How much building-level detail can public flood insurance claims support? Evidence from Eastern Nebraska

Public flood insurance records omit addresses, preventing direct links between insured losses and individual buildings and limiting their use in local assessments of flood losses. Thus, whether building and terrain attributes can support physically plausible allocations remains uncertain. This study evaluates allocation using these attributes and the geographic transferability of severity models for 423 National Flood Insurance Program claims from the March 2019 flood in Dodge and Douglas Counties, Nebraska. Matched claims were fractionally allocated among candidate residential buildings within cells defined by ZIP Code Tabulation Area and flood zone in each county. Rankings were assessed against an inherited inundation classification using area under the receiver operating characteristic curve (AUC) within cells, spatial block bootstrapping, and sensitivity analyses. Severity models were evaluated separately in cells excluded from training. After temporal screening, the baseline combining ground elevation and distance to the Special Flood Hazard Area achieved AUCs of 0.632 in Dodge and 0.589 in Douglas. Both spatial uncertainty intervals included 0.50, the value expected by chance. Agreement varied among cells, and additional terrain variables and coefficient tuning provided no consistent improvement. Gamma, lognormal, and Tweedie models also predicted severity poorly across cells. Unresolved provenance of the inundation classification and incomplete historical inventories constrain interpretation. These findings indicate modest, geographically heterogeneous agreement with classified inundation. Accuracy in locating claims remains unmeasured. Contemporaneous inventories, independently assessed inundation data, and verified links between claims and buildings are needed to establish whether disaggregated public claims can support decisions about individual buildings.

Authors

Institutions

Publication Details

Journal
Stochastic Environmental Research and Risk Assessment
Published
2026-11-19
DOI
https://doi.org/10.1007/s00477-026-03404-x
Primary Topic
Flood Risk Assessment and Management
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

How much building-level detail can public flood insurance claims support? Evidence from Eastern Nebraska

Zhenghong Tang, Jesse Andrews
Stochastic Environmental Research and Risk Assessment
Flood Risk Assessment and Management
article

How much building-level detail can public flood insurance claims support? Evidence from Eastern Nebraska

Zhenghong Tang, Jesse Andrews
article en

Abstract

Public flood insurance records omit addresses, preventing direct links between insured losses and individual buildings and limiting their use in local assessments of flood losses. Thus, whether building and terrain attributes can support physically plausible allocations remains uncertain. This study evaluates allocation using these attributes and the geographic transferability of severity models for 423 National Flood Insurance Program claims from the March 2019 flood in Dodge and Douglas Counties, Nebraska. Matched claims were fractionally allocated among candidate residential buildings within cells defined by ZIP Code Tabulation Area and flood zone in each county. Rankings were assessed against an inherited inundation classification using area under the receiver operating characteristic curve (AUC) within cells, spatial block bootstrapping, and sensitivity analyses. Severity models were evaluated separately in cells excluded from training. After temporal screening, the baseline combining ground elevation and distance to the Special Flood Hazard Area achieved AUCs of 0.632 in Dodge and 0.589 in Douglas. Both spatial uncertainty intervals included 0.50, the value expected by chance. Agreement varied among cells, and additional terrain variables and coefficient tuning provided no consistent improvement. Gamma, lognormal, and Tweedie models also predicted severity poorly across cells. Unresolved provenance of the inundation classification and incomplete historical inventories constrain interpretation. These findings indicate modest, geographically heterogeneous agreement with classified inundation. Accuracy in locating claims remains unmeasured. Contemporaneous inventories, independently assessed inundation data, and verified links between claims and buildings are needed to establish whether disaggregated public claims can support decisions about individual buildings.

Stochastic Environmental Research and Risk AssessmentVol. 40(12)
Texas Tech University (US), University of Oklahoma (US)
Flood Risk Assessment and Management
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon