Do water-saving ground cover rice production systems increase grain yields at regional scales?

Field Crops Research 150 (2013) 19–28

Contents lists available at SciVerse ScienceDirect

Field Crops Research journal homepage: www.elsevier.com/locate/fcr

Do water-saving ground cover rice production systems increase grain yields at regional scales? Meiju Liu a,b , Shan Lin a,∗ , Michael Dannenmann b , Yueyue Tao a , Gustavo Saiz b , Qiang Zuo a , Sebastian Sippel b , Jianjun Wei c , Jun Cao c , Xianzhong Cai c , Klaus Butterbach-Bahl b a

College of Resource and Environmental Science, China Agricultural University, Beijing 100193, China Institute of Meteorology and Climate Research, Atmospheric Environmental Research (IMK-IFU), Karlsruhe Institute of Technology, Garmisch-Partenkirchen 82467, Germany c Shiyan Municipal Bureau of Agriculture, Hubei Province 442000, China b

a r t i c l e

i n f o

Article history: Received 4 April 2013 Received in revised form 5 June 2013 Accepted 6 June 2013 Keywords: Water-saving rice Regional evaluation Ground cover rice production system Yield and yield component Soil redox potential Leaf ␦13 C

a b s t r a c t “Ground cover rice production system” (GCRPS) is an innovative production technique that uses significantly less water than traditional paddy cultivation (Paddy). Consequently, this system may allow for expansion of rice crops to regions with limited water availability. Earlier studies have reported contradictory grain yields and yield performances of GCRPS versus Paddy systems in experimental plots. However, the actual effects of using GCRPS on yields under real farming practices on heterogeneous environments are still unknown. In this study, we compared grain yields and yield components between GCRPS and Paddy systems by sampling paired adjacent farmer fields at 36 representative sites in the region of Shiyan, central China, which is typical for many mountainous areas across China. Furthermore, we characterized soil physico-chemical properties, soil redox potential, stable carbon isotopic composition of plant leaves, and monitored soil temperature during the growing season. Our study revealed the following findings: (1) Across all sites GCRPS significantly increased grain yield by on average 18%. Statistical analysis allowed us to classify three different groups of yield performance within the 36 paired sites: (a) group of significant increase (SI; n = 22) with increases in yields on average 32%, (b) group showing no significant increase (NI; n = 9), here yields increased on average 6%, and (c) sites with grain yields showing a small (−8%), but non significant decrease (ND; n = 5). (2) Shoot dry biomass, number of productive tillers, spikelets per square meter and percentage of filled grains were significantly larger for GCRPS as compared to Paddy systems. (3) No significant differences in soil physical and chemical properties were found for the 0–20 cm layer between GCRPS and Paddy systems. (4) Significantly higher soil temperatures observed in GCRPS during the first month after transplanting were only found in the SI sites, which showed that higher temperature during this critical period was the decisive factor for GCRPS-induced yield enhancement. (5) The average ␦13 C of plant leaves and soil redox potential were significantly higher in GCRPS than Paddy for the SI group only. In-detail analyses of the 5 pairs showing decreases in yields (ND) between GCRPS and Paddy systems revealed the lack of significant effects observed in some key parameters such as soil temperatures during the first month, ␦13 C of plant leaves and soil redox potential. These facts strongly suggested that unnecessary excess water was used, thus hampering GCRPS-induced increases in soil temperature and grain yields, and unequivocally signaling that appropriate water management by farmers is crucial for the successful implementation of GCRPS. Our study demonstrates the large potential of GCRPS to increase grain yields in regions where rice growth is both limited by low temperatures and water scarcity. © 2013 Elsevier B.V. All rights reserved.

1. Introduction

∗ Corresponding author. Tel.: +86 10 62733636; fax: +86 10 62731016. E-mail address: [email protected] (S. Lin). 0378-4290/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.fcr.2013.06.005

Rice is arguably the most important staple food for almost half the world’s population, representing about 20% of its overall energy intake (IRRI, 2007). Recent estimates calculate that 79 million hectares of irrigated lowland rice fields produce about 75% of the world’s rice (Maclean et al., 2002; Qin et al., 2006), which

20

M. Liu et al. / Field Crops Research 150 (2013) 19–28

represents 24–30% of the world’s freshwater resources used for irrigation (Bouman et al., 2006). It is estimated that an annual increase in rice production by 8–10 million t is needed over the next 20 years in order to meet forecasted needs (IRRI, 2011). Furthermore, it is anticipated that 15–20 million ha of irrigated rice will suffer from water scarcity by 2025 due to increasing population growth and associated water demands for urban and industrial use (Tuong and Bouman, 2003; Belder et al., 2005; Bouman, 2007). China is the world’s largest rice producer with 197 million t of paddy rice, a figure that accounts for 35% of global rice production. To meet current national demands, an area of 29 million ha is currently under rice cultivation, consuming about 70% of its total agricultural water resources (FAOSTAT, 2011). Increasing water scarcity has the potential to further exacerbate conflicts on water resources over the coming decades, especially in the middle and northern part of China (Abdulai et al., 2005; Yong, 2009). At current levels of water usage, the total water shortage is estimated to be 30–40 billion m3 year−1 and may be even larger in dry years (MWR, 2007). By 2050, China’s total water deficit could reach 400 billion m3 , which roughly represents 80% of the current annual capacity (Tso, 2004). Moreover, both the poor quantity and quality of water resources threatens not only economic development and quality of life, but it is also exerting a negative impact on food security (Yong, 2009). In order to meet unprecedented challenges in rice production derived from severe deficits in water resources, a number of watersaving technologies have been proposed worldwide. One of these technologies is the ground cover rice production system (GCRPS, Shen et al., 1997; Lin et al., 2002). The GCRPS was developed in 1990 by the Agricultural Bureau of Shiyan (Central China) and it has been constantly expanding in the Shiyan region ever since (Shen et al., 1997). In this cultivation system, soil surface is covered with a 5–7 ␮m thick plastic film, traditional lowland rice cultivars are used and grown at soil water saturation with no standing water layer during the entire growth period (Qu et al., 2012). The proposed advantages provided this technique are numerous and include the following: (i) The adoption of GCRPS in water deficit and cool mountainous regions could preserve heat and effectively alleviate low-temperature stress on early growth stage after transplantation (Shen et al., 1997; Qu et al., 2012); (ii) GCRPS can save water through improved water use efficiency by means of reduced evaporation and seepage during the growing season. Previous research has reported that GCRPS water use efficiency yields up to 0.8–1.0 kg grain m−3 water (Tao et al., 2006), whereas in conventional Paddy systems water use efficiency is on average 0.4 kg grain m−3 (Tuong et al., 2005); (iii) This technique could minimize environmental pollution due to herbicide application because plastic film largely prevents weed germination and development (Peng et al., 1999; Wu et al., 1999). While there are obvious advantages associated with the adoption of GCRPS, contradictory yield performances using GCRPS have been reported under different experimental settings (Shen et al., 1997; Tao et al., 2006; Yang and Zhang, 2010; Qu et al., 2012). GCRPS has shown similar or even reduced grain yields compared with the traditional flooded paddy system in areas where water and temperature were not limiting factors for crop growth (Liang et al., 1999; Wu et al., 1999) and also in a study conducted on markedly sandy soils (Tao et al., 2006). On the other hand, and compared to Paddy systems, grain yield increases have been observed using GCRPS in areas where seasonal water shortage and low-temperature during early growth stages were the main restricting factors (Shen et al., 1997; Jin et al., 2002; Liu et al., 2009; Qu et al., 2012). However, a better and more comprehensive understanding of the effects of GCRPS on yields and yield components is currently hampered by the limited experimental design of earlier work. These studies report findings from a small number of experimental fields that do

