Water saving rice cultivation using sheet-pipe subsurface irrigation

Abstract

Water saving in rice cultivation has assumed paramount importance, especially in the context of climate change. The introduction of sheet-pipe technology in Indonesia heralded as an innovative subsurface irrigation and drainage system, is poised to revolutionize how to manage this vital resource. Our study was designed with two primary objectives: first, to investigate how rice plants respond when water levels are deliberately reduced using the sheet-pipe technology; and second, to comprehensively analyze water productivity and water use efficiency in comparison to conventional flooded rice cultivation systems. We conducted two distinct experiments: one employing sheet-pipe subsurface irrigation (SSI) and the other utilizing conventional flooded irrigation (CFI). In the SSI setup, the water level was maintained at a depth of 5–10 cm below the soil surface 20 days after transplanting to harvesting. With this setting, the soil moisture was maintained at around 85–95 degrees of saturation. On the other hand, the CFI approach involved water flowing directly over the soil surface, with the water level consistently maintained at a mere 2–3 cm above it. Interestingly, while the SSI method did lead to a reduction in yield, it has significant benefits. Our results showed that a reduction in yield was observed for the SSI 15.5–18.6 % lower compared to the conventional method (CFI). However, the SSI is environmentally benefit compared to the conventional method by reducing 37.5–50.5 % in water irrigation, increasing water use efficiency (WUE) up to 70.8 %, and improving 3.2–10.4 % in water productivity. Our findings reveal that optimizing water conservation may have a disadvantageous effect on rice yield, indicating the importance of optimal water level. Future research to find the optimal water level that balances yield production and environment is required, especially to adapt to dry and warming climate change in the future.

1. Introduction

Water management plays a pivotal role in the sustainable production of rice, one of the world's most essential staple crops. The cultivation of rice is deeply intertwined with global efforts to achieve the United Nations Sustainable Development Goals (SDGs), particularly those related to poverty alleviation, food security, and responsible resource use. As the global population continues to expand, and climate change poses unprecedented challenges to agriculture [1,2], efficient and sustainable water management in rice production becomes increasingly critical [3].

Improving the efficiency of irrigation water usage in agriculture is a critical imperative for ensuring the long-term sustainability of agricultural production [4]. The looming specter of water scarcity in agriculture, coupled with the inexorable impact of climate change and intensifying competition for water resources across industrial and domestic sectors, underscores the urgency of this challenge. Within the realm of agriculture, rice farming stands out as one of the most voracious consumers of water resources [5]. The demands placed on paddy fields are comprehensive, spanning from initial land preparation to the continuous inundation of fields during plant cultivation. Notably, this inundation brings forth a set of advantages, including improvements in soil chemical fertility, the accumulation of organic matter, and the enhanced availability of vital micronutrients [6,7].

However, it's imperative to recognize that this flooded water regime also stimulates the activity of methanogenic bacteria, leading to an increased emission of methane gas, a potent greenhouse gas originating from agricultural fields [8]. This uptick in methane emissions is substantial, reaching up to 90 % higher levels when compared to water-saving irrigation systems with reduced water usage [9]. Such an increase in methane emissions carries direct implications for the amplification of global warming potential. Moreover, the inundation system also exacerbates water loss, particularly through percolation, with water loss rates surging by as much as 18.7 % due to elevated hydrostatic pressure [10].

The influence of flooded water in rice fields extends to crop yields, with empirical data revealing significant variability within the flooded system. While some studies have reported substantial yield increases under flooded irrigation conditions [11,12], others employing alternative cultivation systems, such as the System of Rice Intensification (SRI) with intermittent irrigation, have yielded contrasting results. The SRI system, combined with intermittent irrigation, not only boosts yields but also enhances water productivity [[13], [14], [15], [16]]. Consequently, the optimization and controlled management of irrigation systems remain paramount concerns in rice farming, with profound implications for land and water productivity, as well as water use efficiency.

Several technologies have been developed to address the challenge of controlling water levels in paddy fields. These innovations include automatic control systems [17] driven by intelligent algorithms [18] and fuzzy logic-based approaches [19]. Currently, a noteworthy technology that has gained traction in Indonesia is the sheet-pipe technology [20]. Sheet-pipe technology, also known as subsurface irrigation and drainage, is an innovative and efficient water management system used in rice production and other agricultural contexts. It offers a more controlled and precise approach to water management compared to traditional flood irrigation methods. Typically installed at depths of 30–50 cm below the soil surface and positioned horizontally using mole drainer machines, these sheet pipes automatically convert into perforated pipes [21]. Unlike traditional flooded rice cultivation, where fields are continuously submerged, sheet-pipe technology enables precise control over water delivery [22]. Nevertheless, the adoption of sheet-pipe technology in Indonesia has predominantly centered on drainage systems, with no documented studies on its application for subsurface irrigation.

