More rice with less water – evaluation of yield and resource use efficiency in ground cover rice production system with transplanting

Europ. J. Agronomy 68 (2015) 13–21

Contents lists available at ScienceDirect

European Journal of Agronomy journal homepage: www.elsevier.com/locate/eja

More rice with less water – evaluation of yield and resource use efficiency in ground cover rice production system with transplanting Yueyue Tao a,b , Yanan Zhang a , Xinxin Jin a , Gustavo Saiz c , Ruying Jing a , Lin Guo a , Meiju Liu a , Jianchu Shi a , Qiang Zuo a , Hongbin Tao a , Klaus Butterbach-Bahl c , Klaus Dittert b , Shan Lin a,∗ a

College of Resources and Environmental Sciences, China Agricultural University, Beijing 100193, China Department of Crop Science, Section of Plant Nutrition and Crop Physiology, Faculty of Agriculture, University of Goettingen, 37075 Goettingen, Germany c Institute of Meteorology and Climate Research, Karlsruhe Institute of Technology (KIT-IMK-IFU), 82467 Garmisch-Partenkirchen, Germany b

a r t i c l e

i n f o

Article history: Received 3 January 2015 Received in revised form 8 April 2015 Accepted 22 April 2015 Available online 15 May 2015 Keywords: Water-saving rice Ground cover rice production system Grain yield Water use efficiency Nitrogen use efficiency Shoot ı13 C

a b s t r a c t Adoption of the innovative water-saving ground cover rice production system (GCRPS) based on transplanting of rice seedlings under high soil moisture conditions, resulted in an overall increase in grain yield compared to previous reports on GCRPS employing direct seeding. However, there is a lack of quantitative information on water and nitrogen use efficiency as affected by water and nitrogen management in GCRPS-transplanting. To close this knowledge gap, we conducted a two-year field experiment with traditional paddy rice (Paddy) and GCRPS-transplanting under two soil moisture conditions (GCRPSsat and GCRPS80% ), combined with 3 nitrogen fertilizer management regimes (0, 150 kg urea-N/ha as basal fertilizer for Paddy and GCRPS, 150 kg urea-N/ha in 3 splits for Paddy or 75 kg urea-N/ha plus 75 kg N/ha as chicken manure for GCRPS). Grain yield, water and nitrogen use efficiency, stable isotope 13 C of plant shoots and yield components were evaluated. The study showed: (1) compared to Paddy, both GCRPSsat and GCRPS80% produced significantly more grain yield, while no significant difference in grain yield was found between both GCRPS treatments. (2) Irrigation water use efficiency was increased by 140% in GCRPSsat and >500% in GCRPS80% , while total water use efficiency was improved by 52–96% as compared to Paddy. (3) ı13 C of plant shoots was significantly higher in GCRPS than in Paddy, and showed significant positive correlations with total and irrigation water use efficiencies. (4) Compared to Paddy, agronomic N use efficiency was significantly higher in both forms of GCRPS. However, N recovery rates were only significantly higher in GCRPS than in Paddy when all urea nitrogen was applied as basal fertilizer before transplanting. With improved fertilizer N management, i.e., split N application in Paddy or combined application of urea and chicken manure in GCRPS, there were no significant differences. Overall, this quantitative evaluation of water use efficiency highlights that the use of GCRPS involving transplanting of seedlings has a great potential to reduce irrigation water input, increase grain yield and resource use efficiency. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Rice is the major staple food for more than 3 billion people worldwide. The present and future of global rice production largely depends on irrigated rice production systems in Asia, which provide about 75% of the world’s rice supply (Cantrell and Reeves, 2002; Qin et al., 2006). However, growing larger amounts of rice with less water is one of the major challenges for food security

∗ Corresponding author. Tel.: 86 10 62733636; fax: 86 10 62731016. E-mail address: [email protected] (S. Lin). http://dx.doi.org/10.1016/j.eja.2015.04.002 1161-0301/© 2015 Elsevier B.V. All rights reserved.

faced by humanity in the 21st century. Traditionally, irrigated rice grows under continuous flooding or submerged soil conditions, thus requiring much more water than other cereals. It was recently estimated that rice production consumes more than 45% of total freshwater resources in Asia and approximately 30% of the world’s freshwater used for irrigation (Bouman and Tuong, 2001; Bouman et al., 2007). Nevertheless, the rapidly increasing population and associated water demands for urban and industrial use have put a strain on the fresh water resources worldwide (Bouman, 2007). It has been estimated that about 15–20 million hectares of Asia’s irrigated rice will suffer from water scarcity by the year 2025 (Tuong and Bouman, 2003). In the light of diminishing water resources

