Multi-Risk Assessment in Post-Mining Lignite Areas ^†

Abstract

As Europe addresses decarbonization, huge coal mining areas shall be reclaimed and appropriately valorized, creating the post-mining areas. This work assesses their resilience and safety by methodologically addressing multi-hazards threatening closed and abandoned mines. The current work modifies an indicator-based semi-quantitative multi-hazard method towards post-mining risk assessment, described as the (1) identification of the hazards, (2) assessment of the interactions, and (3) quantification of the level of interactions. The methodology can appropriately quantify the multi-hazard severity for various scenarios. However, it has limitations, not considering appropriately the hazards’ sequence and missing an objective criterion for scenario comparison.

1. Introduction

To reduce CO₂ emissions and increase the use of renewable energy sources, based on the European Green Deal, Europe is moving towards independence from coal (and lignite) as the main fuel for energy production. As a result, huge areas committed to coal mining will be reclaimed to benefit local communities. The appropriate valorization of these post-mining areas is a pillar for the just transition to climate neutrality. At the European (e.g., Germany, Czech Republic, Poland) and global levels (e.g., Canada, Australia), plans are already being developed and implemented on potential decarbonization practices. One obstacle to these plans for sustainable reclamation is convincingly quantifying the safety and resilience of post-mining areas, i.e., identifying the hazards and assessing the risks of closed mines and post-mining activities [1]. Mining hazards, as a basis for risk assessment, are challenging for stakeholders because they can occur after a halt in mining operations and even after several years.

Varnes [2] (p. 10) defined hazard as the probability of a potentially catastrophic event occurring in a given area and over a given period and risk as the product of the hazard with its consequences. Areas with complexity in infrastructure and activities, such as mines, tend to show vulnerability to multiple hazards. Recently, a tendency has been developed to focus on multi-hazard analysis and multi-risk assessment, aiming to examine the interactions among different types of hazards (mainly between natural and technological [3,4]). This undertaking necessitates a thorough comprehension of the interdependencies among hazards. A key element of the present work is an advanced multi-hazard analysis with hazards that can occur either one after the other or simultaneously with and without a dependent relationship. Multi-hazard analysis cannot be treated simply as the sum of the individual hazards but needs to investigate the correlation between their occurrences [3,5]. The difference between single and multi-hazard is important, as the consequences of interrelated hazards are often more significant than those deriving from individual hazard summation [5].

This paper aims to identify and analyze the interaction between hazards and enhance post-mining areas’ methodological knowledge. A semi-quantitative multi-hazard method developed by Liu et al. [4] is adapted in the current work to be adjusted for post-mining areas. This method analyzes each hazard individually, builds triggering interactions between hazards using adjusted factors, and aims to quantify the multi-hazard intensity of different multi-hazard scenarios.

2. Methodology

Three categories of methodologies are encountered broadly to explore the interaction of hazards and calculate the multi-hazard intensity: qualitative, semi-quantitative, and quantitative. Qualitative methods are primarily based on engineering judgment and historical data; they identify the most catastrophic hazard scenarios through interaction matrices and diagrams and assess the multiple hazards of the area [3]. Semi-quantitative methods either use correlation coefficients or calibrate hazard occurrence characteristics to capture the interaction of hazards. Quantitative methods are based on numerical approaches adapted to investigate the interaction of hazards. Research using quantitative methodologies is limited as they present greater complexity and require reliable quantitative data as input. Multi-hazard assessment at local scales remains a significant challenge due to the lack of data, and interactions between different types of hazards.

The methodology of the current work is based on a semi-quantitative method, which can be described in three main steps. The initial step involves the identification of the main hazards in the study area, as typically conducted in risk assessment. In the case of coal mines after closure and during valorization, the hazard categories are related to the topography and location, also considering the former mining operations, outlining three major categories: mining, natural, and technological hazards. Mining hazards refer to catastrophic events caused due to mining operations, such as landslides, settlements, and flooding. Natural hazards are related to physical phenomena such as earthquakes, rainfall, fire, and storms. Technological hazards originate from technological and industrial conditions involved in mining operations, such as explosions, fires, and toxic chemical exposure [6,7].

