Members Can Post Anonymously On This Site
Decisions from the Intermediate Ministerial Meeting 2021
-
Similar Topics
-
By NASA
4 Min Read NASA’s Hubble Finds Strong Evidence for Intermediate-Mass Black Hole in Omega Centauri
This NASA Hubble Space Telescope image features the globular star cluster, Omega Centauri. Credits:
ESA/Hubble, NASA, Maximilian Häberle (MPIA) Most known black holes are either extremely massive, like the supermassive black holes that lie at the cores of large galaxies, or relatively lightweight, with a mass of under 100 times that of the Sun. Intermediate-mass black holes (IMBHs) are scarce, however, and are considered rare “missing links” in black hole evolution.
Now, an international team of astronomers has used more than 500 images from NASA’s Hubble Space Telescope — spanning two decades of observations — to search for evidence of an intermediate-mass black hole by following the motion of seven fast-moving stars in the innermost region of the globular star cluster Omega Centauri.
Omega Centauri is about 10 times as massive as other big globular clusters – almost as massive as a small galaxy – and consists of roughly 10 million stars that are gravitationally bound. ESA/Hubble, NASA, Maximilian Häberle (MPIA)
Download this image
These stars provide new compelling evidence for the presence of the gravitational pull from an intermediate-mass black hole tugging on them. Only a few other IMBH candidates have been found to date.
Omega Centauri consists of roughly 10 million stars that are gravitationally bound. The cluster is about 10 times as massive as other big globular clusters — almost as massive as a small galaxy.
Among the many questions scientists want to answer: Are there any IMBHs, and if so, how common are they? Does a supermassive black hole grow from an IMBH? How do IMBHs themselves form? Are dense star clusters their favored home?
The astronomers have now created an enormous catalog for the motions of these stars, measuring the velocities for 1.4 million stars gleaned from the Hubble images of the cluster. Most of these observations were intended to calibrate Hubble’s instruments rather than for scientific use, but they turned out to be an ideal database for the team’s research efforts.
This image shows the central region of the Omega Centauri globular cluster, where NASA’s Hubble Space Telescope found strong evidence for an intermediate-mass black hole candidate. ESA/Hubble, NASA, Maximilian Häberle (MPIA)
Download this image
“We discovered seven stars that should not be there,” explained Maximilian Häberle of the Max Planck Institute for Astronomy in Germany, who led this investigation. “They are moving so fast that they would escape the cluster and never come back. The most likely explanation is that a very massive object is gravitationally pulling on these stars and keeping them close to the center. The only object that can be so massive is a black hole, with a mass at least 8,200 times that of our Sun.”
Several studies have suggested the presence of an IMBH in Omega Centauri. However, other studies proposed the mass could be contributed by a central cluster of stellar-mass black holes, and had suggested the lack of fast-moving stars above the necessary escape velocity made an IMBH less likely in comparison.
An international team of astronomers used more than 500 images from NASA’s Hubble Space Telescope – spanning two decades of observations – to detect seven fast-moving stars in the innermost region of Omega Centauri, the largest and brightest globular cluster in the sky. These stars provide compelling new evidence for the presence of an intermediate-mass black hole (IMBH) tugging on them. Only a few other IMBH candidates have been found to date. This image shows the location of the IMBH in Omega Centauri. If confirmed, at its distance of 17,700 light-years the candidate black hole resides closer to Earth than the 4.3-million-solar-mass black hole in the center of the Milky Way, which is 26,000 light-years away. Besides the Galactic center, it would also be the only known case of a number of stars closely bound to a massive black hole. This image includes three panels. The first image at left shows the globular cluster Omega Centauri, a collection of myriad stars colored red, white, and blue on the black background of space. The second image shows the details of the central region of this cluster, with a closer view of the individual stars. The third image shows the location of the IMBH candidate in the cluster. ESA/Hubble, NASA, Maximilian Häberle (MPIA)
Download this image
“This discovery is the most direct evidence so far of an IMBH in Omega Centauri,” added team lead Nadine Neumayer of the Max Planck Institute for Astronomy in Germany, who initiated the study, together with Anil Seth from the University of Utah, Salt Lake City. “This is exciting because there are only very few other black holes known with a similar mass. The black hole in Omega Centauri may be the best example of an IMBH in our cosmic neighborhood.”
If confirmed, at a distance of 17,700 light-years the candidate black hole resides closer to Earth than the 4.3-million-solar-mass black hole in the center of the Milky Way, located 26,000 light-years away.
Omega Centauri is visible from Earth with the naked eye and is one of the favorite celestial objects for stargazers living in the southern hemisphere. Located just above the plane of the Milky Way, the cluster appears almost as large as the full Moon when seen from a dark rural area. It was first listed in Ptolemy’s catalog nearly 2,000 years ago as a single star. Edmond Halley reported it as a nebula in 1677. In the 1830s the English astronomer John Herschel was the first to recognize it as a globular cluster.
The discovery paper led by Häberle et al. is published online today in the journal Nature.
Scientists think a massive object is gravitationally pulling on the stars within Omega Centauri, keeping them close to its center. Credit: NASA’s Goddard Space Flight Center, Lead Producer: Paul Morris
Download this video
The Hubble Space Telescope has been operating for over three decades and continues to make ground-breaking discoveries that shape our fundamental understanding of the universe. Hubble is a project of international cooperation between NASA and ESA (European Space Agency). NASA’s Goddard Space Flight Center in Greenbelt, Maryland, manages the telescope and mission operations. Lockheed Martin Space, based in Denver, Colorado, also supports mission operations at Goddard. The Space Telescope Science Institute in Baltimore, Maryland, which is operated by the Association of Universities for Research in Astronomy, conducts Hubble science operations for NASA.
Facebook logo @NASAHubble @NASAHubble Instagram logo @NASAHubble Media Contacts:
Claire Andreoli
NASA’s Goddard Space Flight Center, Greenbelt, MD
claire.andreoli@nasa.gov
Ray Villard
Space Telescope Science Institute, Baltimore, MD
Bethany Downer
ESA/Hubble.org
Science Contact:
Maximilian Häberle
Max Planck Institute for Astronomy, Heidelberg, Germany
Share
Details
Last Updated Jul 10, 2024 Editor Andrea Gianopoulos Location NASA Goddard Space Flight Center Related Terms
Astrophysics Astrophysics Division Black Holes Goddard Space Flight Center Hubble Space Telescope Missions Stars The Universe Keep Exploring Discover More Topics From Hubble
Hubble Space Telescope
Since its 1990 launch, the Hubble Space Telescope has changed our fundamental understanding of the universe.
Monster Black Holes Are Everywhere
Hubble’s Star Clusters
Time Travel: Observing Cosmic History
View the full article
-
By NASA
Earth Observer Earth Home Earth Observer Home Editor’s Corner Feature Articles Meeting Summaries News Science in the News Calendars In Memoriam More Archives 23 min read
Summary of the 2023 GEDI Science Team Meeting
Introduction
The 2023 Global Ecosystem Dynamics Investigation (GEDI) Science Team Meeting (STM) took place October 17–19, 2023, at the University of Maryland, College Park (UMD), in College Park, MD. Upwards of 80 people participated in the hybrid meeting (around 50 in-person and the rest virtually). Included among them were GEDI Science Team (ST) members, collaborators, and stakeholders – see Photo. The primary goals of the meeting included providing a status update on the GEDI instrument aboard the International Space Station (ISS), receiving final project updates from the inaugural cohort of the GEDI completed ST, and understanding the present status and future goals of data product development.
After a short mission status update, the remainder of this article will summarize the content of the STM. For those desiring more information on these topics, some of the full meeting presentations are posted online. Readers can also contact the GEDI ST with specific questions.
Photo. GEDI Science Team Meeting in-person and virtual attendees. Photo credit: Talia Schwelling Mission Status Update: GEDI Given New Lease on Life
A lot has changed since the publication of the last GEDI STM summary. (See Summary of the GEDI Science Team Meeting in the July–August 2022 issue of The Earth Observer [Volume 34, issue 4, pp. 20–24]). When the GEDI ST convened in November 2022, the fate of GEDI was hanging in the balance, with the plan being to release GEDI from the ISS at the end of its second extension period.
NASA saved the instrument, however, and a new plan went into effect: in order to extend its tenure on the ISS, the GEDI mission entered a temporary period of “hibernation” in March 2023 after nearly four years in orbit. This hibernation period and movement of the instrument from Exposed Facility Unit (EFU)-6 (operating location) to EFU-7 (storage location) made way for another mission – see Figure 1. (UPDATE: After being in storage for roughly 13 months, the GEDI instrument was returned to its original location on the Japanese Experiment Module–Exposed Facility (JEM–EF) on Earth Day this year, April 22, 2024, and is now once again back to normal science operations using its three lasers.)
