Enhanced SWAP Model for Simulating Evapotranspiration and Cotton Growth Under Mulched Drip Irrigation in the Manas River Basin

Abstract

Model-based simulation of farmland evapotranspiration and crop growth facilitates precise monitoring of crop and farmland dynamics with high efficiency, real-time responsiveness, and continuity. However, there are still significant limitations in using crop models to simulate the dynamic process of evapotranspiration and cotton growth in mulched drip-irrigated cotton fields under different irrigation gradients. The SWAP crop growth model effectively simulates crop growth. However, the original SWAP model lacks a dedicated module to consider the impact of mulching on cotton field evapotranspiration and cotton dry matter mass. Therefore, in this study, the source codes of the soil moisture, evapotranspiration, and crop growth modules of the SWAP model were improved. The evapotranspiration and cotton growth data of the mulched drip-irrigated cotton fields under three irrigation treatments (W1 = 3360 m³·hm⁻², W2 = 4200 m³·hm⁻², and W3 = 5040 m³·hm⁻²) in 2023 and 2024 at the Xinjiang Modern Water-saving Irrigation Key Experimental Station of the Corps were used to verify the simulation accuracy of the improved SWAP model. Research shows the following: (1) The average relative errors of the simulated evapotranspiration, leaf area index, and dry matter weight of cotton in the improved SWAP crop growth model are all <20% compared with the measured values. The root means square errors of the three treatments (W1, W2, and W3) ranged from 0.85 to 1.38 mm, from 0.03 to 0.18 kg·hm⁻², and 55.01 to 69 kg·hm⁻², respectively. The accuracy of the improved model in simulating evapotranspiration and cotton growth in the mulched cotton field increased by 37.49% and 68.25%, respectively. (2) The evapotranspiration rate of cotton fields is positively correlated with the irrigation water volume and is most influenced by meteorological factors such as temperature and solar radiation. During the flowering stage, evapotranspiration accounted for 62.83%, 62.09%, 61.21%, 26.46%, 40.01%, and 38.8% of the total evapotranspiration. Therefore, the improved SWAP model can effectively simulate the evaporation and transpiration of the mulched drip-irrigated cotton fields in the Manas River Basin. This study provides a scientific basis for the digital simulation of mulched farmland in the arid regions of Northwest China.

1. Introduction

Cotton is an important economic crop in the Manas River Basin. Approximately 63% of the water resources in this basin are used for irrigating cotton fields [1], while over 90% of the water in the cotton fields is lost because of evaporation [2]. Subsurface drip irrigation planting mode has changed the traditional process of water movement in farmland [3]. In contrast, it reduces soil water evaporation and increases soil temperature and moisture content [4,5], while also improving the soil pore structure, reducing soil water leakage, and creating favorable water and heat conditions for cotton root development [6,7]. Therefore, accurately quantifying the evapotranspiration of farmland and the growth dynamics of crops in the Manas River Basin is the basis for comprehensively understanding the water consumption structure of farmland and improving agricultural techniques. This is critical for precision agriculture and informed decision-making [8,9]. To understand the farmland evaporation process, the dynamics of land evaporation at the regional scale were analyzed by combining satellite remote sensing and meteorological data. However, the accuracy of remote sensing data is relatively low and the associated costs are high. Therefore, these data are not applicable to the measurement of evaporation in mulched drip irrigation farmland in the Manas River Basin [10,11]. Hydrogen and oxygen isotope technology is also used to calculate evapotranspiration at the field scale. Previous studies have investigated non-productive water loss in corn and explored the dynamics of the regional water cycle. However, frequent manual sampling, data measurement, and analysis constrain the broader application of this technology in mulched drip-irrigated cotton fields [12,13]. Additionally, methods such as the vorticity correlation and evapotranspiration meter methods are commonly used to calculate evapotranspiration [14,15]. However, there are relatively few quantitative studies on evaporation and cotton growth in mulched cotton fields in the Manas River Basin.

Compared with other measurement methods that require destructive sampling in the field on the farm, crop growth models offer a nondestructive approach to analyzing cotton growth dynamics and farmland evaporation conditions, with enhanced real-time responsiveness and sustained continuity [16]. In recent years, researchers from around the world have developed different crop models to study the responses of farmland evapotranspiration and crop growth dynamics to irrigation, climatic conditions, and agricultural management measures. Prominent models in this domain include AquaCrop [17], ORYZA2000 [18], DSSAT [19], and WOFOST [20]. The SWAP (Soil Water Atmosphere Plant) model is a comprehensive model that simulates the movement of soil water and salts, heat transfer, evaporation from farmland, crop growth, and yield prediction [21,22]. It is widely applied to simulate evapotranspiration and crop growth at the field scale.

Several recent studies have underscored the utility of this model in diverse agroecological environments. For instance, Vianna et al. [23] integrated the SWAP model with Standalone and Tipping-Bucket approaches into the SAMUCA model to evaluate irrigation efficiency and sugarcane growth in Brazil, reporting high simulation accuracy with a root mean square error (RMSE) of only 6%. Likewise, Delgado et al. [24] applied the SWAP model to simulate quinoa growth under varying soil textures and irrigation volumes in arid regions of Morocco and Belgium. Their findings confirmed the model’s ability to capture growth responses under both deficit irrigation and mildly saline conditions with high accuracy. Furthermore, Yuan [25] successfully used the SWAP model to simulate water and salt fluxes across farmland and fallow land in the Hexi Irrigation District.

Current studies indicate that the SWAP model has been optimized to enable its application in modernized farmlands that are constantly improving. For instance, Zhao et al. [26] modified the evapotranspiration and soil water parameters in the original SWAP model to simulate soil water and heat dynamics in mulched maize fields. These improvements accounted for the interception effects of plastic mulch on soil temperature and precipitation. However, optimizing model parameters requires a complex calibration procedure. Vincent and Davies [27] highlighted the role of planting holes in altering soil aeration and root zone structures. Notably, in arid and semi-arid regions, planting holes may partially mitigate the insulating effect of plastic mulch, enabling limited exchange of heat and moisture between the soil and the atmosphere. Given these interactions, incorporating the combined effects of plastic mulching and planting holes is essential for accurately simulating soil thermal and hydrological regimes in cotton fields. Enhancing the source code level can reduce calibration workload and support high-precision simulation of mulching drip-irrigated cotton fields in the Manas River Basin under the influence of mulching and planting holes.

