Introduction

Mining activities cause environmental damage and changes to the earth’s surface and underground. Abandoned mines contaminate drainage water, soil, air and ecologic variety (Bell et al. 2001; Bhattacharya et al. 2006; Ribeiro et al. 2010; Si et al. 2010). Pollution of drainage water is continuously generated and dispersed into the area downstream of the mine (Oh and Lee 2011). Acid mine drainage (AMD) is generally defined as the product formed by the atmospheric oxidation (i.e., by water, oxygen and carbon dioxide) of the relatively common iron-sulfur minerals such as pyrite (FeS2) and pyrrhotite (FeS) in the catalyzing presence of Acidithiobacillus ferrooxidans, an acidophilic bacterium, and any other products generated as a consequence of these oxidation reactions (USEPA 1994a, b; Mills 1995; Akcil and Koldas 2006; Ozcelik 2007; Sanliyuksel Yucel and Baba 2013a). Nearly all metallic and most coal mines contain pyrite or other sulfides that oxidize when exposed to the surface environment, generating sulfuric acid (Nordstrom 1982). Acid mine lakes (AML) are a world-wide problem because of the acid mine drainage system. It is vitally important to evaluate surface water quality and environmental damage through acid mine drainage using remote sensing methods (Paktunc 1999; Sabins 1999; Schroeter and Glasser 2011).

A multidisciplinary approach to reduce ecological and economic problems is required for all mining consequences within active and abandoned mines (Schroeter and Glasser 2011). A systematic and multidisciplinary approach to mapping, monitoring and controlling the impact caused by mining activities is necessary to understand the character of hazardous events in an area (Woldai 2001). Information about the temporal and spatial data of different parameters (geological, hydrochemical etc.) is essential in each phase of mining for the success of both the productivity and for reclamation of the land.

Satellite data have been successfully used to monitor the effects of AMD in mining areas such as coal mining and in areas with abundant mining waste (Spahn 1983; Paterson 1997) and to locate abandoned mine sites in northern Ontario in Canada with contaminant AMD problems (Robitaille et al. 1991; Paterson 1997). Peichang et al. (2004) used GIS and remote sensing to identify acid mine lakes in an Ohio coal mining area. Latifovic et al. (2005, 2006) investigated changes in forest areas due to mining activities near Fort McMurray in Canada. Yenilmez et al. (2011) investigated pollution levels at an abandoned coal mine site in Cankiri (north-central Turkey) which were evaluated with respect to topography and surface runoff pathways derived using GIS tools. Oh and Lee (2011) used GIS to generate ground subsidence hazard maps of abandoned coal mines in Samcheok, Korea. Schroeter and Glasser (2011) studied physicochemical properties of post mining lakes demonstrating the use of multitemporal analysis Landsat TM5/ETM + 7 satellite data as basis for a monitoring system for geochemistry of mining lakes.

Study area

The study area is geographically situated approximately 17 km southwest of the county of Can, which is located in Canakkale province of northwestern Turkey (Fig. 1). This region is associated with intense mining activities because of its geological and tectonic properties. Therefore, open/underground mining operations for coal production and industrial raw materials are present. Also, this region and its vicinity have important water resources.

Fig. 1
figure 1

Location map of study area

Full size image

Canakkale province is situated in a climatic transition between Mediterranean and Black Sea climate zones. The summers are hot and dry, and the winters are relatively wetter and colder. The long term (1970–2011) mean annual temperature was 15.2 °C with a minimum of −11.2 °C (February, 2004) and a maximum of 39 °C (July, 2007) based on data collected at Canakkale Meteorological Station located about 40 km northwest of the study area. Precipitation in the region is mostly in the form of rain and the highest precipitation is observed in winter. The mean annual precipitation rate was 594.6 mm between the years 1970 and 2011.

