YUAN Peng-li,WANG Jin-ping,,GUO Can,GUO Zi-yuan,GUO Yao,CAO Cou-gui,

1 Laboratory of Crop Ecophysiology and Farming System in the Middle Reaches of Yangtze River,College of Plant Science and Technology,Huazhong Agricultural University,Wuhan 430070,P.R.China

2 Hubei Collaborative Innovation Center for Grain Industry,Yangtze University,Jingzhou 434023,P.R.China

Abstract The rice-crayfish farming model has been rapidly developed and become an economically viable method to supply food in China in recent years.However,its environmental and economic sustainability has not been thoroughly investigated.This study uses a survey in 2016 and a field experiment in 2017 in Qianjiang,Hubei Province,China to assess the relative economics of concurrent rice-wheat (RW),rice-crayfish (RC),and crayfish monoculture (CM) models in waterlogged land areas.The field survey indicated that the RC model had a higher benefit-cost ratio (3.5:1) than the RW(2.0:1) and CM (3.1:1) models and the RC model protected farmers’ enthusiasm for grain production facing unfavourable weather conditions.The field experiment aimed to explore nitrogen management strategies in RC fields.In the experiment,four levels of nitrogen concentration gradient -0 kg N ha-1 (0 N),75 kg N ha-1 (75 N),150 kg N ha-1 (150 N) and 225 kg N ha-1 (225 N),were set in a 2-year-old rice-crayfish (RC2) field,an 8-year-old rice-crayfish (RC8) field,and a RW field as a control.The field experiment results suggested that the peak rice yield in RW,RC2,and RC8 occurred when 225 N,150 N and 75 N were used,respectively.In RC2 and RC8,however,residual feed-nitrogen that was not used by crayfish was utilized by rice plants.Thus,an optimal amount of nitrogen in RC fields was proposed to improve the nitrogen use efficiency and reduce environmental pollution by nitrogen fertilizer.Farmers use less nitrogen but have higher net income in RC than in RW and CM.It is necessary to sustainably develop integrated farming technologies(i.e.,proper field configurations for rice fields) to effectively sustain rice production.The results also showed that the RC farming model was a viable diversification option for rice farmers in waterlogged land.

Keywords:rice-crayfish,benefit-cost ratio,lodging,nitrogen application

1.lntroduction

Rice provides food for more than half of the world’s population,and the annual per capita rice consumption of many Asian countries exceeds 100 kg (Langerwischet al.2018).UNDESAPD (2015) estimated the global population would increase from 7.4 billion to 9.7 billion by 2050,which will drive the demand for rice in the coming decades.Rice is the biggest grain crop in China.Its total annual planting area was 30.3 million ha in 2020,accounting for 18.5% of the world’s rice planting area,second only to India (FAO 2020).Moreover,its production accounted for 26.7% of the world total in 2020,ranking the first among all rice-producing countries (FAO 2020).Nearly 200 million tons of rice are consumed annually worldwide,demonstrating the important role that rice production plays in ensuring food security.

However,there are labor shortages in rural farms due to urban expansion (Cai and Chen 2000).In 2014,18 million workers migrated from rural areas to cities in China,most of whom were 18 to 40 years old (Li and Zhu 2014).The quantity and quality of labor in rice production have significantly declined (Gaiet al.2014),despite the current major income source of most Chinese farmers is grain production.In recent years,the Chinese government has ceaselessly issued various preferential policies such as raising farmgate grain prices to promote grain production,increase farmers’ income and close the income gap between urban and rural residents.However,farmers’ income throughout the country continued to grow slowly for many years,and income growth in major grain-producing areas is lower than the national average,which has seriously affected farmers’ livehood and lowered their enthusiasm for growing grain crops (Guo and Wan 2013).

