Votre recherche

Abstract Variability in tropical cyclone activity in the eastern Pacific basin has been linked to a wide range of climate factors, yet the dominant factors driving this variability have yet to be identified. Using Poisson regressions and a track clustering method, the authors analyze and compare the climate influence on cyclone activity in this region. The authors show that local sea surface temperature and upper-ocean heat content as well as large-scale conditions in the northern Atlantic are the dominant influence in modulating eastern North Pacific tropical cyclone activity. The results also support previous findings suggesting that the influence of the Atlantic Ocean occurs through changes in dynamical conditions over the eastern Pacific. Using model selection algorithms, the authors then proceed to construct a statistical model of eastern Pacific tropical cyclone activity. The various model selection techniques used agree in selecting one predictor from the Atlantic (northern North Atlantic sea surface temperature) and one predictor from the Pacific (relative sea surface temperature) to represent the best possible model. Finally, we show that this simple model could have predicted the anomalously high level of activity observed in 2014.

Abstract We analyze, using Poisson regressions, the main climate influences on North Atlantic tropical cyclone activity. The analysis is performed using not only various time series of basin‐wide storm counts but also various series of regional clusters, taking into account shortcomings of the hurricane database through estimates of missing storms. The analysis confirms that tropical cyclones forming in different regions of the Atlantic are susceptible to different climate influences. We also investigate the presence of trends in these various time series, both at the basin‐wide and cluster levels, and show that, even after accounting for possible missing storms, there remains an upward trend in the eastern part of the basin and a downward trend in the western part. Using model selection algorithms, we show that the best model of Atlantic tropical cyclone activity for the recent past is constructed using Atlantic sea surface temperature and upper tropospheric temperature, while for the 1878–2015 period, the chosen covariates are Atlantic sea surface temperature and El Niño–Southern Oscillation. We also note that the presence of these artificial trends can impact the selection of the best covariates. If the underlying series shows an upward trend, then the mean Atlantic sea surface temperature captures both interannual variability and the upward trend, artificial or not. The relative sea surface temperature is chosen instead for stationary counts. Finally, we show that the predictive capability of the statistical models investigated is low for U.S. landfalling hurricanes but can be considerably improved when forecasting combinations of clusters whose hurricanes are most likely to make landfall. , Key Points Estimates of missing storms are not sufficient to account for the increase in hurricane activity in the eastern tropical Atlantic Recent upward trends, artificial or not, affect the selection of key determinants of tropical cyclone activity, especially the SST variable Despite previous results to that effect, the May–June NAO does not provide predictive skill for Atlantic landfalling hurricanes

Abstract We develop a portfolio credit risk model that includes firm‐specific Markov‐switching regimes as well as individual stochastic and endogenous recovery rates. Using weekly credit default swap premiums for 35 financial firms, we analyze the credit risk of each of these companies and their statistical linkages, putting emphasis on the 2005–2012 period. Moreover, we study the systemic risk affecting both the banking and insurance subsectors.

Abstract Large scale flood risk analyses are fundamental to many applications requiring national or international overviews of flood risk. While large‐scale climate patterns such as teleconnections and climate change become important at this scale, it remains a challenge to represent the local hydrological cycle over various watersheds in a manner that is physically consistent with climate. As a result, global models tend to suffer from a lack of available scenarios and flexibility that are key for planners, relief organizations, regulators, and the financial services industry to analyze the socioeconomic, demographic, and climatic factors affecting exposure. Here we introduce a data‐driven, global, fast, flexible, and climate‐consistent flood risk modeling framework for applications that do not necessarily require high‐resolution flood mapping. We use statistical and machine learning methods to examine the relationship between historical flood occurrence and impact from the Dartmouth Flood Observatory (1985–2017), and climatic, watershed, and socioeconomic factors for 4,734 HydroSHEDS watersheds globally. Using bias‐corrected output from the NCAR CESM Large Ensemble (1980–2020), and the fitted statistical relationships, we simulate 1 million years of events worldwide along with the population displaced in each event. We discuss potential applications of the model and present global flood hazard and risk maps. The main value of this global flood model lies in its ability to quickly simulate realistic flood events at a resolution that is useful for large‐scale socioeconomic and financial planning, yet we expect it to be useful to climate and natural hazard scientists who are interested in socioeconomic impacts of climate. , Plain Language Summary Flood is among the deadliest and most damaging natural disasters. To protect against large scale flood risk, stakeholders need to understand how floods can occur and their potential impacts. Stakeholders rely on global flood models to provide them with plausible flood scenarios around the world. For a flood model to operate at the global scale, climate effects must be represented in addition to hydrological ones to demonstrate how rivers can overflow throughout the world each year. Global flood models often lack the flexibility and variety of scenarios required by many stakeholders because they are computationally demanding. Designed for applications where detailed local flood impacts are not required, we introduce a rapid and flexible global flood model that can generate hundreds of thousands of scenarios everywhere in the world in a matter of minutes. The model is based on a historical flood database from 1985 to 2017 that is represented using an algorithm that learns from the data. With this model, the output from a global climate model is used to simulate a large sample of floods for risk analyses that are coherent with global climate. Maps of the annual average number of floods and number of displaced people illustrate the models results. , Key Points We present a global flood model built using machine learning methods fitted with historical flood occurrences and impacts Forced with a climate model, the global flood model is fast, flexible and consistent with global climate We provide global flood hazard (occurrence) and risk (population displaced) maps over 4,734 watersheds