The majority of creeks originating from the study area are drained into Kocacay Stream which is the second largest stream in Canakkale province and it flows northeast through the counties of Can and Biga and eventually drains into the Sea of Marmara approximately 45–50 km north of Can. The hydrology of the surface drainage network in Kocacay demonstrates a variable flow regime, which mainly depends on the seasonal precipitation pattern. Based on long-term monthly averages, the mean annual flow rate of Kocacay Stream is found to be 1.28 m3/s at gaging station (Baba et al. 2009). The highest flows are observed during spring months where snowmelt is augmented by spring rains.

Many mining companies have been operating coal deposits for power generation and district heating in the Can region since 1980. Generally, small and medium-scale mining companies operate for short periods and abandon the operational land without providing any rehabilitation. As a result of human intervention in the natural structure and topography of the earth surface, large holes and deterioration form in these areas. Artificial lakes occur due to both surface discharge and underground leakage into abandoned open pit mines and they are characterized by specific limnological development changes. High sulfur content lignite exists in these areas. Furthermore, these lakes gain acidic character due to acid generation from pyrite oxidation. AMLs are more dangerous than AMD as a result of the massive volume of water that potentially poses a threat to the environment. This study focuses on nine acid mine lakes (AML 1, 2, 3, 4, 5, 6, 7, 8 and 9) (Table 1; Figs. 2, 3) which are topographically situated within range of 128–245 m elevation above mean sea level. The acid mine lakes drainage adits from both sides discharge acid mine drainage into the Kocacay Stream. The weathering product of sulfide oxidation is the formation of iron hydroxide, a red/orange-colored precipitate found in dozens of miles of streams affected by AMD (Sanliyuksel Yucel et al. 2013).

Table 1 Coordinates and elevations of acid mine lakes (UTM-WGS84)
Full size table
Fig. 2
figure 2

Drainage system of study area

Full size image
Fig. 3
figure 3

Snapshots from AMLs a AML 1, b AML 2, c AML 3, d AML 4, e AML 5, f AML 6, g AML 7, h AML 8, i AML 9

Full size image

Geological and hydrogeological setting

The study area mainly consists of acidic and agglomeratic tuff, andesite, trachyandesite and andesitic tuff which are called the Can volcanics (Ercan et al. 1995). The volcanic successions of Can volcanic rocks are enriched with high sulfur content, due to alteration processes and structural controls. Widespread and intense zones of silicified, argillic and propylitic alteration can be observed in the Can volcanics (Sanliyuksel Yucel et al. 2012). In the Early–Middle Miocene; terrestrial units were deposited with the calc-alkaline volcanic activity (Siyako et al. 1989). These terrestrial units consist of bituminous shale, siltstone, claystone, sandstone, tuff, and coals, named the Can formation, which is up to 250 m thick and unconformably overlies the Miocene-aged andesitic volcanic rocks (Siyako et al. 1989) (Fig. 4). The Can formation is rich in lignite (also known as brown coal) and 17 m thick (Bozcu et al. 2008). The sulfur content of lignite varies between 0.21 and 14.36 wt % (Gurdal 2011).

Fig. 4
figure 4

Geological map of study area (Sanliyuksel Yucel and Baba 2013a) and location of AMLs

Full size image

The hydrogeology of the study area is primarily dominated by volcanic units. Silicified zones of these volcanics are typically situated as a cap over the volcanic rocks. The silicified volcanics in the study demonstrate fractured aquifer characteristics. Dense argillic alteration, demonstrating impervious properties due to its high clay content, can also be seen below the silicified alteration zone. These clayish zones act as a barrier beneath the fractured aquifer. In general, most of the springs in the study area surface out from the boundary between fractured aquifer (silicified zone) and impervious boundary (argillic zone). The flow rates of these springs range between 0.01 and 3 L/s (Baba and Gunduz 2010). Also, many drills were done in volcanic units in the study area. Flow rates of these drills ranged from 0.2 to 0.7 L/s in volcanics. Hydraulic conductivity values of volcanics units ranged from 1 × 10−7 to 1 × 10−8 m/sn (SRK 2012). According to these data, volcanics can be described as less-permeable and impermeable aquifers. Generally clay alteration of volcanics has a high storage capability, but this part of volcanic unit is not a good aquifer. In addition, Can formation has a less permeable unit in study area. Flow rates of springs in Can formation range from 0.01 to 0.5 L/s in study area. The depths of drillings in Can formation are range from 100 to 150 m depth. The flow rates of these drills are range between 1 and 1.5 L/s (Baba et al. 2009). Groundwater flow direction is SW to NE and water table is generally followed by topography. Groundwater level changes average of 2 m between in dry and wet period.

