This study evaluates the capability of atmospheric CO2 column measurements under cloudy conditions using an airborne intensity-modulated continuous-wave integrated-path-differential-absorption lidar operating in the 1.57-μm CO2 absorption band. The atmospheric CO2 column amounts from the aircraft to the tops of optically thick cumulus clouds and to the surface in the presence of optically thin clouds are retrieved from lidar data obtained during the summer 2011 and spring 2013 flight campaigns, respectively. For the case of intervening thin cirrus clouds with an average cloud optical depth of about 0.16 over an arid/semi-arid area, the CO2 column measurements from 12.2 km altitude were found to be consistent with the cloud free conditions with a lower precision due to the additional optical attenuation of the thin clouds. The clear sky precision for this flight campaign case was about 0.72% for a 0.1-s integration, which was close to previously reported flight campaign results. For a vegetated area and lidar path lengths of 8 to 12 km, the precision of the measured differential absorption optical depths to the surface was 1.3 – 2.2% for 0.1-s integration. The precision of the CO2 column measurements to thick clouds with reflectance about 1/10 of that of the surface was about a factor of 2 to 3 lower than that to the surface owing to weaker lidar returns from clouds and a smaller CO2 differential absorption optical depth compared to that for the entire column.
© 2015 Optical Society of America
Changes in atmospheric carbon dioxide (CO2) concentrations are estimated to produce approximately a 1.82 W/m2 global radiative forcing  and are key factors in the net radiative heating to the Earth’s climate system and in determining future global climate change . The U.S. National Research Council has identified the need for an Active Sensing of CO2 Emission over Nights, Days and Seasons (ASCENDS) space mission  to reduce uncertainties in and to gain a better understanding of CO2 sources and sinks.
In preparing for the ASCENDS mission, NASA Langley Research Center and Exelis Inc. have jointly developed and demonstrated the capability of CO2 column measurements with an intensity-modulated continuous-wave (IM-CW) lidar. Results of CO2 column measurements from aircraft flight campaigns using Exelis’ prototype IM-CW lidar called the Multifunctional Fiber Laser Lidar (MFLL) have been very encouraging [4, 5]. MFLL operates in the 1.57-μm CO2 absorption band with one laser wavelength positioned on the CO2 absorption line center (online) at 1571.112 nm, and two other laser wavelengths (offlines) positioned ± 50 pm on either side of the absorption line. CO2 column differential absorption optical depth (DAOD) values are estimated from combined online and offline measurements using the Integrated Path Differential Absorption (IPDA) approach. The signal-to-noise ratio (SNR) for clear sky IPDA measurements of CO2 DAOD for a 10-s average over vegetated areas with 7 km range was found to be as high as 1300, resulting in an error 0.077% or equivalent XCO2 column precision of ~0.3 ppm [4, 6]. In order to obtain precise measurements of atmospheric column CO2 mixing ratios (XCO2), a ranging capability is required for the ASCENDS lidar to account for CO2 column changes due to surface topographic variations and to discriminate cloud and aerosol returns from surface returns. Precise range measurements using the IM-CW lidar approach were also achieved [4, 5], and the uncertainties had been shown to be below the sub-meter level .
Measuring CO2 column density in the presence of thin cirrus clouds, broken clouds, or cloud decks is a challenging task for both active and passive remote sensors. Passive CO2 retrievals avoid cloudy pixels and use only clear sky measurements . The purpose of this study is to extend the atmospheric CO2 column measurements of the IM-CW lidar from clear skies to cloudy conditions. Under thin cirrus conditions, IM-CW lidar returns from the surface can be clearly separated from those from thin cirrus clouds because the received lidar returns contain vertically-resolved range information as a result of the range-encoded intensity modulation of transmitted lidar signals [4, 5]. In the presence of low-level optically-thick scattered clouds or cloud decks, the lidar returns from the cloud tops are strong enough to provide a measurement of the CO2 column DAOD from the instrument to the clouds. These capabilities will enhance the ability of airborne and space-based IPDA lidars to provide CO2 column measurements in the presence of many types of clouds.
