1 Introduction

Sinkholes induced by mine dewatering are a severe geohazard in karst areas [1, 2]. Identification of potential responsible factors and determination of the extent of damage have been challenging [3], and still today are subject of many researches [4,5,6]. Since the mining zone of influence act (ZOI) was passed in Maryland, United States in the 1990s, the extent of influence related to a quarry dewatering has been evaluated to help determine quarry’s responsibility for sinkholes [7, 8]. In addition, karst collapses still happened in roads and commercial areas outside of the quarry ZOI [9,10,11]. The graphical and probability integral methods are often used to determine the zone of influence for mining-induced subsidence [12]; however, they are not suitable for karst collapses induced by mine dewatering. According to the Chinese geotechnical engineering code, the maximum radius of influence for dewatering in a foundation pit is three times the depth of the foundation pit. This empirical method grossly underestimates the influence area in karst areas and may not identify the responsible factors for sinkhole occurrence. Although the characteristics of collapse are discussed in the literature [13], the root causes have not been analyzed, especially where mine dewatering or foundation pit dewatering took place [14, 15]. Sinkholes discussed in this paper are those induced by dewatering activities. However, sinkholes can also be caused by other factors such as earthquakes, blasting, vibration, and loading. Lei et al. [6] provides detailed discussions on sinkhole formation processes associated with these factors.

The sinkholes in Chaoshan of Tongling City, Anhui Province, China, had occurred for more than 20 years, and extensive geological work including drilling and geophysical exploration have been carried out [16]. Figure 1 shows the regional geology of the study area and the sinkhole distribution. There is an apparent temporal correlation between those collapses and mine dewatering activities. It is generally believed that mine drainage is the main factor inducing these karst collapses. Nevertheless, these conclusions come from empirical experience and speculation, and are not substantiated with real monitoring data.

Fig. 1
figure 1

Regional geology of Chaoshan study area (see Table 1 for legend of lithology)

Full size image

To overcome these shortcomings, real-time monitoring of karst groundwater-gas pressure was conducted in combination with geochemical and isotopic analyses. After recognition of two groundwater subsystems, the authors defined the main engineering activities that induced the collapses, and provided a basis for management, development and utilization of the mines in this region.

2 Geological setting

2.1 Regional geological conditions

The Caoshan sinkhole area is located in Anhui province of China with subtropical humid monsoon climate. The annual average temperature is 16.2 ℃. The average annual precipitation is 1376 mm with rainy season in the spring and summer. The highest annual rainfall of 2174 mm occurred in 1983, and the maximum monthly and daily precipitations are 926.9 mm and 204.4 mm, respectively.

This area is characterized by hilly and plain landforms entrenched by rivers. Ground elevation varies from 12 to 28.6 m above mean sea level (a.m.s.l.). The Yanghe River is the largest river in the study area. Structurally, the Qingshan anticline is the main fold structure with a NE strike. The Zhucun syncline is on the southeast side of the Qingshan anticline (Fig. 1).

Table 1 presents the lithology, which consists of Quaternary and Triassic formations. The magmatic rocks (δ) are widely distributed in the study area.

Table 1 Strata in Chaoshan area
Full size table

The Nanlinghu Formation and Dongmaanshan Formation form karst aquifers in the study area. Their characteristics are as follows:

  • Nanlinghu Formation (T1N) is distributed in the center of the study area. Lithology consists of limestone and marble. Karst is intensely developed with 63% of boreholes drilled in this formation encountering cavities. Borehole pum** tests indicate that the specific yield is 0.44 l/s.m, and the hydraulic conductivity is 0.226 m/d. Capacity of water storage is medium. The water chemistry belongs to Ca-SO4-HCO3 type with salinity varying from 0.1 to 0.3 g/l. Unfilled karst cavities were primarily observed at elevations from 36 to 157 m. Water inrush occurred in shallow roadway and did not occur in deep roadways. Karst cavities are hydraulically connected through structural fractures.