Methods

Satellite data

Satellite image data are relatively inexpensive and have large coverage and an increased volume of information. Moreover, they are multi-temporal and available on a large scale (Anderson et al. 2004; Ozesmi and Bauer 2002; Schroeter and Glasser 2011). This study is based on time series data sets from Landsat, Landsat 5, Landsat 7, Quickbird and Worldview (USGS 2006a, b, c) satellites. This investigation uses cloud-free images between May and November of the years 1977, 1985, 1987 and 1999 from Landsat data and between March and July of the years 2003, 2008 and 2011 from Quickbird and Worldview. The Landsat TM (Thematic Mapper) and Landsat ETM+ (Enhanced Thematic Mapper Plus) data have a geometric resolution of 30 per 30 m and per pixel. The Landsat MSS (Multispectral Scanner) data have a geometric resolution of 60 per 60 m and per pixel. Quickbird data have a geometric resolution of 60 per 60 cm and per pixel. Worldview data have a geometric resolution of 50 per 50 cm and per pixel. Landsat MSS, Landsat TM and Landsat ETM +, Quickbird and Worldview ground resolution of multispectral bands is suitable for scaled images up to a maximum of about 1:50.000 for 1:25.000 and 1:5000. In 2009, the Landsat archive images became freely available to research users. Landsat images were downloaded from USGS website; Quickbird and Worldview images were purchased from NIK System Inc. Co. Remote Sensing Department.

Change detection and visualization

In this study, the satellite images were adjusted geometrically with the software ENVI 5.0 in order to transform them into precisely scaled images for GIS implementation. Then they were transformed to UTM coordinates (WGS84-UTM Zone 35 N). By this process, all data and layers were overlaid precisely.

The satellite images were imported into ArcGIS 9.3 software and acid mine lake borders were digitized manually. This process was generated for Quickbird and Worldview images directly, which have high resolution data. Landsat satellite images were of lower resolution; therefore the “stretch type” adjustment of these images was changed to “standard deviations” and gamma “stretch values” were entered as “−1”. This process generated sharpened images. The acid mine lake borders were easily and accurately determined. After this adjustment process acid mine lakes on Landsat satellite images were digitized (Fig. 5). Border change detection of acid mine lakes was provided by the digitizing process.

Fig. 5
figure 5

Digitizing of Landsat satellite image (1999, AML 2)

Full size image

Using ArcGIS 9.3 software, satellite images were analyzed in time series, borders of acid mine lakes were digitized and converted into vector data format. The areal and perimeter changes were computed and presented via tables and graphics. Thereby, thematic maps of the acid mine lakes were created and visualized using cartographic rules.

Water chemistry

Nine acid mine lakes, where water samples were taken, were monitored in August 2012 in the study area. Water samples were transferred to polyethylene bottles of 500 and 50 mL capacity and kept in portable coolers. The 500 mL samples were used in anion and cation analysis and the 50 mL samples were used for some elements. Physical and chemical analysis of water included primary field parameters (pH, redox potential, temperature, salinity and electrical conductivity), major anions and cations (sodium, potassium, calcium, magnesium, chloride and sulfate) and elements (aluminum, iron, manganese and sulfur). All water samples were filtered through 0.45 µm filter paper and stored in polyethylene bottles at 4 °C. The physical field parameters were instantly measured in situ at water points with a multi-parameter probe WTW model 340i. The major cations and elements were analyzed with inductively coupled plasma mass spectrometry (ICP-MS), SO4 and Cl was analyzed by ion chromatography (IC) at Acme Analytical Laboratories in Canada.

