⚡ electrical engineering

Assessing Urban Resilience to High-Impact Security Threats Using Bayesian-Fuzzy GIS Analysis: Evidence from District 5 of Tehran

This paper proposes and validates a novel integrated Bayesian-Fuzzy GIS approach to assess urban resilience against high-impact security threats at the parcel level, demonstrating through a case study of Tehran's District 5 that combining target attractiveness analysis with resilience mapping effectively identifies critical vulnerabilities and prioritizes interventions to significantly improve urban resilience.

Original authors: Mahdi Bitarafan

Published 2026-07-23

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Mahdi Bitarafan

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a city not just as a collection of buildings and roads, but as a living organism with a nervous system. Sometimes, this organism faces sudden, violent shocks—like a natural disaster or a security threat. Urban resilience is the city's ability to take a hit, keep its heart beating, and bounce back to normal without collapsing. Think of it like a superhero's durability: can the city absorb the punch, stay standing, and keep the lights on?

For a long time, scientists have studied how cities handle nature's wrath, like earthquakes or floods. But there's a different kind of danger: high-impact security threats. These aren't accidental; they are intentional attacks designed to cause chaos, damage, or fear. The tricky part is that not every part of a city is equally at risk. Some spots are like giant magnets for trouble because they are crowded, symbolic, or full of important people. This is called target attractiveness. If you know where the "magnets" are, and you know which parts of the city are "wobbly" (low resilience), you can figure out exactly where to build a shield. This paper dives into that exact puzzle, using a mix of math, expert opinions, and computer maps to see how to make a city tougher against intentional attacks.

The Detective Work: Mapping Tehran's Weak Spots

In this study, a researcher named Mahdi Bitarafan acts like a high-tech detective trying to solve a mystery in District 5 of Tehran. The goal? To create a super-detailed map that shows exactly which tiny pieces of land (called "parcels") are most likely to be targeted by bad actors and which ones are too weak to handle the shock.

Instead of guessing, the author uses a clever three-step recipe to bake the perfect resilience map:

The "Magnet" Test (Bayesian Analysis): First, the study asks, "Which places do attackers love?" By mixing old records of past attacks with the opinions of 14 experts (including engineers, planners, and crisis managers), the study calculates a "target attractiveness" score. It turns out that places like police stations, religious sites, military bases, and crowded public squares are the biggest magnets. Interestingly, even though crowded public areas have a lot of historical attacks, experts rated police and religious sites as even more attractive targets due to their symbolic power.
The "Weight" Test (Fuzzy SWARA II): Next, the study asks, "What makes a city strong?" It's not just one thing. The experts had to weigh different factors like land use, population density, building design, and how easy it is to get to a hospital or fire station. Because human opinions can be fuzzy (like saying "this is sort of important"), the study uses a special math tool called Fuzzy SWARA II to handle that uncertainty. The result? Land use (what a building is used for) was the most important factor, followed closely by population and building architecture.
The "What-If" Test (Monte Carlo Simulation): To make sure the experts didn't just get lucky with their rankings, the study ran a computer simulation 50,000 times, slightly shaking up the numbers each time. The result was reassuring: the top three factors (Land Use, Population, and Building Design) stayed on top almost every single time. This means the findings are robust, not just a fluke.

The Map: Before and After the Fix

Once the math was done, the researcher drew two maps of District 5.

The "Before" Map (Baseline):
This map showed the city's current state without any new security measures. The results were a bit worrying. Only about 8.86% of the area had a "very high" level of resilience, and 29.17% was "high." That means a huge chunk of the district—about 31.43% (the sum of "low" and "very low")—was sitting in a fragile state, ready to crumble if a major threat hit. The most dangerous spots were the ones that were both highly attractive to attackers (like a busy market or a government building) AND low in resilience (maybe because the roads are too narrow for ambulances or the buildings are old).

The "After" Map (Post-Intervention):
Here is where the story gets exciting. The researcher didn't just point out the problems; they simulated what would happen if the city made specific, smart changes. Imagine putting up movable metal poles (bollards) to stop cars from ramming into buildings, adding flower boxes that double as barriers, installing more cameras, and making sure emergency exits are clear.

When these changes were applied to the computer model, the city transformed:

The area with very high resilience skyrocketed from 8.86% to 38.34%.
The area with high resilience grew to 30.79%.
Together, the "safe" zones jumped from 38.03% to a whopping 69.13%.
Meanwhile, the dangerous "low resilience" zones shrank dramatically, dropping from 23.99% down to just 3.53%.

The Takeaway: Why This Matters

The study suggests that you don't need to rebuild the whole city to make it safer. By focusing on the specific "weak links"—the parcels that are both attractive targets and structurally weak—you can make a massive difference.

The paper rules out the idea that we should just look at one thing, like how many people live there or how many police stations exist. Instead, it proves that you have to look at the combination of the target's appeal and the area's ability to survive. It also shows that while some areas can be fixed quickly with better design (like adding bollards or cameras), the "very low" resilience areas (which barely changed in the simulation) might need much bigger, long-term fixes like redesigning the whole street layout.

In short, this paper gives city planners a playbook. It says: "Don't guess where to put your security budget. Use this map to find the spots where the danger is highest and the protection is lowest, and fix those first." It turns the scary, complex idea of urban security into a clear, step-by-step guide for making cities tougher, one block at a time.

