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.
This slide presentation reviews the precursor analysis for flight and ground based anomaly risk significance. It includes information on accident precursor analysis, real models vs. models, and probabilistic analysis.
System properties such as “safety” and “dependability” cannot, in practice, be proven, and must be argued in an “assurance case” aimed at supporting risk-acceptance decisions that have to be made by system acquirers and/or regulatory authorities. The paper is concerned with applications of the “assurance case” idea early in design and development of new systems, when (apart from dedicated testing) the only available operating experience information derives from previous (non-identical) systems. Much of the discussion is based on an evolving acquisition model at the US National Aeronautics and Space Administration; previously, most major systems were developed in-house, but some major systems will now be developed by and acquired from commercial providers. Key points discussed include the following. (1) By promoting a particular kind of focused discussion between acquirers and providers, the use of assurance cases should be particularly valuable under the new acquisition model. (2) In principle, objectives-driven (sometimes called “performance-based”) approaches to assurance of performance have significant advantages in cases where they are applicable. (3) For truly novel systems, completeness of the safety analysis is a significant issue; it is important for the assurance case to include a commitment by the provider (or applicant) to seriously pursue analysis of operating experience, so that previously unrecognized hazards can be identified and addressed. (4) Inquiries into major accidents often point to deficiencies in management oversight in all parts of the life cycle; management processes need to be addressed in the formulation and the implementation of an assurance case. Under the new acquisition model, these considerations imply a serious reconsideration of the way in which the development process is managed by both providers and acquirers.
System properties such as “safety” and “dependability” cannot, in practice, be proven, and must be argued in an “assurance case” aimed at supporting risk-acceptance decisions that have to be made by system acquirers and/or regulatory authorities. The paper is concerned with applications of the “assurance case” idea early in design and development of new systems, when (apart from dedicated testing) the only available operating experience information derives from previous (non-identical) systems. Much of the discussion is based on an evolving acquisition model at the US National Aeronautics and Space Administration; previously, most major systems were developed in-house, but some major systems will now be developed by and acquired from commercial providers. Key points discussed include the following. (1) By promoting a particular kind of focused discussion between acquirers and providers, the use of assurance cases should be particularly valuable under the new acquisition model. (2) In principle, objectives-driven (sometimes called “performance-based”) approaches to assurance of performance have significant advantages in cases where they are applicable. (3) For truly novel systems, completeness of the safety analysis is a significant issue; it is important for the assurance case to include a commitment by the provider (or applicant) to seriously pursue analysis of operating experience, so that previously unrecognized hazards can be identified and addressed. (4) Inquiries into major accidents often point to deficiencies in management oversight in all parts of the life cycle; management processes need to be addressed in the formulation and the implementation of an assurance case. Under the new acquisition model, these considerations imply a serious reconsideration of the way in which the development process is managed by both providers and acquirers.
The objective of this paper is to present a method for importance analysis in parametric probabilistic modeling where the result of interest is the identification of potential engineering vulnerabilities associated with postulated anomalies in system behavior. In the context of Accident Precursor Analysis (APA), under which this method has been developed, these vulnerabilities, designated as anomaly vulnerabilities, are conditions that produce high risk in the presence of anomalous system behavior. The method defines a parameter-specific Parameter Vulnerability Importance measure (PVI), which identifies anomaly risk-model parameter values that indicate the potential presence of anomaly vulnerabilities, and allows them to be prioritized for further investigation. This entails analyzing each uncertain risk-model parameter over its credible range of values to determine where it produces the maximum risk. A parameter that produces high system risk for a particular range of values suggests that the system is vulnerable to the modeled anomalous conditions, if indeed the true parameter value lies in that range. Thus, PVI analysis provides a means of identifying and prioritizing anomaly-related engineering issues that at the very least warrant improved understanding to reduce uncertainty, such that true vulnerabilities may be identified and proper corrective actions taken.
