SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Live footage shows the Launch Control Center (LCC) communicating with the STS-35 Space Shuttle. Scenes include various playback launch views of STS-35. Also shown are panoramic views of the Shuttle on the launch pad, main engine start, ignition, liftoff and booster separation and various Long Range Tracker views.
Footage shows several views of the launching of the Mars Observer Titan III satellite. Scenes include different angle and camera views of the launch. Also shown are panoramic views of the launch vehicle on the launch pad.
Live footage of the STS-41 deployment of the European Space Agency's Ulysses probe is presented. Richard N. Richards, Commander, Robert D. Cabana, Pilot, Mission Specialists William M. Shepherd, Bruce E. Melnick and Thomas D. Akers are shown aboard the Space Shuttle.
The landing of the Discovery Orbiter at Kennedy Space Center (KSC) is shown from many different vantage points, including footage of the landing taken with infrared cameras.
The most significant highlights of flight day 4 of STS-111 are the lifting of the Leonardo Multi Purpose Logistics Module (MPLM) out of Endeavour's payload bay, and its temporary attachment to the Unity module of the International Space Station (ISS). The lifting and attachment are shown. Also shown is the testing of cameras on the CANADARM2 robotic arm on the ISS, an ISS extravehicular activity (EVA) suit, and an interview with the outgoing Expedition 4 crew of the ISS. Earth views include the ISS and Earth limb, and cloud formations.
The most important highlight of flight day 10 of STS-111 was the undocking of the Leonardo Multi Purpose Logistical Module (MPLM) from the Unity module of the International Space Station (ISS). The undocking is shown, as is Leonardo resting in Endeavour's payload bay. The video also includes an interview, and footage of the West Africa and the Atlantic Ocean, Gibraltar and the Mediterranean Sea, Italy, Mexico and the Pacific Ocean, Hispanola, an unknown land mass, and unknown mountains.
The STS-113 eleventh flight day begins with Flight Engineer John Herrington and ISS Science Officer Peggy Whitson answering questions from the media. Earth Views from the Space Shuttle Endeavor payload bay, as the Space Shuttle Endeavor Expedition Five Crew travel over the Baja Peninsula are also presented.
The activities of the STS-113 (Jim Wetherbee, Commander; Paul Lockhart, Pilot; Michael Lopez-Alegria, John Herrington, Mission Specialists) crew during flight day 10 are shown. Also seen are the outgoing Expedition 5 (Valeri Korzun, Commander; Peggy Whitsun, ISS Science Officer/Flight Engineer; Sergei Treschev, Flight Engineer) and incoming Expedition 6 (Kenneth Bowersox, Commander; Donald Pettit, Nikolai Budarin, Flight Engineers) crews of the International Space Station (ISS). A panoramic view of the undocking of the Space Shuttle Endeavor from the International Space Station is also presented.
This presentation summarizes the evolution of chemical modeling in the GMAO, showing how the parallel tracks for mission support, using forecasts over several days, and chemistry-climate coupling, using multi-decadal simulations, are converging in the future. The presentation emphasizes the new GEOS Composition Forecasting system (GEOS CF) and its potential role in future reanalyses of NASA's EOS Aura data and the testing needed to use this in chemistry-climate simulations.
Explore the source record for details and available documents.
Abstract not provided.
Explore the source record for details and available documents.
Immediately following the breakup of the 2005 Arctic spring stratospheric vortex, a tropical air mass, characterized by low potential vorticity (PV) and high nitrous oxide (N2O), was advected poleward and became trapped in the easterly summer polar vortex. This feature, known as a "Frozen-In Anticyclone (FrIAC)", was observed in Earth Observing System (EOS) Aura Microwave Limb Sounder (MLS) data to span the potential temperature range from approximately 580 to 1100 K (approximately 25 to 40 km altitude) and to persist from late March to late August 2005. This study compares MLS N2O observations with simulations from the Global Modeling Initiative (GMI) chemistry and transport model, the GEOS-5/MERRA Replay model, and the VanLeer Icosahedral Triangular Advection isentropic transport model to elucidate the processes involved in the lifecycle of the FrIAC which is here divided into three distinct phases. During the "spin-up phase" (March to early April), strong poleward flow resulted in a tight isolated anticyclonic vortex at approximately 70-90 deg N, marked with elevated N2O. GMI, Replay, and VITA all reliably simulted the spin-up of the FrIAC, although the GMI and Replay peak N2O values were too low. The FrIAC became trapped in the developing summer easterly flow and circulated around the polar region during the "anticyclonic phase" (early April to the end of May). During this phase, the FrIAC crossed directly over the pole between the 7th and 14th of April. The VITA and Replay simulations transported the N2O anomaly intact during this crossing, in agreement with MLS, but unrealistic dispersion of the anomaly occurred in the GMI simulation due to excessive numerical mixing of the polar cap. The vortex associated with the FrIAC was apparently resistant to the weak vertical hear during the anticyclonic phase, and it thereby protected the embedded N20 anomaly from stretching. The vortex decayed in late May due to diabatic processes, leaving the N2O anomaly exposed to horizontal and vertical wind shears during the "shearing phase" (June to August). The observed lifetime of the FrIAC during this phase is consistent with time-scales calculated from the ambient horizontal and vertical wind shear. Replay maintained the horizontal structure of the N2O anomaly similar to NILS well into August. The VITA simulation also captured the horizontal structure of the FrIAC during this phase, but VITA eventually developed fine-scale N2O structure not observed in MLS data.