2. Instrumentation and methodology
The basic architecture of MFLL for CO2 column measurements is illustrated in Fig. 1. In [4, 5], Dobler et al. and Lin et al. describe this instrument and the corresponding measurement approach in detail and provide a list of the system parameters, so only a brief discussion will be presented here. Distributed feedback lasers (DFBs) are locked to the CO2 online and offline wavelengths and fed into their corresponding intensity modulators to impart their unique ranging-codes (e.g., swept frequency in the cases discussed in this paper) to each laser beam. A swept-frequency technique with 500-KHz bandwidth and center frequency around 350 KHz [4, 5] is used in our IM-CW lidar signals, which enables the ranging measurements and the ability to separate the cloud and surface returns. The intensity modulated lasers are then simultaneously amplified using an Erbium Doped Fiber Amplifier (EDFA) to increase the transmitted power. A tiny fraction of the amplified laser power is picked off via an optical tap inside of the EDFA and sent to a detector as reference signals Pr for transmitted laser-power normalization and calibration at the online and offline wavelengths. The lidar backscattered returns of the simultaneously transmitted online and offline signals from clouds and surface are recorded by the MFLL receiver and used as science signals for the CO2 DAOD measurements. Digitized science and reference signals from analog-to-digital convertors (ADC) are processed by correlating these data with their corresponding range-encoded waveforms through the use of matched filters. The magnitude and position of the synthetic pulses of these correlation outputs represent the power and time-delay, respectively, of the received signals. Thus, range estimation and the discrimination of surface returns from cloud returns are achieved based on the outputs of the correlation.
CO2 DAOD values can also be retrieved from a combination of online and offline signal powers using the IPDA technique. A brief discussion of the technique is given here. Details of the instrumentation, environmental impact, and signal processing can be found in [4, 5]. Assuming the transmitted laser power is Pt, the received laser power Pλ at wavelength λ from a backscatterer at range R collected by lidar receiver can be expressed as:4, 5]. Thus, the transmission ratio of online and offline:Eq. (1), this ratio and DAOD can be calculated from the online and offline power measurements  as:Eq. (1) and considering surface returns in optically thin (δ < 1) cloudy areas as well as those in neighboring clear sky regions, the cloud optical depth of the thin clouds can be estimated from the offline power ratio of the cloudy area surface return (Pcloudy) to the clear area surface return (Pclear):Eq. (5) can be used to estimate dust optical depth.
3. Airborne flight campaign data
The data used in this study were collected with the MFLL during the ASCENDS 2011 summer and 2013 winter airborne flight campaigns. The objectives of these campaigns were to demonstrate lidar atmospheric CO2 measurement capabilities in preparation for the ASCENDS space mission. Encouraging results for technology demonstration and CO2 column measurements under various environmental conditions were obtained during the past two field campaigns [4–7, 9–11], and the results of the MFLL cloud studies are described in this paper. The ranging capability of the MFLL system was first demonstrated in the 2011 flight campaign using the swept-frequency IM approach, which provided a clear separation of lidar surface returns from layers of intermediate backscatterers such as clouds and aerosols.