not account for the impacts of GCRPS on yields under actual farming practices at regional scales. Therefore, the goals of this study were: (i) to evaluate the grain yield and yield components of GCRPS at the regional scale in Central China; (ii) to identify environmental and management factors determining the success of GCRPS application at regional scale under real farming practices. 2. Materials and methods 2.1. Sampling region characteristics The study was situated in region of Shiyan, Hubei province, central China (32◦ 02 to 33◦ 10 N, 109◦ 44 to 111◦ 04 E, see Table A1), where GCRPS was introduced at the end of the last century (Shen et al., 1997; Liang et al., 1999). Shiyan is located in the QinBaShan Mountains with peaks reaching a maximum altitude of 2740 m a.s.l. According to Smit and Cai (1996) this area is in the northern subtropical agro-climatic zone of China’s eastern monsoon region. In most of these mountainous regions rice growth is limited by both low temperatures at the start of the growing season, and severe seasonal and regional water scarcity (Shen et al., 1997). Specifically, spring drought periods can cause severe reductions in forecasted yields and may affect – as it was the case in the year 2011 – the date of rice transplanting and overall crop growth. The annual average temperature of the study region is 15.3 ◦ C and total average annual rainfall is 829 mm (average calculated for the 1961–2009 period from seven meteorological stations located in the respective counties of Shiyan (Zhou et al., 2008). Annual rainfall shows a pronounced seasonality, with approximately 45% of the rainfall occurring during the summer period (June to August). The total sunshine hours per year are 1835. Given that GCRPS has only been introduced two decades ago and this growing technique has implications for farming activities, labor demand and costs, GCRPS and traditional lowland rice cultivation are often spatially interwoven, i.e. some farmers have adopted the technique while others have not (Zhou et al., 2008). However, in most cases the adoption of GCRPS by individual farmers is well documented by the local administration so that for the selected sites and fields it was possible to trace specific land management records. 2.2. Site and field selection Site selection was performed by experienced staff members from the local Agricultural Bureau in Shiyan, with specific attention being paid to cover different rice growing areas at varying altitudes, on contrasting soil types and over a range of time spans since adoption of the GCRPS technique. The latter information, as well as data on fertilization regimes and soil and crop management, was obtained through interviews to farmers. A total of 49 sites with paired treatments consisting of GCRPS versus permanent flooding paddy fields (hereafter referred to as GCRPS and Paddy) were selected for soil-plant sampling and for in situ field measurements. The distance between the paired plots were in most cases less than 100 m with only 9 out of 49 sites being more than 250 m apart (Table A1). Geographical coordinates of the sites and fields were recorded by GPS (Garmin Colorado 300) and altitudes were obtained using the Global Digital Elevation Model (GDEM) provided by NASA and METI (2008). 2.3. Production management of Paddy and GCRPS Traditional Paddy systems imply that rice seeds are sown and raised in external nurseries for one month before being transplanted to the fields. Fields get ploughed puddled and leveled while flooded. Based on the survey conducted to individual farmers,

M. Liu et al. / Field Crops Research 150 (2013) 19–28

approximately an average of 100 kg N ha−1 was applied as compound NPK fertilizer to surface and incorporated to a soil depth of 20 cm before transplanting. At both tillering and grain filling stages, additional doses of 40 kg N ha−1 were given as urea in order to increase rice milling quality and protein content (Wopereis-Pura et al., 2002; Leesawatwong et al., 2005) and yield, all of which resulted in a total N application of approximately 180 kg N ha−1 . Under this traditional rice paddy cultivation, the field is kept flooded with about 3–5 cm water depth until two weeks before harvest. The average plant density at harvest was 23 ± 1 hills m−2 (n = 108) with 4–5 rice seedlings being transplanted at each hill. Management of water and nutrients significantly differed for GCRPS. One week before transplanting GCRPS fields was flooded for field preparation (puddling and leveling). The day before transplanting, compound fertilizer containing about 150 kg N ha−1 was applied to the soil surface in a single dose and incorporated into the soil by plowing, which was followed by leveling. Since the plastic film covers the soil surface, topdressing is not used in GCRPS, i.e. farmers usually broadcast all the fertilizer before transplanting. Following field preparation the GCRPS field was covered with transparent or black plastic polyethylene film of 5–7 ␮m thickness, which was perforated in regular patterns to allow transplanting of 1-month-old rice seedlings. The soil remained water-saturated without standing water layer on the soil surface of planting area and kept the water in the ditch during the first week after transplantation. After this initial stage, the soil was usually kept at 80–90% of its maximum water holding capacity for the remaining growing period. The average plant density of GCRPS at harvest was 21 ± 1 hills m−2 (n = 108) while 2–3 rice seedlings were transplanted at each hill.

2.4. Soil sampling methodology and analytical procedure Soil sampling of all 49 pairs was undertaken with the aid of a soil auger (3.5 cm diameter) at the 0–20 cm depth interval before any field preparation during March to April 2011. Nine individual sampling locations were performed at each site and each three individual samples were mixed as one replicate. Soil bulk density analysis was determined according to the method described by Blake and Hartge (1986) taking three spatial replicates at the middle of 0–20 cm depth from one soil profile dug by hand in each field. These samples were subsequently dried at 105 ◦ C for 24 h and weighed. Samples from 0 to 20 cm depth interval were used for analysis of physical and chemical soil properties. Soil texture was determined using the pipette method (Gee and Bauder, 1986). Soil pH was measured in 1:2.5 soil–water solution using a combined electrode pH meter (HI 98121, Hanna Instruments, Kehl am Rhein, Germany). Extractable soil NO3 − -N and NH4 + -N was estimated from 1:10 soil–CaCl2 (0.01 M) extracts using an autoanalyser (AA3, Bran & Luebbe, Nordstadt Germany). Sub-samples were powdered in a ball mill (MM200, Retsch, Haan Germany). Removal of inorganic carbon in some samples (indicated by pH values > 7) is a prerequisite to obtain correct soil organic carbon content. To avoid possible losses of soil organic carbon through direct acid washing (Harris et al., 2001), an acid fumigation method was applied for SOC analyses as originally described by Harris et al. (2001) and modified by Walthert et al. (2010). Elemental carbon and nitrogen were determined in duplicate using a Costech Elemental Analyzer (Costech International S.p.A., Milano, Italy) fitted with a zero-blank autosampler coupled via a ConFloIII to a Thermo Finnigan Delta Plus-XL (Thermo Scientific, Waltham, MA, USA) using continuous-flow isotope ratio mass spectrometry. Soil temperature at 5 and 15 cm depth was recorded at hourly intervals during the whole vegetation period at all plots by the installation of calibrated temperature sensors with integrated logger (EBI-20T, Ebro Instruments, Germany).