In fact, the technology of subsurface irrigation, predominantly utilizing drip irrigation for multi-crops such as rice and wheat [23,24], has garnered widespread adoption. Extensive research has demonstrated its capacity to save more water irrigation and enhance water use efficiency significantly. Despite its efficacy, the subsurface drip irrigation necessitates proximity between lateral pipes, typically ranging between 33.75 and 67.5 cm [23], leading to substantial equipment expenses and operational costs. On the other hand, the adoption of sheet-pipe technology allows for significantly wider lateral pipe spacing, spanning up to 4 m [20], thereby reducing the dependency on lateral pipes and mitigating associated expenses. Therefore, this study initiates the utilization of sheet-pipe technology for subsurface irrigation, aimed at precisely controlling water levels within rice fields. The objectives were to examine the rice plants' response to the deliberate reduction in water levels facilitated by sheet-pipe technology, to analyze its water productivity and water use efficiency, and then compare it to the traditional flooded rice cultivation systems.

2. Materials and methods

2.1. Experimental setup

The study was conducted at Kinjiro Farm in Bogor, West Java, Indonesia for two planting seasons. The first planting season started on April 2, 2021, and ended on July 18, 2021, with a total of 107 planting days and the second planting season was on October 8, 2022 to January 23, 2023 with a total of 105 planting days. The rice variety used was IPB 3S, which was planted using the system of rice intensification (SRI) method. This rice variety was used since it is one of the new plant type (NPT) rice varieties currently planted by farmers in Indonesia with a relatively dense panicle and fewer numbers of tillers but most of them are productive. Some SRI elements were: the seedling time was 14 days, and the spacing between plants was 30 × 30 cm² with a single transplanting in every hill. The fertilizers used were a combination of compost and chemical fertilizers. A dose of 1 ton/ha of compost and 100 kg/ha of urea, 75 kg/ha of phosphorus (SP-36), and 50 kg/ha of KCl for chemical fertilizers were applied in the first season. The doses of fertilizers were applied based on the recommendation of the Ministry of Agriculture Officer by considering the location with modification. Based on Sato et al. [25]. 2011, with the utilization of organic fertilizer, chemical fertilizer can be reduced 50 % doses from its standard. However, in the second season, more doses of fertilizers were given to anticipate more fertilizers dissolution into drainage water because the second season was conducted in the wet season. Thus, 2 ton/ha of compost, 300 kg/ha of urea, 75 kg/ha of phosphorus (SP-36), and 100 kg/ha of KCl were applied in the second season.

Two experimental boxes with dimensions of 2 x 2 × 0.5 m³ were used to compare two different irrigation systems: sheet-pipe subsurface irrigation (SSI) and conventional flooded irrigation (CFI) systems (Fig. 1a and b). In the SSI setup, water was distributed through sheet-pipes situated 40 cm below the soil surface (as depicted in Fig. 1a). The sheet-pipe is made from high-density polyethylene (HDPE), a material adept at being molded into a pipe shape using a piper machine. Essentially, it comprises a perforated sheet with a standardized pattern featuring 22 holes per 3 cm along its length, each hole measuring 2 mm in diameter. Upon transformation into a pipe, it achieves a diameter of 5 cm with a thickness of 1 mm. Conveniently, the sheet pipe is conveniently packaged in rolls, each extending to 100 m in length [21]. In each box, two sheet-pipe lines were set up. These sheet-pipes were connected to both inlet and outlet pipes. Initially, during the first 20 days after transplanting (DAT), the water level is meticulously maintained at the soil surface level. Subsequently, it is adjusted to hover at a height of 5–10 cm below the soil surface. According to Hasanah et al. [26], this water level was the optimum water level to mitigate greenhouse gas emissions with the lowest methane emission and maintained productivity. Notably, ten days prior to harvest, the experimental box was dried.

Fig. 1.

The experimental boxes: a) the SSI treatment, b) the CFI treatment.

Conversely, in the CFI configuration, irrigation water was delivered through an inlet pipe situated above the experiment box's soil surface. The water level within the box was consistently upheld at a height of 2–3 cm above the soil surface, maintaining continuous flooding. This water level was monitored through a bucket placed at the inlet, serving as an indicator (as illustrated in Fig. 1b). The bucket's water level was impeccably regulated by means of a floating valve that automatically opens and closes at specific water levels. The bucket remains closed to prevent water surface evaporation. This ingenious system ensures automated irrigation, eliminating the need for electric pumps, and it maintains the water level based on the evapotranspiration rate observed in the field. This model is aptly referred to as an evapotranspirative irrigation system [27].

