Uydu görüntü verileri kullanılarak orman yangın analizi

Abstract

Doğal dengenin tesisinde kritik öneme haiz nitelikteki ormanların, çevresel ve ekolojikfaydaları itibariyle ciddi biçimde korunması gerekmektedir. Orman yangınları etkilerive ortaya çıkardığı sonuçları bakımından tüm dünyada en vahim doğal afetlerin başınıçekmektedir. Dünya genelinde orman yangınlarından bahsedildiğinde, öncelikli olarakAkdeniz ülkeleri, Avustralya, Kanada ve Amerika Birleşik Devletleri hatıragelmektedir. Söz konusu ülkeler yangınlarla mücadele açısından başarılı sayılabilir.Bu başarıda, yangın tehlikesinin gerçekçi tahmini, doğru bir şekilde oluşturulmuşplanlama sistemleri ve tehlikenin olduğu yerlere acil müdahale imkânların da rolübüyüktür. Türkiye'de, orman yangınlarının oluşumunda en elverişli şartlara haiz olanAkdeniz ikliminin etkisinde yer almaktadır. Büyük ölçüde kuru-yarı nemli, yarı nemlive yarı kurak iklim şartlarının hakim olan Türkiye için, orman yangını bütün bölgedekiorman alanları için yüksek risk ve ciddi tehlike barındırmaktadır.Bütün Dünya'da ve benzer şekilde ülkemizde de ormanlara yapılan tahribatın başınıorman yangınları çekmektedir. İstatistiki bilgiler irdelendiğinde ülkemizde ormanyangınlarının değişik yapıya sahip olduğu ancak ilerleyen senelerde nüfus sayısındakiartışa bağlı olarak yangınlar da da fark edilir bir artışın yaşandığı göze çarpmaktadır.Bu durum; doğal orman yangınına ait yangın öncesi ve yangın sonrası Landsat 8 OLI / TIRS görüntülerikullanılarak, yanan alan belirlenmiştir.Çalışma kapsamında söz konusu yangında, OGM verilerine göre 540 ha ormanlıkalanın yok olduğu tespit edilmiştir. Ancak yapılan analizler sonucu, yanan alan(NDVI) değeri 1372 ha , yanan alan (BAI) değeri 1403 ha ve yanan alan (NIR) değeri1588 ha olarak hesaplanmıştır. Buna göre, en doğru ve kapsamlı sonucun NIRbandının maskesinin filtrelenmesinden elde edildiği görülmüştür.