Recently,a new rice-crayfish farming model has emerged in the lower-middle reaches of the Yangtze River basin,mainly in Hubei,Hunan,Jiangxi,Anhui,and Jiangsu provinces.This farming model has significantly increased farmers’ income.Experimental research has shown that compared with rice monoculture,the total output,profit and ratio of output to input in the rice-crayfish model increased by 46 818.0 CNY ha-1,40 188.0 CNY ha-1and 100.0%,respectively (Siet al.2017).Consequently,the rice-crayfish model has rapidly expanded in the region.In 2017,the area of paddy fields used for crayfish breeding was about 566 667 ha,accounting for 70.83% of the total crayfish aquaculture area (MARABFet al.2018).The total area of ricecrayfish farming reached 278 000 ha in Hubei Province in 2017 (Hubei Statistics Office 2018).Farmers in Qianjiang City in Hubei Province,China,located in the hinterland of Jianghan Plain,are the main rice-crayfish model users in the region.The rice-crayfish farming model,originated in 2000,was adopted in 32 900 ha in 2017,accounting for 91% of the total rice planting area in Qianjiang (Suet al.2018).

This study conducted a survey in 2016 to evaluate the benefit-cost ratio in the rice-crayfish farming model in different villages in Qianjiang.The survey results suggested that farmers put significant amounts of feed into rice fields to increase crayfish yield.Therefore,it is hypothesized that feed residues and crayfish feces were used as part of nitrogen fertilizers in paddy fields.In 2017,a gradient test of nitrogen fertilizer was conducted in rice fields with different crayfish breeding years in Guanshan Farm in Qianjiang.

2.Data and methods

2.1.Study sites

Qianjiang is located in the Jianghan Plain hinterland in the central-southern Hubei Province (30°09′-30°35′N,112°31′-112°59′E),a region with a north subtropical monsoon humid climate.The historical large-scale subsidence and its sealing accumulation haveformed lowlying flat land,plain,and lake landforms with numerous rivers and lakes in this region.The height of the land varies from 26 m above sea level in the southeast to 38 m in the northwest.

The region from the Han River and Dongjing River to the lake area contains three main landforms which form three planting belts:a high-elevation and concentrated cotton planting area;the middle flatland with dual flood and drought zones;and a low-humidity lakeside ricegrowing area.The region is located in the Middle-Lower Yangtze Plains,well known as“the land of fish and rice”.Many lakes and rivers in this region make it particularly suitable for rice growing and freshwater fish farming.Rice growing areas are concentrated in the southwest of this region with a planting area of 59 060 ha (Hubei Statistics Office 2018)

2.2.Farm survey

We compared the input and output characteristics of the rice-wheat (RW) rotation,rice-crayfish co-culture (RC),and crayfish monoculture (CM) models using a farmer field survey.A total of 236 farmers were surveyed in 2016 in the main rice-growing regions of Qianjiang.The sample included 100 traditional RW farmers,100 RC farmers,and 36 CM farmers.The survey included 12 villages in 9 towns (Appendix A),with a random selection of farmers in each village.Each pair of RM,RC,and CM farmers were located in the same village to ensure similar climate and soil conditions.The rice and crayfish yields for each farmer were determined using data collected during harvest.Rice and crayfish yields were measured as air-dried weight and fresh weight,respectively,and are expressed as t ha-1.The survey was also used to determine the percentage of ditches used for crayfish in each paddy field (ditch percentage).Nitrogen (N)fertilizers,pesticides,and feeds applied by each farmer in the field were recorded during the rice,wheat,and crayfish growing season.The fertilizer-N was calculated as N per ha per year.

Rice lodging of these farmers in 2017 was investigated.At the end of 2016,some farmers transformed RW fields into RC,which decreased the number of RW farmers surveyed in that year.Therefore,the investigation only included RW farmers and RC farmers who remained the same farming model in 2016 and 2017 and eliminated those who changed farming model.

2.3.Field experiment

Experiment designField experiments were conducted at the Guanshan Farm of Qianjiang (30°10′N,112°42′E)during the 2017 rice growing season (June-October).This experiment sought to explore whether the amount of nitrogen fertilizer can be reduced in RC fields and then propose a reasonable amount of nitrogen fertilizer.A split-plot experimental design was used with differing planting models in the main plots and a different nitrogen gradient in the sub-plots.The main plots include those having been breeding crayfish in paddy for 2 and 8 years,respectively,and traditional rice-wheat paddy fields as a control.The nitrogen concentration gradient set in the sub-plot was:0 kg N ha-1(0 N),75 kg N ha-1(75 N),150 kg N ha-1(150 N) and 225 kg N ha-1(225 N).Each main plot included three identical rice field models belonging to different farmers,and the nitrogen fertilizer gradient test was repeated three times in each paddy field.Nine cooperator-farmers were involved in the experiment,and the nitrogen gradient test was conducted in each paddy field during the rice-growing season.The soil is sticky soil,with a pH value ranging from 6.2 to 7.5.The total soil organic matter contents ranged from 26.34 to 31.60 g kg-1.Fig.1-A and B show the temperature and precipitation of a rice-growing season in the study area.