Technical Summary: Assessing Urban Resilience to High-Impact Security Threats Using Bayesian-Fuzzy GIS Analysis

Problem Statement
Urban areas face emerging high-impact security threats—intentional acts such as explosions, attacks on critical infrastructure, and aggressive events targeting crowds—that challenge the continuity of infrastructure, emergency response capabilities, and social stability. While significant research exists regarding urban resilience to natural disasters and infrastructure failures, there is a notable gap in evaluating resilience to intentional security threats at a fine spatial resolution. Existing approaches often treat target attractiveness (susceptibility to attack) and facility vulnerability (capacity to withstand impact) as separate concerns, or they focus on individual buildings rather than district-level spatial patterns. Furthermore, traditional scoring methods often fail to account for the uncertainty inherent in expert judgments and the complex interplay of physical, social, and institutional factors in dense urban environments. This study addresses the need for a spatial decision-making framework that integrates target attractiveness with resilience capacity at the parcel level to prioritize interventions effectively.

Methodology
The research proposes an integrated Bayesian-Fuzzy GIS approach, validated through a case study of District 5 in Tehran, a heterogeneous urban area containing diverse residential, commercial, transport, and institutional functions. The methodology proceeds through four distinct stages:

Bayesian Target Attractiveness Analysis: To quantify the sensitivity of different urban land uses, the study synthesizes historical incident data with expert opinions. Using a Bayesian evidence synthesis model, historical priors are combined with expert-rated attractiveness scores and confidence levels. A balanced weighting factor ( $\lambda = 0.50$ ) ensures equal consideration of historical evidence and expert judgment to generate a Bayesian Target Attractiveness Index (BTAI) for various target categories.
Fuzzy SWARA II Weighting: To determine the relative importance of resilience criteria under conditions of uncertainty, the Fuzzy SWARA II (Step-wise Weight Assessment Ratio Analysis) method is employed. A panel of 14 interdisciplinary experts (civil engineers, urban planners, crisis managers, and urban managers) evaluated criteria across physical, social, emergency accessibility, and organizational dimensions. Triangular Fuzzy Numbers (TFNs) were used to handle linguistic variables and vagueness in expert judgments.
Robustness and Sensitivity Analysis: The stability of the derived weights is verified using Monte Carlo simulations (50,000 iterations). Random perturbations ( $\pm 10\%$ ) are applied to the initial weights to test the rank stability of the criteria and sub-criteria, ensuring the weighting structure is not overly sensitive to minor fluctuations in expert opinion.
GIS-Based Spatial Mapping and Intervention Modeling: The standardized weights are applied within a Geographic Information System (GIS) to create a Baseline Resilience Map at the parcel level. This map categorizes spatial units into five resilience classes (Very Low to Very High). Subsequently, a Post-Intervention Map is generated by simulating the impact of specific physical, managerial, and design interventions (e.g., movable bollards, protective urban furniture, improved emergency access, CCTV coverage) on the relevant GIS layers.

Key Contributions
The paper makes three primary contributions to the field of urban security and resilience:

Parcel-Level Spatial Resolution: Unlike previous studies focusing on buildings or broad districts, this approach assesses resilience at the parcel level, allowing for the detection of micro-variations in vulnerability and capacity.
Integration of Attractiveness and Resilience: The study uniquely combines a Bayesian analysis of target attractiveness (threat sensitivity) with a Fuzzy-SWARA weighted resilience index (capacity to absorb shock). This dual-layer approach prevents the misallocation of resources to either highly attractive targets that are already resilient or low-resilience areas that are not significant targets.
Evidence-Based Intervention Prioritization: By comparing baseline and post-intervention maps, the framework provides a quantitative basis for prioritizing specific planning measures. It moves beyond diagnostic assessment to demonstrate the potential efficacy of targeted interventions.

Results

Target Attractiveness: The Bayesian analysis identified police/law enforcement, religious/symbolic, military/security, and governmental/institutional facilities as the most attractive targets for high-impact security threats. While "civilian and crowded public targets" had high historical priors, expert judgment elevated the ranking of symbolic and institutional targets due to their psychological and political impact.
Resilience Criteria: The Fuzzy SWARA II analysis determined that Land Use (0.185), Population Characteristics (0.148), and Building Architecture (0.122) are the most critical criteria for urban resilience. The Monte Carlo analysis confirmed high robustness, with the top three criteria maintaining their rank in 96.9% of simulations.
Baseline Resilience: In District 5, the baseline assessment revealed significant spatial heterogeneity. Only 8.86% of spatial units exhibited "Very High" resilience, while 31.43% fell into "Low" or "Very Low" categories. Moderate resilience was the most common class (30.54%).
Intervention Impact: The simulation of targeted interventions resulted in a substantial improvement in the district's resilience profile. The combined share of "High" and "Very High" resilience areas increased from 38.03% (baseline) to 69.13% (post-intervention). Conversely, the "Low" resilience category decreased from 23.99% to 3.53%. However, the "Very Low" category remained largely unchanged (7.44% to 7.50%), indicating that some areas face structural limitations requiring more radical, long-term planning solutions beyond immediate protective measures.

Significance and Claims
The paper claims that urban resilience to high-impact security threats is a multivariate spatial notion that cannot be addressed through single-aspect analysis or protection-focused approaches alone. By integrating Bayesian evidence synthesis, Fuzzy logic, and GIS, the study demonstrates that resilience planning can be transformed from a general diagnostic tool into a precise mechanism for intervention prioritization.

The authors assert that the most critical planning focus should be on spatial units where high target attractiveness intersects with low or very low resilience. These areas represent the highest risk and require immediate, comprehensive interventions. The study concludes that while physical and design interventions (such as bollards, surveillance, and emergency access improvements) can significantly enhance resilience in many areas, persistent "Very Low" resilience zones require long-term structural reorganization, including land-use planning and infrastructure renewal. The proposed framework offers a scalable, evidence-based tool for urban managers to allocate scarce resources effectively, balancing immediate security needs with long-term urban development goals.

The Detective Work: Mapping Tehran's Weak Spots

The Map: Before and After the Fix

The Takeaway: Why This Matters

More like this