Model simulations of ozone (O3) driven by meteorological reanalyses are useful for filling observational gaps and interpreting observed O3 variability and trends. However, the transport circulation of reanalysis products is impacted by changes to the observing system (the data assimilated into the reanalyses). We examine the impacts of these changes on simulated O3 from two models, Global Modeling Initiative (GMI) Chemistry Transport Model (GMI CTM) and Modern‐Era Retrospective Analysis for Research Applications version 2 (MERRA‐2) GMI Replay (M2 GMI Replay) simulation, using observations from global ozonesondes (>50,000 profiles) and satellites from 1980 to 2016. Both models are constrained by meteorology from the NASA MERRA‐2 reanalysis, and both use versions of NASA's GMI chemical mechanism. We focus on an observing system change affecting simulated O3 after 1998, associated with the assimilation of temperature and humidity data from new microwave profiling satellites. A large post‐1998 O3 increase, mainly confined to 15–20‐km altitude, of ~10 Dobson units (DU) in midlatitudes occurs in the GMI CTM, worsening the bias compared to observations. In contrast, an increase in M2 GMI Replay simulation O3 of ~10 DU is observed only near −60° latitude, reducing the bias compared to observations. The GMI CTM O3 high biases display a Quasi‐Biennial Oscillation (QBO)‐like periodicity that result from excessive transport from the tropical stratosphere to the midlatitude lower stratosphere during the QBO westerly phase. We quantify O3 discontinuities caused by MERRA‐2 observing system changes and demonstrate how the MERRA‐2 Global Modeling Initiative Replay simulation dampens the effects of these changes and QBO‐driven artifacts on simulated lower stratospheric and total O3. We caution against using simulations driven by a reanalysis to derive multidecadal O3 trends, especially prior to 1998.
In modern Machine Learning, model training is an iterative, experimental process that can consume enormous computation resources and developer time. To aid in that process, experienced model developers log and visualize program variables during training runs. Exhaustive logging of all variables is infeasible, so developers are left to choose between slowing down training via extensive conservative logging, or letting training run fast via minimalist optimistic logging that may omit key information. As a compromise, optimistic logging can be accompanied by program checkpoints; this allows developers to add log statements post-hoc, and "replay" desired log statements from checkpoint---a process we refer to as hindsight logging. Unfortunately, hindsight logging raises tricky problems in data management and software engineering. Done poorly, hindsight logging can waste resources and generate technical debt embodied in multiple variants of training code. In this paper, we present methodologies for efficient and effective logging practices for model training, with a focus on techniques for hindsight logging. Our goal is for experienced model developers to learn and adopt these practices. To make this easier, we provide an open-source suite of tools for Fast Low-Overhead Recovery (flor) that embodies our design across three tasks: (i) efficient background logging in Python, (ii) adaptive periodic checkpointing, and (iii) an instrumentation library that codifies hindsight logging for efficient and automatic record-replay of model-training. Model developers can use each flor tool separately as they see fit, or they can use flor in hands-free mode, entrusting it to instrument their code end-to-end for efficient record-replay. Our solutions leverage techniques from physiological transaction logs and recovery in database systems. Evaluations on modern ML benchmarks demonstrate that flor can produce fast checkpointing with small user-specifiable overheads (e.g. 7%), and still provide hindsight log replay times orders of magnitude faster than restarting training from scratch.
