Prediction of River Flooding using Geospatial and Statistical Analysis in New York, USA and Kent, UK

Thursday, 18 December 2014
Antonios Marsellos, University of Brighton, School of Environment and Technology, Brighton, BN2, United Kingdom, Katerina Tsakiri, University of Brighton, School of Computing, Engineering, and Mathematics, Brighton, United Kingdom and Martin Smith, University of Brighton, School of Environment and Technology, Brighton, United Kingdom
Flooding in the rivers normally occurs during periods of excessive precipitation (i.e. New York, USA; Kent, UK) or ice jams during the winter period (New York, USA). For the prediction and mapping of the river flooding, it is necessary to evaluate the spatial distribution of the water (volume) in the river as well as study the interaction between the climatic and hydrological variables. Two study areas have been analyzed; one in Mohawk River, New York and one in Kent, United Kingdom (UK). A high resolution Digital Elevation Model (DEM) of the Mohawk River, New York has been used for a GIS flooding simulation to determine the maximum elevation value of the water that cannot continue to be restricted in the trunk stream and as a result flooding in the river may be triggered. The Flooding Trigger Level (FTL) is determined by incremental volumetric and surface calculations from Triangulated Irregular Network (TIN) with the use of GIS software and LiDAR data. The prediction of flooding in the river can also be improved by the statistical analysis of the hydrological and climatic variables in Mohawk River and Kent, UK. A methodology of time series analysis has been applied for the decomposition of the hydrological (water flow and ground water data) and climatic data in both locations. The KZ (Kolmogorov-Zurbenko) filter is used for the decomposition of the time series into the long, seasonal, and short term components. The explanation of the long term component of the water flow using the climatic variables has been improved up to 90% for both locations. Similar analysis has been performed for the prediction of the seasonal and short term component. This methodology can be applied for flooding of the rivers in multiple sites.