The soil has physical characteristics as presented in Table 1. The soil texture was silt loam with a silt content of more than 50 %. Soil permeability is 2.02 cm/h with a bulk density of 0.83 g/cm³. Soil moisture conditions for saturated water content, field capacity, and permanent wilting point are 0.466, 0.432, and 0.207 m³/m³, respectively. Based on the data, the relationship between soil moisture (soil water content) and soil water potential was then determined. It was represented by a soil water retention curve (SWRC). The van Genuchten model, which is the most widely used SWRC model, was used to represent this relationship [28,29]. The model is represented by Equation (1) as follows [30]:

$$\frac{\theta -{\theta }_r}{{\theta }_s-{\theta }_r}={\left[1+\left(\alpha {\left|h\right|}^n\right)\right]}^{-m}$$ (1)

where θ is the volumetric water content (m³/m³), θ_r is the residual water content (m³/m³), θ_s is the saturated water content (m³/m³), h is the water pressure (kPa), α is a scaling parameter that is inversely proportional to mean pore diameter (cm⁻¹), n is the soil water characteristic curve index (shape parameter of the curve), and m = 1 − 1/n. The parameters for the van Genuchten model (Equation (1)) were calibrated using MS Excel solver integrated within MS Excel, which leveraged the observed data provided in Table 1. The solver is a powerful tool technique for the optimization process and the development of the rating curve [31]. Fig. 2 illustrates the output of this fitting process, revealing the optimized values for van Genuchten's parameters, with α set at 0.006 cm⁻¹ and n at 1.246.

Table 1.

Soil physical properties of the box experiment.

No	Parameters	Value	Unit
1	Dry Bulk Density	0.83 ± 0.04	g/cm3
2	Particle Density	2.21 ± 0.10	g/cm3
3	C-Organic	5.09 ± 0.31	%
4	Organic Content	8.78 ± 0.54	%
5	Permeability	2.02 ± 0.09	cm/hour
6	Soil Texture:
	Sand	22 ± 2	%
	Silt	55 ± 5	%
	Clay	23 ± 7	%
	Type name	Silt Loam
6	Water content:
	Saturated (SAT)	0.466 ± 0.013	m3/m3
	Field Capacity (FC)	0.432 ± 0.001	m3/m3
	Permanent Wilting Point (PWP)	0.207 ± 0.005	m3/m3

Note: Two soil samples were collected and were analyzed in a certified laboratory. The data in the table are the mean ± SD.

Fig. 2.

A soil water retention curve by the van Genuchten's model.

2.2. Weather, soil, and water parameters measurements

In the location, specific sensors were used to measure several parameters such as weather, and soil attributes including moisture, temperature, electrical conductivity (EC), and water level. The automatic weather station (AWS) Vantage Pro2 by Davis Instruments Corp. Inc, USA was used to measure weather parameters such as air temperature, air humidity, solar radiation, wind speed, and rain every 30 min. These invaluable weather records were stored in the console, offering readily accessible data for analysis. The daily weather parameter data were used to determine the reference evapotranspiration (ETo) according to the Penman-Monteith model, which is adopted by the FAO as a standard model [32].

Simultaneously, the soil parameters were measured using advanced 5-TE sensors, seamlessly integrated with the EM50 Data Logger from Meter Group Inc., USA and it was placed at 5 cm below the soil surface to measure soil moisture, soil temperature and soil EC. The water levels were expertly gauged with the aid of pressure sensors and e-Tape sensors, thoughtfully linked to the EM50 datalogger. Irrigation was quantified through the utilization of a water meter strategically positioned post water storage. The process of irrigation was automated, functioning in tandem with a floating valve system to regulate water levels, as previously outlined. Similarly, drainage operations were automated by configuring the maximum height of the outlet pipe. When the elevation of the plot surpassed the designated maximum height of the outlet pipe, drainage commenced. It's noteworthy that any irrigation or drainage rates below the recorded minimum of 0.03 m³/h were not accounted for. Consequently, to ensure accuracy, adjustments to these variables should be made through a process of variable optimization. The collected data were then used for water balance analysis through optimization process in both irrigation regimes as well as quantifying water productivity and irrigation use efficiency.

2.3. Plant performances monitoring

At regular intervals of every three days, we conducted assessments of the morphological attributes of the rice crop. Within each plot, we quantified parameters such as plant height, tillers, and panicle numbers, obtaining data from a total of twelve samples per plot. Upon reaching the harvest date, a comprehensive examination and weighing process were carried out, encompassing key factors such as biomass (straw), grain yield, and panicle numbers across all plots. These recorded values were subsequently converted to standardized units, measured in tons per hectare, for grain yield and biomass weight, ensuring consistency and ease of comparison.

