In arid regions, the optimal utilization of river channels to collect rainwater resources can effectively alleviate the problem of water scarcity. This paper employs visual MODFLOW software to simulate the infiltration volume of rainwater under different schemes, with the objective of identifying the optimal infiltration volume. The results indicate that the infiltration volume of rainwater is 0.6, 79.7, 36.5, 62, and 10,000 m3 under different design schemes. COMSOL software is employed to simulate the infiltration situation of rainwater under different schemes, and the scheme with the most favourable infiltration effect is identified. The simulation results indicate that scenarios 2, 4, and 5 demonstrate superior early-stage infiltration performance. Ultimately, the optimal infiltration scheme is selected by integrating the infiltration volume, infiltration effect, construction, economic, and ecological factors to achieve the objective of enhancing rainwater harvesting and utilization. This study offers effective and reasonable measures for addressing the challenge of water resources management in arid regions. It provides suitable reference schemes for river channel rainwater storage projects and offers innovative approaches to rainwater resource utilization, thereby contributing to the alleviation of water scarcity in arid areas.

Water is an indispensable material basis for the survival of species on Earth and a vital resource for the development of modern society, playing a significant role in the sustainable development of the national economy and society, and in ensuring the well-being of the people (Lingyue & Hong 2000). Water scarcity, deterioration of the water environment, and frequent occurrences of floods and droughts are the main water-related challenges faced by China (Xinghua 2007). Although floods are disasters, they can become a usable water resource when managed and utilized properly by humans (Qixin 2023). Currently, China is in a stage of urbanization. Rapid urbanization has led to significant changes in the urban underlying surface (Tong et al. 2021), with a sharp increase in the imperviousness of the ground (Kunyang et al. 2021) and a continuous decrease in the urban green space ratio (Song et al. 2021). Rapid urbanization has also led to a dramatic increase in water use for industrial and agricultural development, forcing excessive exploitation of groundwater. This over-exploitation has caused ecological and environmental issues such as soil salinization (Song et al. 2023) and seawater intrusion (Chenzhe et al. 2022), severely damaging our natural ecosystem. Rapid urbanization has left most cities facing water shortages (Huang & Zhou 2013), with nearly 20% of cities in a severe state of water scarcity. The rapid evolution of human society has disrupted the original ecological balance, leading to frequent urban flood disasters (Yunjie 2017). Additionally, the massive loss of rainwater leads to insufficient replenishment of urban groundwater, further exacerbating the water crisis in cities (Zhangjun et al. 2011). The discrepancy between demand and supply of water resources is becoming increasingly pronounced, necessitating the implementation of efficacious strategies to enhance the collection and utilization of rainwater resources and to achieve the sustainable utilization of water resources to meet current and future needs.

Yongqiang et al. (2005) elucidated the implications of flood resource utilization, elucidating the contradiction between water supply and demand. They also proposed a methodology for the process and a risk management strategy, which initiated a surge in rainwater and flood resource utilization in China's hydrological research. Yanwei et al. (2011) proposed an innovative approach to regulating rainwater resources: low impact development (LID). This approach employs a range of technical tools and design methods that offer novel insights into the utilization of rainwater resources. Shiguo et al. (2005) conducted a study to identify the optimal conditions for utilizing reservoirs to divert water from the Nenjiang River. They also proposed specific and feasible measures for the use of reservoirs and other projects to utilize rainwater floods. Xingchao's (2018) analysis of the concept and basic principles of underground reservoirs and the advantages of applying them to sponge cities demonstrated the irreplaceable significance of underground reservoirs in the construction of sponge cities. According to recent studies, Lulu (2019) constructed a SWMM model to determine the effectiveness of stormwater flood control and the hydro-ecological effects of various LID facilities placed in the study area. Yinghui et al. (2013) employed the hydrodynamic module of MIKE11 software to construct a river model of the Oujiang River basin. They then analysed and calculated the parameters of flood level and flood propagation time for each characteristic section of the main stream. Shuang et al. (2013) established a SWAT model in the Nansihu basin with the objective of simulating non-point source nitrogen and phosphorus pollution. Furthermore, a response study of lake sedimentation was conducted.

Brouwer & Van (2004) conducted a research assessment on the effects of rain and flood utilization in the Netherlands, examining the ecological, economic, and social impacts. Harris & Kennedy (1999) studied the potential for urban rain and flood resource development and utilization from a water supply perspective and incorporated it into urban development planning. Kumar et al. (2022) developed a transient groundwater flow model using Modflow for their study area. They predicted groundwater levels and recharge and determined the relationship between groundwater and cultivated land area. Vasistha & Ganguly (2022) studied the seasonal and depth variation of water quality in two natural lakes of Haryana using traditional methods in collaboration with GIS modelling system using inverse distance weighing (IDW) method.

The utilization of hidden storage resources from rain and flood is a new approach and method in water management, transforming rain and flood water into a natural resource that can be comprehensively utilized by humans through various artificial means (Guochun 2023). This study introduces a novel approach to improving the collection and utilization of rainwater resources from river channels. This approach differs from previous measures, such as the installation of LID facilities and underground reservoirs to collect rainwater resources. In this paper, the Visual MODFLOW software is employed to simulate the rainwater infiltration under different scenarios with the objective of identifying the scenario with the most optimal infiltration. The COMSOL software was employed to simulate the infiltration of rainwater under a range of scenarios, with the objective of identifying the scenario that yielded the most effective infiltration. Finally, the most suitable and effective infiltration scheme is selected by combining the infiltration volume and infiltration effect, as well as construction, economy, ecology, and other aspects. The results provide a scientific basis for addressing water scarcity in the arid regions of Western Liaoning, offer appropriate reference schemes for hidden storage projects in river channels, and present new ideas and methods for the utilization of rain and flood resources, so as to achieve the alleviation of the problem of water resources shortage in the arid areas.

