Earth observations from space began over 60 years ago (National Academies of Sciences, Engineering, and Medicine 2008; Thies and Bendix 2011). They have facilitated and accelerated an unprecedented progress in understanding and quantifying key Earth system processes. These observations have provided a unique global perspective on fundamental components of the Earth system, including the structure of the atmosphere, weather patterns and extremes, the ozone layer, sea level change, the water and carbon cycles, aerosol emissions, ice sheets, and Earth’s radiative energy budget. Furthermore, the accuracy of short- and medium-range weather forecasts has improved by more than 50% in the satellite era owing to the constantly increasing number and quality of observations and advancements in data assimilation techniques (Bauer et al. 2015; Eyre et al. 2020, 2022). Obtaining new and valuable information requires development of more advanced technologies and retrieval approaches. The importance of such developments was recently stressed in the National Academies of Sciences, Engineering, and Medicine (NASEM) 2017–27 Decadal Survey for Earth Science and Applications from Space (ESAS 2017).1 ESAS 2017 identified new key targeted observables essential for further advancing our knowledge of the Earth system. The atmospheric planetary boundary layer (PBL) was among these key observables.
The PBL is the lowest atmospheric layer, strongly affected by the diurnal cycle of solar insolation, where complex nonlinear interactions between the atmosphere, ocean, land, cryosphere, and biosphere occur at a wide range of spatial and temporal scales. The richness of possible PBL states, the intermittency of small-scale fast-evolving 3D PBL structures, the potential for large gradients, the proximity of the surface, and a typically low thickness of that layer make it extremely difficult to measure from space. On the other hand, the PBL is considered a high-priority observable since it cuts across several major themes identified by the panels who developed the ESAS 2017, including the Weather and Air Quality, Climate, Water Resources, Ecosystem, and Integrating themes. In particular, the PBL was the most important objective for the Weather and Air Quality panel of the ESAS 2017. This is because PBL is at the heart of fundamental science challenges: cloud–climate feedback, severe storm prediction, the exchanges of energy, water, and carbon between the atmosphere, ocean, land, and ice through PBL turbulent fluxes, and finally the depth and mixing of PBL air significantly influence air quality and human health.
Since current spaceborne instruments are strongly limited in their sensitivity to the PBL, the ESAS 2017 recommended the incubation of future mission concepts for probing the 3D PBL temperature and water vapor structure and the 2D structure of PBL height as they are critically important for understanding the underlying PBL processes. Following that ESAS 2017 recommendation, the National Aeronautics and Space Administration (NASA) PBL Incubation Study Team Report (Teixeira et al. 2021) identified “(i) the most critical PBL science questions and applications topics in the context of Earth system science, (ii) specific PBL needs from a data assimilation, modeling and prediction perspectives, (iii) the critical geophysical observables and their associated spatial and temporal measurement requirements so as to address the key PBL science questions and applications topics, (iv) the observational gaps from the current program of record, and (v) practical yet effective emerging measurement approaches and technologies to address measurement requirements from space using a range of system architectures.”
According to the PBL report, essential components of a future global PBL observing system could include differential absorption lidar and radar in low-Earth orbit (LEO), hyperspectral infrared and microwave sounders in LEO, radio occultation, and geostationary hyperspectral infrared sounding. Recommended measurement requirements include resolutions of 200 m in the vertical, 1 km in the horizontal, and 1–4 h in time. The PBL Incubation Study Team Report also highlighted the essential role of multiscale modeling in developing future PBL observing systems.
Following the report, our team at the Jet Propulsion Laboratory (JPL) started exploring in simulation mode the capability of some instrument technologies in targeting small-scale PBL structures. The tested technologies included Global Navigation Satellite System–Radio Occultation (GNSS-RO), differential absorption radar (DAR), infrared sounders (IR), Multi-angle Imaging SpectroRadiometer (MISR), Visible-Shortwave Infrared Spectrometer (VSWIR), and microwave sounders (MW).
Most of these instruments could be built with a variety of characteristics, for example, the spectrometers and sounders can be constructed with a variety of wavelength ranges, spectral channels, and spatial sampling strategies in mind, and with different noise characteristics. Furthermore, it would be prohibitively expensive to construct every possible combination of each instrument and to fly comprehensive field campaigns to test each one’s PBL science applicability. Therefore, a key challenge was to develop a reliable PBL retrieval observing system simulation experiment (OSSE) framework to assess different measurement technologies and to determine the best architectures to optimally address the PBL requirements. In this study, we document that effort by describing a PBL retrieval OSSE framework based on large-eddy simulation (LES), and provide conclusions on the analyzed capability of and potential future developments for six different instrument technologies.
Methodology
The PBL retrieval OSSE framework.
An OSSE is a modeling experiment used to evaluate the potential of a new observing system in measuring a feature of interest before the actual system can be built (Zeng et al. 2020). A traditional OSSE consists of a nature run (i.e., synthetic truth) that realistically simulates atmospheric structures, a data assimilation system, and a modeling component that mimics a future measurement technology. Recently, NASA expanded that definition to include two new types: retrieval and sampling OSSEs (Zeng et al. 2020). The retrieval OSSE, which is the focus of this study, quantifies the accuracy of a prospective measurement of a geophysical quantity of interest.
Our PBL retrieval OSSE framework (Fig. 1) comprises three major components: (i) simulated PBL conditions, often referred to as nature runs, (ii) instrument simulators that predict the observations produced by potential instruments, and (iii) retrieval algorithms that convert these synthetic observations back into PBL properties. Additionally, we investigate potential synergies between instruments that could strengthen their mutual capabilities. In this idealized study, we make no assumptions about the orbiting platforms, orbit parameters, or sampling strategies. The main focus is to evaluate the ability of the instruments to measure basic PBL characteristics. Below we describe the main components of the framework.
Schematic of the PBL retrieval OSSE framework applied in this study.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0129.1
Nature runs.
Numerical weather prediction (NWP) OSSEs heavily rely on general circulation models (GCMs). However, with typical resolutions of O(10) km, GCMs can only resolve the largest quasi-two-dimensional flow structures (Zeng et al. 2020). While using models with O(1) km resolutions can be a partial remedy, typical PBL structures are three-dimensional (3D) and range between O(1) m and O(1) km. In practice, only LES can represent complex PBL processes with high fidelity, and this tool has been essential for studying 3D turbulence and convection for the last five decades (Stoll et al. 2020). LES can explicitly resolve most of the 3D kinetic energy cascade, including circulations leading to the formation of individual clouds (Stoll et al. 2020). Modern LES models can produce realistic physics-based 3D fields of temperature, moisture, clouds, and winds, and have been validated against many observations for typical PBL regimes in all main climate zones, supporting their prominent role in process-based PBL investigations (see Table A1 and references therein). An important drawback of LES models is that they cannot represent larger mesoscale structures (i.e., exceeding their domain size) and are usually run for short [i.e., O(10) h] times. Nevertheless, given their unique ability to represent all key processes affecting the evolution of the PBL, we argue that they are an essential component of PBL retrieval OSSEs.
To first approximation, PBL regimes are determined by the planetary-scale atmospheric circulations: Hadley, midlatitude, and polar cells. Those PBL regimes have well-recognized mean latitudinal patterns, transitioning from warm and humid deep precipitating layers in the tropics, through drier shallow layers in subtropics, a large variability of dry and moist midlatitudinal layers, to cold shallow layers in the subpolar and polar regions. Using that perspective, we built a library of typical PBL regimes following previous modeling studies based on field campaigns. Our approach prioritizes sampling different PBL types rather than simulating their spatially continuous global coverage.
While LES models are commonly used to resolve turbulent structures in the lowest troposphere where a large spatiotemporal variability occurs, they typically neglect the upper atmosphere as irrelevant for the PBL dynamics. Furthermore, they often apply periodic lateral boundary conditions to ensure the continuity and coherence of the resolved turbulent structures within their domains. For those reasons, many of the high-resolution simulations in our library extend up to several kilometers in altitude, except for deep convection cases and those applying interactive radiation, which required extending the LES domain above the troposphere. The LES can account for large-scale horizontal advective tendencies but typically neglects mesoscale (i.e., domain-size) or larger horizontal gradients of the physical fields. To produce full atmospheric columns required to simulate satellite observations, we combine LES outputs with upper-atmospheric profiles from Modern-Era Retrospective Analysis for Research and Applications (MERRA-2). The upper parts of the profiles are fixed in time and have no spatial variability that is practically negligible at the horizontal and temporal scales represented by the LES. The reanalysis profiles are selected to match average conditions at the time and place that the LES case represents. Note that a subset of the LES cases from our library neglects the diurnal cycle as they were designed to represent the observed statistics of quasi-steady PBL regimes (see appendix).
The LES and MERRA-2 transitions in the vertical are smoothed to avoid generating artificial gradients at the top of the LES domain. MERRA-2 grid cells are larger than our typical LES domains, so our library has zero horizontal variability above the LES. However, the LES cutoff altitudes were selected to include 80%–90% of atmospheric water vapor and, according to field campaign evidence (Bedka et al. 2021), the above-LES fields are relatively uniform such that our framework captures most of the horizontal variability. Note that various approaches to the formulation of a full atmospheric column are possible, also accounting for a subdomain horizontal and temporal variability of the upper atmosphere. However, we here focus on investigating the sensitivity of our instruments to the PBL structures evolving on hourly time scales for fixed upper-atmospheric conditions (i.e., only the LES part changes with time and space).
