1. C ONNECTING FOR A R ESILIENT A MERICA 1 Public- Private Partnerships: Key Drivers of Disaster Supply Chains Dr. Ramesh Kolluru Dr. Mark Smith

2. C ONNECTING FOR A R ESILIENT A MERICA 2 Public- Private Collaboration Partnership Public Sector Private Businesses 4 Owns 85% of CIKR assets Drives 98% Supply Chains

3. C ONNECTING FOR A R ESILIENT A MERICA 3 5 Louisiana Business Emergency Operations Center State Initiative To support disaster management in Louisiana by developing an accurate understanding of economic impacts to critical infrastructures and major economic drivers, as well as facilitating the coordination of businesses and volunteer organizations with the public sector through enhanced resource and information management. A Partnership Of

4. C ONNECTING FOR A R ESILIENT A MERICA 4 Public-Private Partnerships The NIMSAT Institute establishment of public-private partnerships through the LABEOC seeks to mitigate the risks we face as a nation to the intricate interdependencies between CIKR assets and the public and private sector supply chains that depend on these assets. Through the Louisiana BEOC, the State of Louisiana will : –Facilitate communication between public and private sectors –Enable business and industry to identify roadblocks to recovery –Reduce overall economic impacts to Louisiana businesses and industry –Mobilize government resources - Ex.- Fuel Supply-Demand Project *The long term goal would be to create and replicate this model in Louisiana, the Gulf Coast and the Nation.

5. C ONNECTING FOR A R ESILIENT A MERICA 5 Current Issues in Public- Private Partnerships Can government resources be utilized? What types of information are useful and can be shared? What are the motivators and roadblocks for such partnerships? Is there a governance structure that all sectors could utilize in engaging with their counterparts or other stakeholders in response? What is the capability maturity model? Modeling, Simulation and Analysis of Operational Capability for Simple and Complex Events

6. C ONNECTING FOR A R ESILIENT A MERICA 6 Challenges & Progress Information Sharing Between the Sectors  Government-to-Business Information Sharing  Business-to-Government Information Sharing ** The LABEOC has begun to engage these partners in willingness to participate Motivators or Roadblocks - Concerns between sectors - LABEOC facilitate these partnerships - Legal policy - Integrated monitoring systems ** The LABEOC managers have begun to coordinate and solve these issues. Since the LABEOC is a neutral party there is a receptiveness.

7. C ONNECTING FOR A R ESILIENT A MERICA 7 Challenges & Progress Capability Maturity Model - Complex Emergencies - Real Time Plans - Measuring performance between interactions ** LABEOC plans to engage all sectors through a information hub website to ensure coordinating communications Established private sector supply chain models already created - CPFR Model – (Collaborative Planning, Forecasting and Replenishment **Industry Standard framework in development

8. C ONNECTING FOR A R ESILIENT A MERICA 8 Fuel Supply-Demand Project To enhance Energy sector resiliency through improved information sharing from the “platform to the pump” that enable communications across the energy supply chain to facilitate processing delivery, and distribution, while enhancing the State Energy Profile. In collaboration with Department of Energy & Louisiana Department of Natural Resources

9. C ONNECTING FOR A R ESILIENT A MERICA 9 Current Development SmartGrid policy Energy Supply-Demand Model Evacuation Fuel Plan Improved LABEOC measures for monitoring energy sector Conduct regional exercises that include the Gulf Coast Region

10. C ONNECTING FOR A R ESILIENT A MERICA 10 Fuel Supply Grant Transportation/Fuel Demand Model –How much fuel is needed to support an evacuation? Behavioral Social Network Model –When will people evacuate? (shadow evacuations?) –Where are they likely to go? Fuel Supply System Model –What is the fuel availability at gas stations? Results for Decision Support –Can we mobilize adequate emergency fuel, if needed?

