Search

Burned area is a critical input to biomass burning carbon emissions algorithms and for understanding variability in fire activity due to climate change. This study presents the spatial and temporal patterns of wildland fires in the Mongolian Plateau (MP) using Collection 6 NASA MCD64A1 500 m global Burned Area product from 2001 to 2021. Both inter- and intra-annual fire trends and variations in two subregions, Mongolia and China's Inner Mongolia, were analyzed. The results indicated that an average area of 1.3 × 104 km2 was consumed by fire per year in the MP. The fire season has two peaks: spring (March, April, and May) and autumn (September, October, and December). The profiles of the burnt area for each subregion exhibit distinct seasonality. The majority of wildfires occurred in the northeastern and southwestern regions of the MP, on the border between Mongolia and China. There were 2.7 × 104 km2 of land burned by wildfires in the MP from 2001 to 2021, 57% of which occurred in spring. Dornod aimag (province) of Mongolia is the most fire-prone region, accounting for 51% of the total burned area in the MP, followed by Hulunbuir, at 17%, Sukhbaatar, at 9%, and Khentii at 8%. The changing patterns of spatiotemporal patterns of fire in the MP were analyzed by using a spatiotemporal cube analysis tool, ArcGIS Pro 3.0.2. The results suggested that fires showed a decreasing trend overall in the MP from 2001 to 2021. Fires in the southern region of Dornod aimag and eastern parts of Great Xing'an Mountain showed a sporadic increasing trend. Therefore, these areas should be priorities for future fire protection for both Mongolia and China. [ABSTRACT FROM AUTHOR]

Modeling forest fire spread is a very complex problem, and the existing models usually need some input parameters which are hard to get. How to predict the time series of forest fire spread rate based on passed series may be a key problem to break through the current technical bottleneck. In the process of forest fire spreading, spread rate and wind speed would affect each other. In this paper, three kinds of network models based on Long Short-Term Memory (LSTM) are designed to predict fire spread rate, exploring the interaction between fire and wind. In order to train these LSTM-based models and validate their effectiveness of prediction, several outdoor combustion experiments are designed and carried out. Process data sets of forest fire spreading are collected with an infrared camera mounted on a UAV, and wind data sets are recorded using a anemometer simultaneously. According to the close relationship between wind and fire, three progressive LSTM based models are constructed, which are called CSG-LSTM, MDG-LSTM and FNU-LSTM, respectively. A Cross-Entropy Loss equation is employed to measure the model training quality, and then prediction accuracy is computed and analyzed by comparing with the true fire spread rate and wind speed. According to the performance of training and prediction stage, FNU-LSTM is determined as the best model for the general case. The advantage of FNU-LSTM is further demonstrated by doing comparison experiments with the normal LSTM and other LSTM based models which predict both fire spread rate and wind speed separately. The experiment has also demonstrated the ability of the model to the real fire prediction on the basis of two historical wildland fires. [ABSTRACT FROM AUTHOR]