Individual hazards are characterized based on their predisposing factors and associated triggering factors. For example, a landslide is influenced by its predisposing factors, such as the topography of the ground and the soil conditions, and can be triggered by climate conditions, e.g., heavy rainfall [8]. Moreover, in a multi-hazard scenario, a landslide can be triggered by other hazards, such as flooding and earthquakes. The second step deals with interactions among hazards and determining multi-hazard scenarios. The three main mechanisms of hazard interaction are (a) trigger (when a hazard triggers another, generating what is called a “cascade (or domino) effect”), (b) influence (when a hazard influences another, without acting as a trigger), and (c) coincidence (when hazards occur simultaneously and independently). A visual representation of potential interactions is a matrix interaction tool or a diagram tool. The final (third) step focuses on identifying the level of interactions between hazards. This assessment is based on the intensity of individual hazards and the level of hazard interaction. Different hazards—from the same category or not—can interact and cause higher consequences.

The multi-hazard analysis demonstrates that interacting hazards can be more destructive. Liu’s method successfully achieves this by employing adjusted principles applied to secondary hazards, effectively increasing their initial intensity. Conversely, triggered secondary hazards undergo an increase in their initial intensity, contingent upon the initial intensity of the primary hazard. Additionally, the proportional growth factors are determined based on whether the probability of a secondary hazard being triggered is low or high. Consequently, secondary hazards that primarily have not been triggered do not have a rise in their initial intensity. Liu’s research on the interaction between natural and technological hazards led to the development of the adjusted principles of hazard intensities (see the following section). The same adjusted principles were applied in the current study. However, it is important to consider the need for future adjustments of these factors to ensure a more precise implementation of this methodology, encompassing natural, mining, and technological hazards.

Previous studies on risk and hazard assessment of coal mines have primarily focused on thoroughly examining individual hazard phenomena. However, abandoned mining areas are typically impacted by not just one hazard but multiple hazards that can occur simultaneously or consecutively. It is challenging to consider multiple hazards and not only assess single ones because the overall risk is underestimated in the latter case.

3. Study Area of the Megalopolis Lignite Mining Complex

The case study is related to the Megalopolis lignite mining complex in the Peloponnese region in southern Greece, which is about to seize its operations. Multi-hazard assessment is challenging in local scale studies due to the lack of data about hazard interactions and their intensities.

According to the main categories of hazards in post-mining areas,

Table 1. Initial and adjusted intensities of hazards.

Hazards	Initial Intensity	Adjusted Intensity
Earthquake (EQ)	3	3
Rainfall (RF)	2	2
Landslide (LA)	4	8.7
Flooding (FL)	3	7.6
Release of gas (RG)	2	5.2

presents the identified natural, mining, and technological hazards that potentially can happen to the study area considering its location, topography, and history of hazard occurrence. For the mining hazards, the mine excavation is composed of several benches [8], and a landslide can occur either on a local scale (bench scale) or on a generalized scale, including larger sliding volumes. Moreover, there are several mechanisms for flooding, such as rising groundwater after mining closure, extreme rainfall, or a landslide that can trigger flooding by sliding volumes. For the natural hazards, seismicity and rainfall are associated with potential earthquakes and moderate height precipitation of the studying area. Finally, regarding technological hazards, a gas release can happen when storage infrastructures or pipelines fail, leading to several types of pollution. In the current work, experts analyze, compare, and rank the available records and historical data according to the triggering and predisposing factors of the Megalopolis lignite mine for the identified hazards to determine initial hazard intensities and their interactions.