Figure 1. NASA’s GEDI instrument was moved from EFU-6 to EFU-7 on the ISS on March 17, 2023, where it remained in hibernation for 13 months until its recent reinstallation to EFU-6 on April 22, 2024. The instrument is once again back to normal science operations using its three lasers. Figure credit: NASA As The Earth Observer reported in 2023, data from GEDI are being used for a wide range of applications, including biomass estimation, habitat characterization, and wildfire prediction (See page 4 of The Editor’s Corner in the March–April 2023 issue of The Earth Observer [Volume 35,Issue 2, pp. 1–4]. This section also reports on GEDI’s extension via out-of-cycle Senior Review in 2022). GEDI data is used to develop maps to quantify biomass that are unique in both their accuracy and their explicit characterization of uncertainty and are a key component in the estimation of aboveground carbon stocks, as absorbed carbon is used to drive plant growth and is stored as biomass – see Figure 2. These estimations help quantify the impacts of deforestation and subsequent regrowth on atmospheric carbon dioxide (CO2) concentration. NASA’s choice to extend the GEDI mission has significantly broadened the capacity to collect more of these important data.
Figure 2. Country-wide estimates of total aboveground biomass in petagrams (Pg) using GEDI Level-4B Version 2.1 dataset (GEDI_L4B_AGB). Figure credit: ORNL DAAC DAY ONE
GEDI Mission Operations, Instrument Status, and Data Level Updates
Ralph Dubayah [UMD—GEDI Principal Investigator (PI)] opened the meeting with a summary of the current status of the mission and GEDI data products. After reviewing the details of GEDI’s hibernation (described in the previous section) he went on to describe what GEDI has accomplished over the past 4.5 years of operations, noting that the instrument collected over 26 billion footprints over the land surface.
All the data collected by GEDI during its first epoch (i.e., before its hibernation) have been processed and released to the appropriate Distributed Active Archives Centers (DAACs) as Version 2 (V2) products. (To learn more about the DAACs and other aspects of Earth Science data collection and processing, see Earth Science Data Operations: Acquiring, Distributing, and Delivering NASA Data for the Benefit of Society, in the March–April 2017 issue of The Earth Observer, [Volume 29, Issue 2, pp. 4–18]. The DAACs – including URL links to each – are listed in a Table on page 7–8 of this issue). The two DAACs directly involved with GEDI data processing are the Land Processes DAAC (LP DAAC) and Oak Ridge National Laboratory (ORNL) DAAC. The LP DAAC houses GEDI Level-1 (L1) data, which consists of geolocated waveforms, and L2 data, which is broken down into L2A and L2B. L2A data includes ground elevation, canopy height, and relative height metrics. (Waveform measurements are described in detail in a sidebar on page 32 of the Summary of the Second GEDI Science Team Meeting in the November–December 2016 issue of The Earth Observer [Volume 28, Issue 6, pp. 31–36].) L2B data includes canopy cover fraction (CCF) and leaf area index (LAI). The ORNL DAAC houses GEDI L3 gridded land surface metrics data, L4A footprint level aboveground biomass density data, and L4B gridded aboveground biomass density data – e.g., see Figure 2.
Dubayah went on to explain that while GEDI hibernated, the mission team would work to enhance existing data products as well as produce new products. Version 3 (V3) datasets for all data products are expected to be released by the fall of 2024, and new data products are in development, including a waveform structural complexity index (WSCI) and a topography and canopy height product that blends data from GEDI and the Ice, Clouds, and land Elevation Satellite–2 (ICESat–2) mission. A new dataset, the GEDI L4C footprint level waveform structural complexity index (WSCI) product, was added to the ORNL DAAC catalogue in May 2024. To further improve data quality and coverage, the GEDI team is hoping to organize an airborne lidar field campaign to southeast Asia in the coming years. Dubayah concluded his updates by highlighting a set of six papers published in 2023 in Nature and Science family or partner journals that focused on the use of GEDI data. Visit our website for a comprehensive list of publications related to GEDI.
After receiving a general update from the mission PI, the next several presentations gave meeting participants a more in-depth look at GEDI science data planning and individual data products. Scott Luthcke [NASA’s Goddard Space Flight Center (GSFC)—GEDI Co-Investigator (Co-I)] presented status updates for the GEDI Science Operating Center (SOC), including the Science Planning System (SPS) and Science Data Processing System (SDPS) automation, development, and processing. In addition, he reported on the status of the L1 geolocated waveform data product and the L3 gridded land surface metrics product. At the time of this meeting, the SPS had completed operations through mission week 223 – almost 4.5 years of data – and was beginning to transition to improving processes on the back end while GEDI hibernates. The SDPS had completed processing and delivery of all V2 data products to the LP DAAC and ORNL DAAC.
Luthcke reported on GEDI’s current observed and estimated geolocation performance, including detailed summaries of component analysis and steps towards improving Precision Orbit Determination (POD), Precision Attitude Determination (PAD), Pointing Calibration, time-tag correction, and Oven Controlled Crystal Oscillator (OCXO) calibration. GEDI passes over Salar de Uyuni, the world’s largest salt flat located in Bolivia – see Figure 3, are being used to assess the PAD high-frequency and low-frequency errors. Estimated errors are shown to be consistent with observed geolocation errors. Finally, Luthcke gave a summary of completed L3 products and new wall-to-wall 1-km (0.62-mi) resolution and high-resolution products.
Figure 3. Salar de Uyuni, the world’s largest salt flat as seen from the International Space Station. Figure credit: Samantha Cristoforetti/ESA/NASA John Armston [UMD—GEDI Co-I] updated attendees on GEDI L2 products. L2A consists of elevation and height metrics, and L2B consists of canopy cover and vertical profile metrics. To assess GEDI ground and canopy top measurement accuracy and improve algorithm performance, the mission team is using data collected from NASA Land, Vegetation, and Ice Sensor (LVIS) campaigns from 2016 to present. Armston reported that L2B estimates of canopy and ground reflectance were completed for the first mission epoch (April 2019–March 2023) and the GEDI team continues to work on algorithm improvements for cover estimates in challenging conditions (e.g., steep slopes). Data users can expect improved waveform processing for ground elevation and canopy height, new reflectance estimation, and revised quality metrics and flags in the L2A and L2B not-yet-released V3 products.
Jim Kellner [Brown University—GEDI Co-I] shared the current status of and planned algorithm improvements to the L4A data product, or the footprint-level aboveground biomass density product. The algorithm theoretical basis document for L4A data products was published in November 2022; it describes how models were developed and the importance of quality filtering. L4A data product development continues in tandem with updates to L2A data and improvements to existing calibration and validation data and ingestion of new data.
Sean Healey [U.S. Forest Service—GEDI Co-I] reviewed coverage and uncertainties of the recently produced V2 L4B data products – see Figure 4. Ongoing GEDI-relevant research includes:
investigating a statistical method called bootstrapping, which may allow more complex types of models; conducting theoretical statistical studies aimed at decomposing mean square error for model-based methods; and developing ways to estimate biomass change over time – which will become more important as the extended mission potentially stretches to a decade. Figure 4. Gridded mean aboveground biomass density [top] and standard error of the mean [bottom] from Version 2.1 of the GEDI L4B Gridded Aboveground Biomass Density product, published on October 29, 2023. Figure credit: ORNL DAAC Competed Science Team Presentations—Session 1
This GEDI STM was the last convergence of the first iteration of the GEDI competed ST. Attendees received final in-person updates on the cohort’s projects and plans for future research. Over the course of the three-day meeting, there were several sections dedicated to Competed ST Presentations. For purposes of organization in this report, each section has been given a session number.
Taejin Park [NASA’s Ames Research Center (ARC) and Bay Area Environmental Research Institute (BAERI)] kicked off the ST presentations with an overview of his group’s progress in enhancing the predictions of forest height and aboveground biomass by incorporating GEDI L2, L3, and L4 data products into a process-based model, called Allometric Scaling Resource Limitation (ASRL), over the contiguous United States (CONUS). The ASRL model effectively captures large-scale, maximum tree size distribution and facilitates prognostic applications for predicting future aboveground biomass changes under various climate scenarios. Park also described collaborative research efforts with international partners to map changes in aboveground biomass in tropical and temperate forests using a carbon management systems (CMS).