In response to these challenges, this study presents an improved version of the SWAP model, developed through source code modification to enhance the evaporation, crop growth, and soil temperature modules. These modifications were based on a detailed analysis of the compensatory effects of varying irrigation levels, plastic mulching, and planting hole configurations on soil temperature in cotton fields. To evaluate performance of the model, empirical data on evaporation, leaf area index (LAI), and cotton dry matter from mulched cotton fields in 2023 and 2024 were used for calibration and validation. The improved SWAP model aims to improve the accuracy of simulating evapotranspiration and cotton growth under mulched drip irrigation across different irrigation gradients. This study established a scientific basis for optimizing irrigation strategies and advancing the implementation of efficient water-saving technologies in plastic mulch cropping systems.

2. Materials and Methods

2.1. Study Area

We conducted field trials for two consecutive years from 2023 to 2024 at the Key Laboratory of Modern Water-saving Irrigation, affiliated with the Turpan Agricultural Science and Technology Bureau (latitude 89°59′47″, longitude 44°19′28″, altitude 412 m).

The experimental station is located in the middle section of the northern slope of the Tianshan Mountains in Xinjiang––in the middle reaches of the Manas River Basin––on the southern edge of the Junggar Basin. The average annual rainfall in this region is 150 mm, primarily occurring during the summer months, while the average annual evaporation is 2100 mm. This region exhibits a temperate continental climate [28]. Groundwater was found at a depth of 8 m. The mean soil bulk density in the experimental pits was 1.43 g·cm⁻³. Meteorological data collected from a weather station (TRM-ZS1, Jinzhou Sunshine Technology Development Co., Ltd, Jinzhou, China) mounted 2 m above the ground at the test site are shown in Figure 1. in Figure 1, and the soil physicochemical properties for the 0–100 cm profile are presented in

Table 1. Soil parameter data at different depths.

Soil Depth /cm	Sand /%	Silt /%	Clay /%	International Soil Texture	Saturated Water Content /%	Field Capacity /%
0–30	31.73	40.94	27.33	Loam	43.7	30.7
30–60	30.54	41.45	28.01	Loam	43.3	30.2
60–100	29.02	40.87	30.11	Loam	44.2	32.3

.

2.2. Experiment Design

The cotton variety used in this experiment was Jinken 1442. Sowing was conducted on 24 April 2023 and 25 April 2024. Each treatment included three test pits, and two replicate sets were established to ensure the reliability of the experiment. The dimensions of each test pit were 2 m × 2 m × 2 m. The bottom was filled with gravel: first 1–2 mm coarse sand layer, then 5–10 mm fine gravel, and finally a 20–30 mm coarse gravel, with each layer being 10 cm thick. and the walls were reinforced with brickwork to ensure stability and isolation. The planting density was maintained at 25 plants·m⁻². Cotton was sown with alternating row spacings of 30 and 60 cm, with a plant spacing of 10 cm. The planting configuration used one mulch film covering two drip irrigation belts, with each drip irrigation belt irrigating two rows of cotton plants. Apart from the water used for sowing, the total irrigation water volume for the entire growth period of cotton was set at three different levels (W1 = 3360 m³·hm⁻², W2 = 4200 m³·hm⁻², and W3 = 5040 m³·hm⁻²), which were 80%, 100%, and 120% of the local field irrigation water volume, respectively. Subsurface drip irrigation was applied using polyethylene drip tape (diameter: 16 mm). Emitters were spaced 30 cm apart, and each had a flow rate of 2 L·h⁻¹. Each emitter was positioned 5 cm from the cotton plant root zone to ensure efficient water delivery to the plants. The amount of water delivered during each irrigation event was measured using a high-precision water meter attached to the drip irrigation pipe. The experimental layout is shown in Figure 2, and the irrigation schedule across different growth stages for both years is presented in

Table 2. Irrigation amounts for different growth stages in 2023–2024.

Year	Growing Period	Date /(mm/dd)	Irrigation Time	Irrigation Quota /m3·hm−2	Irrigation Quota /m3·hm−2	Irrigation Quota /m3·hm−2
2023	Sowing	04/24	1	336	420	504
	Seeding	05/06–06/04	1	336	420	504
	Flowering	06/05–07/08	2	268.8	336	403.2
	Boling	07/09–08/19	6	336	420	504
	Boll opening	08/20–09/25	2	235.2	294	352.8
Total			12	3696	4640	5544
2024	Sowing	04/25	1	336	420	504
	Seeding	05/08–06/06	1	336	420	504
	Flowering	06/07–07/10	2	268.8	336	403.2
	Boling	07/11–08/21	6	336	420	504
	Boll opening	08/22–09/27	2	235.2	294	352.8
Total			12	3696	4640	5544

.

2.3. Data Measurement

2.3.1. Cotton Field Evapotranspiration

The evapotranspiration of cotton was measured using a micro-plantation evapotranspiration instrument (Xi An University of Technology) in the measurement pit [29]. Evapotranspiration at three-day intervals at 10:00, from 25 May 2023 to 22 May 2024, spanning the cotton growing season until harvest. During the flowering and fruiting stages, rainy, cloudy, and sunny days with similar dates were selected, and the evapotranspiration of the cotton field was recorded every 2 h under typical weather conditions [30].

2.3.2. Soil Moisture Content

Soil samples were collected before each watering and dried to measure the soil moisture content at depths of 5, 10, 20, 40, 60, and 80 cm. Soil moisture content was measured at nine-day intervals, with the 2023 monitoring commencing on 6 May and the 2024 monitoring commencing on 8 May [31]. The measured values of the initial soil volumetric water content for each growing period have been calibrated against the reference PICO-32 TDR device (IMKO Micromodultechnik GmbH, Ettlingen, Germany).

2.3.3. Soil Temperature

Glass right-angle thermometers (YF-303, YUNFEI SCI-TECH, Zhengzhou, China) were used to measure the soil temperatures at depths of 5, 10, 15, 20, and 25 cm beneath the surface of each test pit. When transpiration measurements were performed, soil temperature data were collected simultaneously.

2.3.4. Growth Temperature

Meteorological data were used to compare the effective accumulated temperature for cotton growth based on the growing degree day (GDD) method, as proposed by Yang et al. [32].

$${\mathrm{T}}_{\mathrm{C}}=\sum ({\mathrm{T}}_{\mathrm{a}}-{\mathrm{T}}_{\mathrm{b}})$$

(1)

where ${\mathrm{T}}_{\mathrm{C}}$ is the growing degree days for cotton development (°C); ${\mathrm{T}}_{\mathrm{a}}$ is the daily mean temperature (°C); and ${\mathrm{T}}_{\mathrm{b}}$ is the minimum temperature required for cotton to grow (°C), T_b = 10 °C.