Finally, it is important to recognize the sources of uncertainty in any modeling study, including those involving LES. These uncertainties can arise from approximations applied to governing equations, numerical discretizations, or physical parameterizations, and can vary depending on the type of the simulated PBL. While these uncertainties have been quantified in many intercomparison studies, they were not explicitly accounted for in the current study as it would require running large ensembles of simulations covering the mentioned aspects. The appendix provides a list of selected PBL cases along with the references to associated field campaigns and numerical experiments and intercomparison studies that involved LES studies. The models we used to simulate these PBL cases are also specified there.
Given the size of our LES domains, we aim at detecting only sub-domain-size features directly associated with local PBL dynamics (e.g., eddies, clouds or cloud pockets, moisture perturbations, or local PBL height). In other words, our primary focus is on the smallest scales as recommended by the ESAS 2017. For larger-scale features (e.g., synoptic or mesoscale systems, cyclones, fronts, large-scale free-tropospheric advection) and to account for subseasonal or seasonal variability, one may still need to use a complementary NWP retrieval OSSE framework.
JPL instrument and retrieval simulators
Instrument simulators take an atmospheric state, in this case as simulated by the LES and reanalysis, and use radiative transfer codes to generate observable properties such as radiance or radar backscatter. These properties are typically converted to an observation signal, accounting for how an instrument would sample each wavelength. This calculation to convert an atmospheric state to an observation is often called the forward model. In reality, instruments report observations and “work backwards” to obtain the atmospheric state. This is called the inverse problem.
We employed six JPL instrument simulators (Table 1), including both passive and active sensors, to represent the PBL signals they would measure. In the next section, we describe each instrument’s measurement techniques, treatment of uncertainty, results, and future perspectives for PBL applications.
List of instruments participating in the PBL retrieval OSSE along with their basic technical details. Except for MI, the approximate resolutions shown are estimates for future measurements from space. Note also that MI measures cloud-top height rather than T or q profiles.
Instrument results and perspectives
Multiangle imaging.
Overview.
This passive measurement technique relies on multiangle imaging of PBL objects (i.e., clouds) in the solar spectrum, from visible to near-infrared (VNIR) and shortwave infrared (SWIR). Although it does not directly determine high-precision temperature and humidity profiles, it strongly constraints them through recovered cloud properties and provides unique insights into cloud-scale processes. Thanks to multiple high-resolution cameras looking at clouds from different angles, it enables robust operational retrievals of cloud-top heights via stereo techniques along with horizontal wind (Horváth and Davies 2001). In short, MI captures the evolution of individual clouds and cloud systems, which can in turn facilitate future estimations of their dynamical properties.
An important proof of concept using high-resolution MI is the emerging passive technique of computed cloud tomography (CCT). CCT produces a gridded 3D image of a cloud’s inner structure and outer shape (Levis et al. 2015). The latest developments in CCT are about accessing key cloud microphysical properties, effective droplet radius and effective variance, first as new 1D (vertically distributed) unknowns using multispectral information (Levis et al. 2017) then, potentially, as two new unknown 3D fields using additional polarization information (Levis et al. 2020). Successful demonstration of CCT so far (∼5% retrieval error for total cloud water content) has used high-resolution imaging: pixels that are at most a few tens of meters, which is routinely achievable by AirMSPI on NASA’s ER-2 aircraft (Diner et al. 2013). CloudCT (Tzabari et al. 2021) is a European Research Council–sponsored mission consisting of about 10 camera-carrying CubeSat flying in a “string-of-pearls” formation to image simultaneously a cloud scene with many different viewing directions at about 50 m resolution in a swath of 70 km. CloudCT is currently planned for launch at the earliest in 2024.
Results.
Development of CCT using moderate-resolution (i.e., hundreds of meters) MI is justified by the science-driven need for wider swaths that ensure global coverage with reasonably short revisit time for both existing (MISR + MODIS/Terra; Diner et al. 1998; King et al. 2003) and future (AOS) NASA satellite missions. CCT for large/opaque clouds observed at moderate resolution is very different from CCT on smaller/less-opaque counterparts observed at high spatial resolution. This is mainly because voxels that match the larger pixel scales can be optically thick and/or internally variable so the gridded version of a large/opaque cloud is incompatible with the core assumptions of any deterministic 3D radiative transfer (RT) code: that voxels must be optically thin and internally uniform. A first step in a new direction in cloud remote sensing from space was the operative definition of the “veiled core” (VC) of sufficiently large and opaque clouds as the internal region where the detailed (pixel-scale) structure of the clouds does not influence the details of the multiviewing imagery; i.e., its impact is buried in the sensor noise (Forster et al. 2021).
Motivated by the clear need to revisit both the forward modeling and the inverse problem solution in CCT for large/opaque clouds, we investigated the foundational physics of cloud image formation, with a focus on space-based sensors (Davis et al. 2022). We describe the outer shell (OS) of the cloud, between its VC and outer boundary, as a boundary layer for the radiation transport in the same sense that the PBL is a boundary layer for atmospheric fluid dynamics. Specifically, we have in the PBL a strong influence of the presence of a lower physical boundary on the thermal structure and behavior of the flow in space and time. In RT, a cloud is a 3D optical medium and its boundary is not a hard 2D surface; rather, it is a fuzzy region where the volume extinction coefficient gradually vanishes. In essence, the radiation flow enters as a uniform irradiance field from the sun then transitions in the OS from ballistic trajectories to random walks on the way into the cloud. The converse happens on the way out, heading toward the imaging sensors, but now with rich textures that define the cloud image. Using discrete-time random walk theory, Davis et al. (2022) correctly predict the optical thickness of the OS based on the forward-scattering tendency resulting from the strong forward peak in cloud droplets’ phase functions; that prediction confirmed the empirical determination by Forster et al. (2021).
The practical consequence for CCT on large/opaque clouds is that there should be far less tomography unknowns assigned to the VC since there is so little information about its structure in the observed radiance fields. In essence, only low frequencies are of interest, hence a major reduction in the complexity of the inverse problem. Also, the current forward 3D RT model used in CCT should be replaced by some kind of hybrid deterministic 3D RT model with very efficient but approximate transport in the VC and standard 3D RT in the OS.
In the PBL OSSE, we also produced high-fidelity synthetic multiangle imagery of cloud scenes at native LES resolutions using the 3D RT model MYSTIC (Mayer 2009), which is part of libRadtran (Mayer and Kylling 2005; Emde et al. 2016). As customary in photorealistic CGI, MYSTIC is run in its backward Monte Carlo mode with several standard variance-reduction techniques (Buras and Mayer 2011). Examples are shown in Fig. 2 for shallow tropical convection transitioning to deep convection and for subtropical stratocumulus, broken stratocumulus, and cumulus, as seen from above. Those visualizations demonstrate the high level of realism of the LES in producing cloudy PBL regimes.
High-precision visualizations of four sample LESs using the 3D radiative transfer model MYSTIC: (a) tropical continental deep convection (LBA) at time 32,400 s, (b) subtropical marine nondrizzling stratocumulus (DYCOMS-II) at time 10,590 s, (c) subtropical marine precipitating shallow convection (RICO) at time 21,600 s, and (d) subtropical marine drizzling stratocumulus with pockets of open cells (POC) at time 28,845 s. All MYSTIC simulations are performed at a wavelength of 670 nm, showing results for MISR’s forward-facing B camera at −45.6° from due south (viewing azimuth of 180°). The sun is located at 46° zenith and 140° azimuth angle (southeast) corresponding to Terra’s AM equatorial crossing time. (a) LBA has OPAC continental average aerosol with AOD = 0.1 and “evergreen broad forest” surface albedo from IGBP library; (b) DYCOMS-II has OPAC maritime clean aerosol with AOD = 0.1 and ocean bidirectional reflectance distribution function (BRDF) parameterized with Cox and Munk (1954) and Cox (1954) assuming a 2 m s−1 wind speed. (c) RICO and (d) POC share the same aerosol properties as (b) but assume an ocean BRDF for a 10 m s−1 wind speed. Both LES and radiative transfer simulations use periodic horizontal boundary conditions, which are visualized here by a 10% replication of the cloud fields in the x and y directions. To demonstrate high realism of LES, some visualizations were performed for higher-resolution simulations (cf. Table A1) providing a detailed structure of individual clouds.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0129.1
We also demonstrated using MI PBL OSSEs that multiangle imaging from a pair of space-based sensors could retrieve 3D winds for convective plumes in a broken cumulus field. Preliminary results (not shown) suggest that such data can be used to retrieve a domain-average entrainment rate at cloud top for a stratocumulus field along with vertical wind at cloud base on a similar scale.
Perspectives.
