1 Introduction

Drought is a severe environmental stress threatening global food security by adversely impacting sustainable crop production (Ullah et al. 2017; Ilyas et al. 2021). Rice (Oryza sativa L.) is the leading cereal crop and the largest consumer of freshwater in the agricultural sector. However, decreasing freshwater availability due to enormous population growth, climate change, industrialization, and unproductive water loss is the main challenge to its sustainable production (Chareesri et al. 2020; Das et al. 2021a). Drought induces down-regulation of different physiological and biochemical processes in which signal transmitted through several messengers that modulate the production of reactive oxygen species (ROS) in rice causing severe oxidative damage to cellular activities resulting in a significant yield loss (Farooq et al. 2009; Ullah et al. 2019b; Panda et al. 2021). Drought response is also developed at molecular level where changes in gene expression (up- and down-regulation) take place (Farooq et al. 2009). Different genes are induced in response to drought at the transcriptional level, and these gene products are thought to function in tolerance to drought (Farooq et al. 2009). Future climate projection indicates more uncertainties on irrigation water availability and frequent episodes of drought, which will further disturb the widely prevailing rice production systems under inundated conditions where fields are kept flooded during most part of the growth period (Ullah et al. 2017). Conventional rice cultivation system under continuous flooding is water intensive overexploiting freshwater resources and should be replaced with water-saving rice production systems to maintain productivity and to conserve environment without compromising grain yield (Ruiz-Sanchez et al. 2011). Alternate wetting and drying (AWD) is a highly efficient water-saving and environment-friendly irrigation technique for lowland irrigated rice, which reduces irrigation water input by as high as 38% and greenhouse gas emissions by as high as 78% compared with the traditional method of rice cultivation without compromising yield (Maneepitak et al. 2019a; Santiago‐Arenas et al. 2021; Das et al. 2021b). Despite visible benefits, adoption of AWD at larger scale is still a challenge largely due to the associated fear of yield loss, which warrants for agronomic management options to maintain yield. In addition, proper implementation of AWD where soil drying does not go beyond the so-called safe level (field water level is not allowed to drop more than 15 cm below the soil surface [soil water potential ≥ –20 kPa]) is also critical for better results (Carrijo et al. 2017). Another constraint, specifically associated with Thailand, regarding the implementation of AWD in the wet season is continuous rainfall, which makes the maintenance of the dry phase of the AWD cycle difficult; therefore, the adoption of this technique is substantially higher in the dry season rice compared with the wet season rice (Ruensuk et al. 2021). Proper implementation of AWD coupled with better nutrient management option would help boost the confidence of rice-growing farmers inclined to adopt this technique. In this regard, potassium (K) or silicon (Si) when applied in conjugation with optimum nitrogen (N) and phosphorus (P) fertilization has been found effective in enhancing growth, yield, and irrigation water productivity of rice under water-deficit stress (Guntzer et al. 2012; Ullah et al. 2018b, 2019a). Therefore, exogenous soil application of K or Si is a promising approach for sustainable rice production in the context of decreasing irrigation water availability.

Potassium is an indispensable macronutrient and chemical constituent of various crucial biomolecules that regulate many physiological and biochemical functions in plants including enzyme activation, protein synthesis, and carbohydrate metabolism (Sardans and Peñuelas 2021). The direct role of K in regulating stomatal movements, photosynthetic carbon reduction, antioxidant-mediated defense, ion homeostasis, and ROS detoxification in plants is well documented that confers tolerance to plants against drought stress (Waraich et al. 2012). Potassium-induced stomatal regulation helps reduce water loss through transpiration improving water use efficiency of plants (Chaves et al. 2002; Miao et al. 2010). Potassium mediates plant-water relations regulating ionic balance and osmotic adjustment by maintaining lower osmotic potential with higher turgor pressure in plant cell under drought stress (Römheld and Kirkby 2010; Ullah et al. 2019a). High concentration of K in cytosol of root tissues promotes root hair elongation, which in turn accelerates water and nutrient uptake potential of roots resulting in better utilization of soil moisture under limited water supply (Zain and Ismail 2016; Wang et al. 2017). Scavenging ROS through enzymatic and non-enzymatic antioxidant mechanisms is closely associated with the level of cellular K concentration for improved tolerance when cell encounters dehydration (Farooq et al. 2010). Potassium fertilization facilitates proline biosynthesis in plants, which is a well-known osmolyte playing a significant protective role through osmotic adjustment and enhanced antioxidant enzyme activities protecting cells against dehydration (Teixeira and Pereira 2007; Ellouzi et al. 2017). Different metabolic reactions are catalyzed by cellular K, which also controls repairing of tissue damage through scavenging toxic molecules in plant cells (Kanai et al. 2011). It has been reported that K fertilization facilitates various physiological and biochemical mechanisms, such as an enhancement of aquaporins activity, water use efficiency, root growth, and cell membrane stability, resulting in an improved drought tolerance (Jatav et al. 2014; Wang et al. 2017).

