Spatial Prediction Of Human-caused Grassland Fire Ignition Risk In Hulunbeier Region

Grassland fire disaster is one of the most frequent disasters in the grassland of China,which always happens suddenly and makes destructive damages. It generally occurs in theareas with sparse populations, which makes the relief work more difficult. Grassland firedisaster poses a great threat to people’s lives and social development, has a direct impact onthe economic construction and social stability, and has seriously hampered the stable andhealthy development of China’s animal husbandry. At the same time, it also destroys grasslandresources and affects grassland ecosystem. The frequent grassland fires occurred in thetransition zone of grassland and forest are prone to cause forest fires, which can result inmuch greater losses. No matter home or abroad, fire science researchers study forest fire moredeeply than grassland fire, resulting in the lagging behind of grassland fire disaster research.Based on the formation principle of natural disaster risk and regional disaster systemtheory, this paper studied the impact of human activities on Hulunbeier grassland fire ignitionrisk. Firstly, this paper analyzed the spatial and temporal characteristics of the annualoccurrences and burned areas of grassland fire disaster in the study area during1995-2010.Using Gray series forecasting model, the Radial basis function (RBF) neural networkprediction model and Grey RBF neural network prediction model, the annual occurrences andburned areas of grassland fire disaster on time series were predicted. Then,17human activityindicators were selected, by Grey Relational Analysis, the relational degrees betweenindicators and the annual occurrences and burned areas of grassland fire disaster werecalculated. Based on the results above,5closely human activities related factors: ruralpopulation density, stocking density, housing estates distribution, urban areas, and roads, weretook into the establishment of Hulunbeier grassland fire ignition risk prediction model, usingweights of evidence method and GIS spatial analysis under the ArcGIS software. Accordingto the model established, a prediction map was generated with the help of the WofE extensionin ArcView, which divided the research area into the high, medium and low risk areas.Results indicated that Hulunbeier grassland fire ignition risk was closely related to thehuman access to the grassland, especially concerned with rural population density, roads andhousing estates distribution; it also proved that the weights of evidence could express therelationship between grassland fire ignition risk and human activity objectively andquantitatively, which would provide support for grassland fire risk management. The resultsnot only could provide information support for the enactment of the Ordinance of Hulunbeiergrassland fire disaster risk management, but also could improve the grassland fire researchsystem of China further.