Introduction

Underground coal fires are natural hazards which the coal experiences oxidation, thermal accumulation, and eventually spontaneous ignition (Liang et al. 2016), generally occurring in exposed coal seams, coal storage piles, or coal waste piles (Jiang et al. 2017). Coal fires consume non-renewable coal resources and release tremendous amount of toxic gases (e.g., CO and SO2), greenhouse gases (e.g., CO2 and CH4), and trace elements (e.g., As, F, and Hg). In addition, underground coal fires contribute to land subsidence due to volume loss of coal seams (Kuenzer and Stracher 2012), and threaten the lives and properties of local residents. To data, research on underground coal fires in the environment has focused on the assessment of soil gas emissions, underground water, and soil pollution (Song and Kuenzer 2014).

Pone et al. collected and analyzed gas combustion by-products from gas vents in the Witbank and Sasolburg coalfields, and assessed the health risk to local communities (Pone et al. 2007). Carras et al. measured greenhouse gas emissions originating from spoil piles and waste coal dumps, and studied the difference in gas emissions among three areas characterized by the extent of spontaneous combustion (Carras et al. 2009). Temporal and spatial gas (e.g., CO2, CO, Hg, and hazardous pollutants) trends of vent dynamics were extensively investigated (Dindarloo et al. 2015; Engle et al. 2012; Garrison et al. 2017; Hower et al. 2009, 2011, 2013; O’Keefe et al. 2010, 2011). Gaseous emissions may condense in the mouth of gas vents due to the sharp decrease in temperature, and the characteristics of the secondary minerals of the condensates have been investigated, including the components and the formation mechanisms (Dias et al. 2014; O’keefe et al. 2018; Oliveira et al. 2019). Some gaseous emissions are highly soluble, and liquor leaching may lead to severe acid drainage, resulting in pollution of underground water (Querol et al. 2011). In addition, the toxic trace elements (e.g., As, F, Hg, and Se) emitted by the underground coal fires may become enrich in gas vents and soils (Hong et al. 2018; Liang et al. 2018; Li et al. 2018), thus causing soil contamination. Changes in soil properties could also influence the distribution of soil heavy metals and polycyclic aromatic hydrocarbons (PAHs) in coalfields, which in turn could aggravate the destruction of terrestrial ecosystems. Querol et al. found that mobility of certain trace pollutants was enhanced under acidic conditions, such as Li, Ti, Sc, V, Cr, Mn, Co, and Ni. In addition, Al-K-Fe-NH4+-bearing sulfates easily dissolved under acidic conditions (Querol et al. 2008). Garrison et al. observed that the concentrations of monoaromatic hydrocarbons (BTEX) and PAHs in soils were higher in areas with high soil matter contents due to their enhanced absorption capacities (Garrison et al. 2016). Wang et al. observed that soil Hg and Cd exhibited notable correlations with total carbon contents, and suggested that soil Hg and Cd were related to organic matter (Wang et al. 2019). Therefore, the study of soil properties contributes to explaining the formation and distribution of certain soil pollutants.

A small, but growing, body of research has been conducted to investigate the impact of underground coal fires on the soil properties in coal fire areas. Tobin-Janzen et al. found changes in resident microbial populations in boreholes impacted by Centralia, and Pennsylvania anthracite coal fires, and the soil temperature, pH, and chemical composition were analyzed (Tobin-Janzen et al. 2005). Zhang et al. performed physicochemical characterization of soil samples in the mouths of vents, and microbial communities at high temperatures (Zhang et al. 2013). Zeng et al. investigated surface vegetation coverage, heavy metal distribution, and organic matter, and determined that coal fires can affect soil properties via organic matter digestion (Zeng et al. 2018). However, most of the research on soil properties has focused on vents or surface openings, in addition to the soil gas emissions.

The aims of this investigation are to (1) analyze the concentrations of hazardous pollutants in various soils layers in Wuhushan coal fires not confined to vents or surface openings and (2) collect and analyze surface soil samples to conduct the soil physicochemical characterization in the investigated area. In addition, the effect of backfill consisting of sandy soils on the gas emissions and soil properties is examined.

