Assessing the Effects of Land Use Change on Flood Occurrence in the Mordaq Chai Watershed Using the SWAT and Artificial Neural Network Models

Document Type : Research Article

Authors

1 Professor in Geomorphology, Faculty of Planning and Environmental Sciences, University of Tabriz and Iranian Hazardology Association, Tabriz, Iran

2 M. Sc in Geomorphology, Faculty of Planning and Environmental Sciences, University of Tabriz, Tabriz, Iran

3 Postdoctoral Researcher in Geomorphology, Faculty of Planning and Environmental Sciences, University of Tabriz, Tabriz, Iran

4 Ph.D in Geomorphology, Faculty of Planning and Environmental Sciences, University of Tabriz, Tabriz, Iran

Abstract

Over the past few decades, significant changes have occurred in the Mardagh Chay watershed, primarily due to the development of agricultural lands, the expansion of settlements, and other activities. During this period, several events such as surface flows and river overflows have taken place, leading to flooding in settlements and damage to infrastructure. Accordingly, the present study examines and evaluates the impact of land use changes on flooding in the Mardagh Chay watershed in East Azerbaijan Province. First, land use classification was performed using an artificial neural network model, and land use maps were generated for 2001, 2011, and 2021. Subsequently, the SWAT model was applied to the classification results. According to the findings, over a period of twenty years, with the exception of barren lands, all other land use classes showed a dominant increasing trend, with the rangelands of the region experiencing the greatest expansion. The results of the SWAT model also revealed that the runoff trend in the Mardagh Chay basin has largely corresponded with land use changes during this period. In particular, the growth of pasture and built-up areas in the region has had a direct impact on the increase in runoff in the southern areas. Considering the effect of the topography and slope of the basin on flow trends, as well as the greater imperviousness of land surfaces in the southern areas compared to others, the flood potential in these areas has been determined to be significantly higher.
Introduction
Floods are considered one of the most important and abundant geomorphic hazards in the country, which cause a lot of damage every year. Recent events in the course of the Mordaq River have indicated the frequency and intensity of floods in the basin. However, the changes in hydrological and climatic regimes of the basin and what causes these changes in the basin have not been well studied. During the last few decades, significant developments have occurred in the Mordaq chai basin with the aim of developing agricultural lands, expanding settlements and other purposes. During this period, several events have been observed in surface flows and river flooding, including the flooding of settlements and the destruction of infrastructure during the rainy season. The present research has evaluated the impact of land use changes on floods in Mordaq chai basin in East Azerbaijan province.
Material and Methods
Mordaq Chai basin is located at latitudes between 37° 16΄ and 37° 44΄ north, and at longitudes between 46° 21΄ and 46° 30΄ east. The basin area is about 332 Km2. The elevation variations of the basin range from 1567 meters at the outlet of the basin to 3693 meters in Sahand Mountain. In this study land use classification was done using artificial neural network model and land use map was obtained in 2001, 2011 and 2021. Then, the classification results were compared, for which the LCM model was used. The SWAT model has been implemented on the results of land use classification. In addition, the sensitivity assessment of the results has been considered in two phases of calibration and validation, in this regard, land use in 2001 was recalibrated in the statistical period of 1985-1999 and validated in the statistical period of 2000-2001. Also, the land use of 2011 was calibrated in the statistical period of 2002-2009 and validated in the statistical period of 2010-2011 and the land use of 2021 was calibrated in the statistical period of 2012-2019 and validated in the statistical period of 2020-2021.
Results and Discussion
