Location

The Leonardo Event Center

Start Date

5-12-2015 3:24 PM

Description

Increasing rates in sea level rise imply drastic consequences for U.S. coastal populations, infrastructure, ecological systems, and natural re- sources in the coming decades. These direct impacts will lead to negative repercussions in public health, biodiversity, tourism, and other aspects of the global economy. Using hourly tide readings from the past 30 years at 38 gauges along the east coast, we wish to develop a model that will allow us to analyze the trends in this type of data and to accurately and precisely predict sea level change along the east coast. The model devel- oped is an iterative generalized additive model that will use spatial and temporal dependence between gauges and across time, allowing us to predict sea level change all along the east coast, not only at the stations for which we have data. This generalized additive model includes a linear term, a seasonal trend term fit with B-splines, and a term accounting for additional spatial variance with latent factors estimated by confirmatory factor analysis.

Share

COinS
 
May 12th, 3:24 PM

Analyzing Sea-level Change on the East Coast With Spatiotemporally Correlated Data

The Leonardo Event Center

Increasing rates in sea level rise imply drastic consequences for U.S. coastal populations, infrastructure, ecological systems, and natural re- sources in the coming decades. These direct impacts will lead to negative repercussions in public health, biodiversity, tourism, and other aspects of the global economy. Using hourly tide readings from the past 30 years at 38 gauges along the east coast, we wish to develop a model that will allow us to analyze the trends in this type of data and to accurately and precisely predict sea level change along the east coast. The model devel- oped is an iterative generalized additive model that will use spatial and temporal dependence between gauges and across time, allowing us to predict sea level change all along the east coast, not only at the stations for which we have data. This generalized additive model includes a linear term, a seasonal trend term fit with B-splines, and a term accounting for additional spatial variance with latent factors estimated by confirmatory factor analysis.