Materials and methods

Site description

The Wuda coalfield is located in the northwest of Wuda District, and including three coal mines: Wuhushan, Suhaitu, and Huangbaici (Fig. 1), covering an area of 35 km2. The study area is a backfilled region in the northern Wuhushan mine, which belongs to the no. 10 fire zone. The climate in the area is characterized as a strongly continental, semiarid climate, with an average precipitation of approximately 168 mm/a (Song and Kuenzer 2017). The land surface of the Wuhushan coalfield is dominated by bare rocks covering sandy soil, and the vegetation density is very low (Song et al. 2015). The Wuhushan coal mine started construction in 1965. The no. 9, no. 10, no. 12, no. 13, no. 15, no. 16, and no. 17 coal seams are the main coal seams in the Wuhushan mine, among which the no. 9, no. 10, and no. 12 exhibit a high propensity of spontaneous combustion with sulfur contents up to 3.42%. Notably, in-ground coal fires occur when shallow coal seams are exposed at the Earth’s surface (Shan et al. 2019). The destructive and disorderly operation of small-scale private mine results in the spread of coal fires. Since 2006, a coal fire suppression project has been implemented in the area. Coal mines are typically contracted to individuals, and they are permitted to excavate surface rocks, remove spontaneous combustion coal, and sold any mineable coal, but they are responsible for backfilling. However, exaction has destroyed the landscape, and new fissures and cracks have thus been created, which has led to the expansion of the coal fire range. The above scheme has been terminated by the local government, and a fire suppression project was gradually implemented towards standardization in 2011. Follow-up work was organized and implemented by a professional fire-fighting engineering office (Liang 2018). The burning coal was excavated, the ground surface was covered with sandy soil, and sowing grass were sown for ecological restoration purposes. This method obtained promising results, and the area of shallow and visible coal fires was estimated to have decreased by ~60%, while the risk of spreading remains (Jiang et al. 2017). In the research zone, the intensity of shallow coal fires notably decreased, and recovery of vegetation was observed along the edge of the survey zone. However, the overlying soil did not prevent hidden combustion of coal seams. Nearly all the surfaces were devoid of vegetation in the high-temperature zone. Visible surface flames were extinguished, but fumes persisted scattered throughout the central area.

Fig. 1
figure 1

Location of the coal fire in the Wuda Coalfield, Inner Mongolia, China (The source of the figure is from the Google maps)

Full size image

Gas concentrations were measured on a variety of surfaces grouped into three categories based on the degree of combustion of the underground coal fire:

  1. 1)

    Surfaces with spontaneous combustion and obvious features, including cracks (Fig. 3b), sponges (Fig. 3c), fissures (Fig. 3d), and vents (Fig. 3e). This region was defined as area A.

  2. 2)

    Surface with spontaneous combustion but without any obvious features covered by sandy soils. This region is defined as area B.

  3. 3)

    Surfaces characterized by withered plants (Fig. 3e). The vegetation degradation area occurred along the periphery of the investigated area (Fig. 2). It represented the reclamation area, in which coal fires had been extinguished, and grass seeds were sown for ecological restoration. This region was defined as area C.

Fig. 2
figure 2

The sampling location

Full size image

The coal fire-induced landforms can be roughly divided into three types among the backfilled regions in area A, including cracks and vents, sponges, and fissures. Cracks and vents are distributed in the southeastern sampling areas, and obvious black smoke is observed. The maximum width of the cracks is approximately 70cm, and they are possible pathways for the ventilation of the underground coal fires (Kuenzer and Stracher 2012). Compared to the original landform (Engle et al. 2012), the observed vents and cracks contained no visible minerals or tars (Fig. 3b). The soil cover may contribute to the lack of tar and minerals. The fissures are mainly distributed in the northwestern part of the investigated area. The fissures are linear in the surface soil, and they are common outbound pathways of hot gas emissions (Kuenzer and Stracher 2012). The in situ fissures range from a few centimeters, and are extremely narrow (<0.5 cm), and white and yellow salt minerals could be observed besides the fissures. When hot particulate emissions cool, the substances absorb water and change into white salts which mainly consist of alunogen (Querol et al. 2011). The yellow salts mainly comprise native sulfur, which is generated the reaction: SO2+2H2S→3S0+2H2O (Querol et al. 2008). The presence of the sulfur also suggests that the coal fire is located deep below the ground surface, because sulfur only remains stable below 100°C (Gürdal et al. 2015), which is also confirmed by measurement of the temperature in situ (Fig. 4). Sponges are a type of sponge-like contaminated soil (Liang et al. 2018), commonly found on the surface of the backfilled regions, and are mainly distributed in the southeast, as shown in Fig. 3e.