14

Y. Tao et al. / Europ. J. Agronomy 68 (2015) 13–21

for agriculture and increasing demand for rice, water-saving rice production techniques have been sought for many years. A promising water-saving technique for lowland rice production is the so called ground cover rice production system (GCRPS) which was firstly introduced in the mountainous region of Central China in 1990s to overcome both, seasonal water shortages and limitations imposed by lower temperatures at the beginning of the rice growing season. Since water and temperature limitations for growing rice can be found in many regions across China, the technique has the potential to be widely disseminated on more than 4 million hectare in northern China (Lin et al., 2002). GCRPS is an agricultural practice that is substantially different from traditional paddy rice systems. In GCRPS the soil surface is covered with a 5–7 ␮m thickness transparent plastic film, and there is no standing water during the entire growth period (Lin et al., 2002). Indeed, two different GCRPS techniques, namely direct seeding or transplanting of rice seedling, have been used on different soil types and different climatic conditions during the past two decades. For the practice of GCRPS-direct seeding, mainly used on sandy-loam soils with a low buffer and water holding capacity, lowland rice is grown across the entire growing season as directly-seeded upland crop at a soil moisture content of 70–90% maximum soil water holding capacity (Tao et al., 2006). Grain yield, agronomic performance, water and nitrogen use efficiency, and greenhouse gas emissions related to the implementation of this technique have been reported in previous research (Lin et al., 2002; Dittert et al., 2002; Sattelmacher et al., 2005; Tao et al., 2006, 2007; Kreye et al., 2007). Compared to traditional paddy rice production, GCRPS-direct seeding significantly reduced irrigation water consumption. However, under these conditions reductions in grain yield were also commonly observed (Lin et al., 2002; Tao et al., 2006). Moreover with GCRPS-direct seeding, a number of additional disadvantages were found. For example, the transition to more aerobic soil conditions in aerobic rice, which had considerable similarities with GCRPS-direct seeding, led to severe weed infestation, especially during the first weeks following seeding (Zhao et al., 2007). Furthermore, the increase in soil redox potential decreased the availability of micronutrients, which might result in potential deficiency of manganese for rice plants (Tao et al., 2007). Finally, continuous aerobic mono-cropping might increase the risk of soil-borne pathogen infestations leading to deterioration of soil health (Nie et al., 2009). Thus, there was much hesitation to adopt GCRPS-direct seeding in many regions of China. Instead, GCRPS-transplanting has recently become a promising technology in water-saving rice production (Qu et al., 2012; Liu et al., 2013, 2014). GCRPS-transplanting can be readily adopted in areas with typical paddy soils, where transplanting is typically done when rice seedlings are approximately one month old. For this approach the soil moisture is kept close to saturation during the entire growing period (Qu et al., 2012). There have been reports about this technique having a remarkable efficiency by increasing soil temperature, preserving soil moisture and inhibiting weed growth under the thin polythene coverage (Shen et al., 1997). It has been showed that, in areas where either water or temperature were restricting factors for rice production, GCRPS-transplanting significantly increased grain yield by 10–18% in both, a long-term field experiment and at the regional scale where GCRPS-transplanting and paddy rice were compared under real farming conditions (Qu et al., 2012; Liu et al., 2013). Given the potential benefits of GCRPStransplanting for many regions across China, there is an urgent need to quantify its effects on water and nitrogen use efficiency as compared to paddy systems. Furthermore, there is no sufficient information on water use efficiency and the yield performance of GCRPS-transplanting or, whether there is potential to further reduce water input after middle tillering stage, the period when rice plants are less vulnerable to reduced soil moisture. On the other hand, stable carbon isotope composition (ı13 C) of plant has been

a useful indicator of water and carbon balance over longer periods (Farquhar and Richards, 1984). Previous research observed strong relationships between water use efficiency and ı13 C on wheat or barley and mostly applied in plant breeding program (Farquhar and Richards, 1984; Farquhar et al., 1982). However, the association between water use efficiency and ı13 C are not well documented in rice under different production systems. Undoubtedly, there are a few aspects to consider when developing concepts for new fertilizer schemes in GCRPS-transplanting. Since the plastic film covers the soil surface, all N fertilizer has to be applied as a basal dressing before transplanting (Shen et al., 1997). Consequently during their juvenile phase, rice plants may show excessive vegetative growth while in the reproductive phase, growth may be limited by insufficient N availability, which may lead to pre-mature plant senescence and constrain grain yield (Qu et al., 2012). Combination of urea and organic manure, with improved N supply after panicle initiation, could enhance the N uptake during the reproductive phase and increase the grain yield as seen with GCRPS in saturated soil in a long-term experiment (Tao et al., 2014). However so far, there are no studies assessing the effect of combining mineral fertilizer such as urea – the most commonly used fertilizer in China – and manure for their effects on yield potential and nitrogen use efficiency in GCRPS-transplanting under different soil moisture conditions. In response to these research needs, a two-year field experiment was conducted to investigate rice grain yield and yield components, water consumption and N uptake, and ı13 C of plant shoots as an indicator for water stress and soil temperature in the upper Han River basin of China for Paddy and GCRPS. The objectives of this work were: (i) to quantify the water and nitrogen use efficiency of GCRPS-transplanting as compared to traditional paddy rice; (ii) to identify whether irrigation water supply might be further reduced in GCRPS-transplanting after middle tillering stage without reduction in crop yield. 2. Materials and methods 2.1. Site descriptions A field experiment was conducted at a farm (32◦ 07 N, 110◦ 43 E, and 440 m ASL) in Fangxian County of Hubei Province, China during the rice growing season (late April–September) in 2012 and 2013. The soil at the experimental site was a silt loam with a texture of 20.3% sand (0.05–2 mm), 60.0% silt (0.002–0.05 mm) and 19.8% clay (<0.002 mm) and in the 0–20 cm depth layer it had 21.3 g organic matter kg−1 and 1.31 g total N kg−1 . The mountainous region of Fangxian County is exposed to northern subtropical monsoon climate, with an annual mean air temperature of 14.2 ◦ C and an annual average rainfall of 830 mm. The total annual sunshine hours are 1850 ± 150 h and the frost-free period lasts 225 ± 15 days. During the growth period, weather data were collected at a meteorological station (WeatherHawk 500, Campbell Scientific, USA) 30 m away from the center of experimental site. Weekly averages of air temperature and rainfall are shown in Fig. 1. 2.2. Experimental design and field management This field experiment was composed of nine experimental treatments (consisting of three rice production systems combined with three N treatments) arranged in a randomized block design, and replicated thrice. The three production systems were: (1) Paddy, the traditional paddy rice production system, in which fields were flood-irrigated to maintain a 3-cm water layer from transplanting until two weeks before harvest. For water level control five graduated poles were fixed at each plot until two weeks before harvest.

