measurement Uncertainty Relative to Full Atmospheric Column DAOD

14 IPDA Measurement of CO 2 Instrumental Measurement Precision and SNR The relative error on the measurement of DAOD is directly related to the on-line and off-line signal-to-noise ratios (SNR s); Ignore transmit power measurement uncertainty for now. Set DAOD = τ The expression for instrumental uncertainty in determination of τ is 2 () 1 ( P ) ( Poff ) on 2 P on Poff where σ is the standard deviation, P on, P off, are the on-line, off-line measured return power, respectively. Objective: Make σ(τ)/τ as small as possible! If the return power measurement uncertainty refers only to the detection process, then the rhs depends on the values of τ and the SNR for each channel. 2 1/2

15 Airborne LAS Measurements of CO 2 : El Mirage Measured ln(ratio) for column from aircraft altitude to surface: Three altitudes, multiple overpasses at each altitude; On-board Picarro instrument (from P. Tans, NOAA CMDL) for blind comparisons; Concentrated on the CO2 DAOD measurement stability in the 6kft 8kft layer; Calculated delta ln(ratio) values to get DAOD for the 6kft 8kft layer; Repeatability at each altitude of measured DAOD: (2Xspeckle limit) Day to day Repeatability of measured DAOD in this layer: Difference between April 17 and April 19 DAOD: This is 1% of the CO2 DAOD of the entire atmospheric column (that would be measured from high altitude aircraft or from LEO) In ppm equivalent, this is day-to-day repeatability at the level of ~ 4 ppm El Mirage result: We demonstrated a measurement precision level of ~0.5% of the total tropospheric column CO2 DAOD that is obtained using this 2-μm line pair.

16 Layers Measured over Oklahoma Site In the Twin Otter, we measure the DAOD in layers that are a small fraction of the total atmospheric column. The weighting function is calculated using measured atmospheric parameters at the time of the flights

17 Airborne LAS Measurements of CO2 : Oklahoma Measured ln(ratio) from aircraft altitude to surface at each flight altitude; Four altitudes: 7kft, 8kft, 9kft, 10kft (GPS altitude data for each) two overpasses at each altitude: ground tracks within ~100m of each other Emphasis on measurements of 7 kft 10 kft layer DAOD Day to day Variability of measured DAOD in the 7 10 kft layer: For 2 of the 4 days, we have both good atmospheric data and good Validator data at 10kft and 7kft; (data system problem on 8/02, temperature controller problem on 8/03 at the lower altitudes); Difference between July 31 and August 4 measured DAOD: < (overpass average) In situ CO2 profiles (from Ed Browell and Yonghoon Choi) indicated stable 7-10 kft column average CO2 (~ 385 +/- 1 ppm) from day-to-day; Following these tests, proceeded with CO2 column retrievals down to the surface.

18 Signals from 10 kft Pass over Oklahoma Site The ratio holds steady over wide range of surface reflectance (power is equivalent) Residual fluctuation in ratio primarily due to speckle fluctuation, which decreases with increasing integration time. (2.5 sec = ~ 130 m along track) Arrow shows flight track and direction

19 Retrieved CO2 Column Mixing Ratio: 10kft Pass over Oklahoma Site Upper Plot: Retrieved CO2 / 600 m along track resolution (10 s average) Lower Plot: On-line and off-line signals over variable surface reflectance Std Dev due to speckle fluctuations for 600 m resolution: ~ 5 ppm

20 Oklahoma: CO2 Variability in Lower Troposphere The same level of variability in the weighted column [CO2} seen on both days. Column average mixing ratios (overpass average): 7/31: 382 ppmv // 8/04: 393 ppmv The ~ 10 ppm difference in column mixing ratio is consistent with in situ profile data

21 DC-8 Summer 2010 Campaign LAS Instrument SNR Summary Flight Pacific Ocean July 14, 2010 Mojave July 9, 2010 Central Valley July 8, 2010 Surface Retro- Reflectance (average) Off-line SNR 25 kft altitude 1 km average DAOD (1-way) (400 ppm CO 2 ) On-line SNR 25 kft altitude 1 km average sr E E sr E E sr E E+2 Note: Speckle uncertainty for 1 km along-track average: 0.7% (~ 2.7 ppm rms)

22 Mojave/Needles July 9, 2010 Retrieved Column CO2: 20 kft and 8 kft overpasses

23 Results Summary Flights over El Mirage where the atmosphere was stable and surface reflectivity variation low demonstrated the repeatability of the measurement both over multiple passes and between days. Flights over Oklahoma had much more variability in the atmospheric conditions, including CO2 and in the surface reflectivity but demonstrated a precision/stability consistent with the El Mirage results. Summer DC-8 flight campaign results: Obtained high precision DAOD measurements from 8 kft up to 25 kft; Retrieved column CO2 data with along-track resolution ~ 2 km show some variability above fluctuations expected from speckle, SNR; Improvement in laser power normalization: ongoing effort Instrument is structurally & opto-mechanically robust good repeatability Data analysis algorithms have improved and are now more robust. The weighting function enables the ability to measure column CO2 in tropospheric layers defined by the aircraft overpass altitudes, as well as layers with surface as lower boundary.

24 Backup Slides