Kerri Vierling [University of Idaho] shared the results from her team’s projects demonstrating the use of GEDI data fusion products to describe patterns of bird and mammal distributions in western U.S. forests. The focal species for these projects include a suite of vertebrate forest carnivores, prey, and ecosystem engineer species that modify their environments in ways that create habitat for other creatures, e.g., woodpeckers – see Figure 5. Many of these species are of interest for management by a variety of state and federal agencies. Vierling also discussed ongoing analyses identifying biodiversity hotspots and land ownership patterns.
Figure 5. A Female downy woodpecker creates a tree cavity that other organisms may use in the future for habitat. Woodpecker species are great examples of ecosystem engineers. Figure credit: Doug Swartz/Macaulay Library at the Cornell Lab or Ornithology (ML 58304661) Sean Healey presented on his competed ST research on Online Biomass Inference using Waveforms and iNventory (OBI-WAN), a Google Earth Engine application. This forest-carbon reporting tool harnesses GEDI waveforms, biomass models, and statistics to make estimates of mean biomass and biomass change for areas specified by online users. Healey explained the statistical methods applied to operate OBI-WAN and gave context for the use of sensor fusion to provide biomass change information that is critical for monitoring, reporting, and verification.
Keith Krause [Battelle] presented his work evaluating vertical structural similarity of LVIS classic and GEDI large-footprint waveforms. At the GEDI and LVIS footprint scale (20–23 m, or 65–75 ft, spot on the ground), lidar waveforms over forests represent canopies of leaves and branches from several trees. Krause presented results comparing waveforms against each other to show similarities in shape (i.e., if the trees in their footprints have a similar vertical structure). He also described how he used data clustering techniques to group similar waveforms into distinct structural classes. From there, he could map waveforms with similar vertical structure to better understand the spatial distribution of the structural groups.
Breakout Sessions—Session 1
GEDI STMs offer a rare opportunity for members of the competed and mission STs, a variety of stakeholders, and other individuals to convene and discuss ideas and goals for their own research and for the GEDI mission. Toward that end, breakout sessions were held on the first and second day of the meeting – referred to as Session 1 and Session 2 in this report. The individual breakout meetings used a hybrid format allowing in-person and online participants to join the discussion that was most relevant to their interests and expertise.
Chris Hakkenberg [Northern Arizona University (NAU)] led a breakout session on structural diversity, including the horizontal and vertical components. Different structural attributes, (e.g., stand structure, height, cover, and vegetation density) have different – but related – metrics and measurement approaches. Participants discussed biodiversity-structure relationships (BSRs), how to better characterize horizontal structural diversity, and how to define which metrics (i.e., scale, sampling unit, and spatial resolution) are most meaningful in different situations.
Jim Kellner led a session that focused on biomass calibration and validation and how to create the best data products given global environmental variation. Special cases – e.g., mangroves – pose challenges for calibration and validation because they don’t always have as much plot-level data as other environments. Participants discussed how to determine strata while considering climactic and environmental covariates as well as constraints of data availability and consistency.
Competed Science Team Presentations—Session 2
The FORest Carbon Estimation (FORCE) Project is exploring the use of GEDI-derived canopy structure metrics to map forest biomass in the U.S. and Canada. Daniel Hayes [University of Maine] presented comparisons of GEDI metrics and canopy height models derived from airborne lidar and photo point clouds over different forest types and disturbance history in managed forests of Maine. Co-PI Andy Finley [Michigan State University] presented new work that adjusts GEDI L4B biomass estimates to plot data over the continental U.S. from Forest Inventory and Analysis (FIA) program of the U.S. Department of Agriculture’s Forest Research and Development Branch. The project’s next steps are to fuse GEDI canopy structure metrics with other covariates in a spatial model to produce wall-to-wall estimates of biomass for boreal–temperate transition forests in northeast North America.
GEDI data is also being used to study tropical forests. Chris Doughty [NAU] described how he and his team analyzed GEDI L2A data across all tropical forests and found that tropical forest structure was less stratified and more exposed to sunlight than previously thought. Most tropical forests (80% of the Amazon and 70% of southeast Asia and the Congo Basin) have a peak in the number of leaves at 15 m (49 ft) instead of at the canopy top. Doughty and his team have found that deviation from more ideal conditions (i.e., lower fertility or higher temperatures) lead to shorter, less-stratified tropical forests with lower biomass.
Paul Moorcroft [Harvard University] reported on studies of current and future carbon dynamics across the Pacific Coast region based on forest structure and rates of carbon uptake. Moorcroft’s group examined how these ecosystems will behave in the future under different climate scenarios and have plans to conduct similar studies in other regions.
DAY TWO
Naikoa Aguilar-Amuchastegui [World Bank] kicked off day two with his perspective on the importance of streamlining the monitoring, reporting, and validation (MRV) process from scientific estimation to actual use of the data. Once scientific data is generated, end users are often faced with challenges related to transparency and understandability. Scientists can better communicate how to use their datasets properly, by familiarizing themselves with who wants to use their data, why they want to use it, and what their needs are. With this information in mind, data can be presented in more practical ways that allow for a variety of institutions with different standards and frameworks to integrate GEDI data more easily into their reporting. As the GEDI team continues to produce high-quality maps, efforts are underway to connect with end users and provide tutorials, workshops, and other resources.
GEDI Demonstrative Products
Demonstrative products show how GEDI data can be used in practice and in combination with other resources. Ecosystem modeling is one way that GEDI data are being used to address questions about aboveground carbon balance, future atmospheric CO2 concentrations, and habitat quality and biodiversity. George Hurtt [UMD—GEDI Co-I] shared his progress on integrating GEDI canopy height measurements with the Ecosystem Demography model to estimate current global forest carbon stocks and project future sequestration gaps under climate change – see Figure 6. Hurtt emphasized that this unprecedented volume of lidar data significantly enhances the ability of carbon models to capture spatial heterogeneity of forest carbon dynamics at 1 km (0.6 mi) scale, which is crucial for local policymaking regarding climate mitigation.
Figure 6. [Top] Average lidar canopy height at 0.01° resolution, computed by gridding both GEDI and ICESat-2 together, and carbon stocks [middle] and fluxes [bottom] from ED-Lidar (GEDI and ICESat-2 combined). The insets highlight fine-scale spatial distribution and coverage gaps at selected regions (1.5° × 1.5°). Note that the three maps show grid-cell averages aggregated from sub-grid scale heterogeneity for each variable. Figure credit: From a 2023 article in Global Change Biology. There is also great potential for the development and application of methods for mapping forest structure, carbon stocks, and their changes by fusing data from GEDI and the Deutsches Zentrum für Luft- und Raumfahrt’s (DLR) [German Space Operations Center] TerraSAR-X Add-oN for Digital Elevation Measurement (TanDEM-X) satellite mission, which uses synthetic aperture radar (SAR) to gather three-dimensional (3D) images of Earth’s surface. This fusion product is being spearheaded by Wenlu Qi [UMD], who presented on efforts to create maps of pantropical canopy height, biomass, forest structure, and biomass change using the fusion product as well as maps of forests in temperate U.S. and Hawaii.
Data from the GEDI mission are also being used to quantify the spatial and temporal distribution of habitat structure, which influences habitat quality and biodiversity. Scott Goetz [NAU—GEDI Deputy PI] presented on biodiversity-related activities, citing a 2023 paper in Nature that examined the effectiveness of protected areas (PAs) across southeast Asia using GEDI data to compare canopy structure within and outside of PAs – see Figure 7. He also presented an analysis of tree and plant diversity across U.S. National Ecological Observation Network (NEON) sites that showed similar capabilities of GEDI with airborne laser scanning (ALS) for tree diversity.
Figure 7. [Top] Protected Areas (PAs) such as national parks can reduce habitat loss and degradation (from logging) and extractive behaviors such as hunting (shown in red circle), but this figure shows there are a wide range of real-world outcomes based on management effectiveness. [Middle] PAs are aimed at safeguarding multiple facets of biodiversity, including species richness (SR), functional richness (FR) and phylogenetic diversity (PD). PAs often focus on vertebrate conservation, owing to their threat levels and value to humans – including for tourism. This study focused on wildlife in southeast Asia, with mammals shown here representing a variation of feeding guilds and sizes. The same approach is repeated for birds. [Bottom] Wildlife communities inside PAs and in the surrounding landscape may exhibit distinct levels and types of diversity. Figure credit: From a 2023 article in Nature. Competed Science Team Presentations—Session 3
One unique application of GEDI data is using lidar height to improve radiative transfer models for snow processes. Steven Hancock [University of Edinburgh, Scotland] reported on his group’s work studying snow, forest structure, and heterogeneity in forests, explaining that the majority of land surface models used for climate and weather forecasting use one-dimensional (1D) radiative transfer (RT) models driven by leaf area alone. Heterogeneous forests cast shadows and cause the surface albedo to depend upon sun angle and tree height for moderate leaf area indices (LAI), i.e., LAI values from 1-3 – which are common in snow-affected areas. This complexity cannot be represented in 1D models. An RT model can represent the effect of tree height and horizontal heterogeneity to simulate the observed change in albedo with height, which itself spatially varies.