2.3.5. Growth Indicators and Dry Biomass

Every 7–10 days, three uniformly growing cotton plants were randomly selected for physiological and morphological measurements at key growth stages: seedling, budding, flowering–bolling, and fluffing. The measured parameters included plant height and stem diameter. LAI was calculated using Equation (2).

$$\mathrm{L}\mathrm{A}\mathrm{I}=0.83{\displaystyle \frac{\mathrm{L}\mathrm{W}}{\mathrm{A}}}$$

(2)

where L is the leaf length (cm); w is the leaf width (cm); A is the ground area occupied by an individual cotton plant (cm²).

To determine biomass, cotton plant organs were initially blanched in an oven at 105 °C for 30 min to deactivate enzymes, followed by drying at 75 °C for 72 h until a constant weight was achieved. Dry biomass was measured using a high-precision balance with an accuracy of ±0.1 g.

2.4. Modification Model

The SWAP model was developed by Wageningen University in the Netherlands in 1978. It is a comprehensive multidimensional model for soil, water, and plants [33]. The model has demonstrated strong capabilities in simulating both evapotranspiration and plant growth dynamics [34]. However, its original structure does not account for the influence of plastic film mulching, which plays a significant role in modifying soil thermal and moisture regimes by preventing rainfall infiltration, reducing evaporation, and regulating surface temperature [35]. To address these limitations, the SWAP model was improved by modifying key parameters to incorporate the effects of plastic mulching on cotton growth and field-scale evapotranspiration. Model calibration was conducted using field-measured data, including soil moisture, LAI, and cotton dry biomass, under varying irrigation treatments. These enhancements aimed to increase the simulation accuracy of the model in mulched drip-irrigated cotton fields.

2.4.1. Modification of the Soil Temperature Module

In its default configuration, the SWAP model applies Dirichlet boundary conditions at the upper and lower limits of the soil profiles. The top boundary is determined based on air temperature. The bottom boundary is either a fixed temperature value or a Neumann boundary. However, the presence of plastic mulch alters the interaction between air and surface soil temperatures, particularly during the early growth stages. Furthermore, in non-mulched conditions, the crop LAI influences the surface energy balance [36]. When the LAI is less than 1 cm²·cm⁻², plastic mulch has a pronounced effect on the heat flux in the upper 0−10 cm of soil [37,38]. However, when the LAI exceeds 1 cm²·cm⁻², the impact of mulching on soil temperature becomes negligible owing to increased canopy coverage.

To accurately simulate temperature dynamics during the early growth stages (LAI < 1 cm²·cm⁻²), a temperature compensation coefficient was introduced, following Zhang et al. [39]. The coefficient is defined as follows:

$${\mathrm{C}}_{\mathrm{m}}^{\mathrm{i}}={\displaystyle \frac{{\mathrm{G}\mathrm{D}\mathrm{D}}_{\mathrm{a}\left(\mathrm{N}\mathrm{M}\right)}^{\mathrm{i}}-{\mathrm{G}\mathrm{D}\mathrm{D}}_{\mathrm{a}\left(\mathrm{P}\mathrm{M}\right)}^{\mathrm{i}}}{\left[{\mathrm{G}\mathrm{D}\mathrm{D}}_{\mathrm{s}\left(\mathrm{P}\mathrm{M}\right)}^{\mathrm{i}}-{\mathrm{G}\mathrm{D}\mathrm{D}}_{\mathrm{s}\left(\mathrm{P}\mathrm{M}\right)}^{\mathrm{i}-1}\right]-[{\mathrm{G}\mathrm{D}\mathrm{D}}_{\mathrm{s}\left(\mathrm{N}\mathrm{M}\right)}^{\mathrm{i}}-{\mathrm{G}\mathrm{D}\mathrm{D}}_{\mathrm{s}\left(\mathrm{P}\mathrm{M}\right)}^{\mathrm{i}-1}]}}(\mathrm{i}\ge 1)$$

(3)

where ${\mathrm{C}}_{\mathrm{m}}^{\mathrm{i}}$ represents the compensation coefficient for the ith growth stage of cotton; ${\mathrm{G}\mathrm{D}\mathrm{D}}_{\mathrm{a}\left(\mathrm{N}\mathrm{M}\right)}^{\mathrm{i}}$ and ${\mathrm{G}\mathrm{D}\mathrm{D}}_{\mathrm{a}\left(\mathrm{P}\mathrm{M}\right)}^{\mathrm{i}}$ represent the thermal time points of the ith growth stage for non-coated and coated cotton, respectively; and ${\mathrm{G}\mathrm{D}\mathrm{D}}_{\mathrm{s}\left(\mathrm{N}\mathrm{M}\right)}^{\mathrm{i}}$ and ${\mathrm{G}\mathrm{D}\mathrm{D}}_{\mathrm{s}\left(\mathrm{P}\mathrm{M}\right)}^{\mathrm{i}}$ represent the cumulative soil temperatures during the ith growth stage of non-coated and coated cotton, respectively. When i equals 1, both ${\mathrm{G}\mathrm{D}\mathrm{D}}_{\mathrm{s}\left(\mathrm{N}\mathrm{M}\right)}^0$ and ${\mathrm{G}\mathrm{D}\mathrm{D}}_{\mathrm{s}\left(\mathrm{P}\mathrm{M}\right)}^0$ are at 0 °C.

