GIS4048 Module 7: Homeland Security - MEDS Protect

Three Mile Buffer Zone around Boston Marathon Finish Line
with Nearest Hospitals and Finish Line Perimeter Security Checkpoints

Lessons learned from the 2013 Boston Marathon bombing have prompted increased security measures by the Department of Homeland Security. These measures include increased security at ingress and egress points as well as improved surveillance around the event site for monitoring purposes. The first map created for this exercise shows a 3-mile buffer zone around the Boston Marathon finish line, a 500-foot buffer around the finish line, the locations of the ten nearest, currently operational hospitals with emergency rooms, and their 500-foot radius protective buffer zones within which increased security measures prior to, during, and following the marathon can be planned. Fifteen security checkpoints on local and secondary roads leading to the finish line are shown at the outer limit of the 500-foot buffer zone around the finish line as shown on the smaller inset map. Identification verification and backpack checks could take place at those locations.

The second map focused on specific locations suggested for placement of surveillance equipment within the immediate vicinity of the finish line itself. LAS Dataset 3D View and the orthoimagery layer were used to determine where to place the 15 potential surveillance points. Prior to point selection, hillshade was generated for 2:30 pm on April 15, 2013. Shadows impact surveillance equipment's capabilities, and hillshade provided a baseline for the day of the marathon. A difficulty encountered during selection of surveillance locations was determining a set of locations that would provide complete coverage of the finish line vicinity as determined by using the Viewshed tool. Viewshed is an indicator of visibility from other vantage points, in this case, surveillance locations. Believing that cameras placed along roof lines or on the sides of buildings would provide that coverage, initial points were placed in that manner. However, to obtain fairly contiguous coverage, the points had to be adjusted not only vertically but also horizontally. Not knowing the heights of the buildings in the area was a disadvantage, but taking advantage of Google street view as well as researching commercial building heights in general provided some guidelines to estimate reasonable, attainable heights for surveillance cameras. Using 3D GIS techniques to determine locations of surveillance points is significantly more economical, effective, and efficient time-wise than physically selecting, inspecting, and adjusting potential surveillance points.

Suggested Locations for Surveillance Points near Boston Marathon Finish Line

The biggest roadblock that I encountered in trying to assemble a comprehensive security analysis for the marathon came with the inability to complete the line-of-sight portion of the lab. Even with meticulous attention paid to the line-by-line instructions and completing the work in one session (having been forewarned), I could not get a line of sight to show up between any pair of selected points (any surveillance point and the finish line). Redoing the map from the very beginning several times, including completely re-downloading the data again did not improve the situation. This resulted in an incomplete map as a profile graph could not be made for a nonexistent line of sight. Having a profile graph would have enabled a surveillance team to further evaluate potential surveillance points. For instance, the horizontal location of the obstruction can be determined by the aerial view, but how the obstruction possibly could be lessened by the vertical adjustment of the surveillance camera can only be determined by a side view shown with a profile graph.

A 3D version of the map was created in ArcScene using the finish line raster as a surface layer. The orthoimagery layer was draped over it as well. Because I did not have any lines of sight to put in the ArcScene portion of the lab, I added the finishline and suggested surveillance points, hoping that the surveillance points would be placed at their offset heights. That did not happen. Again, the actual line of sights are an essential part of a security analysis like this.

Even though I was unable to complete the lab as intended, the experience was extremely worthwhile. The numerous repetitions of certain steps have helped reinforce key aspects of the lab.

GIS4048 Module 6: Homeland Security - Prepare MEDS

The focus of this week's module was the compilation of a Minimum Essential Data Set (MEDS) to support a standards-based geospatial model for GIS systems as developed by the Department of Homeland Security in the interest of prevention and protection of areas during a time of crisis. Having a standards-based geospatial model founded on a MEDS provides a uniform basis for any community or organization before and during a time of crisis, whether security- or hazard-related. The uniformity of the system ensures that at the time of crisis, the system will be available and useful to anyone involved with crisis planning or recovery. MEDS have already been created for large metropolitan areas through the cooperation of local, state, and federal government agencies.

Data for MEDS are compiled from several different internet sources; shapefiles, rasters, tables, and geodatabases are included. Layers to be included in MEDS are: orthoimagery, elevation, hydrography, transportation, boundaries, structures, land cover, and geographic names. Data must be in North America Datum of 1983 to meet MEDS specifications.

The task for this particular exercise was to identify and put together an essential dataset in preparation for the homeland security crisis of the Boston Marathon bombing. All of the necessary data for this module was downloaded from National Map Viewer and provided. After organizing the map document with a default geodatabase and stipulating the data frame's units (meters), group layers were created to organize the required MEDS layers. Each created group layer was populated with the applicable data layers. The Boston Metropolitan Statistical Area (BMSA) layers were added to the Boundaries group layer. (BMSA includes Boston, Brookline, Cambridge, Chelsea, Everett, Quincy, Revere, Somerville, and Winthrop along with a 10-mile buffer around the outer extent.) The transportation group layer was populated with three road layers (local, primary, and secondary) which were created by using Joins and Relates to join the attributes from the CFCC table to the BMSA Roads layer and then selecting by atttributes to create the three aforementioned layers. The selections were exported data (matching the coordinate system of the receiving geodatabase) and then added to the map as layers which were then adjusted for symbology and labels. Adjustments were made to the Scale Range for each transportation layer as well as the transportation labels so that clarity would reign supreme with any changes in scale.

