Effects of Irrigation Methods on Root Distribution, Water Uptake Patterns, and Water Use Efficiency in Transplanted vs. Direct-Seeded Cotton

Abstract

The transplanted cotton–wheat rotation enables dual cropping but alters root system architecture, typically suppressing taproot growth and promoting shallow lateral and fibrous roots, with unclear implications for irrigation response and water use efficiency (WUE). Field experiments were conducted in 2021 and 2022 to investigate root growth, spatial distribution, and water uptake sources, using a minirhizotron system and stable hydrogen and oxygen isotopes. The study examined the effects of two cultivation modes (direct seeding and transplanting) and three irrigation methods (border irrigation, micro-spray tape irrigation, and surface drip irrigation) on cotton root traits and WUE. Results showed that transplanted cotton roots were predominantly concentrated in the 0–30 cm soil layer (75.35–77.13% of total root length), significantly higher than those of direct-seeded cotton (63.10–74.71%). Under micro-spray tape and drip irrigation, the root length density (RLD) of transplanted cotton was 18.55% and 23.46% higher, respectively, than that of direct-seeded cotton, whereas under border irrigation it was 5.09% lower. Transplanted cotton mainly extracted water from the 0–40 cm soil layer (utilization rate: 65.49%), while direct-seeded cotton primarily relied on water from the 20–60 cm layer (53.20%). Although no statistically significant difference in yield was observed between the two cultivation modes, transplanted cotton exhibited a 15.37% higher WUE than direct-seeded cotton. Moreover, surface drip irrigation substantially enhanced WUE, exceeding that under border irrigation and micro-spray tape irrigation by 37.35% and 14.07%, respectively. This study enhances understanding of root traits in transplanted cotton and demonstrates that irrigation methods regulate WUE by modifying root distribution and water uptake patterns.

1. Introduction

The Yellow River Basin is a major grain-producing region and one of China’s three primary cotton-growing areas [1,2]. In recent years, the basin has experienced growing constraints on arable land and water scarcity, especially in its densely populated eastern regions [3,4]. Traditionally, cotton cultivation in this region has relied on either spring cotton or wheat-cotton intercropping systems, resulting in a situation where it is “sufficient for one crop but insufficient for two crops” [5]. While wheat-cotton intercropping allows for annual double cropping of wheat and cotton, it not only reduces the yield of the preceding winter wheat but also causes severe damage to cotton seedlings during wheat harvest, thereby hindering the mechanization of both wheat and cotton production [4,6].

With the rapid development of factory-based seedling raising and mechanized transplanting technologies, the transplanted cotton cultivation model has gradually matured [7,8]. The integrated approach of factory-raised cotton seedlings followed by mechanized field transplanting after wheat harvest effectively alleviates the land-use conflict between grain and cotton, enables dual harvests of crops, and stabilizes cotton cultivation in the region [4,9]. However, transplanted cotton exhibits distinct root morphology compared to direct-seeded cotton, characterized by abnormal root structures. This morphological difference is not an isolated case but a widespread phenomenon, resulting inevitably from changes in rhizocrine levels in the soil in response to environmental stresses and soil volume, which feed back to the roots and trigger adaptive responses [7,10,11,12]. Direct-seeded cotton typically develops a taproot system, with the main root growing vertically downward and lateral roots extending horizontally, forming an inverted conical root length density distribution [10]. In contrast, transplanted cotton experiences main root degradation, forming root knots at the base, and its root system is predominantly composed of well-developed lateral roots that transition from horizontal extension to a claw-like configuration.

Roots play a crucial physiological role in absorbing and transporting soil moisture. As a root-intensive crop, cotton heavily relies on its root system for the accumulation of above-ground biomass and yield [11,12]. Root distribution has a profound impact on the uptake of soil water and nutrients; a well-developed root system enhances resource use efficiency by accessing a wider range of soil moisture and nutrients [13]. Stable isotopes have been widely used in plant water uptake studies as effective tracers to identify the contributions of different water sources. Isotopic signatures differ significantly among soil water, precipitation, irrigation water, and water from root uptake, soil evaporation, and plant transpiration. Hydrogen and oxygen isotopes in water molecules allow for a detailed investigation of root water uptake patterns, facilitating the understanding of how plant roots selectively absorb water in complex soil environments [14,15].

Numerous studies have shown that cotton root growth and water uptake characteristics are jointly influenced by genotype and agronomic practices, with irrigation methods having a particularly significant impact [16,17]. Different irrigation regimes induce distinct soil wetting processes and water redistribution patterns. Roots, due to their extensive surface area interacting with the surrounding soil, are the first organs to respond to changes in the soil microenvironment [14]. Border irrigation, micro-spray tape irrigation, and drip irrigation are the three most common irrigation methods in the Yellow River Basin. Border irrigation, the most widely used, suffers from severe water waste and low irrigation efficiency [18]. Micro-spray tape irrigation has gained popularity due to its ease of assembly, low risk of clogging, and low initial cost [19]. Drip irrigation, a mature water-saving technology in the region, offers high uniformity and water efficiency but has been reported to inhibit deep root penetration [20]. However, current research on the effects of irrigation on root growth and water uptake has predominantly focused on direct-seeded cotton. In transplanted cotton, root disturbance during transplanting inevitably damages fine roots, leading to a reduction in initial hydraulic conductance [7]. During early plant establishment, this physiological constraint may increase the crop’s reliance on frequent and localized irrigation to maintain an adequate water supply. Moreover, given the distinctive root system architecture of transplanted cotton, irrigation methods that primarily wet deeper soil layers, such as border irrigation, may not effectively match root distribution, raising concerns that such approaches could result in water loss and reduced water use efficiency compared with irrigation strategies that maintain moisture in the surface soil layer.

