Soil Moisture Dynamic Simulation Under Different Irrigation Patterns in Sandy Loam Farmland of the Middle Heihe River Basin

LI Haochen¹, HU Guanglu¹,², WANG Tao³, CHEN Ning³, LI Jianan⁴, FAN Yalun¹

¹School of Environmental and Municipal Engineering, Lanzhou Jiaotong University, Lanzhou 730070, Gansu, China
²Engineering Research Center of Water Resource Comprehensive Utilization in Cold and Arid Regions, Ministry of Education, Lanzhou 730070, Gansu, China
³Daman Water Conservancy and Electricity Management Office, Ganzhou District, Zhangye 734000, Gansu, China
⁴Gansu Institute of Soil and Water Conservation, Lanzhou 730020, Gansu, China

Abstract

Farmlands in arid regions are characterized by low soil fertility, poor water-holding capacity, substantial leakage, and low crop yields. Understanding soil moisture movement patterns in sandy loam farmland is crucial for water conservation and crop yield improvement. This study focused on newly reclaimed sandy loam maize farmland in the middle reaches of the Heihe River, establishing three experimental plots with different irrigation patterns: flat mulching irrigation, ridge mulching irrigation, and subsurface drip irrigation under plastic film. The HYDRUS-2D model was employed to simulate soil moisture migration processes under these different irrigation regimes. The results demonstrated that: (1) The simulated values from the HYDRUS-2D model showed strong agreement with measured data, with R² exceeding 0.864 and RMSE below 0.006 cm³ cm⁻³, validating the model's feasibility and reliability for simulating soil moisture dynamics in sandy loam farmland. (2) Compared with flat mulching irrigation, ridge mulching irrigation increased root zone soil volumetric water content by approximately 20% while reducing seepage losses by 13.3%, despite a 2099 m³ ha⁻¹ reduction in irrigation volume. (3) Subsurface drip irrigation under film reduced irrigation water consumption by 50% and leakage by 50.7% compared with flat mulching irrigation. (4) The "frequent irrigation with small amounts" characteristic of subsurface drip irrigation enabled more direct and efficient water replenishment to the crop root zone, substantially increasing root zone soil volumetric water content while further reducing leakage. Therefore, subsurface drip irrigation under film is recommended for sandy loam farmland in the middle Heihe River basin to achieve water savings and yield increases. The HYDRUS-2D model parameter system developed in this study can also serve as a reference for simulating irrigation water dynamics in similar sandy loam farmlands across northern China.

Keywords: sandy loam; soil volumetric water content; HYDRUS-2D model; irrigation optimization; maize farmland; middle reaches of the Heihe River

1.1 Study Area Overview

The study area is located in Pingchuan Town, Linze County, Zhangye City, Gansu Province, in the middle reaches of the Heihe River. The terrain is flat with an average elevation of 1374 m. The region exhibits typical temperate continental arid climate characteristics, with an average annual precipitation of 117 mm (concentrated in May-September, accounting for approximately 85% of annual rainfall) and annual evaporation reaching 2390 mm—about 20 times the precipitation. The average annual temperature is 7.6°C, with maximum and minimum temperatures of 39.1°C and -27.3°C, respectively. The annual sunshine duration is 3045 hours, and accumulated temperature ≥10°C is 3085°C d. The annual radiation total is 6.1×10⁹ J m⁻². Local farmland types consist of two categories: long-cultivated old farmland along the Heihe River banks and newly reclaimed farmland developed from desert areas since the 1990s. This study focused on newly reclaimed sandy loam farmland located in the oasis-desert transition zone in northern Linze County (100°07′E, 39°21′N), near the Gansu Linze Farmland Ecosystem National Field Scientific Observation and Research Station. Daily precipitation and evaporation data from May-September 2023 are shown in [FIGURE:1].