The Hydrography group layer was populated with three feature classes that had been provided, and Orthoimagery and Elevation group layers were populated with their respective raster datasets. Working with the Landcover raster was quite a lot of fun. After using Extract by Mask to limit the area covered by the raster to only that of the BMSA, the symbology was altered using more meaningful colors imported with a color map created from the National Land Cover Database 2006 Legend. Those colors portray the types of landcover much more adequately than the random colors. Labels for the landcovers were also changed to match the color map labels.

In order to add geographic names from the Geographic Names Information System data file, the text file format was first changed from CSVDelimited to Delimited(|) format to create columns of information. The x,y coordinates of the table were then used as input to create a feature class. After saving the output to the geodatabase, the layer was added to Geographic Names group layer and then projected to the State Plane system (the re-projected layer was also added to the Geographic Names group layer). From this re-projected layer, only the features that were entirely within the 6 counties of the BMSA were selected using Select by Attributes and then Select by Location. This final selection of geographic names was exported as data, added to the map, and moved into the Geographic Names group layer as well before removing the other layer.

The creation of these seven group layers was essentially the staging portion of the geospatial model to follow. In order to make the layers accessible to others who might need to use them, each group layer was saved as a layer file. With the preparation part complete, the next step is to protect ~ the focus of next week's module.

GIS4048 Module 5: Homeland Security - Washington, DC Crime

The purpose of this module was to perform as a District of Columbia Metropolitan Police Department GIS Analyst who is responsible for crime analysis. The goal was to help allocate resources by analyzing January 2011 crime statistics for DC. Specifically, the project involved determining the following:

• whether current patrols are effective as is or need to be altered based on crime patterns in proximity to police stations
• patterns of aggravated assault, homicide, and theft-related crimes

The first map depicts each existing police station's location and percentage of the District of Columbia's total crime for the month. The locations of individual crimes were included for a visual comparison of crimes in relation to existing police stations. Proposed police station locations were based on analysis of crimes in proximity to existing stations, proximity of stations to other stations, and availability of land or existing building for a new station.

Proposed police station locations were based on the data showing that most crimes occur 1/2 to 1 mile from a police station. This was done by applying a multiple ring buffer around each station before performing a spatial join between the crime data layer and the buffers. The 6th and 7th District police stations both had relatively high percentages of crime occuring close to them. Having lived in southeast Washington in the past, I know that that area's reputation for higher crime rates in general is common knowledge to District residents. (One of my roommates there was mugged not once, but twice. In the same night.) The proposed location south of the 7th District police station was selected because of the 7th’s highest rate of crime occurring close to it along with the numerous crimes that have occurred in extreme southern DC south of the 7th District where no police stations are located currently. (A large empty parcel is situated at Mississippi and 10th Place SE). Wherever an additional station to reduce crime around the 7th District is located, it should be east of the Anacostia River to minimize response time.

Also in southeast Washington, the 6th District has a relatively high crime rate occurring near it as well, but perhaps personnel could be transferred from the 6th Substation, 1st District Station, or 1st District Substation, all of which have significantly lower crime rates than the 6th District.

Additionally, DC would benefit from having another station located south of the 3rd District police station (12.56% of total crimes) and west of the Asian Liaison Unit with its 9.1% crime rate occurring in its neighborhood. There is a lot of overlap of police station buffer zones north of the 3rd District station and east of the Asian Liaison Unit, but the areas south of the 3rd District and west of the Asian Liaison Unit have a significant number of crimes and there is no other police station from that area all the way out to the DC boundary. The area of 20th St. NW and L St NW has commercial buildings which might have sufficient space available to accommodate a substation. Percentages of crime, police station locations, and proposed station locations area shown on the first map:

Analysis of Crimes Committed in District of Columbia in January 2011

The density maps of burglaries, homicides, and sex abuse crimes were created using the kernel density method with a radius of 605 (area in square kilometers). The density maps reveal similar patterns to those shown with the previous map: higher concentrations of crime in the southeast quadrant of DC as well as central DC.

Density Maps of Burglaries, Sex Abuse, and Homicides
in Washington DC (January 2011 data)

In general, theft in its various forms, burglary, and robbery were the most common crimes. Because these maps were created by a DC MPD department GIS analyst and are intended for internal use, a locator map was not included. The first map is a bit busy with the locations of individual crimes shown, but seeing those locations of specific crimes helps determine patterns, and the visual aspect reinforces the statistics and written analysis. Some portions of the District of Columbia are noticeably free of crime incidents. Among other reasons, this may be due to those areas being green spaces where fewer incidents of robbery and theft-related crimes would be expected. Another reason is that there are many federal parks (Presidents’ Park, The National Mall, The Ellipse, etc.) in DC which are patrolled by US Park Police. I suspect that crimes handled by US Park Police are not included in the DC police department statistics.

This module was very extensive and very informative. Again, the value of GIS and the wealth of information that can be illustrated with GIS was demonstrated. Throughout this module, much experimentation was done to find optimal color schemes, labeling techniques, layer order, transparency, and so forth to determine the most optimal combination to illustrate the information clearly and completely. A feature that was especially invaluable for this particular analysis was the Identify feature. As in other modules, this experience has left me with the desire to know how to accomplish other things in ArcGIS.