We foresee MI PBL retrieval OSSEs to be essential for both maturing cloud tomography techniques and development and testing of new PBL-cloud remote sensing algorithms based on multiangle imaging. Those new algorithms, which are under development and involve time-lapsed cloud scenes, will facilitate extracting new key cloud properties. Such algorithms include the following:
Tridirectional/two-time technique for deriving cloud-top height and cloud-top vertical velocity in cumulus cloud fields
Nadir-only/two-time technique for estimating cloud-top horizontal wind divergence in stratocumulus, from there, cloud-top entrainment rate and vertical wind at cloud base
Three-dimensional tomographic reconstruction of convectively driven clouds with full multiangle imaging supplemented with either bispectral (VNIR + SWIR) or multipolarization data for their sensitivity to cloud microphysical properties
These emerging methods for retrieving properties of cumuliform and stratiform clouds in the PBL will enable new PBL science questions to be addressed and extend validation opportunities for atmospheric models.
VSWIR.
Overview.
VSWIR sensor capacities have greatly improved recently. For example, NASA’s aging MODIS sensors have 19 channels spanning 620–965 nm wavelength with spatial resolution of 0.25–1 km. By comparison, the EMIT instrument covers 380–2,500 nm with hundreds of channels and 0.03–0.08 km footprints. As described in Richardson et al. (2021), we built a retrieval emulator from fitting a set of optimal estimation retrievals in which forward-model radiances had uncertainties added using the EMIT prelaunch instrument noise model. Uncertainties in aerosol optical depth (up to 0.3) and due to different surface types were included by changing their values in the training set.
Similar capabilities are now flying and will be extended in future missions such as ESA’s Copernicus Hyperspectral Imaging Mission for the Environment (CHIME) and for NASA’s Surface Biology and Geology (SBG) designated observable. As the SBG name suggests, VSWIR missions often target Earth’s surface. However, both the surface and atmosphere scatter and absorb sunlight, so researchers must account for both to reliably retrieve either. In particular, water vapor absorption features allow retrieval of total column water vapor (TCWV). Recent software developments such as JPL’s Imaging Spectrometer Optimal Fitting (ISOFIT), which has been validated in the field (Thompson et al. 2018, 2019) simultaneously retrieve both surface and atmosphere properties. Presently, these retrievals only work for clear-sky footprints, and primarily over land where the surface is bright enough to reflect sufficient sunlight.
Results.
We have recently demonstrated use of VSWIR vapor retrievals for PBL science using AVIRIS-NG flights (Thompson et al. 2021) and the PBL OSSE (Richardson et al. 2021). Currently only TCWV is retrieved with no vertical detail. However, both LES and field-campaign data (Bedka et al. 2021) report that subkilometer horizontal vapor variability when integrated over the whole column is almost identical to the horizontal variability within the PBL, since vapor is well-mixed at higher altitudes. This suggests that horizontal variability of clear-sky PBL water vapor can be retrieved from space at 30–60 m resolution. Such capability would provide new opportunities to study processes such as convective onset, which is linked to horizontal variability in water vapor (Couvreux et al. 2009).
Furthermore, weather and climate models parameterize the subgrid water vapor distribution, which generally requires using scale-aware relationships. In both AVIRIS-NG data and our PBL retrieval OSSE, we quantified this using the second-order structure function exponent (ζ2) (Richardson et al. 2022), a parameter that depends on the physics of the field. For example, it is zero in a white-noise field with no structure, while Kolmogorov derived that ζ2 = 2/3 for a passive tracer in 3D isotropic turbulence. Measurement of ζ2 could reveal PBL processes and allow modelers to identify how subgrid water vapor variance changes depending on thermodynamic conditions and gridcell size.
We evaluated PBL OSSE vapor retrievals in subtropical and midlatitude shallow convective environments. Figures 3a and 3b show a reliable surface retrieval and TCWV biases of order 1 mm. However, within each LES case there is strong correlation between the local true TCWV and the retrieved TCWV, meaning that the spatial standard deviation of PBL vapor can be calculated to within ±30% and usually better. There are biases in retrieval TCWV that depend on whether surfaces were urban or vegetation types, so if an observed scene consists of a mixture of the two then the estimated standard deviation is artificially increased. However, the retrieval simultaneously obtains surface type, so in practice it will be trivial to identify scenes where more accurate retrievals of PBL water vapor statistics are possible.
(a) Example retrievals of surface reflectance spectra for different vegetation types and (b) retrieved TCWV as a function of the true LES TCWV; colors represent each LES case used, with a black dashed 1:1 line. (c),(d) Calculated vs true ζ2 for 23 LES snapshots calculated (a) parallel to the solar azimuth (flying “into” the sun) or (b) perpendicular to the solar azimuth (flying “across” the sun). Colors refer to solar zenith; vertical lines show values associated with a white-noise field (no structure) or Kolmogorov turbulence. The equation below the panels relates the structure function (S2) to its exponent (ζ2).
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0129.1
A major complication is that sunlight typically travels diagonally through the atmosphere, often entering the PBL several kilometers horizontally distant from the nominal surface footprint. Our airborne tests addressed this by only using measurements where solar zenith angle was under 15°, but this would severely restrict a global satellite mission. By ray tracing within LES output, we found that it is still possible to retrieve horizontal variance with little change in uncertainty. Most affected were the structure function results; the retrieved TCWV field is predominantly distorted in the same horizontal direction as the sunlight travels. The structure function exponent can be derived in any direction, and Figs. 3c and 3d show that by calculating only perpendicular to the solar path, we can eliminate the average ζ2 bias. This means that, for example, if the sunlight is passing south–north through the atmosphere, then the statistical properties of the PBL’s horizontal water vapor distribution can still be determined in the east–west direction.
Perspectives.
Our work identified priorities for improving VSWIR PBL applications. First, the ISOFIT retrievals assumed standard atmospheric structures with a fixed vertical profile shape for specific humidity. Retrievals should better account for local vertical moisture and temperature profile structure. Second, airborne campaigns using VSWIR in concert with other sensors could confirm or refute the Figs. 3c and 3d results. Confirmation would imply that upcoming surface-targeted missions could have outputs that are useful for PBL science. Finally, if VSWIR measurements were to occur coincidentally with other sensors, their high spatial resolution could result in useful subfootprint information to improve other retrievals.
In summary, anticipated VSWIR capacity can provide unique information about the horizontal spatial structure of clear-sky PBL water vapor over land, providing atmospheric applications for several proposed missions that nominally target surface objectives.
DAR.
Overview.
DAR is an emerging approach to measure water vapor profiles within cloudy and precipitating volumes inaccessible to other remote sensing techniques (Nehrir et al. 2017; Battaglia and Kollias 2019). The approach uses three radar frequencies between 155.5 and 174.8 GHz, on the wings of a water vapor absorption line to perform in-cloud water vapor sounding. An airborne demonstrator, the Vapor In-cloud Profiling Radar (VIPR), has been developed over the last several years at JPL and has demonstrated precision on the order of 0.8 g m−3 and 200 m vertical resolution during limited deployments (Roy et al. 2020, 2022). Although there have been several preliminary performance assessments for spaceborne applications (Millán et al. 2020; Battaglia and Kollias 2019), it is not yet known how well these measurement characteristics scale to various cloud regimes and from a spaceborne platform. Addressing this uncertainty was the focus of the study. A more complete description of the DAR results is provided in Roy et al. (2021).
We account for both random and systematic error sources in our analysis of DAR following the approach outlined in Roy et al. (2022). Random uncertainty in the radar reflectivity is calculated using the standard approach (Doviak and Zrnić 1993) from the signal-to-noise ratio and the number of averaged radar pulses. An advantage of the DAR approach is that random uncertainties can be expressed analytically given an expression of uncertainty in the radar reflectivities. These random uncertainties are shown as the horizontal error bars in Fig. 4c. Note that DAR retrievals do not depend on radar calibration. Nevertheless, there are several factors that are included in our forward modeling but not accounted for in the retrieval that result in systematic uncertainty. These factors include nonuniform beamfilling, multiple scattering, incorrect assumptions regarding the humidity profile shape, and nonlinearity in the frequency dependence of the hydrometeor single-scattering properties. The systematic biases are shown as the difference between the filled and open circles in Fig. 4c. For details on these calculations the reader is referred to Roy et al. (2022).
Example of (a) shallow convection as simulated by LES, (b) simulated observed DAR reflectivity at 155.5 GHz, and (c) DAR retrieval of water vapor density and its uncertainty.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0129.1
Results.
The end-to-end simulation assessment of DAR measurement capabilities consists of three primary components: development of a flexible DAR instrument simulator, development of a flexible three-frequency DAR retrieval algorithm, and characterization of measurement uncertainty across a diversity of cloud regimes.
We developed an instrument simulator to convert the LES-based data into simulated radar observables. To begin we coupled a multiple scattering radar simulator to the Microwave Limb Sounder millimeter-wave spectroscopic model (Read et al. 2004) with surface and cloud scattering properties (Hogan and Battaglia 2008). Spherical cloud and precipitation species were modeled using the Mie theory, while complex ice crystal shapes associated with snow were modeled using the discrete dipole approximation (Roy et al. 2021). The resulting forward model produces radar reflectivity profiles at the native LES resolution. Once these profiles are generated for each scene, the instrument model accounts for high-level instrument parameters including transmit power, duty cycle, pulse duration, antenna radiation pattern, range resolution, receiver noise figure, and orbital altitude. Application of the instrument model to the scene produces realistic simulation of measured radar power at the radar resolution and including realistic measurement noise.