The soil temperature module of the SWAP model, improved with the temperature compensation coefficient, is calculated as follows:

$$\mathsf{\Delta }\mathrm{T}={\displaystyle \frac{{\mathrm{C}}_{\mathrm{m}}\times \left[{\mathrm{T}}_{\mathrm{s}\left(\mathrm{P}\mathrm{M}\right)}-{\mathrm{T}}_{\mathrm{s}\left(\mathrm{N}\mathrm{M}\right)}\right]\times [{\mathrm{T}}_{\mathrm{a}\left(\mathrm{N}\mathrm{M}\right)}-{\mathrm{T}}_{\mathrm{b}\mathrm{a}\mathrm{s}\mathrm{e}}]}{{\mathrm{T}}_{\mathrm{s}\left(\mathrm{N}\mathrm{M}\right)}-{\mathrm{T}}_{\mathrm{b}\mathrm{a}\mathrm{s}\mathrm{e}}}},$$

(4)

$${\mathrm{T}}_{\mathrm{a}(\mathrm{P}\mathrm{M})}={\mathrm{T}}_{\mathrm{a}\left(\mathrm{N}\mathrm{M}\right)}+\mathsf{\Delta }\mathrm{T},$$

(5)

$${\mathrm{T}}_{\mathrm{a}(\mathrm{P}\mathrm{M})}={\displaystyle \frac{{\mathrm{T}}_{\mathrm{m}\mathrm{a}\mathrm{x}}+{\mathrm{T}}_{\mathrm{m}\mathrm{i}\mathrm{n}}}{2}},$$

(6)

where ΔT is the adjustment temperature of the model for every 1 °C increase in soil temperature under the film covering condition (°C); T_base is the base temperature for cotton growth and was set to 10 °C in this study; Ts_(NM) and Ts_(PM) represent soil temperatures under non-mulched and plastic-mulched conditions, respectively (°C); Ta_(NM) and Ta_(PM) denote air temperatures under non-mulched and plastic-mulched conditions (°C), respectively; and T_max and T_min are the maximum and minimum temperature outputs of the SWAP model (°C), respectively.

For later growth stages, where LAI ≥ 1 cm²·cm⁻², the influence of mulching on soil temperature was negligible. In these cases, the observed surface temperature was directly used as the top boundary condition, and a Neumann boundary with zero heat flux was applied at the bottom of the soil profile.

2.4.2. Modification of the Evapotranspiration Module

In the SWAP model, the upper boundary condition for soil moisture is typically set as a Neumann boundary. In this study, the bottom of each experimental pit was sealed using anti-seepage measures, ensuring that the water input into the mulched cotton fields was derived solely from irrigation and rainfall infiltration. However, owing to the interception effect of the cotton canopy and the presence of plastic film, not all incoming water infiltrated the soil. Specifically, plastic film was found to block approximately 20% of rainfall from reaching the soil surface [40]. Rainfall interception was calculated using Equation (6). The potential evapotranspiration of cotton fields based on daily weather data was also calculated by directly applying the Penman–Monteith method, and the potential crop and soil evaporation were also calculated, as follows:

$${\mathsf{\Delta }\mathrm{P}}_{\mathrm{f}}=\left[{\displaystyle \frac{1.2·\mathrm{B}·\mathrm{P}}{1.2\mathrm{a}·\mathrm{L}\mathrm{A}\mathrm{I}+\mathrm{B}·\mathrm{P}}}\right],$$

(7)

$$\mathrm{B}=1.2(1-{\mathrm{e}}^{-\mathrm{kgrLAI}}),$$

(8)

$${\mathrm{T}}_{\mathrm{p}}={\displaystyle \frac{\left(1-{\mathrm{W}}_{\mathrm{f}\mathrm{r}\mathrm{a}\mathrm{c}}\right)[{\mathrm{V}}_{\mathrm{c}}\frac{{∆}_{\mathrm{v}}}{{\mathsf{\lambda }}_{\mathrm{w}}}\left({\mathrm{R}}_{\mathrm{n}}-\mathrm{G}\right)+\frac{{\mathrm{P}}_1{\mathsf{\rho }}_{\mathrm{a}}{\mathrm{C}}_{\mathrm{a}}}{{\mathsf{\lambda }}_{\mathrm{W}}}(\frac{{\mathrm{e}}_{\mathrm{s}\mathrm{a}\mathrm{t}}-{\mathrm{e}}_{\mathrm{a}}}{{\mathsf{\gamma }}_{\mathrm{a},\mathrm{c}}})]}{{∆}_{\mathrm{v}}+{\mathsf{\gamma }}_{\mathrm{a}}[1+\frac{{\mathsf{\gamma }}_{\mathrm{s},\mathrm{m}\mathrm{i}\mathrm{n}}}{{\mathsf{\gamma }}_{\mathrm{a},\mathrm{c}}{\mathrm{L}\mathrm{A}\mathrm{I}}_{\mathrm{e}\mathrm{f}\mathrm{f}}}]}},$$

(9)

$${\mathrm{E}}_{\mathrm{p}}={\displaystyle \frac{\left(1-{\mathrm{V}}_{\mathrm{c}}\right)\frac{{∆}_{\mathrm{v}}}{{\mathsf{\lambda }}_{\mathrm{w}}}\left({\mathrm{R}}_{\mathrm{n}}-\mathrm{G}\right)+\frac{{\mathrm{P}}_1{\mathsf{\rho }}_{\mathrm{a}}{\mathrm{C}}_{\mathrm{a}}}{{\mathsf{\lambda }}_{\mathrm{W}}}(\frac{{\mathrm{e}}_{\mathrm{s}\mathrm{a}\mathrm{t}}-{\mathrm{e}}_{\mathrm{a}}}{{\mathsf{\gamma }}_{\mathrm{a},\mathrm{s}}})}{{∆}_{\mathrm{v}}+{\mathsf{\gamma }}_{\mathrm{a}}(1+\frac{{\mathsf{\gamma }}_{\mathrm{s}}}{{\mathsf{\gamma }}_{\mathrm{a},\mathrm{s}}})}},$$

(10)

$${\mathrm{L}\mathrm{A}\mathrm{I}}_{\mathrm{e}\mathrm{f}\mathrm{f}}={\displaystyle \frac{\mathrm{L}\mathrm{A}\mathrm{I}}{0.3\mathrm{L}\mathrm{A}\mathrm{I}+1.2}},$$

(11)

where ΔP denotes the total interception amount in the cotton field (mm·d⁻¹); P is the total rainfall amount (mm·d⁻¹); $\mathrm{a}$ is the experience coefficient (mm·d⁻¹); ${\mathrm{W}}_{\mathrm{f}\mathrm{r}\mathrm{a}\mathrm{c}}$ represents the proportion of time during which the crop canopy is moist in a day; ${∆}_{\mathrm{v}}$ is the slope of the water vapor pressure curve (kPa·°C⁻¹); ${\mathsf{\lambda }}_{\mathrm{w}}$ is the latent heat of vaporization (J·kg⁻¹); ${\mathrm{V}}_{\mathrm{c}}$ represents the vegetation coverage rate; ${\mathrm{R}}_{\mathrm{n}}$ represents the net radiation flux at the surface of the crop canopy (J·m⁻²·d⁻¹); G represents the soil heat flux (J·m⁻²·d⁻¹); P₁ is the time unit conversion coefficient (86,400 s·d⁻¹); ρ_a is the air density (kg·m⁻³); C_a is the specific heat capacity of moist air (J·kg⁻¹·°C⁻¹); e_sat is the saturated vapor pressure (kPa); e_a is the actual vapor pressure (kPa); ${\mathsf{\gamma }}_{\mathrm{a},\mathrm{c}}$ is the uniform canopy air resistance (s·m⁻¹); LAI_eff is the effective leaf area index; γ_s,m is the minimum stomatal resistance (s·m⁻¹); γ_a is the humidity constant (kPa·°C⁻¹); γ_s is the moist soil resistance (s·m⁻¹); and γ_a,s is the soil surface aerodynamic resistance (s·m⁻¹).