In plant nutrition, Si is considered an “anomaly” as it is apparently not essential for plant growth and development, but an increasing evidence in the literature shows that soluble Si is beneficial to plants, especially in alleviating the adverse effect of drought stress on most of the cereal crops (Cuong et al. 2017; Hoseinian et al. 2020). Silicon improves soil water holding capacity to ensure an increase in plant available water for maintaining photosynthetic carbon assimilation inducing drought tolerance in plants (Kuhla et al. 2021). The functional role of Si has been well documented in rice, which minimizes cell damage either by deposition in the cell wall and intercellular spaces or inducing lignin biosynthesis for regulating plant metabolism (Ullah et al. 2018b; Sirisuntornlak et al. 2019; Sathe et al. 2021). Plants uptake Si in the form of monosilicic acid and rice being a hyper-accumulator of Si has a greater capability of Si acquisition in culm/stem and leaf, which help enhance mechanical strength with improved light harvesting and distribution capacity (Zargar et al. 2019). Silicon is highly effective in mitigating abiotic stress-mediated cellular damage and in improving crop yield (Yan et al. 2018). Exogenous soil application of Si and microbial inoculation has been reported to minimize drought stress-induced damage on growth, yield, and water productivity of rice (Das et al. 2021a). Silicon in soil produces organic acid, phytohormone, biosurfactant, allelochemicals, and exopolymer in plant root, which help protect plant from heavy metal stress and nutrient deficiency (Ahmed et al. 2011; Etesami 2018). Exogenous soil application of Si alleviates drought stress by (i) improving photochemical efficiency with better chlorophyll integrity, enhancement of CO2 assimilation rate, and activation of antioxidant defense capacity (Rizwan et al. 2015), (ii) enhancing hydraulic conductance of roots and water use efficiency, nutrient acquisition, and osmotic adjustment (Tripathi et al. 2015; Chen et al. 2018; Zargar et al. 2019; Schaller et al. 2021), and (iii) reducing stomatal conductance (Vandegeer et al. 2021). The role of Si in reducing stomatal conductance is not well established in rice. In contrast, Liu et al. (2014) reported that Si enhanced transpiration rate under water-deficit stress through up-regulating aquaporin gene expression. Similarly, Ming et al. (2012) reported higher root hydraulic conductance, stomatal conductance, and leaf transpiration rate of Si-supplemented rice plants with improved water uptake to induce tolerance against water-deficit stress. It has been well documented that Si is highly beneficial in enhancing growth and yield of various agronomic and horticultural crops under water-deficit stress (Guntzer et al. 2012; Ullah et al. 2018b; Sirisuntornlak et al. 2019; Alam et al. 2021; Chakma et al. 2021).

Drought-driven yield loss could be minimized by using various approaches, but judicious management of K and Si nutrition synchronized with efficient water management might be a promising strategy for optimizing growth and yield of rice under water-deficit stress. Moreover, the popular establishment method of rice through transplanting seedlings from nursery to well-puddled soil is also a water-intensive approach associated with a high amount of unproductive water loss along with a high demand of labor (Ullah et al. 2017). In contrast, direct seeding (wet) is increasingly gaining popularity for its water- and labor-saving potential where three basic field operations, such as puddling, transplanting, and maintaining a 3–5 cm of standing water throughout the growing season, are avoided (Ullah et al. 2017). A handful of published work evaluating the individual effect of K or Si on rice is available, but to the best of our knowledge, no published literature is available dealing with a synchronized use of K and/or Si under various levels of AWD and cultivation methods to optimize yield and water productivity of rice. It was hypothesized that proper combination of K and Si fertilizer would improve growth, yield, and water productivity of rice cultivated through wet direct seeding method by using the “safe” AWD irrigation. Therefore, the objective of this study was to evaluate the impact of K and Si fertilizer applied alone or in combination on growth, yield, and water productivity of rice cultivated through wet direct seeding or transplanting method subjected to water-deficit stress.

2 Materials and Methods

2.1 Experimental Set-up

A polyhouse experiment was conducted at the Asian Institute of Technology (14.0791° N, 100.6114° E), Bangkok, Thailand, during 2020. Black plastic pot (height: 30 cm, bottom diameter: 28 cm, and top diameter: 36 cm) was filled with 10 kg dry soil collected from the research farm of the Asian Institute of Technology consisting of 12% sand, 30% silt, 58% clay, and 2.6% organic matter. The soil is slightly acidic (pH of 6.0) with an inherent Si content of 0.0056% along with exchangeable P, K, Ca, and Mg contents of 0.0044%, 0.0304%, 0.33%, and 0.0359%, respectively. Field capacity of the soil was calculated following Datta et al. (2009) where 39.6% soil moisture content was determined at 100% field capacity corresponding to 0 kPa soil water potential. Temperature inside the polyhouse ranged from 22 to 36 °C, while relative humidity fluctuated between 70 and 85% throughout the growing period.

2.2 Experimental Treatment and Design

The factorial experiment consisted of three factors: six fertilizer doses including different combinations of K and Si applied along with N and P and the control (NP100 [control], NP100 + K100, NP100 + Si100, NP100 + K75 + Si25, NP100 + K50 + Si50, and NP100 + K25 + Si75), three soil water potential levels maintained through AWD irrigation (0, –15, and –30 kPa), and two cultivation methods (wet direct seeding and transplanting). The subscript values of 100, 75, 50, and 25 under a specific nutrient represent 100%, 75%, 50%, and 25% of the recommended field application dose of that particular nutrient. The experiment was arranged in a completely randomized design with three replications in which each pot containing one single plant was considered as an experimental unit for an individual treatment combination.

2.3 Crop Husbandry and Nutrient Management

Seeds of Pathumthani 1 rice variety (Oryza sativa L. ssp. indica) were procured from the Pathum Thani Rice Research Center, Pathum Thani, Thailand. Pathumthani 1 is a photoperiod-insensitive, drought-susceptible variety (Cha-um et al. 2010), and its maturity period ranges from 110 to 120 days cultivated through broadcasting (direct seeding) method and 115 to 125 days cultivated through transplanting method (Ullah et al. 2018a; Santiago-Arenas et al. 2019, 2020). The required commercial fertilizers, such as urea (46% N), triple superphosphate (46% P2O5), potassium chloride (60% K2O), and monosilicic acid (20% Si), were collected from a local market. Each pot was fertilized with 0.71 g of NP (16:20; 156 kg ha–1) as a basal dose at 1 day before sowing or transplanting along with 0.43 g of urea (94 kg ha–1) during panicle initiation stage (80 days after seeding/sowing or 65 days after transplanting). These N and P doses were treated as the 100% recommended field application dose of N and P (Maneepitak et al. 2019b). For K and Si fertilization, potassium chloride at 200 kg ha–1 (120 kg K ha–1) (Ullah et al. 2019a) and monosilicic acid at 300 kg ha–1 (60 kg ha–1 soluble Si) (Ullah et al. 2018b) were applied as a basal dose in which each respective pot received 0.91 g potassium chloride and 1.36 g monosilicic acid. These K and Si doses were used for the baseline calculation of 100% K and 100% Si. Other doses of K and Si (75%, 50%, and 25%) were calculated based on the doses of 100% K and 100% Si. Seeds were surface sterilized with 10% H2O2 for 10 min for disinfection and breaking of dormancy, rinsed thoroughly with distilled water, and then soaked in distilled water for 24 h before sowing (Ullah et al. 2017). Pre-germinated seeds were sown in plastic pots (three pre-germinated seeds in each pot) for the wet direct seeding method, while three healthy seedlings were transplanted into each pot from the nursery tray at 15 days after sowing for the transplanting method. Finally, one seedling was kept in each pot for both cultivation methods.