We developed a new three-frequency retrieval algorithm that extends on the two-frequency approach originally employed in the VIPR (Roy et al. 2022). The addition of the third frequency allows for a simultaneous derivation of a water vapor content and the hydrometeor attenuation profile. This additional degree of freedom successfully mitigates hydrometeor related biases that were observed in real-world VIPR data. In addition, the flexible retrieval algorithm permits accurate estimation of integrated water vapor between physically separated cloud layers by interpolating the H2O spectroscopic parameters in the clear-air region.
The DAR measurement sampling and uncertainty characteristics were evaluated across a range of cloud types typifying the transition from stratocumulus to deep convection that occurs along the subtropical trade winds (Roy et al. 2021). Figure 4 shows an example of DAR synthetic observations and retrievals for shallow subtropical precipitating convection and their comparison with ground truth. We find that, generally, in-cloud water vapor profiles can be derived with a precision of 10%–20%, with a vertical resolution of 200 m, and an along-track resolution of 100–200 km. We find that the column integrated water vapor can be derived with an uncertainty of 10%, with an along-track resolution of tens of kilometers.
Perspectives.
We gleaned two key system-level requirements from the PBL retrieval OSSE. First, the DAR profiling requires very high range resolution to mitigate biases related to partial cloudiness at cloud edges. We found that 50 m resolution was required to mitigate these biases, which is roughly an order of magnitude finer resolution than CloudSat (Tanelli et al. 2008), but is readily achievable while maintaining sensitivity by employing high-duty-cycle pulse compression techniques (Roy et al. 2021). Second, we found that for scattered cumulus cloud regimes, scanning capabilities of roughly ±3° are necessary. This scanning must be accompanied by onboard automation to target the scan at optimal cloud targets and avoid clear sky scenes. Cloud targeting information could be based on forward looking measurements from a microwave radiometer, or by leveraging the operational geostationary imagers, and therefore be an example of multi-instrument synergy for PBL science.
IR.
Overview.
Spaceborne measurements of spectrally resolved radiances in the thermal infrared (650 to 2,700 cm−1, or 15.4 to 3.7 μm) allow profiling of temperature, water vapor, cloud properties, and numerous trace gases. Current IR sounders in low-Earth orbit offer spatial resolution of 12–16 km (Menzel et al. 2018). Reductions in footprint size would be an advantage for capturing small-scale spatial variations and mitigating the impact of clouds in the field of view (Di et al. 2021). IR sounders in the current NASA Program of Record offer vertical resolution of around 1 km for temperature and water vapor in the PBL (e.g., Irion et al. 2018). The vertical resolution for a given sounder depends on the details of the surface and atmospheric state as well as the spectral resolution and noise characteristics of the instrument. Future IR sensors with higher spectral resolution and/or reduced noise could offer improved vertical resolution.
Results.
We performed simulations to quantify the vertical resolution of retrievals of PBL water vapor for a range of instrument characteristics, with varying combinations of spectral resolution and instrument noise. The temperature, water vapor, and surface conditions from the LES were input into the k-Compressed Atmospheric Radiative Transfer Algorithm (kCARTA) (DeSouza-Machado et al. 2020), which outputs the Jacobians used in our information content analysis. These Jacobians are the partial derivatives of the radiance with respect to the atmospheric and surface parameters needed to calculate the sensitivity of the radiance to changes in these parameters.
We focused on clear-sky simulations and assumed that the temperature profile is perfectly known, recognizing that this presents a best-case scenario. While IR measurements can be sensitive to the atmosphere below clouds (Kulawik et al. 2006; Susskind et al. 2003), the presence of clouds is a clear limitation for sensitivity to the PBL. Examples shown here utilize the 1,200–1,500 cm−1 spectral region, which provides a wide range of water vapor optical depths and therefore allowing for sensitivity to variations in water vapor at a range of altitudes.
Figure 5 shows the near-surface vertical resolution as a function of spectral resolution and instrumental noise for five example sets of atmospheric conditions from the LES, including shallow subpolar (MPACE), midlatitudinal (ARM), subtropical (DYCOMS, BOMEX), and tropical (LBA) convective PBL types. This presents the best-case vertical resolution of a water vapor retrieval for an IR instrument with the specified spectral resolution and instrumental noise. The effective vertical resolution is related to the calculation for degrees of freedom of signal as described in Rodgers (2000). The calculation requires knowledge of how the spectrum responds to changes in atmospheric conditions (quantified via Jacobians calculated from kCARTA), error covariance, and a priori covariance. The error covariance was calculated as a percentage of the Infrared Atmospheric Sounding Interferometer (IASI) instrumental noise which was taken from Crevoisier et al. (2014). The prior covariance represents uncertainty about the atmospheric state before the observation occurs. It was built from a database of atmospheric moisture profiles taken by radiosondes launched during the MAGIC campaign (Kalmus et al. 2015) with an assumed exponential relaxation of 100 hPa to generate necessary off-diagonal terms.
Information content analysis for IR: (bottom right) typical temperature profiles of an infrared footprint for various PBL types as indicated by the legend. The PBL height is typically associated with a temperature inversion or a change in the stratification. In all cases, clouds are neglected. (top),(bottom left),(bottom center) Estimated achievable near-surface vertical resolution as a function of instrument noise and spectral resolution for each of the PBL types. Values are shown for measurements in wavenumbers from 1,200 to 1,500 cm−1, where the x axis is a fraction of the IASI instrument noise and the y axis is a function of spectral resolution, with IASI close to 0.5 cm−1 (IASI-NG is 0.25 cm−1). Near-surface vertical resolution was calculated as half the altitude by which 2 cumulative degrees of freedom are obtained.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0129.1
The PBL height ranges from around 0.9 to 2.5 km as defined by the height of either temperature inversion or of strong changes in the stratification. The thermal contrast (i.e., temperature difference between the surface and the atmosphere immediately above it) varies between around 1 and 3 K for the maritime and continental cases, respectively. Thermal contrast also impacts vertical resolution. The in-flight instrument noise for IASI (Crevoisier et al. 2014) was used as a reference point. For an IASI-like sensor (spectral resolution of 0.5 cm−1), the vertical resolution for water vapor retrievals in the PBL would be between 0.8 and 1.3 km for the subtropical and tropical case. Future sensors could offer improvements in both spectral resolution as well as noise, thereby providing improvements in vertical resolution of PBL measurements. The planned IASI Next Generation instruments, with spectral resolution of 0.25 and a factor-of-2 reduction in radiometric noise compared to IASI, would improve the vertical resolution of PBL water vapor measurements in these examples to around 0.7 and 1 km, respectively.
Perspectives.
These example simulations can be used to quantify the trade-offs in vertical resolution for PBL associated with different instrument characteristics. Results suggest that vertical resolution of 0.5 km is possible with a spectral resolution of 0.1 cm−1 combined with radiometric noise a factor of 3 lower than IASI. The approach used here could be used to extend the analysis to joint retrievals of temperature and water vapor, and/or to quantify vertical resolution achievable with combinations of IR sensors with other measurement techniques. Note that the major advantage of IR sounding, compared to the other observing technologies discussed in this paper, is that it is fairly feasible to obtain horizontal resolutions on the order of 1 km from space, uniquely satisfying the horizontal resolution requirements for a future PBL mission as discussed in Teixeira et al. (2021).
MW.
Overview.
Passive microwave radiometers provide thermodynamic information of the surface and atmosphere in all weather conditions. Existing microwave temperature and humidity sounders typically have one limited bandwidth channel at the wing of the 60 and 183 GHz absorption line with sensitivity to PBL temperature and water vapor. This is not for lack of information content that additional channels would provide. Information on the temperature and water vapor profile in the PBL is encoded in the shape and magnitude of the microwave spectrum around absorption lines; the weak 22 GHz water vapor line and the wings of the 60/118 GHz oxygen lines and 183 GHz water vapor line. Current satellite systems typically have just a few discrete channels near these lines and do not fully resolve the information across the spectrum. Their typical horizontal resolutions are on the order of 10 km and they are not designed to measure PBL features.
Results.
We first explored the benefits of fully resolving the absorption line spectra by employing a simple perturbation approach to a stratocumulus-topped marine PBL, which is one of the most challenging regimes to measure due to the presence of strong gradients and a shallow layer. Over-ocean brightness temperatures at the top of the atmosphere were simulated from 15 to 200 GHz using a radiative transfer model with the Liebe et al. (1993) atmospheric absorption model, Rayleigh approximation for cloud water absorption, and Meissner and Wentz (2012) wind roughed surface emission model. We perturbed the PBL profile either by increasing the PBL height with fixed water vapor mixing ratio or by increasing PBL humidity with fixed PBL height. These changes resulted in the same change in total precipitable water vapor, but had different PBL structures and most importantly produced unique microwave spectral signatures in the spectral window regions related to the pressure and temperature where the water vapor is concentrated. The perturbations modified four different spectral regions that could be measured with a spectrally resolving radiometer.
For the 15–35 GHz spectrum, the deeper PBL moves water vapor toward lower pressures, increasing the signal close to the 22 GHz line center resulting in a narrower line shape compared to the case with increased PBL humidity without the height change (Fig. 6d). It is useful to note here that the 23.8 GHz band used by most current sensors for retrieving precipitable water vapor was specifically chosen to reduce the dependence of water vapor scale height on the retrieval (Keihm et al. 1995), which is reflected in this example by the hinge points in the spectrum. Fully resolving the 22 GHz line shape provides new information on PBL structure beyond that obtained with current microwave systems.
