Dynamic competitive removal mechanism for petroleum hydrocarbon under consideration of soil water during the thermal enhanced soil vapor extraction

This project focuses on designing an agrivoltaic system integrated with smart irrigation to promote sustainable agriculture and efficient resource management. The system combines solar photovoltaic (PV) panels with agricultural land to generate renewable energy while cultivating crops, thereby achieving dual land utilization. The solar panels provide electricity for powering irrigation pumps and IoT based sensors, which monitor soil moisture, weather data, and crop conditions in real time. A smart irrigation mechanism uses this data to automate water distribution, ensuring crops receive the right amount of water at the right time, minimizing waste and reducing water consumption. Additionally, the shading effect of solar panels reduces soil evaporation and heat stress on crops, leading to improved crop yield and resilience against climate variability. Surplus solar energy can be stored or sold back to the grid, creating an extra source of income for farmers. This project aims to demonstrate an energy efficient, water saving, and climate smart agricultural model that is scalable for both small and large farming operations, addressing the pressing challenges of food security, water scarcity, and sustainable energy production.

Abstract. Estimating latent heat fluxes in semi-arid environments remains challenging due to the strong spatial heterogeneity of soils and plants, land management practices, and limited observational data. In particular, accurately predicting the partition of evapotranspiration into evaporation and transpiration from observations remains very challenging. Land surface models (LSMs) can be used as a tool in this regard, when their validation is possible, but recent studies have indicated that LSMs generally overestimate soil evaporation. This study evaluates the performance of the land surface model ISBA within the SURFEX platform using data from two contrasting sites during the Land surface Interactions with the Atmosphere over the Iberian Semi-arid Environment (LIAISE) field experiment: an alfalfa field subjected to flood irrigation, and a natural grassland which is nearly senescent during the study period. It was found that the ISBA model tended to overestimate the evapotranspiration. Therefore, a dry surface layer (DSL) resistance was implemented in the ISBA model to improve the simulation of evaporation, which has proved successful in other models. The implementation of a DSL resistance led to an improvement in the simulated latent heat flux by reducing bare soil evaporation compared to simulations without a soil resistance. This approach reduced the daily RMSE of the latent heat flux by 29 % and 32 % at the alfalfa and natural grass sites respectively, while marginally increasing the correlation at both sites. Sensible heat flux and net radiation have improved on the order of 10 W m−2, whereas the ground heat flux has deteriorated within the same order. The resulting DSL simulations reduced the overall global error compared to a simulation without a DSL resistance. A sensitivity test of the parameters that drive a DSL resistance in ISBA further improved the simulations, reducing excessive diminution of LE after rain events. The new DSL parameterization helps overcome current problems of ET modeling by reducing bare soil evaporation within LSMs.

To overcome the limited adaptability of current greenhouse environmental models, this study develops targeted models for predicting the spatial distribution of near-surface soil temperature and air humidity. The soil temperature model utilizes a one-dimensional unsteady-state heat transfer approach that accounts for solar radiation, convection, thermal radiation, and evaporation. Concurrently, the air humidity model applies water vapor mass conservation to integrate the effects of soil evaporation, ventilation, and humidification.Experimental validation showed that the soil temperature model effectively captured diurnal variations at different soil depths. The relative root mean square error (rRMSE) was 5% at 0.25 m depth, demonstrating high accuracy, whereas predictions at the more dynamic surface layer (0.05 m) showed a higher error of 10%. The air humidity model produced trends consistent with measured data, albeit with a general underestimation, yielding a mean bias error (MBE) of 4.027 and an RMSE of 8.09%.The proposed models exhibit strong adaptability for simulating humidity and near-surface soil temperature distributions in common greenhouses. This work provides valuable insights and a practical methodological framework for advancing precise environmental control and promoting the development of standardized models in facility agriculture.