Overview of the study area

Physical and geographic profile

Gongyingzi Town, under the jurisdiction of Kazuo Left Wing Mongolian Autonomous County, Chaoyang City, Liaoning Province, is located in the northern part of the Kazuo Left Wing Mongolian Autonomous County. It borders Boluochi Town, Wulan and Shuo Mongolian Township of Chaoyang County to the east, faces Wohugou Township and Shuiquan Township to the south, adjoins Wanshou Street of Jianping County to the west, and neighbours Zhongsanjia Town to the north. The area covers 121.62 km2, situated at 119°84′E, 41°35′N. The terrain of the town is high in the west and south, low in the east and north, with an average elevation of about 350 m. Plains, slopes, and mountains each account for one-third of the area, featuring Sileng Mountain, Louzi Mountain, and Jiutou Mountain, with the Mangniu River flowing from west to east into Shuiquan Township. The location of Gongyingzi Town is shown in Figure 1. The specific simulation area is demarcated along the river valley direction, with the upstream boundary near the observation well ZK5-G1-BK9-Gongyingzi Town in Tazixia Village, extending 1,000 m upstream from the backwater area to the vicinity of Dadiebu Village, and downstream of the seepage interceptor dam near the village of Houtabeiyingzi. The northeast boundary is demarcated by Dadiebu Village-Guanfen, and the southwest boundary is by National Highway 207. The simulation area covers approximately 7.71 km², as shown in Figure 2.
Figure 1

Location map of Gongyingzi Town.

Figure 1

Location map of Gongyingzi Town.

Close modal
Figure 2

Map of the simulation area.

Figure 2

Map of the simulation area.

Close modal

Climate and soil types

The study area is located within Gongyingzi Town, Kazuo County, and features a continental monsoon climate. The main climatic characteristics are as follows: the spring season is characterized by less rain and more drought, the summer is hot with concentrated rainfall, the autumn is clear with ample sunshine, and the winter is bitterly cold with rare snowfall. The annual average temperature is 8.7 °C, with a temperature difference of 1.5° between the north and south of the area. The average annual sunshine duration is 2,807.8 h, and the average frost-free period is 144 days.

The soils of Kazuo County can be divided into three main categories: brown earth, cinnamon soil, and meadow soil, with a zonal distribution of soil types. Cinnamon soil and brown earth are the zonal soils of Kazuo County. The presence of the Daling River and its tributaries has resulted in the formation of meadow soils along their banks. Therefore, the soil type in the study area is classified as meadow soil, primarily distributed on the river floodplains, terraces, and high terraces on both sides of the river. These soils have a moderate organic matter content, high moisture content, and are relatively loose. Meadow soil is distinguished by its high water content and relatively soft soil structure. It is characterized by a high organic matter content, a moderate pH, and a high nutrient content. Due to its high water content and low density, the bearing capacity of meadow soil is typically low. Therefore, it should not be used directly for building foundations without improvement. To enhance stability, drainage and reinforcement measures should be employed.

Hydrogeological conceptual model

Determination of the simulation range in the study area

The selection of the simulation range in the study area was mainly based on the following principles:

  • (1) The characteristics of the terrain and landforms in the study area were fully considered, using natural boundaries as computational boundaries whenever possible, and complete hydrogeological units as the simulation area.

  • (2) Ensure, as much as possible, the distribution of groundwater observation wells on the boundaries to monitor changes in boundary water levels.

  • (3) Avoid the influence of interception dams as much as possible, with boundaries above and below the dam located far from the influence of these dams.

Generalization of the hydrogeological model

  • (1) Generalization of aquifer internal structure

Considering the hydrogeological conditions of the study area, the aquifer is treated as a single unconfined aquifer with the base of the quaternary strata as its bottom. Based on the type, lithology, thickness, and hydraulic characteristics of the aquifer, the model is generalized as a heterogeneous isotropic aquifer, locally considered homogeneous.

  • (2) Generalization of hydraulic characteristics of the aquifer

The groundwater level in the study area varies due to dry and wet seasons, exhibiting non-steady state flow. However, generally, the regional groundwater follows laminar flow and obeys Darcy's law, and can be considered as a non-steady two-dimensional planar flow.

  • (3) Boundary treatment of the study area

In this study, based on the distribution of observation wells (holes) along the boundary of the engineering area, it is advisable to use the time series function provided by Moldflow to define them as ‘specified head boundaries’ that change over time. These can also be approximately generalized as the impermeable boundaries of a shallow groundwater system (controlled by topography, with short and rapidly alternating flow). The boundary conditions should be determined based on specific hydrogeological conditions. Given the hydrogeological conditions of the study area and long-term observation well data on boundary groundwater levels, this study treats the boundaries of the study area as known water level boundaries. The water levels at individual boundary points are then interpolated using contour maps of groundwater levels from different months in 2020.

  • (4) Treatment of source and sink terms

Source terms mainly consider atmospheric precipitation infiltration recharge and river infiltration recharge. Atmospheric precipitation infiltration recharge is divided into zones (excluding the riverbed section), with the precipitation infiltration intensity calculated based on surveyed precipitation infiltration coefficients and precipitation amount per unit area; river infiltration recharge is calculated by dividing the Mangniu River in the simulation area into seven sections, based on river water level, riverbed sediment thickness, river width, and vertical infiltration coefficient of the sediment. Sink terms mainly consider evapotranspiration, agricultural water extraction, industrial water extraction, and groundwater recharge to rivers. Evapotranspiration is calculated based on evaporation intensity; agricultural water extraction includes domestic water usage in each village, calculated based on population numbers and water quotas, while agricultural irrigation water is zoned based on extraction intensity, which is calculated based on the area of crops and irrigation quotas; industrial water extraction is calculated based on surveyed extraction amounts; groundwater recharge to rivers is treated in the same way as river infiltration recharge.

Establishment of the groundwater flow numerical model

Based on the generalized hydrogeological model of the study area, the mathematical model for groundwater flow in the study area is formulated as follows:
In the equation, x, y are spatial coordinates, (m); K(x,y) is the permeability coefficient, (m/day); μ is the degree of water contribution to the submersible aquifer; t is the time variable (day); W(x, y, t) is the source-sink term (m/day); vertical infiltration and evaporation. Z is the aquifer floor elevation (m); h(x, y, t) is the groundwater level to be sought (m); h0(x, y, t) is the initial water level value in the seepage field (m); h1(x, y, t) is the value of the first category boundary water level (m).

