Water vapor is the most dominant of the short-lived greenhouse gases and plays a key role in many atmospheric processes critical to driving Earth’s weather and climate systems (Sherwood et al., 2019). Global and time-resolved measurements of water vapor are important for understanding the hydrological cycle, the thermodynamic structure of the lower troposphere, and broader atmospheric circulation. The 2017 Decadal Survey (DS) for Earth Science and Applications from Space (ESAS) showed a cross-disciplinary need for accurate and high-resolution measurements of water vapor profiles throughout the free troposphere (FT) and the planetary boundary layer (PBL) (National Academies of Sciences, Engineering, and Medicine, 2018). This need was further emphasized by the recent report from the PBL DS Incubation (DSI) study team (Teixeira et al., 2021), which identified key gaps in the Program of Record’s (POR) ability to measure water vapor in the PBL while also summarizing the measurement requirements for an improved understanding of the processes within the PBL across scales. Accurate and high-vertical resolution water vapor profiles and accurate and high-spatial resolution observations of the planetary boundary layer height (PBLH) were some of the top recommendations from both the ESAS DS and DSI study team reports.

Routine measurements of tropospheric water vapor have been made for several decades, and the POR has provided robust observations on a global scale (Teixeira et al., 2021). The PBL DSI report provides an aggregated assessment of the POR capabilities (Table 7-1; Teixeira et al., 2021). The Atmospheric Infrared Sounder (AIRS) (Chahine et al., 2006), Cross-track Infrared Sounder (CrIS) (Bloom, 2001), and Infrared Atmospheric Sounding Interferometer (IASI) (Clerbaux et al., 2009) have provided specific and relative humidity retrievals on the order of 2–3 km vertical and 12–25 km horizontal resolution at 10% accuracy within the PBL (3–7 km in the mid-to-upper troposphere), but they require clear sky and a priori information about temperature and humidity vertical distributions (Maddy and Barnet, 2008; Kahn et al., 2023). The Advanced Microwave Sounding Unit (AMSU) (Aumann et al., 2003) and the Advanced Technology Microwave Sounder (ATMS) (Kim et al., 2014), the IR sounding counterparts, have provided global coverage of specific humidity under clear and cloudy conditions; however, lower degrees of freedom led to even lower vertical and horizontal resolution than those of the infrared sounders. GNSS-RO, an active signal of opportunity counterpart to passive sounders, offers high-resolution profiles of water vapor and temperature (Ao et al., 2012); however, an under-constrained retrieval with coupled moisture and temperature refraction can result in large errors (Nehrir et al., 2017), although emerging microwave sounder/GNSS-RO retrievals show potential to overcome these challenges (Wang et al., 2024). To advance our understanding of the lower-tropospheric process across scales for weather and climate applications, specific humidity profiles with 0.2–1 km vertical and 1–50 km horizontal along track resolutions will be required (Nehrir et al., 2017; Teixeira et al., 2021). Future observing systems that build upon the POR to combine heritage techniques and new active techniques have been identified as potential avenues to meet these requirements (Nehrir et al., 2017; National Academies of Sciences, Engineering, and Medicine, 2018; Teixeira et al., 2021).

Several mature active profiling techniques have been employed for water vapor profiling. Raman lidar exploits the inelastic scattering from water vapor and nitrogen molecules to provide measurements of water vapor and temperature (in addition to aerosol extinction) simultaneously and has ground and airborne heritage (Whiteman et al., 1992; Whiteman et al., 2010; Liu et al., 2014); however, these systems require high-power ultraviolet lasers to overcome the low Raman scattering cross-section and calibration via auxiliary measurements, such as radiosondes. Feasibility studies of Raman lidar for space have been conducted (Di Girolamo et al., 2006; Di Girolamo et al., 2018; Whiteman et al., 2022); however, low subsystem technical readiness and large-power aperture products pose a significant challenge for implementation.

The differential absorption lidar (DIAL) technique, in contrast, directly measures vertical profiles of water vapor with high accuracy without the need for prior information (Schotland R.M., 1965). The DIAL technique relies on elastic scattering to measure the differential attenuation of backscattered signals between absorbing and non-absorbing transmitted wavelengths of laser light. The close spectral and temporal sampling between the online and offline wavelengths yields a direct measure of the vertical profile of water vapor number density with little sensitivity to sources of atmospheric systematic error (e.g., uncertainty in the atmospheric state). Decades of aircraft and surface-based operation (Browell et al., 1979; Ehret et al., 1993; Browell et al., 1998; Wulfmeyer and Bösenberg, 1998; Nehrir et al., 2009; Wirth et al., 2009; Nehrir et al., 2011; Späth et al., 2014; Spuler et al., 2015; Späth et al., 2016; Fix et al., 2019; Bedka et al., 2021; Spuler et al., 2021; Carroll et al., 2022; Millán et al., 2024) have demonstrated a high degree of measurement readiness; however, technological advances are still required to realize DIAL measurement in space (Nehrir et al., 2017; Teixeira et al., 2021).

