Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “statistics of extremes”

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 55 records · Page 3
All sourcesNASA NTRSDOE OSTI
NASA NTRS1994

Statistic analysis of annual total ozone extremes for the period 1964-1988

Annual extremes of total column amount of ozone (in the period 1964-1988) from a network of 29 Dobson stations have been examined using the extreme value analysis. The extremes have been calculated as the highest deviation of daily mean total ozone from its long-term monthly mean, normalized by the monthly standard deviations. The extremes have been selected from the direct-Sun total ozone observations only. The extremes resulting from abrupt changes in ozone (day to day changes greater than 20 percent) have not been considered. The ordered extremes (maxima in ascending way, minima in descending way) have been fitted to one of three forms of the Fisher-Tippet extreme value distribution by the nonlinear least square method (Levenberg-Marguard method). We have found that the ordered extremes from a majority of Dobson stations lie close to Fisher-Tippet type III. The extreme value analysis of the composite annual extremes (combined from averages of the annual extremes selected at individual stations) has shown that the composite maxima are fitted by the Fisher-Tippet type III and the composite minima by the Fisher-Tippet type I. The difference between the Fisher-Tippet types of the composite extremes seems to be related to the ozone downward trend. Extreme value prognoses for the period 1964-2014 (derived from the data taken at: all analyzed stations, the North American, and the European stations) have revealed that the prognostic extremes are close to the largest annual extremes in the period 1964-1988 and there are only small regional differences in the prognoses.

Krzyscin, Janusz W.↗
NASA NTRS1999

Statistical Approach for Determining the Onsets/Durations of ENSO Cycle Extremes

During the interval of 1950-mid 1998, some 16 El Nino and 10 La Nina have been identified on the basis of sea surface temperature in the Nino 3.4 region, these 26 events representing the extremes of the quasi-periodic ENSO cycle. Statistical aspects of these events are examined. Surprisingly, the durations of El Nino and La Nina appear to be strongly bifurcated into shorter and longer duration classes, as do the recurrence periods of El Nino. Moreover, the duration of an El Nino appears to provide a statistically meaningful indication as to when to expect the next onset of El Nino. Because the last El Nino had its onset in April 1997 and was of longer duration, onset of the next El Nino, probably, will not occur until after February 2000.

Wilson, Robert M.↗
NASA NTRS2019

MERRA-2 Data and Analytic Services at NASA GES DISC for Climate Extremes Study

NASA's climate reanalysis datasets from the Modern Era Retrospective-analysis for Research and Applications, Version 2 (MERRA-2) contains numerous long-term atmosphere, land, and ocean data products from 1980-present. MERRA-2 datasets, such as precipitation, soil moisture, and temperature, have been used widely to study extreme events. The native archived MERRA-2 data files are day-file (hourly time interval) and month-file, containing up to 125 parameters in one file. Due to the large number of data files and volumes, it is challenging for users, especially the applications research community, to handle the original hourly data files for long time periods to analyze extreme events. In this presentation, we review MERRA-2 data for studies of extreme conditions, and demonstrate analytic services at the NASA Goddard Earth Sciences Data and Information Services Center (GES DISC). One of the current operational services, 'subsetter', allows users to download only specific data of interest, i.e. data selected by parameter, region, and time period. New services are under development that will provide more 'on-the-fly' statistical calculations when downloading data; improve efficiency when accessing long time-series data. We will provide additional "How-to" resources that include step-by-step instructions on data access and usage. We have tested restructuring of day-files in an optimized data cube, which has significantly improved system performance for accessing long time-series. Overall performance is associated with cube size and structure, data compression method, and how the data are accessed. The optimized data cube structure will enable better online analytic services for statistical analysis and extreme events mining. To demonstrate the service, we use an extreme drought associated with the anomalous 2016 monsoon over southern Asia. This prototype time-series service may be augmented in the cloud infrastructure in the future.