2.4. Data and statistical analysis

Some variables such as crop evapotranspiration, irrigation, and drainage were determined and optimized by MS Excel Solver on a daily basis based on a previously developed method [33,34]. MS. Excel Solver works as a linear programming method by minimizing Equation (2) as the objective function (OB).

$$OB=\sum \left|{\Delta WL}_o-{\Delta WL}_m\right|$$ (2)

Where ΔWL_o is the observed change of water level (in mm) and ΔWL_m is the model of change of water level (in mm). ΔWL_m is determined by Equation (3) as follow:

$${\Delta WL}_m\left(\mathrm{t}\right)=\mathrm{I}\left(\mathrm{t}\right)+\mathrm{R}\left(\mathrm{t}\right)-\text{DR}\left(\mathrm{t}\right)-\text{ETc}\left(\mathrm{t}\right)-\text{DP}\left(\mathrm{t}\right)$$ (3)

where I is irrigation (in mm), R is rainfall (in mm), DR is drainage (in mm), ETc is crop evapotranspiration (in mm), and DP is percolation (in mm). Since the experiment was conducted on the lab scale, DP was set to zero.

During the optimization process, initial rates of DR and I were given at a measured level and then adjusted as earlier mentioned by minimizing OB as well as ETc with the constraints as represented by Equation (4).

$$I\ge 0;DR\ge 0;{ETc}_{\mathit{min}}\le ETc\le {ETc}_{\mathit{max}}$$ (4)

Where ETc_min is crop evapotranspiration minimum (in mm), ETc_max is crop evapotranspiration maximum (in mm). ETc_min and ETc_max were calculated by multiplying ETo to minimum and maximum crop coefficient (Kc) value by the FAO [32]. Since the numbers to be optimized are limited, the optimization process was divided into four growth stages, i.e., initial, vegetative, mid-season, and late-season stages, respectively. The Solver operated iteratively through a search algorithm until all constraints and optimality conditions were satisfied. The optimal variables values (DR, I, ETc) were obtained when it has reached the convergence condition and to be unique solutions. Then the Kc value in each treatment was calculated based on optimized ETc and ETo by Equation (5).

$$Kc=\frac{{ET}_c}{{ET}_o}$$ (5)

Total irrigation during each season is then used to determine the water use efficiency index and water productivity. The water use efficiency index was calculated based on Equation (6) as follow [35]:

$$\text{WUE}=\frac{\mathrm{Y}}{\mathrm{I}}$$ (6)

where WUE is the water use efficiency index (in kg/m³).

Water productivity was calculated in two terms, i.e., total yield per total water input (irrigation + rainfall) and total yield per total crop evapotranspiration as formulated in Equations (7), (8) as follow [36]:

$${\text{WP}}_{\mathrm{I}+\mathrm{R}}=\frac{\mathrm{Y}}{\mathrm{I}+\mathrm{R}}$$ (7)

$${\text{WP}}_{\text{ET}}=\frac{\mathrm{Y}}{\text{ETc}}$$ (8)

where WP_I + R is water productivity concerning water input in kg grain/m³ water, WP_ET is water productivity concerning total evapotranspiration in kg grain/m³ water, and Y is total yield in ton/ha.

Statistical analysis was performed by t-test (α = 0.05) to observe the effect of the treatments on the plant performances, such as number of panicles per hill, panicles weight per hill, main panicle length, total number of grains, number of filled grains, number of unfilled grains and 1000-grains weight.

3. Results

3.1. Actual field conditions under SSI and CFI treatments

In Fig. 3a and b, an overview of weather parameters during the seasons in terms of air temperature, air humidity, solar radiation, and reference evapotranspiration is presented. Throughout the growing seasons, air temperature and humidity exhibit remarkable stability in both seasons. The inaugural season exhibits higher average air temperature, solar radiation, and reference evapotranspiration in comparison to the subsequent season. During the first season (Fig. 3a), the mean values for air temperature, solar radiation, and evapotranspiration were 26.78 °C, 12.07 MJ/m²/d, and 2.65 mm, respectively. In contrast, the corresponding figures for the subsequent season were slightly lower at 25.96 °C, 11.25 MJ/m²/d, and 2.35 mm (Fig. 3b). Conversely, the relative humidity during the first season was lower than that of the second season, with values of 81.38 % and 84.59 %, respectively. The peak air temperature reached 35.2 °C during both the initial and subsequent seasons. Specifically, in the first season, the maximum air temperature was observed in April and May of 2021. In contrast, during the second season, the highest air temperature was recorded in November 2022.

Fig. 3.

Weather parameters during the season: a) first season, b) second season.

The water level dynamics over a growing season revealed distinct characteristics between the CFI and SSI treatments as depicted in Fig. 4a and b. Throughout the seasons, the CFI regime consistently maintained a water level notably higher than the soil surface, with only a brief drop observed around 102 days after transplanting, resulting in a temporary descent of 5.2 cm below the soil surface in the first season. On average, the water level at CFI remained approximately 1.5 cm and 2.3 cm above the ground for the first (Fig. 4a) and second seasons (Fig. 4b), indicative of continuous inundation. In contrast, the SSI system exhibited a different profile, with the water level consistently positioned below the surface. The average depth of the water level at the SSI was measured at 7.2 cm and 2.3 cm below the soil surface for the first and second seasons, signifying a comparatively drier condition when compared to the CFI.