In previous studies, a variety of unsupervised anomaly detection techniques for anomaly detection were applied to SSME (Space Shuttle Main Engine) data. The observed results indicated that the identification of certain anomalies were specific to the algorithmic method under consideration. This is the reason why one of the follow-on goals of these previous investigations was to build an architecture to support the best capabilities of all algorithms. We appeal to that goal here by investigating a cascade, serial architecture for the best performing and most suitable candidates from previous studies. As a precursor to a formal ROC (Receiver Operating Characteristic) curve analysis for validation of resulting anomaly detection algorithms, our primary focus here is to investigate the model fidelity as measured by variants of the AIC (Akaike Information Criterion) for state-space based models. We show that placing constraints on a state-space model during or after the training of the model introduces a modest level of suboptimality. Furthermore, we compare the fidelity of all candidate models including those embodying the cascade, serial architecture. We make recommendations on the most suitable candidates for application to subsequent anomaly detection studies as measured by AIC-based criteria.
Catastrophic accidents are usually preceded by precursory events that, although observable, are not recognized as harbingers of a tragedy until after the fact. In the nuclear industry, the Three Mile Island accident was preceded by at least two events portending the potential for severe consequences from an underappreciated causal mechanism. Anomalies whose failure mechanisms were integral to the losses of Space Transportation Systems (STS) Challenger and Columbia had been occurring within the STS fleet prior to those accidents. Both the Rogers Commission Report and the Columbia Accident Investigation Board report found that processes in place at the time did not respond to the prior anomalies in a way that shed light on their true risk implications. This includes the concern that, in the words of the NASA Aerospace Safety Advisory Panel (ASAP), "no process addresses the need to update a hazard analysis when anomalies occur" At a broader level, the ASAP noted in 2007 that NASA "could better gauge the likelihood of losses by developing leading indicators, rather than continue to depend on lagging indicators". These observations suggest a need to revalidate prior assumptions and conclusions of existing safety (and reliability) analyses, as well as to consider the potential for previously unrecognized accident scenarios, when unexpected or otherwise undesired behaviors of the system are observed. This need is also discussed in NASA's system safety handbook, which advocates a view of safety assurance as driving a program to take steps that are necessary to establish and maintain a valid and credible argument for the safety of its missions. It is the premise of this handbook that making cases for safety more experience-based allows NASA to be better informed about the safety performance of its systems, and will ultimately help it to manage safety in a more effective manner. The APA process described in this handbook provides a systematic means of analyzing candidate accident precursors by evaluating anomaly occurrences for their system safety implications and, through both analytical and deliberative methods used to project to other circumstances, identifying those that portend more serious consequences to come if effective corrective action is not taken. APA builds upon existing safety analysis processes currently in practice within NASA, leveraging their results to provide an improved understanding of overall system risk. As such, APA represents an important dimension of safety evaluation; as operational experience is acquired, precursor information is generated such that it can be fed back into system safety analyses to risk-inform safety improvements. Importantly, APA utilizes anomaly data to predict risk whereas standard reliability and PRA approaches utilize failure data which often is limited and rare.
Accident Precursor Analysis (APA) serves as the bridge between existing risk modeling activities, which are often based on historical or generic failure statistics, and system anomalies, which provide crucial information about the failure mechanisms that are actually operative in the system and which may differ in frequency or type from those in the various models. These discrepancies between the models (perceived risk) and the system (actual risk) provide the leading indication of an underappreciated risk. This paper presents an APA process developed specifically for NASA Earth-to-Orbit space systems. The purpose of the process is to identify and characterize potential sources of system risk as evidenced by anomalous events which, although not necessarily presenting an immediate safety impact, may indicate that an unknown or insufficiently understood risk-significant condition exists in the system. Such anomalous events are considered accident precursors because they signal the potential for severe consequences that may occur in the future, due to causes that are discernible from their occurrence today. Their early identification allows them to be integrated into the overall system risk model used to intbrm decisions relating to safety.
