Item Details

title

Flood Occurrence and Intensity Prediction in North America

author

Yu, Kejia

abstract

In this paper, we fit models to the occurrence and intensity of flood in North America to study risk for insurance purposes. North America is separated into 625 hydro basins with unique hydro graphic information. We collected basin features and climate data as predictors for flood occurrence and intensity. We chose both property losses and inundated areas as response variables for a flood. Our first goal is a predictive model for flood occurrence and intensity, and our second goal is to use this model with future regional climate model data to explore projected changes in flood risk due to climate change. The models we use are logistic regression and random forests. Logistic regression gave a better result and is chosen as the final model. We used a bias-corrected regional climate model to project future flood risk under a climate change scenario.

subject

hydro basin

property losses

quantile match

simulated climate data

contributor

Erhardt, Robert (committee chair)

Erhardt, Robert (committee member)

Jiang, Miaohua (committee member)

Hepler, Staci (committee member)

date

2020-05-29T08:36:03Z (accessioned)

2020-05-29T08:36:03Z (available)

2020 (issued)

degree

Mathematics and Statistics (discipline)

identifier

http://hdl.handle.net/10339/96829 (uri)

language

en (iso)

publisher

Wake Forest University

type

Thesis