Fig. 4.

Actual water levels during the planting period in the SSI and CFI regimes: a) first season, b) second season.

The water level within both irrigation treatments exhibits a discernible positive correlation with soil moisture, as illustrated in Fig. 5a and b. The continuous flooding employed in the CFI led to the soil moisture consistently residing at saturation levels, maintaining an average soil moisture content of 0.498 m³/m³ and 0.480 m³/m³ for the first and second seasons, respectively. Conversely, in the case of SSI, soil moisture levels displayed significant variations, ranging from saturation to the field capacity condition. Under the SSI in the first season (Fig. 5a), saturated conditions were notably prominent during the early phase of 0–50 DAT, transitioning towards a gradual decrease in soil moisture, moving between saturation and field capacity from 50 to 93 DAT, and ultimately gravitating towards the range between field capacity and the permanent wilting point until the time of harvest. The respective average soil moisture content during these three distinct periods was measured at 0.463 m³/m³, 0.453 m³/m³, and 0.306 m³/m³. Meanwhile, During the second season (Fig. 5b), the soil moisture levels in the SSI remained consistently in between saturated and field capacity levels from the initial stages to the 80 DAT. Subsequently, a notable increase in soil moisture occurred between 85 and 90 DAT, returning to saturated conditions. At the end of the planting period, there was a discernible decline in soil moisture, dropping below the established field capacity levels. The overall average soil moisture for the second planting period at the SSI regime was measured at 0.392 m³/m³.

Fig. 5.

Soil moisture fluctuations in the SSI and CFI regimes: a) first season, b) second season.

Simultaneously, soil temperature fluctuations were observed under both irrigation treatments, and they exhibited a similar pattern. In the CFI, the soil temperature consistently showed a tendency to be higher than that in the SSI, as depicted in Fig. 6a and b (in the primary y-axis). Notably in the first season, soil temperature displayed an upward trajectory during the 10–45 DAT period, followed by a relatively stable phase from 45 to 70 DAT, and then a subsequent decline. The average soil temperature in the first season recorded for the CFI and SSI was 27.3 °C and 26.9 °C, respectively (Fig. 6a). During the second season, there was a notable fluctuation in the soil temperature trend (Fig. 6b). It initially declined from the onset of planting until reaching 80 DAT, followed by a subsequent rise to 95 DAT. Towards the end of the season, the soil temperature exhibited a decrease coinciding with a reduction in the soil moisture. In this season, the average soil temperatures were recorded at 26.0 °C for the CFI and 25.7 °C for the SSI. The elevated soil temperature in CFI can be attributed to the influence of the warmer water temperature applied to the soil surface. Extensive research has indicated that irrigation treatments can significantly impact soil temperature, which, in turn, has implications for parameters such as Chlorophyll content (SPAD) and Leaf Area Index (LAI) [37,38].

Fig. 6.

Soil temperature and electrical conductivity fluctuations in the SSI and CFI regimes: a) first season, b) second season.

Both irrigation systems exhibited a noteworthy impact on soil electrical conductivity (EC) dynamics, as visualized in Fig. 6a and b (secondary y-axis). The CFI, characterized by its sustained inundation, has more consistent and higher soil EC values compared to the SSI. In the first season (Fig. 6a), the soil EC values in the SSI exhibited higher levels only during specific intervals, namely, 10–25 DAT and 72–87 DAT, and in the second season (Fig. 6b), solely during 25–35 DAT. Outside of these periods, the soil EC values in the SSI consistently remained lower than those in the CFI. Notably, the average soil EC for SSI were at 0.647 mS/cm and 0.282 mS/cm for the first and second seasons, respectively. In contrast, the CFI maintained higher average soil EC values, recorded at 0.742 mS/cm for the first season and 0.765 mS/cm for the second season. This observation underscores the CFI sustained soil EC values that were 12.8 %–63.1 % higher than those observed in the SSI. This illustrated that irrigation patterns, as well as drainage practices, play a crucial role in shaping the fluctuations in the soil EC [39]. These variations, in turn, exert a discernible impact on crop productivity [40].

3.2. Plant response to different water irrigation

Plant height in both the CFI and SSI regimes exhibited a similar trend, as illustrated in Fig. 7a and b. From the 30–70 days after planting, there was a significant increase in plant height, followed by a phase of relative constancy from the 70^th to the 101^st days after planting. Notably, the SSI, despite employing lower water irrigation, demonstrated slightly greater plant height towards the end of the growing season, although this difference was not statistically significant. The final plant heights recorded for the SSI were 156.2 cm and 167.1 cm in the first and second seasons, respectively. These plant height averages were 2.8–3.4 % higher than those in the CFI regime.