Thermoacoustic instability (TAI) is a critical issue in modern lean-burn gas-turbine combustors, which is induced by a strong coupling between the resonant combustor acoustics and fluctuations in the heat release rate. This instability may lead to high-amplitude pressure waves that generate undesirable noise levels as well as fatigue stresses in mechanical structures of the combustor. The intense pressure fluctuations due to TAI may also cause large flow perturbations and possibly flow reversal that may lead to flame oscillations, flame liftoff, and even flame blow-out. Hence, there is a strong need for exercising control actions in a timely fashion to mitigate the TAI phenomena. Anomaly detection is an essential prerequisite to the design of a good controller and such a detector must be able to reliably predict a forthcoming TAI. To detect and predict the onset of a TAI from an ensemble of pressure time series, this paper investigates three data-driven methods: Fast Fourier transform (FFT), symbolic time series analysis (STSA), and hidden Markov modeling (HMM). The main focus of the paper is to make a comparative evaluation of these three anomaly detection methods for classification of the current regime of operation into stable and unstable categories as well as for real-time identification of precursors to impending instabilities with short-length time series of measured variables (e.g., pressure oscillations). Finally, the results, generated on experimental data from a multi-nozzle combustor apparatus, have been compared to evaluate the performance of FFT, STSA, and HMM methods for TAI analysis.
The Inductive Monitoring System (IMS) software developed at Ames Research Center uses artificial intelligence and data mining techniques to build system-monitoring knowledge bases from archived or simulated sensor data. This information is then used to detect unusual or anomalous behavior that may indicate an impending system failure. Currently helping analyze data from systems that help fly and maintain the space shuttle and the International Space Station (ISS), the IMS has also been employed by data classes are then used to build a monitoring knowledge base. In real time, IMS performs monitoring functions: determining and displaying the degree of deviation from nominal performance. IMS trend analyses can detect conditions that may indicate a failure or required system maintenance. The development of IMS was motivated by the difficulty of producing detailed diagnostic models of some system components due to complexity or unavailability of design information. Successful applications have ranged from real-time monitoring of aircraft engine and control systems to anomaly detection in space shuttle and ISS data. IMS was used on shuttle missions STS-121, STS-115, and STS-116 to search the Wing Leading Edge Impact Detection System (WLEIDS) data for signs of possible damaging impacts during launch. It independently verified findings of the WLEIDS Mission Evaluation Room (MER) analysts and indicated additional points of interest that were subsequently investigated by the MER team. In support of the Exploration Systems Mission Directorate, IMS is being deployed as an anomaly detection tool on ISS mission control consoles in the Johnson Space Center Mission Operations Directorate. IMS has been trained to detect faults in the ISS Control Moment Gyroscope (CMG) systems. In laboratory tests, it has already detected several minor anomalies in real-time CMG data. When tested on archived data, IMS was able to detect precursors of the CMG1 failure nearly 15 hours in advance of the actual failure event. In the Aeronautics Research Mission Directorate, IMS successfully performed real-time engine health analysis. IMS was able to detect simulated failures and actual engine anomalies in an F/A-18 aircraft during the course of 25 test flights. IMS is also being used in colla
Accident Precursor Analysis (APA) serves as the bridge between existing risk modeling activities, which are often based on historical or generic failure statistics, and system anomalies, which provide crucial information about the failure mechanisms that are actually operative in the system. APA docs more than simply track experience: it systematically evaluates experience, looking for under-appreciated risks that may warrant changes to design or operational practice. This paper presents the pilot application of the NASA APA process to Space Shuttle Orbiter systems. In this effort, the working sessions conducted at Johnson Space Center (JSC) piloted the APA process developed by Information Systems Laboratories (ISL) over the last two years under the auspices of NASA's Office of Safety & Mission Assurance, with the assistance of the Safety & Mission Assurance (S&MA) Shuttle & Exploration Analysis Branch. This process is built around facilitated working sessions involving diverse system experts. One important aspect of this particular APA process is its focus on understanding the physical mechanism responsible for an operational anomaly, followed by evaluation of the risk significance of the observed anomaly as well as consideration of generalizations of the underlying mechanism to other contexts. Model completeness will probably always be an issue, but this process tries to leverage operating experience to the extent possible in order to address completeness issues before a catastrophe occurs.