Given the unique root characteristics of transplanted cotton, comparing its claw-like root morphology with the taproot system of direct-seeded cotton to explore root morphological structures and water uptake mechanisms—particularly how roots adapt to environmental changes and enhance WUE under varying irrigation methods—holds significant scientific and practical value. This study aims to: (1) examine the effects of different irrigation methods on the atypical root system of transplanted cotton versus the conventional taproot system of direct-seeded cotton; (2) clarify the water uptake sources of transplanted and direct-seeded cotton under various irrigation regimes; and (3) assess water consumption, yield, and WUE of the two cultivation models under different irrigation treatments.

2. Materials and Methods

2.1. Field Experiments

2.1.1. Experimental Site

Field experiments were conducted from 2021 to 2022 at the Agricultural Efficient Water Test Site of North China University of Water Resources and Hydropower. The site is located in Zhengzhou City, Henan Province, China (34°50′ N, 113°48′ E; elevation 110.4 m). The region has a warm temperate, subhumid monsoon climate, with an average annual temperature of 15.6 °C, a mean annual precipitation of 542.2 mm, a frost-free period of 209 days, and approximately 1869.7 h of sunshine per year. Two cropping cycles are practiced annually. The soil at the experimental site is sandy, with a depth exceeding 5 m and a groundwater table below 5 m. The bulk density of the 0–150 cm soil layer is 1.35 g·cm⁻³, and the field capacity is 24% (volumetric water content).

Table 1. Basic soil fertility parameters at the experimental field.

Soil Depth/cm	PH	Alkaline Hydrolysis Nitrogen (mg·kg−1)	Available Phosphorus (µg·kg−1)	Available Potassium (mg·kg−1)	Organic Carbon Content (g·100 g−1)
0–20	8.5	54.5	15.3	151.8	0.9
20–40	8.6	31.1	5.6	92.4	0.8
40–60	8.6	20.1	4.1	81.3	0.8
60–80	8.7	18.8	3.3	62.3	0.6
80–100	8.9	9.5	2.9	31.5	0.5

summarizes the initial nutrient distribution in the 0–100 cm soil layer prior to the experiment. Changes in rainfall and reference evapotranspiration (ET₀) during the study period are shown in Figure 1.

2.1.2. Experimental Design

To thoroughly compare the claw-like root morphology of transplanted cotton, direct-seeded cotton was used as a control. A split-plot design was implemented with two factors: planting method and irrigation method. The planting method served as the main plot factor, comprising two modes: direct seeding (P1) and transplanting (P2). The irrigation method served as the subplot factor, consisting of three treatments: border irrigation (I1), micro-spray tape irrigation (I2), and surface drip irrigation (I3). In total, six treatment combinations were tested, each with three replicates, resulting in 18 experimental plots. Each plot measured 160 m² (20 m × 8 m), separated by 0.5 m water barriers.

For direct-seeded cotton, a “three-in-one” planting pattern was adopted (Figure 2a). The cotton cultivar used in this study was the transgenic insect-resistant, short-season variety “Zhongmian 50”. Cotton rows were reserved during wheat sowing, and seeds were sown in these rows around 1 May using a mechanical seeder. After emergence, seedlings were manually thinned to maintain a 20 cm plant spacing. For transplanted cotton, a “greenhouse seedling and field transplanting” model was used (Figure 2b). Cotton seedlings were raised in plastic seedling trays prior to transplanting, with each cell having an upper opening width of 3.5 cm and a height of 4.0 cm. Seedlings were raised in a greenhouse from approximately 1 May, and following wheat harvest (around 10 June), a transplanter was used to transfer seedlings to the field, maintaining a 20 cm spacing.

The preceding wheat crop and the subsequent cotton crop received the same irrigation methods. Technical specifications for each irrigation method are presented in

Table 2. Technical parameters of different irrigation methods.

Irrigation Method	Design Specifications	Field Slope	Irrigation Flow	Equipment Specifications
Border irrigation	Width: 2.1 m; Length: 20 m; 3 rows of cotton per border	0.002	Single-width inlet flow rate: 4.0 L·s−1·m−1	—
Micro-spray tape irrigation	Laying length: 20 m; Strip spacing: 1.4 m	—	Micro-spray tape flow rate: 0.165 m3·h−1·m−1	hole belt-type sprinkler tape with a 40 mm outer diameter; each group has 7 inclined water holes spaced 0.3 m apart; working pressure: 0.3 MPa; spray width: 4 m
Surface drip irrigation	Laying length: 20 m; Pipe spacing: 0.7 m	—	Drip head flow: 2.0 L·h−1	hole drip irrigation tape with a 16 mm outer diameter; dripper spacing: 0.2 m; working pressure: 0.1 MPa

. Micro-spray tapes were placed alternately between cotton rows with 140 cm spacing, while drip irrigation tapes were laid directly on the cotton rows so that each plant had an individual dripper. Irrigation timing was governed by soil moisture levels: irrigation was applied when root-zone soil moisture during the seedling, bud, and flowering–boll stages fell to 70 ± 2% of field capacity. A water meter was used to record irrigation volumes. To prevent the formation of late-season bolls, reduce loss of photosynthates, and ensure proper boll maturation, irrigation was terminated during the boll-opening stage. Therefore, the planting density was the same for both direct-seeded and transplanted cotton, at approximately 71,000 plants ha⁻¹.

Before direct seeding or transplanting, a basal fertilizer was applied at rates of 105 kg·ha⁻¹ P₂O₅, 45 kg·ha⁻¹ K₂O, and 30% of the total nitrogen requirement (300 kg·ha⁻¹). The remaining 70% of nitrogen was applied at the bud stage [21,22,23]. All other agronomic practices—including topping, weeding, and pest management—were performed uniformly across treatments following local high-yielding cultivation practices.

