Haigang Liu

Date of Award

Spring 2019

Document Type

Open Access Dissertation



First Advisor

David B. Hitchcock


Spatio-temporal data are everywhere: we encounter them on TV, in newspapers, on computer screens, on tablets, and on plain paper maps. As a result, researchers in di- verse areas are increasingly faced with the task of modeling geographically-referenced and temporally-correlated data. In this dissertation, we propose two different spa- tiotemporal models to capture the behavior of rainfall and flood data in the state of South Carolina.

Both models are built using a Bayesian hierarchical framework, which involves specifying the true underlying process in the first level and the spatio-temporal ran- dom effect in the second level of the hierarchy. The prior distribution of the param- eters or hyper-parameters is specified in the third stage. The two models differ in the covariance structure of the spatial random effects. In the rainfall spatiotemporal model, we employ a Gaussian process model which has a distance-based covariance. To model the flood data, we use a conditional autoregressive (CAR) model with a proximity matrix.

Another aspect that sets the models apart is the covariates considered. In particu- lar, the precipitation model incorporates a variable related to sea surface temperature (SST) to reflect the effect of El Niño-Southern Oscillation (ENSO) activity, along with monthly maximum temperature among other predictors. In the flood model, a gridded field of precipitation values with a spatial resolution of roughly 4 × 4 km is used as one of the covariates since investigating the dynamics between the rainfall and flood levels is of interest.