In contrast to a snowy study area, Ovidiu Csillik [NASA/Jet Propulsion Laboratory] and his team are developing statistical models to link GEDI relative height metrics to tropical forest characteristics traceable to inventory measurements. This dataset of forest structure variables over the Amazon will be used to initialize a demographic ecosystem model to produce projections of future potential tropical forest carbon, as demonstrated by Amazon-wide simulations using initializations from airborne lidar sampling.
Wenge Ni-Meister [Hunter College of the City University of New York] is working on improving aboveground biomass estimates using GEDI waveform measurements. Ni-Meister and her team are testing models in both domestic and international tropical and temperate forests.
Breakout Sessions—Session 2
Two more breakout sessions occurred on day two:
Sean Healey led a discussion on modes of inference for GEDI data. Inference – formally derived uncertainty for area estimates of biomass, height, or other metrics – can take different forms, each of which includes specific assumptions. In this breakout session, participants considered the strengths and limitations of different inference types (e.g., intensity of computation or the ability to use different models).
Laura Duncanson [UMD—GEDI Co-I] led a discussion about facilitation of open science, in other words, how to make GEDI data more accessible and digestible for data users. While GEDI data area free and publicly available via the LP DAAC and ORNL DAAC, gaining access to said data can be intimidating. Sharing more about existing resources and creating new ones can help remove barriers. The LP DAAC and ORNL DAAC have excellent tutorials on GitHub (a cloud-based software development platform that is primarily Python-based), and Google Earth Engine applications are available for accessing and visualizing GEDI data. Future endeavors may include more webinars, R-based tutorials, workshops, and trainings on specific topics and ways to use GEDI data. More information is available via an online compilation of GEDI-related tutorials.
Perspective: A NUVIEW of Earth’s Land Surface
For the second perspective presentation of day two, meeting attendees heard from Clint Graumann, CEO and co-founder of NUVIEW, a company whose mission is to build a commercial satellite constellation of lidar-imaging satellites that will produce 3D maps of the Earth’s entire land surface. Graumann shared NUVIEW’s intent to produce land surface maps on an annual basis and provide a variety of products and services, including digital surface models (DSMs), digital terrain models (DTMs), and a point cloud generated by laser pulses.
Competed Science Team Presentations—Session 4
Laura Duncanson began the second round of science presentations with her group’s research on global forest carbon hotspots. She discussed her 2023 paper in Nature Communications on the effectiveness of global PAs for climate change mitigation – see Figure 8, which found that the creation of PAs led to more biomass – especially in the Amazon. Within GEDI-domain terrestrial PAs, total aboveground biomass (AGB) storage was found to be 125 Pg, which is around 26% of global estimated AGB. Without the existence of PAs, 19.7 Gt of the 125 Pg would have likely been lost.
Figure 8. PAs effectively preserve additional aboveground carbon (AGC) across continents and biomes, with forest biomes dominating the global signal, particularly in South America. The additional preserved AGC (Gt) in WWF biome classes (total Gt + /− SEM*area). World base map made with Natural Earth. The full set of analyzed GEDI data are represented in this figure (n = 412,100,767). Figure credit: From a 2023 article in Nature Communications. Another unique application of GEDI data has to do with water on the Earth’s surface. Kyungtae Lee [UMD], who works with Michelle Hofton [UMD—GEDI Co-I], reported that GEDI appears to capture the monthly annual cycle of lake elevation, showing good correlation with the ground-based observations. Lee explained that even though the GEDI lake elevation estimates show systematic biases relative to the local gauges, GEDI captures lake elevation dynamics well – especially the annual cycle variations. This work has the potential to expand knowledge of hydrological significance of lakes, particularly in data-limited areas of the world. Stephen Good [Oregon State University] presented a survey of his team’s recent work integrating observations from GEDI into hydrology and hydraulics studies of how vegetation can block and intercept moving water. The team found important nonlinear relationships between inferred canopy storage and canopy biomass and were able to estimate canopy water storage capacities and map these globally.
Finally, Patrick Burns [NAU], who works with Scott Goetz, presented results using GEDI canopy structure metrics in mammal species distribution models across southeast Asia (specifically focusing on Borneo and Sumatra). The team’s early results indicate that GEDI canopy structure metrics are important in many mammal distribution models and improve model performance for another smaller subset of species. In other words, when looking at predictors like mean annual precipitation or forest structure (forest structure being a metric that GEDI data provide), the GEDI-derived structure metrics are more intuitive and help us understand distributional changes and fine-scale habitat suitability. In a region like southeast Asia, for example, which has undergone high rates of deforestation in the recent decades, forest structure may be a more relevant predictor in a species distribution model (SDM) than other metrics like climate or vegetation composition. The team will continue to produce models for additional species and expand the extent of the analysis to include mainland Asia.
DAY THREE
Competed Science Team Presentations—Session 5
Day three began with the meeting’s last round of competed ST presentations. John Armston presented the progress of GEDI L2B Plant Area Volume Density (PAVD) product validation using a global Terrestrial Laser Scanning (TLS) database and fusion of the L2B product with Landsat time-series for quantifying change in canopy structure from the Australian wildfires of 2019–2020. Participants then heard from Jim Kellner on using machine-learning algorithms for L4A aboveground biomass density (AGBD). The performance of machine-learning algorithms on a testing data set was comparable to linear regressions used for the first releases of GEDI AGBD data products on average – although there were important geographical differences associated with machine learning. One application under investigation is using machine learning to identify new potential stratifications for GEDI footprint aboveground biomass density.
Lastly, Jingyu Dai [New Mexico State University (NMSU)], who works with Niall Hanan [NMSU], presented on her analysis of the global limits to tree height. Her study shows that hydraulic limitation is the most important constraint on maximum canopy height globally. This result is mediated by plant functional type. In addition, rougher terrain promotes forest height at sub-landscape scales by enriching local niche diversity and probability of larger trees.
Perspective from the Data Side
As described in the summary of Ralph Dubayah’s introductory remarks, the LP DAAC and ORNL DAAC play essential roles in the dissemination of GEDI data and the success of the GEDI program. Representatives from each of these DAACs addressed the ST to summarize recent GEDI-related activities.
Aaron Friesz [United States Geological Survey (USGS)] represented the LP DAAC and gave an update on the current archive size, distribution metrics, and outreach activities. He also discussed plans to support the growth and sustainability of the community through collaboration activities that will leverage the GitHub application; he described some of the resources that are available. Friesz then highlighted the USGS Eyes on Earth podcast and the Institute of Electrical and Electronics Engineers (IEEE) Geoscience and Remote Sensing Society (GRSS)’s Down to Earth podcast, which have featured Ralph Dubayah and Laura Duncanson, and shared plans to update the current GitHub tutorials and how-to guides in the Earthdata Cloud of GEDI V2 and V3.
Rupesh Shrestha [ORNL] represented the ORNL DAAC and shared the status of GEDI L3, L4A, and L4B datasets archived there. He gave an overview of data tools and services for the GEDI datasets, which can be found on the GEDI website and GitHub tutorials website. GEDI L3, L4A, and L4B are available on NASA’s Earthdata Cloud and various enterprise-level services, such as NASA’s WorldView, Harmony, and OpenDAP. GEDI data usage metrics, data tutorials and workshops, and outreach activities, as well as other published community and related datasets were also highlighted. GEDI L3, L4A, and L4B have been downloaded over four million times collectively.
Neha Hunka [UMD] gave the final presentation of the meeting on biomass harmonization activities. She reported that the GEDI estimates of aboveground biomass are capable of directly contributing to the United Nations Framework Convention on Climate Change Global Stocktake. Hunka and her colleagues’ research is aimed at bridging the science–policy gap to enable the use of space-based aboveground biomass estimates for policy reporting and impact – see Figure 9.
Figure 9. Forest biomass estimates in the format of Intergovernmental Panel on Climate Change (IPCC) Tier 1 values from NASA GEDI and ESA Climate Change Initiative (CCI) maps. Figure credit: Neha Hunka Conclusion
Overall, the 2023 GEDI STM showcased an exceptional array of scientific research that is highly relevant to addressing pressing global challenges and answering key questions about global forest structure, carbon balance, habitat quality, and biodiversity among other topics. As the GEDI instrument enters its second epoch, we are excited to welcome a new competed GEDI science team cohort and look forward to the release of V3 data products later this year.
