Optimal Drip Irrigation Frequency and Volume Mitigates Photosynthetic Impairment and Balances Yield–Resource Trade–Offs for Spring Maize in Arid Sandy Loam Soils

Abstract

Optimizing irrigation amount and frequency is critical for maize production in arid sandy loam soils, which are characterized by severe water scarcity and high leaching risks. This study aimed to determine the physiological and multi-objective synergistic responses of spring maize to different drip irrigation regimes in the sandy soil region of Southern Xinjiang and to identify the optimal water management strategy. A two-year field experiment evaluated four irrigation amounts (W₁: 350 mm, W₂: 500 mm, W₃: 650 mm, W₄: 800 mm) combined with three irrigation frequencies (F₄: 4-day, F₇: 7-day, F₁₀: 10-day intervals). The results indicated that increasing irrigation amount and frequency generally improved canopy physiological performance (leaf area index, SPAD values, net photosynthetic rates, and photosystem II photochemical efficiency), though the response magnitudes varied among specific physiological parameters and irrigation levels. Grain yield increased with irrigation amount and was highest at 800 mm under F₄ or F₇ frequency; however, excessive water inputs led to reduced water productivity. Conversely, nutrient use efficiencies for nitrogen, phosphorus, and potassium exhibited a parabolic response, reaching their maxima under W₃ (650 mm) and F₇ (7-day interval). To reconcile the trade-offs between maximum productivity and resource conservation, the entropy weight method coupled with TOPSIS (EWM-TOPSIS) was applied. The results showed that grain yield and water productivity were the dominant indicators determining overall system efficacy, with the W₃F₇ treatment consistently achieving the highest relative closeness (over 76%) to the ideal solution in both years. While this optimal recommendation relies on the multi-criteria weighting framework of the EWM-TOPSIS model, integrating a 650 mm irrigation quota with a 7-day interval offers a practical agronomic strategy to balance yield, water productivity, and nutrient efficiency in arid agroecosystems.

1. Introduction

Maize (Zea mays L.) is a globally vital cereal crop, playing an irreplaceable role in ensuring food security and meeting escalating global demand [1,2]. In arid and semi-arid regions, such as the temperate continental climate of Xinjiang, China, agricultural productivity is severely constrained by extreme water scarcity and erratic precipitation [3,4]. Furthermore, the widespread sandy loam soils in these regions are characterized by poor water-holding capacity, making them susceptible to rapid water percolation and nutrient movement below the shallow root layer [5]. Consequently, optimizing agricultural inputs is not only a primary agronomic challenge but also a fundamental necessity for sustainable crop production and environmental protection in arid agroecosystems [6,7].

Drip irrigation under plastic film has been widely adopted to mitigate water scarcity and enhance crop water productivity (WP) [3,8]. The dynamic balance of soil water availability—termed the soil moisture regime—is fundamentally dictated by soil physical properties (particularly soil texture) interacting with management practices [9]. In drip irrigation systems, this regime is governed by three interrelated but distinct parameters: total seasonal irrigation amount, irrigation frequency (interval between applications), and single-irrigation event size (quota). Canopy photosynthetic performance serves as the primary engine for photoassimilate accumulation, yet it is exceptionally sensitive to fluctuations in the soil moisture regime [10,11]. Inappropriate irrigation frequencies or inadequate total volumes trigger severe intermittent drought stress, inducing stomatal closure (Gsw) to limit transpiration loss, which subsequently restricts intercellular CO₂ diffusion and leads to stomatal limitation of photosynthesis [12,13,14]. Under extreme soil drying coupled with high atmospheric evaporative demand, non-stomatal factors—specifically the impairment of photosystem II (PSII) reaction centers and thylakoid membrane structures—further reduce light-harvesting and electron transport efficiencies [15,16,17]. Conversely, while increasing total irrigation amount or frequency can alleviate moisture stress, excessive single-irrigation quotas do not endlessly boost growth; instead, they trigger luxury water consumption and restrict soil aeration, diminishing photosynthetic gains [18,19,20,21]. Therefore, fine-tuning the combination of total irrigation amount, application frequency, and single-event quota is essential to maintain an optimal soil moisture regime that enhances canopy physiological function without causing hypoxic or luxury-consumption penalties.

Fertigation further complicates root-zone dynamics due to the contrasting mobility profiles of macronutrients in soil profiles [22,23]. Nitrogen, predominantly present as highly mobile nitrate (NO₃⁻-N), is easily carried by excessive irrigation or low-frequency high-volume wetting fronts through the 60 cm sandy loam layer, resulting in severe deep leaching and groundwater contamination risks [5,24,25]. In contrast, available phosphorus (AP) and available potassium (AK) exhibit strong spatial conservatism and low mobility, tending to accumulate in the shallow surface soil near emitters [26,27]. This spatial mismatch—where nitrogen escapes below the active root zone while phosphorus and potassium remain trapped in topsoil—severely limits the synergistic uptake of N, P, and K by crops [28,29]. Short-interval, moderate-volume drip regimes can maintain an optimal wetting front, suppressing non-beneficial topsoil evaporation while simultaneously intercepting mobile nitrogen within the primary root zone (0–40 cm) and driving P and K diffusion toward the root surface [30,31,32]. Regulating irrigation timing and quota is thus pivotal for synchronizing root-zone water-nutrient dynamics with crop uptake demand [33,34,35,36,37].

Traditional agricultural evaluations often rely on single-factor optimization—typically pursuing maximum grain yield—which frequently leads to resource wastage, reduced resource use efficiencies, and elevated environmental footprints [38,39,40]. Modern sustainable agriculture demands a multi-objective decision-making approach that simultaneously optimizes grain yield, WP, and multi-element nutrient use efficiencies [41,42]. The entropy weight method combined with the Technique for Order Preference by Similarity to Ideal Solution (EWM-TOPSIS) provides an objective, robust mathematical framework to eliminate subjective weighting bias and resolve complex trade-offs among conflicting agronomic performance metrics [43,44]. Although real-time soil moisture monitoring and evapotranspiration models are widely studied, their practical adoption under film-mulched drip irrigation in hyper-arid regions (e.g., Southern Xinjiang) remains limited by sensor placement sensitivity within localized wetting bulbs and low operational accessibility for local growers. To bridge research with field practice, establishing structured, easily adoptable combinations of irrigation amounts and fixed calendar intervals—grounded in regional farming benchmarks—is vastly more practical.

However, comprehensive studies systematically evaluating how combinations of total irrigation amount and frequency (and their corresponding single-event quotas) regulate aboveground physiological mechanisms (e.g., canopy expansion, leaf gas exchange, and PSII photochemical efficiency) and their coupling with multi-objective optimization (yield, WP, and N/P/K use efficiencies) for spring maize in arid sandy loam soils remain limited. Therefore, a two-year field experiment was conducted to (1) evaluate the physiological response mechanisms, specifically focusing on how irrigation amount and frequency modulate leaf area index, SPAD values, photosynthetic capacity, and PSII photochemical efficiency; (2) elucidate the synergistic effects of irrigation amount and frequency on dry matter accumulation, grain yield, WP, and multi-element (N, P, K) nutrient use efficiencies; and (3) establish a robust multi-objective evaluation framework using the EWM-TOPSIS model to identify the optimal drip irrigation strategy that balances high productivity with enhanced resource use efficiency in arid agroecosystems.

