GIS4048 Module 4: Natural Hazards - Hurricanes

Just in time for the beginning of hurricane season, this module presented an opportunity to work as a contract employee of the Federal Emergency Management Agency (FEMA). Of interest to FEMA in this scenario was an inventory of structures with structure damage, wind damage or inundation as a result of Hurricane Sandy’s landfall in New Jersey.

Although only classified as a Category 1 hurricane at landfall on the New Jersey coast, Hurricane Sandy was a catastrophic storm due to its enormous size and other factors which amplified the storm surge. States all along the Atlantic coast were impacted to some degree by Hurricane Sandy. Ten states and the District of Columbia were declared major disaster areas by FEMA, making them eligible for funding from the President's Disaster Relief Fund which is managed by FEMA and other federal agencies.

To meet FEMA’s objectives, the track of Hurricane Sandy was plotted. This was done by adding an Excel file to the map and exporting the data to a geodatabase. The Points to Line tool was used to show Hurricane Sandy’s track. That information is shown along with wind speed and barometric pressure on the map below along with the states and district that were declared disaster areas by FEMA. Graticules were added to this map as well. 4° intervals were selected for the graticules since 2° intervals are too closely spaced and 5° intervals are not as easily interpolated as 4° intervals.

Hurricane Sandy Track and
States/District with Major Disaster Declarations

Geodatabase datasets were created for mosaicking two separate sets of aerial imagery (before and after Hurricane Sandy) and imagery effects tools were utilized for comparing pre- and post-hurricane imagery. Attribute domains were created and used to catalog damage in order to compile an inventory of structures along one block of a street in New Jersey. The difficulty with this part of the exercise was determining the degree of damage. Without seeing elevation views of the structures, the extent of any damage was hard to assess. The following map shows the damage assessment and the pre- and post-imagery of structures which were cataloged along one block. 

Hurricane Sandy Damage Assessment
Coolidge Avenue, Toms River Township, New Jersey (2012)

Information from maps such as these would assist FEMA and other organizations to evaluate the extent of damage after a hurricane and to determine how best to allocate resources.

GIS4048 Module 3: Natural Hazards - Tsunami

The March 2011 tsunami that devastated Japan served as a reminder of the importance of planning for natural and nuclear disasters. GIS provides a comprehensive system for producing maps and graphics to aid in such natural and nuclear disaster planning. This lab assignment tasked students with mapping the evacuation zones around the Fukushima Nuclear Power Point as well as determining the at-risk populations within each of those zones. The lab also included determining evacuation zones based on actual measurements of tsunami "run up" (observed maximum water height above a reference sea level) and identifying at-risk cities, roads, and nuclear power plants in those zones. This information was consolidated in one map:

Evacuation Zones in Vicinity of Fukushima II Nuclear Power Plant

Production of the map began with creating a file geodatabase to keep data organized. Feature datasets were created in the file geodatabase and then populated with imported shapefiles. All of the feature datasets were assigned the same projected coordinate system (WGS 1984 UTM 54N). One Excel file was converted and added to the geodatabase. As it was for several other students, creating raster datasets and mosaicking the DEMs caused plenty of issues. Many thanks go out to Brandon and Aaron for their very helpful discussion board posts which resulted in successful completion of that part of the lab. All of this organized data then was used to complete the second portion of the lab.

The second portion of the lab included creation of another file geodatabase and data creation using selection by location, buffering and clipping, extracting by mask, creating a graph (to determine breakpoints for classes of data), and model building. The model built was very involved and awesome to see in action. The flexibility and ease with which changes could be made to the model were appreciated. Another new technique learned in this lab was how to build label expressions. While many of the processes (especially those used to determine the run-up evacuation zones) are not among those that I use often, with continued practice I expect that they will be easier to use and troubleshoot should the need arise.

Finally, the information was synthesized into one map showing the evacuation zones around the Fukushima Nuclear Power Plant and the evacuation zones for tsunami run-up. This was a lot of information to pack into a single map, but, as the saying goes, "A picture is worth a thousand words". In all honesty, now I am feeling a bit vulnerable about living 75 feet above sea level.

GIS4048 Module 2: Natural Hazards - Lahars

Lahars started off the course’s study of natural hazards. Unfamiliarity with the term lead to some research prior to starting the module requirements. The Javanese word “lahar” used for a volcanic mudflow originated in the 1920s. Its use increased dramatically in the 1980s and 1990s, probably in reaction to the frequent volcanic activity during that time period. In some areas this fast-running mixture of mud caused by rapidly melted snow and ice poses the greatest risk of any volcanic activity. Speeds of 45-50 miles per hour and flow depths of 100 feet where the lahars were confined in valleys have been reported at Mount Rainier. Because of their concrete-like properties, lahars are capable of demolishing most structures. They can also occur during periods when no eruptions are occurring in which case there may be no advance warning of the impending event. For these reasons, the importance of risk assessment and hazard planning is critical. This exercise was created to demonstrate that process.

Acting as a private consultant hired to identify potential inundation zones in the vicinity of Mount Hood, Oregon, the process involved mosaicking digital elevation models and then using the Hydrology Tools to determine drainage flow. Coupling that output with 2010 Census data and Oregon school data, a population analysis and schools-at-risk identification was performed. State and local officials could use this information for hazard planning and response time to help keep area residents and visitors safe from lahars.

After establishing a geodatabase to keep the project organized and efficient, XY tool was used to show the location of Mount Hood. Drainage areas were created from the mosaicked DEMs using some of the Hydrology Tools (Fill to eliminate sinks, Flow Direction to assign direction of flow to individual cells, and Flow Accumulation to calculate how many other cells flow into each cell) to create a stream network. The need for an attribute table for further work called for changing the fill output from a floating point raster to an integer raster. Using the Con tool (conditional statement) reduced the stream network to only those areas with sufficient flow accumulation to qualify as streams. That output was then converted to polyline vector features – areas most likely to become inundated during a lahar. This whole process - creating a stream network from DEMs - was rather fascinating to complete.

Applying buffers to the streams provided areas of potential inundation by lahars which were then used in the selection process with Census population block groups and schools data. This resulted in the areas on which teams should focus for large lahar event hazard planning. From this information the following map was created:

Mount Hood Lahar Hazard Assessment

Populations at risk for being impacted were calculated for the possible inundation zones around drainage basins within the study area only rather than for the entire dataset. The potentially impacted population number outside the study area is much greater than the number calculated and displayed on the map. The values for this map are calculated for the study area only which has a lower population density than the areas to the west, so the block group population ranges may differ from those presented by others. Overall, lahars from Mount Hood have the potential to impact 42,689 people within the study area based on 2010 Bureau of Census figures. Including the area outside the study area 212,791 people could be impacted.

Although it does result in a more cluttered-looking map, especially in areas with small census block groups, the census block group outlines were retained. This was decided primarily to minimize misinterpretation of the map resulting in an underestimate of the actual population with the potential to be impacted by a large lahar.

This particular lab demonstrated the variability in information that could be produced for a client. For example, the lab requirements were focused on lahars, but in reality a hazard-planning team would also want to know what populations would be at risk for other volcanic hazards. In this particular case, one requirement was to provide the number of schools at risk from lahar inundation. However, just outside one of the mile-wide lahar inundation zones was an additional school located closer to Mount Hood. Naturally, one would expect the hazard-planning team to be made aware of this school, but including that information was outside the scope of the lab requirements. This also illustrates the importance of communicating similar situations to a client for further clarification or revision of the data to be provided.

The ability to go back and review previous geoprocessing steps came in quite handy during this lab!