📄 earth_science

Open-Cast Mining Surface Risk Assessment Using GIS and Statistical Modelling: A Multi-Criteria Framework for Sustainable Mine Safety and Environmental Management

This study proposes an integrated GIS and statistical modelling framework utilizing Multi-Criteria Decision Analysis to generate a Surface Risk Index that effectively zonation open-cast mining areas into risk categories based on factors like slope, vegetation, and proximity to settlements, thereby supporting sustainable mine safety and environmental management.

Original authors: SHATRAJIT GOSWAMI

Published 2026-07-10

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: SHATRAJIT GOSWAMI

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the Earth's surface as a giant, complex video game map. In this game, open-cast mining is like a player digging massive, deep holes to grab valuable resources (coal). While this is great for the economy, it leaves the landscape looking a bit like a cratered moon, with steep cliffs, dusty air, and rivers that might be clogged with mud.

For a long time, checking if these mining sites were safe or dangerous was like trying to find a needle in a haystack by walking around with a magnifying glass. Experts had to visit every single spot, measure things by hand, and guess where the next landslide might happen. It was slow, expensive, and often missed the big picture.

The New Game Plan: A Digital Crystal Ball

This paper introduces a smarter way to play: a "Digital Crystal Ball" built using GIS (Geographic Information Systems) and statistical modelling. Think of GIS as a super-powered layer cake. Instead of just looking at the top layer (the ground), the researchers stack up transparent sheets of data:

Sheet 1: Where the mines are.
Sheet 2: How steep the hills are (Slope).
Sheet 3: How much greenery is left (Vegetation/NDVI).
Sheet 4: Where the water flows (Drainage).
Sheet 5: Where people live (Settlements).
Sheet 6: Where the roads are.

By mixing these layers together using a special recipe called the Analytical Hierarchy Process (AHP), the researchers created a Surface Risk Index (SRI). This is like a "danger meter" that gives every spot on the map a score from 0 to 1.

What the Numbers Say (The Scoreboard)

The researchers designed this recipe to be applied to major Indian coalfields like Raniganj, Jharia, Talcher, and Singrauli. However, for this specific study, they didn't run the numbers on live data from these fields yet. Instead, they used a synthetic dataset (a carefully constructed, made-up database) to prove that their method works.

Here is what the "Digital Crystal Ball" revealed in this demonstration:

Green is Good, Red is Bad: Areas with lots of trees (high vegetation) had lower risk scores. Areas with steep slopes, lots of mining activity, and dense populations had high risk scores.
The Recipe Weights: The researchers decided which factors mattered most. In their final recipe, Vegetation (NDVI) was the biggest player, weighing in at 0.22. Land Use was next at 0.20, followed by Distance from the Mine at 0.18. The other factors like drainage, slope, and settlements made up the rest.
The Results: The model was incredibly good at spotting the difference between safe and dangerous zones. It achieved a score (called AUC) of 0.89, which is like getting an A+ on a test. It correctly identified that high-risk zones would tend to cluster right around active mines and busy roads, while low-risk zones would be located far away in green, quiet areas.

What This Paper Rules Out

The paper explicitly argues against the old way of doing things. It says that relying only on field observations and engineering checks is not enough because it misses the "spatial variability"—the way risks change across a huge area. It also rejects the idea that these risks are random. The data shows that risks are not scattered by chance; they are tightly linked to specific factors like how close you are to a mine pit or how steep the ground is.

How Sure Are We?

The authors are very confident in their findings, but they are careful to say how they know.

The Proof: They didn't just look at pictures; they used Multiple Linear Regression to prove the connection. The model explained about 79.4% of the changes in risk (an R² of .794). This means the factors they chose (trees, slope, distance, etc.) are the real reasons the risk goes up or down in their simulation.
The Simulation: The study used a synthetic dataset to show that the framework works. The authors state that the model is "statistically significant" with a p-value of less than .001, meaning the results are extremely unlikely to be a fluke.
The Limit: While the model is robust, the paper notes that this specific study was a methodological demonstration. It suggests that for real-world use, the framework needs to be applied to actual, live mining regions (like the ones listed above) and validated with historical accident data in the future.

The Final Takeaway

This paper doesn't claim to have "solved" mining safety forever. Instead, it offers a transferable toolkit. It's like giving mine managers a new, high-tech compass. By using this GIS and statistical framework, they can create "Risk Zonation Maps" that color-code the land:

Dark Green: Very Low Risk (0.00–0.20)
Light Green: Low Risk (0.21–0.40)
Yellow: Moderate Risk (0.41–0.60)
Orange: High Risk (0.61–0.80)
Red: Very High Risk (0.81–1.00)

This allows governments and companies to see exactly where to build buffer zones, where to stop digging, and where to plant trees to protect nearby villages. It turns a chaotic, dangerous landscape into a map you can actually read, helping to make mining safer and more sustainable for everyone.

