The opening session consisted of a series of presentations from DSCOVR mission leaders and representatives from GSFC and NASA Headquarters (HQ), who gave updates on the mission and the two Earth-viewing science instruments on board. Alexander Marshak [GSFC—DSCOVR Deputy Project Scientist] opened the meeting. He discussed the agenda for the meeting and mentioned that both Earth science instruments on DSCOVR are functioning normally – see Figure 1. At this time, more than 115 papers related to DSCOVR are listed on the EPIC website. Marshak emphasized the importance of making the Earth Science community more aware of the availability of the various EPIC and NISTAR science data products.
Adam Szabo [GSFC—DSCOVR Project Scientist] welcomed the STM participants and briefly reported that the spacecraft, located at “L1” – the first of five Lagrange points in the Sun-Earth system – was still in “good health.” The EPIC and NISTAR instruments on DSCOVR continue to return their full science observations. Szabo gave an update on the 2023 Earth Science Senior Review, which DSCOVR successfully passed with overall science scores of ‘Excellent/Very Good.’ The Senior Review Panel unanimously supported the continuation of DSCOVR for the 2024–2026 period.
Thomas Neumann [GSFC, Earth Sciences Division (ESD)—Deputy Director] welcomed meeting participants on behalf of the ESD. Neumann noted the impressive engineering that has led to 8.5 years of operations and counting. He also commended the team on the continued production of important science results from these instruments – with nearly 110 papers in the peer-reviewed literature.
Following Neumann’s remarks, Steve Platnick [GSFC, Earth Sciences Division—Deputy Director for Atmospheres] welcomed the members of the DSCOVR ST as well as users of EPIC and NISTAR observations. He thanked NASA HQ for its continued strong interest in the mission. Platnick also expressed his appreciation for the mission team members who have worked hard to maintain operation of the DSCOVR satellite and instruments during this challenging time.
Richard Eckman [NASA HQ, Earth Science Division—DSCOVR EPIC/NISTAR Program Scientist] noted that a new call for proposals will be in ROSES-2025 and looks forward to learning about recent accomplishments by ST members, which will be essential in assessing the mission’s performance.
Jack Kaye [NASA HQ, Earth Science Division—Associate Director for Research] discussed the NASA research program that studies the Earth, using satellites, aircraft, surface-based measurements, and computer models. The two Earth science instruments on DSCOVR (EPIC and NISTAR) play an important role in the program. He highlighted the uniqueness of the DSCOVR observations from the Sun–Earth “L1” point providing context for other missions and the ability to discern diurnal variations.
Updates on DSCOVR Operations
The DSCOVR mission components continue to function nominally, with progress on several fronts, including data acquisition, processing, archiving, and release of new versions of several data products. The number of people using the content continues to increase, with a new Science Outreach Team having been put in place to aid users in several aspects of data discovery, access, and user friendliness.
Hazem Mahmoud [NASA’s Langley Research Center (LaRC)] discussed the new tools in the Atmospheric Science Data Center (ASDC). He reported on DSCOVR metrics since 2015 and mentioned the significant increase in using ozone (O3) products. He also announced that ASDC is moving to the Amazon Web Services (AWS) cloud.
Karin Blank [GSFC] covered the EPIC geolocation algorithm, including the general algorithm framework. She highlighted additional problems that needed to be resolved and detailed the various stages to refine the algorithm, emphasizing the enhancements made to improve geolocation accuracy.
Marshall Sutton [GSFC] reported on the DSCOVR Science Operations Center (DSOC) and Level-2 (L2) processing. DSOC is operating nominally. EPIC L1A, L1B, and NISTAR data files are produced daily. EPIC L1 products are processed into L2 science products using the computing power of the NASA Center for Climate Simulations (NCCS). Products include daily data images, including a cloud fraction map, aerosol map, and the anticipated aerosol height image. In addition, Sutton reported that the DSCOVR spacecraft has enough fuel to remain in operation until 2033.
