Contents lists available at ScienceDirect Int J Appl Earth Obs Geoinformation journal homepage: www.elsevier.com/locate/jag Multi-source data fusion and modeling to assess and communicate complex flood dynamics to support decision-making for downstream areas of dams: The 2011 hurricane irene and schoharie creek floods, NY Chris S. Renschler ⁎ , Zhihao Wang Department of Geography, Landscape-based Environmental System Analysis & Modeling (LESAM) Laboratory, 116 Wilkeson Quad, Buffalo, NY 14261, USA ARTICLE INFO Keywords: Flood Modeling HEC-RAS GIS Return period LiDAR ABSTRACT In light of climate and land use change, stakeholders around the world are interested in assessing historic and likely future flood dynamics and flood extents for decision-making in watersheds with dams as well as limited availability of stream gages and costly technical resources. This research evaluates an assessment and com- munication approach of combining GIS, hydraulic modeling based on latest remote sensing and topographic imagery by comparing the results to an actual flood event and available stream gages. On August 28th 2011, floods caused by Hurricane Irene swept through a large rural area in New York State, leaving thousands of people homeless, devastating towns and cities. Damage was widespread though the estimated and actual floods inundation and associated return period were still unclear since the flooding was artificially increased by flood water release due to fear of a dam break. This research uses the stream section right below the dam between two stream gages North Blenheim and Breakabeen along Schoharie Creek as a case study site to validate the ap- proach. The data fusion approach uses a GIS, commonly available data sources, the hydraulic model HEC-RAS as well as airborne LiDAR data that were collected two days after the flood event (Aug 30, 2011). The aerial imagery of the airborne survey depicts a low flow event as well as the evidence of the record flood such as debris and other signs of damage to validate the hydrologic simulation results with the available stream gauges. Model results were also compared to the official Federal Emergency Management Agency (FEMA) flood scenarios to determine the actual flood return period of the event. The dynamic of the flood levels was then used to visualize the flood and the actual loss of the Old Blenheim Bridge using Google Sketchup. Integration of multi-source data, cross-validation and visualization provides new ways to utilize pre- and post-event remote sensing imagery and hydrologic models to better understand and communicate the complex spatial-temporal dynamics, return per- iods and potential/actual consequences to decision-makers and the local population. 1. Introduction 1.1. Problem definition Average global flood losses in 2005 were estimated to be about US $6 billion per year (Hallegatte et al., 2013) and are projected to in- crease to US$52 billion by 2050. Stakeholders around the world are interested in assessing historic and likely future flood dynamics and flood extents especially for watersheds lacking stream gages. This re- search evaluates an approach of combining GIS, hydraulic modeling based on commonly available data, latest remote sensing and topo- graphic imagery by comparing the results to an actual record flood event. Before Superstorm Sandy devastated coastal areas of New York (NY) and other States along the US Eastern Coastline in 2012, there was Hurricane Irene in August 2011 that was a precursor to potential da- mage from a significant event. Hurricane Irene heavily damaged mainly rural inland areas in the same region as Sandy did and was the fifth most significant US flood events after Hurricanes Katrina in 2005, Sandy in 2012, Ike in 2008 and Ivan in 2004. National flood insurance payouts for Irene impacted properties, totaled more than $1.3 Billion, in more than a dozen east coast states (Insurance Information Institute, 2015). Prolonged rainfall before, during and after landfall at the New Jersey shoreline (Avila and Cangialosi, 2011), caused water stages along the Schoharie Creek, NY to peak on August 28, 2011 (Fig. 1), North of the Catskills Mountains. Over 300 mm of rainfall caused the Gilboa Dam in the Schoharie Creek Watershed to overflow and officials opened the Blenheim-Gilboa spillway downstream and for the first http://dx.doi.org/10.1016/j.jag.2017.06.002 Received 9 December 2016; Received in revised form 26 May 2017; Accepted 1 June 2017 ⁎ Corresponding author. E-mail address: rensch@buffalo.edu (C.S. Renschler). Int J Appl Earth Obs Geoinformation 62 (2017) 157–173 0303-2434/ © 2017 Elsevier B.V. All rights reserved. MARK