Climate change, riverine flood risk and adaptation for the conterminous United States

Abstract

Riverine floods are among the most costly natural disasters in the United States, and floods are generally projected to increase in frequency and magnitude with climate change. Faced with these increasing risks, improved information is needed to direct limited resources toward the most cost-effective adaptation actions available. Here we leverage a newly available flood risk dataset for residential properties in the conterminous United States to calculate expected annual damages to residential structures from inland/riverine flooding at a property-level; the cost of property-level adaptations to protect against future flood risk; and the benefits of those adaptation investments assuming both static and changing climate conditions. Our modeling projects that in the absence of adaptation, nationwide damages from riverine flooding will increase by 20–30% under high levels of warming. Floodproofing, elevation and property acquisition can each be cost-effective adaptations in certain situations, depending on the desired return on investment (i.e., benefit cost ratio), the discount rate, and the assumed rate of climate change. Incorporation of climate change into the benefit-cost calculation increases the number of properties meeting any specified benefit-cost threshold, as today’s investments protect against an increasing frequency of future floods. However, because future expected damages are discounted relative to present-day, the adaptation decisions made based on a static climate assumption are very similar to the decisions made when climate change is considered. If the goal is to optimize adaptation decision making, a focus on quantifying present-day flood risk is therefore at least as important as understanding how those risks might change under a warming climate.

New Empirical Models for Flood Loss Prediction and Implications for the Coterminous United States

“…Other important future research directions for improving the accuracy and usefulness of loss estimation include 1) more transparency on the damage functions, probability of damage occurring at various inundation depths and structure characteristics employed and how they are linked to each other in loss estimation; and 2) assessing outcomes with respect to different validated damage functions and incorporating techniques to isolate the role of broadly influential loss variables such as catastrophic events. High-resolution hazard models made available over large spatial domains have heralded a plethora of large-scale estimates of flood risk to inform efficient and equitable flood risk management policy 23,38,44 . However, our findings and caveats highlighted above raise questions about the suitability of extrapolation results that don't consider a large range of uncertainties.…”

“…Thus, for different structures, identical hazard exposure can generate markedly different fractional damage. A key problem in practice is that 𝐿 𝑖 𝑇𝑟𝑢𝑒 is unobserved and 𝐷 𝑇𝑟𝑢𝑒 can't be derived but estimates of 𝐿 𝑖 𝑇𝑟𝑢𝑒 are needed for a wide variety of flood risk management applications 3,5,8,23,38,43,44,58 .…”

“…Without conditional probabilities of damage occurring for different flood hazard exposures, extrapolation may overestimate the cases in which damage occurs 20 . We assess our extrapolation results in light of this concern, but we apply damage functions to all properties with non-zero flood hazard exposure like other papers that have estimated flood losses over large scales 23,38,44,56 . To explore the sensitivity of aggregate loss estimates to different first-floor elevation adjustments and scenarios where damages occur at low depths, we develop three sets of assumptions that reflect low, benchmark (reference), and highdamage assumptions (see Methods).…”

“…In practice, no damage relationship satisfies all four criteria above. In the U.S., depth-damage functions (DDFs) are widely used for predicting flood damages 5,[32][33][34][35][36][37][38][39][40][41][42][43][44] , for policy analysis 6,7,45 , and to inform insurance rates and flood mitigation expenditures 1,7,46 . Current DDFs used in the U.S. were created largely from expert judgement and relate the characteristics of structures, often the number of stories and presence of a basement, and inundation depth.…”

Cartografía de riesgo a inundaciones como herramienta de apoyo a la planificación territorial. Aplicación a la ciudad de Luján (Argentina)

“…Maantay y Maroko (2009) presentan una metodología basada en la utilización de datos catastrales (áreas y unidades residenciales) y datos cartográficos del área de inundación puestos a disposición por la Agencia Federal para el Manejo de Emergencias (FEMA, por si siglas en inglés) en la ciudad de Nueva York, que permite estimar cuantitativamente quiénes son y dónde están las poblaciones potencialmente en riesgo, clasificadas étnicamente, lo que permitiría a los decisores trabajar directamente con las comunidades locales afectadas y, de esta forma, colaborar para que los esfuerzos de planificación, mitigación y recuperación de desastres tengan éxito. Wobus et al (2021) avanzan en el análisis de riesgo a inundaciones en Estados Unidos a partir de un modelo de inundación, utilizando los mapas de inundación de FEMA, tanto de amenaza fluvial como pluvial, la base de datos de diques nacionales producida por el Cuerpo de Ingenieros del Ejército, un modelo de elevación digital de 30 metros de resolución, que obtienen del Servicio Geológico de Estados Unidos, un modelo para representar las defensas contra inundaciones conocidas y las distribuciones actuales de personas y activos, que están detalladas en el mapa de densidad de población de la Agencia de Protección Ambiental (EPA, por sus siglas en inglés) y en el Inventario Nacional de Estructuras de FEMA. Además, incorporan proyecciones de la EPA que exploran los cambios futuros en la población humana, la densidad de viviendas y la superficie impermeable del país.…”