2.4 Establishment of Different Soil Water Potential Levels

Tensiometer (Model 2725ARL Jet Fill Tensiometer, Soil moisture Equipment Corp., CA, USA) was permanently installed into the pot for the whole growing season to monitor soil water potential. The tensiometer was set at 15 cm soil depth as most of the rice roots proliferate within this range of soil depth (Ullah et al. 2019b). Sufficient soil moisture was maintained until 30 day of seedling growth followed by establishment of desired soil water potential levels (0, –15, and –30 kPa) maintained through AWD irrigation. Soil water potential levels of 0, –15, and –30 kPa corresponded to approximately 100%, 80%, and 60% field capacity, respectively. Irrigation water was applied when soil water potential reached at the desired level. Soil water potential of –15 and –30 kPa indicates a 20% and 40% depletion of the maximum amount of water held in the soil, respectively, whereas 0 kPa refers to the maximum amount of water held in the soil after gravitational water drainage stops.

2.5 Data Collection

Data on plant height (cm), flag leaf length (FLL) (cm), root dry matter (RDM) (g plant–1), shoot dry matter (SDM) (g plant–1), leaf greenness (SPAD value), leaf relative water content (LRWC) (%), net photosynthetic rate (Pn) (μmol CO2 m–2 s–1), stomatal conductance (gs) (mmol H2O m–2 s–1), transpiration rate (E) (mmol H2O m–2 s–1), free proline (µg g–1 fresh weight), tiller number plant–1, panicle number plant–1, spikelet number panicle–1, filled grain (%), 1000-grain weight (g), grain yield (g plant–1), harvest index (%), water productivity (kg m–3), and shoot K and Si content (mg g–1 dry weight) were collected. Data on plant height, FLL, SPAD value, LRWC, tiller number plant–1, and panicle number plant–1 were measured before harvest.

Plant height was measured from the ground level to the tip of the topmost leaf/panicle using a meter scale 1 day before harvest. Flag leaf length was measured as the distance from the base to the tip of the uppermost leaf blade of three rice tillers and the mean value is presented as FLL. Root dry matter and SDM were determined by measuring the oven-dried (80 °C) weight of rice root and shoot (excluding panicles) until constant weight was obtained.

SPAD value (leaf greenness) was recorded from the fully expanded three leaves at panicle initiation stage using a handheld chlorophyll meter (SPAD-502 plus, Minolta Corporation, Ltd., Osaka, Japan) and the mean value is presented for each treatment as described by Hussain et al. (2000). At panicle initiation stage, LRWC was determined using the method as outlined by Dasgupta et al. (2015). For determination of LRWC, leaf samples were collected from the fully expanded second leaf from the top of the plant. Leaves were weighed for their fresh weight (FW) immediately after sampling. After collecting FW, leaves were cut into small segments of 2 cm, immersed into distilled water in test tubes, kept overnight in the laboratory, and turgid weight (TW) of the samples measured. The fully turgid leaf samples were then dried in an oven at 80 °C until dry weight was constant followed by the measurement of dry weight (DW) of the samples. The following formula was used for determining LRWC:

$$LRWC\%=\frac{\left(FW-DW\right)}{\left(TW-DW\right)}\times 100$$

Photosynthesis-related parameters, such as Pn, E, and gs of the flag leaf, were measured between 09.30 am and 11.30 am using a portable photosynthesis system (LI-6400XT, Li-COR, Lincoln, NE, USA) at panicle initiation stage. Measurements were started at a concentration of CO2 in air of approximately 370 ± 20 μmol mol–1 in the assimilation chamber. The ambient temperature was 28 ± 1 °C. Artificial illumination from a red-blue 6400-02B LED light source that could release continuous light at 1000 μmol m−2 s−1 photosynthetic photon flux density was used during the measurements (Cha-um et al. 2006). Free proline content was measured from the second fully expanded leaf starting from the top at panicle initiation stage following the method of Bates et al. (1973). For free proline determination, fresh leaf sample was cut into small pieces, ground in the mortar with liquid nitrogen into a powder, and 0.05 g of the sample was homogenized in 1.0 mL of 3% aqueous sulfosalicylic acid followed by centrifugation at 10,000 rpm for 10 min. Then, 200 μL of the extract was put into a test tube with the addition of 200 μL acid ninhydrin and 200 μL glacial acetic acid followed by heating in boiling water for 1 h, and finally the reaction was terminated in an ice bath. After cooling, 200 μL of mixture solution was extracted with 400 μL toluene, mixed vigorously at 6000 rpm for 5 min, kept for complete precipitation, and the upper layer was separated for measuring the value of absorbance at 520 nm. The final concentration of free proline in fresh leaf was calculated from standard curve with absorbance value of different concentrations of standard proline solutions.