Verification of the groundwater flow numerical model

Using the identified model for simulation calculations, 20 April and 20 August 2020 were selected as the time periods for model verification. Representative observation wells were chosen, and the calculated water level values were compared with the actual measured water level values, which are provided by the local hydrological observatory for the observation wells. The observation well water level observation curves are shown in Figures 3 and 4.
Figure 3

Comparison of observed and calculated water levels in observation wells on 20 April 2020.

Figure 3

Comparison of observed and calculated water levels in observation wells on 20 April 2020.

Close modal
Figure 4

Comparison of observed and calculated water levels in observation wells on 20 August 2020.

Figure 4

Comparison of observed and calculated water levels in observation wells on 20 August 2020.

Close modal

As shown in the above figures, the verification results for April show that the maximum error occurred at point G6, with a value of 0.01 m, and the minimum error occurred at point G5, with a value of −0.29 m. From the August verification results, the maximum error occurred at point G6, with a value of 0.13 m, and the minimum error occurred at point G5, with a value of −0.17 m. The relative error of curve fitting is less than 5% of the water level change during the period, indicating that the model is effective.

MODFLOW simulation

Analysis of rain and flood resources

The flood characteristic curve from 1990 to 2020 shows that the annual rain and flood volume is generally above 5 million cubic metres, indicating that the study area has abundant rain and flood resources available for utilization, as shown in Figure 5.
Figure 5

Flood volume characteristic curve of the study area from 1990 to 2020.

Figure 5

Flood volume characteristic curve of the study area from 1990 to 2020.

Close modal

A statistical calculation of the 30-year series of floods in the study area was conducted. According to the typical year method, the rain and flood volume in a wet year is 12.95 million cubic metres, with 10 flood events occurring and a rain and flood interval of 9 days. In an average water year, the rain and flood volume is 10.09 million cubic metres, with eight flood events occurring and a rain and flood interval of 11 days. In a dry year, the rain and flood volume is 4.71 million cubic metres, with five flood events occurring and a rain and flood interval of 18 days. In an extremely dry year, the rain and flood volume is 1.31 million cubic metres, with two flood events occurring and a rain and flood interval of 62 days; as shown in Table 1.

Table 1

Typical annual legal flood volume and number of flood fields statistical table

PrerequisiteTypical yearFrequencyNumber of floodsAnnual rainfall resources (10,000 m3)Flood site intervals (day)
Abundant 1995 25% 10 1,295 
Normal 2004 50% 1,009 11 
Dry 2006 75% 471 18 
Extremely dry 2013 95% 131 62 
PrerequisiteTypical yearFrequencyNumber of floodsAnnual rainfall resources (10,000 m3)Flood site intervals (day)
Abundant 1995 25% 10 1,295 
Normal 2004 50% 1,009 11 
Dry 2006 75% 471 18 
Extremely dry 2013 95% 131 62 

Design of different infiltration schemes

Rain and flood water is characterized by short duration and high flow rates. As a recharge source, it needs to infiltrate into the underground aquifer within a short time. The infiltration conditions of the surface rock and soil layers in the study area are poor and cannot meet the requirements for rain and flood infiltration. Therefore, the following infiltration schemes were designed:

  • Scheme I: Rain and flood infiltration under natural conditions.

  • Scheme II: Rain and flood infiltration under conditions of cleaning the fine sand layer on the shallow surface of the riverbed.

  • Scheme III: Rain and flood infiltration under conditions of not cleaning the riverbed but only excavating pits and ponds.

  • Scheme IV: Rain and flood infiltration under conditions of constructing river dams and excavating pits and ponds.

  • Scheme V: Rain and flood infiltration under conditions of cleaning the fine sand layer on the shallow surface of the riverbed and constructing river dams.

Calculation of rain and flood infiltration for different schemes

In this example, after establishing the groundwater flow model of the study area, three modules of MODFLOW – input, run, and output – were used. In the input module, hydrogeological parameters such as the permeability coefficient and specific yield of the study area were entered. The model was run using MODFLOW2000 and ZoneBudget in the run module. Finally, the simulation results were viewed through the output module, and parameters were iteratively adjusted based on these results. To calculate the annual infiltration volume of rainstorm resources in the study area and determine the maximum water volume that can be infiltrated by the stratum media with each flood event, the model simulated the annual infiltration volume of rainstorm resources in the study area. To eliminate interference, it was assumed that the aquifer thickness is infinitely large and the groundwater level is relatively low, indicating that the study area has sufficient effective space to store rainstorm resources, ensuring that the underground storage space does not affect the infiltration of rainstorm resources. In the established numerical model of groundwater flow, the infiltration of rainstorm resources was generalized as infiltration through injection wells. By observing the contour maps of groundwater levels at the time of model output, it was determined whether all rainstorm resources were infiltrated. The approximate range of infiltration volume was determined through continuous parameter adjustment, and minor adjustments were made within this range to finally determine the numerical value of the infiltration volume. This method was used to simulate calculations for different water volumes separately.

The initial water level for the simulation was set to 16 January 2020, and the simulation period was one hydrological year. Boundary water levels are as shown in Table 2, and various hydrogeological parameters are the same as those in the existing mathematical simulation model under the conditions of the established interception dam.