1.2 Experimental Design

In May 2023, irrigation experimental plots were established in newly reclaimed sandy loam farmland (reclaimed in 2020). Three adjacent 25 m × 25 m plots were designated for flat mulching irrigation, ridge mulching irrigation, and subsurface drip irrigation under film (referred to as flat mulching, ridge mulching, and subsurface drip irrigation, respectively). All plots were planted with seed maize. The flat mulching plot employed flood irrigation, with each irrigation event designed at 8995.5 m³ ha⁻¹ (based on local long-term averages for seed maize), approximately 15-day intervals, totaling 6 irrigation events and 375.36 mm of water. The ridge mulching plot followed the same irrigation timing but with total irrigation reduced to 6896.5 m³ ha⁻¹ according to the Gansu Provincial Industry Water Quota (2023 edition). For subsurface drip irrigation, drip lines were placed 10 cm below the surface (based on soil conditions and maize root development), with 5 drip emitters per plot at flow rates of 5, 10, and 15 cm·d⁻¹, totaling 4497.8 m³ ha⁻¹ across 15 irrigation events.

Following the first irrigation in late May, three sampling points were selected along diagonal transects in each plot. Soil samples were collected before each irrigation/fertilization event from 0-20 cm, 20-40 cm, 40-60 cm, 60-80 cm, and 80-100 cm depths using core samplers and augers to determine soil volumetric water content, residual water content, saturated water content, saturated hydraulic conductivity, particle composition, and bulk density. Flat mulching samples were collected from both bare and mulched areas, while ridge mulching and subsurface drip irrigation samples were taken from mulched zones. All samples were analyzed following standard agrochemical methods. Post-harvest yields were 513 kg ha⁻¹ for flat mulching, 570 kg ha⁻¹ for ridge mulching, and 535 kg ha⁻¹ for subsurface drip irrigation.

1.3.1 Soil Water Characteristic Parameters

Soil water characteristic parameters were predicted using the Rosetta software based on bulk density and clay, silt, and sand contents. To improve simulation accuracy for the study area, measured soil water content data were compared with Rosetta predictions to obtain discrepancy data. Using HYDRUS-2D's parameter inversion function, key parameters including residual water content (θᵣ), saturated water content (θₛ), and saturated hydraulic conductivity (Kₛ) were iteratively optimized until simulated soil moisture distributions closely matched observed values. Soil particle size followed USDA standards: clay (<2 μm), silt (2-50 μm), and sand (>50 μm). Adjusted parameters are shown in [TABLE:2].

Soil water movement was described using the Richards equation, with soil hydraulic properties characterized by the van Genuchten-Mualem model. Root water uptake was simulated using the Feddes model with parameters referenced from previous studies in the region [FIGURE:3].

1.3.2 Free Drainage Boundary Leakage Model

The HYDRUS series models reliably assess deep leakage losses. In this study, leakage was calculated using HYDRUS-2D's boundary water flux module. With left and right boundaries set as zero-flux boundaries, leakage occurred only at the lower boundary. The free drainage boundary flux (velocity) data were exported from HYDRUS-2D and time-integrated to obtain total leakage, expressed in mm d⁻¹.

1.4 Model Construction

The model domain was defined as follows: For flat mulching, a rectangular region of 100 cm (radial) × 100 cm (vertical) with a 60 cm mulched zone centered on the upper boundary. Ridge mulching used a 100 cm × 110 cm domain with 10 cm ridge height and 50 cm mulched ridge top. Subsurface drip irrigation used the same dimensions as flat mulching but with circular water application sections representing emitters at 10 cm depth.

Upper boundary conditions: Bare areas were set as atmospheric boundaries (green lines in figures); mulched areas were zero-flux boundaries due to film impermeability. Vertical boundaries were zero-flux (white lines) assuming negligible horizontal water movement, given the high sand content and dominant vertical infiltration. The lower boundary was set as free drainage (olive green lines) since groundwater depth exceeds 10 m and soils are highly permeable. Observation nodes were placed at 10, 30, 50, 70, and 90 cm depths (red points) to monitor and simulate soil volumetric water content dynamics during the growing season. Simulation time was discretized in days from day 0 to 122, with initial time step of 0.001 d, minimum step of 0.00001 d, and maximum step of 1 d.

1.5 Model Validation

Flat mulching data were used for parameter calibration and model accuracy verification. The calibrated HYDRUS-2D model was then applied to simulate soil moisture dynamics under ridge mulching and subsurface drip irrigation, with results compared against measured data from root zones (0-40 cm). Model performance was evaluated using root mean square error (RMSE), mean relative error (MRE), and coefficient of determination (R²). According to literature [15,28-29], R² > 0.8, RMSE < 0.01 cm³ cm⁻³, and MRE < 10% indicate satisfactory simulation results.

