ABSTRACT
The present study aims to investigate the impact of land-use/land-cover changes on streamflow discharge in the Khorram Rud watershed, located in Lorestan Province, western Iran. This research conducted using the Soil and Water Assessment Tool (SWAT) model, integrated with Geographic Information Systems and remote sensing techniques. Input datasets comprised a DEM, soil maps, LULC maps for 2001 and 2020 (derived from Landsat 5 and 8 satellite imagery), as well as daily and monthly climatic and hydrometric records spanning a 28-year period (1996–2020). Following the successful calibration and validation of the model employing the SUFI-2 algorithm, the net effect of LULC changes was analyzed by comparing the outputs of two distinct scenarios: the "baseline" scenario (representing 2001 LULC) and the "change" scenario (representing 2020 LULC). The analysis of the LULC maps revealed substantial transformations throughout the study period, including a 69.99% reduction in dense forest cover, a 33.96% decline in sparse forest and rangelands, alongside increases of 72.57% in agricultural lands and 61.68% in built-up areas. Simulation results derived from the SWAT model indicated a discernible impact of these changes on streamflow discharge. Consequently, the mean annual streamflow under the LULC change scenario increased by 0.16 m³/s (equivalent to 1.8%) relative to the baseline scenario. Based on the paired-sample t-test, this upsurge was found to be statistically significant at the 95% confidence level (p-value = 0.003). Overall, LULC changes within the Khorram Rud watershed were identified as one of the contributing factors to the observed increase in streamflow discharge.

Extended Abstract
Introduction
In the context of global water scarcity and climate change, land-use and land-cover change (LU/LCC) is recognized as a principal factor influencing hydrological processes at the watershed scale. Activities such as deforestation, agricultural expansion, and urban development alter the physical characteristics of the land surface, thereby transforming critical hydrological parameters including infiltration capacity, evapotranspiration, and runoff generation. These changes primarily operate by reducing soil water infiltration and increasing surface runoff, which can exacerbate phenomena such as flooding, diminish groundwater recharge, and weaken base flow in rivers. The consequences of these transformations can be more severe in mountainous and semi-arid regions such as the Zagros, which possess high ecological sensitivity.
The Khorram Rud watershed in Lorestan Province, due to its hydropolitical role (as a headwater of the Karkheh River), the reliance of local communities on agriculture and livestock for their livelihoods, and its location in a water-stressed region, necessitates sustainable water resource management. Numerous studies in various parts of the world (including Har et al., 2021; Maru et al., 2023; Abdulai et al., 2025) and Iran (Baloei et al., 2021; Rezaei Moghaddam et al., 2025) utilizing hydrological models such as SWAT have confirmed a significant relationship between land-cover change and increased runoff. However, a quantitative assessment of this relationship in the Khorram Rud watershed, considering the changes over the past two decades, presents a scientific and practical necessity. The primary objective is to simulate and quantify the impacts of land-use changes from 2001 to 2020 on the streamflow discharge in the Khorram Rud watershed using the SWAT model. The findings can provide a scientific basis for decision-makers to guide environmental planning and integrated water resource management in this watershed.



Methodology
This study employed an applied-analytical approach through hydrological modeling. The primary modeling tool was the semi-distributed physical model SWAT (Soil and Water Assessment Tool), implemented within the ArcGIS 10.4 environment using the ArcSWAT extension. In this research, the required data were collected and prepared for modeling in four main categories: (1) topographic data based on an ASTER Digital Elevation Model (DEM) with a 30-m resolution for watershed delineation, drainage network extraction, and subdivision into 21 sub-basins; (2) a soil map from the FAO global database; (3) land-use maps for the years 2001 and 2020, which were developed and validated through supervised classification of Landsat satellite imagery into five main classes; and (4) climatic and hydrological data, including daily precipitation and temperature data from the Khorramabad synoptic station and monthly discharge data from the Cham Anjir hydrometric station for a 28-year period (1993 to 2020).
During the implementation and analysis process, following the initial model setup and creation of Hydrological Response Units (HRUs), sensitivity analysis, calibration, and validation of the model parameters were performed using the SUFI-2 algorithm in the SWAT-CUP software, based on observed discharge data. The periods 1996-2012 and 2013-2020 were designated for calibration and validation, respectively. Model performance was evaluated using the statistical indices NSE, R², and PBIAS.
To assess the net effect of land-use changes, two scenarios were defined and executed: a baseline scenario applying the 2001 land use map and a change scenario applying the 2020 land-use map, while all other input data remained constant.