Morphological properties of AML 2 were studied in July 2010. The bathymetry of the lake was measured using a 30 m marine rope marked at 1 m intervals that was tied to 1 kg weight. Lake depth was measured at 15 m intervals along a 235 m long transect (A–A′) across the longest part of the lake for a total of 15 depth measurements.

Results and discussion

Areal and perimeter changes of acid mine lakes

Acid mine lakes 2 and 3 are the oldest lakes, followed by AML 1, AML 6, AML 8, AML 9, AML 7 and finally AML 5 and 4 (Fig. 6). AML 5 is situated on the old stream bed, and formed as a result of prevention of water flow by huge amounts of mine waste (Yucel and Sanliyuksel Yucel 2012). AML 5 had a maximum perimeter value of 1,671 m, in 2008 (Table 2). AML 2 had a maximum surface area in 2003, of 3.34 ha (Table 3). Areal change of acid mine lakes effects from environmental conditions such as annual precipitation, evaporation and drainage patterns (Yucel et al. 2013). Total perimeter and area of all acid mine lakes reached maximum levels of 5,699 m and 12.42 ha, respectively, in 2011 (Figs. 7, 8). Around the AMLs, the effluent drains into nearby rivers and into water sources from the mine tailings which had a total area of 166.26 ha, in the year 2011 (Fig. 9). Total area of acid mine lakes and mine tailings were 178 ha in 2011; they covered 20 per cent of the study area.

Fig. 6
figure 6figure 6

Change detection of AMLs in time period a 1977, b 1985, c 1987, d 1999, e 2003, f 2008, g 2011

Full size image
Table 2 Perimeter values of acid mine lakes
Full size table
Table 3 Areal change of acid mine lakes
Full size table
Fig. 7
figure 7

Time versus perimeter change of AMLs

Full size image
Fig. 8
figure 8

Time versus areal change of acid mine lakes

Full size image
Fig. 9
figure 9

Area of mine tailings around acid mine lakes

Full size image

The deepest point in AML 2 was 13.45 m deep and the average depth of the lake was 7.41 m (Fig. 10). The perimeter and surface area of the lake were 616 m and 15.55 m2, respectively in 2010 (Sanliyuksel Yucel and Baba 2013a).

Fig. 10
figure 10

a Bathymetry of AML 2 (July, 2010), b The vertical profile on A–Al cross-section (Sanliyuksel Yucel and Baba 2013a)

Full size image

Open pit coal mine tailings continue to generate acidity and acidify surface water. Lakes and streams affected by coal mines have limited aquatic life due to low pH and high metal concentrations in water. AMLs are very close to Keciagili Village and all lakes around the village are affected by AMD. AML 2 was discharged into Kocacay Stream in 2008 because of mining activities and resulted in serious environmental problems. It is also worthwhile to mention that in Can basin, fish deaths due to the seepage and discharge of the acid mine lakes into surface waters have been observed (Sanliyuksel and Baba 2011). It is clear that acid mine lakes cause environmental risks due to their presence and areal increments.

Hydrogeochemistry of acid mine lakes

Acidification due to pyrite oxidation is a major problem for water quality in acid mine lakes. The field parameter results of AML measurements are presented in Table 4. These results indicate that the pH of AMLs range from 2.53 to 3.05 (Fig. 11a). The pH of a solution is an important measure for evaluating aquatic toxicity (Cravotta 2008). The electrical conductivity (EC) values of AMLs ranged between 2,890 and 6,750 µs/cm (Fig. 11b). The electrical conductivity indicates the relatively high concentrations of dissolved solids in AMLs, mainly sulfate in the case of acid lakes pH < 3, and iron. The Eh values range between 208 and 240 mV. High acidity and high redox potential values are mainly due to the presence of significant quantities of dissolved Fe3+ which acts as a more effective oxidizing agent than O2, in the most acidic environments (Monterroso and Macias 1998). AMLs have different colors, AML 4 is blue, AML 7 is green and others are orange-red-brown in color. Precipitates associated with AMD can be characterized by particular colors depending on their composition (Bigham et al. 1994; Nordstrom et al. 2000). Fe3+ precipitates give the iron hydroxide the colors of yellow, red, orange and brown during the formation reactions of AMD. Jarosite was identified by XRD results in bottom sediment in AMLs (Sanliyuksel Yucel and Baba 2013b). Jarosite is a straw yellow mineral that is common in the weathered zone of sulfide ore deposits and in acid sulfate soils. These minerals also gives color of AMLs.