2.1.1 Analysis of Measured Soil Volumetric Water Content

Under flat mulching, average soil volumetric water content across all layers ranged 0.096-0.181 cm³ cm⁻³, with minimum values in 80-100 cm depth (0.096 cm³ cm⁻³) and maximum in bare zone surface layers (0.181 cm³ cm⁻³). This reflects the sandy loam texture with low clay content in deeper layers, resulting in poor water-holding capacity that decreases with depth. The largest variance occurred in mulched 80-100 cm layer (0.003), indicating high variability due to rapid infiltration. All mulched zone measurements showed weak variation (CV < 10%), while bare zone 20-40 cm layer showed moderate variation (CV = 11.3%).

2.1.2 Model Accuracy Verification

Comparison of simulated and measured soil volumetric water content under flat mulching across 0-100 cm depth showed good agreement, with points evenly distributed around the 1:1 line [FIGURE:5]. R² values were 0.864 for mulched zones and 0.921 for bare zones, with RMSE of 0.005 cm³ cm⁻³ and 0.006 cm³ cm⁻³, and MRE of 4.06% and 5.32%, respectively. These results confirm that the optimized HYDRUS-2D model provides reliable simulations for soil moisture dynamics in newly reclaimed sandy loam farmland in the middle Heihe River basin.

2.2 Ridge Mulching Mode

The calibrated HYDRUS-2D model achieved R² = 0.864 for ridge mulching, confirming its applicability. Ridge mulching significantly increased 0-40 cm soil volumetric water content in mulched areas compared with flat mulching [FIGURE:8-9]. The average water content followed the order: 0-20 cm > 80 cm > 40-60 cm > 20-40 cm > 80-100 cm. The 0-20 cm layer near the ridge center (root-dense zone) showed highest water content, demonstrating that ridge mulching effectively improved shallow root zone moisture. Although irrigation methods and timing remained unchanged from flat mulching, ridge mulching enhanced water productivity and achieved water savings. However, similar to flat mulching, the plastic film prevented direct water replenishment beneath the mulched area, limiting moisture in the central root zone.

2.3 Subsurface Drip Irrigation Mode

For subsurface drip irrigation, the model achieved R² = 0.864, demonstrating good applicability. This irrigation method directly supplements moisture to the root-dense zone. At 5 cm·d⁻¹ emitter flow, water primarily affected the 0-40 cm layer; at 10 cm·d⁻¹, water reached 40-60 cm; and at 15 cm·d⁻¹, water penetrated to 60-80 cm [FIGURE:10-13]. The 0-20 cm and 20-40 cm layers showed the most significant moisture changes, with higher water content maintained longer after irrigation compared with ridge mulching. Although single irrigation amounts were much smaller than ridge mulching, the direct root zone application resulted in higher 0-40 cm water content and lower 40-60 cm content, effectively controlling water within the root zone and reducing deep percolation.

2.4 Comparison of Simulated Leakage Under Different Irrigation Modes

To ensure water supply and yield stability, local farmers traditionally use flood irrigation, causing deep percolation when soil moisture exceeds saturation. Simulated leakage comparison revealed the following order during the growing season: flat mulching > ridge mulching > subsurface drip irrigation [FIGURE:14]. Ridge mulching reduced maximum leakage by 16.5% and total leakage by 13.3% (from 375.36 mm to 325.47 mm) compared with flat mulching. Subsurface drip irrigation reduced maximum leakage by 55.5% and total leakage by 50.7% (to 184.87 mm) compared with flat mulching, and by 43.3% compared with ridge mulching. The "small amount, frequent irrigation" approach of subsurface drip irrigation substantially reduced water loss.

3.1 Comparison of Different Irrigation Modes

Both ridge mulching and subsurface drip irrigation achieved water-saving and yield-increasing effects compared with flat mulching. Ridge mulching reduced water use by 23.3% while increasing yield by 11.1%; subsurface drip irrigation reduced water use by 50% while increasing yield by 4.3%. The model quantified spatiotemporal water distribution patterns: ridge mulching increased root zone moisture through lateral infiltration but still exhibited significant deep leakage, consistent with high hydraulic conductivity of sandy soils [TABLE:2]. Subsurface drip irrigation, through its "small water, frequent irrigation" strategy, precisely controlled water within the root zone (0-40 cm), reducing deep leakage to only 49.3% of flat mulching and 56.7% of ridge mulching. This makes subsurface drip irrigation ideal for sandy loam farmland in the middle Heihe River basin.