Although mulching significantly reduces soil evaporation [41], evaporation still occurs through planting holes and areas where mulch is absent or disrupted by human activities. This study comprehensively considered the limitation of soil evaporation caused by plastic film during the initial stage of cotton growth and development, the retention effect of the plastic film on rainfall, and the soil evaporation volume in the planting holes without the plastic film. The relationship between the plastic film perforation rate and actual soil evaporation was introduced to accurately quantify the soil evaporation volume. In this study, the planting hole perforation rate was set to approximately 2.52% according to the mainstream seeding machinery used locally. Based on the correlation between the soil evaporation rate and the planting hole perforation rate established by Li et al. [42] under soil conditions similar to those used in this study, the relationship can be expressed as follows:

$$\mathrm{e}=(0.00501{\mathsf{\theta }}^{0.1019}+0.00375){\mathrm{t}}^{-0.5},$$

(12)

where $\mathrm{e}$ is the soil evaporation rate and $\mathsf{\theta }$ is the opening rate of the planting holes.

To determine the actual soil evaporation E_soil in a cotton field under mulched drip irrigation, we can combine the calculation of the planting hole opening rate with the original SWAP model, as expressed in the following equations:

$${\mathrm{E}}_{\mathrm{f},\mathrm{m}}=(\mathrm{e}-0.00375){\mathrm{k}}_{1/2}({\displaystyle \frac{{\mathrm{h}}_{\mathrm{a}\mathrm{t}\mathrm{m}-}{\mathrm{h}}_1-{\mathrm{Z}}_1}{{\mathrm{Z}}_1}}),$$

(13)

$${\mathrm{E}}_{\mathrm{f},\mathrm{a}}=(\mathrm{e}-0.00375){\mathsf{\beta }}_1{\mathrm{t}}_{\mathrm{d}\mathrm{r}\mathrm{y}}^{1/2},$$

(14)

$${\mathrm{E}}_{\mathrm{f},\mathrm{p}}=(\mathrm{e}-0.00375){\displaystyle \frac{\left(1-{\mathrm{V}}_{\mathrm{c}}\right)\frac{{∆}_{\mathrm{v}}}{{\mathsf{\lambda }}_{\mathrm{w}}}\left({\mathrm{R}}_{\mathrm{n}}-\mathrm{G}\right)+\frac{{\mathrm{P}}_1{\mathsf{\rho }}_{\mathrm{a}}{\mathrm{C}}_{\mathrm{a}}}{{\mathsf{\lambda }}_{\mathrm{W}}}(\frac{{\mathrm{e}}_{\mathrm{s}\mathrm{a}\mathrm{t}}-{\mathrm{e}}_{\mathrm{a}}}{{\mathsf{\gamma }}_{\mathrm{a},\mathrm{s}}})}{{∆}_{\mathrm{v}}+{\mathsf{\gamma }}_{\mathrm{a}}(1+\frac{{\mathsf{\gamma }}_{\mathrm{s}}}{{\mathsf{\gamma }}_{\mathrm{a},\mathrm{s}}})}},$$

(15)

$${\mathrm{E}}_{\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l}}=\mathrm{m}\mathrm{i}\mathrm{n}({\mathrm{E}}_{\mathrm{f},\mathrm{m}},{\mathrm{E}}_{\mathrm{f},\mathrm{a}},{\mathrm{E}}_{\mathrm{f},\mathrm{p}}),$$

(16)

where ${\mathrm{k}}_{0.5}$ is the average hydraulic conductivity from the surface to the first soil layer (cm·d⁻¹); ${\mathrm{h}}_{\mathrm{a}\mathrm{t}\mathrm{m}}$ is the soil water head that is in equilibrium with the relative humidity of the air (cm); ${\mathrm{h}}_1$ is the soil water head in the first soil layer (cm); ${\mathrm{Z}}_1$ is the depth of the first soil layer (cm); and ${\mathsf{\beta }}_1$ is the soil-specific parameter (cmd^0.5).

2.4.3. Modification of the Crop Growth Module

The original SWAP crop growth module estimates photosynthesis and respiration primarily based on thermal time. However, the root zone temperature, particularly at shallow depths, plays a key role in regulating cotton development [43]. Plastic mulching significantly increases soil temperature from the surface down to a depth of 5 cm and promotes the accumulation of root biomass in the 30 cm soil layer [44,45]. To reflect this effect, the model was enhanced by incorporating measured soil temperatures at 5 cm depth to calculate effective accumulated temperatures for two key periods: from seedling emergence to boll formation, and from boll formation to boll opening. In addition, mulching improves soil hydrothermal conditions, thereby enhancing leaf growth and canopy light interception. According to Liao et al. [46], these improvements have accelerated crop development. Following Tao et al. [47], the canopy light absorption rate was calculated as follows:

$$\mathrm{P}\mathrm{A}\mathrm{R}=0.5{\mathrm{R}}_{\mathrm{g}}{\displaystyle \frac{\mathrm{s}\mathrm{i}\mathrm{n}{\mathsf{\beta }}_{\mathrm{s}\mathrm{u}\mathrm{n}}(1+0.4\mathrm{s}\mathrm{i}\mathrm{n}{\mathsf{\beta }}_{\mathrm{s}\mathrm{u}\mathrm{n}})}{\int \mathrm{s}\mathrm{i}\mathrm{n}{\mathsf{\beta }}_{\mathrm{m}\mathrm{o}\mathrm{d},\mathrm{s}\mathrm{u}\mathrm{n}}}},$$

(17)

$$\mathrm{B}\mathrm{P}\mathrm{A}\mathrm{R}=\mathrm{P}\mathrm{A}\mathrm{R}{\mathrm{e}}^{-\mathrm{kL}},$$

(18)

$$\mathrm{F}=(1-\mathrm{BPAR}/\mathrm{PAR})\times 100\%,$$

(19)

$$\mathrm{PAR}{\mathrm{a},}_{\mathrm{l}}=\sum {\displaystyle \frac{2\mathrm{F}}{1+\mathrm{F}}}{\mathrm{PAR}}_{\mathrm{s}},$$

(20)

where PAR_a,l is the light absorption rate ofe the cotton canopy; PAR_s is the measured solar effective radiation; BPAR is the photosynthetically active radiation at the bottom of the canopy; and L is the total leaf area.