Four weather regimes occurring over the Atlantic area and their preferred precursors and successors are identified. The first pattern exhibits the typical European blocking dipole, the second one represents an enhanced zonal flow, the third one consists of a positive anomaly over Greenland, and the last one is characterized by a ridge over the eastern Atlantic Ocean. The backward and forward memory times of the different regimes are established. Preferred precursor and successor patterns are identified as the maxima of the composite probability densities of the anomalies occurring before the onsets and after the breaks. The preferred transitions between regimes are also examined. The systematic aspect of this study gives rise to multiple applications, especially for case and composite studies on model prediction errors.
The recent large earthquakes (M9.0 Tohoku, 03/2011; M7.0 Haiti, 01/2010; M6.7 L Aquila, 04/2008; and M7.9 Wenchuan 05/2008) have renewed interest in pre-anomalous seismic signals associated with them. Recent workshops (DEMETER 2006, 2011 and VESTO 2009 ) have shown that there were precursory atmospheric /ionospheric signals observed in space prior to these events. Our initial results indicate that no single pre-earthquake observation (seismic, magnetic field, electric field, thermal infrared [TIR], or GPS/TEC) can provide a consistent and successful global scale early warning. This is most likely due to complexity and chaotic nature of earthquakes and the limitation in existing ground (temporal/spatial) and global satellite observations. In this study we analyze preseismic temporal and spatial variations (gas/radon counting rate, atmospheric temperature and humidity change, long-wave radiation transitions and ionospheric electron density/plasma variations) which we propose occur before the onset of major earthquakes:. We propose an Integrated Space -- Terrestrial Framework (ISTF), as a different approach for revealing pre-earthquake phenomena in seismically active areas. ISTF is a sensor web of a coordinated observation infrastructure employing multiple sensors that are distributed on one or more platforms; data from satellite sensors (Terra, Aqua, POES, DEMETER and others) and ground observations, e.g., Global Positioning System, Total Electron Content (GPS/TEC). As a theoretical guide we use the Lithosphere-Atmosphere-Ionosphere Coupling (LAIC) model to explain the generation of multiple earthquake precursors. Using our methodology, we evaluated retrospectively the signals preceding the most devastated earthquakes during 2005-2011. We observed a correlation between both atmospheric and ionospheric anomalies preceding most of these earthquakes. The second phase of our validation include systematic retrospective analysis for more than 100 major earthquakes (M>5.9) in Taiwan and Japan. We have found anomalous behavior before all of these events with no false negatives. Calculated false alarm ratio for the for the same month over the entire period of analysis (2003-2009) is less than 10% and was d as the earthquakes. The commonalities in detecting atmospheric/ionospheric anomalies show that they may exist over both land and sea in regions of maximum stress (i.e., along plate boundaries) Our results indicate that the ISTF model could provide a capability to observe pre-earthquake atmospheric/ionospheric signals by combining this information into a common framework.