The feasibility of a space-based DIAL has been assessed from a theoretical perspective (Browell et al., 1985; Ismail and Browell, 1989; Nagasawa et al., 1995; European Space Agency, 2001; Di Girolamo et al., 2004; European Space Agency, 2004; Di Girolamo et al., 2008). Ismail and Browell (1989) studied the systematic and random error and retrieval performance for a 720-nm DIAL system with heritage from the airborne Lidar Atmospheric Sensing Experiment (LASE) (Browell et al., 1998). The DLR WAter vapor Lidar Experiment in Space (WALES) concept (Gerard et al., 2004) explored advanced DIAL capabilities at 935 nm with multiple transmit wavelengths to provide sensitivity from the upper troposphere and lower stratosphere (UT-LS) to the PBL. Studies of random and systematic error (Di Girolamo et al., 2004), retrieval capabilities (Di Girolamo et al., 2008), and comparisons to passive retrievals (Wulfmeyer et al., 2005) supported the WALES concept and provided a foundation for future space-based DIAL that will require multi-wavelength retrievals for global coverage across certain moisture regimes. It should be noted that none of these previous concepts or studies considered a tunable transmitter for variable PBL wavelength pairs, and, as such, PBL measurements from these concepts and studies were only optimized over certain latitudes instead of globally.

The NASA Langley Research Center (LaRC) has been advancing water vapor DIAL technologies to enable the first space-based DIAL measurements of water vapor profiles and total precipitable water vapor with cross-cutting capabilities to measure attenuated backscatter profiles, distributions of the PBLH, surface-weighted water vapor columns (XH₂O), and methane columns (XCH₄) (Nehrir et al., 2021). The novel laser technology being developed for water vapor sensing at 823 nm (Burns et al., 2021) allows for simultaneous measurements of XCH₄ with high accuracy and spatial resolution. Prior studies have shown the feasibility of the 820-nm spectral region for ground-based DIAL (Machol et al., 2004); however, this spectral region is relatively unexplored and lacks assessment for space-based implementation. The development of a space-based water vapor DIAL concept with multiple transmitted wavelengths and the ability to tune spectrally across the absorption line wing is being explored by the NASA LaRC to meet the needs of a future PBL observing system (Nehrir et al., 2021; Teixeira et al., 2021). Multi-wavelength operation and spectral tunability are both required for measurement sensitivity from the UT-LS to the PBL and across latitudinal moisture gradients; however, this functionality also imposes technical challenges (European Space Agency, 2001; European Space Agency, 2004; Wirth et al., 2009; Wulfmeyer et al., 2015; Carroll et al., 2022). Here, we study the feasibility of using the 823-nm water vapor line for space-based DIAL while adhering to the confines of emerging laser technology and spacecraft size, weight, power, and mission cost constraints.

In this paper, we present a study of the systematic and random errors associated with 823-nm water vapor DIAL and assess the feasibility of retrieving range-resolved profiles of water vapor throughout the FT and PBL. For an accurate and complete description of this measurement capability, we use the Atmospheric Boundary Layer Lidar Pathfinder (ABLE) of the NASA LaRC (Nehrir et al., 2021) as a source of instrument parameters, method of operation, and wavelength selection. A detailed instrument description is beyond the scope of this paper and the focus of a future publication. The following sections provide a description of the DIAL methodology, characterization of the sensitivity to the sources of systematic error in the transmitter subsystem, and the sensitivity of humidity retrievals to varying levels of random errors resulting from different averaging schemes.

