- • Cracks caused by mining under the potential slide rock mass ruptured the aquiclude.
- • Acid rain supplied nutrients to organisms decomposing organic matrix in the shale.
- • Acid rain reduced strength of the shale layer on which the failure surface formed.
Air pollution in China often exceeds “unhealthy” levels, but Chinese air is not only a threat from being breathed: the pollutants may also be causing fatal landslides. Very acid rain from severe air pollution falls widely in southwest China, where coal is a major energy source. We discuss where acid rain may provide an unsuspected link between mining and the fatal 2009 Jiweishan landslide in southwest China; it may have reduced the strength of a thin, calcareous, black sapropelic shale in Jiweishan Mountain by removing cementing carbonate minerals and sapropel matrix. Mining beneath the potential slide mass may not have directly triggered the landslide, but collapse of abandoned adits drained a perched aquifer above a regional black-shale aquiclude. Inflow of acid, oxygenated water and nutrients into the aquiclude may have accelerated the reduction of strength of the weakest rocks and consequently led to rapid sliding of a large rock mass on a layer of weathered shale left composed largely of soft, and slippery talc.
The rapid industrial development that brought an economic miracle to China since 1978 has come at a major environmental cost: the level of air pollution in China can often be described as “unhealthy” or “very unhealthy” (for example see http://aqicn.org/city/beijing). One of the consequences of air pollution is acid rain. Especially in southwest China, where coal combustion is one of the main energy sources, the air pollutants in addition to fine particulates are SO2 and NO2 (Hu et al., 2010), which are major contributors to acid rain. Between 1986 and 2014 in Chongqing, for example, the minimum rainfall pH was 2.8 in 2012 and the average was between 4.3 and 5.0. The proportion of rain-days with acid rain (pH <7.0) in the yearly total number of rain-days has varied between 33.3 and 95.0% (Hu et al., 2010; CEPB, 1998–2014) (Fig. 1). Table 1 lists the main acid-rain components in Chongqing (Feng and Ogura, 1998; Liao and Tang, 2000; Tang et al., 2013).
|Collected from references||12.08–4.83||0.93–6.29||0.98–3.62||0.49–1.40||0.61–4.25||2.45–26.7||0.24–1.52|
Some spring-water comes from acid-sulfate weathering of pyrite in the shale interbeds.
Although some scientists have shown that rock physical properties can be gradually altered by acid rain reacting with constituent mineral such as calcite and illite (Gupta and Ahmed, 2007; Taghipour et al., 2015) and by microbial weathering (Buss et al., 2005; Wilson, 2005; Shelobolina et al., 2012), few studies have discussed a direct relationship between acid rain and a landslide. Zhao et al. (2011) studied the action of acid rain on clay minerals from the slip zone of a reservoir landslide in the Three Gorges Reservoir in China, and found a transformation of illite to smectite which decreased the shear strength of the sliding-zone soil. However, they did not further link their laboratory experiments with a failure mechanism of a landslide. We suggest in this paper that the disastrous 2009 Jiweishan rock avalanche in Wulong County, Chongqing (Fig. 2), was caused by the coupled actions of tectonism, mining activities, karstification, acid rain, and microbial decomposition. Acid rain may have played a key role in reducing the sliding resistance along the potential failure surface: it supplied nutrients to microorganisms capable of decomposing the matrix of sapropel, and it was capable of dissolving calcite cement from the interbedded shale layer on which the landslide failure surface eventually formed. Without calcite and organic matter, the shale becomes a predominantly soft talc powder, which has low frictional resistance (Moore and Lockner, 2008).
Our re-investigation of this already much studied landslide was spurred by the apparent absence of an identified landslide-triggering event: an absence possibly due to a long delay of many decades between obvious potential triggers and the occurrence of the disastrous landslide. The apparent lack of an initiating event led to an unfortunate misdiagnosis of the potential outcome. The diagnosis was for continuing dangerous rockfalls affecting a limited area at the foot of the cliff where a small town was located. Potentially more than 700 lives were saved by the timely relocation of Tekuang Town, but 74 lives were lost when a failure occurred that was very much larger than predicted, and had a long runout down valley, through occupied mine facilities and farmland (Yin et al., 2011).