This study focuses on the demonstration of CO2 column measurements in the presence of various types of clouds. Thus, the weather without substantial aerosol impacts is considered to avoid the complexity of aerosol scattering on lidar signals. Two flight cases are discussed: one for DAOD retrievals to the surface through optically thin cirrus clouds and the other to the tops of optically thick low clouds. In situ measurements of the atmospheric profiles of CO2 concentration, temperature, pressure, and humidity were obtained during aircraft spirals for comparison with the MFLL remote measurements in each case. This study assumes that the atmospheric profiles are horizontally uniform around the spiral centers and have minimal changes within an hour of the spirals. Thus, these measured profiles were used to obtain in situ-derived CO2 DAOD values through radiative transfer calculations. The selections of atmospheric layers in calculating the in situ-derived CO2 DAOD values were based on the aircraft recorded flight altitudes and lidar measured ranges of individual lidar CO2 samples. With these flight altitudes and lidar ranges and their corresponding in situ measured dry-air pressures, lidar measured CO2 column DAOD values (i.e., CO2 column amounts or CO2 partial air pressures) were converted to column XCO2 based on the CO2 amounts and total dry-air amounts from the pressures. The spectroscopic information for the radiative transfer calculations was from the 2008 HITRAN database. To achieve high accuracy in the calculation of the CO2 absorption cross-sections, the CO2 line absorption parameters from recent measurements by Delvi et al.  were included, and the Voigt approximation for the absorption profile was used as discussed in previous reports [4, 5]. The absorption from neighboring CO2, H2O, and other gas lines were also included in this analysis. Remotely sensed CO2 measurements were adjusted to nadir based on real-time aircraft attitude records . Comparisons of MFLL and in situ-derived DAOD and/or equivalent XCO2 values were typically limited to a region within a lateral distance of 8 km of the spiral location where the in situ profiles were collected. When multiple spirals were conducted during a flight, the spiral data closest in time to the MFLL overpass were used. Detailed information on calculations of in situ-derived CO2 DAOD and comparisons of this derived DAOD with lidar CO2 measurements can be found in [4, 5].
3.1 Estimates of CO2 column concentration through thin cirrus clouds
The thin cirrus cloud case studied here was obtained over an arid/semi-arid region around 1600 LT on 22 February 2013 (0000 UT on 23 February 2013) near Blythe, California. The atmospheric profiles of CO2 concentration, temperature, pressure, and humidity were measured by in situ instruments onboard the NASA DC-8 aircraft during an aircraft spiral. This and the other case discussed next both showed significant (several ppm) vertical and horizontal XCO2 variations, which would introduce certain uncertainties in the comparison of lidar and in situ CO2 measurements. Figure 2 shows the in situ measured XCO2 profile. Planetary boundary layer (PBL) XCO2 values below about 3.5 km were significantly higher than those in the free troposphere, which started at about 400 ppm and then decreased with increasing altitude to about 393 ppm at 12 km. During the winter at mid-latitudes, vegetation normally shuts down its evapotranspiration process and nearly stops its CO2 uptake; however, surface soil and human activities continuously release CO2, which, along with air-mass exchange and transport at the top of and within PBL, causes elevated CO2 concentration in the lower atmosphere. The analyzed leg of this flight was maintained at an altitude of about 12.2 km, and the extended thin cirrus clouds were observed just below the aircraft.
Figure 3 shows the measured atmospheric profile as correlation power of lidar returns. Both online and offline channels clearly showed the thin cirrus clouds just below the aircraft (range close to zero). Higher power returns from the clouds observed by the online channel compared to the offline channel resulted from higher transmitted power at the online wavelength in order to offset the reduction in SNR due to CO2 absorption to the surface [4, 5]. The range to the extended thin cirrus layer was rather constant, indicating the height uniformity of these thin clouds. There were some changes in the range to the surface (lower panels), which resulted from small changes in the surface topography. For this cloud case, our calculation of cloud optical depth using Eq. (5) indicated that the cirrus cloud, as expected, was indeed very thin with an averaged cloud optical depth of about 0.158 (Fig. 4). The variability in the optical depth of these cirrus clouds was very large, as indicated in Figs. 3 and 4, with values ranging from near zero to as large as 0.8 with the median 0.135. The higher the cloud optical depth, the stronger the impact of the cloud on CO2 measurements potentially.