For a given DIAL wavelength, the number of received signal and noise photons can be described through the elastic lidar equation as follows: N r x λ , r =   N t x λ β a λ , r + β m λ , r   Δ r A r 2   η   e − 2 ∫ 0 r σ λ , r ′ n w v r ′ + α a λ , r ′ + α m λ , r ′ d r ′ + N B λ , where the number of transmitted photons per laser pulse is given by N t x , β a and β m are the Mie (aerosol) and Rayleigh (molecular air) backscatter coefficients (m^-1sr⁻¹), respectively, Δ r is the vertical sampling resolution (m), A is the collection aperture area (m²), r is the distance from the spacecraft to the scattering volume, and η is the unitless total optical efficiency of the transmitter and receiver ( η = η t x η r x . The exponential describes the two-way transmission loss due to water vapor absorption and aerosols and molecular extinction along the measurement path, where n w v is the water vapor concentration (molecules m^-3), σ is the absorption cross-section (m²), α a is the aerosol extinction coefficient (m^-1), and α m is the molecular extinction coefficient (m^-1). The number of noise photons, from the solar background and the detection electronics, is given by N B . The geometric overlap term is assumed to be unity and is not shown.

The water vapor number density in each DIAL averaging range bin, Δ R , can be calculated as follows (Measures, 1984): n w v r = 1 2 Δ R σ on r −   σ off r ln N s λ o n , r N s λ o n , r + Δ R N s λ o f f , r + Δ R N s λ o f f , r + C , where N s is the number of signal photons for each of the online and offline wavelengths over the DIAL range bin Δ R . C is a constant that captures higher-order correction terms and is covered in detail by Bosenberg (1998). In Eq. 2, it is assumed that a mean background has been removed from each backscatter profile, such that N s λ , r = N r x λ , r −   N B λ .

For this study, the pressure, temperature, and water vapor profiles used for the atmospheric state are derived from the Air Force Research Lab standard models (Anderson et al., 1986). To calculate the water vapor differential absorption cross-sections (DCSs), the HITRAN2016 spectroscopic database is used (Gordon et al., 2017), and a Voigt line shape is assumed. Future analysis will utilize the HITRAN2020 database; however, initial investigations have found little change at this spectral region.

Operation at the 823-nm water vapor absorption line is driven by the advance of an Er:YAG laser that provides access to the near-infrared (NIR) region via non-linear optical conversion with relatively high electrical-to-optical efficiency. This laser oscillator is more electrically efficient and employs a simpler architecture than parametric optical conversion, which has commonly been used for airborne DIAL (Wirth et al., 2009; Carroll et al., 2022). The Er:YAG laser is injection-seeded using a low-power semiconductor laser stabilized to a methane absorption line and emits pulsed light at 1,645 nm via optical pumping at 1,532 nm and active q-switching. Frequency doubling in lithium triborate (LBO) provides access to water vapor lines at 823 nm for moisture profiling, while the residual fundamental light coincides with 1,645-nm methane absorption lines for column measurements. Figure 1 shows the methane (blue) and water vapor (black) absorption cross-sections at the fundamental and doubled spectra, respectively, for several altitudes. To provide uniform sensitivity from the mid-troposphere (10 km) to the PBL, three DIAL wavelengths (two wavelength pairs) are required. The successive transmission of three wavelengths allows for a single water vapor DIAL retrieval and single methane columnar measurement employing the integrated path DIAL (IPDA) technique.

The fundamental and frequency-doubled wavelengths of 1,645.5518 and 822.7759 nm set the methane IPDA online and water vapor DIAL offline wavelengths, respectively, while 1,645.8420 and 822.9210 nm set the methane IPDA offline and water vapor DIAL online wavelengths, respectively. A tunable third wavelength near 1,645.75–1,645.825 nm (822.875–822.9125 nm) is utilized only for the water vapor DIAL retrieval to provide latitudinal sensitivity to different PBL regimes. A 2-kHz laser pulse repetition frequency (PRF) transmits 666 shots per wavelength per second with wavelength switching on a shot-to-shot basis with high spectral purity. The remainder of this paper focuses on water vapor DIAL at 823 nm; however, an in-depth analysis of methane IPDA at 1,645 nm can be found in some studies (Ehret et al., 2008; Kiemle et al., 2014; Bousquet et al., 2018; Barton-Grimley et al., 2022), and a detailed analysis of the projected performance of the methane retrieval employing this Er:YAG laser will be the focus of a future publication.