Using the benefit of hindsight, we searched for a subtle landslide explanation involving slow, progressive loss in strength over a wide area. In this, we sought evidence from a wide range of disciplines eventually including landslide science, mining technology, geology, biology, biochemistry and atmospheric pollution.
The hypothesis we present in this paper of a plausible progressive failure mechanism for the Jiweishan landslide suggested an unexpected developing landslide risk that may be emerging in southern China. This risk is associated with widely distributed carbonate rocks with interbedded shales and the escalating problem of air pollution in China. But the developing risk may not be unique to China: similar lithological combinations of carbonates with shale interbeds occur in North America, Europe and elsewhere in Asia, in heavily industrialized nations with varying degrees of polluted air and problems with acid rain.
2. Widely distributed limestone slopes in southwest China
Many rock slopes in southwest China comprise gently to tightly folded, thick-bedded Permian to Triassic limestone and dolomite with interbedded shale units. These carbonate rocks often overlie mudstone or siltstone, which contain layers of ore-grade hematite or coal with thicknesses up to 2 m (Fig. 3). Of course, the hematite and coal are mined (Xu et al., 2010; Yin et al., 2011; Wang et al., 2015).
In the humid, subtropical climate of southwest China, a karst landscape has developed on the thick carbonate rocks following their folding and uplift since the late Cretaceous period (Liu and Xie, 2010). On the more erodible, softer units, the land surface has eroded more rapidly to leave steep limestone cliffs with slope angles >60° and even near vertical. On the cliff faces, contact surfaces between the shales and limestones are particularly unstable (Yin et al., 2011; Wang et al., 2015) leading to many rockfalls, some block slides and very rarely, large rock avalanches.
Mining of the ore-grade hematite (as below Jiweishan Mountain) and coal layers (Yin et al., 2000) has been active since the 1920s, leaving a legacy of mined-out areas, which in places affects the stability of their overlying rock mass. At Jiweishan Mountain, collapse of old mines had deformed the overlying rock mass (Fig. 3). Mining changes the state of stress around mined-out areas (Parise and Lollino, 2011; Zheng et al., 2015) and deformation here has involved collapse of the roof and wall of gobs and tunnels. Cracks have propagated into the overlying rock, even to the ground surface. Drainage through cracks greatly changed the flow of groundwater (Libicki, 1982; Rapantova et al., 2007). Groundwater previously trapped above shale aquicludes (Yin et al., 2011) then drained into the mines through cracks. Until the facilities were destroyed by the rock avalanche, water was pumped from the active adit 1 to keep it dewatered.
The largely historical mine collapse (from the late 1950s) was previously reasoned insufficient to have triggered the large failure at Jiweishan Mountain on its own (Yin et al., 2011), and we do not dispute this finding. The Chinese government, however, has strengthened the management and inspection of mines in recent years, including minimizing the deformation of overlying rock masses through strict controls.
The troublesome shale interbeds are between 10 and 30 cm thick, and composed of calcite, quartz, clay minerals and sapropel (fine organic material) in somewhat variable proportions. Fig. 4A shows the X-ray-diffraction-determined mineral components of the shale immediately beneath the basal sliding surface of the Jiweishan rock avalanche. Some layers contain more abundant shell fragments and others more sapropel. Using a CS844 carbon &sulfur analyzer, we measured total organic carbon (TOC) in 4 samples from around the basal failure surface. TOC ranged between 0.46 and 1.9%, but a TOC of 13% is also reported from the basal shale (Deng et al., 2014). Tests using extraction and a Leica Transmission/Reflected Light Fluorescence Microscope indicate an organic-rich matrix in the shale that is mainly sapropel (coming from microorganisms such as algae, fungi and bacteria) and vitrinite (coming from advanced plants and animals). Centimeter-scale vitrinite lenses occur sparsely. Folding of southwest China’s thick carbonate sequence has induced many slickensided bedding-plane shears widely in the weak shale interbeds and they are conspicuous at Jiweishan Mtn (Fig. 4B). For this reason, such shale interbeds would be largely at residual shear strength, were it not for some post-deformational diagenesis (Deng et al., 2014). Any process which undoes the diagenetic cementation can facilitate further deformation on the shale interfaces, and possible sliding failure of the overlying rock mass.
