SEARCH · Engineering Papers

Results for “retrievals”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 289 records · Page 16

All sources NASA NTRS DOE OSTI

NASA NTRS2020

Simultaneous Retrieval of Surface Roughness Parameters for Bare Soils from Combined Active-Passive Microwave SMAP Observations

An active-passive microwave retrieval algorithm for simultaneous determination of soil surface roughness parameters [vertical root-mean-square (RMS) height (s) and horizontal correlation length (l)] is presented for bare soils. The algorithm is based on active-passive microwave covariation, including the improved Integral Equation Method (I²EM), and is tested with global soil moisture active passive (SMAP) observations. The estimated retrieval results for s and l are overall consistent with values in the literature, indicating the validity of the proposed algorithm. Sensitivity analyses showed that the developed roughness retrieval algorithm is independent of permittivity for εₛ > 10 [-]. Furthermore, the physical model basis of this approach (I²EM) allows the application of different autocorrelation functions (ACF), such as Gaussian and exponential ACFs. Global roughness retrieval results confirm bare areas in deserts such as Sahara or Gobi. However, the type of ACF used within roughness parameter estimation is important. Retrieval results for the Gaussian ACF describe a rougher surface than retrieval results for the exponential ACF. No correlations were found between roughness results and the amount of precipitation or the soil texture, which could be due to the coarse spatial resolution of the SMAP data. The extension of this approach to vegetated soils is planned as an add-on study.

correlation length↗

NASA NTRS2020

Optimized Profile Retrievals of Aerosol Microphysical Properties from Simulated Spaceborne Multiwavelength Lidar

This work is an expanded study of one previously published onretrievals of aerosol microphysical properties from space-borne multiwavelengthlidarmeasurements. The earlier studiesand this one weredone in the framework of the NASA Aerosol-Clouds-Ecosystems (now the Aerosol Clouds Convection and Precipitation) NASA mission. The focus here is on the capabilities of a simulated spacebornemultiwavelengthlidar system for retrieving aerosol complex refractive index (m = mr+ imi) and spectral single scattering albedo (SSA(λ)), although other bulk parameters such as effective (reff) radius and particle volume (V) and surface (S) concentrations are also studied. The novelty presented here is the use of recently published, case dependent optimized-constraints on the microphysical retrievals using three backscattering coefficients (β) at 355, 532 and 1064 nm and two extinction coefficients (α) at 355 and 532 nm, typically known as the stand-alone 3β+2α lidar inversion. Case-dependent optimized-constraints (CDOC) limit the ranges of refractive index, both real (mr) and imaginary (mi) parts, and of radii that are permitted in the retrievals. Such constraints are selected directly from the 3β+2α41measurements through an analysis of the relationship between spectral dependence of aerosol extinction-to-backscatter ratios (LR) and the Ångström exponent of extinction. The analyses presented here for different sets ofsize distributions and refractive indices reveal that the direct determination of CDOCareonly feasible for cases where the uncertaintiesin the input optical data areless than 15 %.Forthe same simulated spacebornesystem and yield than in Whiteman et al., (2018), we demonstrated that the use of CDOC as essential for the retrievals of refractive index and also largely improved retrieval of bulk parameters. A discussion of the global representativeness of CDOC is presented using simulated lidar data from a 24-hour satellite track using GEOS model output to initialize the lidar simulator.We found that CDOCare representative of many aerosol mixtures in spite of some outliers (e.g. highly hydrated particles) associatedwith the assumptions of bimodal size distributions and of the same refractive index for fine and coarse modes. Moreover, sensitivity tests performed using synthetic data reveal that retrievals of imaginary refractive index (mi) and SSA are extremely sensitive to β(355).