Accurately partitioning evapotranspiration (ET) into soil evaporation (E) and plant transpiration (T) is fundamental for improving water resource management, yet robust ET partitioning remains challenging. This study proposes a two-step ET partitioning strategy that first extracts pure E and T samples from long-term ET observations and then uses these samples to independently constrain E and T sub-models. The strategy was implemented in three classical two-source ET models: Shuttleworth–Wallace (SW), Priestley–Taylor Jet Propulsion Laboratory (PT-JPL), and FAO-56 dual crop coefficient (FAO56-DK), and was compared against the conventional one-step calibration approach. Results show that the two-step strategy consistently improves the estimation of ET components and the transpiration fraction (T/ET). For the PT-JPL model, RMSEs of E, T, and ET decreased from 0.04, 0.06, and 0.078 to 0.03, 0.03, and 0.04 mm/30 min, respectively. In FAO56-DK, R2 values increased from 0.08, 0.55, and 0.65 to 0.10, 0.65, and 0.75. The RMSE of T/ET declined from 0.21 to 0.18 in SW and from 0.47 to 0.34 in FAO56-DK. The effectiveness of pure samples depends on model structure, with E samples most beneficial for SW, T samples for FAO56-DK, and both for PT-JPL. Overall, these results demonstrate that pure-sample constraints substantially enhance ET partitioning accuracy.

Quantitative evaluation of different soil–water characteristic curve models on bare soil evaporation simulation

This study addresses the problems of severe deep percolation, low water use efficiency, and the difficulty of reconstructing the spatiotemporal soil moisture distribution in the 0-60 cm root zone under traditional empirical drip irrigation for greenhouse strawberries. An intelligent irrigation method is proposed by integrating real-time soil moisture data obtained from frequency domain reflectometry (FDR) sensors with the HYDRUS-2D mechanistic model. High spatiotemporal resolution FDR data are assimilated into the HYDRUS-2D framework to dynamically calibrate key soil-root parameters, enabling two-dimensional simulation and water balance analysis of drip irrigation infiltration, root water uptake, soil evaporation, and deep percolation processes. Compared with SIMDualKc and AquaCrop models, the proposed FDR-HY2D approach achieves higher accuracy in simulating soil moisture dynamics across different irrigation treatments and soil layers, with increased R2 and reduced RMSE and MAE. The model effectively reproduces daily moisture variations and vertical gradients in the root zone. Moreover, intelligent irrigation based on FDR-HY2D shifts water allocation from deep percolation to crop transpiration, significantly improving effective water use and water use efficiency. This study provides a digital decision-support tool for precision irrigation and water-fertilizer management in greenhouse strawberry production.

Abstract Soil salinity plays a critical role in regulating wetland evapotranspiration (ET)—including both soil evaporation and plant transpiration—by directly affecting water movement and indirectly altering plant physiology and ecosystem structure. However, common land surface models (LSMs) overlook salinity effects on ET, contributing to uncertainties in simulating land–atmosphere interactions and climate processes, particularly in the coastal regions. In this study, the Noah‐MP LSM was enhanced to incorporate the effects of soil salinity. Using half‐hourly meteorological forcing data and observed soil properties and salinity from the Dongtan coastal wetland in Shanghai (July–December 2023), we evaluated model performance before and after this improvement. Comparisons with in situ observations show that incorporating salinity significantly reduces the model's overestimation of latent heat flux—by up to 90 W/m 2 during the cold season—and decreases cumulative ET bias by as much as 130 mm, with the error rate reduced by approximately 33%. This improvement is largely attributed to salinity constraints on soil evaporation under sparse vegetation conditions. These findings highlight salinity as a key regulatory factor in wetland hydrothermal dynamics and offer a promising approach to improving LSM accuracy in coastal wetland environments.

Abstract Heatwave events significantly alter ecosystem water and energy balance and are often accompanied by extreme surface temperatures. Understanding how surface temperatures during such events are regulated by soil evaporation ( E ) and vegetation transpiration ( T ) remains limited due to challenges in partitioning total evapotranspiration ( ET ). Here, high‐frequency turbulence methods are used to partition observed ET at 32 National Ecological Observatory Network sites across the contiguous United States. Heatwaves were defined as at least three consecutive days with daily maximum air temperature exceeding the site‐specific 90th percentile of the 2019–2021 record. Across 268 identified events, the T/ET ratio decreased by 32% ± 16% relative to the non‐heatwave baseline of 0.65, with greater reductions at lower biomass sites. The T/ET ratio was typically suppressed below non‐heatwave conditions during the early and middle stages of the heatwave (first two‐thirds of event duration), but was on average higher than non‐heatwave baseline levels during late stages (final third) due to extremely low soil evaporation. Of the studied heatwaves, 71% of these had surface temperatures above 38°C in their late stage; however, heatwaves sustaining higher evaporation fluxes (upper tertile of observed fluxes) during the late stage were associated with relative surface temperature anomalies that were on average 45% lower than those of heatwaves with lower evaporation fluxes (lower tertile). The commensurate surface cooling induced by higher transpiration was only 2% during heatwaves, suggesting that transpiration has a limited ability to mitigate extreme surface temperatures. This study allows for improved prediction of ecosystem feedbacks under extreme thermal stress.