Yield components, such as tiller number plant–1, panicle number plant–1, spikelet number panicle–1, filled grain (%), and 1000-grain weight, were determined at harvest. Grain yield was measured and adjusted to 14% grain moisture content. Harvest index was calculated by the ratio of grain yield to biological yield (total biomass). Water productivity was calculated by the ratio of grain yield (kg) to total water input (m3) per pot as described by Maneepitak et al. (2019b).

For measuring shoot K and Si content, shoot sample (consisted of leaf and culm) from each treatment was collected, rinsed thoroughly with deionized water, and oven dried at 72 °C until constant weight. Then, the sample was cut into small pieces, ground to fine powder, and digested with nitric acid via modified USEPA method 3050B using the Hot Block Digestion System (Tarantino et al. 2017). After that, 50 mg of each sample was suspended in 1 mL of 1:1 of nitric acid:deionized water followed by heating at 105 °C for 15 min. After cooling, 2 mL of concentrated nitric acid was added and again digested for 2 h followed by the addition of deionized water for preparing 10 mL solution. This digested solution was filtered through 0.45 µm pore size PTFE filter and filtered samples were diluted with 2.5% nitric acid before the determination of K and Si content via an inductively coupled plasma-optical emission spectroscopy (ICP-OES; Perkin Elmer Avio 200) following Phukunkamkaew et al. (2021). The content of K and Si standard was calibrated on ICP-OES using 1.0 g L–1 of K and Si standard each (Perkin Elmer, 5% nitric acid).

2.6 Statistical Analysis

The data were subjected to a three-way analysis of variance (ANOVA) and were analyzed using Statistix 10 software program (Analytical Software, Tallahassee, FL, USA). Means of significant treatment effects were separated by conducting post hoc analysis using Tukey’s honest significant difference test. In all analyses, differences were considered significant at P ≤ 0.05. Data for significant treatment effect are presented based on the highest order of factorial combination that was significant in the ANOVA.

3 Results

3.1 Shoot K and Si Content

The two-way interaction between fertilizer dose and soil water potential had a significant effect on shoot K and Si content (Table 1). Both nutrients were better accumulated in shoots when NP100 was supplemented with different combinations of K and Si, while accumulation was significantly lower at NP100 fertilizer dose irrespective of soil water potential levels (Fig. 1A,B). Similarly, both K and Si accumulation in shoot was drastically reduced when soil water potential dropped from 0 to –30 kPa. NP100 + K100 fertilizer dose had 9%, 10%, and 10% higher shoot K content than NP100 at 0, –15, and –30 kPa, respectively (Fig. 1A). Shoot K content exhibited a reduction in the range of 10% (NP100 + K75 + Si25) to 16% (NP100 + K50 + Si50) when soil water potential dropped from 0 to –30 kPa. Shoot Si content was not affected by different splits of Si, and NP100 + Si100 fertilizer dose had 24% higher accumulation of Si in shoot than NP100 at 0 kPa (Fig. 1B). Shoot Si content was reduced by 13–21% for different fertilizer doses upon decreasing soil water potential from 0 to –30 kPa.

Table 1 Significance level in the three-way ANOVA of the effect of fertilizer dose, soil water potential, and cultivation method on growth, physiological and biochemical traits, yield components, grain yield, harvest index, water productivity, and nutrient content of rice
Full size table
Fig. 1
figure 1

Interaction effect of fertilizer dose and soil water potential on shoot K content (A) and shoot Si content (B) of rice. Means followed by the same letters are statistically similar among fertilizer doses within a particular soil water potential based on Tukey’s honest significant difference test at P ≤ 0.05. Bars show means ± standard errors of three replications

Full size image

3.2 Growth Parameters

All growth parameters (plant height, FLL, RDM, and SDM) were significantly affected by the main effect of fertilizer dose, soil water potential, and cultivation method, but the two-way and three-way interactions were not significant (Table 1). Plants were the tallest at NP100 + Si100 fertilizer combination, which were 7% longer than the plants at NP100 (Table 2). Plant height progressively decreased with increasing water-deficit stress, while wet direct-seeded plants were significantly longer than transplanted plants. All combinations of K and Si with NP100 had similar FLL, which was maximized at NP100 + K75 + Si25 with a 17% increase from NP100 alone (Table 2). There was no difference in FLL between –15 and –30 kPa, while wet direct-seeded plants had significantly higher FLL than transplanted plants. Plants fertilized with NP100 + K75 + Si25 had 44% higher RDM than those fertilized with NP100 alone (Table 2). Decreasing soil water potential up to –15 kPa had no effect on RDM, which was reduced by 20% at –30 kPa compared with 0 kPa. Root dry matter of transplanted plants was 39% higher than wet direct-seeded plants. Shoot dry matter exhibited a different trend where no difference was observed among all fertilizer combinations, except for NP100 + K50 + Si50 where SDM was 21% higher than NP100 alone (Table 2). Shoot dry matter progressively decreased with decreasing soil moisture regime (8% and 11% at –15 and –30 kPa, respectively, compared with 0 kPa), while wet direct-seeded plants had significantly higher SDM than transplanted plants.

Table 2 Individual effect of fertilizer dose, soil water potential, and cultivation method on plant height, flag leaf length, root dry matter, and shoot dry matter of rice
Full size table