2.2. Measurement Methods

2.2.1. Root Morphology Distribution

Soil profile sampling was used to assess the overall characteristics and spatial distribution of cotton roots at the seedling, bud, and flowering–boll stages [24]. A soil profile measuring 70 cm in depth and 40 cm in width was excavated in the field. A 10 cm × 10 cm × 10 cm iron sampling box was vertically inserted into the soil profile to collect root samples. Three replicate samples were taken for each treatment [25,26]. Soil samples were soaked in water for 6–8 h and then washed through a 0.1 mm nylon sieve to extract roots with diameters < 2.5 mm. Extracted roots were arranged on a glass plate and scanned using a dual-light-source scanner (Epson Perfection V700 PHOTO, Seiko Epson Co., Suwa, Nagano, Japan). Roots were then oven-dried at 80 °C to constant weight to determine root biomass. The scanned images were analyzed using WinRHIZO Pro software (version 2007b; Regent Instruments Inc., Quebec City, QC, Canada) to obtain root morphological parameters, including root length, diameter, surface area, and volume.

2.2.2. Dynamic Growth of the Root System

Root growth dynamics and branching characteristics were monitored every 5–8 days using rhizotrons. The root observation system (ET-100, Bartz Technology LLC, Carpinteria, CA, USA) was employed. A 200 cm micro-root tube was installed at a 45° angle, both within the cotton row and 35 cm away from it [27,28]. Three replicates were established for each treatment. During each observation, the camera was moved downward in 1 cm increments to capture root images, and the corresponding root depth was recorded. The images were analyzed using WinRhizo Tron software (version 2020a; Regent Instruments Inc., Quebec City, QC, Canada) to quantify key root parameters, including root length, diameter, surface area, angle, and the number of root tips. By distinguishing live from dead roots, root growth rate and mortality rate were also calculated.

2.2.3. Soil Water Content

Soil moisture dynamics were monitored using the ECH2O soil moisture monitoring system (EC-5, Decagon Devices, Inc., Pullman, WA, USA). Probes were installed directly on the cotton rows and at distances of 20 cm and 35 cm from the rows. Monitoring depths included 5, 15, 25, 35, 45, 55, and 65 cm. Soil moisture data from the Row and Inter-row positions were recorded separately and analyzed accordingly using the EM50 data logger, with measurements taken automatically every 30 min.

In addition, soil water content was measured using the TRIME system (TRIME-PICO-IPH, IMKO Micromodultechnik GmbH, Ettlingen, Germany) at depths from 0 to 100 cm, with each 20 cm interval treated as one measurement layer. TRIME measurements were calibrated through the soil drying method, and calibration parameters were loaded into the instrument using TRIME WinCAL software (IMKO Micromodultechnik GmbH, Ettlingen, Germany). TRIME access tubes were installed directly on the cotton rows and 35 cm away from the rows. Data were collected every 3–7 days and averaged across the two access tubes. When the soil moisture in the root-active layer approached 70% of field capacity, measurement frequency was increased to ensure timely irrigation [29,30].

2.2.4. Collection, Determination, and Calculation of Isotope Samples

Hydrogen and oxygen isotope analyses were conducted on precipitation water, soil water, and plant stem water. Sample collection procedures during the cotton growing season were as follows:

Precipitation Sampling

Precipitation was collected using a rain gauge. After each rainfall event, water samples were transferred to 10 mL centrifuge tubes, sealed with Parafilm, and immediately stored under refrigeration.

Soil Sampling

Soil samples were collected concurrently with root sampling. Three plots were randomly selected under each treatment. Samples were taken at depth intervals of 0–5, 5–10, 10–15, 15–20, 20–30, 30–40, 40–60, 60–80, 80–100, 100–120, and 120–150 cm. All soil samples were stored at −20 °C prior to laboratory extraction [31].

Plant Stem Sampling

Plant stems adjacent to soil sampling sites were collected at the same time. Stem segments (5–8 cm) were taken 3–5 cm above the soil surface. To avoid isotopic fractionation and contamination from enriched surface water, the outer bark was removed immediately during sampling. Samples were wrapped in plastic film, placed in sealed bags, and stored at −20 °C.

Extraction of Plant Stem Water and Soil Water

Plant stem water and soil water were extracted using a low-temperature vacuum extraction system (LI-2100, LICA, Beijing, China) for δD and δ¹⁸O determination. Extraction time ranged from 1.5 to 3 h, depending on sample water content and batch size, with extraction efficiency reaching up to 98%. To avoid interference from impurities and organic matter during isotope analysis, all extracted samples were filtered through a 0.22 μm organic membrane filter and stored at 4 °C until analysis.

Determination of δD and δ¹⁸O in Water Samples

Hydrogen and oxygen isotope compositions were analyzed using isotope ratio mass spectrometry (IRMS). Samples were injected into an elemental analyzer (Flash 2000 HT, Thermo Fisher Scientific, Bremen, Germany) for high-temperature pyrolysis, and isotopic ratios were measured using an isotope ratio mass spectrometer (Delta V Advantage, Thermo Fisher Scientific, Bremen, Germany). Results were calibrated to the Vienna Standard Mean Ocean Water (V-SMOW) standard using the equation:

$$\mathsf{\delta }\left(‰\right)={\displaystyle \frac{R_{\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}}-R_{\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{a}\mathrm{r}\mathrm{d}}}{R_{\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{a}\mathrm{r}\mathrm{d}}}}\times 1000$$

(1)

where δ represents the isotope composition (‰), and R_sample and R_standard denote the ratios of heavy to light isotopes (¹⁸O/¹⁶O or D/H) in the sample and standard, respectively.

Calculation of Hydrogen and Oxygen Isotopes

The proportional contributions of different soil water layers to plant water uptake were estimated using the Bayesian mixing model MixSIAR (version 3.1.7) [31].

2.2.5. Yield Measurement

Within each replicate plot, a 5 m × 1.4 m area was randomly selected for yield measurement. Harvesting was performed manually in two rounds. The number of open bolls per plant was recorded during harvest. After collection, seed cotton was air-dried and weighed using an electronic balance with ±0.01 g precision to determine seed cotton yield. Samples were then processed using a portable gin to determine lint yield, and lint percentage was calculated accordingly. Additionally, 100 mature bolls were randomly collected before each harvest to determine the mean boll weight.

