Spatially Layered Defenses Against Terrorism

Principal Investigator: Detlof von Winterfeldt


Year 11 (FY2015) was the second year of a project on spatially layered defenses against terrorist. The purpose of this research is to provide guidance to DHS policy makers on how to improve layered defenses.  The research was originally motivated by the PI’s involvement in a National Academy study on evaluating the effectiveness of the Global Nuclear Detection Architecture (GNDA), a major effort of the Domestic Nuclear Detection Organization (DNDO).  The GNDA employs a geographically layered defense strategy to prevent the theft and transportation of radiological and nuclear materials from a foreign source to a target in the US.  The GNDA concept of a layered defense is similar to the concept of “defense in depth” promoted by the Nuclear Regulatory Agency for nuclear power plant safety. Layered defenses occur in several DHS contexts, for example, in approaches to catching smugglers (in the foreign country, upon crossing the border, in the US) or in bioterrorism (stopping the production of biological materials, preventing the importation in the US, stopping the distribution and spread of the disease after a biological attack).  Modeling layered defense systems involves an extension of traditional probabilistic risk analysis methods (PRA, exemplified by BTRA and RNTRA) as well as innovations in risk analysis for networks (exemplied, for example, by PEM, DNDO’s Probabilistic Effectiveness Model). We will developed these advanced risk analysis models, first for the general case of layered defenses, and then apply them to assessing the effectiveness of features of the GNDA. During the first year of this project we reviewed the existing models for layered defenses used by the DNDO, in particular, the Radiological and Nuclear Terrorism Risk Analysis model (RNTRA) and the Probabilistic Effectiveness Model (PEM).  Some of the results of this review are described in a National Academy of Science report “Evaluating the performance measures and metrics for the Global Nuclear Detection Architecture,” parts of which were drafted by the PI.  Relevant to the current projects, we discussed strengths and weaknesses of the two models in the report and we argued for a more integrated modeling solution – which is the core of the current project.  We also reviewed other literature on layered defenses and search and detection games, for example by Bier and her colleagues and by Morton and we developed an annotated bibliography of these models   Building on existing models, we designed an integrated model that combines a transportation model (how to get nuclear materials from point A to point B) with a risk analysis model (how likely is the interception of these materials at various points along the way).  We have superimposed a decision and resource allocation model over this baseline model to examine optimal allocations of resources to maximize detections and increase deterrence of transporting RN materials from their source to a target in the US.    In February 2015 we learned of an interesting concept that could change the global nuclear detection architecture.  It consists of placing radiological isotope identification devices (RIIDs) on each container at the port of lading – long before it enters the United States.  This would increase the dwell time and thus improve detection and identification tremendously.  Before arrival at US points of entry, the RIIDs would automatically transmit information about any radiological materials in the container, thus making the current system of portals unnecessary.  Since then, we have begun to shift our effort to evaluating this concept. Keywords: Layered defenses, defense in depth, radiological and nuclear terrorism, Global Nuclear Detection Architecture (GNDA), Islamic State of Iraq and Syria (ISIS).