Making Maps: Lab 8 - Damage Assessment

Moving on from coastal flooding last week, we turned to damage assessments this lab. For our lab, we used data from Hurricane Sandy to create a map tracking the changing storm up the Atlantic to the New England States, then we used aerial images and parcel data to do a damage assessment of a section of the impacted New Jersey coast.

The hurricane tracking map was fairly straightforward, with some exploration of symbols and labels. The rest of the lab was quite a bit of new layer and data creation, so attention to detail was important. Working within a geodatabase, I created raster mosaics of pre and post-storm imagery, visually explored the images through the Effects toolbar, created attribute domains with coded values, classified structural damage based on the pre and post imagery, and finally associated parcel data with the structural damage feature class I had created.

The actual damage assessment based on aerial imagery was one of the more challenging parts of the lab. Not because it was technically hard, rather because it can be difficult to decide from an image whether a structure is slightly damaged, very damaged, etc. There was no real standard to go by, just our own judgement and consistency, which definitely gave me a sense of how challenging that kind of work is in "real life." After the assessment of structures in a study area, I created a buffer zone of 100, 200, and 300 meters from the coastline to the study area and used one block of structures to examine damage patterns. Based on my non-ground verified damage assessment, 100% of the structures within 100 m or less were damaged or destroyed, less than 50% had major damage or were destroyed within 100-200 m, and 0% had major damage or were destroyed within 200-300 m from the coastline. As noted earlier, accuracy and reliability are reduced when a damage assessment is only based on imagery. Field surveys would increase accuracy and help with verifying the results.