Occurrence pattern and morphological polymorphism of Chinese weedy rice

WANG Hao-quan,DAl Wei-min,ZHANG Zi-xu,Ll Meng-shuo,MENG Ling-chao,ZHANG Zheng,LU Huan,SONG Xiao-ling,QlANG Sheng

Weed Research Laboratory,College of Life Sciences,Nanjing Agricultural University,Nanjing 210095,P.R.China

Abstract Rice,the main food crop in China,has been sporadically reported to suffer from weedy rice infestation.However,the overall occurrence and distribution pattern of Chinese weedy rice remains unclear because a systematic survey has not been conducted.In order to reveal the infestation of Chinese weedy rice,a field survey was conducted in 999 sampling sites all over the rice-growing regions in China from 2009 to 2016 using seven-scale visual scoring of the level of weed infestation.Weedy rice was found 39% occurrence incidence in a total of 387 sites.The sampling sites with 50% or higher overall weedy rice infestation index mainly radiated from Jiangsu,Heilongjiang,Ningxia and Guangdong to the whole East China,Northeast China,Northwest China and South China.A total of 45 morphological characters from 287 populations (collected simultaneously with the field survey) out of those occurred sites were observed and analyzed using multivariate analysis in common gardens with the same cultivation conditions in 2017 and 2019.Canonical correlation analysis showed that 45 morphological characters were significantly related to the latitude,mean temperature,minimum temperature,precipitation and mean diurnal range factors.The 287 weedy rice populations were divided into three morphological groups with climate-dependent geographical differentiation: strong tiller type only in Jiangsu,large leaf type in South China and Central China and large grain type mainly in North China.Weedy rice seriously infested rice fields and had a geography,climate and cultivated rice type-dependent morphological and biotype differentiation in China.It is suggested to pay attention to the harmfulness of weedy rice and adopt comprehensive control strategies.

Keywords: weedy rice,infestation,morphological characters,diversity

1.lntroduction

Weedy rice (Oryzaspp.),a weed of the same species as cultivated rice,is a complex of weedy populations of the genusOryzain paddy fields (Deloucheet al.2007).With the wide adoption of simplified rice cultivation techniques(especially direct seeding rice),the spread of weedy rice is increasing (Chauhan 2013;Ziskaet al.2015).Until now,weedy rice has been found in almost all rice-growing areas worldwide (Zhanget al.2014;Qiet al.2015;Kraehmeret al.2016).It has become the third most important weed in paddy fields afterEchinochloa crus-galli(L.) Beauv.andLeptochloa chinensis(L.) Nees (Deloucheet al.2007).Weedy rice is associated with cultivated rice,thus competing for resources during the growing season(Kraehmeret al.2016;Qiang 2018).It can significantly decrease the yield and quality of cultivated rice (Shivrainet al.2009) or even lead to no harvest (Diarraet al.1985).As one of the global main rice-producing areas,how China suffers from weedy rice remains unknown because of the lack of systematic and detailed research on overall weedy rice’s occurrence and economic impact (Lianget al.2009).

Weedy rice is similar to cultivated rice in morphology but has high weediness.Its weediness is closely related to biological phenotypic characteristics,mainly reflected in red pericarp,strong competitiveness,early maturity,seed shattering and dormancy (Ziskaet al.2015;Chauhan 2021).In most cases,the pericarp of weedy rice is red,so it is also called red rice in some areas (Grosset al.2010;Li X Yet al.2014).However,weedy rice with white pericarp has occasionally been reported in China,the United States,Southeast Asia,Japan,South Korea,and Costa Rica (Ziskaet al.2015;Cuiet al.2016).The red pericarp of weedy rice is caused by the deposition of proanthocyanidins,which is conferred by the combination ofRcandRdgenes responsible for pigment deposition in rice.TheRcgene regulates the spatiotemporal expression of proanthocyanidin biosynthesis structural genes,whereas theRdgene encodes enzymes involved in the proanthocyanidin biosynthesis pathway (Sweeneyet al.2006;Furukawaet al.2007).The mutation in theRcgene inhibits flavonoid biosynthesis,preventing proanthocyanidin synthesis from depositing in the pericarp,resulting in the white pericarp of some weedy rice biotypes (Sweeneyet al.2007;Wang 2017).