data access↗
NASA NTRS2017

The Relationships Between the Trends of Mean and Extreme Precipitation

This study provides a better understanding of the relationships between the trends of mean and extreme precipitation in two observed precipitation data sets: the Climate Prediction Center Unified daily precipitation data set and the Global Precipitation Climatology Program (GPCP) pentad data set. The study employs three kinds of definitions of extreme precipitation: (1) percentile, (2) standard deviation and (3) generalize extreme value (GEV) distribution analysis for extreme events based on local statistics. Relationship between trends in the mean and extreme precipitation is identified with a novel metric, i.e. area aggregated matching ratio (AAMR) computed on regional and global scales. Generally, more (less) extreme events are likely to occur in regions with a positive (negative) mean trend. The match between the mean and extreme trends deteriorates for increasingly heavy precipitation events. The AAMR is higher in regions with negative mean trends than in regions with positive mean trends, suggesting a higher likelihood of severe dry events, compared with heavy rain events in a warming climate. AAMR is found to be higher in tropics and oceans than in the extratropics and land regions, reflecting a higher degree of randomness and more important dynamical rather than thermodynamical contributions of extreme events in the latter regions.

global warming↗
NASA NTRS1984

Origin of the Greenland Ice Sheet

A climate model resolving the seasonal cycle and the two horizontal dimensions was developed and applied to several problems of current interest. Models of this type are useful when for various reasons a general circulation model experiment is not warranted or not feasible. For example, in cases where the signal to natural variability is small it may be advantageous to first consider such a statistical dynamical model because extremely long runs may be necessary in the application. In this case the simpler statistical dynamical model serves as a pilot study device. The model developed is a thermodynamic model whose solution yields the equilibrium seasonal cycle for the surface temperature field over the globe. The model is essentially a statement of the conservation of heat energy for individual columns of the earth atmosphere system. Various terms such as the infrared radiation flux to space are parameterized with earth radiation budget data from satellites such as Nimbus 6. The primary agent modulating the seasonal cycle amplitude is the heat capacity per unit area which is a strong function of surface type -- ocean surface can store 60 times more heat per unit time than land. By adjusting its few empirical parameters the model can be brought into remarkable agreement with the observed seasonal cycle. The model is then very useful for looking at the dependence of the seasonal cycle of the temperature on such externally defined variables as the Earth's orbital elements (eccentricity, tilt, precession of equinoxes) or the configuration of land-sea geography which can be changed by continental drift.

North, G. R.↗
NASA NTRS2004

GPS FOM Chimney Analysis using Generalized Extreme Value Distribution

Many a time an objective of a statistical analysis is to estimate a limit value like 3-sigma 95% confidence upper limit from a data sample. The generalized Extreme Value Distribution method can be profitably employed in many situations for such an estimate. . .. It is well known that according to the Central Limit theorem the mean value of a large data set is normally distributed irrespective of the distribution of the data from which the mean value is derived. In a somewhat similar fashion it is observed that many times the extreme value of a data set has a distribution that can be formulated with a Generalized Distribution. In space shuttle entry with 3-string GPS navigation the Figure Of Merit (FOM) value gives a measure of GPS navigated state accuracy. A GPS navigated state with FOM of 6 or higher is deemed unacceptable and is said to form a FOM 6 or higher chimney. A FOM chimney is a period of time during which the FOM value stays higher than 5. A longer period of FOM of value 6 or higher causes navigated state to accumulate more error for a lack of state update. For an acceptable landing it is imperative that the state error remains low and hence at low altitude during entry GPS data of FOM greater than 5 must not last more than 138 seconds. I To test the GPS performAnce many entry test cases were simulated at the Avionics Development Laboratory. Only high value FoM chimneys are consequential. The extreme value statistical technique is applied to analyze high value FOM chimneys. The Maximum likelihood method is used to determine parameters that characterize the GEV distribution, and then the limit value statistics are estimated.