Results and Discussion
The findings reveal three key sets of results. First, a comparison of land-use maps from 2001 and 2020 indicates extensive changes within the watershed. Specifically, the area of dense forest decreased by 69.99%, and sparse forest decreased by 33.96%. In contrast, agricultural lands expanded by 72.57%, settlement areas by 61.68%, and grasslands by 106.53%. These changes, which were predominantly concentrated in the northern and eastern parts of the watershed, established the primary context for hydrological transformation.
Subsequently, the accuracy and performance of the SWAT model were evaluated. Following calibration and validation using the SUFI-2 algorithm, the model demonstrated good performance. The Nash-Sutcliffe Efficiency (NSE) and coefficient of determination (R²) values were 0.66 during the calibration period and ranged from 0.80 to 0.85 during the validation period. Furthermore, the Percent Bias (PBIAS) index fell within an acceptable range (±10%), confirming the model's reliability for simulation purposes. However, the P-factor values (0.49-0.66) were below the desirable threshold of 0.7, indicating some uncertainty in peak flow simulations, though the low R-factor values (0.33-0.51) suggested acceptable overall uncertainty.
Finally, the quantitative impact of these land-use changes was investigated on the runoff regime. A comparison of outputs from the baseline scenario (2001 land use) and the change scenario (2020 land use) revealed a significant influence of land-cover alterations. Consequently, the mean annual streamflow increased by 0.16 m³/s (equivalent to 1.8%). Based on the paired t-test, this increase is statistically significant at the 95% confidence level (p-value = 0.003). Furthermore, the slope of the upsurge flow trend was steeper in the change scenario (0.164) than in the baseline (0.161), indicating a gradual intensification over the long term; a trend evident in 76% of the study period years. The analysis of hydrological components revealed that surface runoff increased by 23.5%, while groundwater flow decreased by 13%, demonstrating a shift from an infiltration-dominated to a runoff-dominated system. The time of concentration declined from 4.2 hours in the baseline scenario to 3.7 hours in the change scenario. These findings underscore the pivotal role of land-use change in exacerbating surface runoff and increasing flood potential, particularly during intense rainfall events.
Conclusion
This study, utilizing the integration of remote sensing and hydrological modeling, investigated the impact of land-use changes on the streamflow discharge of the Khorram Rud watershed over a 25-year period (1996-2020). The findings reveal that extensive land-cover transformations, particularly a 69.99% decrease in dense forests, a 33.96% decline in sparse forests and rangelands, and a 72.57% expansion of agricultural lands, have acted as one of the effective factors in increasing the watershed's streamflow discharge.
Based on the SWAT model results, land-use changes led to an increase in mean annual streamflow of 0.16 m³/s (equivalent to 1.8%). Long-term trend analysis indicates that the slope of the increasing discharge trend was steeper in the change scenario (0.164) than in the baseline scenario (0.161), indicating a gradual intensification of land-use change effects over time. This increasing trend was evident in 76% of the study period years, emphasizing its stability and continuity. In terms of the mechanism of effect, the reduction in soil infiltration capacity due to the replacement of forest cover with agricultural lands and settlement areas was identified as the primary driver of these changes. According to the optimized model parameters (such as decreased Manning's roughness coefficient and increased runoff curve number), land-use changes appear to have affected the hydraulic characteristics of the watershed surface. This altered pattern was particularly evident in extreme events such as 2019, in which an 11.1% increase in peak discharge was recorded in the change scenario; however, these values pertain to a single event, and comprehensive statistical analysis considering different return periods is needed for conclusions about the general flood pattern.
The findings can provide a suitable scientific basis for developing sustainable watershed management strategies, environmental planning, and adopting appropriate policies to mitigate the negative effects of land-use changes. Conducting similar studies in homogeneous watersheds and investigating the combined effects of land-use changes and climate change are suggested as future research directions.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.




Authors’ Contribution
All authors contributed equally to the conceptualization of the article and writing of the original and subsequent drafts.

Conflict of Interest
The authors declare no conflict of interest.

Acknowledgments
The authors would like to thank all participants of the present study