Over the course of twenty years, with the exception of barren lands, other land use classes have shown a predominant upward trend, and the pastures in the region have witnessed the most significant expansion in terms of area. This shift highlights notable changes in land use patterns across the area. The increase in pastureland could be attributed to various factors, such as changes in agricultural practices, conservation policies, or even natural shifts in the ecosystem. This growth in pastures may have positive implications for biodiversity conservation, soil erosion prevention, and overall environmental improvement in the region. However, to fully understand the underlying reasons for these changes and their long-term impacts, a more detailed analysis and further data examination are required. According to the physiography and topography governing the basin, the situation is such that in the northern half of the region, we see more natural land cover (including snow, water, barren and pasture) than in the southern half, where human activities (gardens and built-up lands) dominate. The role of Mordaq Chai river in this field has also been significant and for example, most of the gardens of the region are distributed linearly next to this river and its branches, or the villages of the region that have the largest area of built-up land are located in the vicinity of these rivers. In 2001, wasteland, pastures, and gardens accounted for the largest share of existing land uses at 61.21%, 29.1%, and 8.01%, respectively, and snow, built-up areas, and water were next in rank at 1.59%, 0.073%, and 0.016%, respectively. In 2021, the relative share of barren lands has decreased sharply and reached 28.21% of the entire region, and on the other hand, pastures have grown a lot and their relative share in the region has become 54.45%.
Conclusions
The simulation outcomes generated by the SWAT model revealed certain limitations in its application within the Mordaq Chai basin. Specifically, the calibration and validation phases at the overarching basin scale, which incorporated land use data from varying years, yielded suboptimal and unsatisfactory results. This primary deficiency was quantitatively demonstrated by the fact that the observed runoff data points frequently fell outside the simulated uncertainty range predicted by the model. The root cause of this discrepancy has been identified as a fundamental inadequacy in the model's algorithmic representation of snowmelt-driven runoff. The SWAT model appears to systematically miscalculate the volume and timing of runoff originating from the melting snowpack, particularly during the critical hydrological period spanning from late winter through late spring. This improper estimation significantly compromises the accuracy of the overall surface runoff simulation, as snowmelt is a major contributor to the basin's hydrologic regime. Furthermore, the analysis of long-term trends established a strong correlation between the hydrological behavior of the basin and anthropogenic changes to the landscape. The recorded trajectory of runoff in the Mordaq Chai basin has been largely congruent with and influenced by the concurrent pattern of land use and land cover change (LULC) over the same period. A direct causal relationship was observed: the expansion of two specific land use classes—pastureland and built-up/urbanized areas—within the southern regions of the watershed has acted as a primary driver for increased runoff generation. This phenomenon is exacerbated by the underlying physiographic characteristics of the basin. The southern areas are distinguished by a more pronounced topography and steeper slopes, which naturally accelerate flow velocity and reduce concentration times. Coupled with this, the land surface in these southern zones has become increasingly impervious due to urbanization and the degradation of natural vegetation cover. This combination of factors—steeper slopes and reduced infiltration capacity—creates a synergistic effect that drastically elevates the flood potential, making these southern areas disproportionately more vulnerable to high-flow events and flooding compared to other parts of the basin.
Acknowledgements
We are grateful to all the scientific consultants of this paper.

