Thermal Infrared Remote Sensing Inversion And The Characteristics Of Typical Pollutant Emissions From Coalfield Fires

China is the most seriously affected country by coal fires in the world.The coal fire disaster not only burns and damages a large amount of coal resources,endangers coal mine safety and economic production,but also seriously causes a series of landscape security and health environment problems.Although underground coal fires are a key source and emission link of atmospheric pollution and greenhouse gases,the basic research on the environment of coal fires at home and abroad is relatively lacking so far,and there are no official authoritative methods for monitoring and calculating carbon emissions from coal fires to refer to.Therefore,it is necessary to propose quantitative detection and analysis methods adapted to the special combustion and emission situation for the coal field fires,and establish a comprehensive environmental impact evaluation index system,which is of great significance to scientifically improve national carbon emission inventory,effectively evaluate the coal fire hazards and take reasonable prevention and control measures.This study has applied thermal infrared remote sensing to conduct coal fire detection and spatio-temporal quantitative inversion,and carried out the first comprehensive environmental field investigation in a full-scale region.By studying the emission characteristics of typical pollutants and greenhouse gases from coal fires,a series of models and algorithms on remote sensing interpretation of coal fires,environmental assessment and quantitative estimation of carbon emissions were proposed,which were successfully applied to the assessment of coal fire suppression effects on fire risk identification,pollution emission impact and carbon emission contribution in the Sandaoba coal fires of Midong coalfield in Xinjiang.Based on Landsat-8 satellite time series thermal infrared images,the coal fire spatial and temporal evolution was interpreted qualitatively and quantitatively,and the coal fire development changes in thermal intensity,spatial location and areal size were also analyzed.The overall accuracy of the Hotspot Analysis method to extract fire thermal anomalies exceeds 65%,which is determined as the optimal threshold algorithm for coal fire demarcation.The overall thermal anomalies in the fire area exhibits a highly clustered spatial autocorrelation distribution pattern.The proposed Hotspot Sequential Frequency Extraction algorithm finds that the hotspot can be judged as a coal fire pixel when its frequency meets 75%or more of the observation number.The Coal-fire Thermal-island Intensity Ratio was constructed to comprehensively assess the intensity and spatial variation of the coal fire heat island phenomenon,and the overall intensity of the fire area was reduced by more than 70%after the fire control measures were carried out.The migration trajectory of the combustion centroid indicates that there is a large variability in the fire development of different functional sections,and the trend of coal fire risk-intensive center is not only unidirectional expansion,but also changes involving multi-directional,bilateral and round-trip propagation.The hotspots in the fire area are mainly located near the mining area,collapse pits and fissure zones,and the fire extends in a striped pattern along the stratigraphic direction and coal seam outcrops towards the outer sides.After fire control or reduction of mining operations,the fire contour boundary observed by remote sensing shows rapid inward contraction,and the burning area then shows signs of rapid decline.For the first time,a large-scale comprehensive field survey was conducted in the Sandaoba fire area,the emissions of gaseous combustion products and thermal property parameters were measured for the whole fire area.A comprehensive evaluation and classification method adapted to coal fire pollution was proposed to estimate the overall coal fire pollution emissions,and the local air quality and combustion status were comprehensively assessed.The annual flue gases emissions from boreholes in Sandaoba fire area are about 14,170 t CO₂,113.6 t SO₂,and 57.3 t CO.The Air Quality Indices in the study area are all below 60,ranging from 24 to 58,indicating that the coal fire air pollution is in the moderate to severe stage.Flue gases concentrations are geographically diverse,and the degree of regional pollution and fire suppression progress are inversely proportional.The Principal Component Analysis can resolve the thermal physical observations into three common factors:thermal physical intrinsic properties,atmospheric dynamics,and combustion degree.The Factor Analysis extracted four integrated factors representing the evaluation indicators of air pollution risk,which are sulfur oxide pollution>nitrogen oxide pollution>carbon oxide>particle pollution in order of pollution severity.Canonical Correlation Analysis displays a significant correlation between pollution and thermal physical indicators.The overall combustion efficiency decreases with the increase of excess air coefficient.The overall combustion efficiency in the fire area is below 80%,so the coal fire combustion type is dominated by smoldering combustion.The thermophysical properties are the decisive factors influencing the fire pollution emissions from coal fires,followed by meteorological,human interference and external influence of the environment.Based on the above coal fires survey and quantitative inversion of remote sensing,a method for quantifying greenhouse gases emission related to coal fires in a coalfield is proposed as an engineering application,including two carbon emission calculation models of the gas emission flux based method and the coal loss in combustion based method.The acquisition and calculation process of the key parameters of the models are elaborated in detail by means of remote sensing,in-site investigation and laboratory measurement.According to the combustion intensity and characteristics,the coal loss in combustion based method divides the coal fire dynamic process into three combustion stages:developing,stable,and decaying period.And the proportions of the staged combustion cycles are 36.51%,33.33%and 30.16%respectively.The growing,stable and reducing combustion areas in each stage were detected through thermal infrared images.From this,the annual burning coal loss for the whole stage was calculated as 6.82×10⁴ tons,and the corresponding deduced annual total carbon emission was 5.372×10⁵ tons.The carbon emission flux model considers different emission sources,it estimates the CO₂ of borehole emission based on the flue gases analysis,and obtains the fracture area percentage of fissure emission sources through fracture field investigation.For soil diffusion emissions,the chamber method is used to measure carbon emissions flux and through Kriging statistical interpolation model to simulate the global flux distribution.By accumulating the three emission sources the total annual average carbon emission in the fire area derived is 5.052×10⁵ tons.Taking the average of the estimation from the two models and considering the contribution of methane on greenhouse gases emission,the annual carbon equivalent of greenhouse gases emission in Sandaoba fire area is finally obtained as 6.313×10⁵ tons.The dissertation has 42 figures,37 tables,and 138 references.