Fig. 3
figure 3

Landscapes of the research zone. b Crack. c Sponge. d Fissure. e Vent. e Degraded vegetation

Full size image
Fig. 4
figure 4

The distribution of surface temperature and soil humid in the research zone

Full size image

Soil sampling

A soil and gas sampling campaign was carried out in August, 2019. A total of 88 soil samples at depths ranging from 0 to 5 cm was collected with a Teflon tool at closely spaced intervals (10m), of which 30 samples were collected in area A, 46 samples were collected in area B, and the remaining samples were collected in area C.

Surface gas concentration measurements

H2S, SO2, and H2 were measured in situ, and monitoring was performed with a portable gas analyzer containing a self-priming pump. The range and accuracy of the emission analyzer are 0–2000 ppm and ± 0.1%, respectively, for H2 and 0–100 ppm and ± 1%, respectively, for SO2 and H2S. The probe of the sampling device was hammed to a depth of approximately 5 cm at each grid point, and connected to the gas inlet of the analyzer. In area A, a rubber tube was directly inserted into the vents and cracks. Measurement was recorded until the date remained stable during the process of measurement. CO, CO2, CH4, C2H6, C2H4, C2H2, N2, and O2 were analyzed via capillary gas chromatography (GC-4000A, Beijing), and the accuracy assessed via standard gas mixture was less than 5%, while the reproducibility was higher than 10%. Belowground gases were drawn with a gas-tight syringe, and injected into cleaned aluminum foil sampling bags. Moreover, near-surface gases were directly drawn with a gas-tight syringe at a vertical distance of 70 cm from the position corresponding to the belowground gases. The airbags were transferred to the laboratory within 24 h.

Soil chemical analyses

The soil pH was determined with various pH meters, and the organic matter content was measured via potassium dichromate oxidation. The soil humidity was measured with a potable digital hygrometer. The total salt contents were determined via measurement of the electrical conductivity of the soil leaching solution. The distribution of the soil pH and total salt contents in the investigated area were interpolated via the inverse distance weighting method in ArcGIS 10.4, as shown in Fig. 6.

Surface temperature measurement

The surface temperature was measured with an infrared thermal imager (Fortric220s, USA), and the measurement range and accuracy of the imager are −20 to 150 °C and ±0.1 °C, respectively. The highest temperature in the images (e.g., the vent in Fig. 3) was selected as the representative temperature. The surface temperature and soil moisture were also interpolated via the inverse distance weighting method in ArcGIS 10.4, as shown in Fig. 3.

Statistical analysis

A statistical t test was performed to statistically determine whether the soil properties in the different areas, including the soil organic matter content, soil pH, and total salt contents, were significantly different (p≤0.01).

Results and discussion

Analysis of the soil physical properties

The ground temperature in area A is significantly higher than that in area B (Fig. 3b), which indicates that the temperature is mainly transmitted to the surface via convection through the cracks and fractures. The variation in the ground temperature is high among the various fire-induced landforms. The maximum temperature in the cracks reaches 354°C (Table 4), significantly higher than that at the sampling points, which indicates that the fire font may lie in the southeastern part of the investigated area.

The cracks are linearly oriented from SW to NE, and it is inferred that the underground coal fires move from SE to NW based on the cracks that are orthogonal to the fire spreading direction. The average surface temperature of the sponges is the highest, and the sandy soil is very dark gray because of combustion and particle emission. The surface temperature of the fissures ranges from the 36.4 to 261 °C (Table 4), and the average temperature is the lowest among all the points with obvious features (Table 4).