Alexander Cede [SciGlob] and Ragi Rajagopalan [LiftBlick OG] reported on the latest EPIC calibration version (V23) that includes the new flat field corrections based on the lunar observations from 2023 and an update to the dark count model. The EPIC instrument remains healthy and shows no change in parameters, e.g., read noise, enhanced or saturated pixels, or hot or warm pixels. The current operational dark count model still describes the dark count in a satisfactory way.
Liang-Kang Huang [Science Systems and Applications, Inc. (SSAI)] reported on EPIC’s July 2023 lunar measurements, which filled in the area near diagonal lines of the charged coupled device (CCD) not covered by 2021 and 2022 lunar data. With six short wavelength channels ranging from 317 to 551 nm, the two sets of lunar data are consistent with each other. For the macroscopic flat field corrections, he recommended the six fitted sensitivity change functions of radius and polar angle.
Igor Geogdzhaev [NASA’s Goddard Institute for Space Studies (GISS)/Columbia University] reported how continuous EPIC observations provide stable visible and near infrared (NIR) channels compared to the contemporaneous data from Visible Infrared Imaging Radiometer Suite (VIIRS) on NASA’s Suomi National Polar-orbiting Partnership (Suomi NPP) and the NASA–National Oceanic and Atmospheric Administration (NOAA) Joint Polar Satellite System (JPSS) missions. (To date, two JPSS missions have launched, JPSS-1, which is now known as NOAA-20, and JPSS-2, which is now known as NOAA-21.) Analysis of near simultaneous data from EPIC and from the Advanced Baseline Imager (ABI) on the Geostationary Operational Environmental Satellite–R (GOES R) platforms showed a high correlation coefficient, good agreement between dark and bright pixels, and small regression zero intercepts. EPIC moon views were used to derive oxygen (O2) channel reflectance by interpolation of the calibrated non-absorbing channels.
Conor Haney [LaRC] reported that the EPIC sensor was intercalibrated against measurements from the Moderate Resolution Imaging Spectroradiometer (MODIS) on NASA’s Terra and Aqua platforms as well as from VIIRS on Suomi NPP and NOAA-20, using ray-matched pair radiances, and was found to be radiometrically stable when tested against two invariant calibration targets: over deep convective clouds over the tropical Pacific (dark target) and over the Libya-4 site located in the Libyan desert in Africa (bright target). The ray-matched and Earth target EPIC gain trends were found to be consistent within 1.1%, and the EPIC sensor degradation was found to be less than 1% over the seven-year record. Preliminary results intercalibrating EPIC with the Advanced Himawari Imager (AHI) on the Japan Aerospace Exploration Agency’s (JAXA) “Himawari–8” Geostationary Meteorological Satellite were also promising when both subsatellite positions were close—i.e., during equinox.
NISTAR Status and Science with Its Observations
The NISTAR instrument remains fully functional and continues its uninterrupted data record. The presentations here include more details on specific topics related to NISTAR as well as on efforts to combine information from both EPIC and NISTAR.
Steven Lorentz [L-1 Standards and Technology, Inc.] reported that NISTAR has been measuring the irradiance from the Sun-lit Earth in three bands for more than eight years. The bands measure the outgoing reflected solar and total radiation from Earth at a limited range of solar angles. These measurements assist researchers in answering questions addressing Earth radiation imbalance and predicting future climate change. NISTAR continues to operate nominally, and the team is monitoring any in-orbit degradation. Lorentz explained the evolution of the NISTAR view angle over time. He also provided NISTAR shortwave (SW) and photodiode (PD) intercomparison. NISTAR has proven itself to be an extremely stable instrument – although measurements of the offsets have measurement errors. A relative comparison with the scaled-PD channel implies long-term agreement below a percent with a constant background.
Clark Weaver [University of Maryland, College Park (UMD)] discussed updates to a new reflected- SW energy estimate from EPIC. This new product uses generic Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) aircraft observations over homogeneous scenes to spectrally interpolate between the coarse EPIC channels. This approach assumes the spectra from an EPIC pixel is a weighted combination of a solid cloud scene and the underlying (cloud-free) surface. Weaver and his team used a vector discrete ordinate radiative transfer model with a full linearization facility, called VLIDORT, to account for the different viewing/illumination geometry of the sensors. Each pixel residual between EPIC observations at six different wavelengths (between 340 and 780 nm) and the composite high-resolution spectrum from AVIRIS has been reduced by about 50%, since the last report. While the total reflected energy for a single EPIC image can be about 15 W/m2 different than the NISTAR measurement, by 2017 the offset bias was, on average, about 1 W/m2.
