Identification and Analysis of Surface Interrupture Landforms in the Kopet-Dagh Region in Iran, Using Machine Learning Techniques

ABSTRACT
This study investigates surface interrupture patterns in the tectonically active Kopet-Dagh region along the Iran–Turkmenistan border, utilizing advanced machine learning algorithms to enhance geomorphometric analysis for optimized water resource management and sustainable land-use planning. The methodology employed integrates high-resolution Digital Elevation Models (DEMs), Python-based analytical tools, and advanced statistical methods to extract key terrain attributes such as slope, aspect, and lineaments. Machine learning algorithms, including random forest and multivariate regression models, were applied to model surface interrupture patterns, achieving accuracies of 85% and 78%, respectively. The results reveal that surface interrupture predominantly occurs in areas with high lineament density (covering 49.51% of the study area), influenced by factors such as slope, fracture density, and surface-water flow patterns. The random forest model identified slope and fracture density as the most influential predictors, demonstrating the efficacy of combining modern machine learning methods with traditional geomorphometric analyses. This research provides an innovative framework for understanding surface water dynamics in tectonically active regions, offering valuable tools for water resource managers and land-use planners to optimize conservation and allocation strategies. The findings also contribute significantly to the expansion of knowledge of geomorphological and tectonic processes in semi-arid mountainous environments and support the development of sustainable water management strategies.
Extended Abstract
Introduction
Surface interrupture is a pivotal phenomenon in Earth sciences that significantly influences landscape evolution and surface water dynamics. This study examines surface interrupture patterns in the Kopet Dagh region of northeastern Iran, a tectonically active area, to support effective water resource management. It is motivated by the need to deepen understanding of tectonic and geomorphological controls on hydrology in semi-arid, mountainous regions. The primary objective is to identify and analyze surface interrupture patterns by integrating geomorphometric analysis with advanced machine learning techniques. Two research questions are addressed: (1) What are the main factors controlling surface interrupture patterns in the Kopet Dagh region? (2) How can machine learning improve the accuracy of modeling these patterns for water management? By proposing a novel analytical framework, this study aims to fill critical gaps in the understanding of surface water dynamics in this region.
 
Methodology

Terrain attributes (slope, flow direction derived via the D8 algorithm, curvature, etc.) were derived from a 10 m Digital Elevation Model (DEM) using the open-source RichDEM Python library, which provides high-performance tools for hydrological analysis.
Lineaments were extracted from the DEM using automated edge-detection algorithms, resulting in a lineament density covering about 49.5% of the study area.
Fracture density was calculated by dividing mapped fracture lengths by unit area (mean ≈ 0.4355; median ≈ 0.4768). Flow accumulation was also modeled using the D8 algorithm to understand dominant drainage directions.
Machine learning analyses included (a) multiple linear regression (yielding R² ≈ 0.399), (b) Random Forest analysis (to assess predictor importance), and (c) K-means clustering (to identify distinct spatial patterns of geomorphic attributes). These analyses were combined with conventional geomorphological interpretation to form a robust framework for modeling surface water flow and storage in the region.

These methods were integrated into a comprehensive analytical framework for modeling and interpreting surface interrupture patterns and their implications for water resources.
 
Results and Discussion
The DEM analysis revealed that the Kopet Dagh study area (≈ 14,193 km²) has a rugged mountainous topography with elevations ranging from about 354 to 2,962 m. This terrain configuration, shaped by active tectonics, strongly influences surface water flow. Lineament analysis showed a high density of structural lineaments (average density ≈ 0.495), indicating pervasive tectonic deformation. The dominant flow direction is to the southeast, reflecting the regional gradient and channel network configuration. Fracture density (mean ≈ 0.4355) emerged as a key structural indicator of subsurface controls on infiltration and water storage.
The Random Forest analysis confirmed that fracture density and slope are the most influential predictors of surface interrupture patterns. These findings imply that regions with high fracture and lineament densities are critical for hydrological connectivity. In practice, this suggests targeted water management strategies should focus on areas of intense tectonic fabric. More broadly, the integration of machine learning with geomorphometric analysis proved effective in identifying the factors that shape surface water dynamics, offering a powerful approach for optimizing water resource allocation.
 
Conclusion
This study employed an integrated approach combining high-resolution DEM analysis, automated image processing, and machine learning to evaluate surface interrupture landforms in the Kopet Dagh region. The results emphasize that fracture density and slope are the primary controls on surface interrupture, while roughly half of the study area exhibits intense lineament coverage indicative of active tectonics. These insights offer valuable guidance for sustainable land-use planning and water resource management in similar semi-arid mountainous regions.
However, limitations of this study include the moderate DEM resolution and limited field data for model validation. Future research should incorporate higher-resolution topographic datasets and field-based observations to refine the models and further improve prediction accuracy for water management applications.
 
Funding
There is no funding support.
Authors’ Contribution
Authors contributed equally to the conceptualization and writing of the article. All of the authors approved thecontent of the manuscript and agreed on all aspects of the work declaration of competing interest none.
 
Conflict of Interest
Authors declared no conflict of interest.
 
Acknowledgments
We are grateful to all the scientific consultants of this paper.