خلاصة:
در سالهای اخیر آتش سوزی در نواحی رویشی رشد فزاینده ای داشته و آثار مخرب زیادی به جا گذاشته است. این مقاله به بررسی روش های پایش و پیش بینی آتش سوزی نواحی رویشی ایران و جهان پرداخته است. مطالعات مربوط به آتش‌سوزی در کشور ما از روش تحلیل سلسله مراتبی برای وزن‌دهی به فاکتورهای موثر و سیستم اطلاعات جغرافیایی در وقوع آتش‌سوزی جنگل‌ها استفاده شده است.در برخی مطالعات دیگر، از رگرسیون و الگوریتم درخت تصمیم‌گیری برای انتخاب متغیرهای موثر در آتش‌سوزی و همچنین مدل‌سازی خطر آتش‌سوزی استفاده شده است و در روش‌های پیشرفته‌تر از تلفیق سیستم استنتاج فازی و شبکه عصبی، هوش مصنوعی و ماشین بردار پشتیبان برای پیش‌بینی آتش‌سوزی‌های آینده استفاده شده است. استفاده از تصاویر محصول اتش در ایران از جمله تحقیقات نوپا محصوب می شود . که مزیت آن نسب به روش های دیگر این است که در عین ارزان بودن، بسیار سریع نیز هستند، و نتایج حاصل از آن نیز قابلیت به روز شدن بالاتری دارند.
نتایج بررسی های انجام‌شده در کشورهای دیگر نشان می‌دهد که اغلب نوع پوشش گیاهی، شیب، جهت جغرافیایی، فاصله از جاده‌ها، توپوگرافی و کاربری اراضی، موثرترین فاکتورها در پایش وقوع حریق جنگل ها بوده‌اند و ادغام لایه‌ها معمولا بر اساس سلسله مراتب و ضریب خطر در وقوع آتش‌سوزی انجام شده است. از فعالیت های قابل ذکر دیگر کشور ها،تهیه نقشه پایش و پیش بینی آتش سوزی های فعال به جهت شناسایی به موقع آتش سوزی با تصاویر ماهواره ی و شاخص های طیفی مناسب می باشد که نتایج مطلوبی را در جهت تصمیم گیری بهینه برای اطفا و جلوگیری از اتش سوزی را به همراه داشته است. شاخص سیستم رتبه بندی خطر آتش سوزی ایالات متحده نسبت به دیگر روش های مشابه در پایش اتش سوزی جنگل ها و مراتع را شاخصی جامع و مطلوب مطرح می کند.
In recent years, fires in vegetation areas have increased and have left a lot of destructive effects. Due to the fact that forest fires are affected by various factors. Therefore, this article examines the methods of monitoring and predicting fires in vegetative areas of Iran and the world. A review of research conducted in Iran shows that studies related to fire in our country have used the method of hierarchical analysis to weigh the effective factors and GIS in the occurrence of forest fires. The selection of effective variables in fire as well as fire risk modeling has been used and in more advanced methods the combination of fuzzy inference system and neural network, artificial intelligence and support vector machine has been used to predict future fires. Satellite images and remote sensing have been discussed. It should be noted that the use of fire product images in Iran is one of the nascent researches. Its advantage over other methods is that while they are cheap, they are also very fast, and the results are more up-to-date.
The results of studies conducted in other countries show that most types of vegetation, slope, geographical direction, distance from roads, topography and land use have been the most effective factors in monitoring forest fires and the integration of layers is usually based on hierarchy and risk factor in The occurrence of fires has been done. Also, indicators that are effective from environmental and climatic factors have been used in fire monitoring and forecasting, from the notable activities of other countries, preparation of fire monitoring and forecasting map It is active for timely detection of fires with satellite images and spectral indicators, which has resulted in favorable results for optimal decisions to extinguish and prevent fires. Researchers in recent research have examined the evaluation of monitoring indicators. They set fire to the U.S. Fire Risk Rating System as a comprehensive and desirable indicator compared to other similar methods in monitoring forest and rangeland fires.
Summary of research on fire potential detection in different parts of the world shows that GIS is very effective for developing information, management and forecasting of forest fire activities, so that creating a database in GIS consisting of variables affecting the occurrence of fire risk areas. The occurrence of forest fires is very useful. In most cases, vegetation type, slope, geographical direction, distance from permanent roads and rivers, topography and land use are the most effective factors in the occurrence of fire and the integration of layers is usually based on the hierarchy and risk factors in the occurrence of fire. In other studies, the fuzzy hierarchical analysis method, in other words, the combination of hierarchical analysis and fuzzy sets has been used to model the risk of fire. Satellite TM images have also been used to identify past fires. The general results of these studies show that in hot climates with dry vegetation, high slope, south direction and close to roads and residential areas, the potential for fire risk is high. To evaluate the accuracy of the method and model used in preparing the fire potential map, the map of critical fire areas is usually compared with the map of areas that have caught fire in the past, and if the two match, the model used seems desirable. In some other studies, a combination of fuzzy analysis system and neural network, artificial intelligence and support vector machine has been used to predict fire behavior management. Also, in more advanced methods, logistic regression and decision tree algorithm have been used to select and map effective variables in fire as well as fire risk modeling. Other studies have predicted the risk of fire and prepared a fire risk map by satellite imagery and helped to improve and correct the correct detection of satellite imagery by changing existing algorithms. And in their latest study, the combination of fire weather indicators with satellite imagery helped to detect the occurrence of fire in a timely and accurate manner.
Studies related to the potential of fire in Iran have been mostly in the field of zoning of fire risk from various factors and GIS as well as hierarchical analysis. To predict the occurrence of forest fires, artificial neural networks and climate data have been used. Predictions of fires in forests and pastures have also been made. And in the latest studies, they examined satellite fire products for accurate and timely forecasting. The results show that Madis sensor images have potential and good potential in fire detection and monitoring, and due to the newness of research in this field, there is a lot of activity. Each of the methods of modeling and fire risk assessment that were expressed in this study, advantages And have their own disadvantages; The use of indigenous models for fire hazard mapping definitely results in better results than non-indigenous models, so to use any method to improve the validity of models and methods of fire prediction and monitoring, all environmental and climatic factors suitable for the indigenous Consider instrumentation.
(
, 
, 
)
