Analysis of the impact of roughness coefficient changes due to land use changes on the hydraulics of the Kashafrood River

In recent years, the effects of land use change have been considered as one of the most important concerns in management of watersheds. The Kashafrood basin in Mashhad has been affected by land use change, urbanization and the development of agricultural water resources. Therefore, conducting research with the help of a physical and distribution-based model in order to evaluate land use changes on hydraulic components is required. Land use changes are among the factors that influence the natural cycle in the ecosystem.

Saadati et al. (2006) simulated the effect of land use changes on surface runoff in the Kesilian Basin in two time periods with different land uses using the SWAT model. Runoff (With precipitation variables and land use type) was simulated with R² = 0.69, and it was found that the types of land uses have different abilities to produce runoff, also the Soil and Water Assessment Tool (SWAT) does not have adequate accuracy in the simulation of minimum runoff (Saadati et al. 2006).

Akhavan et al. (2009) estimated the amount of water (sum of surface runoff and deep underground water supply) and green water (actual transpiration and soil moisture) in the Hamedan-Bahar watershed using the SWAT model. The result of the simulation was satisfactory in most of the stations, especially the output of the basin (Akhavan et al. 2009). Satellite images can provide analysts with accurate and up-to-date information on land cover. The status of land use changes can be specified using data such as maps, aerial photos, satellite imagery, and so on (Shariat Panahi et al. 2010).

Nobert & Jeremiah (2012) evaluated the hydrological response of Wami Basin to land use changes. In order to overcome the problem of lack of observational information on the basin, they used satellite information for the use layer as input for the SWAT semi-distributed rainfall-runoff model, according to which images related to the years 1987–2000 were prepared. The results showed a decrease in forest lands, an increase in agricultural lands and urban areas. Also, these changes have led to an increase in surface flow (runoff) and a decrease in the river base flow.

One of the best ways to evaluate land use changes and riverbed and riverside use in the past is to use aerial photos and satellite imagery. To this end, high-ground resolution spot or Quickbird satellite images are suitable (Ebrahimi & Yazdani 2013).

Aldeen (2013) applied the ArcSWAT model to estimate sediment load in the lands on the left bank of the Mosul Dam in northeast Iraq and finally stated that the factor of runoff coefficient has a significant influence on runoff and sediment rates. This factor is variable at the basin level and its value can vary depending on the type of soil in the land use area. Moreover, the amount of runoff coefficient, in addition to the vegetation cover of the area, depends on the soil texture characteristics of that area (Aldeen 2013).

Borzou et al. (2014) evaluated the effect of land use on runoff rate and sediment yield in the Gilard basin in Tehran Province using the SWAT model. After running the model, the simulated results were compared with the measured values. A comparison of runoff and sedimentation rates for three different types of land uses indicated that due to land use change from rangelands and forests to agricultural lands and residential areas, there was a 30% increase in surface runoff and sediment yield also increased by 4 times. HajiHosseini et al. (2015) used SWAT software to investigate the effect of land use changes on runoff in the Hirmand transboundary basin during 2012–1990. They came to the conclusion that the increase in the cultivation of aquatic products in the region has reduced the runoff of the Hirmand River and the agricultural development in this plain is associated with a decrease every year on average. Farrokhzadeh et al. (2015) evaluated the effect of land use on the suspended load of the Yalfan Hamedan watershed using the SWAT model. The output results of the model with two different land uses (in 1988 and 2012) show the great effect of land use change on the suspended load of the basin so that land use change from 1988 to 2012 led to an increase of about 30.62% and it was found that the model has a good ability to simulate the sediments of this basin. In Northwest China, in the headwater region of the Heihe River, the combined impact of climate variability and land use change has been examined and the near future has been predicted while studying a decade ago or the near past by Zhang et al. (2016). The results of research showed that the joint effects of land use change and climate variability, the surface runoff, groundwater discharge, ET and stream flow are increased (Zhang et al. 2016). Zuo et al. (2016) investigated the impact of land use change and climate change on the runoff and sedimentation rate of the Huangfuchan Basin in China using the SWAT hydrological model and four scenarios. The results showed a significant decrease in both annual runoff and sedimentation of the basin.

