1 Introduction

Carbon dioxide, methane, and water vapor, as important greenhouse gases (GHG), play significant role in the global warming [1,2,3,4]. Carbon dioxide is responsible for 55–60% of current anthropogenic radiative forcing on warming impact [1, 2]. Methane as a powerful greenhouse gas accounts for around 30% of the rise in global temperatures since the industrial revolution [3]. Water vapor, known as the most abundant greenhouse gas, can amplify the warming caused by other greenhouse gases [4]. Precise measurements of GHG concentration, especially the vertical profile in the atmosphere with sufficient spatial and temporal resolution, can well advance our understanding of the climate forcing processes.

Many strategies based on different techniques have been developed for remote sensing of vertical concentration profiles of atmospheric trace gases. These observations include space-based platform, such as satellite platform [5], aircraft (or balloon) platform [6], and ground-based observations platform, including Fourier transform spectrometer [7], and laser heterodyne radiometer (LHR) [8]. Satellite platforms provide the advantages of covering large scales over the earth, though they suffer from limited precision and limited observation capability at lower altitudes [9]. The aircraft (or balloon) platform well fills the gap between ground- and space-based observations [10], but it cannot provide continuous measurements due to limited flight emissions [11]. FTIR method, as a promising complement to the ground-based observations with high precision and accuracy, is heavyweight with high cost and large dimensions [12]. Further improvement of spectral resolution in FTIR will increase the instrument volume. The LHR method can provide long-term and continuous remote measurements of trace gases in terms of spectral resolution, sensitivity, and transportability compared to FTIR technique [13,14,15].

LHR is based on extracting absorption information from a broadband light by beating with a narrow-band local oscillator. The LHR method is capable of providing high spectral resolution up to ~ 10–4 cm−1 [13]. In recent decades, LHR technology has been rapidly developing due to the maturity of tunable lasers and high-speed detectors. Many previous investigations have reported retrieval of the vertical profiles of atmospheric CH4, CO2, and H2O based on near-infrared and mid-infrared LHR approaches. Wang et al. developed an ICL-based LHR sensor to measure the vertical profiles of atmospheric CH4 and H2O at 3.5 µm [16]. Weidmann et al. used an external cavity cascade laser as local oscillator in a wavelength range from 8.08 to 8.93 µm for measuring atmospheric ozone (O3), nitrous oxide (N2O), methane (CH4), water vapor (H2O), and dichlorodifluoromethane (CCl2F2) [14]. However, the laser source in the mid-infrared region is extremely costly. For reducing the cost, distributed feedback (DFB) diode laser in the 1.2–1.7 µm wavelength range is suitable for atmospheric sensing, as reported by Wilson et al. [17] and Deng et al. [12]. Deng et al. coupled three DFB lasers at 1.277, 1.571, and 1.654 µm to realize quasi-simultaneous measurements of atmospheric O2, CH4, CO2, and H2O. Multiple DFB diode lasers are separately controlled by time-division multiplexing ramp signals for scanning the absorption lines of O2, CH4, CO2, and H2O. The detection efficiency of an LHR sensor using multiple diode lasers operated in different periods might be influenced by rapid and random fluctuation of solar radiation due to clouds [18,19,20].

For improving the detection efficiency using compact sensor equipment, here, we propose an all-fiber, near-infrared LHR based on two DFB diode lasers emitting at 1653 and 1568 nm as local oscillators for the simultaneous measurement of atmospheric gases (CH4, CO2, and H2O) within one scanning cycle. The two DFB diode lasers were simultaneously operated within a scanning period to further simplify the system structure of the LHR sensor. The duration of one scanning cycle was experimentally determined to be 2 s. An atmospheric transmission spectrum of multi-gas species was recorded within 20 s after averaging ten times, which highly improved the detection efficiency of our LHR sensor. The light powers of the two local oscillators were finely regulated to be mostly identical for generating nearly the same interference on the LHR signal, for further removing the measurement interference by another local oscillator and accurately retrieving the absorption spectra of atmospheric CH4, CO2, and H2O. An intercomparison was carried out to validate the developed LHR sensor compared with the measurement data by the Japanese GOSAT satellite [15, 21].