21

Soil temperature dynamics at 15 cm depth were similar to those at 5 cm depth between adjacent Paddy and GCRPS fields with greater fluctuations observed in 5 than in 15 cm depth (data not shown). 2.5. Plant sampling methodology and analytical procedure Due to the severe drought in Spring 2011 and associated shortage of irrigation water during the rice transplanting period, not all farmers were finally able to use the pre-selected sites for rice cultivation. Therefore, we could not obtain biomass data for 13 out of 49 sites, resulting in a final replication of 36 paired sites at a regional scale. Aboveground plant biomass, grain yield and yield components (productive tiller number, spikelets per panicle, percentage of filled grain, thousand-grain weight, harvest index) were measured for all 36 sites and paired fields at maturity in mid September 2011. At each field, three replicated sampling spots of 3 m2 were randomly chosen and harvested for determination of grain and straw dry yield. Before the final harvest, plant material from eight hills was collected at every sampling spot for the determination of yield components. Plant space and row distance were recorded for the calculation of plant density. Panicle numbers were counted. Filled and unfilled spikelets were separated by submerging them in tap water (Peng et al., 2004). Thousand-grain weight, spikelets per panicle, percentage of filled grain (100 × filled spikelet number/total spikelet number), and the harvest index (100 × grain yield/shoot dry matter) were also determined. At maximum tillering stage, tiller numbers were counted and the latest expanded leaves were sampled for ␦13 C determination, leaf SPAD (Soil Plant Analysis Development) values were measured using a chlorophyll meter (SPAD-502, Konika Minolta Sensing Inc., Japan) in 24 replicates for each plot. Soil redox potential was also measured in situ by means of a redox potential meter (ORP Testr 10, USA) in 12 replicates for each plot. Preparation and isotopic analyses of leaves were performed as described for soil samples. 2.6. Statistical analyses Shapiro–Wilk tests were applied to check for normal distribution. Non-parametric tests were applied if the data was not normally distributed. First, t-test was used to test for significant differences between GCRPS and Paddy at the regional scale. Then parametric and non-parametric tests were applied to each site/single pair of fields. According to individual statistical analyses of yields observed at each paired fields, three groups of data were identified and separated, i.e. (1) group of significant yield increase under GCRPS cultivation (SI), (2) group of non significant increase (NI) and (3) group of non significant decrease (ND). Correlation analyses were used to examine relationships between yield and yield components using Pearson’s correlation coefficients. All statistical analyses and calculations were performed using parametric (paired and two-side t-test) and non-parametric (Wilcoxon matched pairs rank sum test; two-sided) tests and GLM procedure of Statistical Analysis System (SAS, version 8.2). 3. Results 3.1. Grain yield and yield components at the regional scale We sampled 36 paired sites composed of 2 adjacent fields (1 GCRPS and 1 Paddy field) at a large regional scale. For each field 3 replicated plots were sampled. There were neither significant differences in soil texture, bulk density nor in soil chemical properties between GCRPS and Paddy at the regional scale (Table 1). The grain and straw yield of all 36 pairs ranged from 4.27 to 9.97 t ha−1 and 4.11 to 9.89 t ha−1 for GCRPS, 3.63 to 8.77 t ha−1

22

M. Liu et al. / Field Crops Research 150 (2013) 19–28

12

10 Grain Straw

Grain yield (t ha )

-1

-1

10

Yield (t ha )

GCRPS Paddy

***

***

8 6 4

8 6 4 2 0

2

10-15

GCRPS

Paddy

GCRPS

15-30

>30

Soil clay content (%)

Paddy

Fig. 1. Grain and straw yields under the ground cover rice production system (GCRPS) and the Paddy system (n = 108). *** Significant at P < 0.001.

and 3.07 to 10.55 t ha−1 for Paddy, respectively (Fig. 1). The average grain and straw yield for 108 replicated GCRPS sampling plots were 7.15 t ha−1 and 6.75 t ha−1 respectively, which represent 18% and 14% significantly higher yields than those observed in Paddy systems (6.05 and 5.92 t ha−1 ). Increases in yields due to GCRPS were similar for the wide range of clay contents of soils of this region, illustrating that the successful implementation of GCRPS does not depend on soil textural properties (Fig. 2). Grain yields were positively correlated with shoot dry matter, number of productive tiller and spikelets per square meter (p < 0.001; Table 2). Furthermore, grain yield was also positively correlated with the percentage of filled grains and the harvest index (p < 0.001), but showed no significant relationship to thousandgrain weight. Shoot dry matter, number of productive tillers, spikelets per square meter and percentage of filled grains were significantly higher for GCRPS than for Paddy sytem (Table 3). The spikelets per panicle, thousand grain weight and harvest index tended to be higher in GCRPS than in Paddy systems, but statistical significance was not reached. Within one month after transplanting and before reaching maximum tillering stage, the average daytime soil temperature at 5 cm

Fig. 2. Average grain yield of GCRPS and Paddy system for different soil clay contents (n = 11 for clay content of 10–15%; n = 16 for clay content of 15–30%; n = 9 for clay content of >30%). Error bars represent the standard error of means.

depth was 1.8 ± 0.1 ◦ C higher in fields under GCRPS compared to Paddy fields (Fig. 3a). The difference in accumulated effective temperature between GCRPS and Paddy for the period between transplanting and maximum tillering stage was 58.0 ± 5.5 ◦ C during the daytime and 39.8 ± 5.8 ◦ C during the nighttime, while no significant effect of GCRPS could be demonstrated after the maximum tillering stage (Fig. 3b). The minimum and maximum air temperature during the entire vegetation period was 8.3 ◦ C and 39.8 ◦ C (Fig. 3c). 3.2. Grain yield and yield component in groups separated by yield success According to individual site-specific statistical analyses at each site, the grain yield of 22 pairs (SI) was significantly larger in GCRPS than in Paddy systems, while for 9 pairs (NI) there was an insignificant trend toward increased yield at GCRPS fields and for other 5 pairs (ND) tended to decrease under GCRPS (Fig. 4a). Based on these criteria, the groups of SI, NI and ND were used for further data analyses and interpretations. It is worth noting that none of the three

Table 1 Mean soil texture (clay, silt and sand content), bulk density (BD), mineral nitrogen (Nmin), pH and soil organic carbon- and nitrogen content in the top soil layer 0–20 cm of ground cover rice production system (GCRPS) and Paddy system. Treatments

Clay (%)

Silt (%)

Sand (%)

BD (g cm−3 )

Nmin (kg N ha−1 )

pH

C-content (g kg−1 )

N-content (g kg−1 )

GCRPS Paddy LSD0.05 F-Value

21.1 21.1 4.3 0.0183ns

55.7 55.4 5.8 0.0153ns

23.2 23.5 6.4 0.002ns

1.21 1.22 0.07 0.3418ns

21.9 23.2 7.3 0.023ns

6.72 6.81 0.4 0.0538ns

16.0 14.1 2.1 2.95ns

1.57 1.43 0.20 2.61ns

ns: not significant. For soil texture, n = 36; for BD, Nmin, pH, C- and N-content, n = 108.