To understand the sensitivity of the DIAL retrieval to different parts of the troposphere, the DCS can be examined. Figure 2 shows the DCS for each line pair over the sideline tuning range, emphasized in Figure 1, calculated with the US standard atmosphere (a–b) and example profiles of the DCS with the bounding moist tropical and dry arctic atmospheres (c–d); Figure 3 provides atmospheric state definitions. The online–sideline DCS (Figure 2A) peaks in the upper FT and, therefore, exhibits higher sensitivity to measurements in this region, with a broadly uniform response to sideline wavelength tuning. Figure 2B shows that the sideline–offline pairs peak in sensitivity in the lower FT, extending into the PBL, and the overall sensitivity is decreased as sideline tuning progresses to shorter wavelengths. Combining the FT and PBL line pairs provides vertical coverage from ∼10 km down to the surface. The methods of line-pair splicing have been explored for space (Di Girolamo et al., 2008) and are routinely implemented from aircrafts (Di Girolamo et al., 2008; Wirth et al., 2009; Carroll et al., 2022). A combination of optical depth and retrieval uncertainty is used to optimally combine wavelength pairs into a single retrieval, where the intermediate regions are filled in using a weighted mean. These methods are operational in the HALO technology testbed instrument, documented by Carroll et al. (2022), and have been evaluated as a robust method for space-based operations.

Sources of error within the DIAL measurement can be systematic, from instrumental and atmospheric states, random, or from background photons and shot noise. The primary sources of systematic errors result from uncertainty in the laser line locking, broadening of the laser line, and spectral impurity (Ismail and Browell, 1989), all relating to uncertainty in the knowledge of the absorption cross-section. Uncertainty in the knowledge of the atmospheric state, i.e., temperature and pressure, can also result in systematic errors, where the magnitude of the error is line dependent. Random errors can originate from shot noise on the backscattered and solar background signals and detector dark counts (Wulfmeyer and Craig, 2001), both serving to decrease retrieval precision. An error in the water vapor number density can be directly evaluated by calculating the absolute error in the DCS for uncertainty in a specific parameter. The same parameter uncertainty can be combined with the atmospheric state to calculate the specific humidity retrieval error. This analysis uses the standard atmosphere at its two extremes, high-humidity tropical (TRO) and low-humidity sub-Arctic winter (SAW) (Figure 3), to quantify errors with respect to retrievals of specific humidity.

Uncertainty in the DIAL retrieval can arise from systematic (correlated) and random (uncorrelated) errors. Sources of systematic error can stem from the transmitter and receiver subsystems, such as laser frequency uncertainty or detection non-linearity, uncertainty in assumed atmospheric temperature and pressure states, and absorption cross-section accuracy. The random error discussed here is dominated by noise derived from solar background photon counts, detector dark counts, and shot noise. The simulation results given in Section 3 indicate that the systematic and random error magnitude resulting from the current space-based water vapor DIAL technology developments is within the requirements laid out by several community reports (National Academies of Sciences, Engineering, and Medicine, 2018; Teixeira et al., 2021). The requisite instrument specifications (power aperture product and laser spectral fidelity) needed to achieve the required SNR and retrieval performance are being developed. Additional future technology developments and maturation (increased laser power and detection efficiency) will provide increased performance to further increase science capabilities.

Table 1 and Figure 9 provide an encompassing view of the threshold and baseline systematic error. The threshold characteristics of the Er:YAG transmitter considered, i.e., a laser linewidth uncertainty of <250 MHz, center frequency stability of ± 100 MHz, and a spectral purity of 99.5% give a maximum systematic error of 2.317%, could be achieved over the 822.875–822.921-nm wavelength range with a minor peak of 1.833% at 822.9145 nm. Figure 9A shows a minimum error of 1.25%, providing upper and lower bounds (the tropical atmosphere exhibited the highest cumulative error due to the higher moisture loading). The baseline configuration with <100 MHz, center frequency stability of ± 5 MHz, and a spectral purity of >99.9% yield a maximum systematic error of 1.395% over the 822.875–822.921-nm wavelength range with a minor peak of 1.119% at 822.921 nm. Figure 9 shows that the baseline parameters have a flat error response to sideline tuning, <0.1%, whereas the threshold parameters exhibit a non-linear error response and have a change of >0.5% between the polar and tropical wavelength operating points. Combined, the estimated systematic error for all wavelengths of potential operation is below the baseline and threshold requirements provided by Nehrir et al. (2017) for nominal instrument parameters.