The retrieved DAOD of CO2 column to the surface and its equivalent XCO2 values are shown in Fig. 5. CO2 column DAOD estimates for both clear (blue points) and cloudy (presence of intervening thin cirrus clouds; red points) conditions for this flight leg are plotted separately. Their mean and standard deviations are listed in the figure to compare with those derived from in situ measurements (also listed). The mean difference in DAOD retrievals between cloudy and clear conditions (Fig. 5(a)) is about −0.0013 or −0.20%, which corresponds to an equivalent XCO2 difference of −0.7 ppm (Fig. 5(b)). For airborne CO2 column measurements, the thin clouds are very close to the sensor and their returns are significant compared with surface returns. While for space CO2 remote sensing case, because similar lidar powers would illuminate on cirrus and surface and cirrus reflectance would be one to two orders of magnitude smaller than that of the surface, the power returns from cirrus would be much weaker than those from surface. Thus, the difference in the CO2 column measurements between clear and cirrus cases would be much smaller (within 0.02% for δ ~0.16) as shown by space modeling studies . The clear sky precision of these measurements with 0.1-s integration time was as high as about 0.72% (or about 2.8 ppm). A 10-s integration would reduce the error to about 0.072% (or 0.28 ppm), which is consistent with our previous results . Compared to in situ derived CO2 DAOD column values, the lidar DAOD measurements were slightly smaller (about 0.004). Although the equivalent column XCO2 (about 395.0 ppm) obtained from lidar CO2 DAOD measurements were about 2.4 ppm lower than the in situ-derived values (397.4 ppm), this difference is within 1-σ (2.83 ppm) of the observational uncertainties of our 0.1-s integration measurements. Spatiotemporal differences in remotely sensed and in situ observations are expected to account for the absolute differences between the two measurements. Because clouds significantly reduce received lidar signals and degrade CO2 measurements, the required SNR of CO2 column measurements in cloudy conditions is a factor of e-δ scaled down value of corresponding clear conditions under current ASCENDS plan. Also, only weather conditions with δ less than 0.7 are required for CO2 retrieval to avoid unacceptable CO2 measurement error.
3.2 Retrieving CO2 columns to optically thick low level clouds
The data for this case were obtained over an agriculturally vegetated area on 10 August 2011 in the vicinity of West Branch, Iowa. As with the previous case, in situ measurements of the atmospheric profiles were obtained at about 1730 LT (2330 UT or 23.5 UT) during an aircraft spiral centered on the West Branch tall tower. Figure 6 shows the in situ measured XCO2 profile. In the free troposphere, the XCO2 values were generally >385 ppm, had a tendency of slowly increasing with altitude, and reached about 392 ppm at 12 to 13 km. The low XCO2 (~365 ppm) values in the PBL below ~2 km reflected the CO2 drawdown due to the active agriculture growing season, especially for corn, during the period of this flight campaign. The aircraft flew over the spiral location several times at different altitudes. Multiple layers of clouds from high thin cirrus and scattered mid-level clouds to thick low-level fair weather cumulus clouds were observed during this flight. These complicated environments of different kinds of clouds increase uncertainties in atmospheric CO2 column retrievals compared to those over pristine clear scenes.
Figure 7 shows the aircraft altitude profile around West Branch, IA on 10 August 2011. The descending spiral in the vicinity of the West Branch tower is indicated by the quick decrease in altitude from the highest flight altitude after 23.0 UT. Red segments in the figure represent the 7 flight legs within 8 km of the spiral center. These 7 legs are shown in order from low to high altitudes in the figure. For the low level flight legs around the spiral location, the range and DAOD values from the aircraft to the thick low cumulus clouds were small (e.g., leg 1 in Fig. 8) and thus, the precision of the DAOD measurements was not be very high because the SNR of DAOD is proportional to the product of DAOD value and SNR of the online and offline power ratio . To make an assessment of DAOD measurements to thick low clouds, four high altitude legs are considered. Within these four legs, observations showed that the early portion of leg 4 had heavy high clouds (c.f., right picture of Fig. 8) and almost no low clouds were observed for leg 6. Thus, DAOD columns to clouds were retrieved for the late part of leg 4 (about 22.29 UT) and the entirety of legs 5 and 7. The flight altitudes for these legs were approximately 7.8, 9.4, and 12.5 km, respectively.