2. Materials and Methods

2.1. Experimental Area

The field experiments were conducted at the Academician Expert Workstation of Modern Agriculture in Aral, Xinjiang, China (40°37′ N, 81°12′ E, 1009 m a.s.l.). The experimental site features a typical temperate continental climate, with a long-term mean annual temperature of 10.7 °C and an average annual precipitation of approximately 50 mm. The soil texture within the 0–60 cm profile is classified as sandy loam, characterized by a bulk density of 1.50 g cm⁻³ and a field capacity of 0.23 cm³ cm⁻³. The soil layer below 60 cm is loamy sand. For the entire 0–100 cm soil profile, the average bulk density and field capacity are 1.53 g cm⁻³ and 0.21 cm³ cm⁻³, respectively. The groundwater table is at a depth greater than 5 m. Prior to the experiment, the initial chemical properties of the topsoil (0–20 cm) were determined as follows: average nitrate-nitrogen (NO₃⁻-N) content of 27.27 mg kg⁻¹, available phosphorus (AP) of 5.46 mg kg⁻¹, and available potassium (AK) of 54.58 mg kg⁻¹. The total precipitation during the experimental periods (maize growing seasons) in 2024 and 2025 was 42.33 mm and 17.60 mm, respectively. The dynamic changes in daily temperature, precipitation, and relative humidity throughout the growing seasons are illustrated in Figure 1.

2.2. Experimental Design and Management Practices

A two-factor randomized complete block design (RCBD) was employed, combining four drip irrigation amounts (W₁: 350 mm, W₂: 500 mm, W₃: 650 mm, W₄: 800 mm) and three drip irrigation frequencies (F₄: 4-day, F₇: 7-day, F₁₀: 10-day intervals), yielding 12 treatment combinations. The experimental field was organized into three spatial blocks (replicates) aligned perpendicular to the field gradient. Each block contained a complete set of all 12 treatments, which were randomly assigned to individual plots within the block using a computer-generated random sequence (Figure S1).

Each experimental plot covered an area of 100 m² (10 m × 10 m). To eliminate lateral soil water and nutrient movement between adjacent treatments, a 1 m wide unirrigated isolation strip was established between neighboring plots, and a 3 m wide buffer zone surrounded the perimeter of the entire experimental site. The field experiment was conducted on the exact same physical plots in both 2024 and 2025 without plot relocation to preserve treatment continuity.

Maize cultivar ‘Xinyu 108’ was planted using a wide–narrow row planting pattern (30 + 80 cm), with a narrow row spacing of 30 cm, a wide row spacing of 80 cm, and a plant spacing of 14.5 cm, resulting in a planting density of 130,000 plants ha⁻¹. A Venturi injector was used for fertigation. The applied fertilizers included urea (N ≥ 45%), monoammonium phosphate (N–P₂O₅ ≥ 12–60%), and agricultural potassium sulfate (K₂O ≥ 50%). The total application rates of nitrogen, phosphorus, and potassium (N–P₂O₅–K₂O) were 400, 200, and 200 kg ha⁻¹, respectively. These seasonal nutrient application rates were established based on local farming surveys and official agronomic recommendations from the Xinjiang Production and Construction Corps (XPCC) for high-yielding drip-irrigated spring maize in Southern Xinjiang, taking into account the high planting density. Fertilizers were applied seven times over the entire growing season. Specifically, 5% of the total fertilizers were applied with irrigation water after seedling emergence. During the jointing, tasseling, and grain-filling stages, fertigation was performed twice per stage, accounting for 30%, 35%, and 30% of the total fertilizer amount, respectively. The experiment utilized inline flat emitter drip tapes with an emitter spacing of 30 cm and a flow rate of 2.6 L h⁻¹. The specific irrigation layout is illustrated in Figure 2.

2.3. Sampling and Data Calculation

2.3.1. Weather Data

An automatic weather station (HOBO U30, United States) recorded wind speed, air temperature, relative humidity, precipitation, and photosynthetically active radiation (PAR, 400–700 nm) at 10 min intervals.

2.3.2. Reference Evapotranspiration (ET₀) and Theoretical Crop Water Requirement (ET_c)

Daily reference crop evapotranspiration (ET₀, mm day⁻¹) was calculated using the standard FAO-56 Penman–Monteith model based on meteorological variables (daily mean air temperature, relative humidity, wind speed at 2 m height, and solar radiation) recorded by the automatic weather station (Section 2.3.1):

$$E{T}_0={\displaystyle \frac{0.408\Delta \left(R_n-G\right)+\gamma \frac{900}{T+273}{u}_2\left(e_s-e_a\right)}{\Delta +\gamma \left(1+0.34u_2\right)}}$$

(1)

where $R_n$ is net radiation (MJ m⁻² day⁻¹), $G$ is soil heat flux density (MJ m⁻² day⁻¹), $T$ is mean daily air temperature (°C), $u_2$ is wind speed at 2 m height (m s⁻¹), $e_s-e_a$ is saturation vapor pressure deficit (kPa), $\Delta$ is the slope of the saturation vapor pressure curve (kPa °C⁻¹), and $\gamma$ is the psychrometric constant (kPa °C⁻¹). Standard crop evapotranspiration ($E{T}_c$, mm) was calculated as $E{T}_c=K_c\times E{T}_0$, where $K_c$ values were set to 0.60 (initial stage), 1.20 (mid-season), and 0.80 (late season) in accordance with FAO-56 guidelines calibrated for drip-irrigated spring maize in Northwest China. Total seasonal theoretical $E{T}_c$ was estimated at 628.58 mm in 2024 and 582.45 mm in 2025, which served as an independent benchmark to compare with the soil water balance results.

2.3.3. Leaf Area Index and Aboveground Dry Matter

Starting from 27 days after emergence of spring maize, plant samples were collected every 10–20 days. Three representative plants were randomly selected from each treatment to determine leaf area and aboveground dry matter. Leaf area was measured by determining the length and maximum width of each fully expanded leaf per plant using a tape measure (0.1 cm precision), and the leaf area index (LAI) was subsequently calculated.

$$LAI=0.75\rho {\displaystyle \frac{{\displaystyle \sum _{i=1}^m{\displaystyle \sum _{j=1}^n{L}_{ij}{W}_{ij}}}}{\mathrm{m}}}$$

(2)

where 0.75 is the correction coefficient for maize leaf area; ρ is the plant density; Lᵢⱼ and Wᵢⱼ are the length and maximum width (cm) of the j-th leaf of the i-th maize plant, respectively; and m is the number of measured plants.

At each sampling event, three plants were randomly selected from each plot and separated from the base of the stem and the underground parts. The various organs (stem, leaf, bract, and ear) were separated, placed in an oven at 105 °C for 1 h for deactivation, and then dried at 75 °C to constant weight. After drying and cooling, the mass was weighed using an electronic balance (0.01 g precision). The average dry matter weight per three plants in each plot was calculated and then multiplied by the planting density to determine the population aboveground dry matter (kg·ha⁻¹).

2.3.4. Leaf SPAD Value

At the jointing, grain-filling, and maturity stages, the SPAD value of the spring maize ear leaf (the top first fully expanded leaf at the jointing stage) was measured at the upper, middle, and lower positions using a SPAD-502Plus portable chlorophyll meter. The average of the three readings was taken as one data point, and each treatment was replicated three times.

2.3.5. Gas Exchange Parameters and Chlorophyll Fluorescence Parameters

On clear, sunny days during the jointing, grain-filling, and maturity stages of spring maize, leaf gas exchange parameters—including net photosynthetic rate (P_n), transpiration rate (T_r), and stomatal conductance (G_sw)—were measured between 11:00 a.m. and 12:00 p.m. using a LI-6400 portable photosynthesis system (LI-COR Inc., Lincoln, NE, USA). Measurements were conducted on the ear leaf (or the uppermost fully expanded leaf at the jointing stage) with three replicates per treatment. To minimize environmental fluctuations and ensure comparability among treatments, chamber microclimatic conditions were standardly controlled across all measurements: photosynthetically active radiation (PAR) was set to a saturating light intensity of 1400 μmol m⁻² s⁻¹ using the internal LED light source, and sample cell CO₂ concentration was maintained at 400 μmol mol⁻¹ via a CO₂ injector. During the 1 h measurement window, ambient air temperature ranged from 28 to 32 °C, and chamber vapor pressure deficit (VPD) was maintained within 1.8–2.5 kPa. Plot measurement sequences were fully randomized to prevent systematic temporal bias.

