Coal fire identification and state assessment by integrating multitemporal thermal infrared and InSAR remote sensing data: A case study of Midong District, Urumqi, China

https://doi.org/10.1016/j.isprsjprs.2022.06.007Get rights and content

Abstract

Coal fire disasters generally occur in major coal-producing countries. They destroyed many coal resources and triggered various environmental problems. The identification and state assessment of coal fire areas are paramount to coal fire management. In this paper, a method for identifying coal fires by integrating land surface temperatures (LSTs) from multitemporal thermal infrared remote sensing and subsidence information from multitemporal interferometric synthetic aperture radar (MT-InSAR) is proposed. The method primarily consists of estimating the LSTs and subsidence values, normalizing and equidistantly partitioning the LSTs, estimating the thermal anomaly frequency (TAF), extracting high-subsidence areas, determining the thermal anomaly frequency threshold (TAFT), and identifying coal fire areas according to the TAFT. The coal fire ratio index (CRI) was developed to quantitatively evaluate the severity of coal fires and its variations. Midong District, Urumqi, Xinjiang, China, was selected as the study area. Forty-five Landsat 8 images and sixty-one Sentinel-1 SAR images were used to retrieve LST and subsidence time series, respectively. The proposed method and CRI were applied to identify and evaluate coal fires in the study area. The results demonstrate the satisfactory reliability of these methods compared with field surveys. The maximum CRI was 6.145 × 10-4 (July 28, 2019), and the minimum CRI was 2.685 × 10-5 (January 13, 2020). Coal fires in summer were more severe than those in winter. The CRI profile presented seasonal and annual periodicities and was affected by the local climate. The soil cohesion and humidity are higher in winter due to possible snowmelt, which weakens the interaction between oxygen and coal seams and the subsequent combustion of coal. The LST, thermal anomaly, and deformation information were combined in a time series overlay analysis to reveal the differences in the LSTs and subsidence values among mining areas, coal fire areas, and areas with both mining and coal fires. The results verify the applicability and reliability of the proposed method.

Introduction

A coal fire refers to a burning phenomenon that occurs when an underground coal seam is subjected to combustion conditions under the influence of natural or human factors (Song and Kuenzer 2014). Coal fires are common in coal-producing countries, such as China, the United States, India, Australia, and Indonesia (Biswal et al., 2019, Deng et al., 2021, Deng et al., 2020, Engle et al., 2012, Kuenzer et al., 2008, Liu et al., 2019). The burning of coal fires is accompanied by severe environmental pollution, geological disasters, and the loss of coal resources (Biswal et al., 2019, Song and Kuenzer, 2014, Syed et al., 2018). Hence, the effective identification and assessment of coal fire areas are crucial in coal fire management. In China, coal fires occur widely throughout Xinjiang, Inner Mongolia, Shaanxi, Shanxi, Qinghai (Song and Kuenzer 2014). The coal fields in Xinjiang are mostly inclined, shallow coal fields in which the coal seams are relatively thick and appear in widespread outcrops. The climate in this region is dry with little rain. Consequently, coal fire disasters are likely to occur (Deng et al., 2021, Deng et al., 2020, Liu et al., 2019).

The current methods for detecting coal fires mainly include direct field surveys, geophysical prospecting, geochemical prospecting, drilling methods, and remote sensing detection (Li et al., 2021, Qi et al., 2013, Shao et al., 2018, Syed et al., 2018). The first four types of methods can directly delineate coal fire areas. However, these techniques require substantial human and material resources. Moreover, their detection range is limited and efficiency is low. It is difficult to apply them to the detection of large-scale coal fire areas. In contrast, methods for detecting coal fires based on remote sensing technology have the advantages of a wide observation range, a low cost, and a high efficiency. According to the type of observation information used, remote sensing detection can be divided into two categories: coal fire detection based on thermal anomalies and coal fire detection based on subsidence information. The former mainly uses unmanned aerial vehicle (UAV)-borne or spaceborne thermal infrared sensors to identify thermal anomalies caused by coal fires. The latter mainly uses multitemporal interferometric synthetic aperture radar (MT-InSAR) to extract regional subsidence information produced by coal fires for identification.