NASA Aerosol-Clouds-Ecosystems↗

NASA NTRS2020

Simultaneous Retrieval of Surface Roughness Parameters for Bare Soils From Combined Active–Passive Microwave SMAP Observations

An active–passive microwave retrieval algorithm for simultaneous determination of soil surface roughness parameters [vertical root-mean-square (RMS) height ( s ) and horizontal correlation length ( l )] is presented for bare soils. The algorithm is based on active–passive microwave covariation, including the improved Integral Equation Method (I(exp 2) EM), and is tested with global soil moisture active passive (SMAP) observations. The estimated retrieval results for s and l are overall consistent with values in the literature, indicating the validity of the proposed algorithm. Sensitivity analyses showed that the developed roughness retrieval algorithm is independent of permittivity for E(sub s) > 10 [-]. Furthermore, the physical model basis of this approach (I(exp 2) EM) allows the application of different autocorrelation functions (ACF), such as Gaussian and exponential ACFs. Global roughness retrieval results confirm bare areas in deserts such as Sahara or Gobi. However, the type of ACF used within roughness parameter estimation is important. Retrieval results for the Gaussian ACF describe a rougher surface than retrieval results for the exponential ACF. No correlations were found between roughness results and the amount of precipitation or the soil texture, which could be due to the coarse spatial resolution of the SMAP data. The extension of this approach to vegetated soils is planned as an add-on study.

Correlation length↗

NASA NTRS2019

An Ensemble of Bayesian Neural Networks for Exoplanetary Atmospheric Retrieval

Machine learning (ML) is now used in many areas of astrophysics, from detecting exoplanets in Kepler transit signals to removing telescope systematics. Recent work demonstrated the potential of using ML algorithms for atmospheric retrieval by implementing a random forest (RF) to perform retrievals in seconds that are consistent with the traditional, computationally expensive nested-sampling retrieval method. We expand upon their approach by presenting a new ML model, plan-net, based on an ensemble of Bayesian neural networks (BNNs) that yields more accurate inferences than the RF for the same data set of synthetic transmission spectra. We demonstrate that an ensemble provides greater accuracy and more robust uncertainties than a single model. In addition to being the first to use BNNs for atmospheric retrieval, we also introduce a new loss function for BNNs that learns correlations between the model outputs. Importantly, we show that designing ML models to explicitly incorporate domain-specific knowledge both improves performance and provides additional insight by inferring the covariance of the retrieved atmospheric parameters. We apply plan-net to the Hubble Space Telescope Wide Field Camera 3 transmission spectrum for WASP-12b and retrieve an isothermal temperature and water abundance consistent with the literature. We highlight that our method is flexible and can be expanded to higher resolution spectra and a larger number of atmospheric parameters.

Adam D. Cobb↗

NASA NTRS2021

Evaluation of GEOS Precipitation Flagging for SMAP Soil Moisture Retrieval Accuracy

The precipitation flag in the Soil Moisture Active Passive (SMAP) Level 2 passive soil moisture (L2SMP) retrieval product indicates the presence or absence of heavy precipitation at the time of the SMAP overpass. The flag is based on precipitation estimates from the Goddard Earth Observing System (GEOS) Forward Processing numerical weather prediction system. An error in flagging during an active or recent precipitation event can produce either 1) an overestimation of soil moisture due to short-term surface wetting of vegetation and/or surface ponding (if soil moisture retrieval was attempted in the presence of rain) or 2) an unnecessary nonretrieval of soil moisture and loss of data (if retrieval is flagged due to an erroneous indication of rain). Satellite precipitation estimates from the Integrated Multisatellite Retrievals for GPM (IMERG), version 06, Early Run (latency of ~4 h) precipitationCal product are used here to evaluate the GEOS-based precipitation flag in the L2SMP product for both the 1800 local time (LT) ascending and 0600 LT descending SMAP overpasses over the first five years of the mission (2015–20). Consisting of blended precipitation measurements from the Global Precipitation Mission (GPM) satellite constellation, IMERG is treated as the “truth” when comparing to the GEOS model forecasts of precipitation used by SMAP. Key results include (i) IMERG measurements generally show higher spatial variability than the GEOS forecast precipitation, (ii) the IMERG product has a higher frequency of light precipitation amounts, and (iii) the effect of incorporating IMERG rainfall measurements in lieu of GEOS precipitation forecasts are minimal on the L2SMP retrieval accuracy (determined vs in situ soil moisture measurements at core validation sites). Our results indicate that L2SMP retrievals continue to meet the mission’s accuracy requirement [standard deviation of the unbiased RMSE (ubRMSE) less than 0.04 cu. m/cu. m].