Table 2 Pearson’s correlation coefficients for grain yield, shoot dry matter (DrM), number of productive tillers (PT), spikelets per panicle (SP), spikelets per square meter (SM), percentage of filled grains (PFG), harvest index (HI) and thousand-grain weight (TGW) (n = 216). DrM Grain yield DrM PT SP SM PFG HI TGW *

***

0.88 – – – – – – –

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

PT

SP ***

0.64 0.69*** – – – – – –

SM ***

0.41 0.19** −0.10ns – – – – –

PFG ***

0.79 0.75*** 0.73*** 0.41*** – – – –

HI ***

0.41 0.20** 0.05ns 0.10ns 0.03ns – – –

TGW ***

0.38 −0.06ns 0.03ns 0.44*** 0.21** 0.52*** – –

0.14ns 0.18* −0.08ns −0.03ns −0.14ns 0.11ns −0.01ns –

M. Liu et al. / Field Crops Research 150 (2013) 19–28

23

Table 3 Average yield of shoot dry matter (DrM), number of maximum tiller numbers (MTN), number of productive tiller numbers (PT), spikelets per panicle (SP), spikelets per square meter (SM), percentage of filled grains (PFG), harvest index (HI) and thousand-grain weight (TGW) in ground cover rice production system (GCRPS) and Paddy system (n = 108). Treatment

DrM (T ha−1 )

MTN (no. m−2 )

PT (no. m−2 )

SP (no. panicle−1 )

SM (no. ×103 m−2 )

PFG (%)

HI (%)

TGW (g)

GCRPS Paddy LSD0.05 F-Value

13.9 12.0 1.03 13.79***

437 372 59 4.83*

245 213 19 11.5**

143 142 10 0.31ns

34.5 31.4 3.07 4.02*

66.1 61.8 4.1 4.15*

51.2 51.0 2.1 0.42ns

27.2 26.7 0.9 1.67ns

*

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

Table 4 Average of soil clay, silt and sand content, bulk density (BD), mineral nitrogen (Nmin), pH and soil organic carbon- and nitrogen content in the top soil layer (0–20 cm) based on the comparison of three groups. Treatments

Clay (%)

Sand (%)

BD (g cm−3 )

Nmin (kg N ha−1 )

pH

C-content (g kg−1 )

N-content (g kg−1 )

2.4A 2.3A

26.3 ± 2.8A 26.7 ± 2.7A

1.23 ± 0.02A 1.23 ± 0.02A

21.7 ± 1.6A 22.0 ± 2.5A

6.45 ± 0.12A 6.71 ± 0.11A

15.1 ± 0.6A 13.9 ± 0.6A

1.50 ± 0.05A 1.39 ± 0.05A

4.9A 4.1A

16.5 ± 3.6A 17.2 ± 3.8A

1.15 ± 0.03A 1.16 ± 0.04A

23.8 ± 4.1A 28.2 ± 3.7A

7.23 ± 0.15A 7.07 ± 0.18A

17.9 ± 0.6A 14.8 ± 0.6A

1.77 ± 0.05A 1.50 ± 0.06A

5.3A 5.2A

21.6 ± 6.3A 20.9 ± 5.6A

1.21 ± 0.04A 1.24 ± 0.06A

19.1 ± 2.6A 19.4 ± 2.2A

7.01 ± 0.21A 6.82 ± 0.26A

14.4 ± 1.1A 13.9 ± 1.0A

1.53 ± 0.12A 1.46 ± 0.11A

Silt (%) a

Group of significant yield increase (SI) GCRPS 20.1 ± 2.0A 53.6 ± 19.9 ± 1.5A 53.4 ± Paddy Group of non significant increase (NI) 22.8 ± 3.0A 60.7 ± GCRPS Paddy 23.2 ± 3.2A 59.6 ± Group of non significant decrease (ND) GCRPS 22.4 ± 4.6A 56.0 ± Paddy 22.6 ± 4.5A 56.6 ±

Within each column, different capital letters indicate statistically significant difference between GCRPS and paddy systems (p < 0.05). a For the SI group: soil texture, n = 22; BD, Nmin, pH, C- and N-content, n = 66. For the NI group: soil texture, n = 9; BD, Nmin, pH, C- and N-content, n = 27. For the ND group: soil texture, n = 5; BD, Nmin, pH, C- and N-content, n = 15.

statistical groups showed any significant difference in average soil texture, mineral nitrogen content, total organic carbon and nitrogen content between GCRPS and Paddy fields for the entire dataset (Table 4). The average grain yield in the SI group was significantly higher (average 32%) under GCRPS compared to Paddy systems (p < 0.0001; range 8–88%). For the NI group, yields tended to increase by 6% under GCRPS (P > 0.555), and for the ND group there was an average decrease of 8% under GCRPS (P > 0.4253) (Fig. 4b). Straw dry weight, shoot dry matter, maximum tiller number, productive tiller number, spikelets per square meter and percentage of filled grains were significantly larger for GCRPS than for Paddy systems in the SI group, while no significant differences were found between GCRPS and Paddy for the NI and ND groups (Table 5). The temporal course of hourly soil temperature was similar in GCRPS and Paddy fields for all three groups (Fig. 5a). However, we observed a significant increase in the average hourly soil temperature only for the SI group, while no such differences were found for the NI and ND groups (Fig. 5a). The observed differences in accumulated effective temperature at 5 cm soil depth between both systems were highest in the SI group, while the lowest

differences were observed in the ND group (Fig. 5b). For the SI group it increased dramatically between the two systems for the time interval spanning transplanting to maximum tillering stage. In contrast, differences in accumulated effective temperature barely increased for the NI and ND groups during the entire growth period (Fig. 5b). Values of leaf ␦13 C ranged from −29.0‰ to −29.3‰ at the maximum tillering stage. For the SI group, plant leaf ␦13 C (Fig. 6a) and soil redox potential at 5 cm (Fig. 6b) were significant higher in GCRPS than in Paddy systems, while no significant difference was found in the ND group. 4. Discussion Here we provide strong evidence based on a robust dataset that GCRPS significantly increased rice grain yields as compared to adjacent paddy fields in the mountainous area of Shiyan, Central China. Out of 36 paired sites, 22 of them showed significant yield and biomass increases, while only at 5 sites a non significant reduction in yield was observed for GCRPS (Figs. 1 and 4a). Previous studies reporting reduced or unchanged grain yields under

Table 5 Observed yield of straw and shoot dry matter (DrM), number of maximum tillers numbers (MTN), number of productive tillers numbers (PT), spikelets per panicle (SP), spikelets per square meter (SM), percentage of filled grains (PFG), harvest index (HI), thousand-grain weight (TGW) among three groups between GCRPS and Paddy system. Straw yield (T ha−1 )

DrM (T ha−1 )

Group of significant increase (SI)a GCRPS 7.08 ± 0.15A 14.4 ± 0.27A Paddy 5.54 ± 0.16B 11.1 ± 0.25B Group of non significant increase (NI) 6.44 ± 0.23A 13.8 ± 0.38A GCRPS 6.90 ± 0.24A 13.9 ± 0.34A Paddy Group of non significant decrease (ND) 5.89 ± 0.21A 12.1 ± 0.44A GCRPS Paddy 5.82 ± 0.32A 12.6 ± 0.54A

MTN (no. m−2 )

PT (no. m−2 )

SP SM (no. (no. panicle−1 ) ×103 m−2 )

PFG (%)

HI (%)

TGW (g)

440 ± 17A 353 ± 16B

250 ± 5.5A 203 ± 6.2B

145 ± 2.5A 140 ± 2.7A

35.8 ± 0.84A 30.3 ± 0.82B

65.0 ± 1.1A 60.3 ± 0.97B

50.8 ± 0.39A 50.2 ± 0.60A

27.1 ± 0.18A 26.8 ± 0.27A

440 ± 19A 425 ± 20A

241 ± 7.3A 244 ± 5.0A

143 ± 3.1A 149 ± 6.2A

34.4 ± 1.2A 36.2 ± 1.2A

67.3 ± 2.2A 62.8 ± 2.6A

53.3 ± 1.1A 50.3 ± 1.2A

27.8 ± 0.32A 26.1 ± 0.40A

407 ± 32A 355 ± 27A

228 ± 7.1A 206 ± 9.2A

138 ± 5.3A 148 ± 5.3A

29.0 ± 1.3A 33.5 ± 2.5A

69.3 ± 1.0A 68.8 ± 1.2A

50.9 ± 0.92A 53.9 ± 0.72A

26.6 ± 0.92A 27.0 ± 0.41A

Within each column, different capital letters indicated significant difference between GCRPS and paddy treatment based on the means of each group (p < 0.05). a For group of significant increase, n = 66; for group of non significant increase, n = 27; for group of non significant decrease, n = 15.