2.3. Data Processing and Analysis

Experimental data were organized using Microsoft Excel (version 2016, Microsoft Corp., Redmond, WA, USA). Analysis of variance (ANOVA) was performed using SPSS (version 24.0, IBM Corp., Armonk, NY, USA). The Bayesian mixing model MixSIAR (version 3.1.7) was used to evaluate plant root water uptake sources. Figures and charts were produced using Origin (version 2021, OriginLab Corp., Northampton, MA, USA).

3. Results

3.1. Differences in the Two-Dimensional Distribution of Root Length Density (RLD) Between Direct-Seeded Cotton and Transplanted Cotton

3.1.1. Two-Dimensional Distribution of RLD Under Border Irrigation

Figure 3 shows the two-dimensional distribution of RLD for direct-seeded and transplanted cotton under border irrigation. The results indicate that the roots of transplanted cotton were primarily concentrated in the 0–30 cm soil layer, accounting for 77.13% of the total root length, whereas direct-seeded cotton exhibited only 63.10% within this layer. Although both planting methods exhibited the highest RLD in the 0–10 cm layer, transplanted cotton reached a maximum RLD of 0.64 cm·cm⁻³, compared with 0.52 cm·cm⁻³ for direct-seeded cotton, indicating a stronger concentration of shallow roots. Within the 0–30 cm soil layer, the average RLD of direct-seeded cotton was 13.80% lower than that of transplanted cotton. In contrast, in the 30–70 cm soil layer, direct-seeded cotton exhibited an average RLD that was 70.00% higher than transplanted cotton. Overall, across the 0–70 cm profile, direct-seeded cotton displayed a slightly higher average RLD (0.30 cm·cm⁻³) than transplanted cotton (0.28 cm·cm⁻³). These findings suggest that, under border irrigation, direct-seeded cotton tends to distribute more roots in the deeper soil layers (30–70 cm), whereas transplanted cotton roots are more concentrated in the shallow layers (0–30 cm).

3.1.2. Two-Dimensional Distribution of RLD Under Micro-Spray Tape Irrigation

Figure 4 presents the two-dimensional distribution of RLD for direct-seeded and transplanted cotton under micro-spray tape irrigation. Under this irrigation method, transplanted cotton exhibited an RLD that was 18.55% higher than that of direct-seeded cotton. The overall root distribution pattern under micro-spray tape irrigation was similar to that under border irrigation; however, transplanted cotton showed a more pronounced accumulation of roots in the shallow soil layer (0–30 cm). Specifically, the average RLD of direct-seeded cotton in the 0–30 cm soil layer was 26.42% lower than that of transplanted cotton. In contrast, the difference between the two planting methods became smaller in the 30–70 cm soil layer. These results indicate that, relative to border irrigation, micro-spray tape irrigation further enhances the RLD of transplanted cotton, particularly in the shallow soil layer.

3.1.3. Two-Dimensional Distribution of RLD of Direct-Seeded and Transplanted Cotton Under Surface Drip Irrigation

Figure 5 shows the two-dimensional distribution of RLD for direct-seeded and transplanted cotton under surface drip irrigation. Under this irrigation method, the RLD of transplanted cotton was 23.46% higher than that of direct-seeded cotton. Both planting methods exhibited their highest average RLDs under surface drip irrigation compared with border and micro-spray tape irrigation, suggesting that drip irrigation is the most conducive to root development. For transplanted cotton, the average RLD under surface drip irrigation increased by 31.33% and 20.15% relative to border and micro-spray tape irrigation, respectively. For direct-seeded cotton, the increases were 0.96% and 15.37%, respectively. These results demonstrate that surface drip irrigation is especially favorable for the root growth of transplanted cotton.

3.2. Analysis of Variance of Root Morphological Properties in Direct-Seeded and Transplanted Cotton

Table 3. Differences in total root length, total root surface area, total root volume, and average root diameter of direct-seeded and transplanted cotton under different irrigation methods ¹.

Year	Treatment		Total Root Length (mm)	Total Root Surface Area (mm2)	Total Root Volume (mm3)	Average Root Diameter (mm)
2021	P1	I1	71 ± 4.92 d	7.44 ± 0.94 d	0.27 ± 0.05 d	0.44 ± 0.02 e
		I2	83.89 ± 5.04 bcd	10.66 ± 0.68 abcd	0.48 ± 0.17 bcd	0.51 ± 0.03 bcd
		I3	100.96 ± 8.92 abcd	11.55 ± 2.17 abc	0.65 ± 0.11 abc	0.57 ± 0.02 abc
	P2	I1	88.29 ± 8.67 bcd	8.74 ± 0.51 bcd	0.38 ± 0.05 cd	0.46 ± 0.02 de
		I2	110.95 ± 9.02 ab	11.07 ± 0.96 abc	0.69 ± 0.06 ab	0.55 ± 0.02 abc
		I3	125.97 ± 6.29 a	12.92 ± 0.28 ab	0.78 ± 0.04 a	0.56 ± 0.07 abc
2022	P1	I1	79.14 ± 3.3 cd	8.13 ± 0.37 cd	0.37 ± 0.04 cd	0.46 ± 0.04 de
		I2	96.77 ± 11.66 abcd	11.14 ± 0.36 bcd	0.58 ± 0.08 abcd	0.51 ± 0.04 bcd
		I3	106.18 ± 17.58 abc	11.21 ± 1.23 abc	0.7 ± 0.05 ab	0.58 ± 0.06 ab
	P2	I1	81.4 ± 3.83 cd	8.67 ± 1.22 bcd	0.48 ± 0.05 bcd	0.48 ± 0.04 cd
		I2	100.17 ± 9.69 abcd	10.55 ± 0.79 abcd	0.59 ± 0.05 abcd	0.55 ± 0.02 abc
		I3	124.21 ± 7.62 a	14.18 ± 1.55 a	0.78 ± 0.03 a	0.6 ± 0.03 a
F-Values
Y			NS	NS	NS	NS
P			27.413 **	7.889 **	19.430 **	NS
I			44.915 **	47.980 **	68.548 **	32.512 **
Y × P			6.608 **	NS	NS	NS
Y × I			NS	NS	NS	NS
P × I			NS	NS	NS	NS
Y × P × I			NS	NS	NS	NS

