About

Antonia Sebastian is an Assistant Professor and the Director of the Institute for the Environment's Sustainable Triangle Field Site at the University of North Carolina at Chapel Hill. She holds a Ph.D. in Environmental Engineering from Rice University, earned in 2016, and a B.S. in Civil Engineering from Rice University, earned in 2011. Her research focuses on advancing the quantification of complex hazards and their associated risks to social and engineered systems. She studies how human and environmental changes influence the occurrence and frequency of extreme weather and climate hazards, as well as their impacts on communities. Her work is primarily computational, utilizing methods such as geostatistical analysis, machine learning, and numerical modeling to better understand natural hazards and inform risk mitigation strategies.

Research topics

• Computer Science
• Geology
• Geography
• Environmental science
• Cartography
• Environmental resource management
• Oceanography
• Meteorology
• Geotechnical engineering
• Archaeology

Selected publications

• Review article: Exploring methods capturing vulnerability dynamics in the context of flood hazard research

  2026-01-08 · 3 citations

  articleOpen access

  Abstract. Flood vulnerability is highly dynamic, shaped by evolving social, economic, physical, and environmental conditions. Yet, most flood risk assessments still treat vulnerability as static, overlooking how these characteristics evolve and interact with one another. Here, we investigate how dynamic vulnerability in the context of flood risk is considered in 67 studies that use (a) indicator-based, (b) curve-based, (c) dynamic simulation models, (d) qualitative analysis, or (e) statistical analysis of sub-dimensions of vulnerability. Specifically, we examine the conceptual focus (ex-post, during the event, ex-ante), the type of dynamics (event-related, underlying, or complexity-caused), the dimensions of vulnerability captured, and the sources of data used. We find that curve-based approaches were used to address all types of dynamics and conceptual foci, but often in connection with quantitative impact modelling. Dynamic simulation models offered the richest representations of dynamics due to behavioural and systemic complexity but required significant data and computational resources, and faced challenges of model calibration and validation. Indicator-based approaches were effective in capturing underlying socio-economic and environmental changes, though often at coarse temporal resolution. Qualitative methods provided deep insights into the processes and contexts shaping vulnerability. Statistical analyses overlapped conceptually with indicator approaches but tended to focus more narrowly on event-related processes and specific sub-dimensions of vulnerability. Based on our review, we identify three priorities for advancing dynamic flood vulnerability research: leveraging scenarios to explore future change, improving causal inference, and improving data availability and resolution. Additionally, we encourage the flood risk research community to look to other hazard communities or research disciplines that work on systemic or dynamic processes, which might offer inspiration or novel approaches to assessing flood vulnerability dynamics.

  PublisherOA PDFDOI

• Coastal squeeze reduces nitrogen removal services provided by wetlands: insights from an interdisciplinary framework

  Environmental Research Letters · 2025-09-29

  articleOpen access

  Abstract Natural landscapes provide valuable ecosystem services that increase community resilience to environmental change. We present a novel interdisciplinary framework to quantify and spatially evaluate the value and fate of coastal wetlands in the context of sea level rise (SLR) and future land use (FLU) plans. We apply our framework in New Bern, NC, USA, where we project changes in nitrogen removal ecosystem services provided by wetlands and undeveloped open spaces during heavy rainfall events under current sea levels and with 0.15–1.5 m (0.5–5 ft) of SLR. These landscapes currently provide $90 000 USD worth of nitrogen removal ecosystem services annually. Areas currently designated for conservation are especially valuable, contributing 53% of annual services despite making up only 13% of New Bern’s total land area (107 km 2 ). We show that these Conservation designations are expected to lose over 60% of their wetlands with 0.90 m (3 ft) of SLR, reducing New Bern’s expected annual benefit by 56%. Wetland migration to higher elevations is inhibited largely by existing urban development, though we locate potential wetland migration corridors that extend into Developed and Urban Transition FLU designations. Application of our framework can help to maintain ecosystem services and reduce the pressures of coastal squeeze across changing coastal landscapes.

