Fourier Transforms and Auto-correlation in Identifying Column Integrated CO2 Mixing Ratio from a Continuous Wave Lidar System for the ASCENDS Mission

Transcription

1 Fourier Transforms and Auto-correlation in Identifying Column Integrated CO2 Mixing Ratio from a Continuous Wave Lidar System for the ASCENDS Mission Doug McGregor, Jeremy Dobler, Jeff Pruitt, Grant Matthews ITT Exelis This document is not subject to the controls of the International Traffic in Arms Regulations (ITAR) or the Export Administration Regulations (EAR).

2 Overview What is ASCENDS? Why is it important? The Instrument Modulation Early Modulation Processing New Modulation Processing Results of Modulation Resolving Range In Situ Collection Results vs. In-Situ Conclusions, current and future efforts 2

3 What is ASCENDS? What is this instrument? ITT Exelis Multifunctional Fiber Laser Lidar (MFLL) is one of several Airborne Instrument Demonstrators (AIDs) for Active Sensing of CO2 Emissions over Nights, Days, and Seasons (ASCENDS). Why is ASCENDS important? Understanding the carbon cycle is critical understanding long term climate. We know CO2 sources, but we really don t understand CO2 sinks. In-situ systems either aren t persistent, or are geographically sparse. Passive sounders need solar illumination for measurements. How is ITT Exelis MFLL novel? True simultaneous multi-line spectroscopic measurement of CO2 and O2 eliminate most environmental noise (common mode) and enable true ppm calculations. CW lasers, amplifiers, and modulators leverage long-life telecomm technology. We re illuminating the ground with colors both on and off absorption lines, and using fancy modulation & demodulation to sort out how much comes back from each. This lets us measure column CO2. 3

4 Multifunctional Fiber-Laser Lidar Engineering Development Unit (EDU) Top and Bottom: Intensity Modulation is applied to each CW output color for multi-line spectroscopy. In between: Both on-line and off-line colors are incident on the same detectors. We separate the signals though digital processing. 4

5 Modulation Techniques and Instrument History 1 st Gen: Pure modulation tones Began with off-the-shelf lock-in amplifiers Great with a perfectly empty path Vulnerable to multipath different superposition for different tones Vulnerable to electrical-system frequency response 2 nd Gen: Swapping Modulation Tones All channels swapped (rolled) tones every 1/500 th of a second Solved uneven frequency-response Solved uneven multipath effects Still vulnerable to multiple returns like cloud decks 3 rd Gen: Exotic Modulation Schemes Swept, Stepped, and Pseudo-Noise Modulation Greater bandwidth allows for separating returns by range! 5

6 Early Modulation Processing To Simulate a lock-in amplifier Multiply the incoming signal sample by sample with The original modulation tone (Sin) A 90 degrees out of phase version of the modulation tone (Cos) Add the mean values of those arrays in quadrature Works Perfectly for pure tones Requires pre-shifting when the waveforms have time-dependency Only provides data for a single phase shift, unless it slowly loops 6

7 Fourier Modulation Processing To perform a correlation (like a lock-in iterated for all possible phases) Multiply the Fourier transform of the incoming signal sample by sample with The FFT of the complex conjugate of the original modulation tone (Sin) The FFT of the complex conjugate of the 90 degrees out of phase version of the modulation tone (Cos) Take the Inverse FFT of both results Add the mean values of those arrays in quadrature Provides a lock-in value for all ranges simultaneously (many times faster) 7

8 Resolving Range and Discriminating Against Clouds The swept encoding was used for the DC-8 flights and demonstrated the ability for the CW measurement to both determine the range to the surface and discriminate returns from clouds. Plots from Wallace Harrison (NASA LaRC) 8

9 In-Situ Collection Typical science flights have multiple segments at altitudes ranging from 10 kft to 40 kft in 5 kft intervals over the target of interest. Spiraling down at the center of the flight track provides a full-column in-situ collection. With temperature, pressure and relative humidity information, we generate a modeled optical depth. The modeled OD is then compared to the remote lidar measurements. 9

10 Courtesy of Dr. Ed Browell (NASA LaRC) 10

11 Conclusions, Current and Future Work ITT Exelis has demonstrated a Continuous Wave (CW) Intensity Modulated (IM) Integrated Path Differential Absorption (IPDA) measurement that is capable of discriminating against thin clouds and aerosols in the path. The demonstration unit has shown absolute comparisons to in situ measurements traceable to WMO standards to sub ppmv levels with standard deviations of <2ppmv We have also demonstrated the ability to determine range to the surface without the need for a separate altimeter We are currently developing a Global Hawk compatible demonstrator under a NASA ESTO IIP We ve also made the first O 2 measurements at 1.26um using an all fiber Raman amplifier transmitter We are beginning a follow on ACT to scale the power of the O 2 Raman amplifier to 5 W, and demonstrate an XCO 2 measurement using the combined CO 2 and O 2 measurements. 11