On the same days as the gas exchange measurements during the jointing, grain-filling, and maturity stages, chlorophyll fluorescence parameters of the ear leaf were measured using a Li-600 fluorescence stomatal meter (LI-COR, USA). The parameters included the maximum photochemical efficiency of PSII (F_v/F_m), the actual photochemical efficiency of PSII (ΦPSII), and the photochemical quenching coefficient (qP).

$$F_v/F_m={\displaystyle \frac{F_m-F_o}{F_m}}$$

(3)

$$\varphi PSII={\displaystyle \frac{F_m^{\prime }-F_s}{F_m^{\prime }}}$$

(4)

$$qP={\displaystyle \frac{F_m^{\prime }-F_s}{F_m^{\prime }-F_o^{\prime }}}$$

(5)

2.3.6. Grain Yield, Water Consumption and Water Productivity

Grain yield was determined at physiological maturity. To eliminate border effects, all sampling was conducted strictly within the central zone of each plot, leaving outer guard rows on both sides and a 1.0 m buffer zone at both ends of each plot unharvested. In each plot, an identical sampling area—a 3 m long section spanning three plastic mulch strips (totaling 12 rows)—was harvested. The actual number of maize plants within each sampling area was precisely recorded. All ears within the designated areas were harvested, air-dried naturally, and manually threshed, with total grain weight measured using an electronic balance. The final grain yield per unit area (kg ha⁻¹) was calculated by converting the harvested grain weight based on the exact sampling area ratio.

Before sowing, after harvest, and both before and 24 h after each irrigation event, systematic soil sampling was conducted across the 0–150 cm soil profile (divided into 0–20, 20–40, 40–60, 60–80, 80–100, 100–120, and 120–150 cm layers). In each plot, three soil cores were collected from locations exhibiting uniform crop growth at varying horizontal distances from the drip line—specifically directly under the drip tape, beside the plant, and in the middle of the wide row. Soil gravimetric water content was subsequently determined using the standard oven-drying method.

Seasonal crop water consumption was evaluated using the soil water balance equation applied to the 0–150 cm soil profile:

$$E{T}_{\mathrm{app}}=P+I+U-R\pm \Delta W$$

(6)

where $E{T}_{\mathrm{app}}$ represents the apparent seasonal crop evapotranspiration (mm), $P$ is precipitation (mm), $I$ is irrigation amount (mm), $U$ is groundwater upward contribution (mm), $R$ is surface runoff (mm), and $\Delta W$ is the change in soil water storage in the 0–150 cm profile between sowing and physiological maturity (mm). Surface runoff ($R$) was zero due to the leveled field and low drip discharge rates (2.6 L h⁻¹). Upward groundwater recharge ($U$) was negligible owing to the deep water table (>5.0 m).

Intensive soil moisture monitoring before and 24 h after irrigation during the grain-filling stage (Figures S2 and S3) indicated that under deficit and moderate irrigation regimes ($W_1$ to $W_3$), moisture replenishment was primarily retained within the 0–80 cm root layer, and soil water content remained below field capacity (gravimetric FC = 15.33% in 0–60 cm), resulting in negligible deep percolation ($D\approx 0$). However, under high-volume single events (particularly $F_{10}{W}_4$ and $W_4$), post-irrigation topsoil moisture temporarily exceeded field capacity, indicating potential gravitational drainage during post-irrigation redistribution (estimated at ~10–12.5% of the single application quota). Because profile sampling was conducted at critical phenological stages rather than continuously after all individual irrigation events to avoid soil disturbance, $E{T}_{\mathrm{app}}$ for $W_4$ represents an upper-bound estimate encompassing actual crop ET and secondary deep percolation losses. Water productivity ($WP$, kg ha⁻¹ mm⁻¹) was calculated as $WP=GY/E{T}_{\mathrm{app}}$.

Water productivity (WP: kg ha⁻¹ mm⁻¹) was defined as the ratio of grain yield to crop water consumption, calculated as:

$$WP={\displaystyle \frac{GY}{ET}}$$

(7)

where GY is the grain yield (kg ha⁻¹), and ET is the crop water consumption (mm).

2.3.7. Nutrient Use Efficiency

Based on the total nitrogen, total phosphorus, and total potassium contents in various organs (stem, leaf, bract, and grain) of spring maize plants at maturity, the N, P, and K uptake per plant was calculated and then converted to kg·ha⁻¹. Subsequently, the nutrient use efficiency was calculated as follows:

$$NUE={\displaystyle \frac{GY}{NU}}$$

(8)

$$PUE={\displaystyle \frac{GY}{PU}}$$

(9)

$$KUE={\displaystyle \frac{GY}{KU}}$$

(10)

where NUE, PUE, and KUE represent nitrogen use efficiency, phosphorus use efficiency, and potassium use efficiency (kg·kg⁻¹), respectively; NU, PU, and KU represent the uptake amounts of nitrogen, phosphorus, and potassium (kg·ha⁻¹), respectively; and GY is the grain yield (kg·ha⁻¹).

2.3.8. Comprehensive Analysis Based on the EWM-TOPSIS Model

To comprehensively evaluate the overall benefits of maize production under different treatments (Grain yield, WP, NUE, PUE, and KUE), this study employed the entropy weight method (EWM) combined with the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) to construct a comprehensive evaluation model. The specific calculation procedures are as follows:

Data standardization: Assuming the experiment consists of m treatments (m = 12 in this study) and each treatment is evaluated by n indicators, the original data matrix is constructed as X = (x_ij)_m×n, where x_ij represents the value of the j-th indicator for the i-th treatment. To eliminate discrepancies in dimensions and magnitudes among different indicators, the original data were first normalized using the range standardization method.

$$z_{ij}={\displaystyle \frac{x_{ij}-\min (x_j)}{\max (x_j)-\min (x_j)}}$$

(11)

where z_ij is the dimensionless value after standardization; max (x_j) and min (x_j) denote the maximum and minimum values of the j-th indicator across all treatments, respectively. To avoid logarithms of zero in subsequent steps, the standardized data were transformed as follows: z′_ij = z_ij + 10⁻⁵.

Weight determination using EWM: The proportion of the i-th treatment for the j-th indicator (p_ij) was calculated as:

$$p_{ij}={\displaystyle \frac{z_{ij}^{\prime }}{{\displaystyle {\sum }_{i=1}^m{z}_{ij}}}},i=1,\dots ,m$$

(12)

The information entropy of the j-th indicator (e_j) was calculated as:

$$e_j=-k{\displaystyle \sum _{i=1}^m{p}_{ij}}\ln (p_{ij}),k={\displaystyle \frac{1}{\ln (m)}}>0$$

(13)

The objective weight of the j-th indicator (w_j) was determined as:

$$w_j={\displaystyle \frac{1-{\mathrm{e}}_j}{{\displaystyle {\sum }_{j=1}^n(1-{\mathrm{e}}_j)}}}$$

(14)

Comprehensive evaluation using TOPSIS: The weighted normalized decision matrix A = (a_ij)_m×n was constructed using the calculated weights wj and the standardized matrix:

$$a_{ij}=w_{ij}\times z_{ij}^{\prime }$$

(15)

The positive ideal solution (A⁺) and the negative ideal solution (A⁻) were identified as follows:

$$A^{+}=(a_1^{+},a_2^{+},\cdots ,a_n^{+})=(\max (a_1),\max (a_2),\dots ,\max (a_n))$$

(16)

$$A^{-}=(a_1^{-},a_2^{-},\cdots ,a_n^{-})=(\min (a_1),\min (a_2),\dots ,\min (a_n))$$

(17)

The Euclidean distances of each treatment from the positive and negative ideal solutions (D_i⁺ and D_i⁻) were calculated as:

$$D_i^{+}=\sqrt{{\displaystyle \sum _{j=1}^n{\left(a_{ij}-a_j^{+}\right)}^2}}$$

(18)

$$D_i^{-}=\sqrt{{\displaystyle \sum _{j=1}^m{\left(a_{ij}-a_j^{-}\right)}^2}}$$

(19)

The relative closeness, denoting the comprehensive evaluation index (S_i), was calculated for each evaluation object:

$$S_i={\displaystyle \frac{D_i^{-}}{D_i^{+}+D_i^{-}}}$$

(20)

where S_i∈ [0, 1]. A larger S_i value indicates that the comprehensive water and fertilizer benefit of the given drip irrigation treatment is closer to the optimal level, whereas a smaller value indicates poorer performance. Based on the S_i values, the 12 treatments were ranked to identify the optimal irrigation strategy.