Scholars have conducted a considerable amount of research on coal fire identification and modelling using UAV thermal infrared technology. It identifies coal fire areas by detecting abnormal land surface temperatures (LSTs) (He et al., 2020, Li et al., 2018, Shao et al., 2021). This technology is highly flexible and benefits from a high spatial resolution and a high observation efficiency. However, despite their ability to achieve high-resolution monitoring, UAVs struggle with problems such as a small data collection range, a high cost of multitemporal data acquisition, the sensitivity of flights to weather conditions, and difficulties associated with airspace applications (Kelly et al., 2019, McKenna et al., 2017, Wang et al., 2022). Due to these drawbacks, achieving long-term and large-scale monitoring using UAVs is challenging. Compared with UAV detection, the spaceborne thermal infrared remote sensing has the advantages of a wide observation range, a low cost, the ability to acquire data without flight restrictions, and a consistent imaging interval. In addition, global data from many satellites are available for free. These data provide a basis for detecting coal fires based on thermal infrared remote sensing (van der Meer et al., 2012, Zhou et al., 2019). Accordingly, many scholars have used data from the Moderate Resolution Imaging Spectroradiometer (MODIS), Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), Landsat 5 Thematic Mapper (TM), Landsat 7 Enhanced Thematic Mapper Plus (ETM + ), and Landsat 8 to conduct various studies on the identification and analysis of coal fires (Huo et al., 2014, Jiang et al., 2011b, Kuenzer et al., 2008, Martha et al., 2010, Mishra et al., 2011, Roy et al., 2015, Song et al., 2015). These studies have shown that using spaceborne thermal infrared remote sensing to identify coal fire areas is feasible. However, the identified fire areas may deviate from the actual situation due to the impacts of the imaging season, the quality of the data, topographic undulations, the surface coverage, the coal fire depth, etc.

Subsidence in coal fields where mining has ceased and coal fires are ongoing is mainly caused by coal fires. As a new type of active remote sensing technology, MT-InSAR can effectively detect surface subsidence and provide new perspectives for identifying coal fire areas (Ferretti et al., 2000, Ferretti et al., 2001, Hooper, 2008, Hooper et al., 2012, Yu et al., 2013). To date, scholars have applied the MT-InSAR method in conjunction with GPS and field survey data and comprehensively verified that MT-InSAR can play a valuable role in identifying coal fires (Jiang et al., 2011a, Kumar et al., 2020, Liu et al., 2019). Specifically, these studies have shown that MT-InSAR technology can effectively identify the ground subsidence in these areas caused by underground coal fires. However, in coal fields containing coal fires and active mining, subsidence is caused by both processes. Thus, the subsidence area cannot be simply regarded as a coal fire area. In addition, surface subsidence due to coal fires is not readily apparent in the early stage of combustion. Accordingly, accurately identifying coal fires based on single-source subsidence information is particularly challenging. Therefore, scholars have proposed an analytical method that combines multisource remote sensing data to identify coal fires. It comprehensively analyse the thermal infrared and radar remote sensing data (Xu et al., 2021b, Yan et al., 2020). Nevertheless, researchers in these studies employed only a small amount of data to jointly analyse subsidence and thermal anomalies, whereas field survey data and subjective interpretations are necessary to identify coal fire areas. Therefore, we should further explore the characteristics of coal fire areas to increase the accuracy and objectivity of identifying these areas.

In addition to identifying coal fire areas, evaluating the severity of coal fires is another essential aspect of coal fire management. However, coal fires mainly occur below the ground surface. Using direct observations to obtain information about the state of a coal fire is difficult. Traditional methods mostly use geophysical and geochemical exploration, which can only detect fires in small areas and requires considerable manual labour and material resources. Hence, these techniques struggle to meet the requirements of macro-quantitative analysis (Li et al., 2021, Qi et al., 2013, Shao et al., 2018). Alternatively, coal fire areas can be feasibly identified by combining multisource remote sensing data, as this approach has the advantages of a wide observation range, a high efficiency, and a low cost (Biswal et al., 2019, Syed et al., 2018). Unfortunately, no research on the use of remote sensing information to retrieve and evaluate the severity of coal fires has been conducted.