Ralph Dubayah concluded the STM with a summary of hibernation period goals and a farewell to this iteration of the competed ST. He extended a heartfelt thank you and farewell to Hank Margolis [NASA Headquarters, emeritus] who has been the NASA Program Scientist for the GEDI mission since 2015. Thank you, Hank. We will miss you.
Talia Schwelling
University of Maryland, College Park
tschwell@umd.edu
View the full article
-
By NASA
Earth Observer Earth and Climate Earth Observer Home Editor’s Corner Feature Articles News Science in the News Calendars In Memoriam More Meeting Summaries Archives 22 min read
Summary of the Ninth DSCOVR EPIC and NISTAR Science Team Meeting
Introduction
The ninth Deep Space Climate Observatory (DSCOVR) Earth Polychromatic Camera (EPIC) and National Institute of Standards and Technology (NIST) Advanced Radiometer [NISTAR] Science Team Meeting (STM) was held virtually October 16–17, 2023. Over 35 scientists attended, most of whom were from NASA’s Goddard Space Flight Center (GSFC), with several participating from other NASA field centers, U.S. universities, and U.S. Department of Energy laboratories. One international participant joined the meeting from Estonia. A full overview of DSCOVR’s Earth-observing instruments was printed in a previous article in The Earth Observer and will not be repeated here. This article provides the highlights of the 2023 meeting. The meeting agenda and full presentations can be downloaded from GSFC’s Aura Validation Data Center.
Opening Presentations
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.
Figure 1. Sun-Earth-Vehicle (SEV) angle (red curve) and the distance between Earth and the DSCOVR satellite (blue curve) versus time starting from the DSCOVR launch on February 15, 2015 to April 1, 2024. These two measurements are used to track the location and orientation, respectively, of DSCOVR. The spacecraft changes its location by about 200,000 km (~124,274 mi) over about a 3-month period, and its SEV gets close to zero (which would correspond to perfect backscattering). The gap around the year 2020 was when DSCOVR was in Safe Mode for an extended period. Figure credit: Adam Szabo (Original figure by Alexander Marshak, with data provided by Joe Park/NOAA) 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.
EPIC Calibration
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.
Total Column Ozone
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.
Aerosols
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.
Clouds
Yuekui Yang [GSFC] explained the physical meaning of EPIC cloud effective pressure (CEP) in an “apples-to-apples” comparison with CEP measurements from the Global Ozone Monitoring Experiment 2 (GOME-2) on the European Operational Meteorology (MetOp) satellites. The results showed that the two products agreed well.
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.
Figure 2. Daytime variability of the median liquid cloud optical thickness over the ocean for different seasons of the year derived using EPIC L2 data. The various colored curves represent data collected in different seasons of the year. The black curve represents the annual average – which is most useful for calculations of cloud optical thickness. Figure credit: Alfonso Delgado Bonal 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.
Ocean
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.
Vegetation
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.
Sun Glint
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.
Figure 3. EPIC image taken over Mexico on July 4, 2018. The red, white and blue spot over central Mexico is the result of Sun glint reflecting off high clouds containing ice crystals. EPIC is particularly well suited for studies of ice clouds that cause Sun glint, because unlike most other instruments, it uses a filter wheel to take images at multiple wavelengths, which means the image for each wavelength is obtained at a slightly different time. For example, it takes four minutes to cycle from red to blue. During that time, Earth moves by ~100 km (~62 mi) meaning each image will capture a slightly different scene. Brightness contrasts between images can be used to identify glint signals. Image credit: Tamas Vanai 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.
Conclusion
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.)
Alexander Marshak
NASA’s Goddard Space Flight Center
alexander.marshak@nasa.gov
Adam Szabo
NASA’s Goddard Space Flight Center
adam.szabo@nasa.gov
View the full article
-
By NASA
Rahul Ramachandran (ST11) met with the World Food Program’s Head of Geospatial Support Unit. The focus was on his team’s work in Geospatial AI Foundation Models, specifically discussing the upcoming second version of the HLS Foundation Model. This new iteration promises an advanced architecture and extended training on global time sequences, offering unprecedented capabilities. The World Food Program’s Geospatial Support Unit expressed keen interest in leveraging this model to develop applications that could transform their operations. Ramachandran invited the World Food Program to join this open effort, highlighting the potential for these collaborations to revolutionize geospatial analytics and support global humanitarian efforts.
View the full article
-
By NASA
Earth ObserverEarth and Climate Earth Observer Home Editor’s Corner Feature Articles NewsIn Memoriams Science in the News MoreMeeting Summaries Archives 26 min read
Summary of the 2023 Precipitation Measurement Mission Science Team Meeting
Andrea Portier, NASA’s Goddard Space Flight Center/Science Systems and Applications, Inc., andrea.m.portier@nasa.gov
Introduction
The annual Precipitation Measurement Mission (PMM) Science Team Meeting (STM) took place September 18–22, 2023, in Minneapolis, MN. The PMM program supports scientific research and applications, algorithm development, and ground-based validation activities for the completed Tropical Rainfall Measuring Mission (TRMM) and current Global Precipitation Measurement (GPM) mission, including the GPM Core Observatory. Participants (including 137 in person and 22 virtual attendees) joined the meeting from a variety of affiliations including NASA, the Japan Aerospace Exploration Agency (JAXA), universities, and other partner agencies—see Photo.
The meeting included 46 plenary presentations spread across 7 thematically focused sessions and 77 poster presentations split between 2 sessions, with both oral and poster sessions covering mission and program status, partner reports, GPM algorithm development, and scientific results using GPM data.
The meeting also included a series of splinter sessions for precipitation working groups. The working groups included NASA–JAXA Joint Precipitation Science Team, the Committee on Earth Observation Satellites–Precipitation Virtual Constellation, GPM Mentorship Program, and topically focused groups on Applications, Hydrology, Land Surface, Latent Heating, Multisatellite, GPM Intersatellite Calibration (XCAL), Ground Validation (GV), Particle Size Distribution (PSD), and Oceanic Areas. These working groups were a combination of invitation-only, in-person, and hybrid meetings. Owing to the distributed nature of these meetings, summaries of their proceedings are not included in this article.
This article highlights current updates on the GPM mission and summarizes scientific results conveyed during the 2023 PMM STM. The meeting agenda and full presentations can be accessed through the 2023 PMM Science Team Meeting Files. Note that this is a password protected page; readers interested in accessing these files will need to reach out via the GPM Contact Form on the website to receive the access code.
Photo. Attendees of the 2023 PMM STM in front of the McNamara Alumni Center in Minneapolis, MN. Photo credit: Chris Kidd/GSFC and University of Maryland, College Park (UMD) Status Report and Updates on PMM: Perspectives from NASA and JAXA
The PMM missions are the fruit of long partnerships between NASA and JAXA. The PMM Science Team (ST) includes more than 20 international partners. The subsections that follow highlight the status of the PMM program and related activities that were conveyed by NASA and JAXA PMM Science Program Management Teams.
NASA
Will McCarty [NASA Headquarters (HQ)—GPM Program Scientist] presented the NASA HQ perspective regarding PMMs – present and future. He explained that current missions continue to drive the focus for precipitation science, and that future missions will continue to link the thermodynamic and dynamic factors of precipitation science by targeting additional temporal information. McCarty introduced several current and upcoming missions and programs, including satellite launches [e.g., NASA’s Time-Resolved Observations of Precipitation structure and storm Intensity with a Constellation of Smallsats (TROPICS), an Earth Venture Instrument (EVI), and the Investigation of Convective Updrafts (INCUS), an Earth Venture Mission], instruments [e.g., NASA’s Polarized Submillimeter Ice-cloud Imager (POLSIR), also an EVI, which will be deployed on two CubeSats], and field campaigns [e.g., NASA’s Investigation of Microphysics and Precipitation for Atlantic Coast-Threatening Snowstorms (IMPACTS) and Convective Processes Experiment Cabo Verde (CPEX-CV) experiments]. He then briefly discussed the second (2017) Earth Science Decadal Survey and provided an overview of the future Earth System Observatory (ESO), which will have interconnected core missions (e.g., the Atmosphere Observing System (AOS)). He also discussed the Planetary Boundary Layer (PBL), which the Decadal Survey classifies incubation targeted observable. McCarty concluded by noting that the future PMM ST call may be integrated by combining mission science from multiple satellites.