2 Measurement principle

The optical path in laser heterodyne radiometer fundamentally consists of three channels including the solar radiation path, an output radiation from local oscillator, and a light coupler for generating a heterodyne signal. The LHR method converts the infrared light at the frequency of 1014 Hz to the radio frequency signal at 108 Hz [13]. In the LHR system, the electromagnetic field of the solar radiation ES(t) and the local oscillator (ELO(t)) can be expressed as [11, 22]

$$ E_{{\text{S}}} \left( t \right) = A_{{\text{S}}} {\text{cos}}\left( {\omega_{{\text{s}}} t + \varphi_{{\text{s}}} } \right) $$
(1)
$$ E_{{{\text{LO}}}} \left( t \right) = A_{{{\text{LO}}}} {\text{cos}}\left( {\omega_{{{\text{LO}}}} t + \varphi_{{{\text{LO}}}} } \right), $$
(2)

where As, ωs, and φs are the amplitude, frequency, and phase of the solar radiation, respectively. ALO, ωLO, and φLO are the amplitude, frequency, and phase of the local oscillator, respectively. The light signal with the two lights mixing on a square-law photodetector is proportional to the total optical intensity that is the square of the total electric field amplitude E(t) = ES(t) + ELO(t) as expressed [20, 23]

$$\begin{aligned} \mathrm{i} & =\mathrm{\alpha }{\left({\mathrm{E}}_{\mathrm{S}}+{\mathrm{E}}_{\mathrm{LO}}\right)}^{2}=\mathrm{\alpha }{\left[{\mathrm{A}}_{\mathrm{S}}\mathrm{cos}\left({\upomega }_{\mathrm{s}}\mathrm{t}+{\mathrm{\varphi }}_{\mathrm{S}}\right)+{\mathrm{A}}_{\mathrm{LO}}\mathrm{cos}\left({\upomega }_{\mathrm{LO}}\mathrm{t}+{\mathrm{\varphi }}_{\mathrm{LO}}\right)\right]}^{2} &=\alpha \left\{{A}_{S}^{2}\frac{1+\mathrm{cos}2\left({\upomega }_{\mathrm{s}}\mathrm{t}+{\mathrm{\varphi }}_{\mathrm{S}}\right)}{2}+{\mathrm{A}}_{\mathrm{LO}}^{2}\frac{1+\mathrm{cos}2\left({\upomega }_{\mathrm{LO}}\mathrm{t}+{\mathrm{\varphi }}_{\mathrm{LO}}\right)}{2}+{\mathrm{A}}_{\mathrm{S}}{\mathrm{A}}_{\mathrm{LO}}\mathrm{cos}\left[\left({\upomega }_{\mathrm{S}}+{\upomega }_{\mathrm{LO}}\right)\mathrm{t}+{\mathrm{\varphi }}_{\mathrm{S}}-{\mathrm{\varphi }}_{\mathrm{LO}}\right] +{\mathrm{A}}_{\mathrm{S}}{\mathrm{A}}_{\mathrm{LO}}\mathrm{cos}\left[\left({\upomega }_{\mathrm{S}}-{\upomega }_{\mathrm{LO}}\right)\mathrm{t}+{\mathrm{\varphi }}_{\mathrm{S}}-{\mathrm{\varphi }}_{\mathrm{LO}}\right]\right\},\end{aligned}$$
(3)

where α is the photodetector’ sensitivity. The high-frequency components (at 2ωs, 2ωLO and ωs + ωLO) and the constant DC components are filtered out by the radio frequency electronics. The heterodyne beat signal at the intermediate frequency (ωs-ωLO) is proportional to the product of the electric field amplitudes (AL0 × AS). The weak absorption information of trace gas in the solar radiation can be amplified by the high light power of the local oscillator. The retrieving spectra of the target gas are achieved by tuning the frequency of the local oscillator.

3 Experimental setup

Figure 1 shows the schematic of the developed LHR instrument with two DFB diode lasers emitting at 1571 nm (DFB1) and 1653 nm (DFB2) as local oscillators. A sun tracker (EKO Instruments CO, LTD, Model STR-21G) was used to track the solar position. The solar radiation was chopped at a modulation frequency of 500 Hz by a mechanical chopper before collected by a fiber collimator (Thorlabs, RC08FC-P01). The solar radiation was then coupled with the light emitted from the local oscillators using a fiber coupler. The light from the two DFB lasers was coupled together after being passed through a fiber optic attenuator for avoiding the saturation of heterodyne signal detected by a photodetector (Thorlabs, PDA10CF-EC). The light signal from the photodetector was processed by several radio frequency electronics, including RF amplifiers (Mini- Circuits, ZX60), a bandpass filter (Mini-Circuits, BBP-30+), and a Schottky diode (HEROTE, DHM020BB). The heterodyne signal was finally demodulated at the modulation frequency of 500 Hz using a lock-in amplifier (Stanford Research Systems, SR830) and was sampled by a data digitalization card (National Instrument, model USB-6361). A laptop was used to perform data processing and result display.