  • The Dongmaanshan Formation (T2d) is in the eastern part of the area. The lithology consists of marble and dolomite. Karst is intensely developed with 60% of borehole encountering cavities. Most of karst cavities are filled with clay. Based on borehole pum** test, the specific yield is 3.58 l/s.m, and the hydraulic conductivity is 4.6 m/d at 2.58 m of drawdown. Capacity of water storage is strong. The water chemistry is Ca-SO4-HCO3 type with salinity of 0.453 g/l.

The types of groundwater are typical of Quaternary pore water and karst water in the area. Precipitation is the main source of recharge. There is a strong hydraulic connection between surface water and karst water. The groundwater flows from southwest to northeast. Mine dewatering has resulted in groundwater cones of depressions.

2.2 Anthropogenic engineering activities

Because of strong tectonic activities, copper, gold, and pyrite are widely distributed in the study area. Mining of these minerals in three mines is the main human activity (Fig. 1). There is also a water supply wellfield (Quantang water supply) in the area. Table 2 shows characteristics of the mines and the wellfield. The **nhua Copper Mine has experienced three stages of mining, occurrence of collapse, and closure. Affected by mine dewatering, the groundwater level in the area has dropped significantly over two decades. According to reports on the water supply wellfield in Dongmaanshan Formation (T2d), the groundwater level elevation was − 2.1 m a.m.s.l. in July 1996, and − 25 m in July 2014, with a decline of approximately 23 m.

Table 2 Summary of engineering activities in the area
Full size table

2.3 Characteristics of karst sinkholes

The Chaoshan sinkhole area encompasses two villages: **nmin village and Chaoshan village. There are 86 sinkholes within an area of 2 km2. The direct economic loss was approximately 7 million Yuan, and 107 households were affected by the karst collapses.

The sinkholes occurred in three periods: 1990s, 2006–2007, and 2011–2016. The total number of sinkholes in **nmin village is 34, which first appeared in September and October 1990 and peaked in 1997 and 2006–2007, respectively. The remaining 52 collapses occurred in Chaoshan village from 2011 to 2016. The sinkhole shapes are usually circular or oval on surface and range from 4 to 26 m in diameter and from 0.3 to 10 m in depth. Earth fissures are common around the collapse pits.

3 Research methods

The karst aquifers of Nanlinghu Formation and Dongmaanshan Formation are distributed in the core of Zhucun syncline with a total thickness of more than 400 m. Previous studies have considered them as one karst water system [15, 16]. Implications of such an understanding is that all three mines are potentially responsible for occurrence of these sinkholes. The focus of this paper is to refine the karst groundwater system through field surveys of geological cross-sections, monitoring of karst groundwater dynamic conditions, and hydrochemical isotopic characteristics.

3.1 Field survey and measurement of geological cross-sections

Karst phenomenon exposed by drilling is limited in a covered karst area. Bedrock outcrops of T1N and T2d occur in the southwest of the study area. The strata strike, direction of groundwater, and analysis on the karst characteristics suggest an aquiclude between T1n and T2d Formations.

3.2 Hydrochemical isotopic analysis

Hydrogen and oxygen isotopes are the elements making up water and directly reflect the circulation process of groundwater. They are ideal environmental tracers for water circulation. The analysis results of these stable isotopes are expressed by the delta value, which is the relative difference between the isotope ratio of the sample and the standard [17, 18]. The average ocean water is an international standard for comparison, often referred to as Standard Mean Ocean Water (SMOW). The water sampling points is extended from the study area to 5 km southwest so that the springs and underground rivers in the Nanlinghu and Dongmaanshan Formations were sampled.

3.3 Karst groundwater-gas pressures monitoring

Six groundwater monitoring wells were installed in the sinkhole area after the November 2011 collapse event. Monitoring data were collected over 3 months with pressure transducers at a frequency of once a day. The monitoring effort was to better understand the hydraulic connection between T2d and T1n Formations, and the dewatering ZOI related to the three mines, i.e., **nhua Copper Mine, Lion Pyrite Mine, and Chao Gold Mine.