IMERG-precipitationCal↗

NASA NTRS2020

Ground-based retrievals of aerosol column absorption in the UV spectral region and their implications for GEMS measurements

Quantifying the spectral variation of column aerosol absorption in the ultraviolet (UV) and visible (Vis) wavelengths is required for accurate satellite-based aerosol and trace-gas retrievals. Retrievals of the column-averaged imaginary part of refractive index and single scattering albedo (SSA) in the UV–Vis range have been performed at Yonsei University, Seoul, Korea, since 2016 by combining co-located measurements from the NASA Aerosol Robotic Network (AERONET) Cimel sun-sky photometer, the Ultraviolet Multifilter Rotating Shadowband Radiometer (UV-MFRSR), the SKYNET Prede sky radiometer, and the NASA Pandora sun spectrometer. We investigated the spectral variation of column-averaged imaginary part of refractive index for UV–Vis wavelengths to refine models used in our aerosol retrieval algorithm to process measurements from the upcoming Geostationary Environment Monitoring Satellite (GEMS). The retrieved imaginary part of refractive index for highly absorbing fine pollution particles (BC), dust (DS), and non-absorbing (NA) particles in the selected UV–Vis range (380–440 nm) showed 0–20%, 30%, and 0–40% of spectral dependence, respectively. Retrievals of Ozone Monitoring Instrument (OMI) measurement data using the improved aerosol model showed improved correlation with AERONET data compared to the old algorithm that did not properly account for aerosol absorption effects. These results corroborate the advantage of using local climatology derived from ground-based UV–Vis spectral aerosol absorption measurements for satellite GEMS aerosol retrievals over East Asia. Moreover, this study reveals that spectral variations in the UV column aerosol absorption in East Asia differ from those in other regions.

ultraviolet (UV) and visible (Vis) wavelengths↗

NASA NTRS2021

Assessment of the Advanced Very High Resolution Radiometer (AVHRR) for Snowfall Retrieval in High Latitudes Using CloudSat and Machine Learning

Precipitation retrieval is a challenging topic, especially in high latitudes (HL), and current precipitation products face ample challenges over these regions. This study investigates the potential of the Advanced Very High Resolution Radiometer (AVHRR) for snowfall retrieval in HL using CloudSat radar information and machine learning (ML). With all the known limitations, AVHRR observations should be considered for HL snowfall retrieval because 1) AVHRR data have been continuously collected for about four decades on multiple platforms with global coverage, and similar observations will likely continue in the future; 2) current passive microwave satellite precipitation products have several issues over snow and ice surfaces; and 3) good coincident observations between AVHRR and CloudSat are available for training ML algorithms. Using ML, snowfall rate was retrieved from AVHRR’s brightness temperature and cloud probability, as well as auxiliary information provided by numerical reanalysis. The results indicate that the ML-based retrieval algorithm is capable of detection and estimation of snowfall with comparable or better statistical scores than those obtained from the Atmospheric Infrared Sounder (AIRS) and two passive microwave sensors contributing to the Global Precipitation Measurement (GPM) mission constellation. The outcomes also suggest that AVHRR-based snowfall retrievals are spatially and temporally reasonable and can be considered as a quantitatively useful input to the merged precipitation products that require frequent sampling or long-term records.

Mohammad Reza↗

NASA NTRSDate not supplied

Assessment of the Accuracy and Consistency of Cloud Property Retrievals from Passive Radiometers with Varying Spectral Information

Over the past few decades, space-based passive radiometers have been used to study the spatial distribution of clouds and some of their basic properties. Modern radiometers have significantly higher spatial resolution and collect more spectral information (i.e., have more spectral bands) than early instruments, and these advances have allowed researchers to characterize cloud physical and microphysical properties with greater detail. There is significant interest in utilizing as much of the existing satellite record as possible to study global and regional changes in cloud occurrence and distribution as well as the associated changes in reflected shortwave and outgoing longwave fluxes. Cloud property retrieval algorithms such as the NASA Satellite Cloud Optical and Radiative Property System (SatCORPS) have traditionally utilized as much spectral information as possible in order to maximize the accuracy of the retrievals. However, changes to the retrieval algorithm over time in order to take advantage of progressively more spectral information results in inconsistencies in cloud properties which in turn results in inconsistencies with products derived from these properties. It is not possible to implement modern state-of-the-art algorithms to earlier instruments which lack diverse spectral information, but many instruments since the early 2000s generally have approximately similar spectral bands in the visible (~0.65 microns), longwave infrared (~11 microns), and in many cases, the midwave infrared (~3.8 microns) and water vapor absorption (~6.7 microns), portions of the electromagnetic spectrum. Different versions of SatCORPS have been implemented with either the full set of spectral information collected by imagers such as GOES-16 and MODIS or with a reduced set of spectral bands in order to be more consistent with past imagers that lack such diverse spectral information. This study uses independent cloud property retrievals from the Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) to assess the accuracy of the retrievals (e.g., cloud fraction, thermodynamic phase, altitude) from the different retrieval methods and to demonstrate the advantages and disadvantages of using a reduced set of common spectral bands in order to achieve temporal and spatial consistency across different instruments.