Ridge mulching improves soil aeration and root development, enhancing water and nutrient uptake efficiency. The ridge furrow collects water, increasing hydraulic head during irrigation and promoting lateral water movement to the root zone. However, strong infiltration in sandy soils means large single irrigation amounts still cause substantial deep percolation before uniform lateral distribution can occur. Subsurface drip irrigation addresses this by delivering water directly to the root zone at high frequency, mitigating drought stress caused by severe water percolation in sandy soils and maintaining optimal moisture levels for crop growth.

3.2 Model Guidance for Similar Soil Farmland

The HYDRUS-2D model demonstrates strong regional adaptability for simulating water, heat, and solute transport in unsaturated porous media. This study validated the model using measured soil moisture data from sandy loam farmland, achieving R² = 0.864-0.921 and RMSE < 0.006 cm³ cm⁻³ across different irrigation modes. The model accurately captured rapid water percolation characteristics of sandy soils and is particularly suitable for analyzing deep leakage under various irrigation patterns. Similar studies have successfully applied HYDRUS-2D in the Hetao Irrigation District and other arid regions of northern China. The parameter system developed here can be extended to similar sandy loam farmlands, providing scientific guidance for irrigation management departments to optimize irrigation amounts and frequencies, achieve precise root zone moisture control, and promote transformation toward "high-efficiency water saving and stable high yield."

3.3 Outlook

Crop evapotranspiration varies with growth stage, meteorological conditions, soil texture, and irrigation regime, affecting model parameters. While this study simulated soil moisture dynamics under different irrigation modes, it did not deeply explore precipitation effects on evapotranspiration or differences in root growth and water uptake characteristics among irrigation modes. Future research should incorporate these factors to adjust model parameters and achieve true optimization of water use efficiency and crop yield. Additionally, integrating HYDRUS-2D with remote sensing data could enhance regional-scale simulation capabilities for water productivity distribution patterns in arid oasis farmlands.

4 Conclusions

Based on field experiments and model calibration, this study analyzed spatiotemporal variation characteristics of soil volumetric water content under different irrigation modes in sandy loam farmland of the middle Heihe River basin. The main conclusions are:

1) The calibrated HYDRUS-2D model showed good agreement with field measurements (R² = 0.864-0.921, RMSE < 0.006 cm³ cm⁻³), demonstrating reliable performance for simulating soil moisture dynamics in this region.

2) Ridge mulching increased root zone soil volumetric water content by ~20% and reduced leakage by 13.3% compared with flat mulching, despite 2099 m³ ha⁻¹ less irrigation water. Subsurface drip irrigation reduced irrigation water by 50% and leakage by 50.7%.

3) The "frequent irrigation with small amounts" approach of subsurface drip irrigation directly replenished the root zone, increasing root zone water content by 16.5% while reducing leakage by over 43% compared with ridge mulching. This makes it the recommended irrigation mode for local sandy loam farmland.

4) The HYDRUS-2D parameter system developed in this study has broad applicability for similar sandy loam farmlands in arid regions, providing a scientific basis for optimizing irrigation management and promoting water-saving, high-yield agriculture.

References

[1] Hu Guanglu. Research on Distribution Pattern and Influential Factors of Water Productivity in Oasis of Arid Areas. Gansu People's Publishing House, 2011.

[2] Chen Haizhi, Hu Guanglu, Li Jianan, et al. Climate change and its impact on maize yield in oasis in the middle reaches of the Heihe River from 1967 to 2020—A case study of Linze County. Journal of Gansu Agricultural University, 2023, 58(2): 137-144, 154.

[3] Yang Na, Mao Xiaohan, Li Yan, et al. Meta-analysis of soil organic carbon and its active fractions in response to straw and film mulching. Environmental Science, 2025, 46(4): 2292-2300.

[4] Zheng Haofei, Pang Lei, Sun Jianhao, et al. Responses of productivity to film mulching patterns and chemical fertilizer application rates in maize||pea intercropping system. Chinese Journal of Eco-Agriculture, 2024, 32(10): 1639-1649.