Ten additional monitoring wells were installed from 2014 to 2020 to measure karst groundwater-gas pressures at a much high frequency of 10-min intervals. Figure 2 shows relocations of these monitoring wells. Six monitoring wells were placed within groundwater runoff zone in the Dongmaanshan Formation (T2d) near **nhua Copper Mine, whereas four monitoring wells were in Nanlinghu Formation (T1n) and near the Chao Gold mine and Lion Pyrite Mine. Construction details of the monitoring wells are presented in Table 3.

Fig. 2
figure 2

Distribution of karst groundwater-air monitoring wells (see Table 1 for legend of lithology)

Full size image
Table 3 Features of monitoring borehole
Full size table

4 Result discussions

4.1 Division of karst groundwater subsystems

The results of hydrogen and oxygen isotope analyses are informative in dividing the groundwater subsystems. Figure 3 shows the isotope results of twenty-two isotopic water samples. The data points fall into three distinctive parts. Part A is closer to the global SMOW line, indicating that groundwater flow is primarily affected by rainfall. The water samples in Part A area were all collected in the T1n Formation. Parts B and C are below the SMOW line, which means groundwater flow is affected by both rainfall and evaporation. The water samples in Parts B and C were collected in the T2d Formation.

Fig. 3
figure 3

Deuterium and oxygen isotope analysis results

Full size image

Based on the dynamic characteristics of groundwater and measurements of groundwater level, a groundwater divide is present near the contact between T1n and T2d Formations. Figure 4 shows a geological cross-section, along which the groundwater level of WGC02 is higher than that on either side, such as ZK03 in the T1n and ZK01 in the T2d. The location of the cross section is shown in Fig. 2. Geological profile measurements on outcrops and drill cores show that the boundary area of T1n Formation is a layer of limestone of approximately 30 m with few karst features. On the other hand, the boundary area of T2d Formation is a medium-thick layer of dolomite with high rate of karst cavities encountered in boreholes, and underground river and springs are present in the southern karst area. As shown in Fig. 5, the groundwater in T2d Formation flows from west to east, whereas the groundwater in T1n Formation flows westward to the groundwater cone of depression around Lion Pyrite Mine. The strata strike, direction of groundwater, and analysis on the karst characteristics suggest an aquiclude between T1n and T2d Formation. The karst water system can be divided into Nanlinghu karst water subsystem and Dongmaanshan karst water subsystem. As shown in Fig. 5, the subsystem divide is not exactly along the formation contact due to irregular distribution of monitoring wells.

Fig. 4
figure 4

Geological cross-section from west to east in the collapse area (water level gauged on November 17, 2014)

Full size image
Fig. 5
figure 5

Contours of groundwater level in the study area (see Table 1 for legend of lithology)

Full size image

4.2 Root cause analysis of sinkhole occurrence

The analysis on influence source of karst collapse is based on the division of the karst water system, and then the human engineering activities are discussed in each karst water subsystem.

4.2.1 Influence of mining in Nanlinghu karst water subsystem

Chao Gold Mine and Lion Pyrite Mine have been active for more than 30 years in the Nanlinghu karst water subsystem. A groundwater cone of depression has been formed with the Lion Pyrite Mine as the center. The influence extent has reached the boundary between T1n and T2d with a radius of approximately 910 m. There are four monitoring wells in this subsystem to monitor the karst groundwater-gas pressures. The monitoring results are shown in Fig. 6. The groundwater dynamics is closely related to rainfall, and the amplitude of groundwater level variation is proportional to the distance from the center of the groundwater cone of depression.

Fig. 6
figure 6

Groundwater-gas pressure variations in monitoring wells ZK07, SJ03 and ZK3

Full size image

Monitoring well ZK07 is 300 m away from the Lion Pyrite Mine. A barometer is used to monitor the air pressure changes associated with the groundwater fluctuation in the sealed borehole. If the rainfall recharge is greater than the mine drainage, the groundwater level rises and compresses the air in the cavities to produce positive pressures. If the rainfall recharge is less than the mine drainage, the groundwater level drops, and negative pressure is generated in the cavities. Most of the air pressure values in the monitoring well suggest negative pressure conditions. The maximum negative pressure is − 2.86 kPa and the maximum positive pressure is 0.6 kPa. The variation range of groundwater level near the mining area is difficult to erode the overlying Quaternary deposits. Monitoring well SJ03 is 580 m away from the Lion Pyrite Mine. Its change trend is consistent with monitoring well ZK07 in response to rainfall. However, the groundwater level at SJ03 rises sharply with 4 m of variation range.