Fig.1 Temperature and rainfall during the rice-growing season(June-October) in 2016 (A) and 2017 (B) and comparison of rice yields in 2016 and 2017 (C).Min temperature and Max temperature represent the minimum and maximum temperature values of the day,respectively.

Management practicesThe plot area was 5 m×6 m,and plots were segregated by a 1-m wide alley which was wrapped with plastic film inserted into the soil to a depth of 0.50 m to form a barrier.The rice variety tested was Taiyou 390.Phosphorus (75 kg ha-1as single superphosphate) and potassium (150 kg ha-1as KCl)were added as base fertilizers prior to transplanting.Nitrogen fertilization (0,163,326 and 489 kg ha-1as urea with 46% nitrogen) was applied in a ratio of 5:5,at the pre-transplanting and early tillering stage,respectively.Two rice seedlings were transplanted per hill with consistent irrigation,insecticide spraying,and other field operations.

Sampling and measurementsRice yields were measured by following the procedure described by Yoshidaet al.(1972).To avoid border effects,plants in two rows on each side of the plot were abandoned.Grain yield was measured from a harvest area of 3.0 m2(in each plot,three 1 m2areas were randomly chosen)and adjusted to 14% moisture.

Soil samples were collected from the surface soil (0-20 cm) in each plot at the pre-transplanting,tillering,heading,and harvest stage,respectively.Five soil cores from each plot were mixed to obtain one composite sample and then passed through a sieve(2 mm diameter).One portion was air-dried and passed through a 100-mesh sieve to analyze the total nitrogen while the other portion was immediately stored at 4°C (for less than one week) for ammonium nitrogen analysis.All visible roots and fresh litter material were removed from the soil samples.The total nitrogen content was determined using the micro-Kjeldahl approach (Page 1982).Ammonium-N and nitrate-N were extracted with a 2 mol L-1KCl aqueous solution and then analyzed by indophenol blue and phenol disulfonic acid colorimetry,respectively (Xuet al.2010).

2.4.Statistical analysis

Rice yield in rice-crayfish fields was calculated as:Rice yield=Total rice yield from rice planting area/(Rice planting area+Ditch area).To compare rice yields between RC and RW,the yield change (%) in RCvs.RW was calculated as:Rice yield change (%)=(Yield in RC-Yield in RW)/Yield in RW×100% (Huet al.2016).OriginPro 8.0 was used to prepare all graphs.Multiple linear regression analyses were performed in IBM SPSS Statistics 15.0 Software.One-way analysis of variance (ANOVA) atα=0.05 was conducted to test the significance of differences across the survey and field experiment data.

3.Results and discussion

3.1.Basic information of farmers in different farming models

Farmers adopting rice-crayfish (RC) and crayfish monoculture (CM) had a much higher education level compared with rice-wheat (RW) farmers (Fig.2-A).Junior high education accounted for 66.67,60.00,and 16.70% in RC,CM,and RW,respectively,while high school education accounted for 7.07,20.00%,and 0,respectively.Compared with the traditional RW model,aquaculture in paddy fields requires a high level of knowledge.Other studies have reported that education enables farmers to better manage their farms by improving the decision-making abilities (Asadullah and Rahman 2009),and also improves their access to information which could allow them to pay and charge better prices for their materials bought and products sold,respectively (Lockheedet al.1979).