[5] Qiao Yibo, Wu Pengnian, Wang Yanli, et al. Simulation analysis on water leakage and nitrogen leaching of winter wheat farmland under micro-sprinkler irrigation in Huang-Huai-Hai Plain. Journal of Agro-Environment Science, 2023, 42(7): 1540-1553.

[6] An Ning, Guo Bin, Zhang Dongmei, et al. Desert vegetation composition and spatial distribution of soil nutrients in the middle section of Hexi Corridor. Arid Zone Research, 2024, 41(3): 432-443.

[7] Li Jianan, Zhou Chengqian, Hu Guanglu, et al. Characteristics and influencing factors of soil macropores in the root zone of sand-fixing plants in the desert-oasis transition zone. Arid Zone Research, 2024, 41(12): 2015-2026.

[8] Chen Yansi, Huang Chunlin, Hou Jinliang, et al. Extraction of maize planting area based on multi-temporal Sentinel-2 imagery in the middle reaches of Heihe River. Remote Sensing Technology and Application, 2021, 36(2): 324-331.

[9] Hu Guanglu, Tao Hu, Jiao Jiao, et al. Runoff trend and attribution analysis of the Zhengyi Gorge in the middle reaches of the Heihe River. Arid Zone Research, 2023, 40(9): 1414-1424.

[10] You Debao, Wang Jianlin, Lv Mingqiang, et al. Evapotranspiration of maize field in irrigation area in Heihe middle reaches using the Penman-Monteith method. Acta Agriculturae Boreali-Sinica, 2015, 30(S1): 139-145.

[11] Jia Angyuan, Zhang Yongyong, Zhao Wenzhi, et al. Characteristics of soil macropores and water infiltration in oasis croplands in an arid region. Acta Pedologica Sinica, 2022, 59(2): 486-497.

[12] Ma Bin. Research on Remote Sensing Estimation Model of Water Use Efficiency of Oasis Farmland Shelterbelt System. Shihezi University, 2022.

[13] Lu Rukun. Methods for Agrochemical Analysis of Soils. China Agricultural Science and Technology Press, 2000.

[14] Gao Zhenguo, Zhong Ruilin, Yang Shuai, et al. Recent progresses in research and applications of Hydrus model in China. Soils, 2022, 54(2): 219-231.

[15] Wu Yuanzhi, Huang Mingbin. Analysis of influential factors for maize root water uptake based on Hydrus-1D model. Transactions of the Chinese Society of Agricultural Engineering, 2011, 27(S2): 66-73.

[16] van Genuchten M T. A closed-form equation for predicting the hydraulic conductivity of unsaturated soils. Soil Science Society of America Journal, 1980, 44(5): 892-898.

[17] Li Dongsheng, Ji Xibin, Zhao Liwen, et al. Simulation of seed corn farmland soil moisture migration regularity in the midstream of the Heihe River Basin. Arid Zone Research, 2015, 32(3): 467-475.

[18] Sun Xuemei, Zhang Yimin, Zhang Zhiwei, et al. Simulation of nitrogen transport and transformation in dry crop peanut land profiles in a red soil zone. Journal of Agro-Environment Science, 2024, 43(7): 1557-1567.

[19] Han M, Zhao C, Feng G, et al. Evaluating the effects of mulch and irrigation amount on soil water distribution and root zone water balance using HYDRUS-2D. Water, 2015, 7(6): 2622-2640.

[20] Cheng Guo, Zhang Nan, Zhang Lei, et al. Optimization of drip irrigation scheme for apple with dwarf rootstock and dense planting in Southern Xinjiang with HYDRUS-2D. Water Saving Irrigation, 2024(9): 121-128.

[21] Guo Jiawei, Zhou Hui, Shi Haibin, et al. Analysis of dry discharge effect of salt wasteland based on HYDRUS-2D. Journal of Irrigation and Drainage, 2023, 42(S1): 206-212.

[22] Ding Yuntao, Cheng Yu, Zhang Tibin, et al. Modeling of dynamics of deep soil water and root uptake of maize with mulched drip irrigations using HYDRUS-2D. Agricultural Research in the Arid Areas, 2021, 39(3): 23-32.

[23] Xiao Qingli. Soil Water Dynamics and Hydrological Relation of Cropland and Shelterbelt in the Oasis Zone of the Heihe River. Northwest A & F University, 2016.

[24] Busscher W. Simulation of field water use and crop yield. Soil Science, 1980, 129(3): 193.