Abstract Tropical forests play a vital role in the global carbon cycle and land–atmosphere interactions. Estimating tropical forest carbon–water dynamics is challenging due to observational and modeling uncertainties. This study leverages the “Trends and drivers of the regional scale terrestrial sources and sinks of carbon dioxide” (TRENDY) project models and satellite observations to assess changes (2003–2021) in vegetation carbon, gross primary production (GPP), evapotranspiration (ET), and net biosphere production (NBP) in the Amazon and Congo. Atmospheric CO 2 , climate variability, and land use and land cover changes constrain these variables between 1700 and 2021 with the overall increasing trends of carbon stock and fluxes. The models overestimate vegetation carbon and GPP, while ET and NBP are consistent with observations. Fire‐activated models predict lower values for vegetation carbon and GPP, ET, and NBP, aligning more closely with observations. The higher ET from fire‐activated models may result from enhanced soil evaporation due to increased canopy openings. Fire‐inactivated models could well estimate the magnitudes of NBP. The high vegetation carbon in nitrogen‐enabled models points to simulation uncertainties and imbalance in model numbers regarding the nitrogen cycle. Although the nitrogen cycle enhances water use efficiency in both the Amazon and Congo, the models show a higher sensitivity to the nitrogen cycle in the Congo. This study highlights the challenges in accurately representing tropical biogeochemical cycles and the values of satellite products in model evaluations, underscoring the need for standard modeling protocols that address biogeochemical components (e.g., nutrient cycles) to better resolve process‐based representations.

Despite the promise of graph neural networks (GNNs) in hydrological forecasting, existing approaches face critical limitations in capturing dynamic spatiotemporal correlations and integrating physical interpretability. To bridge this gap, we propose a spatial-temporal graph neural network (ST-GNN) that addresses these challenges through three key innovations: dynamic graph construction for adaptive spatial correlation learning, a physically-informed feature enhancement layer for soil moisture and evaporation integration, and a hybrid Graph-LSTM module for synergistic spatiotemporal dependency modeling. The temporal and spatial modules of the spatio-temporal graph neural network exhibit a structural symmetry, which enhances the model’s representational capability. By integrating these components, the model effectively represents rainfall-runoff processes. Experimental results across four Chinese watersheds demonstrate ST-GNN’s superior performance, particularly in semi-arid regions where prediction accuracy shows significant improvement. Compared to the best-performing baseline model (ST-GCN), our ST-GNN achieved an average reduction in root mean square error (RMSE) of 6.5% and an average improvement in the coefficient of determination (R2) of 1.8% across 1–8 h forecast lead times. Notably, in the semi-arid Pingyao watershed, the improvements reached 13.3% in RMSE reduction and 2.5% in R2 enhancement. The model incorporates watershed physical characteristics through a feature fusion layer while employing an adaptive mechanism to capture spatiotemporal dependencies, enabling robust watershed-scale forecasting across diverse hydrological conditions.

Abstract. Understanding carbon flux dynamics in tropical ecosystems is crucial for evaluating their role in global climate regulation. This study investigates the temporal variability of the net ecosystem exchange (NEE) of CO2 and its interactions with key meteorological variables in a tropical forest ecosystem of the Pantanal, Brazil. Using high-resolution hourly data from a flux tower, we apply Detrended Fluctuation Analysis (DFA) and Detrended Cross-Correlation Analysis (DCCA) to analyze diurnal to seasonal cycles of NEE, latent heat (LE), sensible heat (H), global radiation (Rg), air and soil temperature (Tair and Tsoil), relative humidity (RH), and vapor pressure deficit (VPD). The results reveal a strong diurnal coupling between solar radiation, temperature, and carbon fluxes, with peak CO2 uptake occurring at midday. A key novel finding is a marked shift to strong anti-persistence in NEE at the weekly scale during the dry season, a pattern supported by concurrent reductions in LE and RH and increases in H and VPD. This highlights that water limitation is a critical driver of carbon release and reveals a previously unidentified regulatory mechanism in the ecosystem's carbon cycle. These findings underscore the sensitivity of carbon dynamics to hydrometeorological conditions and underline the necessity of multi-scale analysis. Uncertainties remain regarding the role of extreme droughts and floods, as well as land-use dynamics, which merit further investigation.

Climate change is increasingly constraining agricultural productivity, particularly for smallholder farmers in semi-arid regions. Rising demand for water and other agricultural inputs necessitates the use of process-based modeling tools to optimize agricultural practices and support water management. The limited application of the AquaCrop model to silage maize in the Souss-Massa region underscores the need for site-specific calibration to improve model reliability and optimize crop management practices. This study aims (i) to evaluate, for the first time, the ability of the AquaCrop model in simulating canopy cover (CC), total soil water content (SWC), and silage maize biomass in the Souss-Massa region, using data collected from 17 fields during the 2022–2024 growing seasons, and (ii) to study the effects of management practices such as mulching, shifting sowing dates, and irrigation management scenarios on silage maize yield and water productivity as climate change adaptation strategies. AquaCrop demonstrated high performance in estimating CC, SWC, and final above−ground biomass, with coefficients of determination (R 2 ) ranging from 0.93 to 0.98 and Nash-Sutcliffe Efficiency (NSE) above 0.94. Root Mean Square Error (RMSE) varied slightly, from 7.0-7.25% for CC, 5.71-7.56 mm for SWC, and 0.74-1.12 t ha -1 for biomass. Scenario analysis indicated that synthetic mulch reduced actual evapotranspiration (ET c act ) by 17% and improved water productivity by 35%. Advancing the sowing date by 40 days improved above−ground biomass by 8% and a 14% in transpiration−based productivity (WP Tr ). Irrigation triggered at 120% depletion of readily available water (RAW) reduced soil evaporation by 41%, improve ET−based water productivity by 14% and maintains 95 % of the reference yield compared to farmers’ irrigation practices. Application of a 75% ETc (crop evapotranspiration under standard conditions) deficit-irrigation strategy represents an optimal trade-off, reducing water use by 26%, maintaining 94% of biomass. These results confirm that the AquaCrop model is a valuable tool for designing management practices that enhance water conservation and productivity in semi-arid regions.

Runoff induced erosion and thereby sediment yield present significant economic and environmental challenge. However, there are very limited research available addressing surface runoff and sediment yield in Ajima catchment located in the Upper Blue Nile basin in Ethiopia. This study aims to model runoff, sediment yield and conservation measures in Ajima catchment using the SWAT hydrological model. The essential model parameters for surface runoff and sediment yield were identified using model sensitivity analysis. Monthly streamflow and sediment yield data for the period 1990-2013 obtained from MoWE were used to calibrate and validate the model. The model's performance was assessed using statistical measures such as the coefficient of determination (R2) and Nash-Sutcliffe Efficiency (NSE), percent bias (PBIAS) and Kling-Gupta Efficiency (KGE). Sediment yield reduction strategy was proposed in the model by changing topographic, physiographic, and land use characteristics to recommend soil and conservation practices suitable for the area which are established as scenarios.The results showed that the Soil Evaporation Compensation Factor (ESCO) for streamflow and the USLE management parameter (USLE-MP) for sediment yield were parameters most sensitive to the model. During the model's calibration and validation, the model showed significant correlations between computed and observed data for streamflow (R2 = 0.69 and 0.77; NSE = 0.76 and 0.86; PBIAS = 23.85 and - 11.84; KGE = 0.73 and 0.83) and sediment yield (R2 = 0.84 and 0.87; NSE = 0.76 and 0.79; PBIAS = 34.88 and 26.00; KGE = 0.55 and 0.71), respectively. The catchment's annual estimated average runoff is approximately 748mm, and average yearly sediment yield is 112.70 tons/ha. Terracing decreased sediment yield from 112.70 tons/ha to 48.10 tons/ha, and residue management decreased it to 107.17 tons/ha. Therefore, the most successful conservation technique is found to be terracing, which lowers sediment production by over 57%. This study provides valuable information that is helpful for ongoing efforts to regulate erosion and sedimentation in the study region. We suggest practices of implementing terracing for areas highly prone to erosion. In regions with moderate to low erosion risk, techniques such as contour farming, strip cropping, and residue management are recommended as cost-effective strategies to reduce sediment production and soil degradation.