The mean age of farm owners were 49,47,and 58 in RC,CM,and RW,respectively (Fig.2-B).Previous research has shown that farmer age can influence their knowledge in soils (Barbero-Sierraet al.2016)and pesticides (Ibitayo 2006).Older farmers are in general less efficient since they are less adaptable to new technological developments (Seyoum 1998).The main occupation of the farmers was rice cultivation,but many had part-time jobs,such as animal husbandry,fish farming,and gardening.Most farmers had extensive experience in rice cultivation.RC farming,on the other hand,had only been implemented for an average of 5.9 years.The smallest RC area was 0.3 ha,and the largest was 8.93 ha (Appendix B).The areal distribution of RC varied across geographical environment,with the area mainly being 2-3 ha in plain lake areas,while under 1 ha in hilly areas (Appendix B).

Fig.2 Selected characteristics of rice-crayfish (RC),and crayfish monoculture (CM),and rice-wheat (RW) farmers.A,education percentage.B,farmers’ age.

RC requires more knowledge and labor inputs than RW due to its complex system combining farming with aquaculture.Therefore,the RC farming model requires knowledgeable and capable agricultural producers to be successfully implemented.

3.2.Management features in different farming models

The results of survey research showed that the average amount of nitrogen fertilizer applied was 153 and 340 kg ha-1in RC and RW models,respectively,and the median nitrogen dosage was 146 kg ha-1in RC and 331 kg ha-1in RW (Fig.3-A).The high-yield farmers who grew wheat used more fertilizer than those applying the RC model.Most farmers believed that crayfish feed residue can reduce the application of nitrogen fertilizer in fertile fields.It has been shown that in rice-turtle co-cultures,some of the unused N and P feed by turtles was absorbed by rice plants (Zhang Jet al.2016).The average amount of feed was 2 273 and 3 586 kg ha-1in RC and CM,according to 5% nitrogen content,the feed nitrogen amount is 114 and 179 kg ha-1in RC and CM,respectively (Fig.3-B).It was discovered during the survey that feed was applied in April and May each year to harvest crayfish by RC farmers.However,in the CM model,in addition to the feed in April and May,farmers also placed large amounts of feed in June and July to increase crayfish yield.

In rice-fish farming systems,aquaculture trenches occupy part of the rice fields,which may lead to a decline in rice production (Vromantet al.2002).In the 2016 survey,it was found that the areal distribution of the RC system was different across geographical environments(Appendix B).In plain lake areas,due to flat terrain and abundant water resources,most RC areas were 2-3 ha,while the average was 0.5 ha in hilly regions (Appendix B).Thus,the effect of ditch proportion on rice yield in different geographical environments was analyzed.In paddy fields (2 ha),the ditch proportion was 15% which would not decrease rice yield (Fig.3-C).The optimal ditch proportion was 10% for a paddy area of 2-8 ha(Fig.3-D).Wuet al.(2012) suggested that the ditch and pond model,which supplies shelter for fish during lowirrigation periods,may limit rice yield due to reduced ricestand area.However,rice planting had a positive edge effect,which would fortify the effects from increased edgezones and thus compensate for yield loss in smaller ricestands (Du and Wang 1998).In the crop community,the edge line crop condition was superior to the condition of inner crops (Hadjichristodoulou 1993).This was because compared with the inner group the edge group has less mutual shading,more sufficient light,better ventilation,larger root extension range,stronger root systems,and therefore a higher level of absorption and more advanced individual development.Songet al.(1995) showed that marginal winter wheat areas demonstrated numerous advantages to the amount of light,temperature,water,and wind,as represented by more ears per ha and more grains per ear than middle areas.Therefore,the perimeter/area ratios simultaneously decreased as field area increased,which decreased the edge effect(Laurance 2008).For example,Kiviniemi and Eriksson(2002) reported that the richness of plant species in the grasslands of Sweden increased near the edge of small fragments but decreased in large fragments.To avoid affecting the yield of rice-crayfish fields,the appropriate ratio of aquaculture ditch was selected according to different paddy areas.Since the edge effect decreases as the paddy field area increases,the ditch proportion should not exceed 15% in paddy fields ≤2 ha,and was no more than 10% in those of 2-8 ha.

Fig.3 Field management and field engineering in different farming models.A,annual amount of nitrogen fertilizer per farmer household in the rice-crayfish and rice-wheat farming models.B,the annual amount of feed applied per farmer household in the rice-crayfish and crayfish farming models.C and D,the relationship between changes in rice yield (%) and ditch proportion (%)in rice-crayfish fields with a total area of ≤2 ha and 2-8 ha,respectively.