Y. Tao et al. / Europ. J. Agronomy 68 (2015) 13–21

(2) GCRPSsat , where soil water content was kept near saturation from the time of transplanting until two weeks before harvest. The raised beds of 1.56 m width remained without a visible water layer, while the furrows between the raised beds and around each plot were filled with water; and (3) GCRPS80% , which was irrigated in the same way as GCRPSsat before middle tillering stage, while after this period, irrigation was applied in the furrows when soil water content fell below 80% of the soil water capacity according to the capacitance probe reading. Three N treatments were examined: (1) N0: zero-N fertilizer; (2) N1: 150 kg urea N ha−1 given as basal fertilizer in both Paddy and GCRPS; and (3) N2: for the Paddy system: 150 urea kg N ha−1 given in split application rates of 90, 30 and 30 kg N ha−1 at transplanting, early tillering and panicle initiation stages, respectively. However in GCRPS, N2 fertilization included 75 kg urea N ha−1 plus 75 kg N ha−1 as chicken manure, all applied as basal application. All treatments received the same amount of phosphorus (45 kg P2 O5 ha−1 as Ca(H2 PO4 )2 ) and potassium (45 kg K2 O ha−1 as KCl). The size of each plot was 90 m2 (9 m × 10 m). In order to minimize horizontal water and nutrient flow between neighboring plots, each plot was confined by small walls made from brick and cement measuring 0.4 m in width and 0.9 m in depth. In addition they were isolated by impermeable film lining. The soil water content of all GCRPS plots was monitored using a capacitance probe (Diviner 2000, Sentek, Australia). The access tube (1 m length and 0.05 m diameter) was placed in the center of a raised bed in each plot. Readings from the capacitance probe were taken every 2 days from the middle tillering stage to harvest, at increments of 0.1 m from soil surface to a depth of 0.6 m. To control the water flow, each plot was instrumented with two manual gate valves both connected with a water meter (GB/T778-96, Haiquan, Ninbo, China) to monitor the amount of water inflow and outflow. The irrigation water was supplied to each individual plot by stable irrigation pipelines from a well near the experimental site. Due to the incorrect installation of impermeable film, water leakage occurred in some plots in 2012. After harvest in 2012, all plots were equipped with impermeable film. Thus, in this paper we only present the water related-data measured in 2013. The hybrid lowland rice cultivar ‘Yixiang 3728’ was used for this experiment. Two 25 days old rice seedlings were transplanted per hill with a constant spacing of 18 × 26 cm on 8 May in 2012 and on 28 April in 2013. Before transplanting, all plot were ploughed, leveled and separated into five raised beds (1.56 m in width and 9.4 m in length) surrounded by 0.15-m-wide and 0.15-m deep furrows. All fertilizers were applied and incorporated into the soil of each raised bed. Then the soil surface of GCPRS plots was covered with 5-␮m thick and 1.7-m wide transparent polyethylene film and then holes were punched. Hand-weeding was used to regularly control weed growth in paddy systems during the entire growing period in both years. There were no herbicide applications as there was no weed growth under the covering plastic sheet in GCRPS. An insecticide (Ethofenprox) was sprayed in mid-August of both years to fight rice planthoppers (Laodelphax striatellus) in all plots. Harvest was done on September 16, 2012 and September 10, 2013. After harvest, the rice straw from all plots was also removed and weighed.

15

At maximum tillering stage, plant shoot samples were collected and analyzed for stable carbon isotope composition (ı13 C) using a combustion elemental analyzer (Costech International S.p.A., Milano, Italy) fitted with a zero-blank auto-sampler coupled via a ConFloIII to a Thermo Finnigan Delta Plus-XL (Thermo Scientific, Waltham, MA, USA). At maturity, grain yield and straw dry yield were determined from a 10-m2 core plot in the center of the raised beds. Before final harvest, 8 hills (0.4 m2 ) were selected randomly in the center raised bed to determine yield components – the number of productive tillers, the number of spikelets per panicle, percentage of filled grains and thousand-grain weight. The number of productive tillers was counted, and panicles were hand-threshed afterwards. Filled spikelets were separated from unfilled spikelets by submerging them in tap water (Peng et al., 2004). The number of spikelets per panicle, thousand-grain weight and percentage of filled grains (100 × filled spikelets number/total spikelets number) were calculated. Harvest index was calculated as the ratio of dry grain yield to total dry biomass at harvest. Grain and straw were first ground in a micro hammer mill (FZ102, Taisite, Tianjin, China), and then by a ball mixer mill (MM200, Retsch, Haan, Germany). The nitrogen concentration of plants was determined by elemental analyzer (EA1108, Fisons Instruments, Milan, Italy).

2.3. Plant sampling and analyze Observation of tiller numbers was done in 5-day intervals. Shoot dry matter was measured at the stages of middle tillering, maximum tillering, panicle initiation, flowering and maturity. On each sampling date, 8 hills (0.4 m2 ) were harvested. Plant samples were washed with distilled water and oven-dried at 70 ◦ C to constant weight.

Fig. 1. Seven-day averages of daily air temperature and daily rainfall during the rice growing season in 2012 (a) and 2013 (b). Weekly amount of irrigation water during the rice growing season in 2013 (c). Arrows indicate the middle tillering stage.

16

Y. Tao et al. / Europ. J. Agronomy 68 (2015) 13–21

Table 1 Cumulative amounts of irrigation water, rainfall and total water consumption in Paddy and GCRPS during the growing season of 2013. Treatment

Irrigation (mm)

Rainfall (mm)

Water consumption (mm)

Paddy GCRPSsat GCRPS80%

753 ± 47 344 ± 28 119 ± 5

638 638 638

1391 ± 47 982 ± 28 757 ± 5

2.4. Calculations and analysis Water use efficiency (WUEt , kg grain m−3 water) was calculated as the grain yield per total water received from irrigation and rainfall. Irrigation water use efficiency (WUEi , kg grain m−3 water) was calculated as the grain yield per amount of irrigation water (Sinclair et al., 1984). Fertilizer-N use efficiency was evaluated by measuring agronomic efficiency (AE, kg grain kg−1 N) and N recovery efficiency (NRE, %), which were calculated according to Novoa and Loomis (1981) as: AE =

3. Results Temperature dynamics and rainfall distribution during the rice growing season showed the same trend in the two reported years (Fig. 1a and b). The average daily temperature were 23.1 ◦ C and 23.5 ◦ C, and total precipitation were 626 and 638 mm during the rice growing season in 2012 and 2013, respectively. Consumption of irrigation water was obviously much lower in GCRPS than in Paddy. Moreover after middle tillering stage, GCRPS80% received less irrigation water than GCRPSsat (Fig. 1c). The average amount of irrigation water was 344 mm in GCRPSsat and 119 mm in GCRPS80% in 2013, which was substantially less than 753 mm measured in Paddy (Table 1). The total water supplies to GCRPSsat and GCRPS80% were 982 mm and 757 mm, which was 70.6% and 54.4% of 1391 mm given in Paddy (Table 1).