Christopher R. Yost↗

NASA NTRS2022

Polarization performance simulation for the GeoXO atmospheric composition instrument: NO2 retrieval impacts

NOAA’s Geostationary Extended Observations (GeoXO) constellation will continue and expand on the capabilities of the current generation of geostationary satellite systems to support US weather, ocean, atmosphere, and climate operations. It is planned to consist of a dedicated atmospheric composition instrument (ACX) to support air quality forecasting and monitoring by providing similar capabilities to missions such as TEMPO (Tropospheric Emission: Monitoring Pollution), currently planned to launch in 2023, and Ozone Monitoring Instrument (OMI), TROPOMI (TROPOspheric Monitoring Instrument), and GEMS (Geostationary Environment Monitoring Spectrometer) currently in operation. As the early phases of ACX development are progressing, design trade-offs are being considered to understand the relationship between instrument design choices and trace gas retrieval impacts. Some of these choices will affect the instrument polarization sensitivity (PS), which can have radiometric impacts on environmental satellite observations. We conducted a study to investigate how such radiometric impacts can affect NO2 retrievals by exploring their sensitivities to time of day, location, and scene type with an ACX instrument model that incorporates PS. The study addresses the basic steps of operational NO2 retrievals: the spectral fitting step and the conversion of slant column to vertical column via the air mass factor (AMF). The spectral fitting step was performed by generating at-sensor radiance from a clear sky scene with a known NO2 amount, the application of an instrument model including both instrument PS and noise, and a physical retrieval. The spectral fitting step was found to mitigate the impacts of instrument PS. The AMF-related step was considered for clear sky and partially cloudy scenes, where instrument PS can lead to errors in interpreting the cloud content, propagating to AMF errors and finally to NO2 retrieval errors. For this step, the NO2 retrieval impacts were small but non-negligible for high NO2 amounts; we estimated that a typical high NO2 amount can cause maximum of 0.25 x 10^15 molecules/cm^2 for a PS of 5%. These simulation capabilities were designed to aid in the development of a GeoXO atmospheric composition instrument that will improve our ability to monitor and understand the Earth’s atmosphere.

trace gases↗

NASA NTRS2022

Multi-Frequency Radiometer-Based Soil Moisture Retrieval and Algorithm Parameterization Using In Situ Sites

L-band brightness temperature (TB) has been shown to provide the best sensitivity to soil moisture (SM) although C- and X-band based products offer a longer time-series from satellite-based measurements. Currently, global coverage SM is routinely produced from spaceborne measurements using all three frequency bands, but despite continued validation efforts of the products, the relative characteristics and performance of these observations have not been fully established. Therefore, this study focused on the parametrization of SM retrieval algorithms at L-, C- and X-bands using TB observations from the L-band radiometer on NASA's SM Active Passive (SMAP) mission and the C- and X-band channels of JAXA's Advanced Microwave Scanning Radiometer 2 (AMSR2) onboard the GCOM-W satellite. These can be applied in global SM retrieval algorithms using either one of the frequencies or a combination of them. The reference in situ SM data was obtained from 12 core validation sites across various land cover types around the world. The investigation highlighted the known challenges of retrieving SM from C- and X-band data compared to the higher sensitivity of the L-band data. Even with a site-specific retrieval algorithm parameterization, the mean correlation of the C- and X-band retrievals for the core validation site SM measurement were much lower than that for L-band, being 0.52 (0.54) and 0.45 (0.47) for vertical (horizontal) polarization, respectively, while for the L-band retrieval the corresponding values were 0.81 (0.77). The parameterization exercise showed that matching the C- and X-band TB measurements with an emission model was not difficult; the problem was relating the observations to SM under the influence of large roughness and vegetation effects. As a result, parameter optimization produced values for some sites that were not realistic or did not allow any practical sensitivity to SM at C- and X-band. Considering the L-band observations, the parameter optimization resulted in superior bias performance as compared to the operational SMAP product parameterization, but the sensitivity to SM changes (R and unbiased root mean square difference) did not improve markedly, or in some cases degraded at the expense of a smaller bias.