2.4.4. Parameter Calibration

The SWAP model requires input parameters related to both soil and crop characteristics [48]. Soil hydraulic parameters were derived from the measured soil physical and chemical properties. The crop parameters included the following: the initial dry weight of the crop, which was determined by drying and measuring the cotton plants at the seedling stage; the maximum relative increase in LAI, which was calculated based on the LAI and growth-stage temperature; the specific leaf area, which was calculated based on the leaf area and dry weight of the cotton during the four growth stages; the leaf lifespan under optimal conditions, which was obtained by measuring the lifespan of a specific cotton leaf; the efficiency of converting total dry matter into the dry weight of each organ of cotton, which was calculated by sampling and calculating the proportion of the dry weight of each organ of cotton during different growth periods; and the maximum CO₂ assimilation, which was calculated by fitting the light response curve obtained from measuring the upper and lower limits of suitable temperatures during the four growth periods of cotton. The model was calibrated using field data from 2023 and validated using data from 2024. Calibration was guided by the high-sensitivity input parameters identified by Ravensbergen et al. [49]. The final calibrated values are listed in

Table 3. Crop parameters used in SWAP.

Serial Number	Parameter	Initial Value	Correction Value	Source of Parameter
1	TSUME	750.0	886.6	Measured
2	TSUMAM	1550.0	1045.1	Measured
3	TDWI	75.0	6.185	Measured
4	RGRLAI	0.012	0.0400	Measured
5	SLA (0—0.55—0.86—1.17—1.75—2)	0.003—0.003—0.0015—0—0—0	0.003—0.003—0.0035—0.0025—0.024	Measured
6	SPAN	47.0	60.0	Measured
7	RML (kg CH2O kg d−1)	0.0300	0.0300	J.G. Kroes [50]
8	RMO (kg·kg−1)	0.60	0.50	J.G. Kroes [50]
9	CVL	0.72	0.70	Cheng, et al. [51]
10	CVR	0.72	0.70	Measured
11	CVO	0.85	0.50	Measured
12	CVS (kg·kg−1)	0.69	0.80	Measured
13	FR (0.53—0.87—1.19—1.77)	0.2—0.2—0—0	0.19—0.13—0.06—0.10	J.G. Kroes [50]
14	KDIF	1.0	0.6	J.G. Kroes [50]
15	KDIR	0.75	0.2	J.G. Kroes [50]
16	AMAX (kg·ha·h−1) (0—1—1.7—2)	30.0—30.0—30.0—0.0	30.0—40.0—40.0—30.0	J.G. Kroes [50]

and

Table 4. Soil hydraulic parameters in the Van Genuchten Mualem (VGM) model.

Depth /cm	Residual Water Content /cm3·cm−3	Saturated Water Content /cm3·cm−3	Shape Parameter /cm−1	Parameter of Curve	Saturated Hydraulic Conductivity /cm·d−1	Hydraulic Conductivity Shape Factor
0–30	0.030	0.437	0.006	1.619	64.89	0.5
30–60	0.030	0.433	0.008	1.652	80.35	0.5
60–100	0.035	0.442	0.010	1.586	71.64	0.5

.

2.5. Evaluation of Simulation Effects

The performance of the improved SWAP model was evaluated by comparing the simulated outputs with the corresponding field-measured values. Two statistical indicators were used to assess model accuracy: root mean square error (RMSE) and mean relative error (MRE) [52]. Lower RMSE and MRE values indicate better agreement between the simulated and observed data, which reflects higher model accuracy. These evaluation metrics were calculated as follows:

$$\mathrm{R}\mathrm{M}\mathrm{S}\mathrm{E}=\sqrt{{\displaystyle \frac{\sum _{\mathrm{i}=1}^{\mathrm{i}=\mathrm{N}}{\left({\mathrm{M}}_{\mathrm{i}}-{\mathrm{S}}_{\mathrm{i}}\right)}^2}{\mathrm{N}}}},$$

(21)

$$\mathrm{M}\mathrm{R}\mathrm{E}={\displaystyle \frac{1}{\mathrm{N}}}{\sum }_{\mathrm{i}=1}^{\mathrm{N}}\left|{\displaystyle \frac{{\mathrm{M}}_{\mathrm{i}}-{\mathrm{S}}_{\mathrm{i}}}{{\mathrm{S}}_{\mathrm{i}}}}\right|\times 100\%,$$

(22)

where M_i is the simulated value; S_i is the measured value; and N is the total number of both simulated and measured values.

These metrics were used to evaluate the accuracy of the model in simulating key parameters, including evapotranspiration, LAI, and cotton dry matter accumulation, across different irrigation treatments and cotton growth stages.

3. Results

3.1. Simulation of Evapotranspiration in Mulched Cotton Fields

3.1.1. Simulation of the Temporal Course of Evapotranspiration in Cotton Fields

The simulation results of evapotranspiration intensity and daily cumulative evapotranspiration of mulched cotton fields under clear, cloudy, and rainy conditions based on the improved SWAP model for the W2 treatment are shown in Figure 3 and Figure 4. The RMSE and MRE of the improved model ranged from 0.99 to 1.91 mm and 10.57% to 18.52%, respectively, indicating a significant improvement in the model’s simulation accuracy. As shown in

Table 5. Simulation evaluation of daily cumulative evapotranspiration in cotton fields under the W2 treatment in 2023 and 2024.