3.3.The cost-benefit ratio in different farming models

According to the survey study,the net income of farmers who bred crayfish was much higher than those who only planted wheat and rice.Based on the different inputs and outputs in wheat,rice,and crayfish farming,it can be summarized that rice-crayfish farmers had the highest net income,while rice-wheat farmers had the lowest (Table 1).Net income for wheat and rice in RW was around 1 875 and 15 825 CNY ha-1,respectively,with a cost-benefit ratio of 1:2.0 (Table 1;Fig.4-A).In RC,costs included rice payment of 10 875 CNY ha-1,plus crayfish breeding costs of 18 195 CNY ha-1,totaling 29 070 CNY ha-1.The benefits of RC included the value of the reduced cost for fertilizer.The RC benefits were about 3.5 times its costs(Fig.4-A).The CM costs were 31 590 CNY ha-1and its net income was 66 360 CNY ha-1.Overall,CM benefitcost ratio is 3.1 (Fig.4-A).Aquaculture is characterized by high input and high output costs.The production costs for RC and CM farmers were higher than those for RW farmers due to expendtiure on crayfish seeds,crayfish feed,labor,and other necessary items for raising crayfish,such as utilities,crayfish tools,field engineering renovation and escape prevention facilities.

Table 1 Cost and benefit (CNY ha-1) in rice-wheat,crayfish monoculture,and rice-crayfish farming models in 2016 and ricecrayfish model in 2017

Low grain price is the main factor that leads to the low cost-benefit ratio of the RW system.Grain prices average 2 CNY kg-1,while that of crayfish is around 35-55 CNY kg-1(MARABFet al.2018).Compared with the CM system,RC made more efficient use of its resources and improved the cost-benefit ratio.

Rice is a thermophilic plant (Zsoldos and Karvaly 1978) that grows in summer and autumn,especially the middle-season rice.Crayfish become active,prey,and grow at a water temperature of 15°C.From early April to the end of May,the average water temperature is between 20 and 25°C,which encourages the rapid growth of crayfish (Huner 2002).As temperatures begin to rise in June,crayfish begin to dig,burrow,mate,and breed,and their reproduction peaks in September and October which leads to poor growth (Arce and Diéguez-Uribeondo 2015).Research has shown that high water temperatures,such as those often observed in rice fields in the summer,accelerated crayfish development,which led to many crayfish maturing at a size,below the market standard(Alcorloet al.2008).This wasted a large amount of human and material resources,making the cost-benefit ratio of CM significantly lower than that of RC (Fig.4-A).In the intensive CM system,farmers apply large amounts of feed in June,July,and even in August,September,and October to increase crayfish yield.Crayfish yield was much lower in other months compared with that in April and May,due to the influence of temperature and reproduction (Fig.4-B).

Fig.4 Cost-benefit ratio (C:B) (A) and the yield of wheat,rice and crayfish in the rice-wheat,rice-crayfish and crayfish monoculture farming models in 2016 (B).Bars are SD.

3.4.Analysis of the factors influencing the economic benefits of the rice-crayfish model

We conducted a regression analysis for the RC model,investigating how net income,the dependent variable,is influenced by a range of independent variables,including equipment,fertilizer,pesticide,rice seed,crayfish seed,labors,and others,while controlling age,education,and farm size (Table 2).The multiple linear regression model is:Net income=90 478.4-0.077×Equipment-0.175×Fertilizer-0.288×Pesticide+0.04×Rice seed-0.017×Crayfish seed+0.058×Others.These seven independent variables (equipment,fertilizers,pesticides,rice seeds,crayfish seeds,labor,and others) accounted for 75% of the variations in net income of RC.The multiple linear regression model had a good fit and reliable prediction results (P＜0.01).It is shown that after controlling for the effects of farmers’ age,education level,and farming size,pesticide and fertilizer input significantly negatively affected RC farmers’ net income (P＜0.05).Among the control variables,farm size significantly affected net income (P＜0.05).Studies have shown that the integrated planting and breeding model of rice fields reduced the occurrence of farmland pests and diseases due to animal activities,and then reduced the use of pesticides (Xieet al.2011;Huet al.2016).Studies have shown that the amount of fertilizer used per unit area dropped sharply as the scale of farms increased,and large-scale farms had higher crop yields than their small-scale counterparts (Juet al.2016).The high labor cost and low level of machinery on small farms hinder the application of precise fertilization technology and management based on scientific knowledge.Wuet al.(2018) found that for every 1% increase in farm size,the use of fertilizers and pesticides per hectare would be significantly reduced by 0.3 and 0.5%,respectively,and agricultural labor productivity would increase by almost 1%without significantly reducing crop yields.Our survey results showed that the farm size of the RC model in the plain lake area was mainly 2-3 ha (Appendix B).This has effectively promoted the transformation of small-scale farms to large-scale farms in China,and effectively controlled the use of chemical fertilizers and pesticides.