Table 2

Known boundary water level values given by infiltration programmes

Boundary pointB1B2J18J22B5B6B7BK5BK6
Jan 301.86 296.632 289.7 286.29 283.36 279.76 271.86 271.4 271.71 
Feb 301.86 297.002 289.06 286.22 283.3 279.7 271.8 271.86 271.97 
Mar 300.86 296.772 289.16 286.24 282.64 279.04 271.14 271.32 272.43 
Apr 300.86 296.392 289.02 285.82 282.67 279.07 271.17 271.37 272.48 
May 300.36 295.882 288.79 285.85 282.62 279.02 271.12 271.39 272.42 
Jun 299.86 295.862 288.52 285.71 282.49 278.89 270.99 272.26 272.29 
Jul 299.44 295.832 288.19 285.5 282.54 278.94 271.04 272.25 272.5 
Aug 299.31 295.822 288.29 285.74 282.62 279.02 271.12 271.86 272.57 
Sep 299.8 296.3 288.78 285.08 282.5 278.9 271 272.1 271.75 
Oct 299.7 296.3 288.75 285.01 282.5 278.9 271 272.07 271.71 
Nov 299.5 296.9 288.73 285.08 282.5 279 271.3 272.11 271.76 
Dec 299 296.4 288.75 285.08 282.5 278.9 271.2 272.11 271.72 
Boundary pointB8B9J23J29B10G3ZK5G1BK9
Jan 273.26 281.46 282.61 284.81 288.384 296.63 302.53 302.85 302.12 
Feb 273.26 281.4 282.37 284.57 288.464 296.63 302.53 302.85 302.12 
Mar 273.24 280.74 282.71 284.91 288.564 296.57 302.73 302.62 302.32 
Apr 273.27 280.77 282.74 284.94 288.664 296.44 302.73 302.24 302.32 
May 273.22 280.72 282.69 284.89 288.764 296.33 302.23 302.24 302.82 
Jun 273.09 280.59 282.56 284.76 288.864 295.92 302.73 302.22 302.32 
Jul 273.14 280.64 282.61 284.81 288.964 295.47 302.72 302.19 301.9 
Aug 273.22 280.72 282.69 284.89 288.864 295.66 302.59 302.18 301.77 
Sep 273.1 280.6 282.64 284.84 288.764 295.38 302.22 302.67 302.2 
Oct 273 280.6 282.59 284.79 288.664 295.37 302.22 302.64 302.18 
Nov 272.6 280.6 282.63 284.83 288.304 295.43 302.67 302.65 302.24 
Dec 272.9 280.7 282.62 284.82 288.384 295.43 302.67 302.65 302.26 
Boundary pointB1B2J18J22B5B6B7BK5BK6
Jan 301.86 296.632 289.7 286.29 283.36 279.76 271.86 271.4 271.71 
Feb 301.86 297.002 289.06 286.22 283.3 279.7 271.8 271.86 271.97 
Mar 300.86 296.772 289.16 286.24 282.64 279.04 271.14 271.32 272.43 
Apr 300.86 296.392 289.02 285.82 282.67 279.07 271.17 271.37 272.48 
May 300.36 295.882 288.79 285.85 282.62 279.02 271.12 271.39 272.42 
Jun 299.86 295.862 288.52 285.71 282.49 278.89 270.99 272.26 272.29 
Jul 299.44 295.832 288.19 285.5 282.54 278.94 271.04 272.25 272.5 
Aug 299.31 295.822 288.29 285.74 282.62 279.02 271.12 271.86 272.57 
Sep 299.8 296.3 288.78 285.08 282.5 278.9 271 272.1 271.75 
Oct 299.7 296.3 288.75 285.01 282.5 278.9 271 272.07 271.71 
Nov 299.5 296.9 288.73 285.08 282.5 279 271.3 272.11 271.76 
Dec 299 296.4 288.75 285.08 282.5 278.9 271.2 272.11 271.72 
Boundary pointB8B9J23J29B10G3ZK5G1BK9
Jan 273.26 281.46 282.61 284.81 288.384 296.63 302.53 302.85 302.12 
Feb 273.26 281.4 282.37 284.57 288.464 296.63 302.53 302.85 302.12 
Mar 273.24 280.74 282.71 284.91 288.564 296.57 302.73 302.62 302.32 
Apr 273.27 280.77 282.74 284.94 288.664 296.44 302.73 302.24 302.32 
May 273.22 280.72 282.69 284.89 288.764 296.33 302.23 302.24 302.82 
Jun 273.09 280.59 282.56 284.76 288.864 295.92 302.73 302.22 302.32 
Jul 273.14 280.64 282.61 284.81 288.964 295.47 302.72 302.19 301.9 
Aug 273.22 280.72 282.69 284.89 288.864 295.66 302.59 302.18 301.77 
Sep 273.1 280.6 282.64 284.84 288.764 295.38 302.22 302.67 302.2 
Oct 273 280.6 282.59 284.79 288.664 295.37 302.22 302.64 302.18 
Nov 272.6 280.6 282.63 284.83 288.304 295.43 302.67 302.65 302.24 
Dec 272.9 280.7 282.62 284.82 288.384 295.43 302.67 302.65 302.26 

The main cause of flooding in the study area is heavy rainfall, which often concentrates within 1–3 days, and a flood event typically lasts about 3–5 days. During a flood event, the flood volume is mainly concentrated within 24 h, accounting for about 70–80% of the total flood volume. Therefore, when using MODFLOW software for numerical simulation, it is only necessary to simulate the rain and flood infiltration volume within 24 h.

  • (1) Rain and flood infiltration under natural conditions

To simulate the infiltration of a single flood event, different infiltration volumes were evenly distributed to each injection well to simulate the effect of infiltration over 24 h. The depth of the groundwater level after infiltration was analysed. If the water level is below the ground elevation, it indicates that the infiltration capacity of a single flood event is greater than the given infiltration volume. For example, with an infiltration volume of 10,000 m3, the contour line of equal water level after infiltration is shown in Figure 6. Observing the groundwater level 24 h after simulation, it was found that the rain and flood resources did not all infiltrate into the aquifer, with local water levels higher than the ground elevation as indicated by the red-circled area, suggesting that the single flood event infiltration volume is less than 10,000 m3.
Figure 6

Groundwater level contour map for an infiltration volume of 10,000 m3 under natural conditions.

Figure 6

Groundwater level contour map for an infiltration volume of 10,000 m3 under natural conditions.

Close modal

Through simulations with different water volumes, it was found that when the water volume is 5,000 m3, the groundwater level 24 h after simulation indicates that all rain and flood resources have infiltrated into the aquifer, with water levels below the ground elevation.

From these results, it can be seen that in the software model, under natural conditions without cleaning the riverbed in the study area, the single flood event rain and flood resource infiltration volume is greater than 5,000 m3 but less than 10,000 m3. For accurate calculation of infiltration volume, this range was further adjusted in small increments and numerically simulated using the software. The final result was 6,000 m3. The contour line of equal water level under this condition is shown in Figure 7.
Figure 7

Groundwater level contour map for an infiltration volume of 6,000 m3 under natural conditions.