The soil humidity is correlated with the combustion degree of the underground coal fire. The contours of the soil humidity are dense in areas where the combustion degree is high (Fig. 3). The soil near the sponges obtained the highest water contents, followed by the fissures (Table 4). Humid soil is almost undetectable in the degraded area due to the influence of solar radiation. In area B, at 25% of the points, humid soil is detected, mostly ranging from 2.1 to 10.1% (Table 4), and humid soil mainly occurs at the center of the high-temperature zone (Fig. 3). Research has shown the coal may contain more than 20% moisture by weight (Gürdal et al. 2015), and the vapor may seep to the surface along fissures within the bedrock by the buoyancy generated via heating, and it may accumulate in soil near the fissures and vents, thus increasing the soil humidity. The average soil humidity near fissures and sponges is higher than that near the vents and cracks (Fig. 3), because the former are more likely to act as outbound ventilation pathways (Kuenzer and Stracher 2012).

Analysis of the belowground gas concentration

Coal fire gases are transported to the surface through two primary mechanisms: advection through vents and fractures (vent emissions) and diffusion through the overburden (O’keefe et al. 2018). Gases are released in area A via advection, while diffusion dominates in area B. CO2, CO, and CH4 are the dominant gases in the research zone with additional components such as H2, SO2, H2S, C2H4, C2H2, and C2H6. These gaseous products may be largely generated via the thermal degradation of the biopolymers occurring in coal seams. The CO2 concentrations are relatively high in the vicinity of the fire front (Fig. 6), with the highest CO2 concentration reaching 11 vol%. The concentrations of CO2 seem to exhibit no correlation with the surface temperature, which is contrary to the conclusions of Hower et al. (2009). This probably occurs because gas emissions are more likely dominated by the underground fire itself, not the surface temperature. In addition, the distance to the fire or the presence of obstacles also influences the ground temperature and gas emissions (Garrison et al. 2017). The variation in the belowground CO2 concentration in area A is wide (Table 1), and is influenced by the fire depth, ventilation area, and relative position of the fire front to the vent (O’keefe et al. 2011; Song et al. 2019, 2020). The average CO2 concentration in area A is 5.4 vol% (Table 1), significantly exceeding the OSHA 8-h limit of 5 vol%. In the area B, the average CO2 concentration is of the same order of magnitude as that in area A (Table 2), which indicates a relatively high permeability, of the thin calcite-cemented sandstone occurring below surface soil layer. The variation differs within a range of 3 orders of magnitude in area B (Table 2), and there are two main reasons for the high variation. On the one hand, cracks, vents, and fissures are established before landfilling, but may remain still active despite the sandy soil cover. On the other hand, dense fissures and tectonic discontinuities occur within various strata of the formations.

Table 1 Principal statistical parameters of multiple gases in area A
Full size table
Table 2 Principal statistical parameters of multiple gases in the area B
Full size table

The vegetation cover is low in the investigated area (Fig. 4f), of which areas A and B are covered with little or no vegetation. Kuenzer et al. found that toxic gases and underground heat are responsible for the deterioration of vegetation (Kuenzer et al. 2007). A strongly negative relationship is found between the soil CO2 and O2 concentrations in areas A and B (Fig. 5), which indicates that the soil O2 is displaced by CO2 produced by the underground coal fire. Patil et al. suggested that a minimum soil O2 concentration ranging 12–14% could sustain a healthy root environment, or the plant would wither due to a decreased root respiration (Patil et al. 2010). However, the maximum O2 concentration in the degraded area the least affected by underground coal fires is lower than 10 vol% (Table 3). Therefore, anoxic conditions are also a major factor causing ecological environmental damage.