GYf − GYuf Napplied

NRE =

Nf − Nuf × 100 Napplied

where GYf and GYuf are grain yield in fertilized and unfertilized plots (kg ha−1 ), respectively, and Napplied is the amount of fertilizerN applied (kg ha−1 ). Nf and Nuf are the amount of N uptake in fertilized and unfertilized plots (kg ha−1 ), respectively. Analyses of variance were performed with the general linear model (GLM) procedure of statistics analysis system (SAS Institute, 2001). The statistical model used included sources of variation due to year, system and nitrogen, and interactions of year × system, year × nitrogen, system × nitrogen, and year × system × nitrogen. Level of significance in figures is given by ns, *, ** for not significant, significant at P < 0.05 and P < 0.01, respectively. Whenever treatment differences were found, the statistical significances were calculated based on the least significant differences (LSD) at the 0.05 probability level.

Fig. 2. Average rice grain yield as affected by rice growing system and nitrogen application including F-statistics on effects of year (Y, 2012 and 2013), growing system (S) and nitrogen application (N) on grain yield. * Significant at 0.05 probability level; ** significant at 0.01 probability level; *** significant at 0.001 probability level; ns: not significant. Bars labeled with the same capital letters show no significant difference (P < 0.05) between systems for each N level; bars labeled with the same lowercase letters show no significant difference (P < 0.05) between N treatments within Paddy and each type of GCRPS. Vertical bars represent standard error of mean.

Fig. 3. Irrigation water use efficiency (a, WUEi ), total water use efficiency (b, WUEt ) and the ı13 C of plant shoot at maximum tillering stage (c, ı13 C) as affected by rice growing system (Y, 2013, n = 3). Bars labeled with the same capital letter show no significant difference (P < 0.05) between systems for each N level. Vertical bars represent standard error of mean.

Y. Tao et al. / Europ. J. Agronomy 68 (2015) 13–21

17

Fig. 4. Relationships between ı13 C values of plant shoot at maximum tillering stage and irrigation water use efficiency (a, WUEi ) and total water use efficiency (b, WUEt ).

3.1. Grain yield, water and nitrogen use efficiency No interaction effects on grain yield among years (Y), production systems (S) and nitrogen (N) were observed (Fig. 2). Based on pooled data of the two experimental years, there were significantly higher grain yields in GCRPS than in Paddy with N fertilization. On the other hand in the absence of N fertilization, GCRPS tended to show slightly higher grain yields than Paddy (Fig. 2). No significant

differences were found in grain yields of GCRPSsat and GCRPS80% (Fig. 2). Nitrogen fertilizer application significantly improved grain yields in both GCRPS and Paddy as compared to N0. Within N1 and N2, there were no significant differences, i.e., in Paddy, there was no difference between urea single dressing and split application and in GCRPS, and there was no difference between single 150 kg urea-N application and the combination of 75 kg urea-N and 75 kg chicken manure-N (Fig. 2). Irrigation water use efficiency (WUEi ) ranged from 1.0 to 1.2 kg grain m−3 water in Paddy and, depending on the mode of GCRPS, significantly increased to the range of 2.3–7.8 kg grain m−3 water (Fig. 3a). Both, GCRPSsat and GCRPS80% significantly improved total water use efficiency (WUEt ), while GCRPS80% showed significantly higher total water use efficiency than GCRPSsat (Fig. 3b). ı13 C, natural stable carbon isotope abundance in plant shoots sampled at maximum tillering was significant higher in GCRPS than in Paddy, with ı13 C values ranging from −27.9‰ to −28.7‰ in GCRPS, and from −28.5‰ to −29.0‰ in Paddy (Fig. 3c). There was no significant difference in shoot ı13 C between GCRPSsat and GCRPS80% (Fig. 3c). Furthermore, ı13 C in plant shoots were significant positively correlated to the irrigation water use efficiency and total water use efficiency (Fig. 4). Agronomic use efficiency of nitrogen fertilizer ranged from 3.7 to 3.9 kg grain kg−1 N in Paddy and from 10.5 to 11.6 kg grain kg−1 N in GCRPS (Fig. 5a). No significant differences in AE were found between GCRPSsat and GCRPS80% (Fig. 5a). When all urea nitrogen was applied as basal fertilizer before transplanting, nitrogen recovery efficiency was significantly increased in GCRPS as compared to Paddy; furthermore, it was significantly higher in GCRPS80% than in GCRPSsat when all urea nitrogen was applied as basal fertilizer (Fig. 5b). No significant differences were found among treatments of improved N schemes, i.e., between Paddy with N split application and GCRPS with combined urea and chicken manure (Fig. 5b). 3.2. Yield related parameters

Fig. 5. Agronomic efficiency (a, AE) and N recovery efficiency (b, NRE) as affected by rice growing system and nitrogen application (Y, 2013, n = 3). Bars labeled with the same capital letters show no significant difference (P < 0.05) between systems for each N level; bars labeled with the same lowercase letters show no significant difference (P < 0.05) between N treatments within Paddy and within each mode of GCRPS. Vertical bars represent standard error of mean.