Figure 7

Groundwater level contour map for an infiltration volume of 6,000 m3 under natural conditions.

Close modal

From the above results, it is evident that in the software model, under the natural condition of not cleaning the riverbed in the study area, the single flood event rain and flood resource infiltration volume is greater than 6,000 m3 but less than 10,000 m3. To accurately calculate the infiltration volume, this range was slightly adjusted and numerically simulated using the software, with the final result being 6,000 m3.

Based on the above analysis, in a wet year with 10 flood events annually, the infiltration capacity (maximum infiltration) of the stratum media in the study area is 60,000 m3; in an average water year with eight flood events, the infiltration capacity is 48,000 m3; in a dry year with five flood events, the infiltration capacity is 30,000 m3.

  • (2) Rain and flood infiltration under the condition of only cleaning the riverbed

Under the condition of only cleaning the shallow fine sand layer of the riverbed, taking an infiltration volume of 780,000 m3 as an example, the contour line of equal water level after infiltration is shown in Figure 8. Observing the groundwater level 24 h after simulation, it was found that all rain and flood resources have infiltrated into the aquifer, with local water levels below the ground elevation as indicated by the red-circled area. This indicates that the single flood event infiltration volume under this condition is greater than 780,000 m3.
Figure 8

Groundwater level contour map for an infiltration volume of 780,000 m3 under the condition of only cleaning the riverbed.

Figure 8

Groundwater level contour map for an infiltration volume of 780,000 m3 under the condition of only cleaning the riverbed.

Close modal

Through continuous parameter adjustments and simulations with different water volumes, it was found that when the water volume is 830,000 m3, the groundwater level observed 24 h after the simulation indicated that not all rain and flood resources infiltrated into the aquifer, with local water levels higher than the ground elevation. This suggests that the single flood event infiltration volume under this condition is less than 830,000 m3. From these results, it can be concluded that in the model, the rain and flood infiltration volume under the condition of cleaning the shallow fine sand layer of the riverbed in the study area is greater than 780,000 m3 but less than 830,000 m3. For accurate calculation of the infiltration volume, further small adjustments within this range were made, and numerical simulation calculations were performed using the software, with the final infiltration volume determined to be 797,000 m3.

Based on the above analysis, in a wet year with 10 flood events annually, the infiltration capacity (maximum infiltration) of the stratum media in the study area is 7.97 million m3; in an average water year with eight flood events, the infiltration capacity is 6.376 million m3; in a dry year with five flood events, the infiltration capacity is 3.985 million m3.

  • (3) Rain and flood infiltration under the condition of not cleaning the rivered plus excavating pits and ponds

Under the condition of not cleaning the riverbed plus excavating pits and ponds, taking an infiltration volume of 350,000 m3 as an example, the contour line of equal water level after infiltration is shown in Figure 9. Observing the groundwater level 24 h after simulation, it was found that all rain and flood resources have infiltrated into the aquifer, with local water levels below the ground elevation as indicated by the red-circled area. This indicates that the single flood event infiltration volume under this condition is greater than 350,000 m3.
Figure 9

Groundwater level contour map for an infiltration volume of 350,000 m3 under the condition of not cleaning the riverbed plus excavating pits and ponds.

Figure 9

Groundwater level contour map for an infiltration volume of 350,000 m3 under the condition of not cleaning the riverbed plus excavating pits and ponds.

Close modal

Through continuous adjustments and simulations with different water volumes, it was observed that when the water volume is 380,000 m3, the groundwater level 24 h after the simulation indicated that not all rain and flood resources infiltrated into the aquifer, with local water levels higher than the ground elevation. This suggests that the single flood event infiltration volume under this condition is less than 380,000 m3. From these results, it can be concluded that in the model, under the condition of not cleaning the riverbed plus excavating pits and ponds in the study area, the single flood event rain and flood resource infiltration volume is greater than 350,000 m3 but less than 380,000 m3. For accurate calculation of the infiltration volume, further small adjustments within this range were made, and numerical simulation calculations were performed using the software, with the final infiltration volume determined to be 365,000 m3.

Based on the above analysis, in a wet year with 10 flood events annually, the infiltration capacity (maximum infiltration) of the stratum media in the study area is 3.65 million m3; in an average water year with eight flood events, the infiltration capacity is 2.92 million m3; in a dry year with five flood events, the infiltration capacity is 1.825 million m3.

  • (4) River dam plus excavating pits and ponds without cleaning the rivered

Under the condition of constructing a river dam plus excavating pits and ponds without cleaning the riverbed, taking an infiltration volume of 600,000 m3 as an example, the contour line of equal water level after infiltration is shown in Figure 10. Observing the groundwater level 24 h after simulation, it was found that all rain and flood resources have infiltrated into the aquifer, with local water levels below the ground elevation as indicated by the red-circled area. This indicates that the single flood event infiltration volume under this condition is greater than 600,000 m3.
Figure 10

Groundwater level contour map for an infiltration volume of 600,000 m3 under the condition of constructing river dam plus excavating pits and ponds without cleaning the riverbed.

Figure 10

Groundwater level contour map for an infiltration volume of 600,000 m3 under the condition of constructing river dam plus excavating pits and ponds without cleaning the riverbed.

Close modal

Through simulations with different water volumes, it was observed that when the water volume is 630,000 m3, the groundwater level 24 h after the simulation indicated that not all rain and flood resources infiltrated into the aquifer, with local water levels higher than the ground elevation. This suggests that the single flood event infiltration volume under this condition is less than 630,000 m3. From these results, it can be concluded that in the software model, under the condition of ‘river dam construction plus not cleaning the riverbed and cleaning pits and ponds’ in the study area, the single flood event rain and flood resource infiltration volume is greater than 600,000 m3 but less than 630,000 m3. For accurate calculation of the infiltration volume, further small adjustments within this range were made, and numerical simulation calculations were performed using the software, with the final result determined to be 620,000 m3.

Based on the above analysis, in a wet year with 10 flood events annually, the infiltration capacity (maximum infiltration) of the stratum media in the study area is 6.20 million m3; in an average water year with eight flood events, the infiltration capacity is 4.96 million m3; in a dry year with five flood events, the infiltration capacity is 3.10 million m3.

  • (5) Cleaning the riverbed plus river dam construction

Under the condition of cleaning the riverbed plus constructing a river dam, taking an infiltration volume of 1.03 million m3 as an example, the groundwater level 24 h after simulation showed that all rain and flood resources infiltrated into the aquifer, with water levels below the ground elevation. This indicates that the single flood event infiltration volume under this condition is greater than 1.03 million m3.

Through simulations with different water volumes, it was observed that when the water volume is 1.08 million m3, the groundwater level 24 h after the simulation indicated that not all rain and flood resources infiltrated into the aquifer. The contour line of equal water level after infiltration is shown in Figure 11, with local water levels higher than the ground elevation as indicated by the red-circled area, suggesting that the single flood event infiltration volume is less than 1.08 million m3.
Figure 11

Groundwater level contour map for an infiltration volume of 1.08 million m3 under the condition of cleaning the riverbed plus constructing river dam.

Figure 11

Groundwater level contour map for an infiltration volume of 1.08 million m3 under the condition of cleaning the riverbed plus constructing river dam.

Close modal

From these results, it can be concluded that in the model, under the condition of cleaning the riverbed plus constructing a river dam in the study area, the single flood event rain and flood resource infiltration volume is greater than 1.03 million m3 but less than 1.08 million m3. For accurate calculation of the infiltration volume, further small adjustments within this range were made, and numerical simulation calculations were performed using the software, with the final result determined to be 1.062 million m3.

Based on the above analysis, in a wet year with 10 flood events annually, the infiltration capacity (maximum infiltration) of the stratum media in the study area is 10.62 million m3; in an average water year with eight flood events, the infiltration capacity is 8.496 million m3; in a dry year with five flood events, the infiltration capacity is 5.31 million m3.

  • (6) Comparison of different infiltration schemes

The annual infiltration volumes calculated for each infiltration scheme were compared, as shown in Table 3. A comparison chart is presented in Figure 12.
Table 3

Calculation results of annual rainfall infiltration under different infiltration scenarios

Different infiltration programmesAnnual infiltration volume (10,000 m3)
Abundant yearNormal yearDry year
Failure to clean the riverbed 4.8 
Clearing the riverbed 797 637.6 398.5 
Not clearing the riverbed but excavating pit and pond 365 292 182.5 
Not clearing the riverbed but excavating pit and barrage 620 496 310 
Clearing the riverbed and excavating barrage 1,062 849.6 531 
Different infiltration programmesAnnual infiltration volume (10,000 m3)
Abundant yearNormal yearDry year
Failure to clean the riverbed 4.8 
Clearing the riverbed 797 637.6 398.5 
Not clearing the riverbed but excavating pit and pond 365 292 182.5 
Not clearing the riverbed but excavating pit and barrage 620 496 310 
Clearing the riverbed and excavating barrage 1,062 849.6 531 
Figure 12

Comparison of infiltration volumes by infiltration scheme.

Figure 12

Comparison of infiltration volumes by infiltration scheme.

Close modal

From the comparison of the simulation results, it is concluded that without the limitation of underground water storage space and without considering the economic and construction factors, the single flood infiltration volume of Scenario 5 is the highest, with a volume of 10,000 m3. While the single flood infiltration volume of Scenario 1 is the lowest, with a volume of 6,000 m3. Compared with the initial situation, the infiltration volumes of the four designed scenarios can achieve the purpose of enhancing the use of rainwater resources. The optimal infiltration scheme should be determined by combining various factors and local hydrogeological conditions.

COMSOL simulation

COMSOL software was used to simulate the dynamic process of infiltration of rainwater resources into the aquifer for 24 h. Since the simulation area can be divided into six zones based on permeability coefficients, for clear and direct observation of the infiltration process, this paper takes Zone 2 as an example in the COMSOL modelling, generalizing it from the river channel surface to the aquifer as a rectangular soil column ranging from 0 to −5 m. Following the concept of ‘air cells’ proposed by Jing & Shenghong (2003), Xiaoping et al. (2015), etc., a thin layer of ‘air cells’ was arranged on the upper boundary of the fine sand layer. When rainwater enters the ‘air cell’ area, its large permeability coefficient and absence from the actual seepage area do not affect the distribution of infiltration in the aquifer. This example uses the ‘Darcy's Law’ physical field node in COMSOL to observe the changes under different infiltration scenarios by adjusting the infiltration rate and porosity. After constructing the physical model of the study area, the Darcy's Law physical field is used on the basis of the air cells unit as an aid in order to make the model close to the real conditions, and the mass flow rate is controlled to achieve the simulation of different rainfall periods by adding the boundary condition of mass flow. Mass flow boundary conditions control the mass flow rate to achieve the simulation of different rainfall periods.

  • (1) Dynamic processes of rainfall infiltration under natural conditions

In Scheme 1, due to the limited infiltration capacity of fine sand, at the initial stage of rainfall, the air cell is set as a fixed flow boundary. As rainfall continues and exceeds the infiltration capacity, the pressure in the ‘air cell’ becomes positive, and its surface is set as a fixed pressure boundary (input pressure value is 0), simulating the overflow of excess water along the surface. The simulation result is shown in Figure 13.
Figure 13

Map of simulation results for scheme 1.

Figure 13

Map of simulation results for scheme 1.

Close modal

The results presented earlier demonstrate that in Scenario 1, the absence of detention facilities results in a complete absence of rainwater infiltration into the channel during the early stages of rainfall. Consequently, the total loss of rainwater during the early stages of rainfall is inevitable. During the mid to late rainfall period, when the volume of water is increasing, some rainwater is able to infiltrate into the channel. Nevertheless, the quantity of infiltration is considerably less than the total volume of flooding during the rainfall period.

  • (2) Dynamic processes of rainfall infiltration under cleaned riverbed conditions

Scheme 2, which involves cleaning the riverbed, effectively increases the infiltration capacity of the fine sand layer. The ‘air cell’ is set as a fixed flow boundary, with the infiltration rate equal to the rainfall intensity. As rainfall exceeds the infiltration rate, the fixed flow boundary of the air cell is converted to a fixed pressure boundary. The simulation result is shown in Figure 14.
Figure 14

Map of simulation results for scheme 2.