Andrew Lacis [GISS] said that DSCOVR measurements of Earth’s reflected solar radiation from the “L1” position offer a unique perspective for the continuous monitoring of Earth’s sunlit hemisphere. Six years of EPIC data show the seasonal and diurnal variability of Earth’s planetary albedo – but with no discernible trend. Planetary scale variability, driven by changing patterns in cloud distribution, is seen to occur at all longitudes over a broad range of time scales. The planetary albedo variability is strongly correlated at neighboring longitudes but shows strongly anticorrelated behavior at diametrically distant longitudes.
Update on EPIC Products and Science Results
EPIC has a suite of data products available. The following subsections summarize content during the DSCOVR STM related to these products. They provide updates on several of the data products and on related algorithm improvements.
Natalya Kramarova [GSFC] reported on the status of the EPIC total O3 using the V3 algorithm. The absolute calibrations are updated every year using collocated observations from the Ozone Mapping and Profiling Suite (OMPS) on Suomi NPP. EPIC total O3 measurements are routinely compared with independent satellite and ground-based measurements. Retrieved EPIC O3 columns agree within ±5–7 Dobson Units (DU, or 1.5–2.5%) with independent observations, including those from satellites [e.g., Suomi NPP/OMPS, NASA’s Aura/Ozone Monitoring Instrument (OMI), European Union’s (EU) Copernicus Sentinel-5 Precursor/TROPOspheric Monitoring Instrument (TROPOMI)], sondes, and ground-based Brewer and Dobson spectrophotometers. The EPIC O3 record is stable and shows no substantial drifts with respect to OMPS. In the future, the EPIC O3 team plans to compare EPIC time resolved O3 measurements with observations from NASA’s Tropospheric Emissions Monitoring of Pollution (TEMPO) and the South Korean Geostationary Environment Monitoring Spectrometer (GEMS) – both in geostationary orbit. (Along with the EU’s Copernicus Sentinel-4 mission, expected to launch in 2024, these three missions form a global geostationary constellation for monitoring air quality on spatial and temporal scales that will help scientists better understand the causes, movement, and effects of air pollution across some of the world’s most populated areas.)
Jerrald Ziemke [Morgan State University] explained that tropospheric column O3 is measured over the disk of Earth every 1–2 hours. These measurements are derived by combining EPIC observations with Modern-Era Retrospective Analysis for Research and Applications (MERRA2) assimilated O3 and tropopause fields. These hourly maps are available to the public from the Langley ASDC and extend over eight years from June 2015 to present. The EPIC tropospheric O3 is now indicating post-COVID anomalous decreases of ~3 DU in the Northern Hemisphere for three consecutive years (2020–2022). Similar decreases are present in other satellite tropospheric O3 products as well as OMI tropospheric nitrogen dioxide (NO2), a tropospheric O3 precursor.
Algorithm Improvement for Ozone and Sulfur Dioxide Products
Kai Yang [UMD] presented the algorithm for retrieving tropospheric O3 from EPIC by estimating the stratosphere–troposphere separation of retrieved O3 profiles. This approach contrasts with the traditional residual method, which relies on the stratospheric O3 fields from independent sources. Validated against the near-coincident O3 sonde measurements, EPIC data biased low by a few DU (up to 5 DU), consistent with EPIC’s reduced sensitivity to O3 in the troposphere. Comparisons with seasonal means of TROPOMI tropospheric O3 show consistent spatial and temporal distributions, with lows and highs from atmospheric motion, pollution, lightning, and biomass burning. Yang also showed EPIC measurements of sulfur dioxide (SO2) from recent volcanic eruptions, including Mauna Loa and Kilauea (Hawaii, U.S., 2022–2023), Sheveluch (Kamchatka, Russia, 2023), Etna (Italy, 2023), Fuego (Guatemala, 2023), Popocatépetl (Mexico, 2023), and Pavlof and Shishaldin (Aleutian Islands, U.S., 2023). Yang reported the maximum SO2 mass loadings detected by EPIC are 430 kt from the 2022 Mauna Loa and Kilauea eruptions and 351 kt from the 2023 Sheveluch eruption.