In the city of Tirana, Albania, Valmir et al. (2019) conducted a study and showed that the highest volume of runoff in bare lands is 38,829.91 l on 74% slopes and the lowest volume is in forest lands amounting to 12,840.6 l on 64% slopes. Additionally, the highest amount of sediment in bare lands is 515.15 tons per hectare on 62% slopes and the lowest amount is in forest lands amounting to 18.18 tons per hectare on 64% slopes (Valmir et al. 2019). According to research by Wei & Junguo (2019) severe floods have attracted the attention of policymakers and engineers around the world, and land use management was recognized as one of the drivers of severe floods. They also proved that flood peaks are more sensitive to changes in roughness (Wei & Junguo 2019).

Pengfeng et al. (2021) studied the YZA Reservoir in northwest China and they found that the deviation theory to avoid sediment and the plan of coordinated delivery of water and sediment can increase the useful life of a reservoir, and a coordinated plan for the dispatch of water and sediment can effectively alleviate siltation in water diversion channels (Pengfeng et al. 2021).

Ifeanyi et al. (2022) investigated the potential effects of land use cover change on river flow in the upper basin of the Sokoto Rima River in Nigeria. They concluded that ET is a major factor for changes in streamflow due to changes in land use in the catchment. The sensitivity of the model to land use/land cover change is reasonable, but further research is recommended to compare the results with land use/land cover change (Ifeanyi et al. 2022). The main goal of land management is to investigate the effects of land use change and its hydraulic effects on the basin. The precipitation-runoff process of the basin is the result of the influence of many factors such as climate, land use, etc (Jinghua et al. 2022). In 2022, a study was conducted regarding the turbulence in the channels, the simulations well illustrated that the evolution of along-canopy vortices includes the onset of flow instability from small perturbations, vortex merging and growing, vortex equilibrium and secondary instability. The above processes are well indicated by the coherence of velocity time series in longitudinal and transverse components (Yuan et al. 2022). During an effective simulation study of the flow of a medium curved, it was determined that the free-flowing branch can greatly diminish the downstream helical flow strength; overall, the variation tendency of the ratio of helical flow strength to discharge squared is immune to the small range of change in stable inflow (Xiaolong et al. 2022). Junhao et al. (2023) studied the VMD-CNN-AM-BOA-BiLSTM model and turned it into an effective data-driven tool in the practice of hydrological forecasting and practical reference for water resources management and flood warning. The passage of Kashafrood through the vicinity of the residential areas of the study area raises concerns about the risks and damages caused by the changes in land use and its impact on flood hydraulics. The present research provides information that can be used to understand the various causes of the concerns and identify the vulnerable and dangerous points. In addition, the extent of flooding due to the change of use has been determined based on hydraulic studies and by determining the flood bed with a higher return period in order to clarify the intensity and extent of protective measures. In this regard, through the use of the HEC-RAS model, the impact of the change in land use on the flow of hydraulics was determined.

The physical characteristics of the basin as well as the river are very important in hydrological studies and flow hydraulics. They play an important role in determining flow, sedimentation and the extent of erosion.

According to the hydraulic model's need for flood discharge values as input parameters of the model, these values were estimated by experimental methods (Dicken, Fuller, Krieger) which are more useful. The estimated flood values with different return periods were selected from Dicken's method as a more suitable method. In this regard, the data of flood events at the Olang-e-Asadi hydrometric station were used for reasons such as the accuracy, the length of the data collection period and its good location (Table 2 and Table 3 ).

First, all the hydrometer stations that were located in or near the area were checked. But Olang Asadi station was considered as the base station. This was due to its location in the center of the basin, as well as its having a suitable statistical period and the a high similarity in terms of climate, and physical characteristics with the study area.

A₁ is the basin area (km²), A₂ is the station area (km²), Q₂ is the station flood (m³) with different return periods, Q₁ is the basin flood (m³) with different return periods.

The peak flood discharge in the Dicken method is obtained using the basin area and the regional coefficient (Zahrabi et al. 2011). The regional coefficient was determined according to the statistics of hydrometric stations in the region. For this purpose, the recorded statistics of the peak flood discharge of the Olong Asadi hydrometry station in a period of 32 years were statistically analyzed using HyFA software, and the flood values of the station in the period of different returns were selected based on the best fit (three-parameter log-normal).

