Introduction

Geomorphology, as a fundamental branch of earth sciences, examines surface changes and investigates geomorphological processes. Landforms, as the diverse shapes of the Earth's surface, result from these processes and provide crucial information about geological history, erosion, and environmental changes. The study area in this research is the basins of the southern slope of the Sahand mountain range. The Qaleh-e-Chay basin, which originates from Sahand and ultimately enters the Lake Urmia basin. The Sufi-e-Chay basin, on which the Alavian Dam was also built, passes through the cities of Maragheh and Bonab and finally enters the Lake Urmia basin.The southern slopes of the Sahand Mountain Massif in northwestern Iran represent a region rich in geomorphic diversity due to their unique geographical location and exposure to tectonic, geological, and climatic factors. This area includes the Qaleh Chay, Soufi Chay, Mardagh Chay, and Lilan Chay basins, as well as parts of the Qarangho basin, all of which are sub-basins of Lake Urmia. Precise identification and analysis of these landforms can play a critical role in sustainable natural resource management, regional planning, and mitigating natural hazards such as floods. With advancements in remote sensing technologies and digital elevation models (DEMs), more accurate and rapid analyses of these landforms have become feasible. This study aims to identify, classify, and analyze the landforms of the southern Sahand slopes using the Topographic Position Index (TPI) and three advanced algorithms: MLMSR, CMLSR, and SPSR.

Methodology

The study utilized DEM data with a spatial resolution of 30 meters and the TPI to analyze and classify landforms. The TPI, an effective index in geomorphological studies, evaluates the topographic position of each pixel relative to its neighboring pixels. Positive TPI values indicate elevated areas (e.g., peaks and ridges), while negative values denote lower areas (e.g., valleys). Three algorithms—MLMSR (Multi-Layered Morphological Spatial Representation), CMLSR (Complex Multi-Level Summit Recognition), and SPSR (Single Point Summit Recognition)—were employed to process DEM data and extract landforms. Each algorithm applies different methods for analyzing elevation data to identify and classify landforms. The research process involved acquiring DEM data, calculating the TPI, applying algorithms, generating landform maps, and analyzing the results. The algorithms were evaluated for their performance in areas with varying characteristics, such as mountainous and flat regions.

Results and Discussion

The analysis revealed that the southern Sahand slopes encompass ten primary landform types, each with distinct characteristics. Narrow valleys and channels were predominantly observed in steep, mountainous areas in the northern and eastern parts of the region, while plains and flatlands were concentrated in the southern and lower sections. Ridges and elevated plateaus were prominent in higher altitudes, reflecting the influence of tectonic and erosional processes on landform development. A comparison of algorithms showed that MLMSR excelled in identifying peaks and ridges in mountainous areas. SPSR was more effective for precise classification of flat and plain areas, while CMLSR demonstrated satisfactory performance in recognizing complex landforms and conducting multi-scale analyses. The generated maps provided comprehensive information on the distribution and diversity of landforms, serving as a foundation for further studies.

Conclusion

The results showed that the study area consists of various landforms such as narrow valleys, flat plains, hills, ridges and high plateaus due to diverse topographic and geological conditions. Each of these landforms has unique characteristics and their distribution in the region is influenced by factors such as slope, slope direction, altitude, lithology type and tectonic activities. In mountainous and steep areas, narrow valleys and high drainages are most concentrated and these areas indicate intense erosional activities. In contrast, plains and flat areas in the downstream and marginal parts have been formed due to extensive sedimentation processes. Also, hills and high plateaus are seen at medium and high altitudes, indicating the effect of wind and water erosion on the formation of these landforms. The algorithms used in this study each provided different capabilities in identifying and analyzing landforms. The MLMSR algorithm performed better due to its high ability to identify complex shapes such as peaks and ridges. In contrast, the SPSR algorithm was more suitable for flat areas and plains due to its high accuracy in processing pixels. The CMLSR algorithm also provided the ability to analyze landforms at different scales and allowed for the extraction of more details from land structures. In this study, digital elevation model (DEM) data with an accuracy of 30 meters was used to analyze topographic locations. Due to its high accuracy and detail, these data enabled rapid and automatic analysis of landforms and can be used in other similar areas. The analyses performed showed that the TPI index, as an effective tool in distinguishing and classifying landforms, has high capabilities in geomorphological studies.