3.3 Physiological and Biochemical Traits

Leaf greenness (SPAD value), LRWC, Pn, and E were significantly affected by the interactive effect between fertilizer dose and soil water potential (Table 1). SPAD value largely remained similar across fertilizer doses under a particular soil water potential level, while decreasing soil water potential up to –15 kPa had no effect on SPAD value regardless of fertilizer doses (Fig. 2A). At –30 kPa, a significant reduction in SPAD value was evident for all fertilizer doses, except for NP100 + K100 combination where decreasing soil water potential did not affect SPAD value. Fertilizer dose of NP100 + K100 had an overall higher LRWC regardless of soil water potential levels, which was reduced by 10% for NP100 at 0 kPa and by 9% and 10% for NP100 + K25 + Si75 at –15 and –30 kPa, respectively (Fig. 2B). Leaf relative water content was significantly reduced with decreasing soil water potential irrespective of fertilizer doses where a maximum reduction of 12% was recorded at –30 kPa compared with 0 kPa for NP100 + K25 + Si75 fertilizer dose. Fertilizers application with no K or Si supplementation (NP100) had an overall lower Pn followed by NP100 + K25 + Si75 fertilizer dose regardless of soil water potential levels (Fig. 3A). There was no significant difference in Pn among other fertilizer doses across soil water potential levels. A significant reduction in Pn ranging from 36% (NP100 + K100) to 45% (NP100 + K25 + Si75) was recorded when soil water potential was reduced from 0 to –30 kPa. Transpiration rate remained lower for NP100 + K100 fertilizer dose at all soil water potential levels, while the other fertilizer doses had largely similar E across soil water potential levels (Fig. 3B). The effect of decreasing soil water potential was evident on E of plants at all fertilizer doses with more profound impact at NP100 where E was reduced by 28% when soil water potential was reduced from 0 to –30 kPa. The three-way interaction among fertilizer dose, cultivation method, and soil water potential had a significant effect on gs and proline content (Table 1). Different fertilizer doses had largely similar gs for wet direct seeding method of cultivation at all soil water potential levels and the same was also true for the two cultivation methods within a particular fertilizer dose and soil water potential (Table 3). Transplanted plants had 37% higher gs at NP100 fertilizer dose than NP100 + K100 and NP100 + K75 + Si25 fertilizer doses at 0 kPa, while the same difference was 35% and 53% compared with NP100 + Si100 and NP100 + K25 + Si75 fertilizer doses, respectively, at –15 kPa. Stomatal conductance significantly reduced with decreasing soil water potential regardless of cultivation methods or fertilizer doses, and a maximum decrease of 50% in gs of wet direct-seeded plants was evident at NP100 + K100 fertilizer dose when soil water potential was reduced from 0 to –30 kPa.

Fig. 2
figure 2

Interaction effect of fertilizer dose and soil water potential on leaf greenness (A) and leaf relative water content (B) of rice. Means followed by the same letters are statistically similar among fertilizer doses within a particular soil water potential based on Tukey’s honest significant difference test at P ≤ 0.05. Bars show means ± standard errors of three replications

Full size image
Fig. 3
figure 3

Interaction effect of fertilizer dose and soil water potential on net photosynthetic rate (A) and transpiration rate (B) of rice. Means followed by the same letters are statistically similar among fertilizer doses within a particular soil water potential based on Tukey’s honest significant difference test at P ≤ 0.05. Bars show means ± standard errors of three replications

Full size image
Table 3 Effect of fertilizer dose, soil water potential, and cultivation method on stomatal conductance and free proline content of rice
Full size table

The three-way interaction among fertilizer dose, cultivation method, and soil water potential indicated that plants cultivated through wet direct seeding method had more proline content than transplanted plants across soil water potential levels and fertilizer doses (Table 3). Proline content ranged from 84.6 µg g–1 (NP100) to 103.3 µg g–1 (NP100 + K75 + Si25) at 0 kPa for wet direct-seeded plants, the corresponding range for transplanted plants was 64.0 µg g–1 (NP100 + K25 + Si75) to 86.1 µg g–1 (NP100 + K75 + Si25). Proline content showed a maximum increase of 27% at NP100 + K50 + Si50 fertilizer dose compared with NP100 for wet direct-seeded plants at –15 kPa and 40% increase at NP100 + K100 compared with NP100 for the same cultivation method at –30 kPa. NP100 + Si100 had 25% and 48% higher proline content than NP100 + K25 + Si75 fertilizer dose for transplanted plants at –15 and –30 kPa, respectively. A significant increase in proline content was observed with decreasing soil water potential across fertilizer doses and cultivation methods. A maximum increase of 58% in proline content of wet direct-seeded plants was evident at NP100 + K100 fertilizer dose when soil water potential decreased from 0 to –30 kPa.

3.4 Yield Components, Yield, and Water Productivity

Tiller number plant–1, panicle number plant–1, spikelet number panicle–1, and filled grain percentage were only significantly affected by the main effect of fertilizer dose, soil water potential, and cultivation method (Table 1). The performance of NP100 fertilizer dose was poor for all these yield components, while other fertilizer doses had up to 19% higher tiller number plant–1 and 26% greater panicle number plant–1 for NP100 + K25 + Si75 and NP100 + Si100, respectively, up to 52% higher spikelet number panicle–1 and 13% higher filled grain percentage (for NP100 + K100) compared with NP100 (Table 4). The detrimental effect of decreasing soil water potential was observed on all these yield components with 13%, 15%, 23%, and 11% reduction in tiller number plant–1, panicle number plant–1, spikelet number panicle–1, and filled grain percentage, respectively, at –30 kPa compared with 0 kPa. Wet direct-seeded plants had 6%, 8%, 9%, and 5% higher tiller number plant–1, panicle number plant–1, spikelet number panicle–1, and filled grain percentage, respectively, compared with transplanted plants. The significant interaction between fertilizer dose and soil water potential (Table 1) indicated that NP100 had lower 1000-grain weight across soil water potential levels, while there was largely no difference in other fertilizer doses irrespective of soil water potential levels (Table 5). A maximum increase up to 17%, 16%, and 17% at 0, –15, and –30 kPa, respectively, in 1000-grain weight was recorded for different fertilizer doses compared with NP100. Decreasing soil water potential caused a significant reduction in 1000-grain weight, which was common for all fertilizer doses, except for NP100 + Si100, with different proportions ranging from 7% (NP100) to 14% (NP100 + K100) at –30 kPa compared with 0 kPa.