The range estimates for legs 4, 5 and 7 are shown in Fig. 9. These ranges to the surface (blue points) or clouds (red points) were very similar for individual flight legs, which was a result of generally level flight attitudes, a flat surface, and similar cloud heights. Although improved range resolution can be obtained from spectral domain reordering of periodic modulation waveforms , the range data used in this analysis were processed using the curve fitting technique of the correlation output [4, 5]. The rms error of the range estimates was generally within 3 meters, which is very good and should also be adequate to meet space mission science requirements. DAOD retrievals (Fig. 10) showed consistent high precision results among the three legs although variations in DAOD retrievals to cumulus clouds were much larger than those to the surface owing to smaller DAOD values and weaker lidar power returns from the clouds. The estimated reflectance of thick boundary layer clouds was basically slightly higher than 0.01 sr−1 which was only about 1/10 of the reflectance of vegetated surface. Considerable variations in cloud 3D structure, cloud top height, and particle size distribution may also affect the precision of the CO2 measurements.
Table 1 summarizes lidar DAOD retrievals and their corresponding equivalent XCO2 estimates for 0.1-s integration time along with in situ-derived values. Because of the altitude dependence of XCO2, there were differences in the in situ-derived XCO2 column for the three legs analyzed (from about 381 ppm to about 384 ppm). Relative precision of CO2 DAOD column to surface was about 1.3 to 2.2% for these legs. Increasing integration time to 10-s would increase the precision to less than 0.18%. Compared to the errors of CO2 column to surface, errors in CO2 column to cloud tops increased to about 4.1 to 5.6% with 0.1-s integration time. Lidar measurements of XCO2 column to surface were within 3 ppm of the in situ observations. The differences were about a factor of 2 bigger for the retrievals of XCO2 column to clouds compared to those for XCO2 column to surface due to the reasons mentioned above. These differences are within 1-σ of the standard deviations of the lidar measurements.
This study evaluated the capability of IM-CW laser absorption lidar for CO2 column measurements in cases of thin cirrus and thick fair weather boundary layer cumulus clouds. For thin cirrus clouds, consistent CO2 DAOD and equivalent XCO2 column values to surface for clear and cloudy skies were obtained in a 12-km altitude flight over an arid/semiarid region. The clear sky precision for the flight campaign case studied was about 0.72% for a 0.1-s integration, which is very close to previous flight campaign results. The difference between lidar and in situ derived CO2 values was within 1-σ uncertainty of the CO2 measurements. Under optically thick boundary layer cloud conditions, it was shown that MFLL data can be used to make CO2 column measurements to the tops of the clouds although their reflectance was only about 1/10 of that of vegetated surface. Even in a very complicated multi-layer cloud environment of the mid-west vegetated area case with flight altitudes 8 to 12 km, the precision of the estimated CO2 DAOD column to surface was as high as 1.3 – 2.2% for 0.1-s integration. The precision of CO2 column measurements to thick clouds was about a factor of 2 to 3 lower than that to the surface owing to much weaker lidar power returns from clouds and a smaller CO2 column DAOD compared to those for the surface. More observations for various environmental conditions and cases including the weathers with both single and multiple layered clouds will provide us with more information on the general effects of thin clouds on the measurements of column XCO2 to the surface and on the retrievals of DAOD above thick cloud tops. These results indicate the potential of IM-CW lidar for CO2 measurements over cloudy scenes in space applications.
The authors would like to express their appreciation to D. MacDonnell, D. Garber, D. McGregor, and Y. Hu for their valuable comments and encouragement. This research was supported by the NASA ASCENDS Mission Study and NASA Langley Research Center.
References and links
1. Intergovernmental Panel on Climate Change, Climate Change 2013: The physical science basis, Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change: Chapter 8, G. Myhre, D. Shindel, F. Bréon, W. Collins, J. Fuglestvedt, J. Huang, D. Koch, J. Lamarque, D. Lee, B. Mendoza, T. Nakajima, A. Robock, G. Stephens, T. Takemura, H. Zhan, Eds., Cambridge, UK (Cambridge University, 2013).
2. B. Lin, L. Chambers, P. Stackhouse Jr, B. Wielicki, Y. Hu, P. Minnis, N. Loeb, W. Sun, G. Potter, Q. Min, G. Schuster, and T.-F. Fan, “Estimations of climate sensitivity based on top-of-atmosphere radiation imbalance,” Atmos. Chem. Phys. 10(4), 1923–1930 (2010). [CrossRef]
3. National Research Council, Earth Science and Applications from Space: National Imperatives for the Next Decade and Beyond, Washington, D.C. (National Academies, 2007).