Ott, Rick↗
NASA NTRS2015

Life Beyond the Planet of Origin and Implications for the Search for Life on Mars

Outer space is vast, cold, devoid of matter, radiation filled with essentially no gravity. These factors present an environmental challenge for any form of life. Earth's biosphere has evolved for more than 3 billion years shielded from the hostile environment of outer space by the protective blanket of the atmosphere and magnetosphere. Space is a nutritional wasteland with no liquid water and readily available organic carbon. Moving beyond a life's planet of origin requires a means for transport, the ability to withstand transport, and the ability to colonize, thrive and ultimately evolve in the new environment. Can life survive beyond its home planet? The key to answering this question is to identify organisms that first have the ability to withstand space radiation, space vacuum desiccation and time in transit, and second the ability to grow in an alien environment. Within the last 60 years space technology allowed us to transport life beyond Earth's protective shield so we may study, in situ, their responses to selected conditions of space. To date a variety of microbes ranging from viruses, to Bacteria, to Archaea, to Eukarya have been tested in the space environment. Most died instantly, but not all. These studies revealed that ultraviolet radiation is the near-term lethal agent, while hard radiation is the long-term lethal agent when the organism is shielded from ultraviolet radiation. In fact, bacterial spores, halophilic cyanobacteria and Archaea as well as some lichens survive very well if protected from ultraviolet radiation [1]. Some microbes, then, may be able to survive the trip in outer space to Mars on a spacecraft or in a meteorite. Once on Mars can a terrestrial microbe survive? Although the conditions on Mars are not as harsh as those in space, they are not hospitable for a terrestrial microbe. Studies, however, have shown that certain microbes that can survive in space for several years may also be able to survive on Mars if protected from ultraviolet radiation [1]. Laboratory simulation experiments using a mock-up of the Phoenix lander have shown that microbes transported to the surface of Mars on a spacecraft come off the spacecraft and mix into the Martian regolith [2]. Additionally, studies simulating Martian dust storms demonstrate that microbes can survive in the Martian wind blown dust and be scattered across the Martian surface away from the spacecraft. Would these microbes that may survive on Mars metabolize and propagate? Growth requires liquid water, a carbon source and an energy source. Survival on Mars also requires protection from ultraviolet radiation. In the cold, dry environment of Mars the probability of microbial metabolism and growth at or just beneath the surface is extremely low. Although the probability is low, Mars may be contaminated with potentially live terrestrial organisms. In light of that statistic we must be extremely diligent and cautious in our search for Martian life. If we are not cautious we may find life on Mars and it may be a contaminant from Earth.

Survival in Space↗
NASA NTRS1998

Statistical Aspects of ENSO Events (1950-1997) and the El Nino-Atlantic Intense Hurricane Activity Relationship

On the basis of Trenberth's quantitative definition for marking the occurrence of an El Nino (or La Nina), one can precisely identify by month and year the starts and ends of some 15 El Nino and 10 La Nina events during the interval of 1950-1997, an interval corresponding to the most reliable for cataloging intense hurricane activity in the Atlantic basin (i.e., those of category 3-5 on the Saffir-Simpson hurricane scale). The main purpose of this investigation is primarily two-fold: First, the statistical aspects of these identified extremes and the intervening periods between them (called "interludes") are examined and, second, the statistics of the seasonal frequency of intense hurricanes in comparison to the extremes and interludes are determined. This study clearly demonstrates that of the last 48 hurricane seasons, 20 (42 percent) can be described as being "El Nino-related" (i.e., an El Nino was in progress during all, or part, of the yearly hurricane season--June-November), 13 (27 percent) as "La Nina-related" (i.e., a La Nina was in progress during all, or part, of the yearly hurricane season), and 15 (31 percent) as "interlude-related" (i.e., neither an El Nino nor a La Nina was in progress during any portion of the yearly hurricane season). Combining the latter two subgroups into a single grouping called "non-El Nino-related" seasons, one finds that they have had a mean frequency of intense hurricanes measuring 2.8 events per season, while the El Nino-related seasons have had a mean frequency of intense hurricanes measuring 1.3 events per season, where the observed difference in the means is inferred to be statistically important at the 99.8-percent level of confidence. Therefore, as previously shown more than a decade ago using a different data set, there undeniably exists an El Nino-Atlantic hurricane activity relationship, one which also extends to the class of intense hurricanes. During the interval of 1950-1997, fewer intense hurricanes occurred during El Nino-related seasons (always less than or equal to 3 and usually less than or equal to 2, this latter value having been true for 18 of the 20 El Nino-related seasons), while more usually occurred during non-El Nino-related seasons (typically greater than or equal to 2, having been true for 22 of the 28 non-El Nino-related seasons). Implications for the 1998 and 1999 hurricane seasons are discussed.

