Hurricane coastal flood analysis using multispectral spectral images

Flooding is one of the main hazards caused by extreme events such as hurricanes and tropical storms. Therefore, flood maps are a crucial tool to support policy makers, environmental managers and other government agencies for emergency management, disaster recovery and risk reduction planning. However traditional flood mapping methods rely heavily on the interpolation of hydrodynamic models results, and most recently, the extensive collection of field data. These methods are time-consuming, labor intensive, and costly. Efficient and fast response alternative methods should be developed in order to improve flood mapping, and remote sensing has been proved as a valuable tool for this application. Our goal in this paper is to introduce a novel technique based on spectral analysis in order to aggregate knowledge and information to map coastal flood areas. For this purpose we used the Normalized Diference Water Index (NDWI) which was derived from two the medium resolution LANDSAT/TM 5 surface reflectance product from the LANDSAT climate data record (CDR). This product is generated from specialized software called Landsat Ecosystem Disturbance Adaptive Processing System (LEDAPS). We used the surface reflectance products acquired before and after the passage of Hurricane Ike for East Texas in September of 2008. We used as end member a classification of estimated flooded area based on the United States Geological Survey (USGS) mobile storm surge network that was deployed for Hurricane Ike. We used a dataset which consisted of 59 water levels recording stations. The estimated flooded area was delineated interpolating the maximum surge in each location using a spline with barriers method with high tension and a 30 meter Digital Elevation Model (DEM) from the National Elevation Dataset (NED). Our results showed that, in the flooded area, the NDWI values decreased after the hurricane landfall on average from 0.38 to 0.18 and the median value decreased from 0.36 to 0.2. However for the non-flooded area the NDWI increased after the hurricane landfall. The average value varied from 0.15 to 0.43 and the median value from 0.13 to 0.43. These results demonstrate that these differences can be explored for the mapping of flood areas. As NDWI was developed to quantify the amount of water in the leaf of the plants, the increase of the value is expected within the amount of water that the leaf will absorb. However in flooded areas the amount of water is so high that it is possible that the reflectance follows the water spectral behavior absorbing more than reflecting in the Near Infrared region. Thus, remote sensing techniques showed to be powerful tools since they could characterize flooded areas. However further studies are needed, applying and validating these techniques for other regions and different storms. Optical remote sensing is promising for many applications, since it will be an open door to studies of spatial and temporal analysis of the flood impacts mainly in areas with remote access and with a lack of in situ data.