Table 4 Physical parameters, major ion and some element values of acid mine lakes (mg/L) (August, 2012)
Full size table
Fig. 11
figure 11

a pH, b electrical conductivity values of acid mine lakes (August, 2012)

Full size image

Hydrogeochemical processes play a major role in determining the quality of water draining from the waste rock pile. The wide variety of minerals in waste and host rocks influences the resultant drainage composition. Sulfate is clearly the dominating anion in all AMLs. It results mainly from pyrite oxidation. Sulfate levels were observed to range between 1,055 and 4,600 mg/L with a mean value of 2,677.22 mg/L, which exceeds the water quality standard of 250 mg/L by a factor of 10 (Environmental Protection Agency 2003; WHO 2004; ITASHY 2005) (Table 4). Magnesium and calcium values ranged between 64.9 and 353 mg/L, and between 179 and 525.3 mg/L, respectively.

The geochemical behaviors of Fe and Al mainly control the mobility of metals though adsorption and coprecipitation in the acidic sulfate water (Bigham and Nordstrom 2000; Lee et al. 2002; Lee and Chon 2006; Equeenuddin et al. 2010). Various kinds of ochers, mainly iron oxyhydroxide sulfates, precipitate along the AMD impacted river (Bigham and Nordstrom 2000; Equeenuddin et al. 2010). Iron and aluminum are the most predominant metal species in the AMLs in the study area. It was observed that iron concentrations ranged between 7.19 and 338.17 mg/L (Table 4). The concentration of aluminum was observed to range between 17.18 and 357.47 mg/L (Fig. 12). The high aluminum concentrations are also a clue for the high clay content of the regional rocks. High concentration of Al and Fe result in important increases in acidity and provokes the buffering of the AMD systems. These lakes also have extremely high concentrations of manganese that were observed to range between 20.17 and 164.97 mg/L. High sulfur concentrations were found in all lakes, the maximum level was as high as 1,892 mg/L. It was expected that the sulfur concentrations would be high because it is well known that Can coals contain high sulfur contents (Sanliyuksel Yucel and Baba 2013a).

Fig. 12
figure 12

Iron and aluminum concentrations of AMLs (August, 2012)

Full size image

Conclusions

Surface mining for lignite resulted in new artificial lakes. Acid mine lakes are the most important part of the post mining landscape and have ecological as well as socio-economic functions. The major concern with respect to water quality is whether acidification resulting from mining has been managed successfully. With an increasing demand for efficient environmental monitoring of mining and acid mine lakes, quantitative water monitoring and integration of these data to satellite images and GIS tolls is systematic way for determining border changes of AMLs. This is the first study about change detection and visualization of acid mine lakes using time series satellite image data in geographic information systems in Canakkale province, Turkey. Satellite imagery is providing view of different years’ information on a regional scale on land and changes on water bodies.

The environmental problems are the result of 30 years of mining activity and other pollution sources. The AML 2 and 3 in the Can County will be filled in about 26 years. AMLs cause potential risk of infiltrating the drinking water resources of the surrounding villages. The results show that the number of acid mine lakes increased and these caused environmental risks due to their hydrochemical properties and areal increments. Therefore, a long-term strategy for common water resources management of groundwater and surface water regarding water quantity and quality is necessary. Future work should focus on the determination of AMD treatment alternatives.