3. Effects of acid rain
We suggest that percolation of acid rain through the limestone into the shale has had two effects on the calcareous black shale as follows:
(1) Acid rain has supplied mineral nutrients to the micro-organisms such as bacteria and fungi that utilize the organic carbon in the shale as an energy source. In addition to the carbon energy source and trace amounts of phosphate in the shale (0.004% P2O5at Jiweishan rock avalanche by X-ray fluorescence), the acid rain brings oxygen and other nutrients including nitrogen, potassium and sulfur (Rogoff and Wender, 1957; Du et al., 2008) (Table 1). Various microorganisms can decompose the organic material (Wang and Shao, 2013; Austin and Callaghan, 2014), even degrading coal under aerobic conditions (Ackerson et al., 1990; Sekhohola et al., 2013). Microorganisms are widely used in the coal industry to transform solid coal to fluid coal products (Wilson et al., 1987; Klasson et al., 1993). A comprehensive DNA test on a shale sample taken from immediately beneath the landslide sliding surface indicated hundreds of genera of microorganisms currently present in the shale: the main genera were Bacillus, Virgibacillus and Lactococcus, Buckholderia, Chroococcidiopsis, Leptolyngbya and other two genera belonging to Bacillaceae and Xenococcaceae families (our tests could not further determine their genus). The former six genera of microorganisms are either anaerobic or aerobic bacteria or both, and capable of decomposing sapropel and vitrinite. We cannot confirm that similar bacteria were present before the landslide, but it is a reasonable assumption that such bacteria have always been present in the local environment.
Groundwater drainage through cracks through the shale ensures a constantly renewed supply of nutrient fertilizer and oxygen to facilitate the process of decomposition where previously the shale was a major aquiclude inhibiting karst development in the underlying limestone strata (Yin et al., 2011).
(2) Any surplus acid in water reaching the shale through tens of meters of karstified limestone dissolves calcite from the shale. We conducted a simple experiment immersing a sample of shale in dilute hydrochloric acid (pH 3.0). The acid reacted with calcite (largely microscopic shell fragments) producing constant effervescence for several hours. There was also a mild odor of H2S indicating a minor presence of probably pyrite. The shale after reaction did not disintegrate but was much weaker and, on microscope inspection, appeared to be a black sponge, full of micro-pores. Comparison of Ca2+ in input rainwater with output spring water taken from the foot of the cliff below the source area of the Jiweishan rock avalanche indicated a marked increase in Ca2+ from 2.45–26.7 mg/l in the rainwater to 50.59 mg/l in the spring water through dissolution of calcite from the limestone and shale (Table 1). The pH of the spring water currently lies between 6.5 and 7.0 and so is still acid despite passage through limestone. We were not able to determine whether the residual acid content originated from rain or oxidation of the pyrite. Nor is there a record of the pre-landslide spring-water pH.
Dissolution of the calcite and decomposition of the organic component of the shale has obvious strength-reducing consequences. Together with recrystallized calcite, the sapropel formed part of the matrix binding the mineral particles (Lu et al., 2001; and our experiment above). The organic matrix is also a reason for the apparent more ductile behavior of the shale, and why the shale tensile and compressive strengths reduce with increased organic content (Kumar et al., 2015). Without calcite and organic material, the basal shale at the Jiweishan rock avalanche would be composed largely of detrital talc. Scanning electron microscope images of shale taken recently from the source area of Jiweishan rock indicate densely distributed micro-pores (Fig. 4C). This surface, however, has been etched by acid rain and the action of micro-organisms only since June 2009 when it was first exposed to them. The original interface on which sliding initiated was destroyed by abrasion during the landslide, and so we cannot directly test the hypothesis that the specific loss of resistance to sliding at the base of this landslide was due to loss of matrix cement.
4. Failure process and key role of acid rain in Jiweishan rock avalanche
The above analysis suggests that failure of the Jiweishan Mountain slope was due to the successive actions of tectonism, karstification, mining activity, acid rain and microbial decomposition. The conception and failure of this slope may be described as follows:
(1) The strata in Jiweishan Mountain dip to the Northwest on the northeast limb of the Zhaojiaba anticline. Bedding shear induced by this tectonic folding reduced the shear strength of the shale to a residual value. Shear tests on the shale indicate that bedding shear can reduce the shear strength by about 30% from and cohesion to and (Deng et al., 2014).