[25] Er Raki S, Ezzahar J, Merlin O, et al. Performance of the HYDRUS-1D model for water balance components assessment of irrigated winter wheat under different water managements in semi-arid region of Morocco. Agricultural Water Management, 2021, 244: 106065.

[26] Li C J, Xiong Y W, Cui Z, et al. Effect of irrigation and fertilization regimes on grain yield, water and nitrogen productivity of mulching cultivated maize (Zea mays L.) in the Hetao Irrigation District of China. Agricultural Water Management, 2020, 232: 106065.

[27] Yang Jiugang, He Jiwu, Ma Yingjie, et al. Effects of irrigation frequency and quota on cotton growth and yield with drip irrigation under plastic film. Water Saving Irrigation, 2011(3): 29-32, 38.

[28] Miao Qingyuan, Mi Lina, Qin Lanyu, et al. Developing optimal drip irrigation schedule for alfalfa production in desert using the HYDRUS-1D model. Journal of Irrigation and Drainage, 2024, 43(5): 8-15.

[29] Lei B, Saifeng Z, Xinyu L, et al. Intelligent drip fertigation increases water and nutrient use efficiency of watermelon in greenhouse without compromising the yield. Agricultural Water Management, 2023, 282: 108278.

[30] Peng X, Chen D, Zhou Z, et al. Optimizing drip fertigation management to simultaneously improve the yield, water productivity and fertilizer agronomic utilization efficiency of grapes in different precipitation year patterns. Agricultural Water Management, 2024, 295: 108749.

[31] Zhang Shuying, Zhang Jinzhu, Wang Zhenhua, et al. Optimization of water and nitrogen application system for cotton yield under mulched drip irrigation in Northern Xinjiang. Journal of Drainage and Irrigation Machinery Engineering, 2024, 42(6): 641-648.

[32] Wang Rui. Effects of Plastic Film Mulching on Water Transformation and Utilization Efficiency of Summer Maize. Northwest A & F University, 2022.

[33] Fu Yujuan, Li Yao, Yu Haoyang, et al. Effects of ridging and plastic film mulching on the water balance of summer maize during its growth period. Transactions of the Chinese Society of Agricultural Engineering, 2023, 39(10): 76-87.

[34] Ding Yuntao, Cheng Yu, Zhang Tibin, et al. Effects of the lower limit of drip irrigation on soil water and salt distribution and maize growth in sand-lined farmland. Acta Pedologica Sinica, 2022, 59(3): 733-744.

[35] Li Jiaxin, Huang Shuang, Jiang Zhen, et al. Study on the laws of soil water and salt transport in corn farmland with drip irrigation under shallowly buried film based on HYDRUS-2D. Water Saving Irrigation, 2025: 1-18.

[36] Song Zhenwei, Guo Jinrui, Deng Aixing, et al. Effects of surface tillage regimes on soil moisture and temperature of spring corn farmland in Northeast China. Transactions of the Chinese Society of Agricultural Engineering, 2012, 28(16): 108-114.

[37] Deng Huan, Dai Fei, Shi Ruijie, et al. Water and fertilizer transport law and root response of maize in full film double furrows based on HYDRUS-2D/3D. Transactions of the Chinese Society for Agricultural Machinery, 2022, 53(S2): 100-108.

[38] Wang Tianye, Wang Ping, Wu Zeyu, et al. Progress in the study of ecological resilience of vegetation under drought stress. Advances in Earth Science, 2023, 38(8): 790-801.

[39] Zhuang Haoran, Feng Kepeng, Xu Dehao. Changes, influencing factors and sensitivity of water use efficiency in maize farmland ecosystems based on evapotranspiration separation in the Ningxia irrigated area. Arid Zone Research, 2023, 40(7): 1117-1130.

[40] Shi Jianzhou, Liu Xiande, Tian Qing, et al. Rainfall response of soil water content on a slope of Larix principis-rupprechtii plantation in the semi-arid Liupan Mountains. Arid Zone Research, 2023, 40(4): 594-604.

[41] Ji Jiamen, Cheng Yiben, Chen Linglong, et al. Dynamic changes in soil moisture and its response to rainfall in Pinus sylvestris var. mongolica plantation in Horqin Sandy Land. Arid Zone Research, 2023, 40(5): 756-766.