12 Acknowledgments We would like to thank NASA LaRC for their continued contributions to advancing and evaluating the MFLL instrument concept and processing algorithms for application to the ASCENDS mission. We would also like to thank NASA Earth Science and Technology Office for funding the development of the O 2 Raman amplifier and our current IIP to develop and advanced demonstrator compatible with the Global Hawk platform Special thanks to Dr. Ed Browell and Wallace Harrison (both NASA LaRC) and their teams for the data analysis seen in these slides. 12

13 Supplemental Slides Courtesy of Jeremy Dobler ITT Exelis This document is not subject to the controls of the International Traffic in Arms Regulations (ITAR) or the Export Administration Regulations (EAR).

14 Multifunctional Fiber-Laser Lidar Engineering Development Unit (EDU) High reliability telecom components, CW operation requires low peak power and allows for less expensive laser components and higher MTBF. Simultaneous transmission of all wavelengths eliminates many noise sources from atmosphere, surface and instrument by rendering them common mode. 14

15 Flight Campaign history Flight Test Campaigns LAS Dec. 1 Dec 6, 2004 Newport News, VA First aircraft integration and engineering check flights 4 Engineering Check flights LAS May 21-25, 2005 Ponca City, OK 5 Science Flights: Land, Day & Night DOE ARM Site LAS Jun 20-26, 2006 Alpena, MI 6 Science Flights: Land & Water, Day & Night LAS Oct 20-24, 2006 Portsmouth, NH 4 Science Flights: Land (inc. mountains) & Water, Day & Night LAS May 20-24, 2007 Newport News, VA 8 Science Flights: Land & Water, Day & Night LAS Oct 17-22, 2007 Newport News, VA 9 Science Flights: Land & Water, Day & Night, Clear & Cloudy LAS S ept 21-Oct 30, 2008 Hampton, VA (LaRC) 10 Science Flights: Land & Water, Day & Night, Clear & Cloudy LAS July 10 17, 2009 Hampton, VA (LaRC) 5 Science Flights: Land & Water, Day & Night, Clear & Cloudy LAS July 31 Aug 7 th, 2009 Ponca City, OK 5 Science Flights: Land, Day & Night, Clear & Cloudy DOE ARM Site LAS Aug st, 2009 Hampton, VA (LaRC) 2 Science Flights: Land & Water, Clear & Cloudy LAS May June 2010 Hampton, VA LARC 6 science flights, land water clear and cloudy LAS July-August 2010 Dryden Air Operation Facility 5 science flights diverse land and water, mostly clear Total of >70 individual flight sorties over 7 years and >1000 hours of ground testing 15

16 2011 Was a Busy Year Several upgrades were made to the EDU in 2011 The detector subsystem was modified to support up to a 2 MHz bandwidth and lightened by ~10 lbs The data acquisition system was upgraded to generate and receive at 2MS/s An O2 amplifier was integrated into the EDU Phase sensitive waveforms were implemented Swept Stepped Pseudo-random Noise Each encoding was evaluated on the NASA UC-12 aircraft during 5 science flights in May

17 July and August 2011 DC-8 Flights Completed 7 science flights and one engineering check flight Total of 43 hours of flight data Flight 0 Castle AFB (eng chk flt) Flight 1 Castle AFB (clear conditions) Flight 2 Open Ocean (over thick stratus deck) Flight 3 Rail Road Valley Nevada (some low level and high cirrus clouds) Flight 4 British Columbia (Snow and Ice) Flight 5 Four Corners (power plant plumes) Flight 6 WBTT (Iowa tall tower site) Flight 7 WLEF (Wisconsin tall tower site) 17

18 Data from Dr. Ed Browell (NASA LaRC) 18

19 Amplitude (V) dbm O2 Channel Development Offline Forward Raman Spectrum at Full Pump Power (60A) from 40dB Tap Wavelength (nm) An O 2 Raman amplifier was developed under NASA ACT grant. The amplifier outputs 1.5 W average power with a spectral profile that matches the input seed DFB lasers. The amplifier was integrated into a DC-8 high-bay rack and flown in July and August 2011 for 43 hours of flight operations Online 50kHz Modulation at Full Pump Power (60A) from 40dB Tap (180mA SOA) Time (us) 19

20 O2 Results The normal mode of operation for both the CO 2 and O 2 channels is to use a fixed on and off line shown on the right. During the transit flights we also ran sweeps of the O 2 trough we are using. All of the sweeps were plotted as differential transmission versus wavelength and compared to model predictions from LBLRTM, shown in the bottom plots. 1 Ratio LC/SO vs Wavelength Average of All Sweeps ratio LC/SO Model data