Assuring safety in the NAS with the inclusion of new entrants, such as Advanced Air Mobility (AAM), will require overcoming unique safety challenges that result from combining innovative technologies with novel airspace concepts for moving people and cargo using autonomous vehicles. The focus of the In-time Aviation Safety Management System (IASMS) is to overcome AAM’s safety assurance challenges. The IASMS Concept of Operations (ConOps) describes an interconnected set of services, functions, and capabilities (SFCs) designed to manage operational risks, identify unknown risks, and inform system designs. This paper describes an approach for defining SFCs based on technology trends in research, assessment of known and unknown risks in voluntary safety reports, and causal and contributing factors in aviation accidents and incidents. This approach would identify potential SFCs that further expand the Monitor, Assess, and Mitigate (M-A-M) functionality that represents the enabling framework of the IASMS. Safety implications that will result from integration of AAM in the transformation of the National Airspace System (NAS) were addressed in National Academies committees reports on AAM and IASMS. Development of a ConOps for IASMS was a top recommendation and can be represented as a reframing of safety assurance that builds on real-time alerting such as the Traffic Alert and Collision Avoidance System, and adds the more encompassing in-time temporal parameter in recognition of the different timelines for collecting and assessing safety data for risk mitigations. For example, mining for safety trends from data sources such as the Aviation Safety Information Analysis and Sharing system occurs over a longer time period. Research on AAM operations poses that SFCs can be designed to monitor the safety margin appropriate for AAM including with regards to the distance between current flight parameters and nominal ideal conditions. These in-time comparisons will become more complex as the density of operations increases at least in certain areas and can include planned and actual 4D trajectory, and in-time comparisons having implications on conflict modeling and prediction including expected and actual departure time, fix/waypoint crossing times, and arrival time. These comparisons would be integrated as part of SFCs that redefine and inform new safety margin. An increased safety margin improves management of operational risks while reducing the potential for anomalies. An increased safety margin also has implications for operator confidence in the certainty of its operations and trust in automation. Technology trends in research could be used to refine existing SFCs and define needs for additional SFCs that provide safety improvements to the design and operation of vehicles, airspace design, and operator performance requirements. NASA is developing innovative approaches to safeguard against major accidents and incidents that have occurred in the NAS and those anticipated with the inclusion of envisioned AAM operations. The innovations use operational performance data to monitor, detect, and predict flight variations exceeding safe nominal patterns, such as would be caused by navigational error, severe weather complications, or hijacking of UAS controls. These innovative approaches have high potential to prevent accidents and incidents in the new AAM era. It is anticipated that elements of the innovations will evolve into SFCs for the IASMS. Voluntary safety reports can be monitored to identify anomalies related to design or operational performance risks. Reports could be periodically monitored and assessed for specific topics. Reports might serve as weak signals or precursors indicative of emergent risk such as when combined with other safety information. The architecture could include SFCs that are based on voluntary safety reports recognizing the periodic temporal nature of data analysis. As previously mentioned, aviation accidents with their causal and contributing precursors can inform the need for SFCs in the IASMS. Accidents and incidents at San Francisco International Airport such as Asiana 214 and Air Canada 759 illustrate how combinations of different factors lead to increased risk. These types of precursors and different factors have implications on the types of SFCs that could be needed to monitor and manage different sources and types of design and operational risk. Continuing to assure the safety of AAM as designs and operations gain in complexity can be accompanied by defining SFCs that also increase in complexity. These SFCs can leverage information from findings and recommendations synthesized across on-going research, voluntary safety reports, and accident and incident reports. These SFCs can serve to refine accuracy of algorithms and resolve limitations with current practices. The IASMS architecture represents the framework for the SFCs and their critical role in safety assurance.