¹: For a given trait, treatments with the same letter within a year are not significantly different based on one-way ANOVA followed by LSD multiple range test at p < 0.05. ** indicate significant differences at the 1% level; NS indicates no significant difference at the 5% level. P1 and P2 represent direct-seeded and transplanted planting modes, respectively. I1, I2, and I3 represent border irrigation, micro-spray tape irrigation, and surface drip irrigation, respectively.

presents the variance analysis of total root length, total root surface area, total root volume, and average root diameter of direct-seeded and transplanted cotton at the bud stage under different irrigation methods, as determined by the soil profile sampling method. The data indicate that the planting method significantly influences total root length, total root surface area, and total root volume. Specifically, transplanted cotton showed increases of 17.29%, 9.95%, and 21.26% in total root length, total root surface area, and total root volume, respectively, compared to direct-seeded cotton. However, there were no significant differences in the average root diameter between transplanted and direct-seeded cotton.

Irrigation methods also had a significant impact on total root length, total root surface area, total root volume, and average root diameter. Cotton grown under surface drip irrigation (I3) exhibited the highest values for all root indices, while the lowest values were recorded under border irrigation (I1). Compared with border irrigation (I1) and micro-spray tape irrigation (I2), surface drip irrigation (I3) led to increases of 42.99% and 16.73% in total root length, respectively. Similarly, total root surface area increased by 51.22% and 14.83%, total root volume increased by 92.92% and 23.69%, and average root diameter increased by 25.59% and 8.46%, respectively, under surface drip irrigation.

No significant differences in root system parameters were observed between direct-seeded and transplanted cotton across different years. Furthermore, the interaction effects of year, planting method, and irrigation method on the root system parameters were not statistically significant.

Additional root distribution data for 2021 and 2022 are provided in the Supplementary Material (Figures S1–S4).

3.3. Analysis of Root Water Uptake Sources Based on the MixSIAR Model

The MixSIAR Bayesian isotope mixing model was used to quantitatively calculate the contribution proportions of soil water from different depths to crop water uptake for direct-seeded and transplanted cotton under various irrigation methods during the 2021 and 2022 growing seasons. The model also identified the primary root water uptake depths for both direct-seeded and transplanted cotton. As shown in Figure 6 and Figure 7, the contribution proportion of soil water fluctuated significantly during cotton growth. As the growth period progressed, the main water uptake layer of cotton roots gradually shifted from the surface layer (0–20 cm) to the middle layer (20–100 cm), and the contribution of deep soil water (100–150 cm) to root water uptake also increased over time.

Between different planting methods, the contribution rates of soil water from the 0–20 cm, 20–40 cm, 40–60 cm, 60–100 cm, and 100–150 cm layers to root water uptake were 22.32%, 26.01%, 27.19%, 12.23%, and 12.26%, respectively, for direct-seeded cotton, while those for transplanted cotton were 36.81%, 28.68%, 16.40%, 10.81%, and 7.30%. Transplanted cotton primarily utilized water from the 0–40 cm layer, with a utilization rate of 65.49%, whereas direct-seeded cotton reached only 48.32%. Direct-seeded cotton mainly utilized water from the 20–60 cm layer, with a utilization rate of 53.20%, while transplanted cotton reached only 45.08%. Although the water uptake depth for both direct-seeded and transplanted cotton gradually shifted downward with growth, the downward trend of water uptake depth for direct-seeded cotton was more rapid by comparison.

In terms of irrigation methods, under border irrigation, micro-spray tape irrigation, and surface drip irrigation, the proportion of soil water from the 0–40 cm layer in cotton root water uptake increased sequentially, reaching 52.23%, 56.83%, and 61.66%, respectively. Notably, for transplanted cotton under surface drip irrigation, the proportion of soil water from the 0–40 cm layer was as high as 66.01%. Even under border irrigation, transplanted cotton derived only 7.20% of its water uptake from the 100–150 cm soil layer, indicating that transplanted cotton failed to effectively utilize deep soil water reserves despite the presence of abundant deep-profile moisture.

Between different years, compared with 2022, the water utilization rate of surface soil water (0–20 cm) in 2021 increased by 5.99%, and the variation in contribution rates of soil water from different layers to root water uptake was more pronounced. The absorption and utilization of surface soil water by roots exhibited a pattern of two peaks and two troughs, reaching the lowest level during the boll-opening stage.

3.4. Differences in Water Consumption, Yield, and WUE Between Direct-Seeded and Transplanted Cotton

3.4.1. Differences in Water Consumption Between Direct-Seeded and Transplanted Cotton

Table 4. Differences in water consumption between direct-seeded and transplanted cotton under different irrigation methods ¹.