  PublisherDOI

• Building Back Greener: Quantifying Residential Building Decarbonization Opportunities Following Floods

  SSRN Electronic Journal · 2025-01-01

  preprintOpen access
  PublisherDOI

• Characterizing spatiotemporal trends in extreme precipitation in Southeast Texas

  UNC Libraries · 2025-07-04

  articleOpen access1st authorCorresponding
  PublisherDOI

• Author Correction: Climate change exacerbates compound flooding from recent tropical cyclones

  npj natural hazards. · 2025-04-02

  articleOpen access
  PublisherOA PDFDOI

• Reconstructing repetitive flood exposure across 78 events from 1996-2020 in North Carolina, USA

  2025-02-12

  preprintOpen access

  Measuring flooding through time is crucial for understanding exposure and vulnerability — key components to estimating flood risks and impacts. Yet, historical records of flood inundation are sparse. In this study, we leverage high-resolution geospatial data and address-level records of National Flood Insurance Program (NFIP) claims and policies to reconstruct flood extents in eastern North Carolina (NC) for damaging events that occurred between 1996 and 2020 using random forest machine learning algorithms. We identify and model 78 events, achieving an average Area Under the Curve (AUC) of 0.76 and outperforming flood extent estimates from process-based and remote sensing models when evaluated against NFIP records for six events. We find that approximately 90,000 (2.3%) buildings in our study area flooded at least once, of which over 20,000 (0.53%) flooded more than once. We identify more than twice as many flooded buildings as those that filed flood insurance claims between 1996 and 2020. Furthermore, the data indicate that 43% of previously flooded buildings within the study area are located outside the Federal Emergency Management Agency (FEMA) Special Flood Hazard Area (SFHA). Our results illustrate that flood exposure, especially repetitive exposure, is much more widespread than previously recognized. By generating a comprehensive record of past flood extents using address-level observations of damage, we create a first-of-its-kind geospatial database that can be used to identify locations of repetitive flooding. This represents a crucial first step in examining the dynamic relationships between flood exposure, vulnerability, and risk.

  PublisherOA PDFDOI

• Supplementary material to "Flood risks to the financial stability of residential mortgage borrowers: An integrated modeling approach"

  2025-05-19

  preprintOpen access
  PublisherDOI

• Covariance-informed spatiotemporal clustering improves the detection of hazardous weather events

  2025-07-01

  preprintOpen access

  Abstract. Spatiotemporal clustering can be used to detect weather events in multi-dimensional datasets. This method requires that the resolution of a dataset equivalently resolves fluctuations across space and time, thereby normalizing the dataset for unbiased clustering across three dimensions. Yet, few studies test whether a dataset meets this requirement as there is no standard approach to do so. To address this methodological gap, we present a framework to quantify the relationship between space and time using space time separable covariance modelling. We demonstrate that, by defining a temporal resolution of interest (e.g. hours, days), the equivalent spatial resolution can be empirically derived using a space time metric. We present an application using the unsupervised machine learning method Density-Based Spatial Clustering of Applications with Noise (DBSCAN) to detect heat waves and severe storms across the Southeastern US from 1940 to 2023 from ECMWF Reanalysis version 5 (ERA5) data. We analyse the seasonal behaviour of space time metrics for precipitation and heat index before selecting representative values. We find that both ERA5-derived daily heat index and hourly precipitation are insufficiently resolved for unbiased clustering at their native resolutions (i.e., 0.25 spatial degrees [degree] per day for heat index and 0.25 degree per hour for precipitation). We show that a resolution of 0.39 degree per day (0.05 degree per hour) prevents preferential clustering in either the spatial or temporal dimension for heat index (precipitation). We hypothesize that event identification will improve by resampling the data by the space time metric. Heat wave clusters that were produced using the unbiased resolution were compared against the NOAA Storm Events Database from 2019 to 2023. Recall of heat waves increased from 0.92 to 0.94 using the covariance-informed resolution, demonstrating the importance of normalization prior to weather event reconstruction. Ultimately, the inclusion of temporal geostatistics leads to improved reconstruction of historical weather events and enables evaluation of their scale and variability.

