Explore the vast range of titles available.

Social vulnerability indicators seek to identify populations susceptible to hazards based on aggregated sociodemographic data. Vulnerability indices are rarely validated with disaster outcome data at broad spatial scales, making it difficult to develop effective national scale strategies to mitigate loss for vulnerable populations. This paper validates social vulnerability indicators using two flood outcomes: death and damage. Regression models identify sociodemographic factors associated with variation in outcomes from 11,629 non-coastal flood events in the USA (2008–2012), controlling for flood intensity using stream gauge data. We compare models with (i) socioeconomic variables, (ii) the composite social vulnerability index (SoVI), and (iii) flood intensity variables only. The SoVI explains a larger portion of the variance in death (AIC = 2829) and damage (R2 = 0.125) than flood intensity alone (death—AIC = 2894; damage—R2 = 0.089), and models with individual sociodemographic factors perform best (death—AIC = 2696; damage—R2 = 0.229). Socioeconomic variables correlated with death (rural counties with a high proportion of elderly and young) differ from those related to property damage (rural counties with high percentage of Black, Hispanic and Native American populations below the poverty line). Results confirm that social vulnerability influences death and damage from floods in the USA. Model results indicate that social vulnerability models related to specific hazards and outcomes perform better than generic social vulnerability indices (e.g., SoVI) in predicting non-coastal flood death and damage. Hazard- and outcome-specific indices could be used to better direct efforts to ameliorate flood death and damage towards the people and places that need it most. Future validation studies should examine other flood outcomes, such as evacuation, migration and health, across scales.

Infrastructure development is central to the processes that abate and produce vulnerabilities in cities. Urban actors, especially those with power and authority, perceive and interpret vulnerability and decide when and how to adapt. When city managers use infrastructure to reduce urban risk in the complex, interconnected city system, new fragilities are introduced because of inherent system feedbacks. We trace the interactions between system dynamics and decision-making processes over 700 years of Mexico City’s adaptations to water risks, focusing on the decision cycles of public infrastructure providers (in this case, government authorities). We bring together two lenses in examining this history: robustness-vulnerability trade-offs to explain the evolution of systemic risk dynamics mediated by feedback control, and adaptation pathways to focus on the evolution of decision cycles that motivate significant infrastructure investments. Drawing from historical accounts, archeological evidence, and original research on water, engineering, and cultural history, we examine adaptation pathways of humans settlement, water supply, and flood risk. Mexico City’s history reveals insights that expand the theory of coupled infrastructure and lessons salient to contemporary urban risk management: (1) adapting by spatially externalizing risks can backfire: as cities expand, such risks become endogenous; (2) over time, adaptation pathways initiated to address specific risks may begin to intersect, creating complex trade-offs in risk management; and (3) city authorities are agents of risk production: even in the face of new exogenous risks (climate change), acknowledging and managing risks produced endogenously may prove more adaptive. History demonstrates that the very best solutions today may present critical challenges for tomorrow, and that collectively people have far more agency in and influence over the complex systems we live in than is often acknowledged.

Rapid and accurate maps of floods across large domains, with high temporal resolution capturing event peaks, have applications for flood forecasting and resilience, damage assessment, and parametric insurance. Satellite imagery produces incomplete observations spatially and temporally, and hydrodynamic models require tradeoffs between computational efficiency and accuracy. We address these challenges with a novel flood model which predicts surface water area from the U.S. National Water Model using a convolutional neural network (NWM‐CNN). We trained NWM‐CNN on 780 flood events, at a 250 m resolution with an RMSE of 4.58% on held out validation geographies. We demonstrate NWM‐CNN across California during the 2023 atmospheric rivers, comparing predictions against Sentinel‐1 mapped flood observations. We compared historical predictions from 1979 to 2023 to flood damage reports in Sacramento County, California. Results show that NWM‐CNN captures inundation extent better than the Height Above Nearest Drainage (HAND) approach (25%–36% RMSE, respectively). Plain Language Summary We use machine learning (ML) to map floods quickly and accurately over large areas, which can help with predicting flooded extent, understanding impact, and aiding flood insurance and response. On their own, satellite images don't catch everything because they may be obscured or unavailable at the peak of a flood. Computer models that predict floods require a trade‐off between speed, accuracy, and resolution. Our solution uses ML to learn from the U.S. National Water Model and satellite images from past floods to predict how much of an area will be covered in water. We demonstrate this on floods in California in 2023 caused by atmospheric rivers, and we looked back at floods in Sacramento County from 1979 to 2023. We compared our method to another commonly used model and found ours was more accurate, making it a promising tool for future flood mapping and response planning. Key Points Convolution neural networks (CNN) are suitable for rapid modeling of surface water dynamics for large‐scale inundation mapping We deploy a CNN for continuous flood mapping across all of California during the devastating 2023 atmospheric river (AR) events Inundation extent across Sacramento is more accurately predicted with CNN than the Height Above Nearest Drainage (HAND)

A growing body of evidence suggests that criminal activities associated with drug trafficking networks are a progressively important driver of forest loss in Central America. However, the scale at which drug trafficking represents a driver of forest loss is not presently known. We estimated the degree to which narcotics trafficking may contribute to forest loss using an unsupervised spatial clustering of 15 spatial and temporal forest loss patch metrics developed from global forest change data. We distinguished anomalous forest loss from background loss patches for each country exhibiting potential 'narco-capitalized' signatures which showed a statistically significant dissimilarity from other patches in terms of size, timing, and rate of forest loss. We also compared annual anomalous forest loss with the number of cocaine shipments and volume of cocaine seized, lost, or delivered at country- and department-level. For Honduras, results from linear mixed effects models showed a highly significant relationship between anomalous forest loss and the timing of increased drug trafficking (F = 9.90, p = 0.009) that also differed significantly from temporal patterns of background forest loss (t-ratio = 2.98, p = 0.004). Other locations of high forest loss in Central America showed mixed results. The timing of increased trafficking was not significantly related to anomalous forest loss in Guatemala and Nicaragua, but significantly differed in patch size compared to background losses. We estimated that cocaine trafficking could account for between 15% and 30% of annual national forest loss in these three countries over the past decade, and 30% to 60% of loss occurred within nationally and internationally designated protected areas. Cocaine trafficking is likely to have severe and lasting consequences in terms of maintaining moist tropical forest cover in Central America. Addressing forest loss in these and other tropical locations will require a stronger linkage between national and international drug interdiction and conservation policies.