The ecosystem consisting of the interaction between microorganisms which are livingbeings, animals, plants including trees, and physical environmental factors like light,air, soil and temperature is called a forest. The infinite number of substances and eventsthat bring the forest to existence are in a two-way relationship and interaction witheach other. For this reason, the forest is an ecosystem that manifests itself through thepopulation of many plants and animals.The forests which are playing a critical role in the establishment of natural equilibriumhave to be devoutly protected due to their environmental and ecologic benefits. Forestfires are the leading cause of destruction to the forests worldwide as well as in ourcountry. When referring to forest fires all over the world, primarily Mediterraneancountries, Australia, Canada and the United States come to mind. These countries canbe considered successful in the fight against fires. This success is due to the realisticprediction of the danger of fire, the right planning systems and the opportunities forimmediate intervention in places where there is danger. Turkey is situated under theinfluence of the Mediterranean climate which provides the most favorable conditionsfor the occurrence of forest fires. In Turkey, which is mainly dominated by dry andsemi-humid, semi-humid and semi-arid climatic conditions, forest fires constitute agreat risk and a serious danger for all forest areas in the region.In areas where forest fires have occurred, fires are classified as areas to obtainextensive information on the damage caused by fires. The main purpose of theclassification is to demonstrate the effectiveness of the activities carried out in combatwith fires and to contribute to the measures to be taken against this fight.To achieve success in combating forest fires, it is not enough to use resourceseconomically and effectively by the timely and on-the-spot recruitment of thenecessary measures. In addition, advanced technologies must be used in every phaseof the firefighting process. For this reason, due to the possibilities they offer, RemoteSensing methods have been used for the mapping of fire areas, fire management,determination of the severity of the combustion and the burnt area for a long period oftime.Fire characteristics and plant canopy can be determined in connection with thetemperature differences between the fire surface and the fire surface using the firealgorithms used in Remote Sensing methods.The severity of combustion is defined as the determination of the ecosystem specificeffect of the physical, chemical and biological structure that undergoes a change in thefire. In this context, spectral changes occurring in the infrared bands of the vegetationare taken into consideration in the determination of burning intensity and burned area,and some analyzes are made by the determined indices. Combustion intensity can becalculated by grading difference images generated by taking advantage of satelliteimage indices before and after fire.In the post-fire period, many activities are carried out by taking advantage of thesatellite image indexes for examining forest foliation. In this context, the responsibleinstitutions in our country are carrying out activities for the re-foresting of areas whereforest fire is happening, and forests have been damaged by fire.xviiiOn the other hand, in the literature review, it has been determined that various studieshave been carried out to determine the forest fire characteristics (burning area andintensity of burning) and to determine the rehabilitation of the area in question.For example, the intensity of burning in forest fires in southern Spain has beendetermined by processing LANDSAT TM / ETM satellite imagery. In the satelliteimages obtained before and after the fire, the middle (M) and near (N) infrared (IR)bands were analyzed and the locations of the burned and unburned pixels could bedetermined. The middle (M) and near (N) bands were found to be the most appropriatebands to determine the fire character.Currently, when the fire effects of large-scale fires in various parts of the world areexamined, the dimensions of damage inflicted on vegetation can be categorized byusing the method of Differenced Normalized Burn Ratio (dNBR) which is widely usedin determining the forest burning intensity map.As a result of the analysis based on the fire which took place in the Extremadura regionof Spain in 2009 and had an impact on 3000 hectares (ha) forest area, 15 satelliteimages obtained through LANDSAT TM for forest foliation during the 27 monthsafter the fire were examined. The normalized difference for each image was examinedby calculating the NDVI for the aforementioned forest foliage survey.Like all over the world, forest fires cause serious loss also in our country. For thisreason, one of the main tasks of forest fire fighting organizations is to reduce forestfires to the minimum. Prediction of fire development and fire behavior characteristics(burning substance consumption, burning intensity and spreading rate) is critical interms of firefighting activities.The inadequacy of forest monitoring stations in our country and the spatial size of theforests make it difficult to systematically detect forest fires. Remote Sensing methodsstand out as a vital data source for the operational activities to be carried out in largescale with the possibilities that they have brought.In terms of the effects of forest fires and their consequences, the worst natural disastersin the world are leading. When statistics are analyzed, forest fires in our country havedifferent structures, but in the following years, there is a noticeable increase in firesdue to the increase in the number of the population. This situation; natural decay,erosion, desertification, landslides and mass loss. However, renewal of forests andvegetation damaged by fires is also important for land management.Estimation of forest fires and their behavior is very important in terms of fire fightingmethods. In this context, using satellite imagery greatly facilitates the detection of fireaffectedareas and intensity of burning in large areas.Temperatures reaching extremely high levels in July and August, particularly in theMediterranean region, pose a threat to forest areas when they reach low humidity.Disaster struggle has an important position among remote sensing practices and forestfires, one of the most frequent natural disasters today, can be critically evaluatedthrough satellite imagery.In this study, we investigated the Antalya Kumluca fires, one of the forest fires in theMediterranean region in our country. Pre-and post-fire satellite images were used todetermine the burning area. In this sense, it is observed that the work put forward inour country is not in the desired place. The aim of this work is to detect rapidly the burning area in forest fires that may occur in the future and to contribute to forestrehabilitation applications to be realized through satellite images.For this purpose, forest fire, remote sensing analysis techniques and Landsat 8 OLI /TIRS satellite images which were effective in Kumluca district of Antalya provincebetween 24.06.2016 - 29.06.2016 were evaluated. Using the Landsat 8 OLI / TIRSimages before and after the fire of Antalya-Kumluca forest fire, the burning area wasdetermined.Within the scope of the study, it has been determined that there is no forest area of 540ha according to GGM data. However, the results of the analyzes were calculated as1372 hectares for burning area (NDVI), 1403 ha for burning area (BAI) and 1588 hafor burning area (NIR). Accordingly, it has been shown that the most accurate andcomprehensive result is obtained by filtering the mask of the NIR band.