Fig. 5
figure 5

Correlation between soil CO2 and O2 concentrations in the investigated area

Full size image
Table 3 Principal statistical parameters of multiple gases in the degraded vegetation area
Full size table

Compared to the vents and cracks in the Heshituoluogai coal fire area (Wang et al. 2017), H2 is detected at most points in area A, and the average H2 concentration is 401 ppm (Table 1), which is significantly higher than that in the Heshituoluogai coal fire area. The high concentration of H2 indicates that the underground coal fire occurs in the advanced heating phase. The CO and CH4 concentrations in area A are lower than the CO2 concentration, indicating that the system is approaching near-complete combustion conditions. CO, ranging from 9 to 6008 ppm (Table 1), far exceeds the 8-h OSHA limit of 50 ppm. The CH4 concentration ranges from 4 to 2274 ppm (Table 1). In comparison to Tiptop coal-mine (O’keefe et al. 2018), the high CH4 concentration also indicates that the coal fire occurs in the stage of intense combustion. The CO and CH4 concentrations in the area B are significantly lower than those in area A, while the range of the concentrations is large, ranging from 4 to 3530.3 ppm and 1 to 2276 ppm (Table 2), respectively. C2H4 is detected in almost 76.5% of the analyzed samples in area A, and its concentrations reach up to 74 ppm (Table 1). Compared to C2H4, C2H6 is detected in almost 23.3% of the analyzed samples, and C2H4 and C2H6 are detected only in trace amounts: in fewer than 39% and 7%, respectively, of the analyzed samples (Table 1). H2S and SO2 are highly concentrated in area A, ranging from 0.72 to 26.4, and 2 to 100, respectively (Table 1).

In the degraded vegetation area, H2, SO2, and H2S are not detected. The belowground CO and CH4 concentrations are low, averaging approximately 7.11 ppm and 10.9 ppm, respectively (Table 3). A moderately negative relationship is found between the soil CO2 and CH4 concentrations in area A (r=0.727, p<0.01, n=30), indicating that they have the same origin, and most of the gases may be generated from the combustion of coal. The CO concentration does not seem related to the CO2 concentration in areas A and B. This indicates that the system occurs in an unsteady state, and CO may be rapidly converted into CO2. Although landfills may prevent the emission of CO, SO2, H2S, C2H2, C2H4, and C2H6, the average CO2 and CH4 concentrations are the same order of magnitude in areas A and B (Table 3). This indicates that the degree of greenhouse gas emissions through soil diffusion in the backfilling area also contributes to the burden of the atmosphere. Therefore, the estimation of greenhouse gas emission through soil diffusion in backfilling areas deserves attention.

Analysis of the near-surface gas concentration

The near-surface CO2 concentration is relatively high in area A, and the maximum value is below the OSHA 8-h limit of 0.5 vol% (Table 1). The contours of the near-surface CO2 concentration are dense in the vicinity of the high-temperature zones, but sparse in the area far from the high-temperature zones (Fig. 6). The average concentration of CO2 is 839 ppm in the vicinity of the degraded vegetation area (Table 3), almost unaffected by underground coal fires. Near-surface CH4 and CO are also detected in area A, but their concentrations are lower than their corresponding belowground concentrations (Table 1). The maximum ground concentration of CO is 68 ppm in area A (Table 1), which is slightly higher than the eight-hour OSHA limit of 50 ppm. Obviously, gases are diluted by air as they are emitted, but temperature inversions can cause potential health risks in the surrounding regions of coal fire areas (Hower et al. 2009). The correlation of the belowground CO2 concentration with near-surface CO2 concentration in area A is positive but not significant (R2=0.38, p<0.01, n=30), and in area B, there seems to be no association (R2=0.17, P=0.01, n=58). This suggests that gases generated from the vents, cracks, and fissures mainly affect the near-surface CO2 concentration distribution. The near-surface CO and CH4 concentrations in area B are in line with these same rules. The near-surface CH4 concentrations at same locations are even higher than their corresponding belowground CH4 concentrations due to the influence of vent advection. H2, SO2, H2S, C2H4, C2H6, and C2H2 are below the detection limits in the near-surface zone (Table 2).

