Introduction

Alluvial fans are landforms in many arid and semi-arid regions. Tectonic, climate, channel incision, and avulsion can abruptly change the surface of an alluvial fan and, subsequently lead to the formation of a new fan. With time, alluvial fans exhibit significant differences in terms of weathering, the rate of headward erosion, the drainage pattern, and flooding and depositional processes. At first, young or active fans with distributary drainage pattern develop at the mountain front. If a faulted structure develops, its continuous propagation toward the river stimulates its deflection. The deflection of river and its migration can stimulate the abandonment of the oldest fan and the formation of a new fan along the deflected river. Long-term expo-sure to physiochemical weathering can lead to the degradation of old surfaces. Soil degradation induced by gully erosion represents a worldwide problem in the many arid and semi-arid countries, such as Iran. Iran is recognized as the second in the world in terms of soil erosion where approximately 2.5 billion tons of fertile lands are lost per year. Gully erosion can stimulate multiple environmental hazards, such as desertification, increasing sediment load in rivers and reservoirs, flood, and soil productivity loss. This study assessed: (1) the relationship between the evolution of alluvial fan surfaces and its effects on the geometric variability of these landforms; (2) the relationship between the geometric evolution of different surfaces and changes in the gully erosion pattern and controlling factors.

Methodology

The present study was implemented on southern slopes of the Aladag Mountains. The maximum and minimum altitude is 297 and 885 meters above sea level. Based on the Demarton climate index, the climate of the study area is classified as semi-arid. Drainage pattern, surface roughness and morphology were recorded to distinguish active fan surfaces from inactive fan surfaces.

In this study, the ALOS DSM global digital surface model (DSM) from AW3D30 was used to extract geometric indices. This dataset is a digital surface model (DSM) with global coverage and a horizontal resolution of about 30 meters, which is developed based on 3D topographic data of the world with higher resolution (5 meters).

In order to calculate the vegetation indices, data from the MSI sensor of the Sentinel-2 satellite at the L2A processing level were used. The L2A surface data, after applying atmospheric corrections, provide the reflectance of the land surface, and its synchronized dataset adjusts for radiometric changes resulting from processing line updates and ensures the temporal consistency of the data for time series analyses. Also, to extract the land use and land cover map, the World Cover product was used. In this study, 11 geometric indices were used to investigate the characteristics of active and inactive alluvial fans.

In order to quantitatively assess and analyze the characteristics of gully erosion, data from the Multispectral Imager (MSI) of the Sentinel-2 satellite were used. In order to accurately identify and separate pixels related to gully from other land uses and surface phenomena, a supervised machine learning method with the Support Vector Machine (SVM) algorithm was used. In the next step, the distribution of gully and key indices related to their dimensions and dispersion were calculated.

Results and Discussion

According to the statistical results, gully formation was 2.5 times higher in active surfaces than in inactive surfaces. These differences could be due to the exposure of active surfaces to recent flooding processes and their smaller area compared to inactive surfaces. The significant increase in the average gully length in active surfaces indicates higher flow energy and continued erosion processes. The significant increase in the average gully width in inactive surfaces is due to the processes of gully wall destruction, lateral erosion, and relative filling over time. Length-width ratio was significantly higher in active alluvial fans than in inactive surfaces. The greater dispersion of runoff and the distribution of coarser and more permeable sediments facilitate the conditions for the longitudinal expansion of gully on active surfaces. On the contrary, the presence of highly weathered sediments and the time that has elapsed since the last flooding in inactive surfaces stimulated the lateraldevelopment of gully. A positive correlation between gully width and flow strength indicates the dominance of lateral erosion and gullies in the inactive surfaces compared to active alluvial fans. In active surfaces, the initial energy of the flow, the development of the drainage network, and the elevation conditions are the most important factors affecting the gully erosion pattern. In inactive alluvial fans, an increase in the topographic moisture index, an increase in surface roughness, and the stabilized state of the surfaces are more important in the gully erosion pattern. This comparative analysis shows that although some topographic factors in both alluvial fan generations affect the development of the gully, the evolutionary stages of the alluvial fan, by affecting their geometry, could lead to differences in the relative importance of geometric factors and their role in the gully erosion pattern.



Conclusion

In this study, the geometric characteristics of different alluvial fan surfaces were first evaluated. Then, the morphometric characteristics of the alluvial fans were measured and quantified to determine the pattern of gully erosion in different generations of alluvial fans. Finally, the relationship between the evolution of alluvial fan surfaces, variability in their surface geometry, and the dynamic of gully erosion pattern was investigated. Based on the results, alluvial morphometric indices show significant differences between active and inactive alluvial fans. Active surfaces with higher drainage density, longer gullies, and higher length-to-width ratio indicate younger and more erosionally dynamic systems. In contrast, inactive surfaces with wider and deeper gullies can be an indication of the evolution of gully pattern in the absence of intense erosional activity. Alteration of hydrological connectivity between alluvial fan and contributing catchment is recognized as a principle method to control gully erosion. Reducing hydrological connectivity in contributing catchment andshifting hydrology pathway on alluvial fan, such as restoring vegetation and building check dam at the bottom of gullies in catchment.