The topographic map was entered into the GIS environment and cross-sections along the river were extracted using the capabilities of this environment and the HEC-GEO-RAS extension for the ArcView software environment. HEC-RAS and its companion GIS extension HEC-GeoRAS can aid in the development of flood inundation maps. HEC-RAS is an incomplete, yet easy-to-use software package for determining water surface profiles in a wide variety of streams (El-Naqa & Jaber 2018).

Based on the input information on the topography and morphology of the river in the GIS environment, the required information of the HEC-RAS model was extracted and entered into this model. Afterwards, the river was simulated in different hydraulic regimes. Based on the obtained hydraulic results, the most suitable bed was selected for the river path. Then, by entering the results of the HEC-RAS model into GIS with respect to river topography and water surface profile, a flood zoning map within the floodplain area was prepared for different return periods of the river. The river's roughness coefficient was calculated by the Cowan method (Borzou et al. 2014).

According to the Froud number determined in each section, the flow regime was classified into three categories: supercritical, critical and subcritical. The supercritical regime is the Froud number greater than 1, the critical regime is the Froud number equal to 1, and the subcritical regime is the Froud number less than 1 (KRRWC 2016).

To determine the severity and extent of lateral erosion, the following classes were separated and the river was evaluated by measuring two factors: terrace height (H) and erosion activity (T). To this end, the height of terraces was calculated using HEC – RAS software and was zoned in Arc view software into three classes (H₁ to H₃) and T classes with field surveys were also assigned to four classes (T₁ to T₄).

T₁ indicates the river flows on a calm bed and erosion is evident at the site of arcs; T₂ indicates the effects of erosion are seen in 10–30% of the river path; T₃ indicates the effects of erosion are seen in 30–50% of the river path; T₄ indicates the effects of erosion are visible in over 50% of the river path.

H₁ indicates the depth of wall caused by lateral erosion between 0 and 2 m; H₂ indicates the depth of wall caused by lateral erosion between 2 and 5 m; H₃ indicates the depth of wall caused by lateral erosion between 5 and 10 m.

All the parameters used in river modeling such as hydraulic radius, river bed slope and cross-sectional area are derived from the geometry of the river and only Manning's roughness coefficient is estimated based on the opinion of experts and vegetation characteristics. Based on this, the only parameter that can be subjected to sensitivity analysis is the roughness coefficient. This parameter plays an important role in simulating the flow and determining the characteristics of the flow. Therefore, the effects of the annual land use changes in the vicinity of the river on its hydraulics have been investigated by applying changes to the Manning's roughness coefficient of ±10%, ± 20% and ±30% and comparing the results to the actual conditions.

The ecosystem of the Kashafrood River is considered as one of the most important issues of Mashhad. However, with the development of human settlements on the floodplain, and riverside, as well as the drilling of unauthorized wells, the discharge of construction debris, and the total human interference in the river area, a lot of land use changes and destruction have occurred.

Also, industrial, service and domestic wastewater after entering the Kashafrood is reused for agricultural irrigation of the surrounding lands which results in land pollution. These show the need to organize and manage the river basin properly (KRRWC 2016).

The lengths of erosion classes along the riverbanks are shown in Table 4. As displayed in the table, the greatest erosion length belongs to the H₂ class on both the right and left banks. Moreover, regarding erosion severity, class T₄ has the highest amount of erosion on both banks. Accordingly, the sections that fall into the H₄T₄ –H₃T₃ –H₃T₄ erosion class have high erosion potential. These sections are mostly seen in population centers that threaten riverside resources. So, they need protection programs along the path.

The most important and influential residential area in the Kashafrood is Mashhad Metropolis. The city of Mashhad has experienced a lot of growth and development due to its special feature of accepting immigrants so according to the official statistics, the city's population of 2,807,464 people in 2011 reached 3,347,165 people in 2021. The population of the city is predicted to be 4,446,042 people in 2031 (KRRWC 2016).