The assessment of the level of interaction relies on recorded and historical catastrophic events and their intensities. These data are adapted to simulate the possible impact scope and the corresponding intensity of each hazard. In order to implement the methodology, each hazard is evaluated based on its intensity and the predisposition of the study area. These classifications help prioritize the assessment of hazards by considering the nature of the phenomena and the likelihood of their appearance. There are six intensity classes ranging from 0 to 5; the higher the number, the greater the hazard intensity, and zero means that the hazard has no impact. Precipitation data reserved from the Hellenic National Meteorological Service. The annual precipitation of the Megalopolis area has a moderate value with a low to moderate duration regarding the recordings; thus, the initial intensity was selected as 2. The earthquake’s initial intensity is related to the classification of Greek seismic hazard. According to the latest Greek seismic code, the country was divided into three seismic hazard zones. Regarding the horizontal peak ground acceleration, the study area falls into zone II; therefore, the initial intensity was selected as 3.

Three landslides have occurred in the Megalopolis lignite mine from 1996 to 2001 [8]. Thus, the initial intensity was selected as 4, based on this information from historical data and the related bibliography. In the case of non-repetitive hazards, such as flooding or gas release, the occurrence probability is replaced by the site’s predisposition to experience such phenomena [7] and their initial intensities were selected as 3 and 2, respectively. Table 1 illustrates the initial intensity for each hazard.

The triggering relationships among different hazards are summarized by a hazard matrix, as shown in Figure 1, where hazards are categorized into natural (green), mining (orange), and technological (blue). The interactions among the identified hazards are categorized into two levels: dark blue indicates that the primary hazard (first column) has a high probability of triggering the secondary hazard (first row); light blue indicates that the primary hazard has a low probability of triggering a secondary hazard. The diagonal cells (grey) are intentionally unfilled and represent the interaction of each hazard with itself, which is meaningful because the current study is focused on interactions between different hazards. Natural hazards trigger the mining and technological hazards at a higher level; therefore, they will be the primary hazards in the multi-hazard scenarios.

According to the triggered relationship between hazards (see Figure 1), the initial intensity degree of each hazard is adjusted before the multi-hazard intensity calculation. The increased hazard intensities with the adjusted coefficients express a higher impact when hazards interact. The adjustment principles, based on [4] for natural and technological hazards, are shown in

Table 2. Adjustment principle for hazard intensity [4].

Primary Hazard Situation	Intensity Degree of Primary Hazard	Adjustment Principle
For hazard that needs adjustment, there is no primary hazard that triggers its occurrence	-	Adjusted intensity degree = Initial intensity degree × 1
For hazard that needs adjustment, there is a primary hazard that has a high probability of triggering its occurrence	5	Adjusted intensity degree = Initial intensity degree × 1.5
	3–4	Adjusted intensity degree = Initial intensity degree × 1.4
	1–2	Adjusted intensity degree = Initial intensity degree × 1.3
For hazard that needs adjustment, there is a primary hazard that has a low probability of triggering its occurrence	5	Adjusted intensity degree = Initial intensity degree × 1.3
	3–4	Adjusted intensity degree = Initial intensity degree × 1.2
	1–2	Adjusted intensity degree = Initial intensity degree × 1.1

. In future work, it is necessary to adjust them in post-mining hazards for a more accurate implementation of this methodology. Earthquakes and rainfall maintain their intensity because other hazards cannot trigger them. The landslide intensity is adjusted from 4 to 8.7 (4 × 1.4 × 1.3 × 1.2), and the adjusted intensities for the residual hazards are 7.6 for flooding and 5.2 for gas release. The final multi-hazard intensity (MH) is calculated by summing up every hazard’s adjusted intensity.

Table 3. Tentative multi-hazard scenarios in post-mining areas.

