Comment on Tzampoglou, P.; Loupasakis, C. Hydrogeological Hazards in Open Pit Coal Mines–Investigating Triggering Mechanisms by Validating the European Ground Motion Service Product with Ground Truth Data. Water 2023, 15, 1474

Abstract

The commented paper uses arbitrary and unsubstantiated hypotheses to attribute land subsidence phenomena in the Amyntaion basin to the operations of the Public Power Corporation (PPC) surface coal mine, disregarding, or at least grossly underestimating, the effect of about 600 pumped deep wells for irrigation purposes all over the basin. In addition to the huge difference in the pumped quantities of water from the aquifer, ground water table lowering due to the PPC mine has negligible influence at distances over 500 m from the edge of the mine, while the areas examined in the paper are at distances of several kilometers from the edge of the mine. Furthermore, the authors attribute the landslide that occurred in the mine in 2017 to the steep excavation slopes of the mine and the increased groundwater pore pressure due to reduced peripheral pumping, which is completely inaccurate. To build their case, the authors of the commented paper disregard multiple references in research publications on the above issues, as explained in the main text of this discussion.

1. Introduction

The commented paper by Tzampoglou and Loupasakis [1] purports to provide “an excellent case study to demonstrate the reliability and value of the EGMS product”, which is not the case, in our opinion. The article does not review the reliability of the Earth Observation data from the European Ground Motion Service (EGMS) of the Copernicus European Union’s Earth observation program objectively, but selectively interprets the EGMS data to confirm the main argument of the paper, which is that the mining activities are responsible for the ongoing steady ground subsidence in an area extending at distances more than 2 km from the edge of the mine, and thus affecting the villages in this area. Our claim is that extensive deep pumping for irrigation purposes in the whole valley is responsible for the ground subsidence and damage in the villages.

The “Introduction” section of the commented paper does not adequately place the study in a broad context, and the current state of the research field is not thoroughly reviewed, with key publications not highlighted. Although the article investigates subsidence phenomena in the Amynteon basin and the hydrogeological hazards in Open-Pit Coal Mines, reference [2] concerns environmental contamination, and reference [3] does not refer to mines or open pits but to tailings dam failures. There are also other references that do not correspond to the relevant text. For instance, it is mentioned that “land subsidence slowly affects extensive areas around mines” in references [4,4]. In contrast, reference [4] refers to mine collapse and affected buildings at a 200 m distance within the mining area, and reference [5] refers to a study investigating the possibility of land subsidence occurrence and mentions that, as of the publishing date, no damages had been recorded.

Also, the key publications used to investigate the failure mechanisms triggering natural hazards are those of the authors. Similarly, the authors’ publications are also the key publications cited to relate the mining dewatering wells to the groundwater level drop extending up to 2 km around the mine, and their impact on the villages of Valtonera and Anargiroi.

Overall, the paper is based on unjustified statements and misconceptions, while the arguments are inconsistent throughout the text and have serious shortcomings, as clearly demonstrated in the following paragraphs.

2. The Authors’ Attempt to Attribute Ongoing Steady Subsidence in an Area Extending Several Kilometers Around the Mine to the Mining Activities of PPC, Based on Insufficient Information, Leading to Arbitrary and Unsubstantiated Hypotheses

(a)

The authors claim that farming activities cause minor settlements. However, the article lacks critical information [6,7,8,9] about the locations, density, and depth of the irrigation wells, and ignores the huge difference in pumped quantities for irrigation purposes compared to the pumping quantities for mining operations. According to undisputed data (electricity consumption for irrigation purposes), the approximately 600 irrigation wells, reaching depths up to 120 m all over the basin [10,11], pump about 40–45 million cubic meters of water per year over the last 25 years, while the PPC mine removed only 4–5 million cubic meters of water per year over the last 15 years (until 2020) via a small number of wells at the perimeter of the mine. In the Amynteon basin, the greatest pressure in the aquifer comes from the irrigation wells and is estimated to be in the order of 80–85% [6]. The authors fail to include and evaluate vital data, as key publications are ignored.

Note that the “control points” of vertical displacement measurements provided by Copernicus are located in areas surrounded by a large number of irrigation wells (Figure 1), presenting an average pumping rate of 25 m³/h (open data of irrigation wells’ permits), where there is an absolute absence of mines’ dewatering wells. Based on the pumping rate, the rate of irrigation consumption in Valtonera is estimated to be approximately 0.40 m³/m², which corresponds to an average annual value of 8 million m³ of groundwater abstraction for irrigation use.