Additionally, a sensitivity analysis was conducted by perturbing individual indicator weights by 10–20% and sequentially omitting single indicators. The high consistency of treatment rankings across these scenarios confirmed that the EWM-TOPSIS evaluation model was robust and insensitive to weight fluctuations or indicator selections.

2.4. Data Analysis

All experimental data are expressed as mean ± standard deviation (n = 3). Two-way analysis of variance (ANOVA) was conducted using SPSS 26.0 (SPSS Inc., Chicago, IL, USA) to evaluate the main effects of irrigation amount (W), irrigation frequency (F), and their interaction effects (W × F), with block specified as a random effect. Because inter-annual weather variation significantly influenced crop evapotranspiration demand, statistical analyses were performed separately for the 2024 and 2025 growing seasons. For plant physiological and biomass parameters measured repeatedly across the growing season (e.g., LAI, dry matter accumulation, SPAD, leaf gas exchange, and chlorophyll fluorescence), separate two-way ANOVAs were conducted independently at each specific sampling date/growth stage (jointing, tasseling, grain-filling, and maturity). This stage-specific analytical approach was chosen because plant physiological traits undergo rapid ontogenetic shifts over time, making within-stage mean comparisons (p < 0.05, evaluated via Duncan’s multiple range test or LSD test) the most physiologically meaningful method to identify critical water-stress thresholds across developmental phases. Multi-objective evaluation was executed using the EWM–TOPSIS framework in SPSSAU, and figures were rendered using Origin 2024 (OriginLab Corp., Northampton, MA, USA).

3. Results

3.1. Leaf Area Index and Leaf SPAD Value

3.1.1. Leaf Area Index

Across both growing seasons, the leaf area index (LAI) of spring maize followed a unimodal curve, peaking at the tasseling stage (~70 days post-emergence) before declining (Figure 3). Two-way ANOVA indicated that both irrigation amount (W) and frequency (F) significantly influenced LAI (p < 0.05; Figure 3). Under the same frequency, increasing water input (W₂, W₃, and W₄ vs. W₁) expanded tasseling-stage LAI by 9.5–10.2%, 19.6–19.7%, and 24.1–26.8% across both years, respectively. However, the promotional response of LAI to irrigation frequency was strongly modulated by total water input: shifting from F₁₀ to F₇ and F₄ increased LAI by 3.8–7.2% and 8.5–10.8%, respectively, with these gains being most pronounced under water-deficit conditions and progressively diminishing under higher irrigation regimes (W₃ and W₄).

3.1.2. Leaf SPAD Value

Both drip irrigation amount (W) and frequency (F) significantly affected the leaf SPAD value of spring maize across all growth stages (p < 0.05; Figure 4). Across all treatments, SPAD values exhibited a unimodal dynamic, peaking at the grain-filling stage before declining toward maturity. Based on two-year averages under a given frequency, increasing water input gradually enhanced SPAD values, with the largest relative gains observed at maturity: compared with W₁, SPAD values under W₂, W₃, and W₄ increased by 5.6–11.8% during earlier stages (jointing and grain-filling) and by 22.2%, 36.9%, and 50.4% at maturity, respectively. Similarly, under the same irrigation amount, higher irrigation frequencies (F₄ and F₇) enhanced SPAD values relative to F₁₀ across the jointing (6.4–7.0%), grain-filling (5.1%), and maturity stages (10.8–17.6%). Notably, the W × F interaction was particularly pronounced during late senescence (maturity stage), where high frequency combined with adequate water supply (e.g., F₄W₃/F₄W₄) substantially delayed leaf chlorophyll degradation.

3.2. Gas Exchange Parameters

Both drip irrigation amount (W) and frequency (F) significantly affected leaf gas exchange parameters (P_n, T_r and G_sw) across all growth stages (p < 0.05; Figure 5 and Figures S4 and S5). P_n, T_r and G_sw exhibited a progressive decline from jointing to maturity. Across both years, under a given irrigation frequency, increasing water input markedly enhanced gas exchange, particularly during the critical grain-filling stage: compared with W₁, drip irrigation amounts W₂, W₃, and W₄ increased P_n by 19.3–27.4%, 51.1–53.9%, and 84.7–90.8%; T_r by 4.8–40.8%, 53.0–67.0%, and 68.4–82.8%; and G_sw by 25.4–27.3%, 51.3–54.9%, and 67.5–67.9%, respectively. Under the same irrigation amount, higher irrigation frequencies (F₄ and F₇) also promoted leaf gas exchange relative to F10 during grain filling, increasing P_n by 17.9–19.3% (F₄) and 15.4–16.4% (F₇), T_r by 36.5–36.8% (F₄) and 22.4–27.8% (F₇), and G_sw by 21.6–28.5% (F₄) and 19.3–24.9% (F₇). Notably, high irrigation frequency effectively mitigated stomatal limitations (G_sw) and sustained carbon assimilation (P_n) under deficit irrigation regimes, whereas its beneficial physiological impact diminished under full irrigation.

3.3. Chlorophyll Fluorescence Parameters

Both drip irrigation amount and irrigation frequency significantly affected the F_v/F_m, ΦPSII and qP of maize (p < 0.05;

Table 1. Effects of different treatments on leaf chlorophyll fluorescence parameters at key growth stages in 2024 and 2025.