BackgroundRising temperatures and altered precipitation patterns are expected to have profound impacts on the composition and condition of boreal forests. As a result there are growing needs for climate adaptation strategies in boreal forest management; one potential solution to achieve these goals is the utilization of nature-based climate-informed adaption solutions including afforestation using deciduous species which can help offset carbon emissions and sequester carbon at an increased rate. Deciduous afforestation has the potential to allow mangers to adapt fire-risk, while increasing carbon storage. Here, we investigated the impact of deciduous compared to coniferous afforestation on biomass accumulation in the Canadian boreal using a process-based model (3-PG). 3-PG utilises physiological principals to predict the growth of individual species across a variety of climate scenarios. This approach is valuable for projecting forest growth under changing climate, as it can account for plant responses to environmental factors which may not be captured by empirical models based on historical data. We simulated forest growth under three future climate scenarios to 2080, and compared the aboveground biomass (AGB, tons of Dry Matter per hectare; tDM ha−1) accumulated to baseline estimates using locally adapted coniferous species. In addition, we investigated the modelled effects of converting from conifer to deciduous species on stand level soil water and vapor pressure deficit responses to climate.ResultsWe found that deciduous simulations sequester more carbon under all climate scenarios, with the greatest difference occurring in the warmest scenario (171 tDM ha−1 for coniferous species compared to 347.1 tDM ha−1 for deciduous species). Coniferous species were generally more water stressed than deciduous species; conifers were generally 65.6% more stressed compared to deciduous species in August under the warmest climate scenario, while northern sites were less stressed than southern sites.ConclusionsSimulations such as these highlight the importance of modelling and consideration of different planting scenarios in decision-making to ensure successful resource allocation. They also demonstrate the potential of nature-based adaptation solutions projects, and the role deciduous afforestation can play in provision of habitat, modifying wildfire risk and northern boreal biomass and timber supply.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13021-025-00385-2.

Main conclusionIn the Negev, substantial vapor stems from the wet soil following rain events and therefore cannot be considered as dew but rather as distillation. Distillation provided ~ 35% and ~ 60% of the vapor-driven liquid for the cobbles and rock slabs, respectively, implying that lithobionts may benefit from vapor condensation also in non-dewy deserts.Lithic chlorolichens (lichens with green algae as photobionts) and cyanobacteria cover almost all rock surfaces in the Negev Highlands, where chlorolichens are believed to mainly benefit from non-rainfall water (NRW), i.e., dew and vapor at high relative humidity. Since chlorolichens may also inhabit non-dewy deserts and vapor may also stem from the wet soil (which once condenses is termed distillation), we hypothesized that vapor that stems from the wet soil may also benefit lithic chlorolichens. To evaluate the potential amount accumulated on these rocky surfaces, whether by NRW or soil vapor plus distillation (jointly termed as indirect rain water, IRW), 3-year-long measurements were conducted in the Negev using cloths attached to a pair of rock slabs and a pair of cobbles. Taking 0.05 (reflecting vapor adsorption) and 0.1 mm (reflecting vapor condensation), which allows for net photosynthesis by chlorolichens and cyanobacteria, respectively, we found that: (1) the average number of days with NRW and IRW ≥ 0.05 mm was respectively 128.7 days and 28.0 days (for cobbles) and 37.3 days and 19.3 days (for rock slabs), with dew (which occurs along the year) and distillation (limited to days after rain events) occurring respectively for 36.7 days and 20.0 days (cobbles) and 28.0 days and 6.0 days (rock slabs), (2) average annual amounts of NRW and IRW ≥ 0.05 mm were respectively 11.5 mm and 3.9 mm (for cobbles) and 2.7 mm and 1.8 mm (for rock slabs), with dew and distillation being respectively 4.7 mm and 3.1 mm (for cobbles) and 0.5 mm and 0.9 mm (for rock slabs), (3) average annual daytime duration of > 0.05 mm for NRW and IRW were respectively 307.8 h and 83.9 h (for cobbles) and 81.0 h and 46.7 h (for rock slabs) with dew and distillation lasting respectively 103.8 h and 60.2 h (for cobbles) and 10.3 h and 17.6 h (for rock slabs). Given that daylight duration primarily dictates growth, we may conclude that: (1) cobbles receive substantially higher amounts of NRW and IRW than rock slabs, (2) the amount of distillation received on cobbles (3.1 mm) was not substantially lower than that of dew (4.7 mm). As far as the annual daylight wetness duration for cobble-dwelling lichens is concerned, distillation provided 36.7% of the total duration provided by vapor. Since IRW may occur also in dewless deserts, such as the Mojave Desert, it may partially explain lithic lichen inhabitation in the Mojave and other non-dewy deserts.