Simon Carn [University of Michigan] showed EPIC observations of major volcanic eruptions in 2022–2023 using the EPIC L2 volcanic SO2 and UV Aerosol Index (UVAI) products to track SO2 and ash emissions. EPIC SO2 and UVAI measurements during the 2023 Sheveluch eruption show the coincident transport of volcanic SO2, ash, and Asian dust across the North Pacific. The high-cadence EPIC UVAI can be used to track the fallout of volcanic ash from eruption clouds, with implications for volcanic hazards. EPIC SO2 measurements during the November 2022 eruption of Mauna Loa volcano are being analyzed in collaboration with the U.S. Geological Survey, who monitored SO2 emissions using ground-based instruments during the eruption. Carn finished by mentioning that EPIC volcanic SO2 algorithm developments are underway including the simultaneous retrieval of volcanic SO2 and ash.
Myungje Choi [UMD, Baltimore County (UMBC)] presented an update on the EPIC V3 Multi-Angle Implementation of Atmospheric Correction (MAIAC) algorithm to optimize smoke aerosol models and the inversion process. The retrieved smoke/dust properties showed an improved agreement with long-term, ground-based Aerosol Robotic Network (AERONET) measurements of solar spectral absorption (SSA) and with aerosol layer height (ALH) measurements from the Cloud–Aerosol Lidar with Orthogonal Projection (CALIOP) on the Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) mission. (Update: As of the publication of this summary, both CALIPSO and CloudSat have ended operations.) Choi reported that between 60–90% of EPIC SSA retrievals are within ±0.03 of AERONET SSA measurements, and between 56–88% of EPIC ALH retrievals are within ±1km of CALIOP ALH retrievals. He explained that the improved algorithm effectively captures distinct smoke characteristics, e.g., the higher brown carbon (BrC) fraction from Canadian wildfires in 2023 and the higher black carbon (BC) fraction from agricultural fires over Mexico in June 2023.
Sujung Go [UMBC] presented a global climatology analysis of major absorbing aerosol species, represented by BC and BrC in biomass burning smoke as well as hematite and goethite in mineral dust. The analysis is based on the V3 MAIAC EPIC dataset. Observed regional differences in BC vs. BrC concentrations have strong associations with known distributions of fuels and types of biomass burning (e.g., forest wildfire vs. agricultural burning) and with ALH retrievals linking injection heights with fire radiative power. Regional distributions of the mineral dust components have strong seasonality and agree well with known dust properties from published ground soil samples.
Omar Torres [GSFC] reported on the upgrades of the EPIC near-UV aerosol (EPICAERUV) algorithm. The EPICAERUV algorithm’s diurnal cycle of aerosol optical depth compared to the time and space collocated AERONET observations at multiple sites around the world. The analysis shows remarkably close agreement between the two datasets. In addition, Torres presented the first results of an improved UV-VIS inversion algorithm that simultaneously retrieves aerosol layer height, optical depth, and single scattering albedo.
Hiren Jethva [Morgan State University] discussed the unique product of absorbing aerosols above clouds (AAC) retrieved from EPIC near-UV observations between 340 and 388 nm. The validation analysis of the retrieved aerosol optical depth over clouds against airborne direct measurements from the NASA ObseRvations of Aerosols above CLouds and their intEractionS (ORACLES) campaign revealed a robust agreement. EPIC’s unique capability of providing near-hourly observations offered an insight into the diurnal variations of regional cloud fraction and AAC over “hotspot” regions. A new and simple method of estimating direct radiative effects of absorbing aerosols above clouds provided a multiyear timeseries dataset, which is consistent with similar estimations from Aura–OMI.