Year	Treatment	Jointing Stage			Grain-Filling Stage			Maturity Stage
		Fv/Fm	ΦPS//	qP	Fv/Fm	ΦPS//	qP	Fv/Fm	ΦPS//	qP
2024	F4W1	0.790 ± 0.021 cd	0.503 ± 0.026 e	0.637 ± 0.031 c	0.727 ± 0.021 d	0.346 ± 0.019 d	0.476 ± 0.012 c	0.719 ± 0.007 c	0.292 ± 0.021 d	0.406 ± 0.034 c
	F4W2	0.795 ± 0.009 bcd	0.572 ± 0.017 d	0.719 ± 0.027 b	0.761 ± 0.011 bc	0.461 ± 0.006 c	0.605 ± 0.013 b	0.765 ± 0.006 a	0.455 ± 0.024 c	0.595 ± 0.034 b
	F4W3	0.824 ± 0.012 ab	0.748 ± 0.023 a	0.908 ± 0.041 a	0.789 ± 0.021 ab	0.600 ± 0.029 ab	0.762 ± 0.049 a	0.765 ± 0.015 a	0.549 ± 0.015 a	0.717 ± 0.021 a
	F4W4	0.833 ± 0.016 a	0.756 ± 0.021 a	0.907 ± 0.008 a	0.805 ± 0.010 a	0.628 ± 0.034 a	0.780 ± 0.044 a	0.766 ± 0.007 a	0.559 ± 0.024 a	0.729 ± 0.026 a
	F7W1	0.773 ± 0.026 d	0.494 ± 0.035 e	0.639 ± 0.025 c	0.720 ± 0.012 de	0.342 ± 0.018 d	0.475 ± 0.023 c	0.719 ± 0.009 c	0.295 ± 0.028 d	0.410 ± 0.043 c
	F7W2	0.786 ± 0.018 cd	0.570 ± 0.025 d	0.725 ± 0.047 b	0.749 ± 0.016 cd	0.458 ± 0.004 c	0.612 ± 0.009 b	0.762 ± 0.011 a	0.440 ± 0.017 c	0.578 ± 0.021 b
	F7W3	0.813 ± 0.015 abc	0.730 ± 0.028 ab	0.898 ± 0.050 a	0.784 ± 0.019 ab	0.592 ± 0.042 ab	0.757 ± 0.071 a	0.765 ± 0.003 a	0.542 ± 0.021 ab	0.709 ± 0.027 a
	F7W4	0.827 ± 0.021 a	0.749 ± 0.026 a	0.905 ± 0.016 a	0.797 ± 0.009 a	0.606 ± 0.051 ab	0.760 ± 0.059 a	0.764 ± 0.009 a	0.544 ± 0.004 ab	0.713 ± 0.008 a
	F10W1	0.706 ± 0.012 e	0.422 ± 0.045 f	0.597 ± 0.063 c	0.695 ± 0.016 e	0.320 ± 0.027 d	0.461 ± 0.050 c	0.709 ± 0.008 c	0.287 ± 0.036 d	0.406 ± 0.054 c
	F10W2	0.769 ± 0.019 d	0.509 ± 0.034 e	0.663 ± 0.059 bc	0.730 ± 0.015 cd	0.423 ± 0.026 c	0.580 ± 0.045 b	0.724 ± 0.014 c	0.416 ± 0.034 c	0.576 ± 0.059 b
	F10W3	0.790 ± 0.016 cd	0.675 ± 0.028 c	0.855 ± 0.044 a	0.761 ± 0.023 bc	0.561 ± 0.041 b	0.737 ± 0.045 a	0.741 ± 0.016 b	0.502 ± 0.031 b	0.677 ± 0.028 a
	F10W4	0.813 ± 0.014 abc	0.686 ± 0.024 bc	0.844 ± 0.018 a	0.783 ± 0.029 ab	0.591 ± 0.033 ab	0.755 ± 0.017 a	0.757 ± 0.011 ab	0.522 ± 0.023 ab	0.690 ± 0.027 a
ANOVA
W		**	**	**	**	**	**	**	**	**
F		**	**	**	**	*	ns	**	*	**
Y		**	*	**	**	***	**	**	**	**
W × F		*	ns	ns	ns	ns	ns	ns	ns	**
Y × W		ns	ns	ns	ns	ns	ns	ns	ns	ns
Y × F		ns	ns	ns	ns	ns	ns	ns	ns	ns
Y × W × F		ns	ns	ns	ns	ns	ns	ns	ns	ns
2025	F4W1	0.800 ± 0.010 def	0.525 ± 0.014 d	0.656 ± 0.018 c	0.737 ± 0.013 d	0.401 ± 0.011 cd	0.544 ± 0.017 bc	0.730 ± 0.008 de	0.308 ± 0.014 e	0.421 ± 0.018 e
	F4W2	0.808 ± 0.009 cde	0.620 ± 0.035 b	0.767 ± 0.036 b	0.764 ± 0.008 c	0.464 ± 0.031 c	0.607 ± 0.034 b	0.764 ± 0.011 abc	0.465 ± 0.011 c	0.609 ± 0.008 c
	F4W3	0.826 ± 0.014 abc	0.742 ± 0.013 a	0.898 ± 0.013 a	0.797 ± 0.023 a	0.626 ± 0.065 ab	0.788 ± 0.100 a	0.779 ± 0.004 a	0.545 ± 0.011 a	0.699 ± 0.014 ab
	F4W4	0.836 ± 0.012 a	0.747 ± 0.017 a	0.894 ± 0.033 a	0.807 ± 0.009 a	0.645 ± 0.042 a	0.799 ± 0.060 a	0.777 ± 0.009 a	0.563 ± 0.016 a	0.724 ± 0.027 a
	F7W1	0.790 ± 0.015 ef	0.517 ± 0.026 d	0.655 ± 0.021 c	0.733 ± 0.010 d	0.390 ± 0.019 d	0.532 ± 0.027 bc	0.723 ± 0.010 ef	0.300 ± 0.015 e	0.416 ± 0.026 e
	F7W2	0.804 ± 0.020 cde	0.623 ± 0.033 b	0.775 ± 0.049 b	0.761 ± 0.008 c	0.460 ± 0.025 c	0.604 ± 0.034 bc	0.760 ± 0.013 abc	0.450 ± 0.003 c	0.592 ± 0.011 c
	F7W3	0.819 ± 0.006 abcd	0.733 ± 0.018 a	0.894 ± 0.016 a	0.789 ± 0.013 ab	0.601 ± 0.054 ab	0.763 ± 0.068 a	0.775 ± 0.013 ab	0.543 ± 0.017 a	0.702 ± 0.032 ab
	F7W4	0.834 ± 0.010 ab	0.742 ± 0.030 a	0.889 ± 0.029 a	0.804 ± 0.012 a	0.628 ± 0.038 a	0.781 ± 0.040 a	0.771 ± 0.013 ab	0.557 ± 0.013 a	0.722 ± 0.006 a
	F10W1	0.724 ± 0.004 g	0.450 ± 0.011 e	0.621 ± 0.018 c	0.723 ± 0.011 d	0.372 ± 0.028 d	0.515 ± 0.047 c	0.710 ± 0.007 f	0.286 ± 0.013 e	0.403 ± 0.020 e
	F10W2	0.781 ± 0.012 f	0.576 ± 0.033 c	0.738 ± 0.033 b	0.741 ± 0.008 d	0.404 ± 0.021 cd	0.546 ± 0.034 bc	0.735 ± 0.010 de	0.350 ± 0.016 d	0.477 ± 0.027 d
	F10W3	0.802 ± 0.016 def	0.704 ± 0.020 a	0.878 ± 0.020 a	0.770 ± 0.007 bc	0.563 ± 0.010 b	0.731 ± 0.007 a	0.746 ± 0.001 cd	0.501 ± 0.011 b	0.672 ± 0.014 b
	F10W4	0.812 ± 0.015 bcde	0.707 ± 0.011 a	0.871 ± 0.026 a	0.790 ± 0.009 ab	0.592 ± 0.028 ab	0.750 ± 0.036 a	0.757 ± 0.015 bc	0.550 ± 0.014 a	0.726 ± 0.004 a
ANOVA
W		**	**	**	**	**	**	**	**	**
F		**	**	*	**	**	*	**	**	**
Y		*	**	**	**	*	**	**	**	*
W × F		*	ns	ns	ns	ns	ns	ns	**	**
Y × W		ns	ns	ns	ns	ns	ns	ns	ns	ns
Y × F		ns	ns	ns	ns	ns	ns	ns	ns	ns
Y × W × F		ns	ns	ns	ns	ns	ns	ns	ns	ns

Note: Data are presented as mean ± standard deviation (SD) (n = 3). Different lowercase letters indicate significant differences among treatments at the p < 0.05 level. ANOVA indicates analysis of variance; *** indicates significance at the 0.1% probability level, ** indicates significance at the 1% probability level; * indicates significance at the 5% probability level; ns indicates not significant; the value of each treatment is the average effect value; W: drip irrigation amounts, F: drip irrigation frequencies, Y: year. W₁, W₂, W₃, and W₄ represent drip irrigation amounts of 350, 500, 650, and 800 mm, respectively. F₄, F₇, and F₁₀ represent drip irrigation frequencies of 4-day, 7-day, and 10-day, respectively.