There were no significant differences in straw yield between Paddy and GCRPS within the two N fertilizer levels of 0 or 150 kg urea-N ha−1 . The comparison between N levels revealed the expected finding that nitrogen fertilization significantly improved straw yields in both GCRPS and Paddy as compared to N0 (Fig. 6a). Within schemes with N fertilizer, there were no significant differences in straw yield, i.e., no difference between one basal urea application and splits of N in Paddy or combined use of manure within each system of GCRPS (Fig. 6a). Major effects of growing system and nitrogen were found for number of productive tillers and spikelets per square meter which were both significant, whereas

18

Y. Tao et al. / Europ. J. Agronomy 68 (2015) 13–21

Table 2 F-statistics to assess the effects of year (Y, 2012 and 2013), system (S) and nitrogen application (N) on straw yield (SY), shoot dry matter (DM), productive tillers (PT), spikelets per square meter (SM), percentage of filled grains (PFG), thousand-grain weight (TGW), and harvest index (HI). Source

DF SY

Year (Y) System (S) Nitrogen (N) Y×S Y×N S×N Y×S×N

1 2 2 2 2 4 4

4.16* 5.11* 77.43*** 2.47ns 2.66ns 2.56ns 3.05*

DM

PT

SM

PFG

TGW

HI

10.09** 11.14*** 94.36*** 1.41ns 1.47ns 1.98ns 3.47*

31.09*** 21.78*** 21.22*** 5.71** 2.21ns 2.35ns 2.66*

0.08ns 10.28*** 18.27*** 2.52ns 0.17ns 0.44ns 1.65ns

17.21*** 0.72ns 2.07ns 3.86* 0.59ns 0.82ns 1.30ns

0.39ns 1.78ns 1.25ns 3.40* 1.46ns 0.34ns 2.54ns

0.27ns 0.33ns 16.87*** 2.61ns 2.58ns 1.86ns 1.00ns

ns: not significant. * Significant at 0.05 probability level. ** Significant at 0.01 probability level. *** Significant at 0.001 probability level.

there was no effect on percentage of filled grains and 1000-grain weight (Table 2). The number of productive tillers and spikelets per square meter were significantly greater in GCRPS than in Paddy; while no significant difference was found between GCRPSsat and GCRPS80% (Fig. 6b and c). Splits of N in paddy rice or combined use of manure in GCRPS led to a similar number of productive tillers and spikelets per square meter as one basal urea application (Fig. 6b and c). Tiller numbers and shoot dry matter of rice plants were significantly higher in GCRPS than in Paddy both, with and without N fertilization (Fig. 7). This was even more pronounced with N-fertilized GCRPS (Fig. 7). GCRPSsat and GCRPS80% showed no significant differences in soil temperature during the growing season, thus data on GCRPS are presented as averages of GCRPSsat and GCRPS80% (Fig. 8). Within the first month after transplanting, the average soil temperature with GCRPS was 4.3 ± 0.3 ◦ C higher during the daytime (06:00–18:00) as compared to Paddy (Fig. 8). 4. Discussion 4.1. GCRPS-transplanting significantly improved resource use efficiency and yield Total water input (irrigation plus precipitation) was 1391 mm on average in traditional paddy, which is a typical and comparable value to the local farmer’s practices and for irrigated rice in Asia (Table 1; Bouman et al., 2007). Compared to traditional paddy, ground cover rice production including transplanting of rice seedlings (GCRPS-transplanting) reduced irrigation water input and increased grain yield significantly (Table 1 and Fig. 2). Therefore, irrigation water use efficiency showed a tremendous and significant increase by 141–573% (Fig. 3a). Moreover, N use efficiency was significantly higher in GCRPS-transplanting than in Paddy (Fig. 5). Irrigation water use efficiency was 2–8 folds higher than that reported with GCRPS-direct seeding in an earlier study in which it was only 71% higher with GCRPS-direct seeding than with traditional paddy (Tao et al., 2006). Compared to the paddy system, the difference between GCRPStransplanting and GCRPS-direct seeding on irrigation water use efficiency is closely linked to: (1) soil physical properties; (2) the grain yield performance; (3) possible nutrient disorders. As GCRPSdirect seeding has been tested at a site with sandy soil with low water holding capacity and without hardpan, there was a much greater need for irrigation in order to maintain soil moisture sufficiently high for rice growth (Lin et al., 2002; Xu et al., 2005; Tao et al., 2006). However in the present study, GCRPS-transplanting was evaluated on a loamy-clay soil with much higher water holding capacity and hardpan, i.e., typical paddy soil (Liu et al., 2013, 2014). Before transplanting, land preparation included puddling,

Fig. 6. Average straw yield (a), number of productive tillers (b) and spikelets per square meter (c) as affected by rice production system and nitrogen application (Y, 2012 and 2013, n = 6). Bars labeled with the same capital letters show no significant difference (P < 0.05) between systems for each N level; bars labeled with the same lowercase letters show no significant difference (P < 0.05) between N treatments within Paddy or within each mode of GCRPS. Vertical bars represent standard error of mean.

which is mainly done for weed control and land leveling as well as to increase water retention and reduce soil water permeability (Qu et al., 2012; Datta, 1981). In the present study, compared to Paddy, 54–84% of the water used for irrigation was saved with GCRPS-transplanting (Table 1), while only 46% was saved with GCRPS-direct seeding (Tao et al., 2006). GCRPS-transplanting significantly increased grain yield compared to Paddy (Fig. 2), which confirms previous studies (Qu et al., 2012; Liu et al., 2013, 2014), whereas GCRPS-direct seeding led to a slightly reduced grain yield (Tao et al., 2006). Analysis of yield components revealed that the number of productive tillers had the greatest effect on the grain yield (Table 2 and Fig. 6b). Compared to Paddy, GCRPS-transplanting showed significantly higher number of tillers and dry matter accumulation and this was particularly relevant in the early growing period and finally lead to a positive effect on the number of productive tillers (Figs. 7 and 6b; Qu et al., 2012). The number of productive tillers was also crucial