Jun Wang [University of Iowa] reported on the development and status of V1 of the L2 EPIC aerosol optical centroid height (AOCH) product – which is now publicly available through ASDC – and on improvements to the AOCH algorithm – which focus on the treatment of surface reflectance and aerosols models. He presented applications of this data product for both climate studies of Sahara dust layer height and air quality studies of surface particulate matter with diameter of 2.5 µm or less (PM2.5). In addition, Wang showed the comparisons of EPIC AOCH data product with those retrieved from TROPOMI and GEMS and discussed ongoing progress to reduce the AOCH data uncertainty that is estimated to be 0.5 km (0.3 mi) over the ocean and 0.8 km (0.5 mi) over land.
Yaping Zhou [UMBC] showed how current EPIC O2 A-band and B-band use Moon calibrations due to lack of in-flight calibration and other comparable in-space instruments for absolute calibration. This approach is ineffective at detecting small changes in instrument response function (IRF). This study examined the O2 band’s calibration and stability using a unique South Pole location and Radiative Transfer Model (RTM) simulations with in situ soundings and surface spectral albedo and bidirectional reflectance distribution function (BRDF) measurements as input. The results indicate EPIC simulations are within 1% of observations for non-absorption bands, but large discrepancies exist for the O2 A-band (15.63%) and O2 B-band (5.76%). Sensitivity studies show the large discrepancies are unlikely caused by uncertainties in various input, but a small shift (-0.2–0.3 nm) of IRF could account for the model observation discrepancy. On the other hand, observed multiyear trends in O2 band ratios in the South Pole can be explained with orbital shift – which means the instrument is stable.
Alfonso Delgado Bonal [UMBC] used the EPIC L2 cloud data to characterize the diurnal cycles of cloud optical thickness. To fully exploit the uniqueness of DSCOVR data, all clouds were separated in three groups depending on their optical thickness: thin (0–3), medium (3–10), and thick (3–25). Bonal explained that there is a predictable pattern for different latitudinal zones that reaches a maximum around noon local time – see Figure 2. It was also shown that that the median is a better measure of central tendency when describing cloud optical thickness.
Elizabeth Berry [Atmospheric and Environmental Research (AER)] reported on how coincident observations from EPIC and the Cloud Profiling Radar (CPR) on CloudSat have been used to train a machine learning model to predict cloud vertical structure. A XGBoost decision tree model used input (e.g., EPIC L1B reflectance, L2 Cloud products, and background meteorology) to predict a binary cloud mask on 25 vertical levels. Berry discussed model performance, feature importance, and future improvements.
Robert Frouin [Scripps Institution of Oceanography, University of California] discussed ocean surface radiation products from EPIC data. He reported that surface radiation products were developed to address science questions pertaining to biogeochemical cycling of carbon, nutrients, and oxygen as well as mixed-layer dynamics and circulation. These products include daily averaged downward planar and scalar irradiance and average cosine for total light just below the surface in the EPIC spectral bands centered on 317.5, 325, 340, 388, 443, 551, and 680 nm and integrated values over the photosynthetically active radiation (PAR) and UV-A spectral ranges. The PAR-integrated quantities were evaluated against in situ data collected at sites in the North Atlantic Ocean and Mediterranean Sea. Frouin and his colleagues have also developed, tested, and evaluated an autonomous system for collecting and transmitting continuously spectral UV and visible downward fluxes.
Yuri Knyazikhin [Boston University] reported on the status of the Vegetation Earth System Data Record (VESDR) and discussed science with vegetation parameters. A new version of the VESDR software was delivered to NCCS and implemented for operational generation of the VESDR product. The new version passed tests of physics (e.g., various relationships between vegetation indices and vegetation parameters derived from the VESDR) and follow regularities reported in literature. Analysis of hotspot signatures derived from EPIC and from the Multiangle Imaging Spectroradiometer (MISR) on Terra over forests in southeastern Democratic Republic of the Congo reaffirms that long-term precipitation decline has had minimal impact on leaf area and leaf optical properties.
Jan Pisek [University of Tartu/Tartu Observatory, Estonia] reported on the verification of the previously modeled link between the directional area scattering factor (DASF) from the EPIC VESDR product and foliage clumping with empirical data. The results suggest that DASF can be accurately derived from satellite observations and provide new evidence that the photon recollision probability theory concepts can be successfully applied even at a fairly coarse spatial resolution.