11. 11 C ONNECTING FOR A R ESILIENT A MERICA 11 Private Industry Response to Gustav/Ike Public-Private Partnerships –Mobilized business products and services: $23.8 million dollars –Enhanced Situational Awareness: shortages of fuel CIKR Consequence Models –Reported disruptions to operating capacity of 120 petroleum, natural gas, chemical and electricity facilities (CITGO Refinery, Entergy, Henry Hub, LOOP, Port Fourchon, etc.) –Economic impact to Oil & Gas industry: $7.6B - $8.3B

12. C ONNECTING FOR A R ESILIENT A MERICA 12 Future Research Areas LABEOC has begun research on partnerships Ability to integrate systems and plans between sectors State Emergency Center communication Facilitation on all levels (legal issues, monitoring, planning, etc.) Possible merging of operational systems Regional cooperation Energy sector and Fuel Supply- Demand project

13. C ONNECTING FOR A R ESILIENT A MERICA 13 C ONNECTING FOR A R ESILIENT A MERICA Public- Private Partnerships: Key Drivers of Disaster Supply Chains Dr. Ramesh Kolluru Dr. Mark Smith

14. Ronald T. Eguchi Paul Amyx Charles K. Huyck ImageCat, Inc. www.virtualdisasterviewer.com DHS – S&T Workshop on Emergency Management: Incident, Resource, and Supply Chain Management 5-6, November 2009, UCI, Irvine, CA

15. Presentation Outline The Problem VDV interface for damage assessment Outreach Future activities Summary

16. There is a post-disaster need… … for expert analysis to provide rapid and accurate commentary after significant events … to harness the skills of experts from a wide variety of industries, affiliations, and locations … for a central repository for experts’ interpretations to improve dissemination and discussion of data from current and past disasters After major disasters, it is not always feasible to deploy field teams due to damage and/or access restrictions Multiple field deployments can result in duplication of effort and costs Data is not shared from multiple field teams

17. Korea Japan Philippines Hong Kong Beijing Tibetan Plateau Shanghai Wenchuan Earthquake Chengdu 2008 Wenchuan Earthquake Details (known): May 12, 2008, 2:28pm (Local time) Magnitude: 7.9 (USGS) Location: 30.986°N, 103.364°E Depth: 19km Fault length: approx. 250km Details (unknown): Number of ? dead (missing) ? injured ? buildings collapsed ? buildings damaged ? homeless ? evacuated ? people affected ? Economic loss Damage Estimates: 69,200 dead (20,000 missing) 374,200 injured 5.4 million buildings collapsed 21 million buildings damaged 5 million homeless 15 million evacuated 46 million people affected Economic loss US$80 billion? UNDMT Situation Report No. 8, 14 June 2008

18. VDV Interface for Damage Assessment The login procedure Accepting an assignment Scope of evaluation The notion of a damage scale based on remote-sensed images Damage assessment procedure Results from 2008 Sichuan, China earthquake

19. Figure 1. Virtual Disaster Viewer homepage Virtual Earth navigation controls Virtual Earth base data and imagery Details of project sponsors and participating organizations Pre- and post- disaster satellite imagery Derived layers Field data Legend for expert interpretation results Major sponsors and affiliations www.virtualdisasterviewer.com

22. Grid Cells for Analysis

26. Remote sensing damage scale Damage levelDescriptionExample pre - earthquakeExample post - earthquake RSE - 0 Indistinguishable a) Pre disaster shows building, post- disaster is homogenous cleared area with no evidence of debris, suggesting demolition prior to earthquake. b) Partially built building; (under construction site at the time of the earthquake). c) Building is under cloud cover, or is otherwise obscured by another imagery artefact. RSE - 1 Non / Slight a) No damage seen within image. Pre- and post-event images same shape, size, colour (. b) Slight damage (i.e. tiles from roof small amount of debris) c) Equated to EMS98 level 1 or 2 RSE - 2 Extensive a) Extensive damage seen. b) Possible changes include: Change in building footprint shape. Roof collapsed. Heterogeneous roof colour/texture Substantial rubble. c) New builds that are extensively damaged d) Equated to EMS98 level 3 or 4 RSE - 3 Collapse a) Building collapsed b) Copious rubble evident c) New build which is completely collapsed d) Equated to EMS98 level 5. Remote Sensing for Earthquake Scale (RSE) Damage Scale Descriptions

27. Assigning Damage Level Pre Post

28. Results BUILDING DAMAGE

29. Results LANDSLIDE EXTENT

30. Initial funding from EPSRC (UK), EERI, MCEER (USA) Tool developed by ImageCat Currently being developed for inclusion in EEFIT missions to Indonesia & South Pacific 84 expert volunteers from 8 countries