). Chlorophyll fluorescence parameters exhibited a trend of first increasing and then decreasing from the jointing stage to the grain-filling stage and then to the maturity stage. Under the same irrigation frequency, the F_v/F_m, ΦPSII and qP at the jointing, grain-filling, and maturity stages all initially increased with increasing drip irrigation amount and then stabilized, with no significant difference between W₃ and W₄. Compared with W₁, in 2024, at the grain-filling stage, F_v/F_m increased on average by 4.6%, 8.9%, and 11.3%; ΦPSII increased by 16.3%, 51.7%, and 54.4%; and qP increased by 27.3%, 59.8%, and 62.5%, respectively. In 2025, F_v/F_m increased by 3.3%, 7.34%, and 9.5%; ΦPSII increased by 21.1%, 46.1%, and 47.2%; and qP increased by 10.4%, 43.43%, and 46.5%, respectively. Under the same drip irrigation amount, the F_v/F_m, ΦPSII, and qP at the jointing, grain-filling, and maturity stages all increased with increasing irrigation frequency. Compared with F₁₀, in 2024, at the grain-filling stage, F_v/F_m increased on average by 3.8% and 2.7%; ΦPSII increased by 7.4% and 5.4%; and qP increased by 4.2% and 2.6%, respectively. In 2025, F_v/F_m increased by 2.7% and 2.1%; ΦPSII increased by 10.6% and 7.6%; and qP increased by 7.7% and 5.4%, respectively.

3.4. Aboveground Dry Matter Accumulation

Both drip irrigation amount and irrigation frequency significantly affected the aboveground dry matter accumulation of maize (p < 0.05; Figure 6). Overall, the total aboveground dry matter of maize in 2024 and 2025 showed consistent trends under different drip irrigation amounts and frequencies, with the maximum value achieved under the F₄W₄ treatment in both years. Under the same irrigation frequency, total aboveground dry matter increased with increasing drip irrigation amount. Compared with W₁, drip irrigation amounts W₂, W₃, and W₄ increased by 14.0%, 26.9%, and 35.8% in 2024, and by 16.7%, 40.1%, and 52.4% in 2025, respectively. Under the same drip irrigation amount, the total aboveground dry matter of spring maize at maturity increased with increasing irrigation frequency; however, raising the irrigation amount reduced the differences caused by irrigation frequency. Compared with F₁₀, in 2024, under the W₁ treatment, F₄ and F₇ increased by 12.7% and 12.8%, respectively; under W₂, they increased by 15.8% and 9.1%; under W₃, by 11.9% and 12.1%; and under W₄, by 11.4% and 9.91%, respectively. In 2025, under W₁, F₄ and F₇ increased by 38.2% and 29.9%; under W₂, by 27.5% and 15.3%; under W₃, by 3.7% and 0.5%; and under W₄, by 6.7% and 5.6%, respectively.

3.5. Grain Yield, Water Consumption and Water Productivity

3.5.1. Grain Yield

Both drip irrigation amount and irrigation frequency significantly affected the grain yield of maize (p < 0.05; Figure 7). Under the same irrigation frequency, grain yield increased with increasing irrigation amount. Compared with W₁, grain yield under W₂, W₃, and W₄ increased by 40.9%, 81.8%, and 105.3% in 2024, and by 35.8%, 68.9%, and 92.6% in 2025, respectively. Under deficit to moderate irrigation levels (W₁–W₃), there was no significant difference in grain yield between the F₄ and F₇ treatments, but both produced significantly higher yields than F₁₀. Across W₁–W₃, grain yield under F₄ and F₇ increased by an average of 19.2% and 20.4% in 2024, and 14.9% and 11.1% in 2025 compared to F₁₀, respectively. However, under full irrigation (W₄), no significant differences in grain yield were observed among F₄, F₇, and F₁₀. In 2024, the maximum grain yield was achieved under the F₇W₄ treatment, followed by F₄W₄ and F₁₀W₄. In 2025, the highest average grain yield was obtained under the F₄W₄ treatment, followed by F₇W₄ and F₁₀W₄.

3.5.2. Water Consumption and Water Productivity

Pre- and post-irrigation soil water profile monitoring confirmed that moisture dynamics were predominantly confined to the 0–80 cm root layer, with the deep 100–120 cm soil water content remaining unchanged before and after irrigation (Figures S2 and S3). Both drip irrigation amount and irrigation frequency significantly affected the water consumption (ET) and water productivity (WP) of maize (p < 0.05; Figure 8). Under the same drip irrigation amount, the ET under W₁ and W₂ first decreased and then increased with increasing irrigation frequency, while under W₃ and W₄ it increased with increasing irrigation frequency. Under the same irrigation frequency treatment, ET increased with increasing irrigation amount. On average, compared with the W₁ treatment, the ET under the W₂, W₃, and W₄ treatments increased by 37.3%, 76.5%, and 115.9% in 2024, and by 39.7%, 77.7%, and 116.3% in 2025, respectively.

Under the same drip irrigation amount, WP generally followed the order “F₄ ≈ F₇ > F₁₀”. Compared with F₁₀, the WP under F₄ and F₇ increased by 24.4% and 24.7% in 2024, and by 17.9% and 16.8% in 2025, respectively. Under the same irrigation frequency, the WP under F₄ and F₇ decreased with increasing drip irrigation amount, while the WP under F₁₀ first increased and then decreased with increasing drip irrigation amount. For F₄, compared with W1, the WP under W₂, W₃, and W₄ decreased by 0.9%, 5.8%, and 14.3% in 2024, and by 4.7%, 12.5%, and 19.5% in 2025, respectively. For F₁₀, compared with W₁, the WP under W₂, W₃, and W₄ increased by 13.6%, 21.2%, and 20.7% in 2024, and by 6.1%, 11.0%, and 9.9% in 2025, respectively.

Comparing the measured $E{T}_{\mathrm{app}}$ with the theoretical crop water requirement ($E{T}_c\approx 628.58 \mathrm{mm}$ in 2024 and $582.45 \mathrm{mm}$ in 2025; Section 2.3.2), the seasonal water consumption under $W_3$ showed the highest congruence with crop atmospheric demand. In contrast, $W_1$ and $W_2$ suffered from severe deficit stress, while $W_4$ ($E{T}_{\mathrm{app}}>800 \mathrm{mm}$) resulted in luxury water consumption and secondary percolation losses, which explains the pronounced reduction in $WP$ under excessive irrigation.

3.6. Nutrient Use Efficiency

Both drip irrigation amount (W) and frequency (F) significantly influenced nitrogen, phosphorus, and potassium use efficiencies (NUE, PUE, and KUE) of spring maize, exhibiting a significant interactive effect between the two factors (W × F, p < 0.05; Figure 9). Overall, across all frequency levels, NUE, PUE, and KUE followed a parabolic (unimodal) response to increasing water input, peaking under moderate irrigation (W₃) before declining under the highest water supply (W₄). Across both years, under a given irrigation frequency, increasing irrigation volume from W₁ to W₂, W₃, and W₄ increased NUE by 25.6–32.2%, 40.8–43.1%, and 42.2–45.3%; PUE by 21.9–24.0%, 32.6–38.9%, and 25.8–36.3%; and KUE by 15.4–23.6%, 24.2–51.5%, and 27.1–55.2%, respectively. Similarly, under a given irrigation amount, nutrient use efficiencies initially increased and then decreased with higher irrigation frequencies, with the 7-day interval (F₇) consistently outperforming both low-frequency (F₁₀) and high-frequency (F₄) regimes. Compared with F₁₀, F₄ and F₇ increased NUE by 8.3–17.6% and 15.1–26.2%; PUE by 8.5–10.8% and 18.9–27.2%; and KUE by 6.7–16.9% and 16.6–27.9%, respectively. Notably, the W × F interaction demonstrated that moderate irrigation combined with an intermediate frequency (F₇W₃) achieved the highest synergistic resource use efficiencies, whereas excessive irrigation (W₄) attenuated these gains due to diminishing marginal returns.