This paper focuses on two issues: the identification of coal fire areas and the quantitative analysis of coal fire intensities. Midong District of Urumqi, Xinjiang, China, was selected as the study area. 45 Landsat 8 images and 61 Sentinel-1 SAR images covering the same time span were used to identify and evaluate coal fire areas. The main contributions include the following. (1) A method for fusing the LST information obtained by multitemporal thermal infrared remote sensing with the subsidence information obtained by MT-InSAR is proposed to determine the thermal anomaly frequency threshold (TAFT) and to identify coal fire areas. By using the multitemporal thermal anomaly frequency (TAF) and combining it with the ground subsidence to determine the TAFT, fire areas can be identified more objectively and accurately. (2) A coal fire ratio index (CRI) is introduced to quantitatively evaluate the severity of coal fires and its changes over time. By analysing the CRI values corresponding to remote sensing images at different times, the severity of coal fires and their temporal variations can be quantitatively assessed. (3) The proposed method was used to obtain the spatial distribution of coal fire areas in Midong District, and the reliability of the identification results was verified by a field investigation. (4) By analysing the CRI values across the study area in different periods, the characteristics of the changes in the coal fire intensity were obtained. Meanwhile the influencing factors were revealed.

The remainder of this paper is organized as follows. Section 2 introduces the study area and experimental data. Section 3 presents the methodology, with section 3.1 introducing the basic principles and the overall technical route, section 3.2 describing the coal fire identification method, and section 3.3 describing the CRI and its calculation method. Section 4 describes the experimental results and analysis. Section 5 discusses the differences in the LST and subsidence values among mining areas, coal fire areas and areas with both mining and coal fires, explores the applicability of the proposed method, and compares it with previous studies and deep learning. Section 6 concludes the paper and provides outlooks for future work.

Section snippets

Study area

Midong District, Urumqi, Xinjiang, China, was selected as the study area. The red polygon in Fig. 1(b) shows the specific location of this area. The district is oriented southwest–northeast and is within the boundaries of (87°42′24″E ∼ 87°51′59″E, 43°53′21″N ∼ 43°58′38″N). From west to east, the study area encompasses the Zhongxing Coal Mine, the Zhiqiang Coal Mine, the Sandaoba Fire Area, the Sanyuan Coal Mine, and the Gandiantou Fire Area. Midong District is located east of Urumqi, south of

Basic principles and overall technical route

The combustion of coal fires is accompanied by the continuous release of heat, which can raise the LST in the coal fire area above that in the surrounding non-coal fire areas. This leads to thermal anomalies. In addition, the burning of coal fires reduces the quantity of underground coal resources and eventually causes surface subsidence. Based on these characteristics, coal fire areas can be identified by the combination of thermal infrared remote sensing and MT-InSAR. However, the existing

LST inversion, normalization and density segmentation

In this study, the RTE introduced in section 3.2.1 is used to invert the LSTs corresponding to the Landsat 8 images. LST normalization and density segmentation are the primary processes used to identify coal fire areas and to calculate and analyse the CRI. First, the LST is normalized and then segmented into equal density intervals as described in section 3.2.3. The LST is divided into 7 levels (in research on the urban heat island effect, dividing the LST into 7 levels generally achieves good

Discussion

To understand the dynamic characteristics of the LSTs and subsidence in the coal fire areas and their differences from those in non-coal fire areas, the LSTs and deformation time series are overlaid in this section. Moreover, the influences of coal fires, coal mining and other factors are analysed based on the LST and deformation information at feature points in different regions. Accordingly, the evolutionary characteristics of the subsidence and LST in different regions are revealed.

The

Conclusion and outlook

The burning of underground coal fires results in a considerable loss of coal resources, a variety of secondary disasters and environmental pollution. Compared with traditional coal fire detection methods such as field surveys, remote sensing technology has gradually become quite valuable due to its large spatial coverage and low cost.