Wilson, Robert M.↗
NASA NTRS2002

Incorporating Biological Knowledge into Evaluation of Casual Regulatory Hypothesis

Biological data can be scarce and costly to obtain. The small number of samples available typically limits statistical power and makes reliable inference of causal relations extremely difficult. However, we argue that statistical power can be increased substantially by incorporating prior knowledge and data from diverse sources. We present a Bayesian framework that combines information from different sources and we show empirically that this lets one make correct causal inferences with small sample sizes that otherwise would be impossible.

Chrisman, Lonnie↗
NASA NTRS2017

Classification of Meteorological Influences Surrounding Extreme Precipitation Events in the United States using the MERRA-2 Reanalysis

Extreme precipitation events can have a large impact on society through flooding that can result in property destruction, crop losses, economic losses, the spread of water-borne diseases, and fatalities. Observations indicate there has been a statistically significant increase in extreme precipitation events over the past 15 years in the Northeastern United States and other localized regions of the country have become crippled with record flooding events, for example, the flooding that occurred in the Southeast United States associated with Hurricane Matthew in October 2016. Extreme precipitation events in the United States can be caused by various meteorological influences such as extratropical cyclones, tropical cyclones, mesoscale convective complexes, general air mass thunderstorms, upslope flow, fronts, and the North American Monsoon. Reanalyses, such as the Modern Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2), have become a pivotal tool to study the meteorology surrounding extreme precipitation events. Using days classified as an extreme precipitation events based on a combination of observational gauge and radar data, two techniques for the classification of these events are used to gather additional information that can be used to determine how events have changed over time using atmospheric data from MERRA-2. The first is self organizing maps, which is an artificial neural network that uses unsupervised learning to cluster like patterns and the second is an automated detection technique that searches for characteristics in the atmosphere that define a meteorological phenomena. For example, the automated detection for tropical cycles searches for a defined area of suppressed sea level pressure, alongside thickness anomalies aloft, indicating the presence of a warm core. These techniques are employed for extreme precipitation events in preselected regions that were chosen based an analysis of the climatology of precipitation.

Monsoon↗
NASA NTRS2016

Geoelectric Hazard Maps for the Continental United States

In support of a multiagency project for assessing induction hazards, we present maps of extreme-value geoelectric amplitudes over about half of the continental United States. These maps are constructed using a parameterization of induction: estimates of Earth surface impedance, obtained at discrete geographic sites from magnetotelluric survey data, are convolved with latitude-dependent statistical maps of extreme-value geomagnetic activity, obtained from decades of magnetic observatory data. Geoelectric amplitudes are estimated for geomagnetic waveforms having 240 s sinusoidal period and amplitudes over 10 min that exceed a once-per-century threshold. As a result of the combination of geographic differences in geomagnetic activity and Earth surface impedance, once-per-century geoelectric amplitudes span more than 2 orders of magnitude and are an intricate function of location. For north-south induction, once-per-century geoelectric amplitudes across large parts of the United States have a median value of 0.26 Vkm; for east-west geomagnetic variation the median value is 0.23 Vkm. At some locations,once-per-century geoelectric amplitudes exceed 3 Vkm.