Additionally, in Section 2.2 of the commented paper, “Hydrogeologic Setting”, the authors mention that “It is clear that during the last three decades the operation of the mine and, secondarily, the increase of the agricultural activities radically changed the groundwater dynamics of the basin”; however, in Section 3.1, “The Land Subsidence Mechanism”, it is mentioned that “the general hydrogeological conditions remain practically unchanged since 1992. The open pit keeps on operating as the main draining source while no extra irrigation wells have been drilled at the wider area”. This is contradictory as it is clearly stated initially in the manuscript that there has been an increase in agricultural activities in recent decades, and thus, the groundwater dynamics are influenced by this increase.

In their reply to our comments, the authors state that “Aiming to cover the irrigation water needs the farmers at the perimeter of the mine pump water from the draining canals network using electricity from the old power supply network installed initially for the operation of the drills’ pumps. So, the electricity consumption provides false information about the pumping activities from the aquifers”. This is inaccurate, as documented above, and can be disputed by the open data of the irrigation wells’ permits (see also Figure 1).

(b)

In addition to the huge difference in the pumped quantities of water from the aquifer, ground water table lowering due to the PPC mine has negligible influence at distances over 500 m from the edge of the mine, while the areas examined in the paper are located at distances 2–8 km from the edge of the mine. The authors overlook the critical influencing factors of constant irrigation pumping over the years and the limited influence zone of the mine’s dewatering measures around the mine’s perimeter. Neither the mine’s dewatering measures nor the mine itself can cause ground deformations at distances greater than an influence zone of a maximum of 500 m [8,12,13].

(c)

Many piezometers located closer to the mine show significantly smaller groundwater table drawdown than piezometers located farther away in the valley [10]. It is thus evident that the larger values of groundwater table drawdown in the valley are caused by the operation of hundreds of irrigation wells rather than the peripheral wells of the mine [7]. This statement is also supported by the open data on groundwater table measurements in boreholes located north–northwest of the mine (Figure 2), at Valtonera village and the irrigation area between Valtonera and Pedino villages [14]. This confirms the importance of the hundreds of irrigation wells in lowering groundwater tables.

(d)

The authors claim that “currently the open pit along with the surrounding draining wells operate as an oversized (4 km wide) well, continuously draining a big part of the Amyntaio basin”. This is not correct as the dewatering wells were terminated in April 2020, and thus, the mine was decommissioned [8,9]. Also, according to the National Water Monitoring Network [14], it is evident that the groundwater level has been stabilized (as indicated by measurements conducted from 2018 to 2020, Figure 2).

In their reply to our comments, the authors state that “it is clearly presented in maps and cross sections included at a recent publication [13] written by the, at that time, director of the hydrogeological department of PPC, claiming that it was caused by the practically inactive irrigation drills”. This is also not correct, as in [13], it is clearly stated that the groundwater level dropdown in the valley (away from the mine’s perimeter) is caused by the irrigation wells. Also, the same reference mentions that the negative water balance in the Amynteon aquifer was noticed before the beginning of the mine’s pumping wells.

Also, in the authors’ reply, it is stated that “It is commonly accepted that, the open pit coal mine resembles hydraulically to a huge well with a diameter of 3 to 4 km and appears to be the main cause of the groundwater table level drop not only in its immediate vicinity but also in a wider area, triggering land subsidence phenomena”. This statement is supported only by a PhD thesis and a Technical Report, which, although mentioning this phenomenon, fail to justify it.

(e)

Moreover, the authors mention that “the data from May 1992 relate to the period before the operation of the open pit coal mine [15]”; however, this is not correct as the mine was operating during that time. Based on the authors’ reference [15], the groundwater drawdown during 1986 and 1992 (6-year time period) was 4–5 m and locally 6 m due to the overexploitation of the Amynteon aquifer by farmers for irrigation purposes. Note that dewatering measures were adopted to protect the surface mine during its operation after 1993. Therefore, based on the authors’ reference [15], agricultural activities had a great influence on groundwater, which was evident in the area before the beginning of the mine’s dewatering measures around the mine’s perimeter. The authors fail to include and evaluate vital data from the publication [15] cited in their article, while other key publications, such as [6,7,8,9,14], are ignored.

(f)

Figure 3 in the commented paper presents the groundwater level dropdown between certain periods, for a 23- and 24-year time period. This dropdown should be an outcome of the data processing and evaluation via piezometric maps (authors’ own measurements in a three-year (3) period in 37 points). It is easy to verify that the dropdown contour lines are erroneous and do not correspond to the piezometric maps. For instance, according to the authors’ piezometry, there is practically no groundwater level dropdown in the Valtonera and Anargiroi villages between the years 2014 and 2016. While comparing these measurements to those of 1992, a level dropdown of up to 5 m can be estimated. However, the calculated dropdown level presented by the authors is 10–20 m. It is evident that the outcome of the groundwater level dropdown is not verified from the “Ground truth datasets” used in the article. The same applies to Figure 4 in the commented paper for the incorrect groundwater level dropdown maps. Additionally, the statement in Section 3.1 of the commented paper, “…at Valtonera village. The observed water level drop in this area of around 10 to 20 m…” is inaccurate and inconsistent with the authors’ own water level measurements.