Fig. 7.

Average plant height during the planting period in the SSI and CFI regimes: a) first season, b) second season.

In contrast, the number of tillers and panicles per hill showed distinct variations between CFI and SSI, as depicted in Fig. 8, Fig. 9a and 9b, respectively. The CFI consistently maintained a higher count of tillers compared to the SSI, with this disparity emerging from the 30 days after planting and persisting until harvest. The tiller count increased steadily from the 30^th - 50^th days after planting, remaining relatively stable until harvest. Panicle formation was observed during the generative period after the 66^th and 78^th day following planting in the first and second seasons, with panicle numbers maintaining consistency from the 80–85 days after planting onwards, indicating that ripening occurred around the 80–85 DAT. The first season (Fig. 8, Fig. 9a) yielded a higher abundance of tillers and panicles compared to the subsequent season (Fig. 8, Fig. 9b). Specifically, it was documented that the number of tillers and panicles in the first season surpassed that of the second season by 26 %. This observation underscores that the development of tillers and panicles was influenced not only by irrigation practices but also by the distinct climatic conditions experienced during the two growing seasons. The initial season took place during the transition from wet to dry conditions, characterized by higher total solar radiation and evapotranspiration. In contrast, the subsequent season unfolded during the wet phase, marked by reduced solar radiation and evapotranspiration levels. As a result, the first season facilitated increased tiller formation and biomass production, aligning with findings from prior research. Bouman et al. [36] noted a similar trend, highlighting that the dry season tends to yield greater biomass and productivity, coupled with a higher leaf area index compared to the wet season.

Fig. 8.

Tiller numbers among the SSI and CFI regimes: a) first season, b) second season.

Fig. 9.

Panicles number among the SSI and CFI regimes: a) first season, b) second season.

The response of reducing irrigation water, particularly on panicle formation and grain production, is summarized in Table 2. Notably, there was a significant 23 % reduction in the number of panicles under the SSI regime in the first season and an 11 % insignificant reduction in the second season. Furthermore, when assessing the quality of both panicles and grains, the decrease in irrigation water in the SSI system did not yield statistically significant differences when compared to the CFI, particularly in the first season. Although there was a higher count of grains in the CFI, this difference did not reach statistical significance. This aligns with the findings of another study, which underscored how flooded conditions can enhance production due to a greater grain count and a higher proportion of mature grains compared to other regimes [11]. The larger number of panicles and grains in the CFI consequently led to an increased yield, with the CFI system producing 15.5–18.6 % more yield than the SSI. A similar trend was observed in biomass production, with the CFI generating 8.3 % more biomass than the SSI, particularly in the first season. However, in the second season, the biomass production was comparable in both regimes.

Table 2.

Plant response among the regimes.

Plant parameters	Season 1		Season 2		unit
	SSI	CFI	SSI	CFI
Number of panicles per hill	10 ± 2.8a	13 ± 1.2b	8 ± 1.8a	9 ± 1.4b
Panicles weight per hill	6.67 ± 1.4a	7.08 ± 1.2a	8.08 ± 1.9a	8.50 ± 1.7a	g
Main panicle length	34.29 ± 2.2a	35.3 ± 3.2a	33.1 ± 2.8a	33.8 ± 2.2a	cm
Total number of grains	395.7 ± 16.3a	407 ± 36.8a	215 ± 101.7a	247 ± 59.2a
Number of filled grains	245.7 ± 28.2a	298.0 ± 20.7a	87 ± 44.9a	155 ± 48.8b
Number of unfilled grains	150.0 ± 40.0a	109.0 ± 16.5a	128 ± 74.0a	92 ± 37.6a
1000-grains weight	23.0	24.0	24.8	24.9	g
Biomass	23.05	25.13	22.59	22.03	ton/ha
Yield	4.81	5.91	3.97	4.70	ton/ha

3.3. Water use efficiency and productivity in different regimes

In Fig. 10, the water balance components within each irrigation are elucidated. It's evident that precipitation played a dominant role in contributing to the overall water supply in both systems. The study area is recognized for its high annual rainfall intensity, earning it the reputation of a "rainy city" with an annual rainfall of approximately 3000 mm. In terms of irrigation water, the SSI utilized a comparatively modest 193 mm and 223 mm, representing a 37.5 % and 50.5 % reduction in irrigation water compared to the CFI regime. Water consumption by plants, as measured through crop evapotranspiration, was recorded at 232 mm and 214 mm for the SSI and 295 mm and 280 mm for the CFI in the first and second seasons, respectively. This divergence in crop evapotranspiration had a direct impact on the crop coefficients, with average coefficients of 1.17 and 1.23 for the CFI and 0.92 and 0.93 for the SSI, reflecting the differing water use efficiency in the two regimes. Remarkably, both CFI and SSI experienced notable water losses through drainage, measuring at 732 mm and 919 mm for SSI and 788 mm and 1035 mm for the CFI. These substantial drainage losses primarily resulted from ineffective rainfall management.