Fig. 1
figure 1

Schematic of the developed LHR instrument. DAQ: data acquisition. RF: radio frequency

Full size image

For realizing simultaneous measurement of CO2, CH4, and H2O, all absorption lines should be covered through scanning the wavelength of the two DFB lasers. In this work, the two DFB lasers (DFB1 and DFB 2) were controlled by a sawtooth signal, as shown in Fig. 2a. The absorption lines of CO2, CH4, and H2O were simultaneously detected in one period of the sawtooth signal, as shown in Fig. 2b. The CO2 absorption line was detected by DFB1 in the first half-period of the swath signal. In the latter half-period of the swath signal, the CH4 and H2O absorption lines were scanned by DFB2 laser. Figure 2b shows the simulation transmission spectrum of CO2 in 1571–1571.4 nm, and H2O and CH4 in 1653–1653.8 nm by the reference forward model originally developed the Department of Physics in Oxford University [24]. The input parameters of atmospheric temperature and pressure profile were downloaded from National Centers for Environmental Prediction with temperature from 298 to 212 K and pressure from 1 to 0.1 atm from surface to 78 km [15]. The absorption spectra of CO2, CH4, and H2O were taken from HITRAN database [25]. The transmittance of multiple gases shows no cross talk and is theoretically supported by the spectrum database of HITRAN. The tunable wavelength ranges of DFB1 and DFB2 were controlled to be 1571–1571.4 nm and 1653.2–1653.8 nm, respectively. The absorption of CO2, CH4, and H2O was scanned in sequence, as shown in Fig. 2. The duration of one scanning cycle was experimentally determined to be 2 s with high signal-to-noise ratio after compared with the duration of 1 and 5 s as shown in the supplementary file. This approach significantly improves detection efficiency by avoiding the rapid and unpredictable fluctuations of the solar radiation intensity due to clouds.

Fig. 2
figure 2

a Dependence of the emitting wavelength on the current of DFB1 and DFB2 diode lasers within one scanning period. b Simulation transmission spectrum of CO2 in 1571–1571.4 nm, and H2O and CH4 in 1653.2–1653.8 nm using the reference forward model [23]

Full size image

Two local oscillators were used in the developed LHR sensor for generating heterodyne signal. The heterodyne signal of each molecule can be interfered by another local oscillator. During scanning the frequency of DFB1 laser, the amplitude of CO2 heterodyne signal is composed of the intrinsic CO2 heterodyne signal generated by DFB1 and the supplemental heterodyne signal generated by DFB2. The CO2 absorption line was not scanned by the DFB2 laser, and therefore, the DFB2 laser would affect the baseline of CO2 heterodyne signal, which is shown in Sect. 4.2. The DFB1 laser would generate the same interference on the baseline of CH4 and H2O heterodyne signal in the same way. As shown in Eq. (3), the heterodyne signal is proportional to the light power of the local oscillator. The supplemental heterodyne signal (as the measurement interference generated by another local oscillator) depends on the light power of the local oscillator. The identical light powers of the two local oscillators could generate the same supplemental heterodyne signal that can be removed by dividing the total heterodyne signal in two times in the data process. It is necessary to record the light power ratio of the two local oscillators to determine the proportion of the supplemental heterodyne signal generated by the corresponding laser.

In this work, the essence of multi-gas detection is to use a single tunable signal to simultaneously control the multiple laser sources. A sawtooth scanning signal with amplitude of 1 V was generated by the DAQ card and was split into two channels for controlling two laser driver modules: ITC102 for DFB 1 laser and LDC105 for DFB 2 laser. The modulation coefficients of ITC102 and LDC105 are 40 mA/V and 50 mA/V, which leads to a current sweep range of 20–120 mA and 30–110 mA for DFB 1 laser and DFB 2 laser, respectively. The temperatures of DFB 1 and DFB 2 laser source were set at 25 and 35 °C, respectively. With precise temperature control of laser source as 0.006 °C, the absorption lines of each component were accurately scanned, so that the multiple absorption peaks do not crosstalk.