passive microwave↗

NASA NTRSDate not supplied

A Modeling and Simulation Study for the GeoXO Atmospheric Composition Instrument (ACX): System Level SO2 and O3 Retrieval Performance as a Function of SNR

NOAA’s Geostationary Extended Observations (GeoXO) program is planning to include a hyperspectral UV/visible atmospheric composition instrument in geostationary orbit slated for operations by the early 2030s. Trace gas retrievals provide critical information about our Earth’s system as it relates to both natural and anthropogenic activity. This work focused on the impacts of the signal-to-noise ratio (SNR) on the retrieval uncertainty of SO2 and O3. An end-to-end physics-based imaging system and scene modeling and simulation framework was developed to assess instrument performance considerations in support of the planned GeoXO atmospheric composition instrument (ACX). The framework is composed of three primary components: 1) simulated at-sensor radiance via radiative transfer models of a known atmospheric composition, 2) a noise model driven by previous and planned sensor specifications, and 3) a physical retrieval algorithm. In this analysis, the tropospheric concentrations of SO2 and O3 in the atmospheric column were varied in the boundary layer (0-3 km) and at an elevation of 5 km with a thickness of 1-2 km; the uncertainty in retrieval was studied for various SNR levels for multiple viewing geometries and solar zenith angles. This effort supports the development of the GeoXO ACX instrument by providing a simulated ACX performance assessment of the system-level retrieval performance as a function of SNR, particularly with respect to the wavelength range of interest (approximately 305 – 340 nm) for SO2 and O3 trace gas retrievals.

Monica Cook↗

NASA NTRSDate not supplied

Remote Sensing of Hail from Space: Retrievals, Climatologies, and Challenges Going Forward

In addition to the myriad threats that severe hailstorms pose to society, infrastructure and agriculture, severe hail is difficult to measure in situ, and surface-based hail reporting and detection methods are inconsistent and subject to geographical or societal biases. This motivates the use of spaceborne remote-sensing platforms to retrieve hail and construct climatologies in the most globally consistent way. Passive-microwave algorithms leverage the sensitivity of spaceborne passive-microwave radiometers to scattering by hail, particularly in the channels from 10 to 89 GHz. These retrievals are used to construct global climatologies of severe hail. The Bang and Cecil (2019) retrievals has been applied to several different spaceborne sensors: the Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI), Global Precipitation Measurement Mission (GPM) Microwave Imager, Advanced Microwave Scanning Radiometer for EOS (AMSR-E), and Advanced Microwave Scanning Radiometer 2 (AMSR2) sensors and used to construct [near] global climatologies of severe hail. This retrieval, and others, have been tested using Global Precipitation Measurement (GPM) Ku-band precipitation radar, to assess their effectiveness and regional variability. A successful retrieval and climatology are those that correspond tightly to radar reflectivity and give minimal appearance of regional biases, especially with latitude. Satellite platforms offer consistent observations, even in remote, data-sparse, and oceanic regions that ground-based networks exclude. There are, however, potential disconnects between the processes identified aloft by the satellite and the resultant weather at the ground, leading to uncertainties in the retrievals that may propagate into satellite-based climatologies, particularly in the Tropics, where there are abundant strong - but not necessarily hailing - storms that are strongly represented in the current satellite climatologies. There are ongoing efforts to assess and mitigate the contributing factors to these uncertainties, chiefly among them the effects of non-uniform beam filling in the passive-microwave footprint, and the relationships between the size distributions of hailstones aloft and the dynamic processes and environments with which they interact throughout their trajectories.

