Analysis of wildfire dynamics in the Himalayan region (2016-2025) using remote sensing and geographic information systems


Tezin Türü: Yüksek Lisans

Tezin Yürütüldüğü Kurum: Karadeniz Teknik Üniversitesi, Orman Fakültesi, Orman Mühendisliği, Türkiye

Tezin Onay Tarihi: 2025

Tezin Dili: İngilizce

Öğrenci: ZARAK KHAN AFRIDI

Danışman: Kadir Alperen Coşkuner

Özet:

This study examines the behavior of wildfires using the WFA Pocket Simulator across eight distinct events occurring throughout the Himalayan region, including Nepal, India, Bhutan, Pakistan, and Tibet. By utilizing VIIRS active fire data along with Sentinel-2 L2A imagery sourced from the Copernicus Data Space Ecosystem, we derived the Normalized Burn Ratio (NBR) and its differenced form (dNBR) to quantify the conditions before and after the fires and to outline the burned areas at a spatial resolution of 20 meters. From this data, the rate of spread (ROS) and the burned area were calculated for each individual fire. To provide context for how the fires spread, additional meteorological parameters such as air temperature, wind speed and direction, and relative humidity were sourced from ERA5's post-processed daily statistics. The model's performance was evaluated by comparing the simulated Rate of Spread (ROS) and burned area values to those observed, using bias and an adjusted coefficient of determination (R²). The results showed consistently low R² values, indicating a weak statistical correlation between the simulations and what was actually observed for all the fires. Overall, this study highlights how important local calibration and multi-parameter validation are when simulating wildfires in hilly regions.