Relative change in brightness temperature from 15 to 200 GHz for a perturbation of the boundary layer water vapor mixing ratio height (orange line) and a perturbation in the boundary layer water vapor mixing ratio magnitude without a change in height (blue line). (a) The perturbations in water vapor mixing ratio applied relative to a nominal profile (black line). (b) The corresponding potential temperature. (d)–(f) Individual spectral regions to better highlight the differences; the gray shading represents expected measurement noise.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0129.1
The upper and lower wings of the 50/118 GHz oxygen absorption lines provide temperature information and uniquely respond to the PBL humidity structure through the water vapor absorption continuum, which increases through this frequency range. Increasing the water vapor mixing ratio in the lower troposphere adds attenuation at warmer air temperatures, resulting in a larger brightness temperature change in the spectral windows compared with the case of a PBL height change, which adds attenuation (water vapor) at lower air temperatures. This is particularly reflected in the asymmetry in the peaks at the edges of the oxygen absorption lines (near 47, 110, and 130 GHz), which aids separation of the temperature signal from water vapor, since temperature variations will be symmetric about the line center. Near the 183 GHz water vapor line, the nonlinear spectral kink around 177 GHz in Fig. 6f is indicative of the PBL height change, as the water vapor (absorption) discontinuity moves higher in the atmosphere.
After confirming that microwave sounding can detect relatively weak PBL signals, we have developed a PBL retrieval framework from a spectrum resolving microwave radiometer aided by Bayesian optimization and a spectral angle mapping algorithm. We simulated microwave brightness temperatures from 15 to 200 GHz at 40 MHz spectral resolution from the LES-based geophysical inputs using the radiative transfer model described above. Normally distributed random noise is added to each 40 MHz sample using receiver noise temperatures of 400, 500, and 600 K for the bands of 15–35, 35–90, and 90–200 GHz. These noise temperatures are based on available low-noise amplifiers in these bands. No additional biases are considered as these would be removed with respect to the forward radiative transfer model through the postlaunch calibration process. We produced the retrievals and compared them to the LES geophysical inputs toward resolving PBL vertical structure. We validated the retrievals using standard sounding metrics in addition to PBL metrics, such as PBL height.
Our results showed the potential of the new approach in improving the sounding metrics compared to traditional sounding systems. For the tropical and subtropical PBL regimes, we obtained the approximate retrieval uncertainty of 0.6 g kg−1 for water vapor, 1 K for temperature, and approximately 200 m for PBL height for more distinct PBL types such as stratocumulus-topped, shallow, or morning tropical PBLs. These results are for an unconstrained retrieval algorithm and the maximum benefit from a broad spectrum microwave sounder would be in the context of data assimilation. In data assimilation, a numerical weather prediction model would provide a priori constraints on the atmospheric structure that reduce the solution space, resulting in further retrieval improvements.
Perspectives.
There is significantly more information about the PBL temperature and water vapor structure available in the microwave spectrum than is currently measured by spaceborne radiometer systems. The key innovations to support a PBL mission are technologies to resolve large (>100 GHz) regions of the microwave spectrum (e.g., microwave photonics) at modest spectral resolution (∼1 GHz). While not explored here, it is expected that adding unique observing geometries, such as high spatial resolution or multiangle measurements of the same scene, will increase information content and should be an area of future study.
GNSS-RO.
Overview.
This active remote sensing technique probes atmospheric thermodynamic structure by measuring the bending of the occulted GNSS signal ray propagating through stratified layers (Kursinski et al. 1997). Due to its spaceborne nature and limb sounding geometry, GNSS-RO can provide global high-vertical-resolution (∼200 m) observations at any time of the day. Furthermore, the L-band navigation signals used in GNSS-RO can penetrate through clouds and precipitation. Numerous GNSS-RO missions have been launched since 1995 and the number of RO observations has reached more than 10,000 per day (Irisov et al. 2020; Schreiner et al. 2020).
The GNSS-RO measurements yield vertical profiles of refractivity, which are related to both temperature and moisture. Therefore, the retrievals of temperature and water vapor in the moist troposphere depend on the a priori information on the two fields. While valuable PBL information such as the PBL height can be extracted directly from the vertical structure of the retrieved RO refractivity (Ao et al. 2012; Kalmus et al. 2022), further progress can be made by combining RO with complementary information from other instruments. In our PBL retrieval OSSE, we evaluated the accuracy of GNSS-RO retrievals for key PBL regimes and explored the benefits of combined retrievals with passive microwave radiometer (MW) that has lower vertical resolution (∼1 km). Since the nadir MW brightness temperature measurements can be related to the vertical thermodynamic structure, MW can assist in RO retrievals and help disentangle the contributions of temperature and water vapor in the atmospheric column.
The uncertainty of RO measurements arises from several different sources. In the upper-troposphere-to-middle-stratosphere range (7–35 km), the bending angle measurement uncertainty is mainly contributed by thermal noise, clock error, and imperfect ionosphere calibration (Kursinski et al. 1997). In the middle-to-lower troposphere (<7 km), large vertical refractivity gradients, nonspherically symmetric atmospheric structure, and random small-scale features dominate the bending uncertainty (Zhang et al. 2023). The propagation of uncertainty from error sources to atmospheric profile throughout the retrieval chain has been shown in several studies (Schwarz et al. 2017, 2018), and verified by comparing the RO retrieval with collocated reanalysis statistically, such as through the three-cornered hat method (Rieckh et al. 2021). In general, the RO refractivity measurement uncertainty can reach 0.5% between 10 and 35 km with the COSMIC-2 data and increases to 3% in the lower troposphere (Anthes et al. 2022). For the results shown in this article, the RO measurement uncertainty has been introduced in the bending angles as described next.
Results.
To combine the MW and RO observations, we implemented a joint retrieval algorithm based on the 1DVar technique (Rodgers 2000). In this iterative optimal estimation approach, the field of interest (temperature and water vapor profiles) is retrieved by minimizing the cost function defined by the first-guess estimate (which we take as equal to our a priori) and observation errors. We merged the RO bending angle profiles with MW brightness temperatures using the error covariance matrices encapsulating the uncertainty of the a priori and the two observations. In this study, the covariance matrices are diagonal and include the statistical uncertainty of the bending angle (0.0008 rad) and brightness temperature (0.25 K) following Bormann et al. (2012), and of the state variables following von Engeln et al. (2003).
The classical GNSS-RO forward operator used in weather applications cannot properly handle the super-refraction phenomenon (Ao 2007) caused by sharp refractivity gradients at the top of PBL. To address that issue, we have developed a fast forward Abel integrator based on the algorithm developed in the Radio Occultation Processing Package (Culverwell et al. 2015) that can account for super refraction. Also, we modified the radiative transfer model to calculate the brightness temperature of different frequency channels given the atmospheric structure, surface temperature and near-surface pressure, wind speed, and humidity, and the Earth incident angle and sensor off-nadir angle. The GNSS-RO bending angle and MW brightness temperature were simulated using a forward model (Meissner and Wentz 2012; Rosenkranz 1998). The a priori was derived by smoothing synthetic vertical profiles in 1 km by 1 km windows. Additionally, 2 K bias was added to the a priori temperature to test the 1DVar sensitivity to a priori errors.
Overall, the joint RO–MW retrievals were more accurate than either RO or MW alone (Fig. 7). For temperature retrievals, both MW and RO–MW were able to remove the bias, which propagated uncorrected in the RO retrievals. Clearly, nadir MW observations provided essential new information on true temperature profiles for RO retrievals. The greatest challenge remains probing shallow PBLs with sharp temperature gradients occurring in thin layers. Their retrievals were too smooth for each tested configuration. For moisture profiles, MW does not provide sufficient vertical resolution and the retrievals are too smooth. While RO also struggles to accurately detect sharp vertical transitions for shallow layers and incorrectly detects the amount of PBL moisture, the RO–MW approach demonstrates a remarkable improvement in retrieving the vertical structure of the PBL. Note that the vertical resolution for the MW water vapor retrievals is much finer than for temperature ones (1–2 km versus 3–5 km), which translated into a more accurate retrieval of moisture structure for the combined retrievals.
Profiles of (a)–(c) temperature and (d)–(f) humidity as retrieved by RO, MW, and RO+MW for three types of PBL: (left) tropical deep convection, (center) subtropical shallow convection, and (right) stratocumulus-topped boundary layer.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0129.1
Perspectives.
While GNSS-RO provides high vertical resolution refractivity profiles from the lower troposphere down to the PBL, RO-only temperature and moisture retrievals within the PBL are subject to significant uncertainty. We foresee future PBL applications through a synergy of RO and MW (or similar) observations. There are three potential advantages of joint RO–MW PBL retrievals. First, they can constrain temperature accuracy within 0.4–0.5 K and strongly reduce the errors propagated from the a priori. Second, the retrieved moisture profiles retain high vertical resolution (∼200 m) and more accurately represent vertical gradients within the PBL. Third, the negative refractivity retrieval biases due to super refraction can be significantly reduced, yielding a potentially bias-free PBL water vapor product.