Abstract. Canada's RADARSAT missions improve the potential to study past flood events; however, existing tools to derive flood depths from this remote-sensing data do not correct for errors, leading to poor estimates. To provide more accurate gridded depth estimates of historical flooding, a new tool is proposed that integrates Height Above Nearest Drainage and Cost Allocation algorithms. This tool is tested against two trusted, hydraulically derived, gridded depths of recent floods in Canada. This validation shows the proposed tool outperforms existing tools and can provide more accurate estimates from minimal data without the need for complex physics-based models or expert judgement. With improvements in remote-sensing data, the tool proposed here can provide flood researchers and emergency managers accurate depths in near-real time.

Although the finance literature has devoted a lot of research into the development of advanced models for improving the pricing and hedging performance, there has been much less emphasis on approaches to measure dynamic hedging effectiveness. This article discusses a statistical framework based on regression analysis to measure the effectiveness of dynamic hedges for long-term investment guarantees. The importance of taking model risk into account is emphasized. The difficulties in reducing hedging risk to an appropriately low level lead us to propose a new perspective on hedging, and recognize it as a tool to modify the risk–reward relationship of the unhedged position.

Abstract El Niño‐Southern Oscillation (ENSO) is often considered as a source of long‐term predictability for extreme events via its teleconnection patterns. However, given that its characteristic cycle varies from two to 7 years, it is difficult to obtain statistically significant conclusions based on observational periods spanning only a few decades. To overcome this, we apply the global flood risk modeling framework developed by Carozza and Boudreault to an equivalent of 1,600 years of bias‐corrected General Circulation Model outputs. The results show substantial anomalies in flood occurrences and impacts for El Niño and La Niña when compared to the all‐year baseline. We were able to obtain a larger global coverage of statistically significant results than previous studies limited to observational data. Asymmetries in anomalies for both ENSO phases show a larger global influence of El Niño than La Niña on flood hazard and risk. , Plain Language Summary El Niño‐Southern Oscillation (ENSO) is one of the most important global climate phenomena. It is well‐known to affect precipitation and temperature in many areas of the world. It is therefore very important for researchers (environmental and climate sciences, economics, etc.), risk managers, decision‐ and policy‐makers to understand the influence of ENSO on flooding. Previous studies analyzed the link between ENSO and flooding but because they were based upon 40 years of data, a lot of uncertainties remained as to how ENSO has any significance on flooding. In this study, we use outputs from a climate model large ensemble that provides 1,600 years of simulated data to determine the impacts of ENSO on flooding. But because it is very difficult to run traditional flood models on 1,600 years of data, we rather leverage a machine learning approach to accelerate computations in a context where the focus is on socioeconomic impacts. We find that ENSO is a significant driver of flooding in more regions than what was previously found. Finally, there appears to be a greater global influence of El Niño than La Niña on flooding. , Key Points We simulated an equivalent of 1,600 years of realistic flood events globally using a statistical model forced with climate model outputs We found a statistically significant ( α  = 0.05) influence of El Niño‐Southern Oscillation (ENSO) over 55% of land area for flood occurrence and over 69% for flood impact Asymmetries in anomalies for both ENSO phases show a larger global influence of El Niño than La Niña on flood hazard and risk

Abstract This study presents a firm‐specific methodology for extracting implied default intensities and recovery rates jointly from unit recovery claim prices—backed by out‐of‐the‐money put options—and credit default swap premiums, therefore providing time‐varying and market‐consistent views of credit risk at the individual level. We apply the procedure to about 400 firms spanning different sectors of the US economy between 2003 and 2019. The main determinants of default intensities and recovery rates are analyzed with statistical and machine learning methods linking default risk and credit losses to market, sector, and individual variables. Consistent with the literature, we find that individual volatility, leverage, and corporate bond market determinants are key factors explaining the implied default intensities and recovery rates. Then, we apply the framework in the context of credit risk management in applications, like, market‐consistent credit value‐at‐risk calculation and stress testing.