4 Experimental results

4.1 Individual measurements of CH4, CO2, and H2O

The developed LHR senor was tested by individual measurement of CO2, CH4, and H2O. Figure 3a and b shows the raw individual experimental spectral of atmospheric CO2, CH4, and H2O. The path-integrated absorption of CO2 in Fig. 3a was measured by DFB 1 laser source with a temperature set at 25℃ and a wavelength scanning range of 1571–1571.35 nm. The measured CH4, H2O spectra are shown in Fig. 3b and detected by DFB2 with a temperature set at 35 °C and a wavelength scanning range of 1653.2–1653.80 nm. The sawtooth scanning signal was generated DAQ card with a scanning frequency of 0.5 Hz, leading to a period of 2 s for one scanning cycle. The total scanning duration was determined to 20 s after averaging ten times. The intrinsic absorption spectrum of solar radiation was observed near the CH4 absorption line at 1653.7 nm, which was removed by a multi-peak fit method for data process.

Fig. 3
figure 3

Individual experimental heterodyne-detected atmospheric transmission spectra of a CO2, b CH4, and c H2O

Full size image

The LHR sensor was validated by continuous measurement of atmospheric CO2 in different time scale from 14:00 to 16:10. Figure 4 shows the representative transmittance spectra of atmospheric CO2 from 14:00 to 16:10. Regarding the variation of the solar radiation intensity and the optical path, the raw experimental LHR signal was processed by an intensity normalization method. As shown in Fig. 4, the absorption depth became deeper and the observed peak values were shifted from 0.47 to 0.14 at time from 14:00 to 16:10 with the solar zenith angel increasing form 44.36º to 68.11º. It is well consistent with that the absorption path of the solar radiation is longer with the solar zenith angle increasing [22, 26]. Therefore, the LHR concentration inversion algorithm needs to take into account the experimental time and the solar zenith angle to accurately determine the gas concentration. A typical signal-to-noise ratio of 407(= 0.53/0.0013) for the CO2 transmittance spectra at 14:00 was achieved by dividing the absorption depth (0.53) to the standard deviation (0.0013) of the spectrum baseline, which is sufficient for retrieving the CO2 vertical distributions in the atmosphere. For the data process, it should be noted that the laser frequency might slightly drift caused by the temperature change in the field campaign and should be accurately calibrated. The position of peak value in the measured transmittance spectrum was compared with a reference signal calculated by the reference forward model [24]. A moving step for the spectra peak was determined to calibrate the laser frequency.

Fig. 4
figure 4

Experimental CO2 transmission spectra from 12:00 to 16:00 with the solar zenith angel increasing form 44.36º to 68.11º

Full size image

4.2 Simultaneous measurements of CH4, CO2, and H2O

For realizing simultaneous measurement of CO2, CH4, and H2O, two DFB lasers were coupled together by a fiber optic coupler. LHR signal highly depends on the light power of the local oscillator and the solar radiation. LHR signals excited by the two local oscillators were simultaneously processed by the RF elements, which leads to the supplemental heterodyne signal as described in Sect. 3. For eliminating the measurement interference (the supplemental heterodyne signal) by another local oscillator, the individual background photonic signal at the photodetector (Thorlabs, PDA10CF-EC) excited by each DFB laser source was finely controlled to be close, which generate the same interference on the background LHR signal. The same interference is easily removed in the data process. The photonic signals at the photodetector were finely controlled by regulating the laser power by the two fiber optic attenuators. Figure 5a shows the individual background photonic signals excited by the coupled light between the DFB laser source and the solar radiation, which was directly irradiated on the photodetector without passing through the RF elements. As shown in Fig. 5a, the background photonic signals were nearly same with the current range from 30 to 110 mA in DFB 1 and the current range from 30 to 110 mA in DFB 2 after regulating the laser power by the attenuators. The black line in Fig. 5a presents the merge signal intensity with the two DFB lasers simultaneously irradiating on the photodetector, which is the sum of the individual background signal.

Fig. 5
figure 5

a Two LHR background signals excited by DFB laser are close after finely regulating the laser power without passing through the RF elements. b Separately LHR signal of CO2, CH4, and H2O, and its summation of LHR signal. The baseline (grey) is fitted by third-order polynomial function. c Experimentally determined transmission spectrum with the LHR sensor individually or simultaneously measuring atmospheric CO2, CH4, and H2O

Full size image

As shown in Fig. 5a, the background photonic signals at the photodetector cannot be controlled to be completely same. However, the baselines (red for CO2 and blue for CH4, H2O) of LHR signal shown in Fig. 5b were nearly same after finely regulating the laser power by the fiber optic attenuators. The black line in Fig. 5b was the total LHR signals of CO2, CH4, and H2O simultaneously excited by the two DFB lasers. The absorption baseline of the merge LHR signal (grey in Fig. 5b) was fitted by a third-order polynomial function. The baseline of the CO2 LHR signal (red line) and the CH4, H2O LHR signal (blue line) are identical. The baseline intensity of the merge LHR signal in grey in Fig. 5b is two times of the baseline intensity of the individual LHR signal.