Table 2 Multiple regression analysis of influencing factors of control variables in the rice-crayfish model1)

3.5.Protecting farmers’ enthusiasm for grain production by the rice-crayfish model

In 2017,heavy rainfall occurred in large areas of Qianjiang before rice harvest.According to the meteorological data,176.1 mm of rainfall occurred from September 18 to October 6,2017 (Fig.1-B).The continuous rain during this period significantly reduced rice yield in Qianjiang and increased rice lodging.The cumulative lodging area was 21 500 ha,accounting for 58.9% of the rice-planting area.It was estimated that lodging reduced rice yield by 66 600 t,which corresponded to a loss of 409 million CNY.Low rice price reduced revenue by 210 million CNY,harvest cost increased by 26 million CNY,and the average reduction in net income was 8 940 CNY ha-1(Suet al.2018).Most of the RW farmers in 2016 converted to RC in 2017,but the survey only tracked the RC farmers who did not convert.In 2017,most of the RC farmers reported lodging,and the average rice yield decreased by 19%(Fig.1-C;Appendix C).Reduced yield,rising cost of harvesting,and low grain prices combined to directly decrease the economic output of rice to 870 CNY ha-1(Table 1).Despite this,most farmers stated that they would continue to raise crayfish in paddy fields next year,mainly because the income from crayfish farming in 2017 was high enough to make up for the economic loss of rice (Table 1;Appendix D).Consequently,RC can protect farmers’ enthusiasm for grain production,and more importantly,contribute to food security.In recent decades,farmers have lost interest in farming because of low agricultural income and the shift of high-quality rural labor to cities (Song and Pijanowski 2014;Zhang Yet al.2016).Widespread unused land in rural areas seriously threatens national food security (Chenet al.2009;Tian 2012).The high benefit-cost ratio of the rice-crayfish model has protected farmers’ enthusiasm to grow grain and has injected new vitality into the development of Chinese agriculture.

3.6.Effect of nitrogen fertilizer on rice yield in ricecrayfish fields

According to the 2016 survey,farmers began to realize that residual feed and crayfish waste can be used as part of the fertilizer to reduce the use of nitrogen.However,some farmers continued to use a high amout of input.Therefore,a nitrogen fertilizer gradient test was applied in Guanshan on a 2-year-old rice-crayfish (RC2) field and an 8-year-old rice-crayfish (RC8) field,using a local traditional RW field as a control.The study suggested that the highest rice yields were 225 N,150 N and 75 N in RW,RC2,and RC8,respectively (Fig.5-A).Other studies have shown that over-application of nitrogen fertilizer may actually reduce grain yield by increasing the susceptibility to lodging (Phamet al.2004).Total nitrogen in soil tended to increase in RW,RC2,and RC8,and compared with RW,RC8 showed a significant increase(P＜0.05).During the rice-growing season,the ammonia nitrogen content in soil in RW was significantly lower than that in RC2 and RC8 (Appendix E;P＜0.05),possibly because of the effects of feed residues and crayfish feces on soil nitrogen.A study by Xieet al.(2011) showed the uncontaminated nitrogen in fish feed reduced fertilizer use to increase rice yields in rice-fish co-cultures (RF).Their data showed that 32% of the nitrogen in rice and straw came from fish feed in RF.Neto and Ostrensky (2015)estimated that 18% of animal feed was not consumed and was lost in the aquatic environment and accumulated in the sludge.In the RC model,the nutrients in residual feed and feces could be absorbed and utilized by rice to reduce its demand for nitrogen fertilizer.