Discussion and conclusions
Recent years have witnessed increased efforts toward incubating and developing new techniques for probing the PBL from space. Those efforts are largely motivated by the ESAS 2017 and aim at closing gaps in the current Earth system observation strategy, in which the PBL remains beyond the eyes of operational spaceborne instruments. Both the development of instruments and future mission designs have to rely on OSSEs built around realistic PBL simulations. This requirement is essential for accurately identifying feasible approaches addressing major challenges in measuring the large variability of small-scale PBL features.
Since three-dimensional PBL structures can only be resolved using high-resolution large-eddy simulation (LES) models, we argue that future PBL retrieval OSSEs need to utilize LES models. Their ability to represent virtually all key PBL types has been demonstrated in many past studies. Their small-domain size limitations can be mitigated by using complementary regional and global models. Therefore, we foresee emerging opportunities for collaborations between the observation and modeling communities necessary for formulating a wide range of measurement-specific PBL OSSEs.
The PBL retrieval OSSE results presented here have revealed the potential for current and upcoming instruments to obtain information relevant for NASA’s PBL observation goals. Summarizing each instrument:
- 1)Multiangle imaging can identify cloud properties and wind speeds in the presence of clouds.
- 2)VSWIR reflectance can identify very fine horizontal resolution (<50 m) structure of vertically integrated water vapor within the PBL, but only in clear sky.
- 3)Differential absorption radar can return detailed water vapor profiles within clouds with the vertical resolution of around 200 m and along-track resolution of 100–200 km.
- 4)For IR sounding, the examples discussed suggest that PBL vertical resolutions close to 500 m in clear sky could be possible together with fairly feasible horizontal resolutions of about 1 km.
- 5)Passive microwave sounders with improved spectral resolution could retrieve temperature and water vapor profiles up to practical spatial resolutions of ∼5 km and are less sensitive to clouds.
- 6)GNSS–Radio Occultation can retrieve information at fine vertical resolution of just 200 m, but with coarse horizontal resolution of 100–200 km.
It is notable that each of these measurement techniques has different advantages and limitations. This implies that combinations of instrument technologies will be necessary to fully achieve the PBL science goals. The passive microwave–RO synergistic retrievals presented here represent one example of multiple instruments being used to produce more reliable retrievals, but cross-instrument synergy is definitely not limited to this approach. For example, coincident IR retrievals of clear sky could be combined with DAR profiling of within-cloud water vapor to provide a more complete picture of the PBL. Ongoing and future OSSEs work could further develop a better understanding of these capabilities. All such work will continue to benefit from LES nature runs that provide detailed and realistic PBL fields.
See National Academies of Sciences, Engineering, and Medicine (2018) for more details: https://doi.org/10.17226/24938.
Acknowledgments.
The research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration (80NM0018D0004). We acknowledge funding from the NASA PBL DSI Study Team program. The High Performance Computing resources used in this investigation were provided by funding from the JPL Information and Technology Solutions Directorate. The authors acknowledge the Texas Advanced Computing Center (TACC) at The University of Texas at Austin for providing HPC resources that have contributed to the research results reported within this paper (www.tacc.utexas.edu). Support from Dr. Tanvir Islam on MW simulations is greatly acknowledged.
Data availability statement.
All the PBL regimes listed in Table 1 were simulated based on the provided references describing them in detail. The dataset of large-eddy simulations on which this paper is based is too large to be retained or publicly archived with available resources. However, snapshots for tropical and subtropical PBL regimes over ocean are available at https://amt.copernicus.org/articles/14/6443/2021/amt-14-6443-2021.html. Reanalysis data are available at https://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/.
Appendix: PBL types simulated for the OSSE
Table A1 provides a list of main PBL types simulated for the OSSE.
List of main PBL types simulated for the OSSE. Main references describe basic details of the simulations. However, various modifications were applied to increase PBL variability, as briefly explained in the comments. The LES models used include the Weather Research and Forecasting LES (WRF-LES; Skamarock et al. 2008), System for Atmospheric Modeling (SAM; Khairoutdinov and Randall 2003), JPL-LES (Matheou and Chung 2014), and EULAG (Prusa et al. 2008). Resolutions in parentheses are for the results visualized in Fig. 2.
References
Anthes, R., J. Sjoberg, X. Feng, and S. Syndergaard, 2022: Comparison of COSMIC and COSMIC-2 radio occultation refractivity and bending angle uncertainties in August 2006 and 2021. Atmosphere, 13, 790, https://doi.org/10.3390/atmos13050790.
Ao, C. O., 2007: Effect of ducting on radio occultation measurements: An assessment based on high-resolution radiosonde soundings. Radio Sci., 42, RS2008, https://doi.org/10.1029/2006RS003485.
Ao, C. O., D. E. Waliser, S. K. Chan, J.-L. Li, B. Tian, F. Xie, and A. J. Mannucci, 2012: Planetary boundary layer heights from GPS radio occultation refractivity and humidity profiles. J. Geophys. Res., 117, D16117, https://doi.org/10.1029/2012JD017598.
Battaglia, A., and P. Kollias, 2019: Evaluation of differential absorption radars in the 183 GHz band for profiling water vapour in ice clouds. Atmos. Meas. Tech., 12, 3335–3349, https://doi.org/10.5194/amt-12-3335-2019.
Bauer, P., A. Thorpe, and G. Brunet, 2015: The quiet revolution of numerical weather prediction. Nature, 525, 47–55, https://doi.org/10.1038/nature14956.
Bedka, K. M., and Coauthors, 2021: Airborne lidar observations of wind, water vapor, and aerosol profiles during the NASA Aeolus calibration and validation (Cal/Val) test flight campaign. Atmos. Meas. Tech., 14, 4305–4334, https://doi.org/10.5194/amt-14-4305-2021.
Berner, A. H., C. S. Bretherton, and R. Wood, 2011: Large-eddy simulation of mesoscale dynamics and entrainment around a pocket of open cells observed in VOCALS-REx RF06. Atmos. Chem. Phys., 11, 10 525–10 540, https://doi.org/10.5194/acp-11-10525-2011.
Bormann, N., A. Fouilloux, and W. Bell, 2012: Evaluation and assimilation of ATMS data in the ECMWF system. ECMWF Tech. Memo. 689, 17 pp., www.ecmwf.int/sites/default/files/elibrary/2012/8301-evaluation-and-assimilation-atms-data-ecmwf-system.pdf.
Brown, A. R., and Coauthors, 2002: Large-eddy simulation of the diurnal cycle of shallow cumulus convection over land. Quart. J. Roy. Meteor. Soc., 128, 1075–1093, https://doi.org/10.1256/003590002320373210.
Buras, R., and B. Mayer, 2011: Efficient unbiased variance reduction techniques for Monte Carlo simulations of radiative transfer in cloudy atmospheres: The solution. J. Quant. Spectrosc. Radiat. Transfer, 112, 434–447, https://doi.org/10.1016/j.jqsrt.2010.10.005.
Couvreux, F., F. Guichard, P. H. Austin, and F. Chen, 2009: Nature of the mesoscale boundary layer height and water vapor variability observed 14 June 2002 during the IHOP 2002 campaign. Mon. Wea. Rev., 137, 414–432, https://doi.org/10.1175/2008MWR2367.1.
Couvreux, F., and Coauthors, 2020: Intercomparison of large-eddy simulations of the Antarctic boundary layer for very stable stratification. Bound.-Layer Meteor., 176, 369–400, https://doi.org/10.1007/s10546-020-00539-4.
Cox, C., 1954: Statistics of the sea surface derived from sun glitter. J. Mar. Res., 13, 198–227.
Cox, C., and W. Munk, 1954: Measurement of the roughness of the sea surface from photographs of the sun’s glitter. J. Opt. Soc. Amer., 44, 838–850, https://doi.org/10.1364/JOSA.44.000838.
Crevoisier, C., and Coauthors, 2014: Towards IASI-New Generation (IASI-NG): Impact of improved spectral resolution and radiometric noise on the retrieval of thermodynamic, chemistry and climate variables. Atmos. Meas. Tech., 7, 4367–4385, https://doi.org/10.5194/amt-7-4367-2014.
Culverwell, I. D., H. W. Lewis, D. Offiler, C. Marquardt, and C. P. Burrows, 2015: The Radio Occultation Processing Package, ROPP. Atmos. Meas. Tech., 8, 1887–1899, https://doi.org/10.5194/amt-8-1887-2015.
Davis, A. B., L. Forster, B. Mayer, and D. J. Diner, 2022: Toward cloud tomography from space using MISR and MODIS: The physics of image formation for opaque convective clouds. arXiv, 2011.14537, https://doi.org/10.48550/arXiv.2011.14537.
DeSouza-Machado, S., L. Strow, H. Motteler, and S. Hannon, 2020: kCARTA: A fast pseudo line-by-line radiative transfer algorithm with analytic Jacobians, fluxes, nonlocal thermodynamic equilibrium, and scattering for the infrared. Atmos. Meas. Tech., 13, 323–339, https://doi.org/10.5194/amt-13-323-2020.
Di, D., J. Li, Z. Li, J. Li, T. J. Schmit, and W. P. Menzel, 2021: Can current hyperspectral infrared sounders capture the small scale atmospheric water vapor spatial variations? Geophys. Res. Lett., 48, e2021GL095825, https://doi.org/10.1029/2021GL095825.