  PublisherOA PDFDOI

• Reconstructing Repetitive Flood Exposure Across 78 Events From 1996 to 2020 in North Carolina, USA

  Earth s Future · 2025-07-01 · 6 citations

  articleOpen access

  Abstract Measuring flooding through time is crucial for understanding exposure and vulnerability — key components to estimating flood risks and impacts. Yet, historical records of flood inundation are sparse. In this study, we reconstruct flood extents for 78 damaging events in eastern North Carolina between 1996 and 2020 using high‐resolution geospatial data and address‐level National Flood Insurance Program (NFIP) records. We train random forest models on NFIP‐based labeled flood presence and absence data and a suite of geospatial predictors. Then, we predict the probability of flood damage at every 30 m grid cell within our model domain. Our models achieve an average Area Under the Curve of 0.76 and outperform flood extent estimates from process‐based and remote sensing models when evaluated against NFIP data for six events. We find that approximately 90,000 (2.3%) buildings in our study area flooded at least once, of which over 20,000 (0.53%) flooded more than once. Our estimate is more than double the number of buildings that filed NFIP claims between 1996 and 2020. Furthermore, 43% of flooded buildings are located outside the Federal Emergency Management Agency (FEMA) Special Flood Hazard Area. Our results illustrate that flood exposure, especially repetitive exposure, is much more widespread than previously recognized. By generating a comprehensive record of past flood extents using address‐level observations of damage, we create a first‐of‐its‐kind geospatial database that can be used to identify locations of repetitive flooding. This represents a crucial first step in examining the dynamic relationships between flood exposure, vulnerability, and risk.

  PublisherOA PDFDOI

• Predicting flood damage probability across the conterminous United States

  UNC Libraries · 2025-04-22

  articleOpen access

  Floods are the leading cause of natural disaster damages in the United States, with billions of dollars incurred every year in the form of government payouts, property damages, and agricultural losses. The Federal Emergency Management Agency oversees the delineation of floodplains to mitigate damages, but disparities exist between locations designated as high risk and where flood damages occur due to land use and climate changes and incomplete floodplain mapping. We harnessed publicly available geospatial datasets and random forest algorithms to analyze the spatial distribution and underlying drivers of flood damage probability (FDP) caused by excessive rainfall and overflowing water bodies across the conterminous United States. From this, we produced the first spatially complete map of FDP for the nation, along with spatially explicit standard errors for four selected cities. We trained models using the locations of historical reported flood damage events (n = 71 434) and a suite of geospatial predictors (e.g. flood severity, climate, socio-economic exposure, topographic variables, soil properties, and hydrologic characteristics). We developed independent models for each hydrologic unit code level 2 watershed and generated a FDP for each 100 m pixel. Our model classified damage or no damage with an average area under the curve accuracy of 0.75; however, model performance varied by environmental conditions, with certain land cover classes (e.g. forest) resulting in higher error rates than others (e.g. wetlands). Our results identified FDP hotspots across multiple spatial and regional scales, with high probabilities common in both inland and coastal regions. The highest flood damage probabilities tended to be in areas of low elevation, in close proximity to streams, with extreme precipitation, and with high urban road density. Given rapid environmental changes, our study demonstrates an efficient approach for updating FDP estimates across the nation.

  PublisherDOI

Frequent coauthors

• Philip B. Bedient

  27 shared

• Fangfang Yao

  Ningbo No. 2 Hospital

  22 shared

• Shuai Zhang

  21 shared

• Xiao Yang

  Southern Methodist University

  20 shared

• Samuel D. Brody

  14 shared

• Sebastiaan N. Jonkman

  Delft University of Technology

  14 shared

• Nicholas Clinton

  14 shared

• Bruce Chapman

  Jet Propulsion Laboratory

  14 shared

Education

• PhD, Department of Civil & Environmental Engineering

  Rice University

  2016

• Bachelor of Science, Department of Civil and Environmental Engineering

  Rice University

  2011