Spectral impurity of the laser source can be a prominent contributor to systematic errors; however, Figure 6 provides an indication that irrespective of the selected absorption line, a systematic error of <2% (considering spectral purity of >99.5% and a one-way optical depth of <1.5) will be present. DIAL retrieval sensitivity to multi-mode operation is largest when operating on the side of the absorption line near the line peak (Figure 7). For these conditions, the worst-case scenario of operating on three laser longitudinal modes (fundamental and two higher-order modes) results in a <2% error at the surface. Figure 9 and Table 1 assume the worst-case error of multi-mode operation to yield a peak error of 1.90% at 822.875 nm and a local max of 1.61% at 822.921 nm, both at the surface level. High spectral purity is evident for the Er:YAG system discussed here as stringent spectral purity and frequency stability requirements for methane IPDA at 1,645 nm are >99.9% and 10 MHz, respectively, such that the water vapor requirements will inherently be met. Additionally, by utilizing second harmonic generation to access the 823-nm spectral band for water vapor DIAL (Burns et al., 2021), any spectral impurity resulting from higher-order modes at the fundamental will inherently be reduced at the frequency-doubled output.

Doppler broadening of the backscatter spectra is of low impact for lower-tropospheric DIAL when utilizing sideline wavelengths. However, large aerosol gradients, such as those found in the Saharan air layer, could require corrections to minimize retrieval error, which can be performed using attenuated backscatter measurements provided at the “offline” wavelength (Späth et al., 2020). The post-correction residual error in specific humidity is assessed to be within 1%–2% across the entire wavelength tuning range. Errors resulting from the uncertainty in the atmospheric state inputs from reanalysis products for the calculation of the DCS was found to be <2% and <0.1% for temperature and pressure, respectively, which were within the error budget. Additional systematic contributions resulting from errors in the lidar optical system, such as misalignment between the transmitter and receiver and detection non-linearity between the onside and offline channels, were estimated at a conservative constant 2%. Dedicated engineering can readily reduce these errors to <1%.

Section 3.2 shows that random error is primarily dominated by shot noise on the backscattered photon counts and solar background noise. This error has strong dependence on the wavelength and the vertical and horizontal averaging scales. The system specifications (Table 2) (Nehrir et al., 2021) utilized here to constrain the instrument performance represent the current state-of-the-art for space-based DIAL employing low-energy and high PRF lasers and offer a feasible path to space. Simulated performance over ocean under daytime conditions, a shot noise-limited case, for different vertical and horizontal resolutions over the ocean during daytime conditions ( Δ R = 300 m and Δ R = 1,000 m for the lower and mid/upper troposphere, respectively, and Δ H = 50  km) results in a random error of <10% near the surface and <10% in the mid- and upper troposphere (dependent on moisture loading). As the along-track averaging is reduced to Δ H = 5 km, a surface-level random error of 5%–20% is expected for Δ R = 300 m and >50% for Δ R = 100 m.

Considering an enhanced instrument configuration that employs an Er:YAG amplifier to increase the doubled pulse energy from 3.5 to 6 mJ at 2 kHz PRF and an increased collection aperture diameter of 1 m, a significant increase in daytime performance can be achieved. Figure 15 shows the surface-level retrievals for varying vertical and horizontal resolutions under daytime conditions over open ocean, A = 0.06, and a bright surface, A = 0.5. For Δ R = 300 m, in Figures 14A, C a 5% retrieval error is achieved with Δ H = 25 km, with a 10-km relative increase in along-track resolution for the same error. At this vertical resolution, the 10% baseline error bound is reached for all potential along-track averages >8 km. The largest benefit is observed over high-albedo surfaces due to the relatively higher SNR afforded by the increased power aperture. Figure 15 shows that threshold requirements of 20% error are met for 300 m and >10 km and 200 m and >25 km under conditions that resemble daytime cloudy regions. Collectively, this indicates that improved instrument performance could be met in response to science community needs if mission requirements dictate the potential for a larger and more accommodating satellite bus.

The systematic and random error analysis presented here shows that a space-based DIAL with realistic performance metrics operating at the 823-nm line can meet the measurement requirements for high vertical resolution and accurate profiling of water vapor in the lower–middle troposphere. The technology currently being advanced to enable a space-based water vapor DIAL mission will meet the threshold and baseline requirements laid out by the PBL study team report (Teixeira et al., 2021) and the 2017 Decadal Survey (National Academies of Sciences, Engineering, and Medicine, 2018). In this paper, we showed that tunable operation along the side of the 823-nm absorption line allows for the optimization of the lower-tropospheric water vapor retrievals across different meteorological regimes and latitudes and provides the operational flexibility needed to dynamically optimize random errors and trade precision for the spatial resolution to accommodate different scientific applications. We show that specific humidity measurements with a <10% error can be achieved at 200 m by 50-km and <5% at 300 m by 50-km vertical and along-track resolutions under daytime conditions over open ocean. Combined, the total error budget demonstrated here indicates high impact for space-based DIAL with technologies being advanced for space missions within the next 5–10 years.