24

M. Liu et al. / Field Crops Research 150 (2013) 19–28

(a)

7 6 5 4

20

3

o

30

Difference ( C)

GCRPS Paddy Difference

o

the eventual abandonment of the GCRPS-direct seeding technique under dry conditions (Qu et al., 2012). (ii) GCRPS-transplanting: this agronomic technique is mainly applied in subtropical humid regions presenting clayey to loamy soils. In this GCRPS modality, approximately one month old rice seedlings are transplanted on the hills, while the soil is kept close to water saturation during the entire growing season (Shen et al., 1997; Qu et al., 2012). This later technique has been promoted due to its significant potential for saving irrigation water, and it has therefore been the preferred choice in various regions of China (Fan et al., 2005; Liu et al., 2009; Yang and Zhang, 2010; Qu et al., 2012). The apparent success of the GCRPS in the Shiyan region is closely linked to the environmental limitations existing for rice production in this area. Namely the increasing limitations in irrigation water during spring time and the relatively short vegetative period existing in this mountainous environment. However, a sound assessment of yield performances using GCRPS at a regional scale has so far not been explored. One may assume that the success of GCRPS cultivation with regard to yields may vary with soil properties, production management and climate condition (Jin et al., 2002; Liu et al., 2003; Qu et al., 2012). However, we did not found any significant differences in soil texture and soil chemical properties between GCRPS and Paddy at both regional (Table 1) and group scales (Table 4). Furthermore, we could not detect any significant relationship between soil properties and grain yields (data

2 10

6

Max. Tillering stage

0 -1

-1

0 80

(b)

0600-1800 h 1800-0600 h

60 o

DAET ( C)

Significant increase Non significant increase Non significant decrease

1

Difference (t ha )

Soil temperature at 5 cm ( C)

GCRPS cultivation have been conducted either in subtropical climate that is characterized by having a nearly absolute absence of water shortages and temperature limitation for plant growth (Liang et al., 1999; Fan et al., 2002; Yang and Zhang, 2010), or in areas with markedly sandy soils on which the GCRPS–direct seeding technique was applied (Lin et al., 2002; Tao et al., 2006). Indeed, two different GCRPS techniques have been used under different soil types and climatic conditions over the past 20 years. (i) GCRPS-direct seeding: lowland rice seeds get directly seeded at a soil water content of 80–90% of its water holding capacity in relatively dry areas on sandy soils (Lin et al., 2002; Xu et al., 2005; Tao et al., 2006). However, a range of disadvantages have been reported using this technique such as potential deficiency of the micronutrients Mn and Fe (Lin et al., 2002; Tao et al., 2007; Kreye et al., 2009), difficulty of weed control in the first weeks following seeding (Zhao et al., 2007), and increased infestation under conditions of continuous mono cropping (Peng et al., 2006; Nie et al., 2009). All of which has lead to

40

(a)

4

2

0 -20

0

20

40

60

80

100

Relative increase (%)

20

-2 0

10

Minimum Maximum

40

a

(b) a

a

o

-1

Grain yield (t ha )

8 Air temperature ( C)

GCRPS Paddy

(c)

30

20

a

a

b

6 4 2

10

0 SI 5/15

5/29

6/12

6/26

7/10

7/24

8/7

8/21

NI

ND

9/4

Fig. 3. (a) Average and absolute difference of soil temperature observed at 5 cm depth for GCRPS and Paddy system in daytime (06:00–18:00 h); (b) difference of accumulated effective temperature (DAET) between GCRPS and Paddy system after transplanting in two daily intervals (n = 36); (c) daily minimum and maximum air temperature during the entire vegetation period in 2011 in Fang County (32◦ 03,281 , 110◦ 44,447 ). Error bars represent the standard error of means.

Fig. 4. (a) Relationship between the absolute difference and relative increase of grain yields in GCRPS and Paddy system for each statistically defined sampling sites; (b) average grain yields in GCRPS and Paddy system for the three statistically defined groups: group of significant yield increase (SI), n = 66; group of non significant increase (NI), n = 27; group of non significant decrease (ND), n = 15. Different letters denote significant differences between GCRPS and Paddy system. Error bars represent the standard error of the means.

M. Liu et al. / Field Crops Research 150 (2013) 19–28

GCRPS Paddy

Group of significant increase

40

25

SI

(a)

ND

-28

o

Soil temperature at 5 cm ( C)

20

C (‰)

0 Group of non significant increase

40

-29

a

20

b 0

a

GCRPS Paddy

Group of non significant decrease

40

a

(a)

-30 20 0 5/15

5/17

5/19

5/21

5/23

5/25

5/27

5/29

-40 (b)

o

DAET ( C)

80 60

Eh (mV)

Group of significant increase Group of non significant increase Group of non significant decrease

a

-60 -80 -100 -120

40

-140

20

-160

0 5/15 5/29 6/12 6/26 7/10 7/24

ND

-20

120 100

SI

0

8/7

8/21

9/4

Fig. 5. (a) Average hourly soil temperature in GCRPS and Paddy system for the group of significant yield increase (SI), group of non significant increase (NI) and group of non significant decrease (ND); (b) difference of accumulated effective temperature (DAET) during daytime (06:00–18:00 h) between GCRPS and Paddy system after transplanting for the SI, NI and ND groups (n = 22 for SI; n = 9 for NI; n = 5 for ND). Error bars represent standard error of the means.

not shown). Moreover, the increase in yields observed using GCRPS were not significantly different across a wide range of soil clay contents (Fig. 2). Our data suggests that the observed increase in yields is most likely related to temperature limitations during the period between transplantation and maximum tillering stage (Fig. 3c). Usually, rice plants prefer warm conditions and do not tolerate temperatures below 15 ◦ C, especially during the early growth stage (Tao et al., 2006), which are conditions that correspond well to those prevalent in the Shiyan region. Analysis of yield components may help unraveling the mechanisms that can explain how increases in temperature during the first stages of the rice development may finally affect grain yields. From the point of view of crop production, photosynthetic products before and post anthesis as well as retranslocation of previous photosynthates into the grain are all important processes for grain yield, which is determined by productive tiller number, spikelets per panic and thousand-grain weight (Yang and Zhang, 2010). Our dataset shows that under GCRPS the number of productive tillers as well as the maximum tiller number gets significantly increased (Table 2), which agrees well with similar results reported in previous studies (Tao et al., 2006; Qu et al., 2012). Moreover, we also found that major yield parameters such as spikelets per square meter, and percentage of filled grains (Yang and Zhang, 2010) are in average 5–10% higher for GCRPS as compared to

a GCRPS Paddy

b a

(b)

Fig. 6. (a) Leaf ␦13 C and (b) soil redox potential (Eh) at maximum tillering stage for the group of significant yield increase (SI, n = 66) and the group of non significant decrease (ND, n = 15) in both GCRPS and Paddy system. Error bars represent standard error of the means.