Fig. 10.

Water allocation and Kc values in the SSI and CFI regimes.

Although the SSI regime witnessed a decrease in production, it demonstrated a noteworthy improvement in water productivity (WP_ET) and water use efficiency (WUE), as indicated in Table 3. This enhanced efficiency stemmed from the reduced irrigation demand in comparison to the CFI regime. The SSI managed to conserve a substantial 37.5–50.5 % of irrigation water, a figure exceeding the percentage decrease in production. Consequently, WP_ET and WUE within the SSI system increased by 3.2–10.4 % and 30.3–70.8 %, respectively. This underscores that the SSI exhibited greater efficiency and achieved substantial water savings compared to the CFI regime.

Table 3.

Water productivity and water use efficiency among the regimes.

No	Parameters	Season 1		Season 2		unit
		SSI	CFI	SSI	CFI
1	Average Crop Coefficient	0.92	1.17	0.93	1.23
2	Water Productivity (WPET)	2.07	2.01	1.86	1.68	kg/cm3
3	Water Productivity (WPI + R)	0.50	0.55	0.38	0.37	kg/cm3
4	Water Use Efficiency (WUE)	2.50	1.92	1.78	1.04	kg/cm3
	Increasing WPET of SSI	3.2 %		10.4 %
	Increasing WPI + R of SSI	−8.8 %		2.7 %
	Increasing WUE of SSI	30.3 %		70.8 %

4. Discussion

In the field, it's consistently observed that the water level in the SSI regime remains lower than that in the CFI regime. This sustained low water level in the SSI created aerobic conditions, probably ensuring a higher oxygen availability. This approach aligns with an innovative irrigation system recognized as an alternative, one that permits plant growth in non-flooded and unsaturated soils [36]. Such a regime is particularly well-suited for regions grappling with limited water resources, such as arid and semi-arid areas, and can be effectively integrated with the System of Rice Intensification (SRI) as a viable alternative for rice cultivation [41].

While the SSI yielded a lower rice production compared to the CFI, the percentage of decrease in yield was notably smaller than the percentage of water savings achieved. The decline in production can be attributed to a decrease in the transpiration rate, which occurs when soil moisture drops below the pF limit of 2.54. According to the Feddes model [42], the transpiration rate reaches its optimal level within a specific range of soil water potential values, falling between h2 (the critical water potential for oxygen stress) and h3 (the critical water potential for drought stress). Ahmad et al. [43], specify the pressure head values for rice as −30 cm for h2 and a range of −100 cm to −200 cm for h3, representing high and low potential transpiration respectively. This implies that the ideal pF for rice fields falls between 1.48 and 2.00 for high transpiration and 1.48 to 2.30 for low transpiration. Consequently, under SSI treatment, there is a likelihood of reduced yield due to diminished plant transpiration when the pressure head exceeds the pF value of 2.30 or even 2.54. This aligns with earlier findings suggesting that maintaining the pressure head around −15 kPa (approximately −150 cm) to −25 kPa (around −250 cm) is crucial for sustaining the transpiration process [44,45].

However, the SSI facilitated aeration at certain times under aerobic conditions that is a viable alternative for water conservation in response to climate change, all while maintaining agricultural productivity. As supported by previous research findings [46], the adoption of aerobic farming techniques can result in water savings of up to 148 %, as demonstrated in a two-year experiment across different locations. It is essential, however, to complement aerobic rice cultivation with effective fertilization management to meet nitrogen requirements. Additionally, precise irrigation control is necessary to maintain optimal soil moisture levels and address potential water deficits, thus regulating plant transpiration effectively.

In the SSI regime, crop evapotranspiration was observed to be lower than in the CFI regime, consistent with previous studies demonstrating that aerobic conditions, including SSI, reduce evapotranspiration rates and the leaf area index [47]. This reduction in water loss contributes to increased water productivity and water use efficiency, as evidenced in Table 3. Such results align with studies conducted in Egypt, revealing that subsurface drainage can enhance water use efficiency by 15–20 % without significantly compromising production [48]. The reduced crop evapotranspiration in the SSI system was instrumental in lowering the crop coefficient (Kc) value in comparison to the CFI system. This decline in Kc value can be directly attributed to the absence of standing water characteristic of the aerobic conditions within SSI, which, in turn, contributed to the reduction in crop evapotranspiration (ETc). These findings are consistent with previous research, which demonstrated that Kc values in aerobic conditions tend to be, on average, 19 % lower than those in flooded conditions [49]. In addition, lower ETc such as in the SSI regime contributed to lower crop productivity as early mentioned by a previous study [50].