Y. Tao et al. / Europ. J. Agronomy 68 (2015) 13–21

19

Fig. 7. Number of tillers (Tillers) and dry matter (DrM) per square meter for rice plants under different systems at each nitrogen application (Y, 2013, n = 3). Vertical bars represent standard error of mean.

in the earlier study of Tao et al. (2006) where, in sharp contrast to the present study with GCRPS-transplanting on loamy soil, GCRPSdirect seeding showed significantly lower numbers of productive tillers compared to Paddy. Moreover, crop growth rate, number of tillers and dry matter accumulation were significantly lower than that in Paddy from transplanting to maximum tillering stage (Tao et al., 2006), demonstrating that under the conditions of the earlier experiment, juvenile plant growth in GCRPS-direct seeding was hampered due to severe water deficit. The shift from traditional paddy to GCRPS goes along with a change from strongly anaerobic to clearly more aerobic soil conditions. This is likely to decrease the bioavailability of micronutrients, which may affect plant growth. Problems with respect to micronutrient nutrition have been suspected to constrain growth in GCRPS-direct seeding (Tao et al., 2007; Zuo and Zhang, 2011). For example, manganese concentration of plant shoot was greater than 300 mg kg−1 dry matter in rice shoots with Paddy whereas it was only 15 mg kg−1 dry matter with GCRPS-direct seedling (Tao et al., 2007). Furthermore, the dominant N form in the soil of GCRPS-

Fig. 8. Average and absolute difference of daily soil temperature observed at 5 cm soil depth for GCRPS and Paddy during daytime (06:00–18:00 h) over the rice growing season in 2013. Vertical bars represent standard error of mean (n, Paddy = 3; n, GCRPS = 6).

direct seeding plots and also in aerobic rice was nitrate, which is known to be highly susceptible to leach for example, during seedling establishment (Tao et al., 2006; Ju et al., 2009). Hence, it is considerable evidence that in GCRPS-direct seeding, early plant development was less vigorous, which resulted in lower grain yield and reduced N recovery efficiency (Tao, 2004; Yin et al., 2004). Likewise, weed growth and increased infestation constitute further constraints leading to reduced performance of GCRPS-direct seeding (Zhao et al., 2007; Peng et al., 2006). In contrast in the present study with GCRPS transplanting and more favorable soil conditions, there were no signs of micronutrient deficiency, soil sickness or yield decline that would resemble problems observed in GCRPS-direct seeding and in aerobic rice (Peng et al., 2006; Tao et al., 2007; Nie et al., 2007, 2009; Qu et al., 2012; Tao et al., 2014).

4.2. GCRPS-transplanting has the potential for substantial water saving without yield reductions The adoption of GCRPS80% resulted in a 180% increase in irrigation water use efficiency as compared to GCRPSsat and there were no detrimental effects on grain yield (Figs. 3a and 2). Moreover, GCRPS80% had significantly higher N recovery efficiency than GCRPSsat when all N fertilizer was used as basal fertilization which was common practices (Fig. 5b), which is a novel finding not reported in previous studies. The remarkable increase in irrigation water use efficiency with GCRPS80% was mainly caused by the significant reduction in irrigation water input being only 35% of the need of GCRPSsat (Table 1). Irrespective of this large water saving, yield components such as number of productive tillers, number of spikelets per panicle, percentage of filled grains and thousand-grain weight were comparable in GCRPS80% and GCRPSsat (Table 2 and Fig. 6). This is an outstanding finding as so far, most reports on water-saving rice production techniques, e.g., saturated soil culture or alternate wetting and drying showed that these saving were at the expense of yield reductions (Tao et al., 2006; Bouman and Tuong, 2001). So the present study shows that, despite rice being so sensitive to water supply, flooding is not essentially required for its prosperous growth. However, a key point for successful water-savings seems to be careful consideration of sensitive growth stages, i.e., anthe-

20

Y. Tao et al. / Europ. J. Agronomy 68 (2015) 13–21

sis, when soil water contents must be kept very close to saturation (Humphreys et al., 2005; Ekanayake et al., 1989). The encouraging findings in both WUE and grain yields reported in this study, are highly dependent on attentive water management. In GCRPS80% soil water content was kept near saturation before middle tillering stage, which was particularly important for proper seedling recovery after transplanting. Thus there were similar tiller numbers and dry matter accumulation in the early growth phase in GCRPS80% and GCRPSsat (Fig. 7). Both are important factors determining crop yields (Yang and Zhang, 2010). Furthermore, the number of spikelets per panicle and the percentage of filled grains remained unaffected in GCRPS80% (Table 2), however, both yield components were very sensitive to low water availability during the reproductive period (Ekanayake et al., 1989). It is therefore relevant to consider the seasonal rainfall pattern in this region, characterized by quite even and substantial precipitation in the period of June to August (Liu et al., 2013; Fig. 1a and b), which may have alleviated potential negative effects of water stress during the critical growth stage in GCRPS80% . Therefore, based on the combination of the above circumstances it seems likely that rice plants in GCRPS80% did not experience any critical water deficit after middle tillering stage. While not being in the center of this study, stable isotope ı13 C of plant shoots provide additional valuable information on the effects of water management. ı13 C of plant tissues are highly dependent on environmental conditions, and they reflect the extent to which CO2 assimilation is limited by carboxylation and/or CO2 diffusion in leaves (Farquhar and Richards, 1984; Ehleringer, 1993). More positive ı13 C values are typically observed in plants with a higher WUE which is related to a greater depletion of intra-tissue CO2 concentrations and therefore lower ı13 C discrimination (Farquhar et al., 1982). ı13 C of plant shoots collected at maximum tillering stage were clearly higher in GCRPS than in Paddy (Figs. 3c and 4) indicating greater WUE. Our also show that there were no significant differences in ı13 C between both forms of GCRPS tested here. Our results showed that GCRPS80% attained significantly higher N recovery efficiency than GCRPSsat when both were supplied with 150 kg urea-N ha−1 (Fig. 5b). We hypothesize that reduced water input promoted the utilization of nitrogen in deeper soil horizons (Lahiri, 1980), which was particularly evident when urea nitrogen was applied as a single basal dressing. Compared to basal application, split application of urea or the combined use of mineral fertilizer and manure showed similar effects on crop growth performance and agronomic efficiency (Figs. 2, 6 and 5a). This stands in contrast to previous reports (Zhang et al., 2013; Tao et al., 2014) which is probably to be attributed to differences in soil properties and crop management at the different experimental sites, as well as to timing and rate of N topdressings, which are crucial factors influencing crop growth and grain yield (Sui et al., 2013).