Diner, D. J., and Coauthors, 1998: Multi-angle Imaging SpectroRadiometer (MISR) instrument description and experiment overview. IEEE Trans. Geosci. Remote Sens., 36, 1072–1087, https://doi.org/10.1109/36.700992.
Diner, D. J., and Coauthors, 2013: The Airborne Multiangle Spectropolarimetric Imager (AirMSPI): A new tool for aerosol and cloud remote sensing. Atmos. Meas. Tech., 6, 2007–2025, https://doi.org/10.5194/amt-6-2007-2013.
Doviak, R., and D. Zrnić, 1993: Doppler Radar and Weather Observations. 2nd ed. Dover, 562 pp.
Emde, C., and Coauthors, 2016: The libRadtran software package for radiative transfer calculations (version 2.0.1). Geosci. Model Dev., 9, 1647–1672, https://doi.org/10.5194/gmd-9-1647-2016.
Eyre, J. R., S. J. English, and M. Forsythe, 2020: Assimilation of satellite data in numerical weather prediction. Part I: The early years. Quart. J. Roy. Meteor. Soc., 146, 49–68, https://doi.org/10.1002/qj.3654.
Eyre, J. R., W. Bell, J. Cotton, S. J. English, M. Forsythe, S. B. Healy, and E. G. Pavelin, 2022: Assimilation of satellite data in numerical weather prediction. Part II: Recent years. Quart. J. Roy. Meteor. Soc., 148, 521–556, https://doi.org/10.1002/qj.4228.
Forster, L., A. B. Davis, D. J. Diner, and B. Mayer, 2021: Toward cloud tomography from space using MISR and MODIS: Locating the “veiled core” in opaque convective clouds. J. Atmos. Sci., 78, 155–166, https://doi.org/10.1175/JAS-D-19-0262.1.
Hogan, R. J., and A. Battaglia, 2008: Fast lidar and radar multiple-scattering models. Part II: Wide-angle scattering using the time-dependent two-stream approximation. J. Atmos. Sci., 65, 3636–3651, https://doi.org/10.1175/2008JAS2643.1.
Horváth, A., and R. Davies, 2001: Simultaneous retrieval of cloud motion and height from polar-orbiter multiangle measurements. Geophys. Res. Lett., 28, 2915–2918, https://doi.org/10.1029/2001GL012951.
Irion, F. W., and Coauthors, 2018: Single-footprint retrievals of temperature, water vapor and cloud properties from AIRS. Atmos. Meas. Tech., 11, 971–995, https://doi.org/10.5194/amt-11-971-2018.
Irisov, V., T. Duly, V. Nguyen, and D. Masters, 2020: Radio occultation atmospheric profiling from the Spire nanosatellite constellation and its impact on weather forecasting. Proc. SPIE, 11531, 115310G, https://doi.org/10.1117/12.2574128.
Kalmus, P., S.-M. Wong, and J. Teixeira, 2015: The Pacific subtropical cloud transition: A MAGIC assessment of AIRS and ECMWF thermodynamic structure. IEEE Geosci. Remote Sens. Lett., 12, 1586–1590, https://doi.org/10.1109/LGRS.2015.2413771.
Kalmus, P., C. O. Ao, K.-N. Wang, M. P. Manzi, and J. Teixeira, 2022: A high-resolution planetary boundary layer height seasonal climatology from GNSS radio occultations. Remote Sens. Environ., 276, 113037, https://doi.org/10.1016/j.rse.2022.113037.
Kaul, C. M., J. Teixeira, and K. Suzuki, 2015: Sensitivities in large-eddy simulations of mixed-phase arctic stratocumulus clouds using a simple microphysics approach. Mon. Wea. Rev., 143, 4393–4421, https://doi.org/10.1175/MWR-D-14-00319.1.
Keihm, S., M. Janssen, and C. Ruf, 1995: TOPEX/Poseidon Microwave Radiometer (TMR). III. Wet troposphere range correction algorithm and pre-launch error budget. IEEE Trans. Geosci. Remote Sens., 33, 147–161, https://doi.org/10.1109/36.368213.
Khairoutdinov, M. F., and D. Randall, 2003: Cloud resolving modeling of the ARM summer 1997 IOP: Model formulation, results, uncertainties, and sensitivities. J. Atmos. Sci., 60, 607–625, https://doi.org/10.1175/1520-0469(2003)060<0607:CRMOTA>2.0.CO;2.
Khairoutdinov, M. F., and D. Randall, 2006: High-resolution simulation of shallow-to-deep convection transition over land. J. Atmos. Sci., 63, 3421–3436, https://doi.org/10.1175/JAS3810.1.
Khairoutdinov, M. F., S. K. Krueger, C.-H. Moeng, P. A. Bogenschutz, and D. A. Randall, 2009: Large-eddy simulation of maritime deep tropical convection. J. Adv. Model. Earth Syst., 1 (4), https://doi.org/10.3894/JAMES.2009.1.15.
King, M., and Coauthors, 2003: Cloud and aerosol properties, precipitable water, and profiles of temperature and water vapor from MODIS. IEEE Trans. Geosci. Remote Sens., 41, 442–458, https://doi.org/10.1109/TGRS.2002.808226.
Kulawik, S. S., and Coauthors, 2006: Implementation of cloud retrievals for Tropospheric Emission Spectrometer (TES) atmospheric retrievals: Part 1. Description and characterization of errors on trace gas retrievals. J. Geophys. Res., 111, D24204, https://doi.org/10.1029/2005JD006733.
Kurowski, M. J., K. Suselj, W. W. Grabowski, and J. Teixeira, 2018: Shallow-to-deep transition of continental moist convection: Cold pools, surface fluxes, and mesoscale organization. J. Atmos. Sci., 75, 4071–4090, https://doi.org/10.1175/JAS-D-18-0031.1.
Kurowski, M. J., W. W. Grabowski, K. Suselj, and J. Teixeira, 2020: The strong impact of weak horizontal convergence on continental shallow convection. J. Atmos. Sci., 77, 3119–3137, https://doi.org/10.1175/JAS-D-19-0351.1.
Kursinski, E. R., G. A. Hajj, J. T. Schofield, R. P. Linfield, and K. R. Hardy, 1997: Observing Earth’s atmosphere with radio occultation measurements using the global positioning system. J. Geophys. Res., 102, 23 429–23 465, https://doi.org/10.1029/97JD01569.
Levis, A., Y. Y. Schechner, A. Aides, and A. B. Davis, 2015: Airborne three-dimensional cloud tomography. 2015 IEEE Int. Conf. on Computer Vision, Santiago, Chile, IEEE, 3379–3387, https://doi.org/10.1109/ICCV.2015.386.
Levis, A., Y. Y. Schechner, and A. B. Davis, 2017: Multiple-scattering microphysics tomography. 2017 IEEE Conf. on Computer Vision and Pattern Recognition, Honolulu, HI, IEEE, 5797–5806, https://doi.org/10.1109/CVPR.2017.614.
Levis, A., Y. Y. Schechner, A. B. Davis, and J. Loveridge, 2020: Multi-view polarimetric scattering cloud tomography and retrieval of droplet size. Remote Sens., 12, 2831, https://doi.org/10.3390/rs12172831.
Liebe, H. J., G. A. Hufford, and M. G. Cotton, 1993: Propagation modeling of moist air and suspended water/ice particles at frequencies below 1000 GHz. Proc. AGARD Conf., Palma de Mallorca, Spain, AGARD, 3.1–3.10.
Matheou, G., and D. Chung, 2014: Large-eddy simulation of stratified turbulence. Part II: Application of the stretched-vortex model to the atmospheric boundary layer. J. Atmos. Sci., 71, 4439–4460, https://doi.org/10.1175/JAS-D-13-0306.1.
Matheou, G., D. Chung, L. Nuijens, B. Stevens, and J. Teixeira, 2011: On the fidelity of large-eddy simulation of shallow precipitating cumulus convection. Mon. Wea. Rev., 139, 2918–2939, https://doi.org/10.1175/2011MWR3599.1.
Mayer, B., 2009: Radiative transfer in the cloudy atmosphere. EPJ Web Conf., 1, 75–99, https://doi.org/10.1140/epjconf/e2009-00912-1.
Mayer, B., and A. Kylling, 2005: The libRadtran software package for radiative transfer calculations—Description and examples of use. Atmos. Chem. Phys., 5, 1855–1877, https://doi.org/10.5194/acp-5-1855-2005.
Meissner, T., and F. J. Wentz, 2012: The emissivity of the ocean surface between 6 and 90 GHz over a large range of wind speeds and Earth incidence angles. IEEE Trans. Geosci. Remote Sens., 50, 3004–3026, https://doi.org/10.1109/TGRS.2011.2179662.
Menzel, W. P., T. J. Schmit, P. Zhang, and J. Li, 2018: Satellite-based Atmospheric Infrared Sounder development and applications. Bull. Amer. Meteor. Soc., 99, 583–603, https://doi.org/10.1175/BAMS-D-16-0293.1.
Millán, L., R. Roy, and M. Lebsock, 2020: Assessment of global total column water vapor sounding using a spaceborne differential absorption radar. Atmos. Meas. Tech., 13, 5193–5205, https://doi.org/10.5194/amt-13-5193-2020.