(2) Cracks C1, C2 and a karstified zone (Fig. 5) begin to develop and now define the landslide boundaries. Dating of soils from cracks C1 and C2 (Peng, 2011) indicates silt and sand deposition in karst cavities at least 26 thousand years ago. Before the rapid failure event, cracks C1 and C2 had opened and less than 40% of the area of the karstified zone remained intact (Yin et al., 2011).
(3) Mining of hematite began in the 1920’s and had caused cracks to develop in the overlying rock mass before 1969 (Fig. 3). The mined-out area had reached 52400 m2under the front part of the landslide by January 2008 through mining since 1969 (COTSGT, 2008), and crack C1 on the cliff surface above abandoned Adit #4 had propagated through the failed rock mass and reached 2 m wide by 1969 (Yin et al., 2011). Seepage through the failure surface and into the mined-out areas was adequately maintained by a mean annual precipitation of 1094 mm. Measured groundwater flow into the active mine varied between 26.5 and 49 m3/h depending on recent rainfall until January 2008 (COTSGT, 2008).
(4) Air pollution became severe since 1978, with increasingly acid water reaching the limestone-shale interface with potential to progressively reduce its strength by decreasing the content of organic material and calcite. From here, because the landslide took away the initial sliding surface, we can only speculate that the sliding resistance along the top of the major shale aquiclude was progressively reduced until the overlying rock mass began to slide on a saturated layer of talc.
Acid rain thus may have played a key role in the landslide. Numerical simulation (Xu et al., 2010; Yin et al., 2011) suggests that the deformation induced by the mining activities alone was insufficient to initiate the large-scale rock failure. In the Jiweishan rock avalanche source area, the dip direction and dip angle of the shale are 345° and 19°, respectively (Yin et al., 2011). The sliding direction of the rock mass was 15° east of the dip direction of the strata. Therefore, the apparent dip of the shale along the sliding direction was 18.4°, a little steeper than its angle of internal friction (17° after bedding shear). With the additional resistance of the surviving rock bridges in the karstified zone and along the surface of crack C2, the failed rock mass should not have initiated without further decrease in sliding resistance on the failure surface (Fig. 5). As an interface of talc, however, the internal friction angle could have been as low as between 9 and 13° (Moore and Lockner, 2008).
5. Discussion and conclusions
Although mining under the potential slide rock mass could not induce the landslide alone, cracks caused by collapse of old mines are known to have ruptured an aquiclude. Draining of the previously perched water table changed the flow of groundwater in the overlying rock mass and increased the flow reaching the limestone-shale interface. The increased flow brought oxygenated acid water, rich in nutrients for growth of microorganisms to the interface where failure is known to have initiated. The interface was thereby weakened by decomposition of the shale’s organic matrix and dissolution of calcite cement. Although we suggest that mining accelerated the weakening, we do not see it as an essential element. The essential element would appear to be the action of acid rain on calcareous, black shale, but the effect may be appearing earliest where groundwater flow is accelerated by mine drainage.
Acid rain from air pollution in southwest China is still severe and may be further aggravated because plans for new coal plants in the next decade could double China’s coal consumption (Rohde and Muller, 2015). Therefore, we might anticipate further failures of this kind of slope, even in the absence of mining.
We have presented a hypothesis of how acid rain may have provided a slowly developing failure mechanism of the Jiweishan rock avalanche and suggested that the same process may be degrading the stability of other locations on the widely distributed limestone rocks and their associated bituminous shale interbeds in southwest China. The problem may not be unique to China, because similar lithological combinations occur in other heavily industrialized nations across North America, Europe and other parts of Asia, and microbial decay is endemic on our planet. The mechanism may indicate an unanticipated risk from air-pollution-induced landslides.
Both authors made equal contributions to this paper. Ming Zhang’s work was supported by the funding from the National Natural Science Foundation of China (41472264, 41772334) and China Geological Survey project (DD20179609). Mauri McSaveney’s work was supported by GNS Science and the State Key Laboratory of Geohazard Prevention and Geoenvironment Protection, Chengdu University of Technology, but received no specific funding.