Table 4 Individual effect of fertilizer dose, soil water potential, and cultivation method on yield components of rice
Full size table
Table 5 Interaction effect between fertilizer dose and soil water potential on 1000-grain weight and water productivity of rice
Full size table

The three-way interaction among fertilizer dose, cultivation method, and soil water potential was significant for grain yield (Table 1). Wet direct-seeded plants had similar grain yield at 0 kPa regardless of fertilizer doses and the same was true for transplanted plants at –15 and –30 kPa (Table 6). Fertilizing wet direct-seeded plants with NP100 + K100 resulted in 51% and 40% higher grain yield than NP100 at –15 and –30 kPa, respectively. The corresponding difference for transplanted plants at 0 kPa was 32%. The two cultivation methods had no difference in grain yield across fertilizer doses and soil water potential levels. Harvest index was significantly affected by the three-way interaction among fertilizer dose, cultivation method, and soil water potential (Table 1). Harvest index ranged between 36.2% (NP100 + K50 + Si50 for wet direct-seeded plants) and 53.5% (NP100 + K100 for transplanted plants) at 0 kPa (Table 6). The corresponding respective range at –15 and –30 kPa was 26.1% (NP100 for wet direct-seeded plants) to 45.1% (NP100 + K75 + Si25 for wet direct-seeded plants) and 20.2% (NP100 for transplanted plants) to 31.6% (NP100 + K100 for wet direct-seeded plants). The two cultivation methods had largely similar harvest index across fertilizer doses and soil water potential levels. Harvest index showed a significant reduction with decreasing soil water potential and the maximum decrease (52%) was noted for transplanted plants fertilized with NP100 + Si100 at –30 kPa compared with 0 kPa.

Table 6 Effect of fertilizer dose, soil water potential, and cultivation method on grain yield and harvest index of rice
Full size table

The two-way interaction between fertilizer dose and soil water potential had a significant effect on water productivity (Table 1). There was no difference in water productivity of different fertilizer doses at 0 kPa; however, NP100 + K100 had 27% and 42% higher water productivity than NP100 at –15 and –30 kPa, respectively (Table 5). There was a progressive increase in water productivity with decreasing soil water potential across fertilizer doses ranging from 77% (NP100, NP100 + K75 + Si25, NP100 + K25 + Si75) to 88% (NP100 + K100).

4 Discussion

Crop yield is the final output resulted from cumulative action of all metabolic reactions of plants including photosynthesis, respiration, and transpiration, which are adversely affected by water-deficit stress (Vaghar and Ehsanzadeh 2018). It reduces nutrient acquisition from soil, leaf water potential, tissue water content, and cell turgidity resulting in a decrease in E due to a partial stomatal closure that restricts CO2 entry into the leaf for photosynthesis (Yin et al. 2010). Overproduction of ROS in plants is the consequence of water-deficit stress due to disequilibrium between electron excitation and utilization, which is detrimental to cell membranes, photosynthetic pigments, nucleic acids, and proteins in maintaining optimum cellular function (Lubitz et al. 2019). Reactive oxygen species are small molecule metabolites of oxygen that play dual function in plants: (i) as toxic compounds at high concentrations produced under environmental stress conditions causing oxidative cell injuries and (ii) as important signaling molecules for normal biologic processes regulating plant growth and development, and responses to biotic and abiotic stresses (Rejeb et al. 2014; Del Rio 2015; Huang et al. 2019). Chlorophyll degradation is one of the most significant indicators of plant susceptibility and it has been reported that chlorophyll integrity in the photosystems is damaged by drought stress resulting in an altered chlorophyll ultrastructure, inhibition of Rubisco activity with lower light-harvesting capacity, and lower photosynthetic efficiency (Yin et al. 2010; Zhang et al. 2015; Vaghar and Ehsanzadeh 2018). Drought also trims down photoassimilation and inhibits the translocation of photosynthates through stomatal and mesophyll limitation (Zivcak et al. 2014; Najafabadi and Ehsanzadeh 2017). The promotion of root growth by mineral nutrient application under drought conditions has been reported to facilitate the extraction of water and nutrients from deeper soil layers to improve plant tolerance (Hu and Schmidhalter 2005). Potassium and Si both have significant roles in alleviating the negative impact of drought stress in which plants adopt several intrinsic coping mechanisms to mitigate oxidative damages through strong antioxidant defense systems and osmotic adjustment to optimize cellular processes (Vassileva et al. 2009; Abid et al. 2018).

Water-deficit stress during the panicle initiation stage causes a reduction in LRWC, which inhibits carbohydrate metabolism and cell division of floral organ limiting spikelet fertility with lower filled grain percentage (Boyer and Westgate 2004; Najafabadi and Ehsanzadeh 2017). Spikelet sterility of rice under drought stress is a very common phenomenon, which might be due to decreased pollen viability and receptivity of the stigmatic surface impairing pollination and fertilization of embryo to produce grain (Gonzalez et al. 2019). Grain filling is also affected by impaired photosynthesis with limited available photosynthates, which is not enough for sufficient sink formation (Fu et al. 2011). Remobilization of photoassimilate from leaf to the grain and translocation of photosynthates from source to sink are impaired due to water-deficit stress negatively affecting grain filling with lower filled grain percentage (Rang et al. 2011). In the present study, yield components, such as tiller number plant–1, panicle number plant–1, and 1000-grain weight, were decreased with an increasing severity of water-deficit stress resulting in an increase in the total number of sterile spikelet panicle–1, which might be due to an assimilate shortage and inhibition of photoassimilate translocation to the developing grain during the grain filling stage. It has been reported that limited supply of assimilate to the developing grain (source limitation) is one of the primary causes of grain yield reduction of rice under limited soil water availability (Zhang et al. 2015; Ullah et al. 2020). In the present study, grain yield and yield components of rice were severely impacted by water-deficit stress. Similarly, Yang et al. (2019) also observed that drought stress at the flowering stage of rice had a strong influence on physiological traits, yield components, and yield. Higher water productivity at lower soil moisture regime was observed across all fertilizer combinations in comparison to 0 kPa (Table 5). This is primarily due to lower total water input rather than higher grain yield.

