Much like the prior module, this module looked at how GIS can aid in post-storm analysis using data from Hurricane Sandy. In contrast to the prior module that used DEM's to analyze coastline flooding, this module focused mainly on using storm data to track the path and severity of the hurricane as well as digital imagery to assess structural damage before and after it hit the coast.
I began the module learning how to map the path of the hurricane, the nearby or affected locations of the path, and how to symbolize its severity. Given boundary layers for world countries and the United States (as well as its state boundaries), I added them to a new map in ArcGIS Pro. Using the United States layers, I created a new feature laying to properly symbolize the states where Hurricane Sandy hit or affected (per FEMA). I next added tabular data from Hurricane Sandy's path and used Display XY Data to symbolize the data as points. Using the Points to Line tool, I turned this data into a polyline to better symbolize the path. I most enjoyed learning how to use Symbology to symbolize these points based on their storm type per real storm standards. I had to learn how to create marker symbols and what fonts to use within Symbology to achieve this, as well as some other small skills. Finally, I added graticules and cleaned the map up in Layout to complete the following map design:
As opposed to the more stylistic portion beginning this exercise, the next portion involved more detailed processes. Next in the exercise I was to create a citizen damage assessment on the coast of a small study area in New Jersey. I was provided with a parcel polygon feature layer as well as a polygon outlining the study area. I was also given imagery files from before and after the hurricane hit, however I had to clean it up for use. To do so, I created a mosaic of the images in ArcGIS Pro. By adding a new Mosaic Dataset to my geodatabase and selecting the proper band and pixel types based on the images' metadata, I was able to create the desired layers. I would go on to use these images to assess the damage of structures in the study area post-storm. I would need to create new data to achieve this. I began by adding domains to my geodatabase to ensure integrity. Once I had my domains in place with the proper codes and descriptions, I created a new feature class to record structure damage for the buildings in the study area by parcel. I used tools within the Edit tab of the feature class to record points for each location and evaluated them based on structure damage, wind damage, inundation, and structure type.
The images above show the results of the structure damage analysis. The first image shows the assessment using imagery from before the storm, and the image below using imagery from right after. The points are symbolized on severity of the level of structure damage alone, with the more green values being less damaged and the more red values being more damaged.
With the points created, I was able to draw patterns from the data which could be used to extrapolate to nearby areas. I got a count of buildings within a 100m, 1-200m, and 2-300m distance from the coast that sustained structural damage (as well as counts based on severity). I achieved this by first digitizing the coastline parallel to the study area. I created this feature in much the same way as the points I recorded, however I used a polyline feature type. By using Select by Location with the input feature set as my structure damage layer, the Relationship parameter to within a distance, and the Selecting Features parameter to the coastline layer I digitized, and the distance to 100m, I created a new feature layer from the selection. I then was able to use Select by Attributes to get my totals. For the 100m to 200m analysis, I followed a similar process by selecting by location within 200m. I then selected by location again within 100m, however I chose the Selection Type as remove from the current selection. This isolated only the values within 100m to 200m. I completed a similar process for the 200m to 300m analysis, and again used Select by Attributes on a new feature layer I created to finish the analysis.
My findings showed that buildings within 100m of the coastline sustained major damage or were destroyed. While raw totals of these buildings are less than those in other distances, the density of buildings in this area was also lower. The 1-200m and 2-300m distances showed almost identical results, with the majority of buildings sustaining minor damage or simply being affected. In both distances, only a few buildings were destroyed. The one major difference is that the 2-300m distance is the only one where I indicated that some of the buildings had sustained no damage at all. I would deem the results from the analysis reliable to extrapolate to others, however only on a local scale. Too far away from the study area and the coast shape, storm strength, and topography might be drastically different. Based on the patterns from the three clusters of buildings studied being so similar, this would indeed be a good basis for analysis in nearby areas.
Comments
Post a Comment