Keywords

Main Subjects


©2025 The author(s). This is an open access article distributed under Creative Commons Attribution 4.0 International License (CC BY 4.0)

Achugbu, I. C., Olufayo, A. A., Balogun, I. A., Dudhia, J., McAllister, M., Adefisan, E. A., & Naabil, E. (2022). Potential effects of Land Use Land Cover Change on streamflow over the Sokoto Rima River Basin. Heliyon, 8(7), e09779. https://doi.org/10.1016/j.heliyon.2022.e09779
Arnold, J. G., Moriasi, D. N., Gassman, P. W., Abbaspour, K. C., White, M. J., Srinivasan, R., ... & Jha, M. K. (2012). SWAT: Model use, calibration, and validation. Transactions of the ASABE55(4), 1491-1508. https://doi.org/10.13031/2013.42256
Bakhtiyari, M., Komeh, Z., & Memarian, H. (2018). A Comparison of Fuzzy Analytic Hierarchy Process, Artificial Neural Network and Area Density in Quantitative Evaluation and Landslide Susceptibility Mapping within GIS Framework (Case Study: Simereh Homiyan Watershed). Journal of Geography and Environmental Hazards, 7(3), 19-40. [In Persian] https://doi.org/10.22067/geo.v0i0.67234
Choto, M., & Fetene, A. (2019). Impacts of land use/land cover change on stream flow and sediment yield of Gojeb watershed, Omo-Gibe basin, Ethiopia. Remote Sensing Applications: Society and Environment, 14, 84-99. https://doi.org/10.1016/j.rsase.2019.01.003
Dastres, R., & Soori, M. (2021). Artificial neural network systems. International Journal of Imaging and Robotics (IJIR)21(2), 13-25. https://hal.science/hal-03349542v1
Eastman, J. R., & Toledano, J. (2018). A short presentation of the Land Change Modeler (LCM). Geomatic Approaches for Modeling Land Change Scenarios, 499-505. https://doi.org/10.1007/978-3-319-60801-3_36
Ghazavi, R., & Fazeli, A. (2019). Investigation the effects of land use change on surface runoff using Remote sensing and SWAT model. Journal of Water and Soil Conservation, 25(6), 191-206. [In Persian] https://doi.org/10.22069/jwsc.2019.12697.2739
Haleem, K., Khan, A. U., Ahmad, S., Khan, M., Khan, F. A., Khan, W., & Khan, J. (2022). Hydrological impacts of climate and land-use change on flow regime variations in upper Indus basin. Journal of Water and Climate Change13(2), 758-770. https://doi.org/10.2166/wcc.2021.238
Khoi, D. N., Loi, P. T., & Sam, T. T. (2021). Impact of future land-use/cover change on streamflow and sediment load in the Be River Basin, Vietnam. Water13(9), 1244. https://doi.org/10.3390/w13091244
Maru, H., Haileslassie, A., Zeleke, T., & Teferi, E. (2023). Analysis of the impacts of land use land cover change on streamflow and surface water availability in Awash Basin, Ethiopia. Geomatics, Natural Hazards and Risk14(1), 1-25. https://doi.org/10.1080/19475705.2022.2163193
Neitsch, S. L., Arnold, J. G., Kiniry, J. R., Williams, J. R., & King, K. W. (2005). Soil and water assessment tool: theoretical documentation.
Rahmati, F., Rezaei Moghaddam, M. H., & Nikjoo, M. R. (2021). The Effects of Land use Change on Erosion & Flood in Saqqez basin before & after Shahid Kazemi Dam Construction, Arc Swat model. Geography and Human Relationships, 4(1), 127-144. [In Persian] https://doi.org/10.22034/gahr.2021.285039.1545
Rezaei Moghaddam, M. H., & Rahimpour, T. (2024). Preparation of flood hazard potential map using two methods: Frequency Ratio and Statistical Index (Case study: Aji Chai Basin). Environmental Management Hazards, 10(4), 291-308. [In Persian] https://jhsci.ut.ac.ir/article_95951.html?
Salarin, F., Tatian, M. R., Ghanghermeh, A., & Tamartash, R. (2022). Modeling land cover changes in Golestan province using land change modeler (LCM). Journal of RS and GIS for Natural Resources, 12(4), 47-70. [In Persian] https://doi.org/10.30495/girs.2021.681267
Shirani, K., Heydari, F., & Arabameri, A. (2017). Comparison of artificial neural network and multivariate regression methods in landslide hazard zonation, case study: Vanak Basin, Isfahan province. Watershed Engineering and Management, 9(4), 451-464. [In Persian] https://doi.org/10.22092/ijwmse.2017.101725.1031
Sugianto, S., Deli, A., Miswar, E., Rusdi, M., & Irham, M. (2022). The effect of land use and land cover changes on flood occurrence in Teunom Watershed, Aceh Jaya. Land11(8), 1271. https://doi.org/10.3390/land11081271
Tali, M. P. A. (2011). Land use/land cover change and its impact on flood occurrence: a case study of upper Jhelum floodplain. (Master's thesis). University of Kashmir.
Yari, M., Soltani-Gerdefaramarzi, S., & Ghasemi, M. (2019). Investigation of the Effect of to Land Use Changes on Flood Hydrograph and Fluctuations of Groundwater Level in Part of Ghareh Souh Basin. Journal of Geography and Environmental Hazards, 8(3), 41-58. https://doi.org/10.22067/geoeh.2024.88423.1493
CAPTCHA Image