Potassium has a substantial role on enzyme activation, protein synthesis, photosynthesis, stomatal regulation, and plant-water relations (Sharma et al. 2012; Zain et al. 2014). The present study revealed that most of the physiological parameters, such as leaf greenness, LRWC, Pn, E, and gs, were highly influenced by K fertilization (NP100 + K100) under reduced soil water potential (− 30 kPa). Maximum leaf greenness and LRWC were observed in plants treated with NP100 + K100 indicating that K enhances water and nutrient uptake of rice most likely by penetrating roots deeper in soil profile, thereby enhancing plant’s ability to cope with drought stress (Römheld and Kirkby 2010). This might be due to the K-mediated promotion of root growth resulting in an increase in root surface area, which ultimately enhances uptake of water and mineral nutrients, thereby regulating stomatal functions by maintaining osmotic and turgor potential of cell under water-deficit stress (Zain and Ismail 2016). These enhanced water uptake and mineral acquisition confer greater chlorophyll integrity (stay green) in rice leaf as was also evident during the present study. It is well known that K is a key osmotic driver of plant cell under osmotic stress, while N and Mg are structural and functional components of chlorophyll (a and b) present in photosynthetic apparatus (chloroplast ultrastructure), consisted of grana and stroma. In addition, K and Mg both expedite well-structured organization of grana and stroma lamellae for chlorophyll integrity that improves efficiency of light absorption and Rubisco diffusion to facilitate carbon assimilation (Tränkner et al. 2018). Our results are also in line with Kanai et al. (2011) who observed that K fertilization significantly controlled E and also had a protective role in chlorophyll content of leaf under water-deficit stress. Similarly, a drastic reduction in gs and E was observed under K fertilization (NP100 + K100) indicating that K+ significantly reduces E by decreasing osmotic potential of mesophyll cells (Li et al. 2017). Pumping K+ from guard cell causes a reduction in turgor pressure inducing stomatal closure, which is the initial response of plants exposed to drought stress (Taiz and Zeiger 2010). Stomatal closure reduces transpirational water loss with a simultaneous decrease in stomatal CO2 influx resulting in a decline in Pn (Chaves et al. 2009). Under water-deficit stress, K plays an essential role in activating the ATP synthase enzyme and regulating the stomatal aperture through balancing CO2 entry and water vapor removal from intercellular spaces to optimize CO2 fixation and the utilization of photoassimilate (Waraich et al. 2012; Wang et al. 2017; Li et al. 2018). The present findings are in close agreement with these researchers where plants fertilized with NP100 + K100 exhibited better Pn over other fertilizer doses, especially at the lowest soil water potential of –30 kPa. Osmotic adjustment is the key adaptation of plants at cellular level to minimize the harmful effects of drought stress through maintaining leaf turgor for improved gs and efficient intake of CO2 (Farooq et al. 2009, 2010). Likewise, free proline is the most important compatible solute accumulating in plants exposed to drought stress, which allows plants to maintain a low osmotic potential with improved gas exchange and strong antioxidant systems nullifying the lethal effect of oxidative stress on cells against dehydration (Teixeira and Pereira 2007; Ellouzi et al. 2017). Higher accumulation of free proline in rice leaf in response to K fertilization at the lowest soil water potential (–30 kPa) was evident in the present study, which might be due to K-induced stimulation of biosynthesis of various osmolytes (e.g., free proline, soluble sugar, and glycine betaine) crucial for cellular function under water-deficit stress (Zain et al. 2014; Raza et al. 2014). An improvement in physiological and biochemical performance of rice is closely associated with K-mediated detoxification of ROS, rapid plant signaling systems, activation of antioxidant defense systems, and biosynthesis of osmolytes, which enable plants to cope with drought stress (Demidchik 2014; Jatav et al. 2014).

The highest spikelet number panicle–1 and filled grain percentage were observed in plants fertilized with K along with N and P (NP100 + K100) indicating that K has a significant role in grain formation and in determining the final grain yield of rice (Table 4). An improved water and nutrient supply in plants raised from K fertilization might have resulted in better mobilization and translocation of nutrients towards panicle improving carbohydrate metabolism, which is critical for enhanced grain filling. It was observed that different combinations of K- and Si-fertilized rice plants had a steady increase in yield components, grain yield, harvest index, and water productivity compared with the control (NP100), which were maximized at NP100 + K100 in terms of grain yield and harvest index when soil water potential was the lowest (–30 kPa). Higher grain yield with K fertilization even at severe soil moisture availability could be attributed to more chlorophyll biosynthesis, improved stomata regulation, greater enzyme activity, and enhanced photosynthetic efficiency resulting in more carbohydrate accumulation and translocation for grain formation (Divito and Sadras 2014; Islam and Muttaleb 2016). It has been reported that K fertilization enhances N and P uptake leading to luxurious vegetative growth with delaying physiological maturity, which prolongs grain filling period resulting in more grain yield (Ye et al. 2019). Luxurious vegetative growth inevitably enhances leaf area, which is among the major determinants of biomass accumulation, especially in cereal crops. More assimilate formation and its partitioning to the developing grains helps in maximizing the percentage of filled grains and ultimately grain yield, which is facilitated by longer vegetative growth. Potassium regulates the partitioning of carbohydrates, which is crucial for improving grain filling through remobilization of assimilates from source to sink under suboptimal environments (Zahoor et al. 2017); otherwise, traffic load in phloem due to accumulated carbohydrates might decrease the rate of photosynthesis (Pan et al. 2017; Liang et al. 2017; Zhang et al. 2019). It can be concluded that exogenous soil application of K improves grain yield, harvest index, and water productivity of rice through physiological and biochemical amelioration under water-deficit stress.