Models	Irrigation Treatment	2023		2024
		RMSE /mm	MRE /%	RMSE /mm	MRE /%
Original	W2	1.46	2.54	2.05	2.52
Modified	W2	0.91	1.47	0.98	1.76

, the RMSE and MRE in 2023 were 7.7% and 19.3% lower than those in 2024, respectively. Simulation accuracy was higher in 2023 than in 2024, likely because 2023 served as the calibration year and 2024 as the validation year. This result is consistent with those reported by Zhang et al. [53]. Under the three meteorological conditions, the evapotranspiration intensity in the cotton field followed a single-peak or multi-peak curve trend. Evapotranspiration intensity was relatively low before 08:00 a.m. and after 08:00 p.m., whereas it was strongest between 12:00 p.m. and 05:00 p.m. The peak evaporation intensity of the covered cotton fields was lower on cloudy and rainy days than on sunny days. Minimal variation in evaporation intensity was observed during morning and evening hours across the three weather conditions. Under different weather conditions, the cumulative daily evapotranspiration in cotton fields exhibited an S-shaped curve with slope approaching zero at dawn and dusk. Evapotranspiration accumulation progressed slowly during nighttime hours. The results indicated that evapotranspiration varied significantly under different weather conditions. On sunny days, the evapotranspiration was 2.8 times greater than that on rainy days.

3.1.2. Correlation Between Evapotranspiration in Cotton Fields and Meteorological Factors

The correlations between the evapotranspiration intensity of cotton fields and temperature, wind speed, saturated vapor pressure, relative humidity, rainfall, and solar radiation are shown in Figure 5. Based on the SPSS (IBM SPSS Statistics 26) model, the correlation analysis showed that eevapotranspiration intensity was positively correlated with temperature and solar radiation and negatively correlated with saturated vapor pressure, relative humidity, and rainfall. Correlation between evapotranspiration intensity and meteorological factors varied across parameters. Fitting relationships with saturated vapor pressure, rainfall, and temperature were relatively strong, whereas the correlation with wind speed exhibited lower predictive accuracy. These findings are consistent with those reported previously [54,55]. The ranked strength of the response relationship between cotton field evapotranspiration and meteorological factors was as follows: temperature > solar radiation > rainfall > saturated vapor pressure > air humidity > wind speed. Yang and Hou [56] found that the correlation strength between farmland evapotranspiration and meteorological factors is as follows: sunshine duration, air humidity, rainfall, and wind speed. This supports the conclusion that, across temporal and regional scales, the response relationship between farmland evaporation and temperature and solar radiation is the most sensitive. Temperature and solar radiation directly influence crop leaf stomatal dynamics and soil evaporation volume, thereby determining total farmland evapotranspiration [57,58].

3.1.3. Simulation of Daily Evaporation and Transpiration in Cotton Fields

The measured and simulated evapotranspiration values for the cotton growing seasons in 2023 and 2024, based on the improved SWAP model, are shown in Figure 6. Evaporation in mulched drip-irrigated cotton fields was directly proportional to the irrigation gradient, based on observed evaporation and transpiration conditions. As the cotton growth progressed, evapotranspiration initially increased and subsequently decreased.

In 2023, the total evapotranspiration values under the three irrigation levels were 299.95 mm (W1), 326.07 mm (W2), and 350.53 mm (W3), whereas in 2024, these values increased significantly, reaching 442.52 mm (W1), 536.04 mm (W2), and 655.53 mm (W3). These interannual differences underscore the marked influence of meteorological conditions on evapotranspiration, even under consistent irrigation conditions. Furthermore, the distribution of evapotranspiration across different growth stages showed that the flowering and boll-setting stages consistently contributed the largest share of seasonal evapotranspiration. Particularly, evapotranspiration during this period was 188.48 mm (W1), 202.47 mm (W2), and 214.59 mm (W3) in 2023 and 177.10 mm (W1), 214.47 mm (W2), and 254.36 mm (W3) in 2024. As shown in

Table 6. Model evaluation for ET under the W1, W2, and W3 treatments in 2023 and 2024.

Models	Irrigation Treatment	2023		2024
		RMSE/mm	MRE/%	RMSE/mm	MRE/%
Original	W1	2.02	23.45	1.82	23.36
	W2	1.41	20.13	2.35	22.96
	W3	1.15	19.08	2.18	14.25
Modified	W1	1.00	15.06	0.85	9.53
	W2	1.20	16.33	1.03	8.56
	W3	1.24	15.99	1.38	10.27

, the improved SWAP model significantly enhanced simulation accuracy. Compared with the original model, the improved version reduced RMSE by 14.89%, reaching 56.17%, and lowered mean relative error (MRE) by 16.19%, reaching 62.72%. These results confirm that the optimized model accurately captured the evapotranspiration dynamics of plastic-mulched cotton fields under different irrigation treatments.

3.2. Crop Growth Process Simulation

3.2.1. Simulation of Leaf Area Index

The simulation performances of both the original and improved SWAP models for the LAI are shown in Figure 7. The LAI exhibited a typical unimodal curve in both years, characterized by an initial increase followed by a gradual decline. The most rapid expansion occurred during the bud stage, with LAI reaching its peak during the flowering and boll-setting stages. A positive correlation was observed between the LAI and irrigation volume. As shown in

Table 7. Model evaluation for leaf area index under the W1, W2, and W3 treatments in 2023 and 2024.

Models	Irrigation Treatment	2023		2024
		RMSE	MRE/%	RMSE	MRE%
Original	W1	0.07	16.03	0.16	10.18
	W2	0.25	11.94	0.08	17.10
	W3	0.17	11.07	0.06	17.13
Modified	W1	0.05	12.01	0.03	3.14
	W2	0.15	11.63	0.18	7.60
	W3	0.12	10.46	0.03	7.39

, under all three irrigation treatments (W1, W2, and W3), the RMSE and MRE of the improved SWAP model were 0.03–0.25 and 3.14–20%, respectively. Although both models produced simulations within acceptable accuracy thresholds, the improved SWAP model demonstrated a markedly superior performance. In 2023, the RMSE and MRE for LAI simulation were reduced by 34.7% and 9.3%, respectively, compared with the original model. In 2024, these reductions reached 20% and 59.2%, respectively, with the most significant improvement observed under the W1 irrigation treatment. These simulation results confirm that the improved SWAP model offers a precise representation of leaf area development in plastic-mulched cotton fields.

3.2.2. Simulation of Cotton Dry Weight

The simulation results for cotton dry biomass accumulation under the three irrigation treatments are presented in Figure 8. Both the original and improved SWAP models effectively captured the continuous accumulation of dry matter throughout the growing period. As expected, dry biomass accumulation was positively correlated with increasing irrigation volume. The improved SWAP model closely matched the observed values and significantly outperformed the original model. The RMSE and MRE of the improved model ranged from 55.01 to 69.00 mm and 3.64% to 6.37%, respectively. In contrast, the original model yielded higher values (RMSE, 122.23–193.52 mm; MRE, 9.04–18.60%), demonstrating a less reliable performance. The detailed evaluation metrics are provided in

Table 8. Model evaluation for dry biomass under the W1, W2, and W3 treatments in 2023 and 2024.