- Ackerson et al., 1990
M.D. Ackerson, N.L. Johnson, M. Le, E.C. Clausen, J.L. GaddyBiosolubization and liquid fuel production from coalAppl. Biochem. Biotechnol., 24 (1990), pp. 913-928
- Austin and Callaghan, 2014
R.N. Austin, A.V. CallaghanMicrobial enzymes that oxidize hydrocarbonsFront. Microbiol (2014), 10.3389/fmicb.2013.00338
- Buss et al., 2005
H.L. Buss, M.A. Bruns, M.J. Schultz, J. Moore, C.F. Mathur, S.L. BrantleyThe coupling of biological iron cycling and mineral weathering during saprolite formation, Luquillo Mountains, Puerto RicoGeobiology, 3 (2005), pp. 247-260
- CEPB, 1998–2014
Chongqing Environmental Protection Bureau (CEPB)Bulletin of the environmental condition of Chongqing
- COTSGT, 2008
- Chongqing One Three Six Geological Team (COTSGT), 2008. Report of risk assessment on geological hazards caused by mining activity in Gonghe iron ore plant, Wulong, Chongqing. Assessment Report for Enterprise Production, 27 p. (in Chinese).
- Deng et al., 2014
M. Deng, Q. Xu, B. Han, X. Zhu, G. Cai, G. ZhengThe microscopic structure and rheological characteristics of the slip zone soft rock of Jiweishan landslide in Wulong of Chongqing: ChinaMt. Res., 32 (2014), pp. 233-240
- Du et al., 2008
M. Du, H. Chen, S. JiangBiological liquefaction characteristics of Jurassic weak & non-stick coal in Hengshan, North Shaanxi ProvinceJ. Coal Sci. Eng. China, 14 (2008), pp. 520-522
- Feng and Ogura, 1998
Z. Feng, N. OguraImpacts and control strategies of acid deposition on terrestrial ecosystems in Chongqing area, China: overviews of the cooperative study between Japan and ChinaAdv. Environ. Sci., 6 (1998), pp. 1-8
- Gupta and Ahmed, 2007
V. Gupta, I. AhmedThe effect of pH of water and mineralogical properties on the slake durability (degradability) of different rocks from the Lesser Himalaya, IndiaEng. Geol., 95 (2007), pp. 79-87
- Hu et al., 2010
H. Hu, Q. Yang, X. Lu, W. Wang, S. Wang, M. FanAir pollution and control in different areas of ChinaCrit. Rev. Environ. Sci. Technol., 40 (2010), pp. 452-518
- Klasson et al., 1993
K.T. Klasson, M.D. Ackerson, E.C. Clausen, J.L. GaddyBiological conversion of coal and coal-derived synthesis gasFuel, 72 (1993), pp. 1673-1678
- Kumar et al., 2015
V. Kumar, C. Sondergeld, C.S. RaiEffect of mineralogy and organic matter on mechanical properties of shaleInterpretation, 3 (2015), pp. SV9-SV15
- Liao and Tang, 2000
Z. Liao, L. TangCauses and countermeasures of acid rain in ChongqingEnviron. Protect. Sci., 26 (2000), pp. 1-5
- Libicki, 1982
J. LibickiChanges in the groundwater due to surface miningInt. J. Mine Water, 1 (1982), pp. 25-30
- Liu and Xie, 2010
Y. Liu, S. XieAnalysis on influence factors of the development of karst landform in Southwest ChinaJ. Xuchang Univ., 29 (2010), pp. 143-147
- Lu et al., 2001
F. Lu, L. Sang, J. Wu, Q. LiaoPetrologyGeological Publishing House, Beijing (2001)
- Moore and Lockner, 2008
D.E. Moore, D.A. LocknerTalc friction in the temperature range 25°–400 °C: relevance for Fault-Zone WeakeningTectonophysics, 449 (2008), pp. 120-132
- OZSGT, 2003
- One Zero Seven Geological Team of Chongqing Geological and Mineral Resource Exploration and Development Bureau (OZSGT), 2003. Mining plan of Gonghe iron ore plant. Report for Enterprise Production. 41 p. (in Chinese).
