Forest Fire Probability Prediction Taking Into Account Anthropogenic Load, Lightning Activity, and Weather Conditions

Introduction

Currently, the issue of forest fire danger assessment is relevant. The prediction of surface fires is of greatest importance, since more than 80% of all vegetation fires are surface fires. Almost all crown fires develop from surface fires (Bayham et al., 2020). An important role is played by the creation of a new methodology for predicting the occurrence of forest fires. The basis for the creation of such a methodology should be simple, but adequate to the physics of the process of forest fire ripening, mathematical models, as well as appropriate methodological, information and software.

An analysis of the forest fire danger prediction techniques existing in Russia and abroad shows that almost all the techniques have a weak physical basis and, as a rule, take only weather data into account. Thunderstorm activity and anthropogenic load are not properly taken into account (Grishin, 2002). It should be noted the Canadian, a number of South European methods (Viegas et al., 2000) and the Nesterov criterion (Nesterov, 1949). The advantages of these methods are the simplicity and a fairly good quality of the prediction, but within the territory on which a statistical analysis of forest fire incidents was carried out. The main disadvantage is the neglect of the real cause of forest fires, namely, the drying process of a layer of forest fuel on the underlying surface.

The existing network of weather stations (especially the regions of Siberia, the North and the Far East) does not allow us to talk about any normal coverage of the controlled territory. In this work, when developing a system for predicting the occurrence of forest fires, it is proposed to focus on the prospects for interaction with software that implements global and regional atmospheric models. It should be noted that the use of existing fire danger scales often leads to errors, since forest fire danger is actually estimated only by vegetation conditions (Telitsin, 1987). In addition, in the dry period, differences in the moisture content of forest fuel in the areas are smoothed out, and after 2-3 weeks there is practically no difference in the I and V classes (Telitsin, 1987). This information only confirms the conclusion that it is necessary to develop a new methodology for determining the probability of forest fires.

In the nuclear industry, for example, a probabilistic safety criterion is used (Grishin, 1999), and at present, the need to develop a similar criterion and the corresponding methodology has also arisen in forestry. An analysis of numerous literary sources devoted to the problem of forest fire danger predicting and forest taxation descriptions of specific forestry allows us to conclude that it is necessary to develop a forest fire danger prediction system that would have a spatial resolution at the level of the minimum forest taxation unit (site) and which would allow to obtain results in the system “forestry-quarter”', as the quarterly maps of forest areas, as a rule, do not have a geographic reference.

Simple mathematical models and an approximate analytical formula for determining the drying time of the forest fuel layer were previously developed (Grishin and Baranovskij, 2003). For practical use, the forest-taxation characteristics “completeness” and “bonitet” of the stand are integrated in the model (Baranovskiy and Grishin, 2002), which allows one to take into account the effect of screening of solar radiation by the canopy of the stand. In (Sofronov, 1970), it was noted that there is a dependence of the fraction of solar radiation penetrating under the canopy of the forest stand on these characteristics for various types of forests. It should be especially noted that for this it is enough to use the standard forest taxation descriptions that are available in each forestry.

The purpose of this work is to develop a new probabilistic criterion for forest fire danger and conduct a parametric study of the influence of anthropogenic load, thunderstorm activity and weather conditions on the probability of forest fires.