Models	Irrigation Treatment	2023		2024
		RMSE/kg·hm−2	MRE/%	RMSE/kg·hm−2	MRE/%
Original	W1	184.4	18.60	161.4	9.04
	W2	193.10	15.57	156.11	14.55
	W3	193.52	15.60	122.23	14.63
Modified	W1	56.4	5.70	55.01	4.74
	W2	66.83	3.64	55.76	4.57
	W3	69.00	6.37	58.25	4.09

. These results underscore the enhanced productive capability of the modified model in simulating cotton dry matter dynamics under different irrigation strategies.

4. Discussion

4.1. Simulation and Evaluation of Evaporation and Transpiration in Mulched Cotton Fields

Herein, during model improvement, the proportions of soil evaporation and cotton transpiration were adjusted, thereby enhancing the simulation accuracy. This improvement was attributed to the film-covered drip irrigation planting mode, which not only reduced soil evaporation but also had a compensatory effect on the soil temperature of the farmland [59,60], improved the stomatal opening of the cotton plants, increased the emergence rate, and promoted the transpiration volume of the crops [61]. Zhao et al. [62] found that soil evaporation in the mulched corn field accounted for only 52% of that in non-mulched conditions, and corn transpiration was 1.1 times that in non-mulched conditions. These findings are consistent with the results of this study. The simulation results of both the improved and original models indicate a positive correlation between evapotranspiration in cotton fields and the irrigation gradient. Higher irrigation volume created favorable conditions for cotton growth and accelerates the transpiration rates.

As shown in Figure 6, evaporation trends in cotton fields during the middle and late growth stages in 2024 showed a substantial fluctuation compared to those in 2023. Ramos et al. [63] found that the effect of mulching on soil evaporation during the different growth stages of crops decreased as the crop development stage advanced. Herein, damage to the plastic film and plant senescence during the middle and late cotton growth stages intensified the response relationship between field evaporation and meteorological factors such as irrigation volume, temperature, and rainfall [64].

4.2. Simulation Evaluation of Cotton Dry Matter

Based on the conclusion of Chai et al. [65], this study incorporated the suppressive effect of mulching on early-stage soil evaporation into the model while simultaneously accounting for improvements in soil temperature and moisture content. These variations accelerated cotton growth, particularly during its early development. The experimental results demonstrated a strong positive correlation between irrigation volume and both LAI and dry matter accumulation. Higher irrigation volumes created favorable soil moisture and thermal conditions, thereby stimulating crop growth, a trend corroborated by Xiao et al. [66]. Notably, under deficit irrigation conditions, the regulatory effect of mulching on evapotranspiration, LAI, and dry matter accumulation was particularly pronounced, reinforcing the utility of the mulched drip irrigation model in enhancing on-farm water use efficiency [67]. In terms of improving model simulation performance, the improved SWAP model consistently outperformed the original model in simulating both LAI and dry matter accumulation throughout the growth period. The performance gap widened as the season progressed because of the enhanced simulation of soil hydrothermal conditions in the improved model, which enabled more accurate representation of the late-season physiological processes. These findings are consistent with those of Wu et al. [68]. However, these improvements have limitations. As noted by Li et al. [69], plastic film mulching can accelerate crop development and advance the physiological stages. The improved model predicted earlier leaf senescence and canopy decline than the original model, which continued to simulate active crop growth. This discrepancy led to the improved model overestimating LAI during early growth stages and underestimating it toward the end of the season of the growing season. Particularly, under the 2-year W3 irrigation treatment, LAI values continued to increase slowly at the end of the season, possibly due to reduced physiological demand for water and nutrients in late-stage cotton plants. This residual discrepancy suggests that further refinement of the model is warranted. For large-scale monitoring of the LAI, UAV-based remote sensing technologies may offer higher precision and complement SWAP model applications [70]. In summary, the improved SWAP model demonstrated superior performance in simulating LAI and dry matter accumulation across all irrigation treatments, with RMSE and MRE values significantly lower than those of the original model. These enhancements validate applicability of the model for assessing crop water productivity and optimizing irrigation strategies for plastic-mulched cotton systems.

4.3. Improvement of Limitations Analysis of the SWAP Model

This study improved the soil temperature module of the SWAP model by modifying the measured subsurface temperature. Bayatvarkeshi et al. [71] established response relationships between subsurface and air temperatures using independent algorithms, such as artificial neural networks and complementary methods, thereby reducing the workload associated with temperature measurement. The proportion of precipitation retention by the mulch film and crops is greatly influenced by the growth condition of the crops and the coverage rate of the mulch film [72]. Therefore, the sensitivity of precipitation infiltration to the crop development stage should be emphasized. Herein, the initial SWAP model was mainly improved at the theoretical level, such as through source code. In addition, Xing et al. [73] combined intelligent optimization machine algorithms with the evapotranspiration model, further enhancing the simulation accuracy of the original model for crop growth monitoring and yield estimation. This study only verified the accuracy improvement of the SWAP model at the plot scale. Its simulation accuracy under large-field conditions, extreme weather, different crop types, soil conditions, and agricultural management measures still requires further investigation.

In summary, this study contributes to the body of research on model enhancement in plastic mulch systems by advancing the internal mechanisms of the SWAP model. However, integrating additional factors, such as remote sensing data, crop growth stage-specific rainfall interception, and machine learning-based temperature prediction, offers a broader pathway for improving the applicability and precision of crop models in complex agroecosystems.

5. Conclusions

Based on research into the effects of mulching film coverage on evapotranspiration and cotton growth under different irrigation gradients, this study improved the original SWAP model and analyzed the simulation effects of the improved SWAP model on evapotranspiration and dry mass of cotton in mulched drip irrigation fields. The conclusion is as follows: Evapotranspiration in mulched cotton fields is strongly influenced by irrigation and weather conditions. Saturation vapor pressure, rainfall, and solar radiation have the most significant impact on cotton field evapotranspiration. Dry matter accumulation in cotton is positively correlated with irrigation volume. Compared with the original model, the improved SWAP model significantly enhanced the simulation accuracy of evapotranspiration and cotton dry matter accumulation in mulched drip irrigation cotton fields under different irrigation gradients. This demonstrates the feasibility of using the improved source code of the SWAP model to enhance simulation accuracy and strengthens the application of the model for mulched drip irrigation cotton fields in arid areas.