George Huffman [NASA’s Goddard Space Flight Center (GSFC)—GPM Project Scientist and PMM ST Lead] provided an update on the projected lifetime for GPM. Based on fuel usage alone, GPM should continue to December 2027. However, the amount of solar activity has an impact on that calculation. The Sun is expected to be quite active over the next few years as we approach the Solar Maximum for Solar Cycle 25—which could shorten GPM’s lifetime by as much as four years. He noted that a controlled reentry of the GPM Core spacecraft is planned—and enough fuel has to be kept in reserve to allow this to happen. Huffman discussed a recently developed plan for boosting the orbit of the GPM core satellite—for more details on the plan, see the subsection, “GPM Core Observatory Boost,” later in this article. He added that NASA and JAXA have both approved the plan and deemed its implementation critical for overlap with AOS for instrument intercomparison. The boosting is currently scheduled for November 7–9, 2023.(Update: Since the meeting in September, the GPM orbit boost was executed successfully on the scheduled dates.) The impact of the boosting on radiometer algorithms (e.g., for the GPM Microwave Imager (GMI)) is expected to be less than the impact on the radar algorithms (e.g., for the GPM Dual-Frequency Precipitation Radar, (DPR)). The potential impact on the combined algorithms (i.e., algorithms used to combine data from GMI and DPR) is still being assessed.
Huffman also discussed the status of the GPM data products. He reported that all GPM core data products are using Version 7 (V07). He mentioned that V07 of the Integrated Multi-Satellite Retrievals for GPM (IMERG) Final is out, but IMERG Early and Late data products are pending other actions in the NASA Precipitation Processing System (PPS). (IMERG has 3 classifications of data products: Early (latency of 4 hours), late (latency of 12–14 hours), and final (latency of 3 months).) He noted that the GPM orbit boost requires modifications to V07 core algorithms, and this accentuates the importance of a timely release of V08 algorithms (anticipated early 2026).
Erich Stocker [GSFC—GPM Deputy Project Scientist for Data and Precipitation Processing System Project Manager] discussed the status of GPM data products. He mentioned that radar/combined/IMERG products have transitioned from V06 to V07—but all radiometer products, Level-1 to Level-3, went from V05 to V07 to ensure the version is consistent on all of the products. Stocker continued that the GPM core satellite boost in November 2023 will lead to an outage of radar products for about five months for research and 2–3 months for near real-time (NRT) data products. NRT radiometer products will continue through the boost with only 2–3 days of outage while the satellite reaches its new altitude. He concluded that the initial NRT V07 IMERG processing and V07 retroprocessing of Early and Late IMERG products will start in January 2024.
David Wolff [NASA’s Wallops Flight Facility (WFF)—GPM Deputy Project Scientist for Ground Validation and Ground Validation System Manager] provided an overview of the GPM Ground Validation program and current activities. He stated that the ground validation (GV) program has state-of-the-art ground and remote sensing instruments to acquire precipitation and microphysics data to validate GPM retrievals. He described the ground validation site at NASA’s Wallops Flight Facility (WFF), which includes several radars, disdrometers (an instrument that measures drop-size distribution), and a Precipitation Imaging Processor (PIP) package. Wolff discussed the gauge-only systems, Platforms for In situ Estimation of Rainfall Systems (PIERS), activities for Increasing Participation of Minority Serving Institutions in Earth Science Division Surface-Based Measurement Networks, and pySIMBA – the GPM GV Support Software, an Open-Source Python Package to integrate and Analyze Precipitation Datasets that is available from GitHub. Wolff also provided a brief overview of the successful GPM GV Workshop that was held at Wallops Flight Facility on March 23–25, 2023. He continued by providing GPM Ground Validation Network (VN) updates and discussing VN captures of three-dimensional (3D) polarimetric information within DPR and GMI.
Wolff also noted that the GV program includes field campaigns (e.g., IMPACTS and Marquette, a five-year mini campaign conducted in collaboration with the National Oceanic and Atmospheric Administration’s (NOAA) National Weather Service (NWS)). He also discussed the new S-band radar network in Canada that offers access to high-quality radar data at relatively high latitudes over both land and sea. This data will be used as part of the VN for evaluation of GPM products. He concluded by discussing the Global Hydrometeorology Resource Center (GHRC) that archives past and current field campaign data and provides data quality control, metadata, campaign descriptions, and digital object identifier (DOI) assignments for each instrument/sensor.
Andrea Portier [GSFC—GPM Mission Applications Lead] and Dorian Janney [GSFC—GPM Outreach Coordinator] reflected on the 2022–2023 applications and outreach efforts and also discussed upcoming activities, including the – at the time of the meeting – upcoming tenth anniversary of the GPM Mission in February 2024. The applications team continues its focus on increasing awareness and use of GPM data and products across communities through user-engagement activities, including workshops (e.g., Applying Earth Observation Data for Research and Applications in Sustainable Development held at the 2022 Fall Meeting of the American Geophysical Union (AGU) in San Francisco, CA), trainings (e.g., 2023 GPM Mentorship Program), GPM application case studies, and GPM visualizations. A continuing and integral part of GPM outreach efforts is the numerous activities that reach hundreds of students and adults in a variety of formal and informal settings. This includes cooperative efforts with NASA’s Global Learning and Observations to Benefit the Environment (GLOBE) and hands-on activities at events (e.g., the Earth Day celebration at the Washington, DC’s Union Station). (To read more about the 2023 Earth Day celebration at Union Station, see A Pale Blue Dot in Washington: NASA’s Earth Day Celebration at Union Station, in the July–August 2023 issue of The Earth Observer [Volume 35, Issue 4, pp. 4–12].)
Many of these efforts will be highlighted and amplified during GPM’s tenth anniversary celebration. The GPM Applications and Outreach Team’s planning for the anniversary is underway. The intent is to highlight the vast capabilities of the GPM Mission and how GPM data can be used to address societal applications and improve the understanding of Earth’s water and energy cycles through a series of activities and resources starting in February 2024. These efforts include a reception at GSFC Visitor’s Center, a year-long monthly webinar series, feature articles, applications eBook, and a GPM video, among others. Details of these efforts will be posted through the GPM website.
JAXA
Takuji Kubota [JAXA—JAXA GPM Program Scientist] provided an update and a review of the PMM program status and mission objectives. He emphasized that this update included the perspectives of the Japanese PMM Science Program Management Team, including their roles in the development of DPR and its algorithms, GV, GPM data processing, and GPM data distribution systems. He also gave an update on current activities related to GPM data utilization and application across Japan and Asia. Kubota continued by describing the potential impacts on the DPR instrument because of the proposed orbit boost, noting that the instrument footprints and swath widths will increase proportionately with altitude change accompanied by a slight reduction in radar sensitivity. JAXA is preparing for these impacts with revised codes for L1 algorithms and planning for external calibrations before and after the orbit boost to examine calibrations of the DPR. Kubota also discussed the reprocessing of JAXA’s Global Satellite Mapping of Precipitation (GSMaP) data product (essentially the JAXA equivalent of IMERG) to enable a longer-term precipitation dataset, highlighting its completion in September 2023. GSMaP data is now available back to January 1998. Kubota discussed the future of Japanese precipitation measurements including: Earth Cloud, Aerosol and Radiation Explorer (EarthCARE), scheduled for launch in 2024; Global Observing SATellite for Greenhouse gases and Water cycle (OSAT-GW), planned for launch NET 2024; Advanced Microwave Scanning Radiometer (AMSR) series, which currently includes AMSR2 on the (GCOM-W) and will include AMSR3 on GOSAT-GW; and the previously discussed ESO AOS mission. He concluded with a discussion of JAXA’s plan for observing and celebrating GPM’s tenth anniversary.
Yukari Takayabu [University of Tokyo—JAXA GPM Project Scientist] highlighted results from recent science studies using DPR and GSMaP data products from the JAXA assembled GPM Program Science Team. She noted the use of DPR for extracting high-altitude precipitation information over Africa, capturing low-level precipitation statistics near the center of typhoons, narrowing the blind zone of the DPR to improve shallow precipitation detection in mountainous areas, validation studies of DPR, and retrieving frozen precipitation data using DPR. She concluded her presentation with highlights of GSMaP use for several applications, including the new GSMaP validation work in Japan to observe extreme rainfall, improvements to GSMaP through data-driven approaches, and data assimilation of GSMaP into the JAXA Realtime Weather Watch system.
Nobuhiro Takahashi [Nagoya University] presented an overview of significant updates to the DPM algorithm since the last PMM ST meeting, including changes in the latest V07 processing to accommodate the full-swath Ka-band operations – see Figure 1. He emphasized the impacts on the planning and development of V08 DPR algorithm with respect to the GPM orbit boost (described in George Huffman’s presentation). He noted that the major impacts to the performance of DPR include a degradation of measurement sensitivity and the “rain/no rain” classification. Takahashi concluded by saying that the release of V08 is expected in January 2026.