Figure 14

Map of simulation results for scheme 2.

Close modal

The results demonstrate that this scheme exhibits a superior infiltration effect during the initial stages of rainfall. As the rainfall persists, the infiltration effect becomes increasingly evident in the middle and late stages. The results demonstrate that the cleaning of riverbeds to enhance the infiltration capacity of the fine sand layer has a beneficial impact on rainwater infiltration.

  • (3) Dynamic processes of rainwater infiltration without clearing the riverbed + construction of pits and ponds

Scheme 3, involving the construction of pits, ponds, and river dams, increases the retention time of rainwater, reducing surface runoff and thereby increasing infiltration. In the model, this is achieved by adding surface flow in the air cell and extending the time for the fixed flow boundary in the air cell settings. Other parameters are the same as in Scheme 1. The simulation results are shown in Figure 15.
Figure 15

Map of simulation results for scheme 3.

Figure 15

Map of simulation results for scheme 3.

Close modal

The results indicate that although the scheme increases the retention time of rainwater by constructing pits and ponds, there is still limited infiltration of rainwater into the sloughs during the early stages of rainfall. As the rainfall persists, there is a greater degree of infiltration observed in the middle and late stages of the precipitation event. This option is inferior to option 2, as it fails to effectively infiltrate rainfall in the early stages.

  • (4) Dynamic process of rainwater infiltration without cleaning the river bed + construction of barrage + pit conditions

Scheme 4, involving the construction of pits, ponds, and river dams, increases the retention time of rainwater, reducing surface runoff and thereby increasing infiltration. In the model, this is achieved by adding surface flow in the air cell and extending the time for the fixed flow boundary in the air cell settings. Other parameters are the same as in Scheme 1. The simulation results are shown in Figure 16.
Figure 16

Map of simulation results for scheme 4.

Figure 16

Map of simulation results for scheme 4.

Close modal

This option incorporates stormwater retention facilities in comparison to Option 3, which increases the amount of stormwater retained for a more adequate retention time. As evidenced by the preceding results, there is already a notable infiltration effect in the initial stages of rainfall. Furthermore, the infiltration effect is even more pronounced in the middle and late stages of rainfall. This indicates that the scheme is capable of achieving the objective of collecting rainwater resources in an adequate manner.

  • (5) Dynamic processes of rainwater infiltration under conditions of riverbed cleaning + construction of barrages

The parameter settings for Scheme 5 reference the methods used in Schemes 2, 3, and 4, with the simulation result shown in Figure 17.
Figure 17

Map of simulation results for scheme 5.

Figure 17

Map of simulation results for scheme 5.

Close modal

This scheme enhances both the infiltration capacity of the fine sand layer and the retention time of the rainwater. The results demonstrate that the infiltration capacity for the entire rainfall period under this scheme is superior to that of all other schemes. Furthermore, the infiltration results are more favourable than those observed under the other schemes.

  • (6) Comparison of simulation results for different scenarios

In the model, as rainfall progresses, rainwater continuously infiltrates into the aquifer, which is reflected in the model as a change in the pressure of rainwater on the aquifer over time, and a comparison of the simulation results of the different scenarios reveals that the best infiltration effect is scenario 5, followed by scenarios 2, 4, and 3, and the worst infiltration effect is scenario 1. From the simulation results of the different infiltration schemes, it can be seen that Scenarios 2, 4, and 5 have better infiltration effects at the beginning and middle of the rainfall period. So that the purpose of increasing the infiltration of rainwater into the storage can be well achieved in the case of continuous heavy rainfall, which also shows that the applicability of these three scenarios is very strong, and can be applied to the special circumstances of heavy rainfall and very heavy rainfall.

The study area is an arid region with concentrated rainfall during the year and a paucity of rainwater harvesting facilities. The harvesting of rainwater resources using river tanks is a crucial aspect of water management in arid areas of the country and in cities with limited rainwater harvesting facilities. In order to collect and utilise rainwater in order to reduce the pressure on water supply and demand, a number of rainwater harvesting facilities have been constructed, including rain gardens, underground reservoirs, LID facilities, and so forth. Nevertheless, this is insufficient to address the challenge of water supply and demand. Therefore, there is a need to develop more effective rainwater harvesting facilities to enhance the utilization of rainwater resources. In this study, the maximum infiltration of rainwater for different scenarios was predicted using MODFLOW simulation, and the effect of rainwater infiltration under different scenarios was simulated using COMSOL. It was demonstrated that modifying the permeability of the aquifer and prolonging the retention period of rainwater can effectively enhance the infiltration volume of rainwater into the river channel. The results of the study can be summarized as follows:

  • (1) The results of the MODFLOW simulation indicate that the maximum infiltration volumes for the five scenarios are 0.6, 79.7, 36.5, 62, and 10,000 m3. Among these, the infiltration volume of rainwater under natural conditions is smaller, which is consistent with the actual situation. It can be seen that Options 2, 4, and 5 are capable of achieving the objective of enhancing rainwater harvesting.

  • (2) In the COMSOL numerical simulation results, the rainwater infiltration of all infiltration scenarios at the beginning of the rainfall is not satisfactory. As the rainfall progresses, the amount of rainwater gradually increases, and the rainwater gradually infiltrates into the interior of the aquifer, leading to an increase in the pressure of the aquifer. Among them, Scenarios 2, 4, and 5 showed better collection effects in the simulation results, which basically can achieve the effective collection of rainwater resources at the beginning, middle and end stages of rainfall, which is of great help to the collection of rainwater resources in the study area.

  • (3) Combined with the results of MODFLOW and COMSOL simulation, schemes 2, 4, and 5 are the three schemes with more infiltration volume and better infiltration situation. When implementing the scheme, it is necessary to consider the construction difficulty, maintenance difficulty, economic investment, and ecological effect. After comprehensive analysis, scheme 4 is finally identified as the best infiltration scheme. The identification of this programme provides effective and reasonable measures for the resolution of the issue of water resources utilization in arid areas and serves as a suitable reference programme for the river channel rainwater dark storage project. The programme is designed to collect and utilise rainwater resources, thereby alleviating the problem of insufficient water resources. Furthermore, it encourages the construction and protection of the ecological environment.