Tamás Várnai [UMBC] discussed the EPIC Glint Product as well as impacts of sun glint off ice clouds on other EPIC data products – see Figure 3. The cloud glints come mostly from horizontally oriented ice crystals and have strong impact in EPIC cloud retrievals. Glints increase retrieved cloud fraction, the retrieved cloud optical depth, and cloud height. Várnai also reported that the EPIC glint product is now available at the ASDC. It is expected that glints yield additional new insights about the microphysical and radiative properties of ice clouds.
Alexander Kostinski [Michigan Technology University] reported on long-term changes and semi-permanent features, e.g., ocean glitter. They introduced pixel-pinned temporally and conditionally averaged reflectance images, uniquely suited to the EPIC observational circumstances. The preliminary resulting images (maps), averaged over months and conditioned on cover type (land, ocean, or clouds), show seasonal dependence at a glance (e.g., by an apparent extent of polar caps).
More EPIC Science Results
Guoyong Wen [Morgan State University] discussed spectral properties of the EPIC observations near backscattering, including four cases when the scattering angle reaches about 178° (only 2° from perfect backscattering). The enhancement addresses changes in scattering angle observed in 2020. (Scattering angle is a function of wavelength, because according to Mie scattering theory, the cloud scattering phase function in the glory region is wavelength dependent.) Radiative transfer calculations showed that the change in scattering angles has the largest impact on reflectance in the red and NIR channels at 680 nm and 780 nm and the smallest influence on reflectance in the UV channel at 388 nm – consistent with EPIC observations. The change of global average cloud amount also plays an important role in the reflectance enhancement.
Nick Gorkavyi [SSAI] talked about future plans to deploy a wide-angle camera and a multislit spectrometer on the Moon’s surface for whole-Earth observations to complement EPIC observations. Gorkavyi explained that the apparent vibrational movement of Earth in the Moon’s sky complicates observations of Earth. This causes the center of Earth to move in the Moon’s sky in a rectangle, measuring 13.4° × 15.8° with a period of 6 years.
Jay Herman [UMBC] reported on EPIC O3 and trends from combining Nimbus 7/Solar Backscatter Ultraviolet (SBUV), the SBUV-2 series, and OMPS–Nadir Mapper (NM) data. (OMPS is made up of three instruments: a Nadir Mapper (NM), Nadir Profiler, and Limb Profiler. OMPS NM is a total ozone sensor). Herman compared EPIC O3 data to OMPS NM data, which showed good agreement (especially summer values) for moderate solar zenith angle (SZA). Comparison with long-term O3 time series (1978–2021) revealed that there were trends and latitude dependent O3 turn-around dates (1994–1998). Herman emphasized that global O3 models do not show this effect but rather have only a single turn-around date around 2000.
Alexander Radkevich [LaRC] presented a poster that showed a comparative analysis of air quality monitoring by orbital and suborbital NASA missions using the DSCOVR EPIC O3 product as well as Pandora total O3 column retrievals. Comparison of the June 2023 total column O3 from EPIC data to the same periods in previous years revealed a significant – around 50 DU – increase of total O3 column in the areas impacted by the plume from 2023 Canadian wildfires.
At the end of the meeting Alexander Marshak, Jay Herman, and Adam Szabo discussed how to make the EPIC and NISTAR instruments more visible in the community. The EPIC website now allows visitors to observe daily fluctuations of aerosol index, cloud fraction, and the ocean surface – as observed from the “L1” point, nearly one million miles away from Earth! More daily products, (e.g., cloud and aerosol height, total leaf area index, and sunlit leaf area index) will be added soon.
The 2023 DSCOVR EPIC and NISTAR Science Team Meeting provided an opportunity to learn the status of DSCOVR’s Earth-observing instruments, EPIC and NISTAR, the status of recently released L2 data products, and the science results being achieved from the “L1” point. As more people use DSCOVR data worldwide, the ST hopes to hear from users and team members at its next meeting. The latest updates from the mission are found on the EPIC website. (UPDATE: The next DSCOVR EPIC and NISTAR STM will be held on October 16–18, 2024. Check the website for more details as the date approaches.)
Source : Nasa