Global flood models (GFMs) and earth observation (EO) play a crucial role in characterising flooding, especially in data-sparse, under-resourced regions of the world. However, validation studies are often limited to a handful of historic events and do not directly assess the ability of these products to simulate flood hazard-the probability that flooding will occur in a given location. As a result, it is difficult for stakeholders to decipher the ability of either models or observations to identify flood hazard and make decisions to mitigate for flooding. Here, we leverage flood observations from 20 years of MODIS data to compare the recorded flooding with what would be expected given the hazard simulated by a GFM. We devise an approach, Flood Expectation Per Pixel, and apply it across four large basins in Africa-Congo, Niger, Nile and Volta representing a variety of biomes. We estimate the uncertainty of EO to capture flood events due to burned areas, cloud cover and vegetation, incorporating uncertainty estimates when comparing to modelled hazard. We found that at lower return periods (RPs) (<20 years), the EO data records less flooding than the GFM, suggesting GFMs overpredict frequent flooding. For RPs between 50 and 100 years, GFM and EO data show greater consistency given the uncertainties we consider. For large RPs (100 years) the EO observations show more flooding than expected given the GFM data, potentially due to data errors and non-fluvial flooding, however there are too few observations to draw significant conclusions at these RPs. The EO record indicates that the GFM can differentiate between flood RPs. We find EO and GFM complement each other and thus should be used in tandem to inform strategies to mitigate floods across the hazard spectrum from frequent to extreme flood events.

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 inappropriate Bounded distributions (e.g., Beta) reduce uncertainty estimates for probability of exceedance from short duration inundation time series Example design and price of parametric flood insurance with fractional inundated area in Bangladesh and Argentina accounting for uncertainty

We assess how much of Central America is likely to be agriculturally suitable for cultivating coca (Erythroxylum spp), the main ingredient in cocaine. Since 2017, organized criminal groups (not smallholders) have been establishing coca plantations in Central America for cocaine production. This has broken South America’s long monopoly on coca leaf production for the global cocaine trade and raised concerns about future expansion in the isthmus. Yet it is not clear how much of Central America has suitable biophysical characteristics for a crop domesticated in, and long associated with the Andean region. We combine geo-located data from coca cultivation locations in Colombia with reported coca sites in Central America to model the soil, climate, and topography of Central American landscapes that might be suitable for coca production under standard management practices. We find that 47% of northern Central America (Honduras, Guatemala, and Belize) has biophysical characteristics that appear highly suitable for coca-growing, while most of southern Central America does not. Biophysical factors, then, are unlikely to constrain coca’s spread in northern Central America. Whether or not the crop is more widely planted will depend on complex and multi-scalar social, economic, and political factors. Among them is whether Central American countries and their allies will continue to prioritize militarized approaches to the drug trade through coca eradication and drug interdiction, which are likely to induce further expansion, not contain it. Novel approaches to the drug trade will be required to avert this outcome.

Governments, development banks, corporations, and nonprofits are increasingly considering the potential contribution of watershed conservation activities to secure clean water for cities and to reduce flood risk. These organizations, however, often lack decision-relevant, initial screening information across multiple cities to identify which specific city-watershed combinations present not only water-related risks but also potentially attractive opportunities for mitigation via natural infrastructure approaches. To address this need, this paper presents a novel methodology for a continental assessment of the potential for watershed conservation activities to improve surface drinking water quality and mitigate riverine and stormwater flood risks in 70 major cities across Latin America. We used publicly available geospatial data to analyze 887 associated watersheds. Water quality metrics assessed the potential for agricultural practices, afforestation, riparian buffers, and forest conservation to mitigate sediment and phosphorus loads. Flood reduction metrics analyzed the role of increasing infiltration, restoring riparian wetlands, and reducing connected impervious surface to mitigate riverine and stormwater floods for exposed urban populations. Cities were then categorized based on relative opportunity potential to reduce identified risks through watershed conservation activities. We find high opportunities for watershed activities to mitigate at least one of the risks in 42 cities, potentially benefiting 96 million people or around 60% of the urbanites living in the 70 largest cities in Latin America. We estimate water quality could be improved for 72 million people in 27 cities, riverine flood risk mitigated for 5 million people in 13 cities, and stormwater flooding mitigated for 44 million people in 14 cities. We identified five cities with the potential to simultaneously enhance water quality and mitigate flood risks, and in contrast, six cities where conservation efforts are unlikely to meaningfully mitigate either risk. Institutions investing in natural infrastructure to improve water security in Latin America can maximize their impact by focusing on specific watershed conservation activities either for cleaner drinking water or flood mitigation in cities identified in our analysis where these interventions are most likely to reduce risk.

An analysis of floods or droughts that hit the same place twice shows that using risk management alone does not reduce the effect of extreme events. Addressing the social drivers of hazard impact, equitably, is essential. Repeated disasters reveal exposure inequity.