Fig.5 Rice yields in different nitrogen fertilizer gradients (A) and soil total nitrogen content (B) in rice-wheat (RW),2-year-old rice-crayfish (RC2) and 8-year-old rice-crayfish (RC8) fields.0 N,75 N,150 N and 225 N indicate nitrogen application rates of 0,75,150 and 225 kg N ha-1,respectively.Bars are SD.

According to a report by the IFIA (2011),nitrogen application averaged 305 kg N ha-1yr-1in China compared with the worldwide average of 74 kg N ha-1yr-1.Excessive nitrogen application to rice (Oryza sativaL.) crops in China has increased rice planting costs,caused environmental pollution,reduced grain yield,and contributed to global warming (Penget al.2011).In the production process of aquatic products,feed is the main source of waste and the main cause of the environmental impacts of aquaculture (d’Orbcastelaet al.2009).The natural combination of aquaculture and rice planting can improve energy efficiency,and also reduce environmental pollution.Since used feed and excrement reduce the subsequent fertilizer application,the results of this study suggest that reducing the use of nitrogen fertilizer in RC fields can increase rice yield and achieve a win-win model for agricultural production.

3.7.Rice-crayfish model and food security

In the RC model,farming ditches in the field occupy a part of the rice production area,which may reduce rice production area and threaten China’s food security.However,previous research has shown that aquaculture ditches accounting for less than 10% of the entire paddy field area,which did not decrease the overall rice cultivation and cultivation per unit area of rice (Huet al.2016).

Compared with the rice-wheat,double-cropping rice,and other rotation models,the RC model reduces one season’s crop plating,which may seriously threaten food safety.However,in recent years,the economic efficiency of traditional planting models has proven to be low,which has caused massive labour flow from rural to urban areas and abandonment of large area of arable land (Li 2008).The abandonment rate in the Jianghan Plain was as high as 20-30%from 2001 to 2009,but it decreased to 5% from 2010 to 2017 (Zhanget al.2019).This is mainly because the economic benefits of the RC model adopted in recent years are significantly higher than that of the traditional rice farming methods,which attracts a large number of young laborers to invest in rice production.Our research showed that the average age of RC farmers was lower than that of traditional RW farmers(Fig.2).Even if there is a rice lodging accident,all farmers we surveyed expressed their willingness to develop the RC model next year (Appendix D).The reduction of abandoned land increases the area’s food production,and young labor can promote the application of new agricultural technologies and knowledge.

The RC model originated from the waterlogged land in the Jianghan Plain.Most of the waterlogged land can only grow one season of rice with low yield.Studies have shown that long-term ditching and drainage in cold waterlogged fields promoted the mineralization of soil organic matter,enhanced soil microbial activity,improved soil physical and chemical properties,and subsequently increased yield (Linet al.2016).Therefore,the RC model can improve the soil quality of waterlogged land through field engineering to increase food production.In summary,scientifically developing the RC model is not only unlikely to threaten food security,but also would protect food security.

4.Conclusion

Using Qianjiang City in Hubei as an example,we compared the labor force quality in different farming models (RW,RC and CM),the management characteristics of farmers,and the benefit-cost ratio.The results suggested that the RC model attracted high-quality labor to production due to its high benefit-cost ratio.This helped the exchange and application of technology and information for advanced management practices in the RC model,which increased farming stability of paddy fields.In the lake zone of Jianghan Plain,the RC model effectively utilized local light,temperature,and water resources,and maintained farmers’ enthusiasm for grain production when paddy fields experienced a natural disaster (flooding).The nitrogen compensation effect of residual feed and feces in the RC model suggested that farmers reduce the amount of nitrogen fertilizer application for subsequent rice crops.

Acknowledgements

This research was funded by the National Key Research and Development Program of China (2017YFD0301400)and the Fundamental Research Funds for the Central Universities,China (266620202KPY014).

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on http://www.ChinaAgriSci.com/V2/En/appendix.htm