Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Inference”

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.

At least 775 records · Page 43

Length Scales of Hydrological Variability as Inferred from SMAP Soil Moisture Retrievals

Hydrological variability at a given location is characterized in part by a horizontal length scale – a measure of how far one can travel from that location and still see similar time variations of a hydrological variable of interest. Here, using Level-2 soil moisture retrievals produced by the NASA Soil Moisture Active Passive (SMAP) mission, we compute global distributions of these length scales for the Northern Hemisphere warm and cold seasons (May-September and November-March, respectively). The length scales show significant spatial and seasonal variability, with, as expected, much larger values (e-folding scales of greater than 500 km) often seen in the cold season, when convective rainfall is less prominent. The SMAP-derived length scales are found to be largely consistent with those derived directly, where possible, from precipitation measurements. This suggests a unique value of the retrievals: outside of well-instrumented areas, satellite-based soil moisture datasets have the potential to provide otherwise unattainable estimates of the horizontal length scales of hydrological variability.

Hydrology↗

Genetic Algorithm for Optimization of Neural Networks for Bayesian Inference of Model Uncertainty

The objective of this work was to develop a genetic optimization algorithm that can design a neural network capable of producing uncertainty estimates along with predictions. This algorithm is necessary because the inclusion of uncertainty modeling in a neural network greatly complicates the network’s design space, making the development of a converging model extremely difficult and time consuming. The genetic algorithm presented in this work uses a number of value ranges for various configurable neural network parameters to create a randomly generated population of network architectures. The initially generated population is then evolved over the course of several generations, with the best performing models breeding to produce novel network configurations. Mutations are randomly applied to the network designs to facilitate the development of adaptations beneficial to the task being performed. An experiment was conducted to validate the proposed algorithm, in which the genetic optimizer was tasked with producing a neural network capable of predicting the sound pressure level (SPL) resulting from jet-surface interaction (JSI) noise. The data used for this task was generated at the NASA Glenn Research Center in the Aero-Acoustic Propulsion Laboratory. Starting with an initial population size of 35 randomly generated networks, and evolved over the course of 10 generations, the genetic algorithm produced a design able to predict SPL as a result of JSI noise within 0.272 dB, on average.

Genetic algorithm↗

Revisiting the Effectiveness Of HCHO/NO2 Ratios for Inferring Ozone Sensitivity to Its Precursors Using High Resolution Airborne Remote Sensing Observations in A High Ozone Episode During the KORUS-AQ Campaign

The nonlinear chemical processes involved in ozone production (P(O3)) have necessitated using proxy indicators to convey information about the primary dependence of P(O3) on volatile organic compounds (VOCs) or nitrogen oxides (NOx). In particular, the ratio of remotely sensed columns of formaldehyde (HCHO) to nitrogen dioxide (NO2) has been widely used for studying O3 sensitivity. Previous studies found that the errors in retrievals and the incoherent relationship between the column and the near-surface concentrations are a barrier in applying the ratio in a robust way. In addition to these obstacles, we provide calculational-observational evidence, using an ensemble of 0-D photochemical box models constrained by DC-8 aircraft measurements on an ozone event during the Korea-United States Air Quality (KORUS-AQ) campaign over Seoul, to demonstrate the chemical feedback of NO2 on the formation of HCHO is a controlling factor for the transition line between NOx-sensitive and NOx-saturated regimes. A fixed value (~2.7) of the ratio of the chemical loss of NOx (LNOx) to the chemical loss of HO2+RO2 (LROx) perceptibly differentiates the regimes. Following this value, data points with a ratio of HCHO/NO2 less than 1 can be safely classified as NOx-saturated regime, whereas points with ratios between 1 and 4 fall into one or the other regime. We attribute this mainly to the HCHO-NO2 chemical relationship causing the transition line to occur at larger (smaller) HCHO/NO2 ratios in VOC-rich (VOC-poor) environments. We then redefine the transition line to LNOx/LROx~2.7 that accounts for the HCHO-NO2 chemical relationship leading to HCHO = 3.7 × (NO2 – 1.14 × 1016 molec.cm-2). Although the revised formula is locally calibrated (i.e., requires for readjustment for other regions), its mathematical format removes the need for having a wide range of thresholds used in HCHO/NO2 ratios that is a result of the chemical feedback. Therefore, to be able to properly take the chemical feedback into consideration, the use of HCHO = a × (NO2 – b) formula should be preferred to the ratio in future works. We then use the Geostationary Trace gas and Aerosol Sensor Optimization (GeoTASO) airborne instrument to study O3 sensitivity in Seoul. The unprecedented spatial (250 × 250 m2) and temporal (~every 2 h) resolutions of HCHO and NO2 observations form the sensor enhance our understanding of P(O3) in Seoul; rather than providing a crude label for the entire city, more in-depth variabilities in chemical regimes are observed that should be able to inform mitigation strategies correspondingly.

KORUS-AQ; formaldehyde; NOx; VOC; OMI; GeoTASO; oz↗

Inference of Precipitation in Warm Stratiform Clouds using Remotely Sensed Observations of the Cloud Top Droplet Size Distribution

Drizzle is a common feature of warm stratiform clouds and it influences their radiative effects by modulating their physical properties and lifecycle. An important component of drizzle formation are processes that lead to a broadening of the droplet size distribution (DSD). Here, we examine observations of cloud and drizzle properties retrieved using colocated airborne measurements from the Research Scanning Polarimeter and the Third Generation Airborne Precipitation Radar. We observe a bimodal DSD as the aircraft transects drizzling open-cells whereby the larger mode reaches a maximum size near cloud center and the smaller mode remains relatively constant in size. We review similarities between our observations with droplet growth processes and their connections with precipitation onset. We estimate droplet sedimentation using the cloud top DSD and find a correlation with rain water path of 0.82. We also examine how changes in liquid water paths and droplet concentrations may act to enhance or suppress precipitation.

Droplet size distribution↗

Discrete Recurrent Neural Networks for Grammatical Inference

Recurrent neural networks have recently been shown to have the ability to learn regular and context-free grammars from examples. We show that while conventional analog recurrent networks try to form clusters in activation space to represent discrete states of the grammars during learning, and can be successful in doing so, the clusters so formed tend to become unstable as longer and longer test input strings are presented to the network.

Neural↗

Europa's Differentiated Internal Structure: Inferences from Two Galileo Encounters

Doppler data generated with the Galileo spacecraft's radio carrier wave during two Europa encounters on 19 December 1996 (E4) and 20 February 1997 (E6) were used to measure Euroapa's external gravitational field. The measurements indicate that Europa has a predominantly water ice-liquid outer shell about 100 to 200 klilometers thick and a deep interior with a density in excess of about 4000 kilograms per cubic meter.

Galileo↗

Using Machine Learning to Infer Pre-Entry Properties for Asteroid Threat Analysis

Accurately assessing asteroid threats relies on knowledge of the asteroid’s pre-entry properties such as size, velocity, and mass. Directly measuring these properties can be infeasible due to the sparsity of events and the accuracy and fidelity of various sensors. Current analysis of an asteroid’s pre-entry properties involves modeling the asteroid’s entry into the Earth’s atmosphere. This process can be time consuming and can require manual adjustment of uncertain modeling specific parameters. NASA Ames has developed a genetic algorithm that can help automate asteroid modeling using the Fragment-Cloud Model (FCM). The algorithm generates realistic energy deposition curves based on actual energy deposition curves from real, observed asteroids. By using these synthetic, labeled energy deposition curves, we developed a one-dimensional convolutional neural network that can predict an asteroid’s pre-entry parameters.

ATAP↗