Year	Treatment		Seedling Stage (mm)	Bud Stage (mm)	Flowering-Boll Stage (mm)	Boll-Opening Stage (mm)	Whole Growth Period (mm)
2021	P1	I1	113.18 ± 1.21 b	101.3 ± 0.57 c	204.92 ± 1.04 e	113.23 ± 0.41 c	532.64 ± 1.15 c
		I2	114.96 ± 1.08 a	96.7 ± 0.55 e	207.6 ± 1.07 d	115.72 ± 0.2 b	534.99 ± 1.8 c
		I3	103.69 ± 1.82 e	87.03 ± 0.05 g	206.12 ± 1.09 d	101.04 ± 1.21 e	497.88 ± 4.07 e
	P2	I1	102.73 ± 2.07 e	78.28 ± 0.52 j	187.84 ± 0.09 g	102.3 ± 1.09 e	471.14 ± 2.55 h
		I2	96.27 ± 2.24 g	81.51 ± 0.31 i	181.34 ± 0.31 i	95.8 ± 1.43 f	454.92 ± 3.06 i
		I3	81.26 ± 3.11 i	85.29 ± 0.16 h	184.68 ± 0.12 g	91.07 ± 2.06 h	442.3 ± 5.12 k
2022	P1	I1	115.21 ± 1.9 a	111.79 ± 2.39 a	228.38 ± 0.4 a	116.99 ± 1.93 a	572.37 ± 6.61 a
		I2	107.01 ± 2.49 c	107.12 ± 2 b	218.47 ± 0.23 b	108.03 ± 1.28 d	540.64 ± 2 b
		I3	105.19 ± 2.38 d	99.94 ± 2.01 c	215.57 ± 0.42 c	93.22 ± 0.24 g	513.92 ± 0.22 d
	P2	I1	102.37 ± 2.74 e	100.49 ± 1.61 c	206.59 ± 0.97 d	82.35 ± 0.42 j	491.79 ± 1.69 f
		I2	99.24 ± 2.91 f	98.66 ± 1.26 d	198.44 ± 1.51 f	88.13 ± 0.1 i	484.47 ± 0.16 g
		I3	94.3 ± 3.2 h	89.49 ± 1.17 f	185.44 ± 2.47 h	78.38 ± 0.77 k	447.61 ± 7.61 j
F-Values
Y			5.701 *	886.305 **	1466.549 **	530.110 **	238.988 **
P			311.301 **	728.640 **	4244.305 **	2375.121 **	2799.218 **
I			84.603 **	108.298 **	223.210 **	449.783 **	378.13 **
Y × P			18.178 **	14.098 **	11.692 **	159.518 **	NS
Y × I			13.332 **	28.073 **	174.968 **	4.473 *	20.530 **
P × I			3.556 *	54.399 **	27.723 **	67.465 **	5.653 *
Y × P × I			8.377 **	50.376 **	40.61 **	92.088 **	27.227 **

¹: Since the water consumption of cotton seedlings before wheat harvest is very low, the water consumption of cotton during the seedling stage and the whole growth period in the table does not include the water consumption before wheat harvest. Seedling stage: 10 June to 30 June; Bud stage: 1 July to 25 July; Flowering-boll stage: 26 July to 5 September; Boll-opening stage: 6 September to 15 October. For a given trait, treatments with the same letter within a year are not significantly different based on one-way ANOVA followed by LSD multiple range test at p < 0.05. * and ** indicate significant differences at the 5% and 1% levels, respectively; NS indicates no significant difference at the 5% level. P1 and P2 represent direct-seeded and transplanted planting modes, respectively. I1, I2, and I3 represent border irrigation, micro-spray tape irrigation, and surface drip irrigation, respectively.

summarizes the water consumption of direct-seeded and transplanted cotton at different growth stages and over the entire growing season under various irrigation regimes during 2021–2022. Across planting patterns, the total water consumption of direct-seeded cotton ranged from 497.88 mm (P1I3 in 2021) to 572.37 mm (P1I1 in 2022), whereas that of transplanted cotton ranged from 442.30 mm (P1I3 in 2021) to 491.79 mm (P1I1 in 2022). Overall, direct-seeded cotton consumed 12.54% more water than transplanted cotton. For both planting patterns, the flowering–boll stage accounted for the largest share of seasonal water consumption (38.47–42.01%).

Regarding irrigation regimes, surface drip irrigation markedly reduced cotton water consumption. Relative to border irrigation, surface drip irrigation reduced water use during the seedling, bud, flowering–boll, and boll-opening stages and over the full growth period by 12.76%, 8.32%, 4.54%, 14.06%, and 8.74%, respectively. Compared with micro-spray tape irrigation, the corresponding reductions were 8.59%, 6.15%, 1.78%, 12.09%, and 5.96%.

3.4.2. Differences in Yield and WUE Between Direct-Seeded and Transplanted Cotton

The effects of planting patterns and irrigation methods on cotton yield and WUE are presented in

Table 5. Differences in yield and water use efficiency (WUE) between direct-seeded and transplanted cotton under different irrigation methods ¹.

Year	Treatment		Number of Bolls Per Plant	Boll Mass (g)	Seed Cotton Yield (kg ha−1)	Lint Percentage (%)	WUE (kg m−3)
2021	P1	I1	9 ± 1 bcd	5.11 ± 0.62 cd	3164.37 ± 304.55 de	36.48 ± 1 cd	0.59 ± 0.06 fg
		I2	9 ± 2 bcd	5.11 ± 0.23 cd	3500.89 ± 166.8 cd	36.31 ± 0.17 cd	0.65 ± 0.03 ef
		I3	11 ± 1 a	6.03 ± 0.35 ab	3851.96 ± 108.4 ab	36.88 ± 1.49 bcd	0.77 ± 0.03 bc
	P2	I1	7.67 ± 0.58 d	4.95 ± 0.35 d	3202.77 ± 157.24 de	36.1 ± 0.68 d	0.68 ± 0.03 de
		I2	9 ± 1 bcd	5.81 ± 0.39 abc	3536.88 ± 349.9 bcd	37.51 ± 1.05 ab	0.78 ± 0.08 bc
		I3	10 ± 1 abc	6.07 ± 0.35 a	3828.87 ± 59.37 ab	37.61 ± 0.88 a	0.87 ± 0.01 ab
2022	P1	I1	10 ± 1 abc	5.44 ± 0.68 bcd	3000.55 ± 196.18 ef	37.51 ± 0.63 ab	0.52 ± 0.03 g
		I2	9 ± 1 bcd	5.27 ± 0.5 cd	3564.42 ± 111.82 bcd	36.59 ± 0.25 bcd	0.66 ± 0.02 ef
		I3	10.33 ± 0.58 ab	6.08 ± 0.55 a	3831.3 ± 89.79 ab	37.25 ± 0.76 abc	0.75 ± 0.02 cd
	P2	I1	7.33 ± 1.53 d	5.22 ± 0.51 cd	2855.16 ± 162.45 f	37.26 ± 0.46 abc	0.58 ± 0.03 fg
		I2	9 ± 1 bcd	5.95 ± 0.51 ab	3745.02 ± 134 abc	36.74 ± 0.51 bcd	0.77 ± 0.03 bc
		I3	10 ± 1 abc	5.78 ± 0.34 abc	3947.3 ± 53.21 a	37.39 ± 0.12 ab	0.88 ± 0.02 a
F-Values
Y			NS	NS	NS	NS	NS
P			NS	NS	NS	NS	69.112 **
I			8.622 **	9.108 **	62.529 **	NS	111.283 **
Y × P			NS	NS	NS	NS	NS
Y × I			NS	NS	3.91 *	NS	5.031 *
P × I			NS	NS	NS	NS	NS
Y × P × I			NS	NS	NS	NS	NS