Figure 5c shows the accurate transmission spectra of CO2 (red line), and CH4 and H2O (blue line), which can be used for retrieving the vertical profile of CO2, CH4, and H2O. The black line in Fig. 5c is the determined transmission spectra of CO2, CH4, and H2O by the merger LHR signals simultaneously excited by the two DFB lasers without taking into account the supplemental LHR signal excited by another oscillator. As shown in Fig. 5c, the absorption depth of individual LHR signal (black and blue in Fig. 5c) is two times of the absorption depth determined by the merge LHR signal (black line). Table 1 summarizes the absorption depth of the CO2, CH4, and H2O LHR signal with the LHR sensor individually or simultaneously measuring atmospheric CO2, CH4, and H2O. With finely regulating the laser power, the background LHR signals excited by the two DFB laser sources can be set identically. The accurate absorption depth of trace gases is two times of the retrieved absorption depth determined by the merge LHR signal in this developed LHR sensor. The fluctuation of the solar radiation can affect the amplitude of LHR signal, but with a same effect on the LHR signal CO2, CH4, and H2O, which will not influent the multiple of two times.

Table 1 Summary for the absorption depth of CO2, CH4, and H2O LHR signal with the LHR sensor individually or simultaneously measuring atmospheric CO2, CH4, and H2O
Full size table

4.3 Data retrieval and intercomparison with GOSAT satellite

4.3.1 4.3.1 Data retrieval

The experimental LHR signal is used to convert the atmospheric vertical concentration profile of the target gas. The basic principle of the inversion process algorithm is to use the forward RFM model with different vertical concentration profiles to simulate the theoretical LHR signal. A minimum deviation between the simulated LHR signal and the experimental LHR signal is obtained once a best fit of the vertical concentration profile was found using the inversion process algorithm. The core of the inversion iteration is based on Levenberg–Marquardt (LM) iteration which continuously updates the new concentration profile data [13, 14]. With the experimental LHR signal (y), the iterative state vector (the vertical profile) xi+1 is determined using the following equation based on Levenberg–Marquardt (LM) iteration [15, 18, 27]:

$${x}_{i+1}={x}_{i}+{\left[\left(1+\gamma \right){S}_{a}+{K}_{i}^{T}{S}_{\varepsilon }^{-1}K\right]}^{-1}\times \left[{K}_{i}^{T}{S}_{\varepsilon }^{-1}\left(y-{F}_{i}\right)-{S}_{a}^{-1}({x}_{i}-{x}_{a})\right],$$
(4)

where xa is a priori profile with an a priori covariance matrix Sa, and Sɛ is the measurement covariance matrix. K is the Jacobian matrix and F is a radiative transfer forward model. The relationship between the experimental LHR signal (y) and the forward model can be described as y = F(x) + ɛ, where ɛ is the error vector. In the inverse program, the equation was iteratively calculated to minimize a cost function χ2 defined as

$${\chi }^{2}={(y-F)}^{T}{S}_{\varepsilon }^{-1}\left(y-F\right)+{{\left({x}_{i}-{x}_{a}\right)}^{T}S}_{a}^{-1}\left({x}_{i}-{x}_{a}\right).$$
(5)

The optimal atmospheric concentration profile of the target gas is used as an approximation to the actual atmospheric concentration profile data, once the minimum of the cost function was determined.

The experimental heterodyne signal is the convolution of the actual atmospheric transmittance spectra with the instrument line shape function (ILSF). The ILSF highly depends on the bandwidth of the narrow-linewidth DFB laser, RF elements, and especially the low-pass filter of lock-in amplifier [22, 28, 29]. The linewidth DFB laser is 2 MHz and the center frequency of the bandpass filter is 30 MHz, which are narrower than any absorption linewidth of trace gas. It is difficult to determine the ILSF through measuring each bandwidth of all elements. The ILSF was directly determined by the heterodyne signal between two narrow-linewidth infrared laser beams from two DFB diode lasers, where one beam is at a fixed wavelength and the wavelength of another beam was scanned across the fixed wavelength [21]. Figure 6 shows the determined ILSF using two DFB diode lasers emitting at 1571 nm with different integration time of the low-pass filter in the lock-in amplifier. As shown in Fig. 6, the bandwidth of ILSF was broadened with long integration time constant of lock-in amplifier. The long integration time constant of 100 ms can cause the distortion of the heterodyne signal. The ILSF presents bimodal phenomenon with short integration time constant of 3 ms, which might be due to the unstable of the fixed emitting wavelength at 1571 nm. The long integration time constant in the LHR sensor can provide better instrument performance. Therefore, the ILSF with an integration time constant of 30 ms was used in the LHR sensor with a half width at half maximum of 0.08 cm−1.