3.7. Comprehensive Analysis Based on the EWM-TOPSIS Model

Based on the weights assigned by the entropy weight method (EWM), a weighted normalized matrix was constructed (Tables S1 and S2). The TOPSIS model was then employed to calculate the relative closeness of each treatment to the ideal solution, and the comprehensive benefit ranking was subsequently derived (

Table 2. Comprehensive evaluation and ranking of different treatments based on the EWM-TOPSIS model in 2024 and 2025.

Treatment	2024				2025
	D+	D−	Ci	Ranking	D+	D−	Ci	Ranking
F4W1	0.292	0.206	0.414	10	0.301	0.254	0.458	10
F4W2	0.215	0.270	0.557	8	0.211	0.289	0.578	8
F4W3	0.133	0.334	0.715	3	0.145	0.330	0.695	2
F4W4	0.163	0.337	0.674	6	0.207	0.312	0.601	7
F7W1	0.238	0.256	0.519	9	0.247	0.279	0.531	9
F7W2	0.150	0.311	0.675	5	0.167	0.292	0.637	4
F7W3	0.034	0.428	0.927	1	0.102	0.395	0.795	1
F7W4	0.126	0.362	0.742	2	0.185	0.330	0.641	3
F10W1	0.449	0.000	0.000	12	0.454	0.000	0.000	12
F10W2	0.276	0.178	0.392	11	0.297	0.167	0.360	11
F10W3	0.149	0.307	0.673	7	0.190	0.296	0.610	6
F10W4	0.151	0.336	0.690	4	0.189	0.330	0.636	5

Note: D⁺ and D⁻ represent the Euclidean distances to the positive-ideal and negative-ideal solutions, respectively. C_i represents the relative closeness coefficient, where a higher C_i value (closer to 1) indicates better comprehensive performance. C_i = 0.000 for the F₁₀W₁ treatment occurs because this severe water-deficit treatment defined the negative-ideal solution (D⁻ = 0) across all evaluated indicators. W and F denote drip irrigation amounts and frequencies, respectively. W₁, W₂, W₃, and W₄ represent drip irrigation amounts of 350, 500, 650, and 800 mm, respectively; F₄, F₇, and F₁₀ represent drip irrigation frequencies of 4-day, 7-day, and 10-day intervals, respectively.

). The F₇W₃ treatment (irrigation frequency of 7 days, irrigation amount of 650 mm) exhibited the highest relative closeness in both 2024 and 2025 (92.7% and 79.5%, respectively), consistently ranking first. This indicates that this treatment minimizes redundant consumption of water and nutrients while ensuring crop yield, achieving a perfect balance of multiple objectives. Overall, adopting an irrigation frequency of 7 days combined with a seasonal drip irrigation amount of 650 mm (F₇W₃) enables the optimal spatiotemporal match between water and fertilizer supply and crop demand for spring maize in arid regions, and is therefore recommended as the optimal irrigation regime under the conditions of this experiment.

4. Discussion

4.1. Driving Mechanism of Drip Irrigation Amount and Frequency Regulation on Photosynthetic Physiology and Dry Matter Accumulation of Spring Maize in Arid Regions

Canopy photosynthetic performance serves as the core driver of biomass accumulation, and its extreme sensitivity to root-zone water availability ultimately determines overall crop vigor and yield potential [10,11]. Our results demonstrated that severe water deficit (W₁) or prolonged irrigation intervals (F₁₀) significantly inhibited leaf gas exchange parameters (P_n, T_r, and G_sw) and PSII photochemical efficiency (F_v/F_m, ΦPSII and qP). This physiological impairment was primarily driven by intermittent drought stress. Under water-limited conditions, crops rapidly close their stomata to maintain turgor pressure and minimize transpiration loss; however, this adaptive mechanism subsequently impedes CO₂ diffusion into the intercellular spaces, leading to severe stomatal limitation of photosynthesis [12,13,14]. Specifically, the extreme deficit regime (F₁₀W₁) combined prolonged soil drying with high atmospheric evaporative demand (Tmean = 20–30 °C, RH = 40–70%), inducing visible mid-day leaf rolling and premature canopy senescence. This severe coupled drought triggered not only acute stomatal closure but also non-stomatal impairment of PSII reaction centers, explaining the drastic drop in photoassimilate accumulation.

Chlorophyll fluorescence parameters serve as intrinsic indicators of PSII sensitivity to environmental stress. The findings revealed that F_v/F_m, ΦPSII and qP remained at low levels under W₁ and F₁₀ treatments. These deficits damage the thylakoid membrane structure of mesophyll cells, thereby reducing light-harvesting and electron transport efficiency [15,16,17]. Conversely, increasing irrigation volume and frequency (e.g., F₄W₄) alleviated stress and effectively promoted leaf cell expansion. Specifically, the F4 and F7 treatments significantly enhanced photosynthesis and PSII photochemical efficiency, maintaining a high leaf area index (LAI), SPAD values, and robust PSII activity. This provided a sufficient carbon source for photoassimilate synthesis, ultimately leading to a significant increase in aboveground dry matter accumulation at maturity [45,46,47]. However, the promotional effect of increased irrigation frequency on dry matter accumulation diminished when the irrigation volume reached W₄. This indicates that under ample or excessive water supply, moisture is no longer the primary limiting factor for growth; instead, a plausible explanation is that excessive water might temporarily impair root respiration and metabolic activity due to reduced soil aeration, although soil oxygen status and root respiration were not directly measured in this study [18,19,20].

4.2. Synergistic Regulation Mechanism of Drip Irrigation Volume and Frequency on Yield and Water-Fertilizer Use Efficiency

A core challenge in modern arid agriculture is maximizing the synergistic efficiency of water and fertilizer while ensuring grain yield [27,48]. Our results quantitatively demonstrate an inevitable trade-off: although the maximum irrigation volume (W₄) produced the highest yield, its water productivity (WP) declined significantly under F₄ and F₇ frequencies. This decline stems from an exponential increase in crop water consumption that outpaced the marginal growth in yield, leading to luxury water consumption [21,49,50]. Similarly, nutrient use efficiencies (NUE, PUE, and KUE) followed a parabolic trend, ultimately exhibiting a sharp decline under excessive water input. This corroborates the view that pursuing absolute maximum yield often comes at the expense of resource efficiency and increased environmental costs [39,51]. The intrinsic mechanism of this yield–efficiency trade-off is highly dependent on the spatiotemporal synchronization between root-zone water dynamics and nutrient transport characteristics [23,29]. In fertigation systems, water movement is the primary driver governing nutrient spatial distribution. In light sandy loam soils with low water-holding capacity (gravimetric field capacity of 15.33% in the 0–60 cm layer; Section 2.1), the single-event irrigation quota fundamentally dictates downward hydrological and solute fluxes. Under moderate irrigation regimes (such as F₇W₃), single applications maintained soil moisture within the 0–80 cm root zone without exceeding field capacity, effectively preventing deep percolation (D ≈ 0) and closely matching theoretical crop water demand (ET_c ≈ 582–628 mm). In contrast, under the highest irrigation quota (W₄) and low-frequency large-volume events (F₁₀W₄, ~90 mm/event), post-irrigation soil moisture in the 0–40 cm layer reached 20–25%, significantly surpassing field capacity (Figures S2 and S3). Under gravity, this excess perched water inevitably redistributes downward beyond the active 0–100 cm root zone during the days following irrigation, creating an estimated single-event percolation deficit of ~10.2 mm (~12.5% of the applied quota). This deep percolation loss not only increases apparent water consumption and reduces WP, but also provides direct hydraulic evidence for the downward displacement of highly mobile nitrate (NO₃⁻-N) below the primary root layer (60 cm). This uncouples nitrate availability from root uptake and directly accounts for the pronounced decline in nitrogen, phosphorus, and potassium use efficiencies (NUE, PUE, KUE) observed under excessive irrigation inputs [6,25]. Conversely, available phosphorus (AP) and potassium (AK) exhibit lower mobility in soil. Insufficient single-irrigation volumes can lead to the localized accumulation of these nutrients in the topsoil directly beneath the dripper, causing a spatiotemporal mismatch between water and fertilizer availability [27].