Ongoing research at NASA is driven by a strategic plan defined by the Aeronautics Research Mission Directorate and a vision for future In-Time Aviation Safety Management Systems (IASMS) as described by the National Academies. In both visions, system safety awareness and provision are expanded through increased access to relevant data; integrated analysis and predictive capabilities; improved real-time detection and alerting of domain-specific hazards; decision support, and in some cases, automated risk mitigation strategies. One primary research focus is to develop means by which more timely (i.e., “in-time”) actions may be taken to mitigate precursors, anomalies, or trends that are observed during operations. In this paper, we describe such means as a collection of Services, Functions, and Capabilities (SFCs) that are supported by an underlying information system. For example, an integrated risk assessment capability is envisioned that continuously monitors safety-related metrics and margins and recommends timely operational changes. Assessment functions and/or services can be based on data analytics and predictive models derived from heterogeneous data sets that span relevant indicator metrics and their time histories. Likewise, on-board functions can identify and reduce susceptibility to precursor conditions that have led (and can lead) to aircraft loss-of-control or out-of-control accidents. This paper summarizes development and testing of such an information system tailored to hazards anticipated for future highly autonomous flight missions near and over densely populated areas. Testing is accomplished via simulation and by using small, unmanned aircraft operating over a test range at NASA’s Langley Research Center. Flight plans and test scenarios are defined to emulate several use-cases, including package delivery; reconnaissance; fire management; and urban air taxi vertiport operations. Two test phases are summarized with Phase 1 occurring in (2019-2020) and Phase 2 ongoing (2021-present). Results focus on SFC performance, technology readiness level assessment, and requirements discovery/validation. Companion papers are cited throughout for additional details on the recent testing.
The loss of pilot airplane state awareness (ASA) has been implicated as a factor in several aviation accidents identified by the Commercial Aviation Safety Team (CAST). These accidents were investigated to identify precursors to the loss of ASA and develop technologies to address the loss of ASA. Based on a gap analysis, two technologies were prototyped and assessed with a formative pilot-in-the-loop evaluation in NASA Langley’s full-motion Research Flight Deck. The technologies address: 1) data source anomaly detection in real-time, and 2) intelligent monitoring aids to provide nominal and predictive awareness of situations to be monitored and a mission timeline to visualize events of interest. The evaluation results indicated favorable impressions of both technologies for mitigating the loss of ASA in terms of operational utility, workload, acceptability, complexity, and usability. The team concludes that there is a feasible retrofit solution for improving ASA that would minimize certification risk, integration costs, and training impact.
High pressure can drastically alter chemical bonding and produce exotic compounds that defy conventional wisdom. Especially significant are compounds pertaining to oxygen cycles inside Earth, which hold key to understanding major geological events that impact the environment essential to life on Earth. Here we report the discovery of pressure-stabilized divalent ozonide CaO3 crystal that exhibits intriguing bonding and oxidation states with profound geological implications. Our computational study identifies a crystalline phase of CaO 3 by reaction of CaO and O 2 at high pressure and high temperature conditions; ensuing experiments synthesize this rare compound under compression in a diamond anvil cell with laser heating. High-pressure x-ray diffraction data show that CaO 3 crystal forms at 35 GPa and persists down to 20 GPa on decompression. Analysis of charge states reveals a formal oxidation state of -2 for ozone anions in CaO 3 . These findings unravel the ozonide chemistry at high pressure and offer insights for elucidating prominent seismic anomalies and oxygen cycles in Earth’s interior. We further predict multiple reactions producing CaO 3 by geologically abundant mineral precursors at various depths in Earth’s mantle.
