Applications Week 5: Damage Assessment

For this week's lab, we analyzed imagery from before and after Hurricane Sandy in New Jersey to assess the level of damage in an urban area. We first mapped the path of the storm as it progressed through different levels of storm/hurricane. We created a citizen damage assessment survey that could be used to crowdsource information, photos, and locations of damage from the hurricane. After creating domains for a new point feature class, we added points to this class by comparing the before and after imagery to assess the level of hurricane damage for each structure in a small study area. This part was tricky because the shadows in the post-storm imagery sometimes made it difficult to see debris or damage. It was easy to tell when a structure was completely destroyed, but I often wasn’t sure whether a structure had minor damage or not. On-the-ground photos or other damage assessment information, such as the citizen survey we created, would have been helpful in identifying damage that is not visible from above.

 

The study area with the points I created, organized by the level of structural damage I assessed from the imagery

I then created a line feature for the coastline and created 100 meter, 200 meter, and 300 meter buffers from the coastline to evaluate how many structures with each level of damage were in each buffer zone. I found that there was a clear correlation between proximity to the coast and structural damage, with the percentage of structures that had major damage or were destroyed decreasing reliably with distance. Based on my lack of confidence in my ability to distinguish between “no damage,” “affected,” or “minor damage," I would not extrapolate these exact percentages to nearby areas. However, I would say that based on this data, the majority of structures 100 meters or closer to the coastline had major damage or were destroyed; the majority of structures 100-200 meters from the coastline had little to no damage, with some minor damage, major damage, and destruction; and that structures 200-300 meters from the coastline had little to no damage. This finding would be useful for disaster relief efforts, suggesting that they should prioritize areas within 100 meters from the coastline, with attention also given to areas within 200 meters.