Who is really at risk? Integrating social vulnerability and adaptive capacity into wildfire risk assessment in Pakistan

Abstract Background Wildfire risk assessments in developing countries remain largely hazard-centric, focusing on biophysical drivers such as fuel, weather, and topography while overlooking the socioeconomic and institutional conditions that determine who actually bear the consequences of a fire. Pakistan illustrates this gap: its four provinces span a wide range of ecological and socioeconomic conditions, yet no integrated framework has combined long-term fire observations with social vulnerability and adaptive capacity indicators at the provincial scale. We developed such a framework for Punjab, Sindh, Khyber Pakhtunkhwa (KPK), and Balochistan by combining 22 years of satellite-derived fire observations with socioeconomic and institutional indicators from the 2017 Pakistan Census. Results We constructed four components: a continuous vulnerability index (VI) based on illiteracy, disability, female population share, unemployment, and per capita income; a categorical vulnerability profile (VP) derived using a 60% threshold rule; a multidomain adaptive capacity profile (ACP) spanning health, education, information and communication technology (ICT), and institutional presence; and a fire hazard index (FHI) combining fire radiative power (FRP) and fire count (FC). We integrated these components through a nonlinear risk matrix that captures the amplifying effect of low adaptive capacity on social vulnerability. Balochistan was classified as very high risk, driven by extreme fire intensity (21.33 megawatts (MW)), an adaptive capacity score of 0.15, and high social vulnerability (60% illiteracy), despite having recorded the fewest fire detections of the four provinces. KPK emerged as high risk, Punjab as medium risk, and Sindh as low risk, the most resilient of the four provinces. Conclusions Our findings reveal that wildfire risk in Pakistan is shaped as much by who is exposed and how well they can respond as by fire intensity itself and that low-frequency, high-intensity fire regimes in institutionally under resourced provinces pose disproportionate risk despite modest fire activity. The framework is designed to be transferable to other data-scarce, fire-prone regions and offers a practical, replicable basis for prioritizing wildfire mitigation and adaptive-capacity investment.