Search

Emerging parametric insurance products targeted at regional governments consider an index of flooding as the instrument for payoff and rate setting. Inundation extent from satellite remote sensing may provide a more direct measure of flood risk in this context than hydraulic modeling of flow and inundation. Here, we examine satellite‐based fractional inundated area as a proxy for flood impact that can be used for index insurance payment at a regional scale. Typical methods for estimating return periods from unbounded distributions such as the Generalized Extreme Value distribution are not appropriate for fractional flooded area, which is bounded by 0 and 1. Here we examine alternative bounded distributions (2 parameter and a 4 parameter Beta) to estimate return periods and quantify uncertainty using a bootstrap sampling procedure for the short duration satellite record of fractional flooded area. We consider two examples with distinct flood dynamics (a) a country (Bangladesh) where a flood can cover the majority of the land surface, and (b) a river basin (the Rio Salado basin in Argentina) where the worst flood covered only a modest fraction of the watershed. We explore how a parametric insurance policy based on fractional flooded area could be priced based on a typical approach used in the industry, that accounts for uncertainty for small sample estimation. Our exploratory approach to model selection illustrates how estimating the uncertainty price influences insurance contract pricing and is important to consider the choice of distribution beyond just the traditional measures of goodness of fit. Plain Language Summary: Index insurance, catastrophe bonds, and other types of risk transfer instruments could play an important role in adapting to floods and ensuring sustainable development in a world of increasing flood risk. In this article, we examine how satellite time series of inundation can be used to develop an emerging type of flood insurance, known as parametric or index‐based insurance. Unlike traditional indemnity insurance, which relies on adjusters to estimate loss for individual damage, index insurance uses data ex‐ante to determine payout contracts when pre‐specified thresholds are crossed. Inundation extent from satellite remote sensing may provide a more direct measure of flood risk than data from models or stream gauges. However, typical methods used to estimate return periods for floods from models and gauges are not appropriate for fractional inundated area measurements from satellites. Here we provide a more appropriate method to estimate return periods and quantify uncertainty to price an insurance product leveraging the relatively short satellite record. Example applications for Bangladesh and Rio Salado, Argentina are provided. We show why estimating and pricing uncertainty ultimately influences insurance contract pricing and can help governments select insurance policies that align with their flood adaptation strategy. Key Points: Estimating return periods for fractional inundated area with the oft‐used Generalized Extreme Value distribution is inappropriateBounded distributions (e.g., Beta) reduce uncertainty estimates for probability of exceedance from short duration inundation time seriesExample design and price of parametric flood insurance with fractional inundated area in Bangladesh and Argentina accounting for uncertainty [ABSTRACT FROM AUTHOR]

The lower Brahmaputra River in Bangladesh and Northeast India often floods during the monsoon season, with catastrophic consequences for people throughout the region. While most climate models predict an intensified monsoon and increase in flood risk with warming, robust baseline estimates of natural climate variability in the basin are limited by the short observational record. Here we use a new seven-century (1309–2004 C.E) tree-ring reconstruction of monsoon season Brahmaputra discharge to demonstrate that the early instrumental period (1956–1986 C.E.) ranks amongst the driest of the past seven centuries (13th percentile). Further, flood hazard inferred from the recurrence frequency of high discharge years is severely underestimated by 24–38% in the instrumental record compared to previous centuries and climate model projections. A focus on only recent observations will therefore be insufficient to accurately characterise flood hazard risk in the region, both in the context of natural variability and climate change. This study investigates flood hazards of the Brahmaputra River, Bangladesh. Based on a tree ring reconstruction of seasonal river discharge, climate modelling, and historic documentation of flood events, the authors suggest flood hazard risk is underestimated by ~24–38% in the present day compared to the past 700 years. [ABSTRACT FROM AUTHOR]

The Ganges-Brahmaputra-Meghna river system is the third largest in the world with an annual discharge of approximately 40,000m3/s. The Brahmaputra alone contributes nearly half of the total discharge. Climatic controls on streamflow in the lower Brahmaputra basin are complex and vary seasonally, with glacial and snow melt from the Himalaya and Tibetan Plateau dominating in the fall, winter and spring, and Indian monsoon precipitation acting as the major contributor to summer flow. Here we present a suite of four seasonal and one annual reconstruction of past discharge at the Bahadurabad gauging station in north-eastern Bangladesh based on a careful evaluation of streamflow clustering across different months and streamflow and tree-ring predictor relationships. The spring transition season (May-June), monsoon season (July-September), post-monsoon season (October-December), and annual flow (January-December) reconstructions extend back six centuries from ~1400 to 2014 C.E. The dry winter season reconstruction (January-April) is millennial length and spans ~300 to 2014 C.E. These resulting reconstructions have little predictor overlap allowing us compare low frequency variability and recent trends if any across different seasons and the entire year. The short instrumental data (1956-2014 C.E.) shows increasing dry season flow and decreasing wet season flow, particularly since the 1990s. While these changes are likely associated with increased glacial melt and decreasing monsoon precipitation respectively, they lie within reconstructed paleo-discharge variability estimates. The deltaic flood plains of Bangladesh are extremely vulnerable to catastrophic seasonal floods caused by heavy summer monsoon precipitation. Major flooding events occurred in 1951, 1966, 1987, 1988, and 1998. Our reconstruction indicates unusually high summer discharge in all of these years. The single year highest discharge value over the length of the six-century long reconstruction occurred in 1998 suggesting that this event was highly unusual in the long-term context. Using our reconstructions, we find on average five single-year high discharge events related to flooding occur every century, but these events are more likely to occur during periods of decadal high flow. This suggests a potential to use our multi-centennial reconstruction for flood-hazard risk estimation. [ABSTRACT FROM AUTHOR]