Love, Jeffrey J.↗
NASA NTRS1972

An investigation to improve the Menhaden fishery prediction and detection model through the application of ERTS-A data

The author has identified the following significant results. Preliminary analyses indicate that several important relationships have been observed utilizing ERTS-1 imagery. Of most significance is that in the Mississippi Sound, as elsewhere, considerable detail exists as to turbidity patterns in the water column. Simple analysis is complicated by the apparent interaction between actual turbidity, turbidity induced by shoal water, and actual imaging of the bottom in extreme shoal water. A statistical approach is being explored which shows promise of at least partially separating these effects so that partitioning of true turbid plumes can be accomplished. This partitioning is of great importance to this program in that supportive data seem to indicate that menhaden occur more frequently in turbid areas. In this connection four individual captures have been associated with a major turbid feature imaged on 6 August. If a significant relationship between imaged turbid features and catch distribution can be established, for example by graphic and/or numeric analysis, it will represent a major advancement for short term prediction of commercially accessible menhaden.

Maughan, P. M.↗
NASA NTRS1988

Navigation with noncoherent data - A demonstration for VEGA Venus flyby phase

Deep Space navigation with noncoherent (one-way) data types is demonstrated for the VEGA Venus flyby phase under extreme conditions. Estimates and statistics are computed using one-way Doppler and wideband Very Long Baseline Interferometry (VLBI) data. The behavior of the onboard oscillator is modeled for both spacecraft to obtain useful orbit determination results. Even with this limitation, it is demonstrated that one-way data solutions are comparable with the solutions using both Soviet sparse coherent (two-way) and wideband VLBI data. During the useful life time of VEGA balloons, the two solutions differ by a maximum of 4.7 km in position and 7.6 cm/sec in velocity for VEGA 1 and by a maximum of 8 km and 42 cm/sec for VEGA 2.

Bhat, Ramachandra S.↗
NASA NTRS2016

Accurately Characterizing the Importance of Wave-Particle Interactions in Radiation Belt Dynamics: The Pitfalls of Statistical Wave Representations

Wave-particle interactions play a crucial role in energetic particle dynamics in the Earths radiation belts. However, the relative importance of different wave modes in these dynamics is poorly understood. Typically, this is assessed during geomagnetic storms using statistically averaged empirical wave models as a function of geomagnetic activity in advanced radiation belt simulations. However, statistical averages poorly characterize extreme events such as geomagnetic storms in that storm-time ultralow frequency wave power is typically larger than that derived over a solar cycle and Kp is a poor proxy for storm-time wave power.

Murphy, Kyle R.↗
NASA NTRS2017

Ask-the-expert: Active Learning Based Knowledge Discovery Using the Expert

Often the manual review of large data sets, either for purposes of labeling unlabeled instances or for classifying meaningful results from uninteresting (but statistically significant) ones is extremely resource intensive, especially in terms of subject matter expert (SME) time. Use of active learning has been shown to diminish this review time significantly. However, since active learning is an iterative process of learning a classifier based on a small number of SME-provided labels at each iteration, the lack of an enabling tool can hinder the process of adoption of these technologies in real-life, in spite of their labor-saving potential. In this demo we present ASK-the-Expert, an interactive tool that allows SMEs to review instances from a data set and provide labels within a single framework. ASK-the-Expert is powered by an active learning algorithm for training a classifier in the backend. We demonstrate this system in the context of an aviation safety application, but the tool can be adopted to work as a simple review and labeling tool as well, without the use of active learning.

software↗
NASA NTRS2017

Ask-the-Expert: Active Learning Based Knowledge Discovery Using the Expert

Often the manual review of large data sets, either for purposes of labeling unlabeled instances or for classifying meaningful results from uninteresting (but statistically significant) ones is extremely resource intensive, especially in terms of subject matter expert (SME) time. Use of active learning has been shown to diminish this review time significantly. However, since active learning is an iterative process of learning a classifier based on a small number of SME-provided labels at each iteration, the lack of an enabling tool can hinder the process of adoption of these technologies in real-life, in spite of their labor-saving potential. In this demo we present ASK-the-Expert, an interactive tool that allows SMEs to review instances from a data set and provide labels within a single framework. ASK-the-Expert is powered by an active learning algorithm for training a classifier in the back end. We demonstrate this system in the context of an aviation safety application, but the tool can be adopted to work as a simple review and labeling tool as well, without the use of active learning.

GUI↗