Sarah D Bang↗

NASA NTRSDate not supplied

Airborne HSRL Assessment of CALIOP Aerosol Extinction Retrievals Constrained by Column AOD

Current operational CALIOP aerosol extinction profile retrievals usually rely on accurately specifying the relationship between aerosol extinction and backscattering (i.e. lidar ratio). Uncertainties in the assigned lidar ratios are typically the largest source of systematic error (~30-50%) in the CALIOP retrievals of aerosol extinction, backscatter, and aerosol optical depth. Alternatively, column aerosol optical depth (AOD) can be used to solve the lidar equation and obtain column-equivalent lidar ratios and retrieve aerosol extinction profiles from the CALIOP attenuated backscatter profiles. These derived lidar ratios correspond to the aerosols in the altitude range from 0-7 km. We derive column-equivalent aerosol lidar ratios and aerosol extinction profiles from CALIOP attenuated backscatter profiles constrained using co-located column AODs provided by several instruments/techniques: Synergized Optical Depth of Aerosols (SODA), Ocean-Derived Column Optical Depths (ODCOD), MODIS (dark target), MODIS (Multi-Angle Implementation of Atmospheric Correction-MAIAC), and PARASOL (Generalized Retrieval of Aerosol and Surface Properties-GRASP). These various retrievals of AOD and the column-average aerosol lidar ratios and aerosol extinction profiles derived from CALIOP using these AOD constraints are evaluated using the extensive record of coincident and co-located airborne HSRL measurements of AOD, aerosol lidar ratio, and aerosol extinction profiles acquired during more than 140 HSRL underflights of CALIPSO since 2006. The HSRL technique allows for the independent measurement of extinction and backscatter without the need for external constraints or assumptions regarding the lidar ratios. Initial comparisons with these coincident HSRL aerosol extinction profiles show that the CALIOP aerosol extinction profiles retrieved using these various AOD constraints are generally in better agreement with the HSRL measurements than aerosol extinction profiles from the operational techniques. The quality of agreement depends on the accuracy and magnitude of the AOD constraint. We present these comparisons of AOD, column aerosol lidar ratio, and aerosol extinction profiles for each of the AOD constraints described above.

lidar↗

NASA NTRSDate not supplied

Principal Component and Machine Learning Approach to Gap Fill Hyperspectral Ocean Color Satellite Retrievals

Retrievals of ocean color properties from space are important for monitoring the health of the ocean ecosystem but such retrievals can be limited spatially due to conditions such as clouds, aerosols, and sun glint. Gap filling of ocean color retrievals is typically performed by combining retrievals from multiple satellites or temporally averaging multiple days of retrievals. Despite these techniques large gaps still exist posing challenges for near real time monitoring of events like harmful algae blooms. To address these limitations, we developed a spatial gap filling approach applying machine learning approach to hyperspectral instruments to learn how to perform an atmospheric correction under challenging retrieval conditions. In this approach a principal component analysis is used to decompose the hyperspectral measurements into spectral components that describe the scattering and absorption of the atmosphere mixed with the surface spectral signatures. The coefficients of the principal components are used to train a neural network to predict ocean color properties derived from a standard MODIS ocean color algorithm. We apply the approach to two hyperspectral UV/VIS sensors, the Ozone Monitoring Instrument (OMI) and TROPOspheric Monitoring Instrument (TROPOMI) to show that it can be used to estimate ocean color properties such as chlorophyll, remote sensing reflectance, and fluorescence line height. This method could be used as a gap-filling technique for the future Ocean Color Instrument (OCI) onboard upcoming NASA's Plankton, Aerosol Cloud, ocean Ecosystem (PACE) satellite to provide additional information for monitoring the health of our global oceans. Additionally, it could be applied to the first NASA and Smithsonian geostationary Tropospheric Emissions: Monitoring of Pollution (TEMPO) spectrometer to better understand diurnal variability in inland and coastal ocean ecology.