Figure 1. Evaluation of DPR product improvements from V06 to V07. Dual frequency product has smaller bias than KuPR product. The correlation coefficient improved from V06 to V07.Figure credit: Nobuhiro Takahashi/Nagoya University Kosuke Yamamoto [Earth Observation Research Center (EORC) and JAXA] summarized application activities initiated by the JAXA GPM Program Science Team. He discussed the use of GSMaP precipitation data to support and enhance several application areas, e.g., the operational use of GSMaP for flood and severe weather forecasting as well as the use of GSMaP in operational systems, including the JAXA Agro-meteorology Information Provision System (JASMIN), ASEAN Food Security Information System (AFSIS), and the Japanese’ Coast Guard’s Maritime Domain Awareness (MDA) initiative. Yamamoto also discussed the 2022 Japan–Australia–India–U.S. (QUAD) Joint Leaders’ Meeting Tackling Extreme Precipitation Events Workshop, an online event that took place March 1–3, 2023, and associated workshop reports focusing on the utilization of satellite observations across Pacific Islands.
GPM Algorithm Updates
Presenters during this session provided information and updates on various aspects of the five major algorithms of GPM. Full documentation and detailed updates for each algorithm are available at the Precipitation Data Directory.
Dual Frequency Radar Algorithm
The DPR algorithm team provided updates on DPR-related work, including the further refinement of the path-integrated attenuation (PIA) estimates used in the surface reference technique (SRT). They examined the effects of using the new AutoSnow algorithm – which uses satellite snowfall observations to create snowfall maps – on PIA estimations and changes in the surface type classification. Overall, the changes were small on the estimated precipitation profiles. Other algorithm refinements include the addition of a dry and wet snow category and wind speed. The team is currently examining how to recover Ka-band attenuation from the Ku-band. They stressed that results from this analysis are preliminary, and more work is needed to assess the utility of this technique. Finally, the team is discussing the implications of the GPM orbit boost on the DPR algorithm.
GPM Combined Radar–Radiometer Algorithm
The GPM Combined Radar–Radiometer Algorithm (CORRA) team discussed the changes and improvements to the CORRA V07 algorithm over the previous version. They highlighted the new AutoSnow algorithm and its impacts within CORRA V07. The team also examined the impact of the precipitation particle size distribution (PSD) initial assumptions on the estimation of snowfall as well as a machine-learning based initialization approach that improves the agreement between CORRA and NOAA’s Multi-Radar/Multi-Sensor System (MRMS) snow estimates. In addition, the team continues to examine a radiometer-only module to estimate light precipitation over oceans. This module will be included in the next version (V08) of CORRA. The team is also looking at the consequences of the GPM orbit boost.
Goddard Profiling Algorithm for GMI
The Goddard Profiling Algorithm (GPROF) team continues to work on well-known issues. The V07 update includes improvements in the a priori database to help constrain outputs from GPM constellation radiometers as well as inclusion of the radiometers on TROPICS and NASA’s Temporal Experiment for Storms and Tropical Systems–Demonstration (TEMPEST-D). The two new neural network-based implementations of GPROF in V08 are anticipated in roughly a year. The team reported that they have no issues with the GPM orbit boost.
Integrated Multi-Satellite Retrievals for GPM Algorithm
The IMERG algorithm team reported on V07, which includes a wide range of algorithm changes from V06. V07 includes retrospective reprocessing of the entire TRMM–GPM record and thus supersedes all previous versions. The team also reported that the algorithm changes improve the performance of IMERG estimates both in terms of its precipitation detection and systematic and random bias. The presenters noted improvements over frozen, orographic, and coastal surfaces. The team is now working on priority items that need completing in order to implement V08.
Convective–Stratiform Heating Algorithm
The GSFC Convective–Stratiform Heating (CSH) algorithm team provided an overview on latent heating (LH) retrievals. The presentation highlighted some of the details in updating to V07, including more accurate cloud-resolving model (CRM) simulations (using 3D domain rather than two-dimensional) and new detailed radiation retrievals. V07 is also “terrain aware,” meaning that the algorithm includes added details of radiative heating profiles and eddy transport terms. For V08, the CSH team plans to have a new 3D CRM database with a grid size of 250 m (820 ft) and look-up tables (LUTs) for non-surface raining columns for the tropical/summertime part of the algorithm as well as LUTs for terrain. These V08 improvements are still in development as of this meeting.
Science Results and Data Quality
A large component of the meeting was dedicated to presentations by NASA PMM-funded Principal Investigator (PI) teams on the science research and applications being achieved using PMM data. PI oral presentations were divided into four thematically focused topical sessions: Precipitation Microphysics, Snow and Hail, Storm Analysis, and Data Uncertainty. The subsections that follow highlight scientific results from each of these sessions. The reader is referred to the full reports online for more details.
Precipitation Microphysics
Presenters during this session described various techniques and new methodologies to study microphysical properties of precipitation including shape and size of precipitation particles (e.g., drop size distribution (DSD)), phase identification (e.g., liquid, solid, and mixed phase/melting), scattering properties, and precipitation rate, using both radar and radiometer observations. These property measurements play a pivotal role in improving precipitation retrieval algorithms, allowing scientists and decision makers to better understand and forecast storms.
One presenter in this session discussed new methods for classifying different types of precipitation (e.g., rain, graupel, hail, and dry and wet snow) using DPR precipitation retrievals. The new technique will be implemented into the V08 DPR algorithm. The discussion also covered a technique to establish relationships between GMI brightness temperature and hydrometeor type (e.g., rain, snow, graupel, and hail), leveraging the GPM validation network to construct LUTs of hydrometeor type likelihood – see Figure 2. Another presenter introduced a model to understand how DSD changes near the surface can be used to estimate rainfall rate. The last presenter in this session discussed the development of a precipitation scattering property database—which includes scattering characteristics of about 10,000 different types of ice particles. The database includes scattering cross sections calculated in thousands of orientations for each type of particle. This database is accessible to the public, which helps support the development of physically based scattering calculations and improvement of precipitation retrieval algorithms for both radar and radiometers.
Figure 2. A technique for retrieving hydrometeor information from GMI brightness temperature. In these RGB plots, snow and rain are combined into one category (green), while the individual probabilities are retained in the lookup tables.Figure credit: Dan Cecil/NASA’s Marshall Space Flight Center (MSFC) Snow and Hail
In this session, speakers discussed a broad move toward satellite retrievals for frozen hydrometeors, not just to identify bulk effects (e.g. snow or hail accumulation at the surface), but also to gather information on physical properties of frozen hydrometeors (e.g., where hailstones reside within clouds or what shapes snowflakes take). Understanding frozen hydrometeor properties can significantly improve precipitation and latent heat estimates that are essential for numerical weather forecasting and climate model development.
One speaker applied a method that used DPR and GMI observations to estimate frozen precipitation particle properties for an Olympic Mountain Experiment (OLYMPEX) field campaign case. The results he showed indicated a significant difference in the shapes of snowflakes between land and sea. Another speaker detailed the use of a simple machine learning framework trained on measurements of the use of snowfall and cloud type observations from the CloudSat Cloud Profiling Radar (CPR) to infer surface snowfall from GMI microwave measurements. Other presenters conveyed the results of a study examining different potential indicators of hail within the GPM database. These hail indicators were mapped, and the mean vertical profiles of radar reflectivity and storm structure were contrasted. The final pair of presentations focused on detecting hail in South America and Africa. In South America, hail-producing storms were shown to be strongly linked to local topography – in contrast to hotspots of hail in the U.S. Meanwhile, in Africa, new algorithms for identifying hail in GPM data suggest hail should be common – but this outcome is at odds with ground truth observations. This test case is being used to develop new methods for retrieving hail that include analyzing horizontal profile information within the data.
Storm Analysis
Presenters in this session discussed a variety of applications and assessments of PMM products for analyzing a variety of storms, particularly their cloud, precipitation, and kinematic structures and their structural evolution. The first speaker compared precipitation events simulated in IMERG to the same event with rain gauge observations. They found that while IMERG missed many winter precipitation events in mountainous regions –which rain gauges typically can measure – IMERG also captured summer virga events – which rain gauges typically miss. Another presenter compared IMERG to river catchment and integrated watershed observations and found that IMERG overestimated small precipitation events but underestimated large events. The next presenter showed a comparison IMERG simulations to the multi-instrument MRMS dataset during the lifecycle of precipitation events. The results shown suggest that IMERG errors in precipitation intensity could be improved by inputting other variables (e.g., ice water path or vertical velocity) into the precipitation retrievals. The discussions during this session also covered other plans to use PMM products to study convection in atmospheric river events, in combination with a modeling analysis using different convection schemes. The final pair of presenters spoke about understanding convective-scale drivers of the Inter Tropical Convergence Zone ascent and widening the use of a simple prognostic model that will use PMM data for filling terms in the model. One model weakness is the decay term for the convection cloud shield, which, if determined, could reduce error in climate models, particularly with radiative processes. The final speaker used TRMM Visible and Infrared Scanner (VIRS) data to develop and test a method for identifying and classifying cloud areas (i.e., core, midrange extent, and outer bound split window testing) and determine their relationships to other environmental variables, such as sea surface temperatures and column water vapor.