Paddy (Fig. 1, Table 3) and even this figure rises to 10–30% for the SI group (Table 5). These facts indicate that low temperature in the vegetative growth stage hampered the early formation of tillers (Huang et al., 1999) and thus, resulted in the smaller grain yields obtained in Paddy systems compared to those observed in GCRPS, where such temperature limitation was avoided (Fig. 3). Average daily temperature and accumulated effective soil temperature were increased from transplanting to maximum tillering stage employing the GCRPS (Fig. 3). Higher soil temperatures promoted larger straw yields and the formation of early tiller numbers which resulted in a significant positive effect on the number of productive tillers and the percentage of filled grains observed in GCRPS than Paddy (Tables 3 and 5; Qu et al., 2012). The significant increase of maximum and productive tiller numbers (Tables 3 and 5) was also mirrored by a significant increase in spiklets per square meter. Since formation and subsequent disposal of dispensable tillers are regulatory adaptations of rice plants in response to temperature and water availability (Tao et al., 2006), it shows that GCRPS positively effects plant performance. As discussed before our data show that for most sites GCRPS resulted in strong increases in temperature during the crucial first month after transplanting (Figs. 3 and 5b; Qu et al., 2012). In the few cases where we found a non significant reduction in grain yield under GCRPS cultivation (ND group), we also did not find a positive temperature effect of GCRPS as compared to the adjacent Paddy field. The most likely explanation for a missing temperature effect is that farmers flooded GCRPS fields, since flooding would override the effect of the plastic film on soil temperatures. Besides the accumulated temperature effect, other findings also

26

Table A1 The geographical information of all 36 paired fields (The order of pair number from top to bottom of this table is following the point of Fig. 4a from right to left). Group

Pair

County

Towship

Village

Householder GCRPS

a

a b

7 5 10 2 6 9 12 21 14 17 23 13 26 16 4 1 19 35 27 31 32 30 3 15 28 20 18 11 34 24 36 25 22 29 8 33

Zhushan Zhushan Zhushan Danjingkou Zhushan Zhushan Zhushan Yunxi Yun Yunxi Yunxi Zhushan Fang Yun Yun Shiyan Yunxi Fang Fang Fang Fang Fang Yun Yun Fang Yunxi Yunxi Zhushan Fang Yunxi Fang Yunxi Yunxi Fang Zhushan Fang

Wenfeng Wenfeng Baofeng Liuliping Wenfeng Baofeng Baofeng Guanfang Baoxia Guanfang Guanfang Baofeng Tucheng Baoxia Baoxia Zhangwan Guanfang Chengguan Tucheng Yaohuai Yaohuai Yaohuai Baoxia Baoxia Tucheng Guanfang Guanfang Baofeng Chengguan Guanfang Chengguan Dianzi Guanfang Tucheng Wenfeng Chengguan

Tangwan Tangwan Shangba Haokou Tangwan Shangba Longjing Huilong Gaoqiao Huilong Huilong Longjing Baiji Gaoqiao Gaoqiao Huangtu Huilong Binggong Baiji Yaochang Yaochang Yaochang Fenshuiling Gaoqiao Baiji Huilong Huilong Longjing Binggong Huilong Binggong Tianbaoshan Huilong Baiji Tangwan Binggong

Su Hengtai Su Hengshan Wang Daming Li Yugen Su Hengshan Zhang Xianshe Zhang Zugen Zhong Xiumei Wang Yongcheng Zhu Rongyu Zhu Donggen Du Xiangmei Kuang Jianwen Yu Benxing He Xianchun Zhang Dehua Zhu Rongyu Che Youju Yang Wanying Zhou Facai Zhang Chenghua Zou Hongquan Dong Huanquan Yu Benqing Li Ruwu Ming Lizheng Kai Cailiang Li Daolin Che Haizhao Peng Jinao Du Deyong Xie Yonggui Xie Changquan Liu Yunhua Yang Yingyi Li Wenyun

Su Henglin Wen Shilin Liu Zhenhua Ming Pinghe Su Henglin Zhang Shizhong Li Daolin Liu Xiangcheng Zhan Shiyou Zhu Xingwu Zhu Qinghe Li Daolin Wu Meiqing Zhan Shili He Xiangen Zhang Huibing Kang Caixian Xu Qiang Yu Xingli Lu Dinghua Wang Jie Zhu Danfen Yan Keren Xie Huali Sun Xueyu Zhu Xinghou Zhu Ronglan Zhu Shijun Zheng Huifu Zhu Xingzhi Li Wenyun Zhang Hejun Liu Xiangcheng Yu Lichun Gan Shibing Ji Taifa

SI: group of significant yield increase; NI: group of non significant increase; ND: group of non significant decrease. Distance: distance between GCRPS and Paddy field in each pair.

Longitude E

Latitude N

Longitude E

Altitude (m)

Distanceb

GCRPS

GCRPS

Paddy

Paddy

GCRPS

Paddy

(m)

419 438 426 169 436 426 500 656 652 539 661 505 600 650 650 216 567 467 600 622 618 621 350 650 599 558 529 502 466 584 466 585 656 595 427 466

447 449 425 186 436 424 503 656 652 539 649 504 590 649 649 224 570 462 595 593 593 597 352 649 598 567 527 508 458 546 459 585 656 589 452 460

206 75 60 589 12 108 88 17 10 18 19 54 458 31 75 87 20 114 337 974 1010 1132 32 55 238 297 18 89 145 1381 140 16 37 217 22 100

◦



32 12,047 32◦ 11,926 32◦ 18,408 32◦ 31,404 32◦ 12,009 32◦ 18,406 32◦ 20,456 33◦ 10,422 32◦ 35,963 33◦ 09,264 33◦ 10,426 32◦ 20,484 32◦ 13,835 32◦ 35,958 32◦ 35,972 32◦ 38,063 33◦ 09,530 32◦ 02,337 32◦ 13,840 32◦ 06,729 32◦ 06,42.4 32◦ 06,725 32◦ 41,959 32◦ 35,970 32◦ 13,849 33◦ 09,592 33◦ 09,210 32◦ 20,460 32◦ 02,338 33◦ 09,946 32◦ 02,340 33◦ 08,215 33◦ 10,425 32◦ 13,774 32◦ 11,882 32◦ 02,342

◦



110 17,748 110◦ 17,834 109◦ 59,422 111◦ 03,387 110◦ 17,783 109◦ 59,418 109◦ 56,866 109◦ 45,475 110◦ 15,636 109◦ 44,013 109◦ 45,410 109◦ 56,874 110◦ 42,727 110◦ 15,671 110◦ 15,668 110◦ 37,716 109◦ 44,181 110◦ 43,307 110◦ 42,727 110◦ 27,762 110◦ 27,45.0 110◦ 27,758 110◦ 21,707 110◦ 15,672 110◦ 42,729 109◦ 44,241 109◦ 44,009 109◦ 56,865 110◦ 43,313 109◦ 44,605 110◦ 43,328 109◦ 50,498 109◦ 45,444 110◦ 42,875 110◦ 17,495 110◦ 43,333

