Assessment of wildfires forecast performance in Albania: Case study

Abstract

The harmful impacts of climate change caused by wildfires are substantially harming the people of mainland Europe, as well as damaging species biodiversity and the ecosystem. It can be minimised by improving the effectiveness of fire risk forecasting and mitigation strategies. The aim of this paper was to investigate the accuracy of forest fire forecasts in Albania produced by the FWI (Fire Weather Index) system. During the summer of 2022, observations and data were collected on expected and actual fires in the prefectures of Albania, which were previously divided into four categories according to the level of fire risk: high, moderate, low, and zero. It was determined that, in the summer of 2022, Albania happened a grand total of 620 wildfires. The data were analysed using two indicators: the probability of fire occurrence for a particular prefecture and the number of fires per prefecture. The analysis revealed varying degrees of accuracy in fire predictions across different prefectures, with higher precision observed in high-risk regions but diminishing as the risk level decreased. The most dependable indicator of forecast accuracy, reaching 75%, was observed in high-risk areas during the month of August. Predicting fire localization within moderate-risk zones consistently achieved results above 50% but fell short of the 60% threshold. Overall, the results confirm the effectiveness of using data to predict the probability of fires for prefectures with a high and increased level of the relevant threat. This will make it possible to more effectively deploy and mobilise the resources needed to overcome them and substantially reduce the losses associated with them