The largest variation range of groundwater level was 17.2 m and occurred in monitoring well ZK03, which is 550 m away from the Chao Gold Mine. The depth to groundwater ranges from 16 to 33 m. The groundwater level change is sensitive to extreme precipitation events. For example, in response to 2-day cumulative rainfalls greater than 50 mm that occurred in November 2014, February 2015, and April 2015, the groundwater level at ZK03 raised nearly 10 m. Under normal rainfall conditions, the influence on the amplitude of groundwater level change is weak.

The above observations show that the shallow karst water has been dewatered by mining in the T1n karst water subsystem. The rainfall caused the groundwater level to rise suddenly in the recharge area where the Quaternary sediments are thin. The sudden groundwater level raise caused differential settlement, which resulted in house damages and ground cracks.

4.2.2 Influence of mining in Dongmaanshan karst water subsystem

Human engineering activities in this subsystem include **nhua Copper Mine in the south and the Quantang water supply wellfield 2 km away from the northern study area. Six karst groundwater-gas pressure monitoring wells were installed in the subsystem.

As shown in Fig. 5, all karst collapses are in the Dongmaanshan karst groundwater subsystem. Occurrence of the collapses in **nmin Village was induced by dewatering of **nhua Copper Mine in 1990 and 2006. Re-mining of **nhua Copper Mine began in October 2011 after a water curtain barrier was constructed in the main groundwater runoff zone in 2007. The Chaoshan Village collapses, which are 1.2 km away from the mine, occurred after the re-mining started. The copper mine was closed in 2012. From 2014 to 2016, 2 or 3 collapses of 1–3 m in diameter occurred every year along the Yanghe River in Chaoshan Village. No additional collapses happened after July 2016.

During the groundwater monitoring period from 2014 to 2016, the groundwater level changes have the following characteristics (Fig. 7):

  1. A)

    Greater fluctuations of the groundwater level were noted in monitoring wells that are closer to the water supply wellfield. The monitoring wells near **nhua Village, SJ02 and SJ04, has moderate groundwater level changes. However, the groundwater level near Chaoshan Village at monitoring wells SJ01, ZK01 and WGC02 showed fluctuations of up to 3 m.

  2. B)

    The karst groundwater-gas pressure changes respond to rainfall at the monitoring wells near the sinkholes. Many sinkholes are distributed in west and south of monitoring point ZK01 near Chaoshan Village. Surface runoff from rainfalls directly flows into these sinkholes, resulting in groundwater fluctuations up to 15 m. The groundwater level fluctuation ranges from 9.34 to 21.7 m in monitoring well WGC02.

Fig. 7
figure 7

Dynamic change of groundwater-gas pressure at monitoring wells on both sides of the groundwater divide

Full size image

5 Summary

The Chaoshan sinkholes that occurred in Tongling City of Anhui Province, China were taken as an example to illustrate approaches to perform analysis of their causes. Based on in-situ investigation, karst groundwater-gas pressure monitoring, and isotopic analysis, the karst groundwater system is better understood, and the previously believed one karst system concept is revised in recognizing an aquitard between T1n and T2d Formations. The division of karst water system lays the foundation for the analysis and sets the limit for the ZOI delineation in one direction. The ZOI in other directions are discussed in each karst water subsystem. This technical approach provides essential data for identifying potential responsible party for the existing sinkholes and for prevention and control of karst collapse and management of mines in the future.

The two karst water subsystems, Nanlinghu subsystem and Dongmaanshan subsystem, control distribution of sinkholes. Most collapses occurred in the Dongmaanshan subsystem, which includes **. The influence distance can reach 1.2 km. Such evaluations suggest that any new underground engineering activities should be avoided in the Dongmaanshan groundwater subsystem to prevent occurrence of new sinkholes unless effective measures are taken to prevent the decline of the groundwater level and effective groundwater monitoring is strengthened.