California’s food and economic security depends on water availability, particularly under increasingly extreme climate scenarios. A key component of the water balance is evapotranspiration, the combination of soil and surface evaporation and plant transpiration. Evapotranspiration is influenced by natural drivers (e.g., climate, vegetation cover) and human intervention (e.g., irrigation, land management). Here, we analyze the transition between one of California’s driest years (2022) to an exceptionally wet year (2023) to assess evapotranspiration responses to climate extremes. Despite increased precipitation, total statewide evapotranspiration changed less than 10%. In 2022, human contributions accounted for 30% of statewide evapotranspiration and 80% in managed lands. In 2023, natural evapotranspiration increased, and human contributions fell by 30%, yet still comprised nearly 50% of evapotranspiration in managed areas. Our findings underscore the enduring role of human activity on California’s hydrology, even during wet years, and demonstrate a framework to separate natural and anthropogenic controls on evapotranspiration.

A model for predicting soil evaporation rate with consideration of evaporation surface descent

Intercropping is an emerging potential nature-based solution for sustainable crop production in temperate environments. However, its long-term role in contributing to climate mitigation and adaptation remains unclear. This work presents the first evidence of potential long-term water and carbon effects of barley (Hordeum vulgare L.) and pea (Pisum sativum L.) intercropping versus its barley monoculture for a typical temperate environment in Scotland. Based on experimental data, water (HYDRUS 5) and soil carbon (RothC) models were coupled to project water-carbon dynamics for the short-term during a two-season field trial (2022-2023) and the long-term future (2024-2050) under a worst-case climate scenario (Representative Concentration Pathway, RCP 8.5). The coupled water-carbon model effectively captured the water-carbon dynamics observed in the short-term. Compared to barley monoculture, intercropping increased evapotranspiration up to ∼20% in the short-term, dominated by the dry weather conditions in 2022. Long-term intercropping projected lower interannual variability in evapotranspiration than barley monoculture, but showed higher plant transpiration in dry years, indicating more adaptive water use and hydrological resilience. As intercropping is projected to increase transpiration but reduce soil evaporation compared with barley monoculture, it maintained similar levels of soil water content and storage in the topsoil (0-30cm). In addition, by 2050, soil carbon was predicted to increase in the upper topsoil (0-5cm) of intercropping by 16% (1.91kgm-2) compared to barley monoculture (1.63kgm-2). These novel findings suggest that intercropping could play a critical role in enhancing hydrological resilience and carbon sequestration in temperate environments for sustainable land management.

ABSTRACT Crop‐stage‐specific deficit irrigation (DI) has been widely applied to achieve the optimum agricultural water use in dryland farming areas. However, the water use characteristics during different crop stages have not been fully investigated, considering ET uncertainties. It may hinder the correct decisions on optimum agricultural water management. This study investigated how the root‐zone water budget components varied throughout the growing season in a summer maize field under three different irrigation regimes by using a soil water model, STEMMUS‐ET, with the indirect and direct ET methods. Two successive years of crop‐stage‐specific DI experiments were conducted on a summer maize field in Northwest China to calibrate and evaluate the STEMMUS‐ET model. Results indicate that STEMMUS‐ET simulated the soil water contents, ET , soil evaporation, and root‐zone water budgets well for all irrigation treatments. The influence of using different ET methods on soil moisture content mainly affects shallow soil layers (0–30 cm). The soil evaporation simulation was largely improved by the direct ET method due to the consideration of aerodynamic and surface resistance terms, especially after irrigation. Different irrigation amount has a significant effect on the transpiration but not on the soil evaporation. It is the frequency rather than the amount of irrigation that largely affects soil evaporation. Compared to CK treatment, the DI treatments depleted more soil water storage with less use of irrigation water throughout the growing season. T1, with the reduced irrigation water amount properly at the same irrigation frequency, can significantly improve WUE, increasing it by 9.71% compared to CK. These insights help make reasonable water management in dryland agriculture.