Isotopic measurements have been made on organic sulfur and phosphorus compounds recently discovered in the Murchison meteorite. Carbon, hydrogen and sulfur measurements were performed on individual members of the organic sulfur compounds, alkyl sulfonates; and carbon and hydrogen measurements were made on bulk alkyl phosphonates. Cooper and Chang reported the first carbon isotopic measurements of Murchison organic sulfonates, providing insight into the potential synthetic mechanisms of these and, possibly, other organic species. Hydrogen isotopic measurements of the sulforiates now reveal deuterium excesses ranging from +660 to +2730 %. The deuterium enrichments indicate formation of the hydrocarbon portion of these compounds in a low temperature astrophysical environment consistent with that of dense molecular clouds. Measurements of the sulfur isotopes provide further constraints on the origin and mechanism of formation of these organic molecules. Recently, there has been growing documentation of sulfur isotopic anomalies in meteoritic material. Thiemens and Jackson have shown that some bulk ureilites possess excess S-33 and Thiemens et al. have reported excess S-33 in an oldhamite separate from the Norton County meteorite. Rees and Thode reported a large S-33 excess in an Allende acid residue, however, attempts to verify this measurements have been unsuccessful, possibly due to the heterogeneous nature of the carrier phase. With the recognition that sulfur isotopes may reflect chemistry in the protosolar nebula or the precursor molecular cloud, identification of potential carriers is of considerable interest. In the present study, the stable isotopes of sulfur were measured in methane sulfonic acid extracted from the Murchison meteorite. The isotopic composition was found to be: (delta)S-33 = 2.48 %, (delta)S-34 = 2.49 % and (delta)S-36 = 6.76 %. Based upon analysis of more than 60 meteoritic and numerous terrestrial samples, the mass fractionation lines are defined by Delta-33 = (delta)S-33 -0.50(delta)S-34 and Delta-36 = (delta)s-36 - 1.97 (delta)S-34. From these relationships Delta-33 = 1.24 % and Delta-36 = 0.89 % are observed. These anomalies, particularly the Delta-33, lie well outside the range of analytical uncertainty. They are the largest observed in any meteoritic component and the first found in an organosulfur compound. As discussed by Thiemens and Jackson, due to it's position on the periodic chart, sulfur undergoes chemically induced mass independent isotopic fractionations as does oxygen. Experiments by Mauersberger et. al. show that in such processes, the magnitude of fractionation for the different isotopically substituted species varies with mass and angular momentum; thus, anomalies are expected for both S-33 and S-36, but not necessarily of the same magnitude. Laboratory experiments have also confirmed that chemically produced, mass independent fractionations are mediated by molecular symmetry factors. A chemical source of fractionation requires that the sulfur isotopic anomaly was established in the gas phase, probably in reactions involving symmetric CS2. The discovery of an anomalous sulfur isotopic composition in a specific molecule containing excess deuterium is an important advance in the understanding of the cosmochemistry of sulfur. This evidence suggests that methanesulfonic acid was synthesized by interstellar processes. Further measurements and details of possible synthesis and fractionation mechanisms will be presented.
The 1997-99 ENSO (El nino Southern Oscillation) cycle was very powerful, but also well observed. The best satellite rainfall estimates combined with gauge observations allow for a global analysis of precipitation anomalies accompanying the 1997-98 El Nino and initiation of the 1998-99 La Nina. For the period April 1997 to March 1998 the central to eastern Pacific, southeastern and western U.S., Argentina, eastern Africa, South China, eastern Russia, and North Atlantic were all more than two standard deviations wetter than normal. During the same year the Maritime Continent, eastern Indian Ocean, subtropical North Pacific, northeastern South America, and much of the mid- latitude southern oceans were more than two standard deviations drier than normal. An analysis of the evolution of the El Nino and accompanying precipitation anomalies revealed that a dry Maritime Continent led the formation of the El Nino SST (Sea Surface Temperature), while in the central Pacific, precipitation anomalies lagged the El Nino SST by a season. A rapid transition from El Nino to La Nina occurred in May 1998, but as early as October-November 1997 precipitation indices captured precursor changes in Pacific rainfall anomalies. Differences were found between observed and modeled [NCEP/NCAR (National Centers for Environmental Prediction/National Center for Atmospheric Research) reanalysis] precipitation anomalies for 1997 and 98. In particular, the model had a bias towards positive precipitation anomalies and the magnitudes of the anomalies in the equatorial Pacific were small compared to the observations. Also, the evolution of the precipitation field, including the drying of the Maritime Continent and eastward progression of rainfall in the equatorial Pacific, was less pronounced for the model compared to the observations. One degree daily estimates of rainfall show clearly the MaddenJulian Oscillation and related westerly wind burst events over the Maritime Continent, which are key indicators for the onset of El Nino.