Long-term changes of ozone in the UTLS (Upper Troposphere / Lower Stratosphere) reflect the response to decreases in the stratospheric concentrations of ozone-depleting substances as well as changes in the stratospheric circulation induced by climate change. To date, studies of UTLS ozone changes and variability have relied mainly on satellite and in-situ observations as well as chemistry-climate model simulations. By comparison, the potential of reanalysis ozone data remains relatively untapped. This is despite evidence from recent studies, including detailed analyses conducted under SPARC (Scalable Processor Architecture) Reanalysis Intercomparison Project (S-RIP), that demonstrate that stratospheric ozone fields from modern atmospheric reanalyses exhibit good agreement with independent data while delineating issues related to inhomogeneities in the assimilated observations. In this presentation, we will explore the possibility of inferring long-term geographically and vertically resolved behavior of the lower stratospheric (LS) ozone from NASA's MERRA-2 (Modern-Era Retrospective Analysis for Research and Applications -2) reanalysis after accounting for the few known discontinuities and gaps in its assimilated input data. This work builds upon previous studies that have documented excellent agreement between MERRA-2 ozone and ozonesonde observations in the LS. Of particular importance is a relatively good vertical resolution of MERRA-2 allowing precise separation of tropospheric and stratospheric ozone contents. We also compare the MERRA-2 LS ozone results with the recently completed 37-year simulation produced using Goddard Earth Observing System in "replay" mode coupled with the GMI (Global Modeling Initiative) chemistry mechanism. Replay mode dynamically constrains the model with the MERRA-2 reanalysis winds, temperature, and pressure. We will emphasize the areas of agreement of the reanalysis and replay and interpret differences between them in the context of our increasing understanding of model transport driven by assimilated winds.
Hurricanes Juan (1985) and Harvey (2017) bear remarkable similarities because of the exceptional amount of accumulated precipitation over coastal regions in the Gulf, consequent to the peculiar and rare situation in which a hurricane stalls for several days on a coastal area, displaying minimal progress. This is particularly dangerous because: a) terrestrial drag affect only part of the circulation so that weakening is not as fast as when there is rapid motion inland; b) a part of the circulation remains over water and as such low-level moisture flux convergence is maintained; c) precipitation lasts for a very long time over the same area; d) moisture recycling from flooded areas further contributes to maintaining flood conditions. From the forecast perspective, hurricanes with track singularities such as stalling points, loops, or multiple landfalls over the same region are exceptionally difficult to predict because they result from small variations of opposing mechanisms and are therefore intrinsically unstable. Intensity fluctuations and track singularities are strongly coupled. The representation of fluxes undergoing differential drag (ocean vs. land) inside the same circulation is critically dependent upon resolution. Recent modeling development efforts within the NASA Global Modeling and Assimilation Office (GMAO) have brought the horizontal resolution of the global cubed-sphere Goddard Earth Observing System (GEOS) atmospheric model down to 12 km. This has led to innovative findings in the field of tropical cyclone (TC) modeling within a global framework. The GEOS atmospheric model and data assimilation system (DAS), used to produce the MERRA-2 (Modern Era Retrospective analysis for Research and Applications) reanalysis, has a "Replay" capability that adds a forcing term to the model equations to constrain it to follow a specified trajectory. This capability is used to produce global, downscaled versions of the 50 km MERRA-2 reanalysis at a 12.5 km resolution. The full DAS and replay modes both use the Incremental Analysis Update or IAU for applying an analysis increment to the model's background state four times each day at 00Z, 06Z, 12Z and 18Z. In a full DAS mode, a six-hour free run of the model is made (Predictor Step), writing states that will be used as backgrounds or first-guesses for the analysis centered on 00Z. The Analysis system then produces its best estimate of the 00Z state by blending backgrounds with observations. The difference between this best estimate and the model's original first guess at 00Z is treated as an analysis increment and added uniformly as a forcing term to the model during a second, now forced, run of the 12Z to 03Z window, which produces the assimilated results. Replay mode is very similar, except only the 00Z background is needed from the predictor segment, which is thus half as long, and instead of doing an analysis, a previously analyzed state at 00Z is read in. The increment is then computed and applied to the corrector step in the same way as in the original DAS. When downscaling horizontal spatial filters can be applied to the "analysis" increment used to force the Corrector step, so as to reproduce the full meteorological state or selectively choose or discard certain forcings. In this presentation, full global hindcast simulations of Juan and operational forecasts of Harvey are compared.
A technique for speculatively executing load-dependent instructions includes detecting that a memory ordering consistency queue is full for a completed load instruction. The technique also includes storing data loaded by the completed load instruction into a storage location for storing data when the memory ordering consistency queue is full. The technique further includes speculatively executing instructions that are dependent on the completed load instruction. The technique also includes in response to a slot becoming available in the memory ordering consistency queue, replaying the load instruction. The technique further includes in response to receiving loaded data for the replayed load instruction, testing for a data mis-speculation by comparing the loaded data for the replayed load instruction with the data loaded by the completed load instruction that is stored in the storage location.