4. J. T. Dobler, F. W. Harrison, E. V. Browell, B. Lin, D. McGregor, S. Kooi, Y. Choi, and S. Ismail, “Atmospheric CO2 column measurements with an airborne intensity-modulated continuous wave 1.57 μm fiber laser lidar,” Appl. Opt. 52(12), 2874–2892 (2013). [CrossRef] [PubMed]
5. B. Lin, S. Ismail, F. Wallace Harrison, E. V. Browell, A. R. Nehrir, J. Dobler, B. Moore, T. Refaat, and S. A. Kooi, “Modeling of intensity-modulated continuous-wave laser absorption spectrometer systems for atmospheric CO(2) column measurements,” Appl. Opt. 52(29), 7062–7077 (2013). [CrossRef] [PubMed]
6. E. Browell, M. E. Dobbs, J. Dobler, S. Kooi, Y. Choi, F. W. Harrison, B. Moore, and T. S. Zaccheo, “First airborne laser remote measurements of atmospheric CO2 for future active sensing of CO2 from Space,” Proceedings of the 8th International Carbon Dioxide Conference, Jena, Germany, 13–18 September 2009.
7. J. F. Campbell, B. Lin, A. R. Nehrir, F. W. Harrison, and M. D. Obland, “High-Resolution CW Lidar Altimetry Using Repeating Intensity-Modulated Waveforms and Fourier Transform Reordering,” Opt. Lett. 39(20), 6078–6081 (2014). [CrossRef] [PubMed]
8. D. Crisp, B. M. Fisher, C. O’Dell, C. Frankenberg, R. Basilio, H. Bösch, L. R. Brown, R. Castano, B. Connor, N. M. Deutscher, A. Eldering, D. Griffith, M. Gunson, A. Kuze, L. Mandrake, J. McDuffie, J. Messerschmidt, C. E. Miller, I. Morino, V. Natraj, J. Notholt, D. M. O’Brien, F. Oyafuso, I. Polonsky, J. Robinson, R. Salawitch, V. Sherlock, M. Smyth, H. Suto, T. E. Taylor, D. R. Thompson, P. O. Wennberg, D. Wunch, and Y. L. Yung, “The ACOS CO2 retrieval algorithm – Part II: Global XCO2 data characterization,” Atmos. Meas. Tech. 5, 687–707 (2012). [CrossRef]
9. F. Harrison, S. Ismail, A. Nehrir, B. Lin, E. Browell, D. McGregor, S. Kooi, J. Dobler, J. Collins, Y. Choi, and M. Obland, “Advances in the Measurement of CO2 using Swept-Frequency, Intensity‐Modulated, Continuous‐Wave Laser Absorption Spectroscopy”, 2013 American Geophysical Union Fall Meeting, San Francisco, CA, 9–13 December, 2013.
10. R. Menzies, G. Spiers, and J. Jacob, “Airborne Laser Absorption Spectrometer Measurements of Atmospheric CO2 Column Mole Fractions: Source and Sink Detection and Environmental Impacts on Retrievals,” J. Atmos. Ocean. Technol. 31(2), 404–421 (2014). [CrossRef]
11. J. B. Abshire, H. Riris, C. J. Weaver, J. Mao, G. R. Allan, W. E. Hasselbrack, and E. V. Browell, “Airborne measurements of CO2 column absorption and range using a pulsed direct-detection integrated path differential absorption lidar,” Appl. Opt. 52(19), 4446–4461 (2013). [CrossRef] [PubMed]
12. V. Malathy Devi, D. C. Benner, L. R. Brown, C. E. Miller, and R. A. Toth, “Line mixing and speed dependence in CO2 at 6348 cm−1: Positions, intensities and air- and self-broadening derived with constrained multi-spectrum analysis,” J. Mol. Spectrosc. 242(2), 90–117 (2007). [CrossRef]