Data Uncertainty
Presenters during this session discussed new methodologies to address data uncertainties and bias in precipitation retrievals to improve precipitation estimates for science and applications research. Two of the presenters delved into the details of how the GPROF algorithm has inherent precipitation biases due to different hydrometeor characteristics captured by GMI passive microwave brightness temperature – which may be related to thermodynamic environments. Another PI presented updates for improving uncertainty estimates to enhance hydrological prediction. Specifically, he discussed multiscale precipitation uncertainties in precipitation products, including a new product that combines the Space-Time Rainfall Error and Autocorrelation Model (STREAM) with single-orbit rainfall estimates from the combined GPM data product, called STREAM-Sat. He explained how the uncertainties in these products can influence hydrologic prediction. The session concluded with a discussion of machine learning methods to estimate the probability distribution of uncertainties in passive microwave precipitation retrievals at different temporal and spatial scales.
Discussion of Future Missions, Observations, and Activities Relevant to GPM
This session featured presentations on several other existing and upcoming missions in various stages of development, as well presentations covering the future of precipitation instruments and observations, each with applications relevant to GPM. Each presentation included information on plans to advance and support precipitation science in the near term and the coming decade, as described below.
TROPICS
The TROPICS Pathfinder CubeSat mission provides microwave observations of tropical cyclones with less than a 60-minute revisit time to capture better storm dynamics and improve forecasting. The Pathfinder has demonstrated all mission elements and provided new tropical cyclone imagery (12,000+ orbits and counting). The Cal/Val team hopes to release the data to the public in Fall 2023. (UPDATE: Provisional TROPICS data was released in January 2024.) The TROPICS pathfinder satellite showed that the compact TROPICS design performs comparably to the state-of-the-art sounders. Lessons learned will help the TROPICS Team as they work to improve efforts and operate the TROPICS constellation, which now holds a total of five satellites.
AOS
As discussed in Will McCarty’s remarks, AOS is a key component of the Earth System Observatory that was recommended in the 2017 Decadal Survey. The mission will deliver transformative observations fundamental to understanding coupled aerosol– and cloud–precipitation processes that profoundly impact weather, climate, and air quality. Two AOS projects are in the mission concept and technology development phase (Phase-A): AOS-Storm (to launch late 2020s), with a Ku Doppler radar, microwave radiometers, and backscatter lidar in a 55° inclined orbit; and AOS-Sky (to launch early 2030s) with cloud-profiling Doppler radar, backscatter lidar, microwave radiometer, polarimeter, far infrared (IR) radiometer, and aerosol and moisture limb sounders in polar orbit. (This paragraph reflects what was discussed during the meeting, however, AOS is undergoing changes that will be reflected on the website at a later date.)
GPM Microwave Radiometer Constellation in the Next Decade
The future passive microwave radiometer constellation looks robust, with multiple sensors to be launched in the next decade. Small/CubeSat constellations are becoming a reality, and a plan to incorporate them quickly into the overall precipitation constellation is needed. A point of emphasis was that a sensor in an inclined orbit is a necessity when it comes to providing a reference measurement to support this effort – see Figure 3.
Figure 3. Evaluation of passive microwave (PMW) frequencies and coverage to assess data gaps and needs for the future of precipitation constellation.Figure credit: Rachael Kroodsma/GSFC JAXA Precipitation Measuring Mission (JAXA PMM) Radar
Plans call for JAXA’s next generation of precipitation radar to be deployed as part of the agency’s future Precipitating Measuring Mission (PMM – yes, the same acronym as the Precipitation Measurement Mission). Objectives for this next-generation precipitation radar include Doppler observations, higher sensitivity measurements, and scanning capability. JAXA has collaborated with a Japanese science team and user community to explore the feasibility of a next-generation, dual-frequency precipitation radar. The discussion focused on the importance of measuring convection through Doppler velocities from spaceborne radar. The EarthCARE mission will feature the first Cloud Profiling Radar (CPR) with Doppler capability in space. JAXA has participated in NASA’s AOS Pre-Phase A activities. The synergy between the GPM DPR and PMM/KuDPR is expected to contribute to the construction of a longer-term precipitation dataset by providing overlapping observations.
Update on Cloud Services at NASA GES DISC
NASA’s Goddard Earth Sciences Data and Information Services Center (GES DISC), one of two data archive centers for GPM, is moving its data archive to the cloud – with all GES DISC data and services remaining free to all users. This will offer quick access to and subsetting capability for a large volume of data through multiple data access methods (e.g., Amazon Simple Storage Service) and cloud services. Multidisciplinary NASA data will be in one place – the Earthdata Cloud – and available for online analysis and in the cloud environment. Expanded services (e.g., access to the Common Metadata Repository–SpatioTemporal Asset Catalog (CMR-STAC), Harmony – a collective Earth Observing System Data and Information System (EOSDIS) effort to make data access more consistent and easier across all DAACs and Zarr – a data format designed to store compressed multidimensional arrays and thus well suited to cloud computing) are expected to be implemented in the near future. With the migration of GES DISC data to the cloud, some services may look different with details on the exact changes to services coming soon.
GPM Core Observatory Boost
As George Huffman discussed in his presentation, based on forecasted solar activity, the GPM Core Observatory could run out of fuel as early as October 2025 if the current orbit altitude is maintained. To prolong its operations, NASA and JAXA have decided to boost the GPM Core Observatory orbit by ~35 km (~22 mi), which places GPM at an altitude of ~435 km (~270 mi)) – placing it above the International Space Station orbital altitude. The post-boost operations of the satellite are expected to continue through the early 2030s. The boost is expected to last only 2–4 days and occur in the time window between November 2023 and March 2024 (likely November 7–9, 2023, as stated above), the boost will permanently change the sensors’ Field of Views (FOVs) and likely cause a gap of several months in DPR product delivery.
Precipitation in 2040
Sarah Ringerud [GSFC] and George Huffman led this plenary discussion that explored two questions: What comes next? and What does the cutting edge of precipitation science look like 20 years from now? CubeSats, reduced volume of low-frequency-channel observations, shorter sensor lifetimes, increased sampling, and calibration challenges are recognized as inevitable. Exciting new developments are seen in the opportunity for data fusion and interdisciplinary work. Interagency and private sector collaborations are foreseen as critical points for maintaining optimal monitoring of Earth precipitation.
Conclusion
The 2023 PMM STM brought together scientists from around the world to engage on a range of topics that advance the understanding of precipitation science, algorithms, and contributions to applications. The STM highlighted updates and activities enabled by the PMM scientific community. The closing session provided an opportunity for quick updates from precipitation working group members, who held splinter sessions. These updates were followed by an open discussion and review of PMM action items led by George Huffman. He reminded PMM STM participants of several important and noteworthy items, including updates on the orbit boost and subsequent algorithm adjustments, which will be available on the GPM website and be at the forefront for the project for the next six months; V08 of GPM data products are anticipated by early 2026; the budget reduction for the project – but not for current ROSES projects – will impact activities, including next year’s PMM STM; and the next NASA ROSES call might have a different package of opportunities, not strictly focused on PMM/GPM. He concluded by encouraging the PMM ST to share highlights and publications with the GPM Science Program Management Team as well as to continue to initiate collaborations with other colleagues to keep pushing the boundaries of science and outreach.
The next PMM STM will likely be held in September 2024. Details will be posted on the GPM website once they become available.
Acknowledgements The author would like to recognize the following individuals, all of whom made contributions to this article: Ali Behrangi [University of Arizona], Anthony Didlake [Penn State University], Gerry Heymsfield [GSFC], George Huffman [GSFC], Matthew Igel [University of California Davis], Toshio Iguchi [Osaka University], Dorian Janney [GSFC/ADNET Systems], Chuntao Liu [Texas A&M Corpus Christi], Veljko Petkovic [UMD], Courtney Schumacher [Texas A&M Corpus Christi], and Joe Turk [NASA/Jet Propulsion Laboratory].
View the full article
-
-
Check out these Videos
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.