Storm Surge Analysis

This week's analysis involved predicting storm surges of coastal areas, DEM's, and building footprints.  The first exercise built two DEM's, one of the New Jersey shoreline before Super Storm Sandy and one after the storm hit.  Using the Raster Calculator and subtracting the after from the before DEM, you can visualize the change of the terrain, in this case erosion and sand accretion.  This difficulty I had with this and the next map had more to do with cartography than the analysis.  It's very difficult to fit so much data into a single map on such a broad extent without a viewer being overwhelmed by statistics and colors.

There was a smaller exercise after the creation of the above map detailing storm surge for a specific county.  The county's boundary was given, the DEM's raster was reclassified to include and exclude from the amount of the storm surge, and the resulting raster was then turned into a polygon.  This shows the viewer spatially where a storm surge in and around the county was to take place.  Going a bit further, the polygon was clipped to the county boundary and a field was added in order to calculate the polygon's area relative to the county's area.

In the final analysis of this week's module, we explored a potential storm surge that would affect southwest Florida.  My wife and I plan on moving the the Tampa Bay/St. Pete area in the next few years, so this analysis was especially interesting for me.  Two DEM's were provided; LiDAR with 25' horizontal resolution and a USGS DEM with 100' horizontal resolution.  These were reclassified to include and exclude areas above and below the storm surge threshold, in this case one meter.  We used the REGION GROUP tool to exclude disconnected low lying areas further inland for this exercise.  After using SELECT BY ATTRIBUTES and EXTRACT BY ATTRIBUTES, the connected areas were turned into polygons for the remainder of the analysis.

A SPATIAL JOIN was used for each of the DEM's with a building footprint layer to figure out which buildings were within the storm surge according to each DEM.  Subsequently, a UNION was used to combine these two layers together to later SELECT BY ATTRIBUTES for the remainder of the analysis.  The UNION was exceptionally important for determining which building type lay within one or both of the storm surge areas of the two different DEM's.  For this reason, setting a value of 0 rather than NODATA was key to keeping my head on straight in this fairly involved analysis.  After doing this, it was easy to create different clauses to understand the type of building affected by either of the storm surges.  To no surprise, the LiDAR DEM was more reliable and accurate.  This was used to calculate the percentages of the Errors of Omission and Errors of Commission.  After completing this analysis, I can understand why these statistics are so important when literally billions of dollars are on the line with regards to how accurate a DEM is when used in flood insurance analysis.

I tried a few different layouts before going with six smaller maps.  The area is way too big to have a large extent, and similarly too big to choose just one area to focus on.  I decided on bringing attention to Marco Island and segmented the area into quadrants using to map inlays to visualize the area on a macro and micro level.  I hope it's not too busy!  This exercise definitely got my wheels turning as to how to get a bit more creative in cartographic techniques that I can still use a lot of work on.  I did find a few new feature I haven't used in ArcPro yet, like making a Shapefile polygon to show as the study area on the map and in the final layout.