¹: For a given trait, treatments with the same letter within a year are not significantly different based on one-way ANOVA followed by LSD multiple range test at p < 0.05. * and ** indicate significant differences at the 5% and 1% levels, respectively; NS indicates no significant difference at the 5% level. P1 and P2 represent direct-seeded and transplanted planting modes, respectively. I1, I2, and I3 represent border irrigation, micro-spray tape irrigation, and surface drip irrigation, respectively.

. Planting pattern exerted no significant influence on boll number per plant, boll mass, seed cotton yield, or lint percentage. However, WUE differed significantly between planting patterns, with transplanted cotton exhibiting a 15.37% higher WUE than direct-seeded cotton.

Irrigation method significantly affected boll number per plant, boll mass, and seed cotton yield. Compared with border irrigation and micro-spray tape irrigation, surface drip irrigation increased boll number per plant by 21.57% and 14.81%, boll mass by 15.58% and 8.21%, and seed cotton yield by 26.48% and 7.75%, respectively. Lint percentage did not vary significantly among irrigation methods. Surface drip irrigation also significantly improved WUE, with values 37.35% and 14.07% higher than those under border irrigation and micro-spray tape irrigation, respectively.

When considering the combined effects of planting pattern and irrigation method, transplanted cotton under surface drip irrigation achieved the highest yield and WUE, reaching 3888.09 kg ha⁻¹ and 0.87 kg m⁻³, respectively. In contrast, transplanted cotton under border irrigation produced the lowest yield (3028.96 kg ha⁻¹), whereas direct-seeded cotton under border irrigation recorded the lowest WUE (0.56 kg m⁻³).

4. Discussion

4.1. Root Growth and Water Uptake Sources of Transplanted and Direct-Seeded Cotton Under Different Irrigation Methods

Different irrigation methods have significant effects on the RLD of both direct-seeded and transplanted cotton. Surface drip irrigation promotes root growth in both cotton types, benefiting from its water-saving characteristics. The reduced irrigation quota helps avoid soil compaction, which often occurs with border irrigation, thus improving soil aeration and root respiration [17,32]. Additionally, surface drip irrigation ensures high water distribution uniformity and precise control, reducing the risk of root damage or drought stress, thereby creating a favorable soil microenvironment for root development [16,24]. Surface drip irrigation also minimizes deep percolation, allowing soil nutrients to remain in the active root zone, thus improving water and nutrient use efficiency [33,34]. For transplanted cotton, whose root system is primarily distributed in the 0–30 cm soil layer, surface drip irrigation can maintain high humidity in this zone, significantly enhancing its water and nutrient absorption efficiency.

Both planting patterns and irrigation methods influence the water uptake dynamics of cotton root systems. Under the same irrigation conditions, direct-seeded cotton tends to extract a greater proportion of soil water from deeper layers compared with transplanted cotton (Figure 6 and Figure 7). Previous studies have reported that root water uptake is closely associated with root morphological traits (e.g., root dry mass, root length). Therefore, the differences in the contribution rates of soil water from different depths are partly attributable to variations in root system development between the two planting patterns [14,35]. Direct-seeded cotton exhibited a greater reliance on soil water from the 20–60 cm layer [36]. The soil water in the 20–60 cm layer plays a decisive role in determining seed cotton yield [37]. However, some studies have also shown that the principal root water uptake zone of cotton was located in the 40–80 cm soil layer [38]. This discrepancy is largely attributable to the distinct climatic conditions of the experimental site in Xinjiang, where high evaporative demand leads to drier surface soil, thereby causing the primary root water uptake zone to shift downward. In contrast, transplanted cotton preferentially utilized water from the shallower 0–40 cm soil layer, a pattern that corresponds well with its root distribution (Figure 3, Figure 4 and Figure 5).

Regarding irrigation methods, different irrigation modes exert pronounced effects on the vertical sources of root water uptake in cotton. As a precise and highly efficient irrigation technique, surface drip irrigation substantially enhances soil water use efficiency. For transplanted cotton in particular, surface drip irrigation maintains relatively high moisture levels within the 0–40 cm soil layer, concentrating soil water within the active root zone [25,39]. This condition facilitates water absorption by roots in this layer and consequently improves water use efficiency. In contrast, although border irrigation and micro-sprinkler belt irrigation can also supply adequate water, their higher irrigation quotas and longer irrigation intervals lead to greater fluctuations in soil moisture within the 0–40 cm layer, which is unfavorable for water uptake by transplanted cotton roots. Therefore, surface drip irrigation offers clear advantages in enhancing both the depth and efficiency of water absorption in transplanted cotton [25].

Moreover, surface drip irrigation also exerted a positive influence on the water uptake patterns of direct-seeded cotton. Although the root system of direct-seeded cotton is capable of accessing deeper soil water, plants generally prioritize the use of shallow soil moisture due to the shorter transport pathway [40,41]. Surface drip irrigation, through its precise regulation of water supply, enhanced the utilization of soil water in the 0–30 cm layer, thereby improving the capacity of roots to absorb water from the shallow soil profile [10].