31. Conferences & Workshops Wired Magazine article: April 2009 Imaging Notes Magazine article: Fall 2009 Flagship project for Community Remote Sensing theme: IEEE International Geoscience & Remote Sensing Symposium (IGARSS) 2010

32. Sumatra & Samoa currently being developed Validation of expert analysis & assessing experts’ skills – “superusers” Extend functionality may include –Video field blogs & text summaries of professional findings –Greater use of commenting facilities –Develop analytical tools for automatic interpretation of results

33. VDV developed to fill technological need for a post-disaster data portal Multiple functions developed according to needs of field teams on per-event basis Currently broadening VDV’s outreach for future funding and data partnerships A growing community of expert users and contributors are realizing VDV’s potential

34. Presented: 11/05/09http://teamcore.usc.edu Agent-based Evacuation Modeling: Simulating the Los Angeles International Airport Milind Tambe, Jason Tsai, Matthew E. Taylor, Shira Epstein, Andrew Ogden, Prakhar Garg University of Southern California Gal Kaminka, Natalie Fridman Bar Ilan University Emma Bowring University of the Pacific

35. Presented: 11/05/09Jason Tsaihttp://teamcore.usc.edu Emergency evacuation Training and policy decisions are difficult Scenario: Evacuation of an airport terminal after an event Ideal: Conduct live exercises  Personnel can see what actually happens  Policy-makers can try different rules Issue: Live exercises are difficult  Requires the terminal to be shut down  Requires realistic response of hundreds of people  Unethical to instill real fear/anxiety in people

36. Presented: 11/05/09Jason Tsaihttp://teamcore.usc.edu Evacuation simulation Simulations provide an answer Proposal: Simulation replaces live exercises  Evaluate different policies  Decision-making training for officer  Visual conditioning for officers We propose to build an agent-based model with realistic human behavior and compelling visualization

37. Presented: 11/05/09Jason Tsaihttp://teamcore.usc.edu Our approach We focus on unaddressed issues unique to the domain BDI-style architecture  Standard architecture style for agents Social Comparison Theory mechanics  General theory of how agents impact each other Transition to emergency Model the transition from normal to emergency behavior Massive software visualization  Movie-quality people seen at eye-level

38. Presented: 11/05/09Jason Tsaihttp://teamcore.usc.edu Previous work: Teamcore We have worked in simulations in the past DEFACTO system  Training / coordination tool for fire department incident commanders Robocup Rescue  Search and rescue simulation Helicopter team simulations  Helicopter attack strategy simulation

39. Presented: 11/05/09Jason Tsaihttp://teamcore.usc.edu Previous work: Academic Evac. Simulation Wealth of work in evacuation simulations BDI-style architecture Agent interactivity without underlying social theory Social Psychology simulations Apply theories limited to very specific activities Physics-based simulations Individual behaviors are generalized We propose to use a general psychological model that can model both ‘normal’ and ‘emergency’ behavior

40. Presented: 11/05/09Jason Tsaihttp://teamcore.usc.edu Previous work: Architectural planning Architectural simulations provide a high-level view Metrics are high-level (e.g., time to clear a building) Lacks validated realism in individual behavior Legion software

41. Presented: 11/05/09Jason Tsaihttp://teamcore.usc.edu Previous work: Cinema Cinema simulations focus on dramatic effect and believability Metrics involve believability, director’s desires, etc. Lacks validated realism in individual behavior Massive software

42. Presented: 11/05/09Jason Tsaihttp://teamcore.usc.edu Collaborations We have established key collaborations Los Angeles Airport Police and TSA  First step will be to develop a simulation of Terminal 1 Massive software USC School of Cinematic Arts Cross-institutional team Expertise in crowd simulation, disaster response simulation, and Social Comparison Theory

43. Presented: 11/05/09Jason Tsaihttp://teamcore.usc.edu Challenges Computational speed Hundreds of agents with complex decision mechanisms Parameters / Calibration Data for calibration is sensitive information Techniques for analyzing the data are imperfect Validation Scientific validation Expert user buy-in jasontts@usc.edu http://teamcore.usc.edu