Modelling of fire count data: fire disaster risk in Ghana

Global Wildfire Outlook Forecast with Neural Networks

Wildfire occurrence and spread are affected by atmospheric and land-cover conditions, and therefore meteorological and land-cover parameters can be used in area burned prediction. We apply three forecast methods, a generalized linear model, regression trees, and neural networks (Levenberg–Marquardt backpropagation) to produce monthly wildfire predictions 1 year in advance. The models are trained using the Global Fire Emissions Database version 4 with small fires (GFEDv4s). Continuous 1-year monthly fire predictions from 2011 to 2015 are evaluated with GFEDs data for 10 major fire regions around the globe. The predictions by the neural network method are superior. The 1-year moving predictions have good prediction skills over these regions, especially over the tropics and the southern hemisphere. The temporal refined index of agreement (IOA) between predictions and GFEDv4s regional burned areas are 0.82, 0.82, 0.8, 0.75, and 0.56 for northern and southern Africa, South America, equatorial Asia and Australia, respectively. The spatial refined IOA for 5-year averaged monthly burned area range from 0.69 in low-fire months to 0.86 in high-fire months over South America, 0.3–0.93 over northern Africa, 0.69–0.93 over southern Africa, 0.47–0.85 over equatorial Asia, and 0.53–0.8 over Australia. For fire regions in the northern temperate and boreal regions, the temporal and spatial IOA between predictions and GFEDv4s data in fire seasons are 0.7–0.79 and 0.24–0.83, respectively. The predictions in high-fire months are better than low-fire months. This study illustrates the feasibility of global fire activity outlook forecasts using a neural network model and the method can be applied to quickly assess the potential effects of climate change on wildfires.

The dual effect of vegetation green-up date and strong wind on the return period of spring dust storms

Vegetation phenology changes have been widely applied in the disaster risk assessments of the spring dust storms, and vegetation green-up date shifts have a strong influence on dust storms. However, the effect of earlier vegetation green-up dates due to climate warming on the evaluation of dust storms return periods remains an important, but poorly understood issue. In this study, we evaluate the spring dust storm return period (February to June) in Inner Mongolia, Northern China, using 165 observations of severe spring dust storm events from 16 weather stations, and regional vegetation green-up dates as an integrated factor from NDVI (Normalized Difference Vegetation Index), covering a period from 1982 to 2007, by building the bivariate Copula model. We found that the joint return period showed better fitting results than without considering the integrated factor when the actual dust storm return period is longer than 2years. Also, for extremely severe dust storm events, the gap between simulation result and actual return period can be narrowed up to 0.4888years by using integrated factor. Furthermore, the risk map based on the return period results shows that the Mandula, Zhurihe, Sunitezuoqi, Narenbaolige stations are identified as high risk areas. In this study area, land surface is extensively covered by grasses and shrubs, vegetation green-up date can play a significant role in restraining spring dust storm outbreaks. Therefore, we suggest that Copula method can become a useful tool for joint return period evaluation and risk analysis of severe dust storms.