◦



32 12,131 32◦ 11,896 32◦ 18,396 32◦ 31,045 32◦ 12,014 32◦ 18,438 32◦ 20,496 33◦ 10,417 32◦ 35,968 33◦ 09,273 33◦ 10,422 32◦ 20,485 32◦ 13,73 32◦ 35,948 32◦ 35,946 32◦ 38,043 33◦ 09,519 32◦ 02,375 32◦ 13,758 32◦ 06,940 32◦ 06,940 32◦ 06,987 32◦ 41,978 32◦ 35,954 32◦ 13,748 33◦ 09,467 33◦ 09,201 32◦ 20,457 32◦ 02,409 33◦ 09,341 32◦ 02,413 33◦ 08,211 33◦ 10,420 32◦ 13,722 32◦ 11,870 32◦ 02,396

◦



110 17,662 110◦ 17,802 109◦ 59,387 111◦ 03,491 110◦ 17,778 109◦ 59,361 109◦ 56,835 109◦ 45,466 110◦ 15,633 109◦ 44,017 109◦ 45,399 109◦ 56,840 110◦ 42,990 110◦ 15,687 110◦ 15,704 110◦ 37,767 109◦ 44,181 110◦ 43,298 110◦ 42,919 110◦ 27,166 110◦ 27,168 110◦ 27,093 110◦ 21,675 110◦ 15,702 110◦ 42,822 109◦ 44,121 109◦ 44,004 109◦ 56,922 110◦ 43,279 109◦ 44,075 110◦ 43,297 109◦ 50,507 109◦ 45,467 110◦ 42,998 110◦ 17,492 110◦ 43,340

M. Liu et al. / Field Crops Research 150 (2013) 19–28

SI SI SI SI SI SI SI SI SI SI SI SI SI SI SI SI SI SI SI SI SI SI NI NI NI NI NI NI NI NI NI ND ND ND ND ND

Paddy

Latitude N

M. Liu et al. / Field Crops Research 150 (2013) 19–28

point out toward continuous flooding of GCRPS fields as being the main reason behind the reduction in yields observed in this group. These findings are as follows: (1) Magnitudes and diurnal variations observed in hourly soil temperature between paddy and GCRPS were similar in the ND group, while both minimal and maximal soil temperatures were clearly larger in both systems in the SI group (Fig. 5a); (2) Soil redox potentials were significantly higher under GCRPS cultivation at the SI group, but not at the ND group (Fig. 6b); (3) Finally, ␦13 C signatures of plant leaves remained unchanged between paddy and GCRPS cultivation for the ND group, but they were significantly higher under GCRPS than Paddy in the SI group (Fig. 6). The trend in ␦13 C observed in the ND group indicates comparable photosynthetic discrimination of ␦13 C across GCRPS and Paddy fields because of similar stomata conductance under analogous water levels (Schulze et al., 1998; Tao et al., 2007). Our finding that increases in soil temperatures are crucial for enhanced crop yields under GCRPS can obviously help to explain why GCRPS had no positive effect on grain yield in a tropical region of southern China, as in such location, temperature may not be a limiting factor for plant growth (Fan et al., 2002; Yang and Zhang, 2010). In contrast, a number of studies report increased grain yield under GCRPS in mountainous areas of central China where temperature and water were limited in the early stage of rice development (Shen et al., 1997; Liu et al., 2009; Qu et al., 2012). Overall we suggest a strong potential of GCRPS cultivation to increase rice grain yield at regional scales in climate zones where temperature and water availability are limiting, e.g. in northeast China (Heilongjiang Province, Jilin and Liaoning Province), which in principle present some very favorable soil conditions for rice production, such as large soil organic carbon and nitrogen contents, but also strong temperature limitations and water demands. 5. Conclusions Building on a robust and spatially representative dataset gained under real farming conditions, we demonstrate for the first time that GCRPS can increase grain and straw yield at regional scales under varying edaphic conditions. However, the success of GCRPS application is largely depending on appropriate water management avoiding excess water on GCRPS fields. When these water management guidelines are closely followed, GCRPS is an invaluable and stable tool to increase yields in regions where temperature and water significantly limits rice growth. Acknowledgements This work was supported by the National Natural Science Foundation of China (NSFC 51139006) and the Sino-German Center for Science Promotion (GZ667). We would like to thank the Agriculture Department of Shiyan for providing working facilities and long term cooperation. We thank Jing Ruying, Jiang Zengming, Fan Zhaobo, Yao Zhisheng from Beijing; Li Jiajun, Shen Kangrong, Wang Xiaochun, Liu Jun, Wei Guangjun from Shiyan; Luo Jianhua, Fang Yunsheng from Fang County; Sun Yingming from Yunxi County; Chen Xinju from Yun County; Yang Daming, Shi Yuzhi from Zhushan County; Zhao Tianzhong, Zhu Mingli from Danjinakou County for their help during sampling and the field measurements. Appendix A. See Table A1. References Abdulai, A., Glauben, T., Herzfeld, T., Zhou, S., 2005. Water saving technology in Chinese rice production: evidence from survey data. In: The Future of Rural Europe