Following the argument above, it is evident that there is practically no groundwater level dropdown in the Valtonera and Anargiroi villages. Therefore, it is derived that there is no correlation between the mentioned groundwater level dropdown and the vertical deformations provided by the Copernicus Land Monitoring Service. Thus, all the related statements in the article are incorrect.

(g)

The borehole network where the groundwater measurements occur is important and should be considered in the evaluation process. For the year 1992, no network is presented. According to the authors’ reference [15], in 1992, the network where the measurements occurred was scarce, while there was no network in a ~3 km radius around Valtonera village, nor between Valtonera and Anargiroi villages.

(h)

The authors present the groundwater table contour lines in different time periods. However, the extent of the mine presented in the satellite images does not match the real extent during the respective time period. This is crucial for conducting proper piezometric maps.

3. The Authors Attribute the Landslide Which Occurred in the Mine in 2017 to the Steep Excavation Slopes of the Mine and the Increased Groundwater Pore Pressure Due to Reduced Peripheral Pumping, Which Is Completely Inaccurate

(a)

The authors relate the land subsidence phenomenon and the landslide that occurred in 2017 to the dewatering of the mine. This is inaccurate and inconsistent with the authors’ own water level measurements, which clearly show that there is practically no groundwater level dropdown in the Valtonera and Anargiroi villages, as documented above. Therefore, the correlation with the vertical deformations provided by the Copernicus Land Monitoring Service does not apply, and thus, all related statements in Tzampoglou and Loupasakis (2023) are not supported by the data. Note that the locations where the vertical deformations are presented (see authors’ InSAR data in relation to Figure 1 above) are contiguous with irrigation wells and at a significant distance from the mine’s dewatering wells. Also note that the dewatering measures stopped in April 2020, while the irrigation wells have not yet stopped.

(b)

The authors mention that “the pumping activities directly adjacent to Anargiroi village, according to PPC [13], were essentially stopped in 2016, aiming to eliminate the deformation caused by overpumping at the nearby village”. However, according to the same reference, this is not accurate, as no such information is presented. Thus, the writers’ outcome, where the mine’s pumping activities are related to the landslide, is not supported.

(c)

The authors’ statement “The drop in the groundwater level, combined with the area’s general geotectonic setting, triggered extensive land subsidence in an area extending up to 2 km around the mine, causing damage to villages and infrastructure since 2002” is not correct, because the effect of any large well is related to the magnitude of the drawdown (which in our case is 80–100 m close to the mine’s perimeter) and is independent of the well diameter. Concerning the influence zone, this is also evident in hydrogeological cross sections that include water levels in boreholes [11,13]. Please note that any dropdown in groundwater level can only enhance the stability of slopes and cannot trigger the landslide phenomenon.

(d)

Based on the actual geometric data of the mining operations, all statements in the paper regarding the pit slopes and the geometry of the mine benches are inaccurate. Specifically, based on all available data and satellite images, there is no evidence to suggest that steep slopes were applied, or the excavation of the lower benches of the coal seams was prioritized, or that the excavation of the middle benches was systematically neglected. This is also verified in the Audit Findings report of the committee experts [16].

Moreover, no geometrical data are presented to support this statement. In their reply, the authors do not answer our comments and do not provide any data but only mention that they applied the Pythagorean theorem to calculate the steepness, knowing “the horizontal distance between the base of the slope and the crown of the benches as well as the high of the benches and the total depth of the mine”. This statement is completely irrelevant to our arguments, and, of course, it is obvious that no application of the Pythagorean theorem is related to the overall slope calculation.

(e)

The mine’s pumping was limited at the upper aquifer (i.e., up to a depth of 100 m) that lies above the impermeable lignite strata [9], while the total depth of the mine pit exceeds 200 m, and the sliding occurred along a sub-horizontal lignite-to-clay/marl interface [17] at depth about 200 m. Thus, the groundwater pore pressure at this depth is not influenced by the mine’s pumping activities in the upper aquifer. Consequently, the 2017 landslide is not related to the pumping operations at the perimeter of the mine.

In conclusion, the commented paper is based on inaccurate information and fails to present and evaluate the controversial and diverging aspects, leading to a misinterpretation of the data and, eventually, a misjudgment about the real causes of the described events. We point out that, due to the great importance and potential legal implications of these events for the operator of the mine (PPC), groundwater table levels around the mine have been monitored and studied for many years by geologists and engineers of PPC and by independent experts [10,11]; they all reach conclusions opposing the conclusions presented in the commented paper.