Technical Summary: Open-Cast Mining Surface Risk Assessment Using GIS and Statistical Modelling

Problem Statement
Open-cast mining, while economically efficient and dominant in coal extraction for developing nations like India, generates severe environmental and geotechnical hazards. These include landscape alteration, slope instability, land subsidence, groundwater contamination, and threats to nearby communities. Traditional risk assessment methods, relying heavily on field observations and engineering investigations, are often costly, time-consuming, and fail to capture the spatial variability of risks across large mining regions. There is a critical need for an integrated approach that combines environmental, geological, and socio-economic variables to support sustainable mine safety and management.

Methodology
The study proposes an integrated framework combining Geographic Information Systems (GIS), Remote Sensing, and statistical modelling to assess surface risks. The methodology follows a Multi-Criteria Decision Analysis (MCDA) approach utilizing the Analytical Hierarchy Process (AHP).

Data Integration: The framework synthesizes multiple datasets, including:
- Remote Sensing: Sentinel-2 MSI (10m) for Land Use/Land Cover (LULC) and Normalized Difference Vegetation Index (NDVI); Landsat 8/9 (30m) for thermal analysis.
- Terrain Data: SRTM and CartoDEM (30m) for Digital Elevation Models (DEM), slope, and elevation.
- Vector Data: OpenStreetMap for roads and settlements; HydroSHEDS for drainage networks; and mine boundary data from Eastern Coalfields Limited (ECL) and Mahanadi Coalfields Limited (MCL).
Risk Factor Selection: Seven primary risk indicators were identified: NDVI, LULC, Distance from Mine, Distance from Roads, Slope, Drainage Density, and Settlement Density.
Weighting (AHP): Relative weights were assigned to these factors using AHP. The resulting weights were: NDVI (0.22), LULC (0.20), Distance from Mine (0.18), Drainage Density (0.12), Slope (0.10), Settlement Distance (0.10), and Distance from Roads (0.08). A Consistency Ratio (CR) of <0.10 confirmed the reliability of the weighting.
Surface Risk Index (SRI) Generation: A weighted overlay analysis was performed to generate a Surface Risk Index (SRI), classifying areas into five categories: Very Low, Low, Moderate, High, and Very High risk.
Statistical Validation: The GIS-based model was validated using:
- Pearson Correlation: To assess relationships between SRI and predictor variables.
- Multiple Linear Regression: To quantify the influence of variables on SRI.
- Principal Component Analysis (PCA): To identify dominant risk drivers.
- Variance Inflation Factor (VIF): To check for multicollinearity.
- Spatial Autocorrelation (Moran's I): To detect spatial clustering of risk.
- Receiver Operating Characteristic (ROC): To evaluate predictive performance.

Key Results
The study was applied to major Indian coalfields, including Raniganj, Jharia, Talcher, and Singrauli.

Risk Zonation: The resulting maps successfully classified mining areas into distinct risk zones. High-risk areas were identified around active pits, overburden dumps, and densely populated settlements, characterized by steep slopes and sparse vegetation.
Statistical Findings:
- Correlation: A significant negative correlation was found between SRI and NDVI ( $r = -0.61$ ), indicating vegetation reduces risk. Slope ( $r = 0.49$ ), drainage density ( $r = 0.42$ ), and settlement density ( $r = 0.54$ ) showed positive correlations with risk.
- Regression Model: The multiple linear regression model explained approximately 79.4% of the variance in surface risk ( $R^2 = 0.794$ ). NDVI was a significant negative predictor ( $\beta = -0.38$ ), while slope, settlement density, and LULC were significant positive predictors.
- Spatial Clustering: Global Moran's I analysis ( $I = 0.43, p < 0.001$ ) confirmed significant positive spatial autocorrelation, meaning high-risk zones cluster around active mining infrastructure.
- Model Performance: The ROC analysis yielded an Area Under the Curve (AUC) of 0.89, indicating excellent discriminatory capability between high and low-risk zones.
Hypothesis Testing: The null hypothesis (that environmental and anthropogenic variables do not significantly influence surface risk) was rejected. The alternative hypothesis was accepted, confirming that these variables are significant determinants of mining-related surface risk.

Significance and Claims
The paper claims to provide a scientifically robust and transferable methodology for sustainable mine planning and environmental management. By integrating GIS-based MCDA with statistical validation, the framework overcomes the limitations of traditional site-specific assessments, offering a scalable solution for large mining regions.

The study asserts that the proposed Surface Risk Index (SRI) serves as an effective decision-support tool for:

Sustainable mine planning and environmental monitoring.
Disaster prevention and mine closure management.
Policy formulation, including the establishment of buffer zones and overburden stabilization measures.

The authors position this framework as a contribution to responsible mining practices in developing economies, emphasizing its potential to align mining regulations with broader sustainability goals. The methodology is presented as applicable to various coalfields, with specific examples drawn from the Indian context, but the framework is designed to be adaptable to other mining regions globally.

Technical Summary: Open-Cast Mining Surface Risk Assessment Using GIS and Statistical Modelling

More like this