Irrigation frequency and appropriate water volume play a pivotal role in mediating this contradiction. Compared to low-frequency irrigation, short-interval drip irrigation treatments (F₄ and F₇) distribute limited water more uniformly throughout the growth period by maintaining a smaller single-irrigation quota. Specifically, as a proposed mechanism, a moderate irrigation combination (e.g., F₇W₃) is hypothesized to create a favorable wetting front in sandy loam. We speculate that this moisture regime facilitates the local diffusion of P and K toward the root surface by maintaining adequate soil moisture, while optimizing the wetting extent to retain water and mobile nitrate within the primary root zone (0–40 cm). This theoretical spatial alignment provides a plausible mechanism for the observed high congruence between N, P, and K supply and crop uptake patterns [30,31,32]. Physiologically, the superior resource efficiency under F₇ relative to F₄ stems from a finer balance among soil evaporation, root architecture, and soil aeration. Continuous topsoil wetness under F₄ (4-day interval) exacerbates non-beneficial soil evaporation from unmulched wetting zones and could theoretically restrict soil oxygen availability, potentially favoring a shallower root distribution (0–30 cm). Conversely, the 7-day interval (F₇) allows the thin topsoil surface to form a dry soil mulch that suppresses direct evaporation while preserving moisture in the 20–60 cm layer for crop transpiration. Moreover, moderate wetting-drying cycles under F₇ are hypothesized to enhance soil aeration and potentially stimulate compensatory deeper root proliferation (30–80 cm), which may explain the enhanced water and nutrient capture during the grain-filling stage without exceeding field capacity. Agronomically, while ultra-high-frequency drip irrigation (e.g., 1–2 day intervals or daily micro-irrigation) is theoretically advantageous for coarse soils, it presents notable practical drawbacks under field-scale production in Southern Xinjiang. First, regional water delivery is governed by centralized canal and pump rotational scheduling, where rotational turns across field blocks generally operate on 5- to 10-day cycles, making 1- to 2-day applications logistically difficult and cost-prohibitive. Second, film mulching effectively acts as a vapor barrier, substantially retarding root-zone drying in sandy loam and buffering the crop between 4- and 7-day intervals. Third, persistent ultra-high-frequency watering keeps the surface layer constantly saturated, which can induce superficial root crowding and heighten maize lodging risks during windy summer months. Furthermore, the differences in yield and multi-resource use efficiency between F₄ and F₇ were not significant. Considering practical field operations, a 7-day frequency not only maintained high grain yield and WP while maximizing multi-element synergistic uptake, but also reduced total seasonal irrigation operations by over 40% compared to F₄. From a management perspective, this lower operational frequency offers potential practical advantages in reducing labor inputs and system operating wear without sacrificing crop productivity [37,52].

4.3. Multi-Objective Comprehensive Evaluation and Management Implications Based on the EWM-TOPSIS Model

Traditional irrigation regimes often rely on single-factor optimization (e.g., pursuing maximum yield regardless of cost), which inevitably leads to severe resource wastage, diminished economic returns, and environmental burdens such as groundwater pollution in fragile arid ecosystems [24,40]. In this study, the EWM-TOPSIS model was employed to reduce subjective weighting bias and comprehensively evaluate the conflicting objectives among yield, water productivity (WP), and nutrient use efficiency (NUE). The model identified F₇W₃ (7-day interval, 650 mm irrigation volume) as the optimal strategy among the tested treatments under the specific conditions of this experiment. From an agronomic perspective, F₇W₃ maintains a moderate single-irrigation load, ensuring optimal PSII activity and leaf gas exchange. From the standpoint of environmental sustainability, it effectively curbs the diminishing marginal returns associated with luxury water consumption (W₄). By sacrificing a statistically non-significant fraction of absolute yield, the F₇W₃ regime substantially enhanced WP, NUE, PUE, and KUE, achieving the highest relative closeness coefficient [3,42,44].

Despite these compelling results, the site-specific nature of this study presents certain limitations. The spatiotemporal dynamics of water and fertilizer were evaluated specifically based on a sandy loam profile, where soil texture fundamentally dictates hydraulic conductivity [53]. Consequently, site-specific calibrations may be required when extrapolating the optimal F₇W₃ regime to heavy clay or severely saline–alkali soils. Furthermore, while we established robust links between macronutrient uptake and canopy physiology, the underlying micro-mechanisms, such as root architectural traits and the response of the rhizosphere microbiome to varying wetting–drying frequencies, were not explored in depth [54,55]. Future research should integrate root–soil–microbe interaction models to further refine precision water and nutrient management in arid agroecosystems. Several experimental boundaries should be acknowledged: (1) deep percolation fluxes, dynamic nitrate leaching rates, and in situ 3D root architectural traits (e.g., root length density across depth) were not directly measured using automated drainage lysimeters, suction lysimeters, or minirhizotron imaging; rather, these mechanisms were inferred from profile soil moisture dynamics (0–150 cm) and total crop nutrient uptake patterns; (2) ultra-high-frequency irrigation regimes (1–2 day intervals) were not evaluated due to regional rotational canal infrastructure. Future research should integrate continuous drainage monitoring systems, in situ root-rhizosphere sensors, and automated micro-pulsed irrigation to validate these hydrological and physiological trade-offs at higher temporal resolutions.

5. Conclusions

A two-year field study systematically elucidated the synergistic mechanisms of drip irrigation amount and frequency on spring maize in arid sandy loam soils. Severe water deficit or low-frequency irrigation induced pronounced stomatal limitations and photosystem II (PSII) photochemical damage, significantly inhibiting leaf P_n, T_r, G_sw, F_v/F_m, ΦPSII, qP, and aboveground dry matter accumulation. Conversely, short-interval drip irrigation (F₄ and F₇) not only maintained robust leaf gas exchange and PSII activity but also reduced water consumption under the same total irrigation volume, thereby increasing WP by 17–24% compared to low-frequency irrigation (F₁₀). Furthermore, although the highest irrigation volume (800 mm) achieved the maximum grain yield, it concurrently caused a substantial increase in water consumption. This led to significant decreases in WP and nutrient use efficiencies, thereby revealing an inevitable trade-off between yield and resource efficiency. By employing the EWM-TOPSIS (entropy weight method combined with the Technique for Order Preference by Similarity to Ideal Solution) model to reduce subjective weighting bias, the F₇W₃ treatment (650 mm volume, 7-day interval) was identified as the optimal strategy among all evaluated treatments under the tested conditions. By coordinating yield, WP, and multi-nutrient efficiencies, this treatment achieved the highest relative closeness coefficients in 2024 and 2025 (92.1% and 76.3%, respectively). Ultimately, it effectively balanced robust canopy physiological performance, high grain yield, and the optimization of WP and fertilizer use efficiencies. However, these conclusions and the identified optimal irrigation regime (F₇W₃) are based on a two-year field evaluation under the typical hyper-arid climate and sandy loam soil conditions of Southern Xinjiang. Consequently, the boundary of applicability for this specific regime should be considered within similar ecological zones. To extend these findings to regions with differing soil textures, rainfall patterns, or inter-annual climate fluctuations, future studies should incorporate multi-location long-term field trials alongside process-based crop growth models (e.g., AquaCrop or DSSAT) for regional calibration and dynamic management.