Fig. 6
figure 6

The distribution of CO2 concentration in the investigation area

Full size image

Analysis of the soil total salt contents and soil pH

Area A exhibits the lowest soil pH, with an average value of 5.35, and it mostly ranges from 4.03 to 6.67 (Table 4). Several studies of the soil pH distribution impacted by underground coal fires can be referenced for comparison. Tobin-Janzen et al. found that the soil pH in surface soil samples (at a depth from 0 to 20 cm) collected in boreholes indicated slightly acidic to acidic conditions throughout the Pennsylvania mine fire area (6.35–4.11) (Tobin-Janzen et al. 2005). Wang et al. measured the soil properties around the vents, and found that the soil pH and total salt contents near vents varied although they originated from same coal seam, and the data indicated that the soil pH was correlated with the total salt contents (Wang et al. 2020). The soil pH observed in this study is not consistent with the above reports because of the difference in the environment among the areas. However, these studies all indicate that the soil pH near the vents and cracks is generally acidic.

Table 4 The soil properties in the area A
Full size table

The average soil pH varied greatly among the landforms in area A, and the sponges in the area exhibited the lowest soil pH (0.95–1.99), followed by the vents and cracks (1.57–7.18), while the soil pH near the fissures was the highest (3.09–7.79) (Table 4). The negative correlation between the soil pH and the total salt contents (r=0.859, p<0.01) indicated that the soil pH is significantly influenced by the soil salt in area A (Fig. 7). A source for a high acidity may be hydrolysis of the sulfate, while Fe3+-Fe(OH)2+ and HSO4−-SO42− could maintain a pH of approximately 1, while the dissolution of alunogen and the hydrolysis of other Al sulfates may be responsible for a pH values of approximately 4 (Querol et al. 2011). Most samples retrieved from the fissures indicated near-neutral pH, and the significant leachable content of gypsum (high Ca and SO42−) may be unaffected by pH. A small number of samples collected from the fissures even indicated alkalinity, and NH4Cl and organic condensates could account for this alkalinity (Querol et al. 2011). Sulfate in area A could be generated by the reaction of SO2 with the soil moisture, which is confirmed by the high SO2 concentration in the area (Fig. 7). The average soil pH was 7.19 in area B, and significantly lower than that in area A (Fig. 8). This indicates that the soil cover weakens the acidity, which plays a positive role in reducing environmental hazards. Compared to area A, the total salt contents were significantly lower than those in area B (Fig. 8), and the contours of the SO2 concentration were relatively sparse (Fig. 7). SO2 was even not detected in the degraded vegetation area (Fig. 7). The soil pH in area B was relatively higher than that in area C (Fig. 8), but the t test showed that there was no significant difference between these two groups.

Fig. 7
figure 7

The soil pH and total salt contents distribution in the research zone

Full size image
Fig. 8
figure 8

pH and total salt contents comparison of the research zone

Full size image

Analysis of the organic matter contents

The average organic carbon contents in the degraded vegetation area are the highest, with an average content of 17.62 mg/g, which is significantly higher than that in the other areas (Fig. 9). This probably occurs because the carbonated particles generated via vegetation withering increase the soil organic carton contents by mixing with the soils (Cui et al. 2012; Zhao et al. 2016). In comparison, the organic matter contents in the Huangbaici coal fire area decreased with increasing soil temperature, because the heat effect of the coal fire digested the organic matter (Wang et al. 2020). We draw a nearly opposite conclusion: the average organic carbon contents in the high-temperature zone (area A) seem be higher than those in the surrounding area, and the mean average organic carbon contents in area A are significantly higher than those in area B (Fig. 9). The particles released through the fissures within the strata from the underground coal fires may increase the organic carbon contents (Wang et al. 2019). In addition, the landfilling with yellow sandy soil in area B affects the organic carbon contents because the sandy soil contains relatively low organic contents.

Fig. 9
figure 9

Organic carbon contents comparison of the investigated area

Full size image

Limitations and future research directions

  1. 1)

    Breathing cycles are the common phenomena of certain coal fires. Engle et al. observed that the CO2 vent emissions in Wyoming coal fires significantly differ between two consecutive days, while the CO and H2S showed no significant differences (Engle et al. 2012). Dindarloo et al. found that the CO concentration increased in the middle of the day when outside air temperature increased, and declined in the afternoon when the outside air cooled (Dindarloo et al. 2015). The temporal variability in diffuse gas emissions remains a research gap (Engle et al. 2013). Therefore, the temporal variability of in vent emissions and diffuse gas emissions in coal fire areas should be considered.