In 2022, the water uptake depth across all treatments initially became shallower and subsequently deepened. This pattern was likely attributable to the proliferation of newly developed fine roots in the upper soil layer during the squaring stage, which increased shallow water uptake. As the crop progressed into the flowering and boll-setting stages, continuous root penetration into deeper layers enhanced the contribution of subsoil water to total root water uptake [38]. Compared with 2022, the contribution of surface soil water (0–20 cm) to root water uptake increased in 2021, and the fluctuations in the contribution rates from different soil layers were more pronounced. This response was likely due to the relatively higher rainfall in 2021, which altered the soil moisture distribution across layers and consequently modified the patterns of water absorption by cotton roots [42,43].

4.2. Water Consumption, Yield, and Water Use Efficiency Characteristics of Transplanted and Direct-Seeded Cotton Under Different Irrigation Methods

In this study, the total water consumption of direct-seeded cotton was significantly higher than that of transplanted cotton. This difference is primarily attributable to root injury incurred during the transplanting process [40,44]. Following transplanting, cotton undergoes a recovery period during which plant height and leaf area expand slowly, resulting in relatively low water consumption; hence, transpiration and overall water use remain lower in transplanted cotton. Although the compensatory proliferation of lateral roots at later growth stages led to greater overall root length density in transplanted cotton compared with direct-seeded cotton (Figure S2), its plant height and leaf area remained consistently lower, resulting in reduced transpiration and consequently lower water use. Despite the consistently superior vegetative growth of direct-seeded cotton—with greater plant height and leaf area—its yield was only slightly higher than that of transplanted cotton, and the difference was not statistically significant (Table 5). This may be attributed to the more favorable plant architecture of transplanted cotton, which provides an optimal canopy temperature and ventilation environment. Such conditions help prevent excessive vegetative growth that could suppress reproductive development. As a result, transplanted cotton exhibits more rapid reproductive growth, lower bud and boll abscission rates, and higher boll-setting efficiency [45]. Although transplanted cotton does not exhibit a yield advantage relative to direct-seeded cotton, the overall economic return of the post-wheat transplanting system is substantially higher than that of the wheat–cotton relay intercropping system. This advantage arises because, in the post-wheat transplanting system, the preceding winter wheat crop does not require reserved cotton rows, allowing for significantly higher wheat planting density and consequently higher wheat yields [23,46]. Moreover, the transplanting system facilitates mechanization in both wheat and cotton production, further reducing labor and production costs. However, the relatively shallow root distribution of transplanted cotton also increases its sensitivity to soil moisture fluctuations, making it more susceptible to drought stress (Figure 6 and Figure 7). In fields lacking adequate irrigation infrastructure, the yield of transplanted cotton may therefore be severely constrained.

Differences in irrigation methods exert significant effects on cotton water consumption, yield formation, and water use efficiency. The observed variations can be largely attributed to the distinct soil moisture distribution patterns created by each irrigation method and the consequent differences in root water uptake behavior [16]. Surface drip irrigation consistently maintained a stable moisture environment in the upper soil layers and enhanced the precision of soil water regulation. By ensuring adequate but not excessive water consumption, it also promoted yield formation [17]. Previous studies have also shown that drip irrigation helps regulate cotton plant architecture, promoting the allocation of photosynthates toward reproductive organs and thereby increasing yield [25]. In contrast, border irrigation resulted in substantial water loss through evaporation and deep percolation, leading to the highest overall water consumption in the cotton field but without a corresponding increase in yield. The findings of the present study differ from those reported by [47], who observed that border irrigation produced higher yields than a single drip irrigation event. This discrepancy is largely attributable to the conditions of his experimental site in southern Xinjiang, where soil salinity was high, and only one drip irrigation was applied after sowing. Nonetheless, he still concluded that multiple drip irrigations were more conducive to increasing cotton yield. This issue was especially pronounced for transplanted cotton, for which the 0–40 cm soil layer exhibited the poorest moisture stability under border irrigation, ultimately producing the lowest yield and water use efficiency (Table 5).

Overall, the combination of surface drip irrigation and transplanted cotton appears to be the most effective strategy for maximizing yield and water use efficiency, whereas direct-seeded cotton demonstrates greater adaptability under conditions with less precise irrigation control. Therefore, in the cotton–wheat double-cropping areas of the Yellow River Basin, if economic conditions allow, the cultivation strategy combining surface drip irrigation with transplanted cotton should be regarded as the optimal approach.

However, because the present study evaluated cotton yield and water use efficiency solely from the perspectives of root system development and root water uptake sources, without examining the growth performance of the preceding winter wheat crop, further analysis is required. Integrating winter wheat yield and water use efficiency will be essential for a comprehensive assessment of the overall productivity and water use benefits of the wheat–cotton system.

5. Conclusions

The main conclusions of this study are as follows:

(1) Planting method: Compared with direct-seeded cotton, transplanted cotton exhibited a more concentrated root distribution within the shallow soil layer (0–30 cm). Irrigation method: Both direct-seeded and transplanted cotton achieved the highest total root length, total root surface area, total root volume, and average root diameter under surface drip irrigation, indicating that this irrigation mode is most conducive to root system development.

(2) Planting method: Direct-seeded cotton mainly relies on the moisture in the 20–60 cm soil layer (53.20%), while transplanted cotton mainly absorbs water from the 0–40 cm soil layer (utilization rate: 65.49%). Transplanted cotton absorbs only 7.3% of the water from the soil layer below 100 cm, indicating a low utilization rate of deep soil water. Irrigation method: Under border irrigation, micro-sprinkler belt irrigation, and surface drip irrigation, the proportion of root-absorbed water originating from the 0–40 cm soil layer increased progressively.

(3) Planting method: Transplanted cotton had significantly lower water consumption and significantly higher water use efficiency than direct-seeded cotton, while no significant difference in yield was observed between the two planting methods. Irrigation method: Surface drip irrigation significantly enhanced cotton yield and water use efficiency while reducing water consumption. Among all treatments, transplanted cotton under surface drip irrigation achieved the highest yield and water use efficiency, whereas transplanted cotton under border irrigation exhibited the lowest yield and direct-seeded cotton under border irrigation showed the lowest water use efficiency.