Fig. 6
figure 6

Instrument line shape function with different integration time in the low-pass filter of lock-in amplifier

Full size image

Figure 7a–c shows the heterodyne-detected spectra after intensity normalization and the fitted spectra of CO2, CH4, and H2O, which were recorded at 13:00 on 24th, October, 2021. Finally, the residuals within 0.02 between the processed spectra and the fitted result indicate a well configured fit. Figure 8a–c plots the typical priori profile of CO2, CH4, and H2O in the measurement site (Ningbo, China) downloaded from European Center for Medium-range Weather Forecasts and the retrieved CO2, CH4, and H2O profiles. The surface-level concentration of CO2, CH4, and H2O are determined to be 416 ppm, 1.94 ppm, and 8450 ppm.

Fig. 7
figure 7

Heterodyne-detected spectra of a CO2, b CH4, and c H2O and the fitted spectrum. The residuals are within 0.02

Full size image
Fig. 8
figure 8

Typical a priori profile of a CO2, b CH4, and c H2O in Ningbo city and the retrieval results

Full size image

4.3.2 4.3.2 Intercomparison with GOSAT satellite

Simultaneous measurements of CO2 and CH4 by the LHR sensor was tested and validated by an intercomparison measurements compared with the Japanese GOSAT [21] satellite from October to November, 2021. The intercomparison measurements of outdoor CO2 and CH4 were carried out in Ningbo, China (121°34′ E, 29°48′ N). Effective measurement data in 7 days from GOSAT near the longitude and latitude of Ningbo (121°34′ E, 29°48′ N) within a deviation of 6° radius were selected during October–November, 2021.

Figure 9a shows time-series measurements (averaged at one day) of outdoor CO2 and CH4 using the LHR sensor and the Japanese GOSAT satellite over 7 days. The measurements were taken daily between 11:00 and 15:00, when the gas concentration was stable and was not affected by traffic circulation in the morning and evening peak. Each measurement was taken every 30 min and the duration was 10 min. One experimental spectrum was obtained within 20 s after averaging ten times. The LHR signal decreases due to the influence by cloud contamination and would be no longer considered for further analysis. The concentrations of CO2 and CH4 surface level are determined to be around 415 ppm and 1.9 ppm, which are consistent with the measurement result by the Japanese GOSAT satellite. Some measurement results are biased compared with GOSAT satellite, which are probably caused by large difference of longitudes and latitudes that are covered by GOSAT compared with the longitudes and latitudes of the Ningbo area. The temperature and atmospheric pressure on the specific day are not consistent with the data used by GOSAT, which might cause the measurement deviation. Finally, the consistence of measurement results between the LHR sensor and the Japanese GOSAT well validated our LHR sensor.

Fig. 9
figure 9

Time-series measurements (averaged at 1 day) of outdoor a CO2 and b CH4 using the LHR sensor and the Japanese GOSAT satellite over 7 days

Full size image

5 Conclusions

An all-fiber near-infrared laser heterodyne radiometer was developed by combining two DFB lasers emitting at 1571 nm and 1653 nm. The LHR sensor is capable of simultaneously measuring the multiple gases within one scanning cycle, which is greatly improving the detection efficiency of the sensing system. The background photonic signals, at the photodetector, separately excited by the two local oscillators, were finely controlled to be mostly identical, which is favor of removing the measurement interference on the heterodyne signal excited by each local oscillator. The vertical concentration profiles of CO2, CH4, and H2O in the measurement site were experimentally determined by an inversion process algorithm based on Levenberg–Marquardt method. The measurement of surface-level CO2 and CH4 concentration are around 415 ppm and 1.9 ppm, which was validated by an intercomparison campaign in comparison with the Japanese GOSAT satellite. The developed LHR sensor presents a simple structure based on laser heterodyne radiometer for simultaneous measurement of multi-gases in a short scanning cycle.