5. Conclusion The present study demonstrates for the first time that, GCRPStransplanting combined with saturated soil conditions (GCRPSsat ) significantly increased total and irrigation water use efficiency over traditional Paddy by 52% and 140%, respectively. GCRPStransplanting combined with lower soil water content after the middle tillering stage (GCRPS80% ) further reduced irrigation water input while sustaining grain yields on a comparable level. This resulted in an additional increase in irrigation water use efficiency of 180% over GCRPSsat . This result indicates that in regions with favorable lowland rice soils, GCRPS combined with transplanting is a promising technique with a high potential to increase sustainability of rice production and its resource use efficiency. In suitable regions with favorable rainfall patterns in the period of

June–August, GCRPS with reduced soil water potential during the late growth seems to be an additional option for maximizing irrigation water use efficiency. Acknowledgements This work was supported by the National Natural Science Foundation of China (NSFC 51139006 and 41371286). We would like to thank the Agriculture Department of Fangxian Country and Shiyan City for providing working facilities during this cooperation. We are also grateful to Yao Zhisheng, Bai Zhifeng, Du Yanbing from Beijing; Luo Jianhua, Wang Huaqi, Cheng Daode from Fanxian; Li Jiajun, Liu Jun, Wei Jianjun, Wei Guangjun from Shiyan for their help on the field experiment. References Bouman, B., Tuong, T., 2001. Field water management to save water and increase its productivity in irrigated lowland rice. Agric. Water Manage. 49, 11–30. Bouman, B., Humphreys, E., Tuong, T., Barker, R., 2007. Rice and water. Adv. Agron. 92, 187–237. Bouman, B., 2007. A conceptual framework for the improvement of crop water productivity at different spatial scales. Agric. Syst. 93, 43–60. Cantrell, R., Reeves, T., 2002. The cereal of the world’s poor takes center stage. Science 296, 53. Datta, D., 1981. Chapter 8 Land preparation for rice soils. In: Surajit, K., Datta, D. ˜ (Eds.), Principles and Practices of Rice Production. IRRI, Los Banos, Philippines, pp. 259–261. Dittert, K., Shan, L., Kreye, C., Zheng, X., Xu, Y., Lu, X., Shen, Q., Fan, X., Sattelmacher, B., 2002. Saving water with ground cover rice production systems (GCRPS) at the price of increased greenhouse gas emissions. In: Bouman, B., Hengsdijk, H., Hardy, B., Bindraban, P., Tuong, T., Ladha, J. (Eds.), Water-Wise Rice Production, Proceedings of the International Workshop on Water-Wise Rice Production. ˜ ˜ Los Banos, Philippines. 8–11 April IRRI, Los Banos, Philippines, pp. 197–206. Ehleringer, J., 1993. Carbon and water relations in desert plants: an isotopic perspective. In: Ehleringer, J., Hall, A., Farquhar, G. (Eds.), Stable Isotopes and Plant Carbon–Water Relations. Academic Press, San Diego, pp. 155–172. Ekanayake, I., De Datta, S., Steponkus, P., 1989. Spikelet sterility and flowering response of rice to water stress at anthesis. Ann. Bot. 63, 257–264. Farquhar, G., Richards, R., 1984. Isotopic composition of plant carbon correlates with water-use efficiency of wheat genotypes. Aust. J. Plant Physiol. 11, 539–552. Farquhar, G., O’Leary, M., Berry, J., 1982. On the relationship between carbon isotope discrimination and the intercellular carbon dioxide concentration in leaves. Aust. J. Plant Physiol. 9, 121–137. Humphreys, E., Meisner, C., Gupta, R., Timsina, J., Beecher, H., Lu, T.Y., Gill, M., Masih, I., Guo, Z.J., Thompson, J., 2005. Water saving in rice-wheat systems. Plant Prod. Sci. 8, 242–258. Ju, X., Xing, G., Chen, X., Zhang, S., Zhang, L., Liu, X., Cui, Z., Yin, B., Christie, P., Zhu, Z., 2009. Reducing environmental risk by improving N management in intensive Chinese agricultural systems. Proc. Natl. Acad. Sci. U. S. A. 106, 3041–3046. Kreye, C., Dittert, K., Zheng, X., Zhang, X., Lin, S., Tao, H., Sattelmacher, B., 2007. Fluxes of methane and nitrous oxide in water-saving rice production in north China. Nutr. Cycling Agroecosyst. 77, 293–304. Lahiri, A., 1980. Interaction of water stress and mineral nutrition on growth and yield. In: Turner, N., Kramer, P. (Eds.), Adaptation of Plants to Water and High Temperature Stress. John Wiley and Sons, New York, pp. 341–352. Lin, S., Dittert, K., Tao, H., Kreye, C., Xu, Y., Shen, Q., Fan, X., Sattelmacher, B., 2002. The ground-cover rice production system (GCRPS): a successful new approach to save water and increase nitrogen fertilizer efficiency? In: Bouman, B.A., Hengsdijk, H., Hardy, B., Bindraban, P., Tuong, T., Ladha, J. (Eds.), Water-Wise Rice Production, Proceedings of the International Workshop on Water-Wise ˜ ˜ Rice Production. Los Banos, Philippines. 8–11 April IRRI, Los Banos, Philippines, pp. 187–196. Liu, M., Lin, S., Dannenmann, M., Tao, Y., Saiz, G., Zuo, Q., Sippel, S., Wei, J., Cao, J., Cai, X., 2013. Do water-saving ground cover rice production systems increase grain yields at regional scales? Field Crops Res. 150, 19–28. Liu, M., Liang, W., Qu, H., Zhi, G., Chen, Q., Gong, Y., Butterbach-Bahl, K., Lin, S., 2014. Ground cover rice production systems are more adaptable in cold regions with high content of soil organic matter. Field Crops Res. 164, 74–81. Nie, L., Peng, S., Bouman, B.A., Huang, J., Cui, K., Visperas, R.M., Park, H., 2007. Alleviating soil sickness caused by aerobic monocropping: responses of aerobic rice to soil oven-heating. Plant Soil 300, 185–195. Nie, L., Xiang, J., Peng, S., Bouman, B., Huang, J., Cui, K., Visperas, R., 2009. Alleviating soil sickness caused by aerobic monocropping: response of aerobic rice to fallow, flooding and crop rotation. J. Food Agric. Environ. 7, 723–727. Novoa, R., Loomis, R., 1981. Nitrogen and plant production. Plant Soil 58, 177–204. Peng, S., Huang, J., Sheehy, J., Laza, R., Visperas, R., Zhong, X., Centeno, G., Khush, G., Cassman, K., 2004. Rice yields decline with higher night temperature from global warming. Proc. Natl. Acad. Sci. U. S. A. 101, 9971–9975.