No.	Scenario 1	MH	Min	Max	Norm. MH
1	RF → LA	10.7	2.2	15	0.66
2	EQ → RG	8.2			0.47
3	RF → LA → RG	15.9	3.3	22.5	0.66
4	EQ → LA → FL	19.3			0.83
5	RF → FL → LA → RG	23.5	4.4	30	0.75
6	EQ → LA → FL → RG	24.5			0.79

¹ EQ = Earthquake, RF = Rainfall, LA = Landslide, FL = Flooding, RG = Release of Gas.

represents some tentative multi-hazard scenarios in post-mining areas. These scenarios vary in the number of hazards from 2 to 4. The aim is to compare and determine which of them is the most catastrophic. Scenarios with more hazards illustrate higher multi-hazard intensity (MH). In the current work, the comparison is applied between scenarios with the same number of hazards. Thus, we determined a higher (each hazard has an initial intensity of 5 and is triggered with a high probability from every hazard) and a lower (each hazard has an initial intensity of 1 and is triggered with a low probability from every hazard) limit. This range is transformed to vary from 0 to 1, and the MH is normalized to be included in this range for a better view of the scenarios’ intensity level.

Scenario 1 appears to be more destructive than scenario 2—0.66 vs. 0.47—, of which the intensity is just below the midpoint of the intensity range, 0.47. Their difference is due to the landslide in scenario 1. For scenarios 3 and 4, consisting of three hazards, the MH values are high, with scenario 4 reaching 0.83 (the highest value in the present analysis). The highest value is due to the enhanced values of EQ and FL against RF and RG. Finally, the scenarios comprising four hazards also show high values (0.75 and 0.79), with only one differentiation to the primary natural hazard.

The evaluated method has several positive aspects, including its ease of understanding and application. One significant potential of the method involves the final multi-hazard intensity being directly multiplied with vulnerability to quantify proportional consequences in a study area and the resulting risk. The resulting risk can be used for generating multi-risk maps for each scenario. However, some aspects of the method could be improved. For instance, the calculation of MH is derived from adjusted intensities, neglecting to consider the sequence of hazards in the scenarios (e.g., see scenarios 5 and 6 of Table 3). This may lead to an inaccurate reflection of the final multi-hazard intensity. Another drawback is the lack of an objective method to compare different scenarios. Consequently, identifying the most damaging scenario and effectively addressing its consequences becomes challenging for the stakeholders to mitigate. It is essential to address these limitations to ensure the reliability and effectiveness of the multi-hazard risk mapping process.

4. Conclusions

The present research modified an existing multi-hazard methodology, originally designed to analyze the interaction of natural and technological hazards. In this work, we adapted to address hazards specific to post-mining areas, encompassing natural, mining, and technological hazards. The final outcome of the methodology is the first step (out of two steps) toward post-mining risk analysis for multiple hazards. The multi-hazard intensity as quantified in this work has the potential to be directly used in risk assessment at the next stage (to be presented in future works).

The initial step involved identifying the most significant hazards from each category and determining their respective initial intensities and levels of interaction using an interaction matrix, as determined by the authors. The authors considered the historical data, the literature, and similar cases related to the Megalopolis mining complex to determine the initial intensities and the level of interactions. The initial intensities were modified and increased, and the resulting values were appropriately summed up to create various multi-hazard scenario intensities (MH). The ultimate objective was to identify the most destructive scenario by comparing the normalized multi-hazard (MH) intensities. However, it is important to consider the need for future adjustments of these adjusted properties to ensure a more precise implementation of this methodology for natural, mining, and technological hazards.

Scenarios with more hazards show greater MH, normalized to a range from 0 to 1 for better comparison. Scenario 1, featuring a landslide hazard, is the most destructive, with an MH intensity of 0.66, while scenario 2 follows with an intensity of 0.47. Scenarios 3 and 4, each with three hazards, have high MH values, with scenario 4 reaching the highest intensity of 0.83. However, Liu’s methodology (initial and modified) has limitations, such as inaccuracies in MH calculations (it does not consider the sequence of hazards) and the absence of an objective scenario comparison approach.