Weedy rice has a higher competitive advantage than cultivated rice (Burgoset al.2006;Ahmedet al.2012;Daiet al.2014a;Olajumokeet al.2016).Its strong competitiveness is due to the rapid seed germination,rapid growth at the seedling stage,a stronger root system,efficient nutrient utilization (such as N),and a higher biomass potential (by increasing plant height or producing a large number of tillers) (Chauhan and Johnson 2010;Ziskaet al.2015;Olajumokeet al.2016;Wedgeret al.2019;Svizzero 2021).During germination,the endogenous hormone indoleacetic acid accumulates significantly in the mesocotyl of weedy rice,promoting cell and mesocotyl elongation (Li Let al.2014).Additionally,it also has a higher photosynthetic pigment content,a more active reaction center and a higher net photosynthetic rate,all of which lead to enhance overall photosynthetic activity (Daiet al.2017).As a result,it germinates and grows rapidly at the seedling stage,accumulating competitive advantages over time.Due to differences in morphological characteristics,weedy rice biotypes from different regions vary in their effects on cultivated rice.Weedy rice in Northeast China has the most harmful impact on the growth and yield of cultivated rice in the early stage of vegetative growth;while weedy rice in Southeast China is relatively tall,and its competitive effect is more detrimental at later crop-growth stages (Daiet al.2014a).

Early maturity,seed shattering,and dormancy all contribute to weedy rice becoming a weed in paddy fields,as the first two allow weedy rice to escape harvest and enter the soil seed bank,while dormancy allows seeds in the seed bank to live longer,allowing for the persistent infestation in farmland.Generally,weedy rice has a shorter growth period,matures earlier and shatters more easily than cultivated rice (Daiet al.2014a;Nuneset al.2014;Zhaoet al.2018).Both early flowering and a shorter grain-filling stage contribute to the early maturity(shorter growth period) of weedy rice.In comparison to the flowering strategy,weedy rice grain filling is less plasticity but more conservative,and it is significantly shorter than that of cultivated rice (Zhaoet al.2018).During grain filling,weedy rice grains develop the endosperm more rapidly and have a reduced capacity to scavenge reactive oxygen species from endosperm cells,resulting in early maturity (Zhaoet al.2020,2021).Seed shattering occurs as the abscission layer between the grain and the pedicel deteriorates (Liet al.2006).It appears to be a convergent trait in weedy rice,which evolved convergently in different regions (Ziskaet al.2015).While the phenotypic characteristics of seed shattering are convergent,the genetic characteristics that result in seed shattering are not,which means multiple different evolutionary pathways may lead to similar traits (Thurberet al.2013).Seed shattering(SH) is a complex quantitative trait that is influenced by multiple alleles.Although numerous quantitative trait loci(QTLs) associated with shattering have been identified,including SH4,qSH1,Sh-h,and SHAT1,the major QTLs in different populations may also be distinct (Yaoet al.2015).Instead of re-evolving shattering through restoring functional alleles from their ancestors (for example,T substitutions in the SH4 allele exons),many weedy rice populations have re-evolved shattering through alternative genetic mechanisms (Thurberet al.2010,2011,2013;Zhuet al.2012).Dormancy,like seed shattering,is also regulated by multiple alleles.At the moment,at least 10 interacting QTLs have been identified concerning weedy rice dormancy (Subudhiet al.2012;Zhanget al.2017;Nguyenet al.2019).Numerous (but not all) of these dormancy-related QTLs in weedy rice were co-localized with genes associated with the weedy phenotype,including shattering,pericarp color,and flowering time (Nadiret al.2017).Thus,multiple morphological characteristics contribute to the weediness of weedy rice.Studying morphological traits of different weedy rice populations will help us better understand their weediness and control them.