Y. Tao et al. / Europ. J. Agronomy 68 (2015) 13–21 Peng, S., Bouman, B., Visperas, R., Castaneda, A., Nie, L., Park, H., 2006. Comparison between aerobic and flooded rice in the tropics: agronomic performance in an eight-season experiment. Field Crops Res. 96, 252–259. Qin, J., Hu, F., Zhang, B., Wei, Z., Li, H., 2006. Role of straw mulching in non-continuously flooded rice cultivation. Agric. Water Manage. 83, 252–260. Qu, H., Tao, H., Tao, Y., Liu, M., Shen, K., Lin, S., 2012. Ground cover rice production system increases yield and nitrogen recovery efficiency. Agron. J. 104, 1399–1407. SAS Institute, 2001. Sas User’s Guide: Statistics, Version 8.2, 6th ed. SAS Inst., Cary NC. Sattelmacher, B., Dittert, K., Lin, S., 2005. Paddy rice and the water supply. German Res. 27, 15–17. Shen, K., Wang, X., Luo, X., 1997. Test and demonstration on wet-cultivation with film mulching of rice. Hubei Agric. Sci. 5, 18–22 (In Chinese with English abstract). Sinclair, T., Tanner, C., Bennett, J., 1984. Water use efficiency in crop production. BioScience 34, 36–40. Sui, B., Feng, X., Tian, G., Hu, X., Shen, Q., Guo, S., 2013. Optimizing nitrogen supply increases rice yield and nitrogen use efficiency by regulating yield formation factors. Field Crops Res. 150, 99–107. Tao, H., Brueck, H., Dittert, K., Kreye, C., Lin, S., Sattelmacher, B., 2006. Growth and yield formation of rice (Oryza sativa L.) in the water-saving ground cover rice production system (GCRPS). Field Crops Res. 95, 1–12. Tao, H., Dittert, K., Zhang, L., Lin, S., Römheld, V., Sattelmacher, B., 2007. Effects of soil water content on growth, tillering, and manganese uptake of lowland rice grown in the water saving ground cover rice production system (GCRPS). J. Plant Nutr. Soil Sci. 170, 7–13.

21

Tao, Y., Qu, H., Li, Q., Gu, X., Zhang, Y., Liu, M., Guo, L., Liu, J., Wei, J., Wei, G., Shen, K., Dittert, K., Lin, S., 2014. Potential to improve N uptake and grain yield in water saving ground cover rice production system. Field Crops Res. 168, 101–108. Tao, H., 2004. Yield and nitrogen uptake of lowland rice (Oryza sativa L.) in a water saving ground cover rice production system (GCRPS) in Beijing, North China. In: Thesis. Christian-Albrecht Univ., Kiel, Germany, pp. 256. Tuong, T., Bouman, B., 2003. Rice production in water scarce environments. In: Kijne, J., Barker, R., Molden, D. (Eds.), Water Productivity in Agriculture: Limits and Opportunities for Improvement. CABI Publishing, Wallingford, UK, pp. 53–67. Xu, Y., Zhang, Y., Shen, Q., Xu, Y., Zhang, J., 2005. An innovative method for the treatment of rice straw to improve nitrogen uptake efficiency. Biol. Fertil. Soils 41, 291–294. Yang, J., Zhang, J., 2010. Crop management techniques to enhance harvest index in rice. J. Exp. Bot. 61, 3177–3189. Yin, X., Xu, Y., Shen, Q., Zhou, C., Huang, X., Li, M., Yin, J., Dittert, K., 2004. Nitrogen uptake and use efficiency by rice crops cultivated in waterlogged field and sowed on dry field without or with different mulchings. Acta Pedol. Sin. 41, 983–986 (In Chinese with English abstract). Zhang, Z., Chu, G., Liu, L., Wang, Z., Wang, X., Zhang, H., Yang, J., Zhang, J., 2013. Mid-season nitrogen application strategies for rice varieties differing in panicle size. Field Crops Res. 150, 9–18. Zhao, D., Bastiaans, L., Altin, G., Spiertz, J., 2007. Interaction of genotype and management on vegetative growth and weed suppression of aerobic rice. Field Crops Res. 100, 327–340. Zuo, Y., Zhang, F., 2011. Soil and crop management strategies to prevent iron deficiency in crops. Plant Soil 339, 83–95.