This is Not a Drill: In Richland County, the Foundation of Declared Disaster Response and Recovery is FOSS4G

Status:
Accepted

It is best to prepare for the disaster that never happens. Richland County, SC did not have the luxury of practice before disaster struck. However, during a recent flood event, previous migration to a cloud-based open source stack for data dissemination provided real-time situational awareness for first responders, recovery teams, County staff, and citizens. In addition, the geospatial contributions of corporate citizens were executed efficiently using open source imagery processes. This presentation will describe the technology and processes involved in response and recovery during the flood event in Richland County during the first weeks of October, 2015.

Over the past 8 years, Richland County, SC has been systematically moving external operations toward open source technologies. With business users growing impatient with slow internet map queries bouncing between external and internal resources, the GIS team established an open geostack on a cloud platform. The initial stack was limited to Ubuntu/Postgresql/PostGIS/GeoServer feeding web mapping services. As the implementation progressed, the GIS team developed a variety of browser-based applications, first against the Google Maps JavaScript API, but later favoring Leaflet reinforced by a few additional libraries; namely Wicket, Proj4JS, and JavaScript Topology Suite. A PHP service layer of common web endpoints, including Parcel and Address lookup, completed the fully-open technology stack. The efficiency and effectiveness of this external platform has provided the County with a solid base from which additional apps and services are managed. The effectiveness of this work was tested and proven in October, 2015.

After weeks of extended rain, hurricane Joaquin pushed an unprecedented amount of rain toward central SC. This historic precipitation event of unfathomable flooding resulted in a presidential declaration of affected counties in South Carolina as disaster areas. With 17 deaths attributed to this single event, the calculation of property damage continues. As rescue efforts began, the County GIS team established a GEO EOC to manage internal and external requests. Simultaneously and without request, corporate citizens at Woolpert sent aircraft resources to the area to collect aerial imagery of the flood event. The Woolpert team worked diligently to acquire, process, and publish a high-resolution imagery product that captured the current situation on the ground. By leveraging a state-of-the-art imagery processing pipeline, Woolpert was able to turn an imagery product around less than 24 hours after collection. This pipeline relies heavily on open source imagery and GIS tools such as GDAL, PROJ.4, & OGR just to name a few. Additionally, by combining these technologies with our custom image pyramiding technology, we are able to perform hours of computational processing in mere minutes, and ultimately deliver fresh imagery to the public when it is needed most.

As the unprecedented amount of water began destroying infrastructure, road and dam failures threatened many lives and much property. Thus, in the GEO EOC, the team focussed on quickly developing and deploying a mapping application streamlined for situational awareness. The online map included data being served from the state agency responsible for dam monitoring and safety (SC DHEC), County imagery stores in a Google platform (GME), County vector data published through our cloud stack (chiefly Geoserver) and imagery from Woolpert’s stack. The vector data from SC DHEC and Richland County was updated many times each day to reflect road and dam status. These elements were also pushed into the County 911 system for dispatch and routing. With the fully-open web stack in place, the developers were able to move fast by cannibalizing their most recent application and standing up a new internet map featuring the most relevant data.

In addition to a before-and-after ‘swipe’ app and Flood15 mapping tool, the County also relied heavily on a now DEAD tool from Google (Maps Engine). The County’s GME platform was used to rapidly deploy road and dam updates for responders and citizens in a universally recognized format (Google Maps). Although the resulting tools were different and were developed for separate audiences, the process by which data were updated was centrally organized and controlled with strict procedure. Many sources of ‘ground’ information were funneled to our GEO EOC and processed and verified prior to publishing. Non-traditional methods were incorporated into this process with surprisingly effective results.

Slides (External URL)

Session details
Schedule info
Session Time Slot(s):
Thursday, May 5, 2016 - 15:00 to 15:35

Comments