  2. 2)

    CO2 generally stems from three sources in coal fire areas: native CO2, CO2 produced from coal combustion, and CO2 produced from the oxidizing of native CH4 (Ide and Orr 2011). Ide and Orr used 13C isotope signatures to determine the fractions of CO2 emitted from coal combustion in coal seams (Ide and Orr 2011). Carras et al. found that the emission rate of greenhouse gases originating from low-temperature oxidation of coal waste is similar to the emission rates due to biological activity (Carras et al. 2009). In our study, CO2 and CH4 in areas A and B were qualitatively inferred to originate from coal combustion based on their relatively high concentrations and Pearson correlation analysis, as discussed in the “Analysis of the belowground gas concentration” section. A portion of CH4 and CO2 may originate from microbial activity in the degraded area. Quantitative analysis of the source is required in the research zones, which could determine contribution of the greenhouse gases emitted from coal combustion to the pollution of the local environment.

3) Soil chemical properties are sensitive to gas and heat intrusions as discussed in the “Limitations and future research directions” section, such as the soil pH and organic carbon. The mobility of trace metals and changes in microbial communities is reliably related to the soil pH and organic carbon (Rasheed et al. 2013). For example, the soil Hg and Cd contents were related to the organic matter (Wang et al. 2019), and Fe3+ concentration was relatively high near the vents (Querol et al. 2011). Tobin-Janzen found that the ribotype diversity of microbial populations decreased within individual boreholes with increasing temperature (Tobin-Janzen et al. 2005). Therefore, whether the change of soil biochemistry properties could be adopted a complementary index to monitor underground coal fires deserves further research.

Conclusion

Surveys were conducted in the Wuda coal fire area to investigate the impact of underground coal fires on the surface landform, soil gas emissions, and soil properties. Backfilling of sandy soil imposes certain effect on gas emissions and soil properties. These observations lead to the following conclusions:

  1. 1)

    Little vegetation occurs in the area seriously affected by underground coal fires, such as in area A and area B. Fissures, cracks and vents, and the sponges are the main fire-induced landforms in the research zone.

  2. 2)

    A comparison of the components and concentrations of gases is performed among area A, area B, and the degraded vegetation area. The CO2 and CO concentrations peak at 11 vol% and 6008 ppm, respectively, in area A, far exceeding OSHA safety limits. In area B, the CO2 and CO concentrations are relatively low, with average concentration of 9.84 vol% and 3530 ppm, respectively. SO2, H2S, C2H6, and C2H4 highly concentrate in area A, ranging from 0.72 to 26.4, 2 to 100, 6 to 128, and 2 to 74 ppm, respectively. The soil CO2 concentrations are correlated with CH4 but seem to be unrelated to CO. Low soil O2 concentrations damage the healthy root environment for soil O2 is displaced by CO2. CO and CH4 are detected near the ground surface, while the other gases are below the detection limits. Backfill of soil mitigates the emission of SO2, H2S, C2H2, and C2H6, but does not seen to reduce emissions of the greenhouse gases in the research zone. The near-surface gas concentration mainly originates from the obvious openings. Although coal-fire gas emissions are diluted by air, the maximum of near surface CO concentration exceeds the OSHA 8-h exposure limit, which is a hazard to geologists at the site.

  3. 3)

    Thermal surface and soil moisture anomalies are mainly dominated by obvious openings. There seems no relationship between the surface temperature and gas concentration due to solar heating. The sponges exhibit the lowest soil pH (0.95–1.99), followed by the vents and cracks. A negative correlation is found between the soil pH and the total salt contents in area A (r=0.859, p<0.01). Hydrolysis of Al and Fe sulfates originating from soil salts accounts for the relatively low pH values. The particles released through the fissures may increase the organic carbon contents in area A, while the carbonated particles stemming from vegetation withering cause an increase in the organic carbon contents in the degraded vegetation area. Compared to the area A, the soil in area B attain relatively high soil pH values and relatively low organic carbon contents, which suggests that the soil cover helps improve the quality of the soil environment.