Mitra, A., L. Di Girolamo, Y. Hong, Y. Zhan, and K. J. Mueller, 2021: Assessment and error analysis of Terra-MODIS and MISR cloud-top heights through comparison with ISS-CATS lidar. J. Geophys. Res. Atmos., 126, e2020JD034281, https://doi.org/10.1029/2020JD034281.
National Academies of Sciences, Engineering, and Medicine, 2008: Earth Observations from Space: The First 50 Years of Scientific Achievements. National Academies Press, 129 pp., https://doi.org/10.17226/11991.
National Academies of Sciences, Engineering, and Medicine, 2018: Thriving on Our Changing Planet: A Decadal Strategy for Earth Observation from Space. National Academies Press, 716 pp.
Nehrir, A., C. Kiemle, and M. Lebsock, 2017: Emerging technologies and synergies for airborne and space-based measurements of water vapor profiles. Surv. Geophys., 38, 1445–1482, https://doi.org/10.1007/s10712-017-9448-9.
Prusa, J. M., P. K. Smolarkiewicz, and A. A. Wyszogrodzki, 2008: EULAG, a computational model for multiscale flows. Comput. Fluids, 37, 1193–1207, https://doi.org/10.1016/j.compfluid.2007.12.001.
Read, W. G., Z. Shippony, and W. Van Snyder, 2004: EOS MLS forward model. Jet Propulsion Laboratory Algorithm Theoretical Basis Doc., 248 pp., https://mls.jpl.nasa.gov/data/eos_fwd_mdl_atbd.pdf.
Richardson, M. T., D. R. Thompson, M. J. Kurowski, and M. D. Lebsock, 2021: Boundary layer water vapour statistics from high-spatial-resolution spaceborne imaging spectroscopy. Atmos. Meas. Tech., 14, 5555–5576, https://doi.org/10.5194/amt-14-5555-2021.
Richardson, M. T., D. R. Thompson, M. J. Kurowski, and M. D. Lebsock, 2022: New sampling strategy mitigates a solar-geometry-induced bias in sub-kilometre vapour scaling statistics derived from imaging spectroscopy. Atmos. Meas. Tech., 15, 117–129, https://doi.org/10.5194/amt-15-117-2022.
Rieckh, T., J. P. Sjoberg, and R. A. Anthes, 2021: The three-cornered hat method for estimating error variances of three or more atmospheric datasets. Part II: Evaluating radio occultation and radiosonde observations, global model forecasts, and reanalyses. J. Atmos. Oceanic Technol., 38, 1777–1796, https://doi.org/10.1175/JTECH-D-20-0209.1.
Rodgers, C. D., 2000: Inverse Methods for Atmospheric Sounding: Theory and Practice. World Scientific, 238 pp., https://doi.org/10.1142/3171.
Rosenkranz, P. W., 1998: Water vapor microwave continuum absorption: A comparison of measurements and models. Radio Sci., 33, 919–928, https://doi.org/10.1029/98RS01182.
Roy, R. J., M. Lebsock, L. Millán, and K. B. Cooper, 2020: Validation of a G-band differential absorption cloud radar for humidity remote sensing. J. Atmos. Oceanic Technol., 37, 1085–1102, https://doi.org/10.1175/JTECH-D-19-0122.1.
Roy, R. J., M. Lebsock, and M. J. Kurowski, 2021: Spaceborne differential absorption radar water vapor retrieval capabilities in tropical and subtropical boundary layer cloud regimes. Atmos. Meas. Tech., 14, 6443–6468, https://doi.org/10.5194/amt-14-6443-2021.
Roy, R. J., and Coauthors, 2022: First airborne measurements with a G-band differential absorption radar. IEEE Trans. Geosci. Remote Sens., 60, 5108115, https://doi.org/10.1109/TGRS.2021.3134670.
Schreiner, W., and Coauthors, 2020: COSMIC-2 radio occultation constellation: First results. Geophys. Res. Lett., 47, e2019GL086841, https://doi.org/10.1029/2019GL086841.
Schwarz, J., G. Kirchengast, and M. Schwaerz, 2017: Integrating uncertainty propagation in GNSS radio occultation retrieval: From bending angle to dry-air atmospheric profiles. Earth Space Sci., 4, 200–228, https://doi.org/10.1002/2016EA000234.
Schwarz, J., G. Kirchengast, and M. Schwaerz, 2018: Integrating uncertainty propagation in GNSS radio occultation retrieval: From excess phase to atmospheric bending angle profiles. Atmos. Meas. Tech., 11, 2601–2631, https://doi.org/10.5194/amt-11-2601-2018.
Siebesma, A. P., and Coauthors, 2003: A large eddy simulation intercomparison study of shallow cumulus convection. J. Atmos. Sci., 60, 1201–1219, https://doi.org/10.1175/1520-0469(2003)60<1201:ALESIS>2.0.CO;2.
Skamarock, W. C., and Coauthors, 2008: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, 113 pp., https://doi.org/10.5065/D68S4MVH.
Stevens, B., and Coauthors, 2001: Simulations of trade wind cumuli under a strong inversion. J. Atmos. Sci., 58, 1870–1891, https://doi.org/10.1175/1520-0469(2001)058<1870:SOTWCU>2.0.CO;2.
Stevens, B., and Coauthors, 2005: Evaluation of large-eddy simulations via observations of nocturnal marine stratocumulus. Mon. Wea. Rev., 133, 1443–1462, https://doi.org/10.1175/MWR2930.1.
Stoll, R., J. A. Gibbs, S. T. Salesky, W. Anderson, and M. Calaf, 2020: Large-eddy simulation of the atmospheric boundary layer. Bound.-Layer Meteor., 177, 541–581, https://doi.org/10.1007/s10546-020-00556-3.
Susskind, J., C. Barnet, and J. Blaisdell, 2003: Retrieval of atmospheric and surface parameters from AIRS/AMSU/HSB data in the presence of clouds. IEEE Trans. Geosci. Remote Sens., 41, 390–409, https://doi.org/10.1109/TGRS.2002.808236.
Tanelli, S., S. L. Durden, E. Im, K. S. Pak, D. G. Reinke, P. Partain, J. M. Haynes, and R. T. Marchand, 2008: CloudSat’s cloud profiling radar after two years in orbit: Performance, calibration, and processing. IEEE Trans. Geosci. Remote Sens., 46, 3560–3573, https://doi.org/10.1109/TGRS.2008.2002030.
Teixeira, J., and Coauthors, 2021: Toward a global planetary boundary layer observing system: The NASA PBL Incubation Study Team Report. NASA Rep., 134 pp., https://ntrs.nasa.gov/api/citations/20230001633/downloads/AFridlindPBLTowardsReport.pdf
Thies, B., and J. Bendix, 2011: Satellite based remote sensing of weather and climate: Recent achievements and future perspectives. Meteor. Appl., 18, 262–295, https://doi.org/10.1002/met.288.
Thompson, D. R., V. Natraj, R. O. Green, M. C. Helmlinger, B.-C. Gao, and M. L. Eastwood, 2018: Optimal estimation for imaging spectrometer atmospheric correction. Remote Sens. Environ., 216, 355–373, https://doi.org/10.1016/j.rse.2018.07.003.
Thompson, D. R., and Coauthors, 2019: A unified approach to estimate land and water reflectances with uncertainties for coastal imaging spectroscopy. Remote Sens. Environ., 231, 111198, https://doi.org/10.1016/j.rse.2019.05.017.
Thompson, D. R., B. H. Kahn, P. G. Brodrick, M. D. Lebsock, M. Richardson, and R. O. Green, 2021: Spectroscopic imaging of sub-kilometer spatial structure in lower-tropospheric water vapor. Atmos. Meas. Tech., 14, 2827–2840, https://doi.org/10.5194/amt-14-2827-2021.
Tzabari, M., and Coauthors, 2021: CloudCT 3D volumetric tomography: Considerations for imager preference, comparing visible light, short-wave infrared, and polarized imagers. Proc. SPIE, 11833, 1183304, https://doi.org/10.1117/12.2594134.
vanZanten, M. C., and Coauthors, 2011: Controls on precipitation and cloudiness in simulations of trade-wind cumulus as observed during RICO. J. Adv. Model. Earth Syst., 3, M06001, https://doi.org/10.1029/2011MS000056.
von Engeln, A., G. Nedoluha, G. Kirchengast, and S. Bühler, 2003: One-dimensional variational (1-D Var) retrieval of temperature, water vapor, and a reference pressure from radio occultation measurements: A sensitivity analysis. J. Geophys. Res., 108, 4337, https://doi.org/10.1029/2002JD002908.
Weisman, M. L., and J. B. Klemp, 1982: The dependence of numerically simulated convective storms on vertical wind shear and buoyancy. Mon. Wea. Rev., 110, 504–520, https://doi.org/10.1175/1520-0493(1982)110<0504:TDONSC>2.0.CO;2.
Zeng, X., and Coauthors, 2020: Use of observing system simulation experiments in the United States. Bull. Amer. Meteor. Soc., 101, E1427–E1438, https://doi.org/10.1175/BAMS-D-19-0155.1.
Zhang, H., Y. Kuo, and S. Sokolovskiy, 2023: Assimilation of radio occultation data using measurement-based observation error specification: Preliminary results. Mon. Wea. Rev., 151, 589–601, https://doi.org/10.1175/MWR-D-22-0122.1.