Zachary Fasnacht↗

NASA NTRSDate not supplied

Spaceborne Remote Sensing of Hail: Retrievals, Climatologies, and Challenges Going Forward

Sarah D Bang↗

NASA NTRSDate not supplied

A Principal Component and Machine Learning Approach to Spatially Gap Fill Hyperspectral Ocean Color Satellite Retrievals

Retrievals of ocean color properties from space are important for monitoring the health of the ocean ecosystem but such retrievals tend to be limited in spatial coverage due to conditions such as clouds, aerosols, and sun glint. Gap filling of ocean color retrievals is typically performed by combining retrievals from multiple satellites or temporally averaging multiple days of retrievals but despite these techniques large gaps still exist posing challenges for near real time monitoring of events like harmful algae blooms. To address these limitations, we propose a spatial gap filling approach using machine learning to learn how to perform an atmospheric correction under challenging retrieval conditions. In this approach a principal component analysis is used to decompose the hyperspectral measurements into spectral features that describe the scattering and absorption of the atmosphere as well as the underlying surface. The coefficients of the principal components are then used to train a neural network to predict ocean color properties derived from a standard ocean color algorithm such as the MODIS atmospheric correction algorithm. This machine learning approach is independent of a priori information and does not rely on any radiative transfer modeling. We apply the approach to two hyperspectral UV/VIS instruments, the Ozone Monitoring Instrument (OMI) and TROPOspheric Monitoring Instrument (TROPOMI) to show that it can be used to estimate ocean color properties such as chlorophyll, remote sensing reflectance, and fluorescence line height. This method could be used as a gap-filling technique for the future Ocean Color Instrument (OCI) which will be onboard NASA's Plankton, Aerosol Cloud, ocean Ecosystem (PACE) ocean color satellite to provide additional information for monitoring the health of our global oceans. Additionally, it could be applied to the geostationary satellite Tropospheric Emissions: Monitoring of Pollution (TEMPO) to better understand diurnal variability in ocean ecology.

MODIS atmospheric correction algorithm↗

NASA NTRS2023

High-Resolution Snowstorm Measurements and Retrievals Using Cross-Platform Multi-Frequency and Polarimetric Radars

Many studies have provided microphysical retrievals using radars of different frequencies, platforms, and methodologies. However, little is known about the consistency of retrievals derived from different radar platforms (i.e., airborne or spaceborne vs. groundbased) and their methodologies. This study is the first to directly compare snow mean volume diameter (Dm) retrievals from both nadir-pointing airborne multi-frequency radars and ground-based polarimetric range-height indicator (RHI) radar scans along a coincident airborne flight track. Compared to typical airborne/spaceborne/ground validation studies which at best provide O(1 km) resolution, these two platforms together provide directly collocated, high-resolution O(10m) retrievals. Dm retrieval values were over all both largest and smallest when using two separate dual-wavelength methods from aircraft measurements. A Triple-frequency analysis suggests the possibility that snow aggregates were generally composed of needles. The results shown here can be used as a benchmark for comparing retrieval methodologies.

Edwin L. Dunnavan↗

DOE OSTI2021

Fast Local Spatial Verification for Feature-Agnostic Large-Scale Image Retrieval

Images from social media can reflect diverse viewpoints, heated arguments, and expressions of creativity, adding new complexity to retrieval tasks. Researchers working on Content-Based Image Retrieval (CBIR) have traditionally tuned their algorithms to match filtered results with user search intent. However, we are now bombarded with composite images of unknown origin, authenticity, and even meaning. With such uncertainty, users may not have an initial idea of what the search query results should look like. For instance, hidden people, spliced objects, and subtly altered scenes can be difficult for a user to detect initially in a meme image, but may contribute significantly to its composition. It is pertinent to design systems that retrieve images with these nuanced relationships in addition to providing more traditional results, such as duplicates and near-duplicates — and to do so with enough efficiency at large scale. In this work, we propose a new approach for spatial verification that aims at modeling object-level regions using image keypoints retrieved from an image index, which is then used to accurately weight small contributing objects within the results, without the need for costly object detection steps. We call this method the Objects in Scene to Objects in Scene (OS2OS) score, and it is optimized for fast matrix operations, which can run quickly on either CPUs or GPUs. It performs comparably to state-of-the-art methods on classic CBIR problems (Oxford 5K, Paris 6K, and Google-Landmarks), and outperforms them in emerging retrieval tasks such as image composite matching in the NIST MFC2018 dataset and meme-style imagery from Reddit.

42 ENGINEERING↗