Towards multi-sensor operational monitoring of the European fire regime
| Autor: | Pieter Beck, Lorenzo Busetto, Jesus San Miguel-Ayanz, Roberto Boca, Francesco Boccacci |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2014 |
| Zdroj: | 2014 AGU Fall Meeting, San Francisco (USA), 15-19 December 2014 info:cnr-pdr/source/autori:Pieter Beck, Lorenzo Busetto, Jesus San Miguel-Ayanz, Roberto Boca, Francesco Boccacci/congresso_nome:2014 AGU Fall Meeting/congresso_luogo:San Francisco (USA)/congresso_data:15-19 December 2014/anno:2014/pagina_da:/pagina_a:/intervallo_pagine |
| Popis: | Europe loses approximately half a million hectares of forest to human-caused fires annually. High population densities and fuel loads in areas of forest recovery aid uncontrolled wildfires, despite decreasing use of fire in agriculture. In rural areas such as the Mediterranean, pervasive wildfires can be indicative of socio-economic decline and themselves cause political tension. Since the 2000s, the European Forest Fires Information System maps forest fires in Europe daily, as part of a modular system that also assesses fire risk. This system, developed and maintained by the European Commission, feeds pan-European information to forest fire and civil protection agencies in the EU and provides a consistent quality-checked forest fire record for scientific analysis and global reporting initiatives. The current fire mapping algorithm captures fires larger than 50 ha, owing to its reliance on MODIS imagery. As a result, the system misses about 95% of Europe's forest fires, which contribute 20-25% of the annually burned area, because they are too small for detection. We present a new approach to fire mapping that relies on multiple satellite sensors to improve spatiotemporal detail of mapped fire activity. We implemented the approach to map European wildfires exploiting the 30 m spatial resolution of the Landsat 8 sensor, and the daily overpass of the MODIS sensors. By doing so, detections of smaller fires improved dramatically, although they were somewhat delayed. The algorithm relies on a set of heuristic rules that translate knowledge of the temporal dynamics of fire disturbance, recovery, and confounding factors such as agricultural harvest, into spectro-temporal criteria. By design, it can be adapted to areas with different fire regimes, and ingest information from additional satellite sensors to further improve detection rates and times. We discuss the latter in the context of the Sentinel missions, and their potential contribution to global wildfire monitoring. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|