There are some limitations to the research in this paper. Mainly the lack of sampling and testing of water quality in the study area to determine whether the flood water quality needs to be treated before it infiltrates back into the ground, and attention will be paid to the factors not taken into account in future research.

The authors express gratitude to their colleagues for their assistance with the data used in this study. Additionally, Professor Zhang is acknowledged for their evaluation of the article.

This study is financially supported by the National Natural Science Foundation of China (Granted NO.41702264 and NO.42174177). Hebei Key Laboratory of Resource and Environmental Disaster Mechanism and Risk Monitoring (Grant No.FZ248107).

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.

Chenzhe
L.
,
Tengfei
F.
,
Yushan
F.
,
Zhenyan
W.
,
Bin
L.
,
Chen
Q.
,
Guangquan
C.
,
Xingyong
X.
&
Hongjun
Y.
2022
Origin and implications of pollution in coastal groundwater of the Guangdong Province
.
Journal of Marine Science and Engineering
10
(
10
),
1394
1394
.
Guochun
S.
2023
Feasibility study on resource utilization of rainwater and flood storage in river channels in arid regions of western Liaoning province
.
Water and Water Technology
S1
,
165
169
.
Harris
J. M.
&
Kennedy
S.
1999
Carrying capacity in agriculture: Globe and regional issue
.
Ecological Economics
129
(
3
),
443
461
.
Huang
C. F.
&
Zhou
J.
2013
Strategic research on urban stormwater resource utilization in the resource water shortage cities – base on stormwater data of 58 resource water shortage cities
.
Advanced Materials Research
2577
(
777–777
),
380
385
.
Jing
H.
&
Shenghong
C.
2003
Air element method for modeling drainage holes in seepage analysis
.
Rock and Soil Mechanics
02
,
281
283 + 287
.
Kumar
M. S.
,
Singh
M. K.
&
Susmita
G.
2022
Groundwater flow modeling for Cachar, India using MODFLOW: A case study
.
ISH Journal of Hydraulic Engineering
28
(
2
),
232
242
.
Kunyang
W.
,
Shin-ichi
O.
,
Mitsuyo
S.
&
Yuta
S.
2021
Long-term variations in water balance by increase in percent imperviousness of urban regions
.
Journal of Hydrology
602
,
126767
.
Lingyue
L.
&
Hong
G.
2000
Remark on the relationship between water resources rational allocation, Carrying capacity and sustainable development
.
Advances in Water Science
11
(
03
),
307
313
.
Lulu
L.
2019
Study on Urban Rainwater Resources Utilization and Hydrological Ecological Effects in Loess Area
.
Chang An University, Xi'An, China
.
Qixin
W.
2023
Research on the utilization of rainwater resources
.
Water Conservancy Planning and Design
236
(
06
),
17
20 + 71
.
Shiguo
X.
,
Jianwei
L.
&
Liyu
C.
2005
Analysis of the compensation effect of lake-reservoir-river system in floodwater utilization
.
Journal of Hydraulic Engineering
11
,
90
95
.
Song
L.
,
Xinsu
Z.
,
Yongjiu
F.
,
Huan
X.
,
Li
J.
&
Zhenkun
L.
2021
Spatiotemporal dynamics of urban green space influenced by rapid urbanization and land use policies in Shanghai
.
Forests
12
(
4
),
476
476
.
Song
Y.
,
Gao
M.
,
Xu
Z.
,
Wang
J.
&
Bi
M.
2023
Temporal and spatial characteristics of soil salinization and its impact on cultivated land productivity in the BOHAI Rim region
.
Water
15
(
13
), 2368.
Tong
X.
,
Zhiqiang
X.
,
Fei
Z.
,
Yimin
L.
,
Shouquan
Y.
,
Yangbin
Z.
,
Siqiao
Y.
,
Shi
C.
,
Xuan
L.
,
Sidong
Z.
&
Zhiqun
H.
2021
Permeability control and flood risk assessment of urban underlying surface: A case study of Runcheng south area, Kunming
.
Natural Hazards
111
(
1
),
661
686
.
Xiaoping
H.
,
Qing
X.
&
Shenghong
C.
2015
An air element method for overflow boundary problem in seepage analysis
.
Rock and Soil Mechanics
36
(
08
),
2345
2351
.
Xingchao
W.
2018
Application of underground reservoirs in construction of sponge cities
.
Advances in Science and Technology of Water
38
(
01
),
83
87
.
Xinghua
Z.
2007
Water System Health Circulation in Yanji
.
Beijing University of Technology Municipal Engineering, Beijing, China
.
Yanwei
S.
,
Xiaomei
W.
&
Ca
P.
2011
Review of current research and future directions of low impact development practices for storm water
.
Advances in Water Science
22
(
02
),
287
293
.
Yinghui
Z.
,
Huanbao
C.
,
Xiaoyang
W.
&
Min
D.
2013
Application of MIKE11-based hydrodynamic model in flood control and scheduling of Oujiang River mainstream gradient reservoirs
.
Zhejiang Hydrotechnics
41
(
01
),
48
49 + 53
.
Yongqiang
C.
,
Guangheng
N.
,
Heping
H.
&
Junxian
Y.
2005
Optimization dispatching theory and its application of floodwater utilization
.
Journal of Hydroelectric Engineering
05
,
17
21
.
Yunjie
L.
2017
Optimization and Simulating Evaluation of Rainwater Resources Utilization Measures for Dalian Main Urban Area
.
Dalian University of Technology, Dalian, China
.
Zhangjun
L.
,
Zhilei
Z.
,
Xingjun
H.
,
Xia
Z.
&
Rui
L.
2011
Ecological treatment of urban storm water runoff: Treatment and progress
.
Haihe Water Resources
03
,
39
44
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY-NC-ND 4.0), which permits copying and redistribution for non-commercial purposes with no derivatives, provided the original work is properly cited (http://creativecommons.org/licenses/by-nc-nd/4.0/).