The terms and coordinates of the Meander (Table 6) include the following: ANG refers to the arc angle; R refers to the arc radius; L refers to the arc length; T refers to the chord length; and B refers to the distance between confluence of chords and the location of the arch. The coordinates of the beginning of the Meander: (XPC, YPC); the coordinates of the middle of the Meander : (XPI, YPI); the coordinates of the end of the Meander: (XPT, YPT)

The sensitivity of the HEC-RAS model to changes in Manning's roughness coefficient was evaluated by applying 10, 20, and 30% changes to Manning's roughness coefficient, which indicates land use changes in a 40-year period in the Kashafrood basin. Due to the large number of results, not all of them are presented here, and the results of R², RMSE and AME parameters are shown in Table 9. As presented in Table 9, with the increase of changes in the value of the roughness coefficient, the model outputs have more errors. It should be noted that the highest sensitivity to changes in the roughness coefficient is related to the parameters of the flow level and the width of the upper surface. It is worth noting that an increase or decrease in the value of the roughness coefficient affects the output results and the value of the corresponding parameter is not the same. But in general, the values of RMSE and AME error parameters are equal, and R² also has an acceptable value (Table 9). What is clear from the following results is that by precisely estimating the roughness coefficient value, one can expect the acquisition of good accuracy. In this regard, Shuai et al. (2014) conducted a study on the sensitivity of the HEC-RAS one-dimensional model and referred to the high impact of the roughness coefficient due to land use changes on results. (The statistical data presented in Table 9 shows the comparison between primary and secondary hydraulic conditions.)

Based on the height of the terraces formed along the riverbank and studies of the erosion in the river, four erosion classes were identified. Areas where the effects of lateral erosion of the river are more than 50% are recommended as suitable for geotextile application. The surveys of the lateral erosion status of the river under study demonstrate that river walls are unstable in many places and often have a high potential for erosion so some of the trenches in the riverside are being damaged and advancing to agricultural lands or buildings (Figure 11).

The results of calculations of river dimensions at the final equilibrium state and movement threshold conditions at all river sections revealed that in 71% of the cross-sections, the situation is unstable and in the remaining 29%, the situation is stable. Also, due to land use changes that have taken place, most sections have an unstable situation. The highest values of shear stress and shear velocity in unstable sections are 815.41 N/m² and 6.21 m/s, respectively. On the contrary, the highest values of shear stress and shear velocity parameters in stable sections are 516.7 N/m² and 4.92 m/s, respectively. The minimum value of the Shields parameter is 0.42 in unstable sections and 1.1 in stable sections. As shown in the results, the average value of the parameters presented in the stable sections is higher than in the unstable sections. Such a result also applies to the minimum and standard deviation of the presented parameters. On the contrary, the maximum value of parameters and also the Shields parameter in the unstable sections is higher than the stable ones. Also, by applying changes in river bed roughness coefficient parameters (resulting from the construction operations of government agencies) and changes in the CN coefficient (resulting from land use changes in the basin), the model showed an increase in runoff, which itself affects the hydraulics of the river. It should be noted that the highest sensitivity to roughness coefficient changes is related to the parameters of flow area and upper surface width. It is worth noting that an increase or decrease in the value of the roughness coefficient is effective in the output results and the related parameter value is not the same. But overall, the error parameters of RMSE and AME are equal, and R² also has an acceptable value.

As previously mentioned, human encroachments in the Kashafrood bed and bound have occurred in various ways and their severity and weakness vary from place to place. Along the Kashafrood, based on the height of the terraces formed along the riverbank and studies of the erosion in the river, four erosion classes were identified. Surveys of the lateral erosion status of the river under study suggest that river walls are unstable in many places and often have high potential for erosion so some of the trenches along the river are being demolished and advancing to agricultural lands or buildings.

According to the results, the following are suggested for the management of the Kashafrood River basin:

1. The changes in land use, including dredging and reopening of the river, should be corrected. Also, agricultural and garden activities should be stopped in the river bed and should be carried out in the sanctuary with special criteria.

2. Considering the land use changes in the Kashafrood River basin and the entry of raw sewage into the river, in order to improve the quality of running water resources in the Kashafrood, it is necessary to design and implement aeration structures. It should be noted that the construction of these structures affects the hydraulic conditions of the river.

3. Preventing more usage changes and using geotextile for the stability of the river walls.

We wish to express our gratitude to the great professor Mr Hossein Sedqi, a pioneer of water engineering sciences and also the thesis advisor related to this article, who provided valuable and lasting services to the society and students during his lifetime, may his soul rest in peace.

None.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.