Within the framework of utilizing sheet-pipe technology, it proved not only as an effective drainage system but also as a means for subsurface irrigation. The adoption of subsurface irrigation methods, whether through the current sheet-pipe technology or drip systems, has emerged as a promising avenue for bolstering water use efficiency and conserving water resources across various agricultural sectors. Extensive studies have demonstrated the remarkable potential of subsurface drip irrigation, showcasing water use efficiency enhancements of up to threefold compared to traditional methods, alongside substantial water savings ranging from 40 % to 80 % [23,51] Similarly, our observation employing sheet-pipe technology revealed notable water use efficiency improvements, ranging between 30.3 % and 70.8 % (Table 3), with accompanying water savings spanning from 37.5 % to 50.5 % (Fig. 10). Nonetheless, challenges persist, notably in mitigating potential decreases in production associated with subsurface irrigation. While previous research documented production declines ranging from 11 % to 53 % with drip irrigation [23,24], our findings indicate a narrower but still significant decrease ranging from 15.5 % to 18.6 % (Table 2). Addressing this challenge, including its feasibility applied to the farmers, necessitates a multifaceted approach, including meticulous benefit-cost ratio analyses are required.

The feasibility of adopting sheet-pipe technology extends beyond cost considerations, encompassing its ease of application across diverse crop varieties and growing seasons. This technology offers distinct advantages in effectively managing both irrigation and drainage at the field level. Notably, its demonstrated ability to control water flow underscores its practicality for large-scale implementation. Furthermore, the longevity of sheet pipes, which can endure for decades, positions them as a sustainable long-term investment, thereby enhancing agricultural sustainability. This technology closely resembles the FOEAS (the farm-oriented enhancing aquatic system), as employed in Japan. It effectively controlled the desired groundwater for the two soybean cultivars [52]. While the production outcomes achieved with the SSI system still lag those of the conventional method (CFI system), it offers substantial water savings as a valuable trade-off. The sheet-pipe technology holds promising potential for intercropping, such as with soybeans, thereby offering farmers a valuable avenue to boost their income. The practice of intercropping offers distinct economic advantages when compared to monoculture cultivation [53].

Furthermore, the utilization of sheet-pipe technology enhances soil aeration and organic matter content [54], thereby improving the product quality of rice as well as mitigating methane emission [55]. This quality enhancement is attributed to the technology's ability to elevate soil pH, foster aerobic bacteria, and stimulate soil respiration [56]. These conditions pave the way for utilizing sheet-pipe technology as a greenhouse gas mitigation method by curbing methane gas emissions. This emerging challenge in rice production compels us to not only prioritize yield but also focus on water conservation and emission reduction. Hence, the critical factor in harnessing the full potential of sheet-pipe technology is determining the optimal water level for rice production, enabling the maintenance of significant yields while simultaneously reducing emissions and promoting sustainability in rice farming practices by considering regional climate conditions.

5. Conclusions

The present study has underscored the viability of sheet-pipe technology for subsurface irrigation through water level regulation. However, lowering the water level in sheet-pipe subsurface irrigation (SSI) affected the reduction of panicle number and yield compared to conventional flooded irrigation (CFI). Nevertheless, what makes this approach noteworthy was that the percentage reduction in yield was more modest than the percentage of irrigation water saved through the SSI. This remarkable achievement led to a substantial 37.5–50.5 % reduction in water irrigation, significantly elevating water use efficiency (WUE) by 70.8 %, coupled with a commendable 10.4 % boost in water productivity (WP_ET). The imminent challenge for future research lies in optimizing the water level within the sheet-pipe technology to strike a balance that enables the maintenance yields while concurrently reducing emissions and less water input, thereby fostering sustainability in rice farming practices. This endeavor should be approached with due consideration of regional climate conditions, adding an additional layer of complexity to the task.

Funding

This current research received partial support through a grant for fundamental research, bearing contract number "18791/IT3.D10/PT.01.02/M/T/2023," under the title "Integrated Model of Artificial Intelligence-Based Water and Environment Management for Sustainable Rice Farming," provided by the Ministry of Education, Culture, Research, and Technology of the Republic of Indonesia.

Data availability

Data will be made available on request.

CRediT authorship contribution statement

Chusnul Arif: Writing – review & editing, Writing – original draft, Formal analysis, Data curation, Conceptualization. Satyanto Krido Saptomo: Methodology, Investigation. Budi Indra Setiawan: Methodology, Conceptualization. Muh Taufik: Writing – review & editing. Willy Bayuardi Suwarno: Writing – review & editing. Bayu Dwi Apri Nugroho: Investigation. Masaru Mizoguchi: Methodology, Conceptualization.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Chusnul Arif reports financial support was provided by Ministry of Education, Culture, Research and Technology, Republic of Indonesia. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.