27

in the Global Agri-Food System. XIth International Congress of the European Association of Agricultural Economists, Copenhagen, Denmark, August 23–27. Belder, P., Spiertz, J.H.J., Bouman, B.A.M., Lu, G., Tuong, T.P., 2005. Nitrogen economy and water productivity of lowland rice under water-saving irrigation. Field Crops Res. 93, 169–185. Blake, G.R., Hartge, K.H., 1986. Bulk density. In: Arnold Klute (Ed.), Methods of Soil Analysis. Part 1. Physical and Mineralogical Methods. Madison, WI, pp. 363–375. Bouman, B.A.M., Humphreys, E., Tuong, T.P., Barker, R., 2006. Rice and water. Adv. Agron. 92, 187–237. Bouman, B.A.M., 2007. A conceptual framework for the improvement of crop water productivity at different spatial scales. Agric. Syst. 93, 43–60. 2011. FAOSTAT. Food Agricultural Organization, UN http://faostat.fao.org/ Gee, G.W., Bauder, J.W., 1986. Particle size analysis. In: Klute, A. (Ed.), Methods of Soil Analysis. Part 1. Physical and Mineralogical Methods. Agronomy, 9. , 2nd ed. Soil Sci. Soc. Am. J., Madison, USA, pp. 383–411. Fan, X.L., Zhang, J.P., Wu, P., 2002. Water and nitrogen use efficiency of lowland rice in ground covering rice production system in South China. J. Plant Nutr. 25, 1855–1862. Fan, M.S., Liu, X.J., Jiang, R.F., Zhang, F.S., Lu, S.H., Zeng, X.Z., Christie, P., 2005. Crop yields, internal nutrient efficiency, and changes in soil properties in rice–wheat rotations under non-flooded mulching cultivation. Plant Soil 277, 265–276. Harris, D., Horwath, W.R., Kessel, C., 2001. Acid fumigation of soils to remove carbonates prior to total organic carbon or carbon-13 isotopic analysis. Soil Sci. Soc. Am. J. 65, 1853–1856. Huang, Y.D., Zhang, Z.L., Wei, F.Z., Li, J.C., 1999. Ecophysiological effect o drycultivated and plastic film-mulched rice planting. Chin. J. Appl. Ecol. 10, 305–308 (in Chinese with English abstract). IRRI, 2007. International Rice Research Institute, Los Banos, Laguna, Philippines. http://beta.irri.org/statistics/ IRRI, 2011. International Rice Research Institute, Los Banos, Laguna, Philippines. http://irri.org/news-events/hot-topics/international-landacquisition-for-riceproduction?print=1&tmpl=component. Accessed November 10, 2011. Jin, Q.Y., Ouyang, Y.N., Zhang, G.P., 2002. Yield performance and growth characteristics of rice grown in plastic film mulched dry-land. J. Zhejiang Agric. Univ. 28, 362–368 (in Chinese with English abstract). Kreye, C., Bouman, B.A.M., Reversal, G., Fernandez, L., Vera Cruz, C., Elazegui, F., Faronilo, J.E., Llorca, L., 2009. Biotic and abiotic causes of yield failure in tropical aerobic rice. Field Crops Res. 112, 97–106. Leesawatwong, M., Jamjod, S., Kuo, J., Dell, B., Rerkasem, B., 2005. Nitrogen fertilizer increases seed protein and milling quality of rice. Cereal Chem. 82, 588–593. Liang, Y.C., Hu, F., Yang, M.C., Zhu, X.L., Wang, G.P., Wang, Y.L., 1999. Mechanism of high yield and irrigation water use efficiency of rice in plastic film mulched dryland. Sci. Agric. Sin. 32, 26–32 (in Chinese with English abstract). Lin, S., Dittert, K., Tao, H.B., Kreye, C., Xu, Y.C., Shen, Q.R., Fan, X.L., 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.M., Hengsdijk, H., Hardy, B., Bindraban, P.S., Tuong, T.P., Ladha, J.K. (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, X.J., Wang, J.C., Lu, S.H., Zhang, F.S., Zeng, X.Z., Ai, Y.W., Peng, S.B., Christie, P., 2003. Effects of non-flooded mulching cultivation and nutrient management on crop growth, nutrient uptake and balances in rice–wheat cropping systems. Field Crops Res. 83, 297–311. Liu, J., Liu, M.J., Guan, Y.F., Cao, J., Ji, Z.H., Li, J.J., Wang, X.C., Shen, K.R., Lin, S., 2009. Grain yield and nitrogen uptake affected by the Ground Cover Rice Production System with plastic film covering. J. China Agric. Univ. 15, 9–17 (in Chinese with English abstract). Maclean, J.L., Dawe, D., Hardy, B., Hettel, G.P., 2002. Rice Almanac, 3rd ed. Interna˜ tional Rice Research Institute (IRRI), Los Banos, Philippines. MWR (Ministry of Water Resources, P.R. China), 2007. The 11th Five-Year Plan of National Water Resources Development. Gazette of the Ministry of Water Resources of the P.R. China, pp. 34–48. Nie, L.X., Xiang, J., Peng, S.B., Bouman, B.A.M., Huang, J.L., Cui, K.H., Visperas, R.M., 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. Peng, S.B., Shen, K.R., Wang, X.C., Liu, J., Luo, X., Wu, L., 1999. A new rice cultivation technology: plastic film mulching. Int. Rice Res. News Lett. 24, 9–10. Peng, S.B., Huang, J.L., Sheehy, J.E., Laza, R.C., Visperas, R.M., Zhong, X.H., Centeno, G.S., Khush, G.S., Cassman, K.G., 2004. Rice yields decline with higher night temperature from global warming. Proc. Natl. Acad. Sci. U. S. A. 101, 9971–9975. ˜ Peng, S.B., Bouman, B.A.M., Visperas, R.M., 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.T., Hu, F., Zhang, B., Wei, Z.G., Li, H.X., 2006. Role of straw mulching in noncontinuously flooded rice cultivation. Agric. Water Manage. 83, 252–260. Qu, H., Tao, H.B., Tao, Y.Y., Liu, M.J., Shen, K.R., Lin, S., 2012. Ground cover rice production system increases yield and nitrogen recovery efficiency. Agron. J. 104, 1399–1407. Schulze, E.D., Williams, R.J., Farquhar, G.D., Schulze, W., Langridge, J., Miller, J.M., Walker, B.H., 1998. Carbon and nitrogen isotope discrimination and nitrogen

28

M. Liu et al. / Field Crops Research 150 (2013) 19–28

nutrition of trees along a rainfall gradient in northern Australia. Funct. Plant Biol. 25, 413–425. Shen, K.R., Wang, X.C., Luo, X.S., 1997. Test and demonstration on wet-cultivation with film mulching of rice. Hubei Agric. Sci. 5, 18–22 (in Chinese with English abstract). Smit, B., Cai, Y.L., 1996. Climate change and agriculture in China. Global Environ. Change 6, 205–214. Tao, H.B., 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.B., Dittert, K., Zhang, L.M., Lin, S., Roemheld, 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. Tso, T.C., 2004. Agriculture of the future. Nature 428, 215–217. Tuong, T.P., Bouman, B.A.M., 2003. Rice production in water scarce environments. In: Kijne, J.W., Barker, R., Molden, D. (Eds.), Water Productivity in Agriculture: Limits and Opportunities for Improvement. CABI Publishing, Wallingford, UK, pp. 53–67. Tuong, T.P., Bouman, B.A.M., Mortimer, M., 2005. More rice, less water-integrated approaches for increasing water productivity in irrigated rice-based systems. Plant Prod. Sci. 8, 231–241.

Walthert, L., Graf, U., Kammer, A., Luster, J., Pezzotta, D., Zimmermann, S., Hagedorn, F., 2010. Determination of organic and inorganic carbon. ␦13 C, and nitrogen in soils containing carbonates after acid fumigation with HCl. J. Plant Nutr. Soil Sci. 173, 207–216. Wopereis-Pura, M.M., Watanabe, H., Moreira, J., Wopereis, M.C.S., 2002. Effect of late nitrogen application on rice yield, grain quality and profitability in the Senegal river valley. Eur. J. Agron. 17, 191–198. Wu, L.H., Zhu, Z.R., Liang, Y.C., Shi, W.Y., Zhang, L.M., 1999. The development of the rice film mulching cultivation. J. Zhejiang Univ. 25, 41–42 (in Chinese with English abstract). Xu, Y.C., Zhang, Y.L., Shen, Q.R., Xu, Y., Zhang, J., 2005. An innovation method for the treatment of rice straw to improve nitrogen uptake efficiency. Biol. Fert. Soils 41, 291–294. Yang, J.C., Zhang, J.H., 2010. Crop management techniques to enhance harvest index in rice. J. Exp. Bot. 61, 3177–3189. Yong, J., 2009. China’s water scarcity. J. Environ. Manage. 90, 3185–3186. Zhao, D.L., Bastiaans, L., Altin, G.N., Spiertz, J.H.J., 2007. Interaction of genotype and management on vegetative growth and weed suppression of aerobic rice. Field Crops Res. 100, 327–340. Zhou, S.D., Herzfeld, T., Glauben, T., Zhang, Y.H., Hu, B.C., 2008. Factors affecting Chinese farmers’ decisions to adopt a water-saving technology. Can. J. Agric. Econ. 56, 51–61.