- Parise and Lollino, 2011
M. Parise, P. LollinoA preliminary analysis of failure mechanisms in karst and man-made underground caves in Southern ItalyGeomorphology, 134 (2011), pp. 132-143
- Peng, 2011
G. PengThe Research of the Formative Condition and Instability Mechanism of the Landslide Controlled by the Key Blocks in Southwest Mountainous Area – The Cocktail Mountain Landslide in Chongqing Wulong for ExampleChengdu University of Technology (2011)
- Rohde and Muller, 2015
R.A. Rohde, R.A. MullerAir pollution in China: mapping of concentrations and sourcesPLoS ONE, 10 (2015), Article e0135749
- Rogoff and Wender, 1957
M.H. Rogoff, I. WenderThe microbiology of coal. I. Bacterial oxidation of phenanthreneJ. Bacteriol., 73 (1957), pp. 264-268
- Sekhohola et al., 2013
L.M. Sekhohola, E.E. Igbinigie, A.K. CowanBiological degradation and solubilisation of coalBiodegradation, 24 (2013), pp. 305-318
- Shelobolina et al., 2012
E. Shelobolina, H. Xu, H. Konishi, R. Kukkadapu, T. Wu, M. Blothe, E. RodenMicrobial lithotrophic oxidation of structural Fe (II) in biotiteAppl. Environ. Microbiol., 78 (2012), pp. 5746-5752
- Rapantova et al., 2007
N. Rapantova, A. Grmela, D. Vojtek, J. Halir, B. MichalekGroundwater flow modeling applications in mining hydrogeologyMine Water Environ., 26 (2007), pp. 264-270
- Taghipour et al., 2015
M. Taghipour, M.R. Nikudel, M.B. FarhadianEngineering properties and durability of limestones used in Persepolis complex, Iran, against acid solutionsBull. Eng. Geol. Environ., 75 (2015), pp. 967-978
- Tang et al., 2013
X. Tang, Y. Wang, Y. Wang, H. Zhang, P. Guo, B. HuCanopy leaching characteristics of typical forests during acid rain at Jinyun Mountains, ChongqingJ. For. Res., 26 (2013), pp. 548-553
- Wang et al., 2015
L. Wang, B. Li, Z. Feng, Y. Gao, S. ZhuThe failure patterns and their formation mechanisms of large perilous rocks in thick layered limestone masses in Yangjiaochang Town, Wulong CountyActa Geol. Sin., 89 (2015), pp. 461-471
- Wang and Shao, 2013
W. Wang, Z. ShaoEnzymes and genes involved in aerobical kane degradationFront. Microbiol., 4 (2013), p. 116
- Wilson et al., 1987
B.W. Wilson, R.M. Bean, J.A. Franz, B.L. Thomas, M.S. Cohen, H. Aronson, E.T. GrayJr.Microbial conversion of low-rank coal: characterization of biodegraded productEnergy Fuels, 1 (1987), pp. 80-84
- Wilson, 2005
M.J. WilsonWeathering of the primary rock-forming minerals: processes, products and ratesClay Miner., 39 (2005), pp. 233-266
- Xu et al., 2010
Q. Xu, X. Fan, R. Huang, Y. Yin, S. Hou, X. Dong, M. TangA catastrophic rock-slide-debris flow in Wulong, Chongqing, China in 2009: background, characteristics, and causesLandslides, 7 (2010), pp. 75-87
- Yin et al., 2000
Y. Yin, H. Kang, Y. ZhangStability analysis and optimal anchoring design on Lianziya dangerous rockmassChin. J. Geotechn. Eng., 22 (2000), pp. 599-603
- Yin et al., 2011
Y. Yin, P. Sun, M. Zhang, B. LiMechanism on apparent dip sliding of oblique inclined bedding rockslide at Jiweishan, Chongqing, ChinaLandslides, 8 (2011), pp. 49-65
- Zhao et al., 2011
Y. Zhao, P. Cui, L. Hu, T. HueckelMulti-scale chemo-mechanical analysis of the slip surface of landslides in the Three Gorges, ChinaSci. China, Technol. Sci., 54 (2011), pp. 1757-1765
- Zheng et al., 2015
D. Zheng, J.D. Frost, R.Q. Huang, F.Z. LiuFailure process and modes of rockfall induced by underground mining: a case study of Kaiyang phosphorite mine rockfallsEng. Geol., 197 (2015), pp. 145-157