Keywords
Burned Area; Deep Learning; Wildfires; Reconstruction
Start Date
6-7-2022 12:20 PM
End Date
6-7-2022 12:40 PM
Abstract
The absence of a global long-term burned area dataset significantly hampers analyses of longterm trends in wildfire impacts. This prevents conclusive statements on the role of anthropogenic activity on wildfire impacts over the last century. Here, we construct a 1901-2014 data-driven reanalysis of monthly global burned area at a 0.5° by 0.5° scale. A recurrent neural network is trained with weather-related, vegetational, societal and economic input parameters, and burned area as output label for the 1982-2014 time period. This model is then applied to the whole 1901-2014 time period to create a data-driven, longterm burned area reanalysis. This reconstruction allows to investigate the long-term effect of anthropogenic activity on wildfire impacts, will be used as basis for detection and attribution studies and can help to reduce the uncertainties in future predictions.
Reconstruction of 20th Century Burned Area through Recurrent Neural Networks
The absence of a global long-term burned area dataset significantly hampers analyses of longterm trends in wildfire impacts. This prevents conclusive statements on the role of anthropogenic activity on wildfire impacts over the last century. Here, we construct a 1901-2014 data-driven reanalysis of monthly global burned area at a 0.5° by 0.5° scale. A recurrent neural network is trained with weather-related, vegetational, societal and economic input parameters, and burned area as output label for the 1982-2014 time period. This model is then applied to the whole 1901-2014 time period to create a data-driven, longterm burned area reanalysis. This reconstruction allows to investigate the long-term effect of anthropogenic activity on wildfire impacts, will be used as basis for detection and attribution studies and can help to reduce the uncertainties in future predictions.
Stream and Session
false