Moreover, Si is considered a multi-talented element and it benefits rice under water-deficit stress because of its enhanced water retention capacity (Zargar et al. 2019), increased chlorophyll content (Yin et al. 2010), and improved leaf orientation for maximum light interception (Yoshida et al. 1969). Root growth is another vital indicator of plant tolerance against water-deficit stress, and it was observed that different combinations of Si with K fertilization on rice had better root proliferation over the control (NP100). Similarly, SDM was maximized at NP100 + K50 + Si50 indicating that shoot growth is influenced equally by both K and Si fertilization. This luxurious vegetative growth under combination of K and Si fertilization suggests that Si could have potential role on SDM and RDM accumulation through increased cell division and enlargement of apical meristem. The results of the present study are in close agreement with Ullah et al. (2019a) who observed that root system of rice was positively impacted by K fertilization at 120 kg ha–1, while application of monosilicic acid at 300 kg ha–1 (60 kg ha–1 soluble Si) enhanced rice SDM with improved grain yield by 34–45% (Ullah et al. 2018b). This result might be due to Si-mediated improvement of root hydraulic conductance and root activity resulting in an efficient absorption of water and mineral, especially P, from soil to maintain a higher photosynthetic rate with more dry matter production under water-deficit stress (Chen et al. 2011; Luyckx et al. 2017; Schaller et al. 2021). Similarly, Schaller et al. (2020) reported enhanced soil hydraulic conductivity and improved water holding capacity of Si-fertilized soil, which are the major reasons of Si-mediated improvement in growth and productivity of most crops. The authors argued that it is the increased water retention in soil rather than Si accumulation in plants, which helps plants sustain under stressful conditions (Kuhla et al. 2021). The better growth and yield attributes at lower soil water potential of –30 kPa in Si-fed plants observed in the present study could be attributed to better soil moisture retention and a concomitant increased plant available water in soils compared with plants without Si supplementation. Moreover, Si can enhance root water uptake under drought stress through active accumulation of soluble sugars and amino acids (Zhu and Gong 2014). Silicon-mediated improvement in growth and yield of rice under water-deficit stress has been largely credited to an enhanced water use efficiency, which is an important determinant of crop productivity (Gomaa et al. 2021).

In the present study, a reduction in soil water potential caused a significant decrease in Pn across all fertilizer combinations, and plants fertilized with NP100 + Si100 had 42% higher Pn over the control (NP100) at severe drought stress (− 30 kPa), which was also statistically at par with almost all other combination of fertilizers, indicating that both K and Si had a positive role in maintaining Pn under limiter water availability (Fig. 3A). This improved Pn of Si-fertilized plants (NP100 + Si100) under water-deficit stress might be due to a Si-mediated increase of intercellular CO2 concentration of leaf with moderate gs ensuring an ample supply of CO2 for optimum photosynthesis, which is a pre-requisite for grain formation (Li et al. 2020). This claim has been also supported by Chaves et al. (2009) and Kang et al. (2016) who mentioned that photosynthesis is the physiological basis of plant growth and development and Si improves drought tolerance by reducing oxidative damage of lipid and protein with a significant rise in Pn. Stomatal conductance and E are important parameters influencing plant-water relations (Farooq et al. 2009) and the formation of a silica-cuticle double layer on leaf epidermal tissue might be responsible for a reduction in leaf transpiration (Yoshida et al. 1969). A reduction of cuticular and stomatal transpiration along with an improved photosynthetic C fixation is one of the Si-mediated drought tolerance mechanisms in plants (Zhu and Gong 2014). Despite an overall reduction in growth and yield at lower soil moisture regimes, the beneficial effect of Si was evident on panicle number plant–1 and 1000-grain weight, which were significantly improved with NP100 + Si100. Silicon-mediated water and mineral nutrient supply in plants might have resulted in an enhanced mobilization and translocation of nutrients towards panicle improving carbohydrate metabolism, a pre-requisite for grain filling (Schaller et al. 2021). This better performance in growth and yield of Si-fertilized rice at lower soil moisture regime could be credited to more accumulation of carbohydrate as also reported by Crooks and Prentice (2017) and Das et al. (2021a). Ullah et al. (2018a) also reported a significant improvement in panicle number plant–1, 1000-grain weight, and grain yield of rice with Si application as monosilicic acid. Our results are also consistent with Cuong et al. (2017) who reported that Si application in combination with the recommended dose of NPK fertilizer enhanced grain yield and nutrient uptake of rice. High shoot Si content under Si supplementation in rice was observed in the present study, which might be resulted from the deposition of Si on cell wall strengthening cell membrane and altering membrane permeability (Wang et al. 2020). We observed no notable difference in yield and physiological performance of rice under the two studied cultivation methods. Therefore, Si fertilization holds promise irrespective of cultivation methods for enhancing grain yield of rice under water-deficit condition.

5 Conclusion

Growth, physiological traits, and yield of rice were adversely affected by reduced soil water potential level maintained through AWD irrigation. Exogenous soil application of different combinations of K and Si fertilizers were beneficial even at the lowest soil water potential (− 30 kPa). Growth, yield, and physiological response of rice were poor when no K or Si was applied in combination with N and P (NP100). There was no considerable difference among different K and Si fertilizers combination for most of the evaluated parameters; nevertheless, reducing K share below 50% was not effective for some parameters, such as LRWC, spikelet number panicle–1, and Pn. The two cultivation methods (wet direct seeding and transplanting) had largely similar response. Although the negative effect caused by water-deficit stress cannot be fully compensated by K and Si fertilization in rice, it could be minimized for better growth, yield components, and yield through physiological and biochemical regulation under limited water supply. Thus, combining K and Si with major nutrients (N and P) in a fertilizer management program where the share of K is at least 50% could be a promising approach to minimize the harmful impact of water-deficit stress in rice established through either wet direct seeding or transplanting method.