This study proposed a coal fire area identification method that combines multitemporal thermal infrared and MT-InSAR data. This method fully considers the

Declaration of Competing Interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Bing Yu reports financial support was provided by National Natural Science Foundation of China. Bing Yu reports financial support was provided by Science and Technology Department of Sichuan Province. Bing Yu reports was provided by Chinese Academy of Sciences State Key Laboratory of Geodesy and Earth’s Dynamics.

Acknowledgements

This work was funded by the Youth Science Fund of the National Natural Science Foundation of China (41801399), the Science and Technology Project of Sichuan Province (2018JY0138) and the Open Fund Program of State Key Laboratory of Geodesy and Earth's Dynamics (SKLGED2020-5-1-E).

References (57)

  • R. Niclos et al.

    Evaluation of Landsat-8 TIRS data recalibrations and land surface temperature split-window algorithms over a homogeneous crop area with different phenological land covers

    ISPRS J. Photogramm. Remote Sens.

    (2021)
  • B. Osmanoglu et al.

    Time series analysis of InSAR data: Methods and trends

    ISPRS J. Photogramm. Remote Sens.

    (2016)
  • D.P. Roy et al.

    remote sensing of environment

    (2014)
  • P. Roy et al.

    An approach of surface coal fire detection from ASTER and Landsat-8 thermal data: Jharia coal field, India

    Int. J. Appl. Earth Obs. Geoinf.

    (2015)
  • Z. Song et al.

    Coal fires in China over the last decade: A comprehensive review

    Int. J. Coal Geol.

    (2014)
  • Z. Song et al.

    Analysis of coal fire dynamics in the Wuda syncline impacted by fire-fighting activities based on in-situ observations and Landsat-8 remote sensing data

    Int. J. Coal Geol.

    (2015)
  • T.H. Syed et al.

    Remote sensing of coal fires in India: A review

    Earth-Sci. Rev.

    (2018)
  • F.D. van der Meer et al.

    Multi- and hyperspectral geologic remote sensing: A review

    Int. J. Appl. Earth Obs. Geoinf.

    (2012)
  • Q. Vanhellemont

    Combined land surface emissivity and temperature estimation from Landsat 8 OLI and TIRS

    ISPRS J. Photogramm. Remote Sens.

    (2020)
  • Z.F. Yang et al.

    Locating and defining underground goaf caused by coal mining from space-borne SAR interferometry

    ISPRS J. Photogramm. Remote Sens.

    (2018)
  • B. Yu et al.

    Subsidence detection by TerraSAR-X interferometry on a network of natural persistent scatterers and artificial corner reflectors

    Comput. Geosci.

    (2013)
  • B.W. Zhang et al.

    Mapping the Yellow River Delta land subsidence with multitemporal SAR interferometry by exploiting both persistent and distributed scatterers

    ISPRS J. Photogramm. Remote Sens.

    (2019)
  • P. Berardino et al.

    A new algorithm for surface deformation monitoring based on small baseline differential SAR interferograms

    IEEE Trans. Geosci. Remote Sensing

    (2002)
  • A.K. Bhunia et al.

    A novel feature descriptor for image retrieval by combining modified color histogram and diagonally symmetric co-occurrence texture pattern

    Pattern Anal. Appl.

    (2020)
  • S.S. Biswal et al.

    Delineation and mapping of coal mine fire using remote sensing data - a review

    Int. J. Remote Sens.

    (2019)
  • M. Chhetri et al.

    Deep BLSTM-GRU Model for Monthly Rainfall Prediction: A Case Study of Simtokha

    Bhutan. remote sensing

    (2020)
  • Ferretti, A., Prati, C., & Rocca, F. (2000). Nonlinear subsidence rate estimation using permanent scatterers in...
  • Ferretti, A., Prati, C., & Rocca, F. (2001). Permanent scatterers in SAR interferometry. ieee transactions on...
  • Cited by (10)

    View all citing articles on Scopus
    View full text