Weedy rice is adaptable to a wide range of abiotic and biotic stresses and can rapidly develops herbicide resistance (Nadiret al.2017;Jia and Gealy 2018;Fogliattoet al.2020).Recent studies have reported that weedy rice is highly adaptable to temperature changes(cold and high temperature),intermittent water supply(drought),anoxic germination,soil salinity,climate change(such as elevated CO2concentrations) (Fogliattoet al.2020).Moreover,some weedy rice populations exhibit broad-spectrum resistance to rice blast and sheath blight(Jia and Gealy 2018).Due to the competitive advantages conferred by its morphological and physiological characteristics,weedy rice is adaptable to different cropping systems and changing environmental conditions(Nadiret al.2017;Jia and Gealy 2018;Fogliattoet al.2020).Because the sensitivity to herbicides of weedy rice and cultivated rice are very similar,no safe and effective chemical herbicide can specifically control weedy rice in paddy fields except if herbicide-resistant cultivars are used so far (Chauhan 2013;Busiet al.2017).However,herbicide-resistant rice has the risk of resistance genes drifting to weedy rice due to bidirectional gene flow between weedy rice and cultivated rice (Luet al.2016;Merottoet al.2016;Zhanget al.2018).The frequency of gene flow between cultivated and weedy rice varies considerably between studies only at a 0–2.3% rate(Sanchez Olguinet al.2009;Luet al.2016).However,when gene flow occurs on a constant annual basis,it appears inevitable that one or more genes (particularly resistance genes) will eventually penetrate weedy rice(Merottoet al.2016;Daueret al.2018).Introgression of foreign genes into the weedy rice gene bank may increase its genetic diversity and confer some trait modifications with or without adaptive value,resulting in unpredictable biosafety problems (Heet al.2014;Luet al.2016).Additionally,the reverse gene drift from weedy rice to crops may produce hybrid progeny with weedy characteristics such as early maturity,shattering,and dormancy.These more adaptable hybrid progenies are likely to gradually “wild” to form new weedy rice types (Zhanget al.2018;Svizzero 2021).In short,the morphological,physiological,and genetic advantages of weedy rice confer it a high capacity for environmental adaptation and rapid evolution,which contributes to its difficulty in control (Chauhan 2013,2021).

Weedy rice has rich morphological variability and phenotypic plasticity (Schwankeet al.2008;Andreset al.2015),making its identification and management quite difficult.According to the differences of hull color and awn,weedy rice in the United States was divided into two main biotypes: straw hull awnless and black hull awned (Noldinet al.1999;Shivrainet al.2010).Their hybridization may provide opportunities for the emergence of new weedy rice biotypes (Londo and Schaal 2007).Based on morpho-physiology,isozymes and random amplified polymorphic DNA (RAPD) markers,152 biotypes of weedy rice collected from Asia,South America and North America were classified into four types:indicatype similar to cultivars,indica-type similar to common wild rice,japonica-type similar to cultivars,japonica-type similar to wild rice (Suhet al.1997).Several studies deal with the morphological diversity and classification of Chinese weedy rice (Menget al.2005;Maet al.2011;Chen 2014;Daiet al.2014b).Weedy rice in Jiangsu was classified asjaponica,tallindicaand dwarfindicatypes based on morphological characteristics such as plant height,tillers,and panicle traits (Daiet al.2014b).According to the length and color of the awn,weedy rice in Northeast China was divided into long red awn,short red awn,long white awn,short white awn,and awnless types.According to the color of the hull,weedy rice could be also divided into three types: yellow hull,yellowbrown hull,and purple hull (Menget al.2005).According to plant height and flowering period,89 Chinese weedy rice populations were divided into four biotypes: Group I composed of tall plants that flower early;Group II with tall plants flowering late;Group III with short plants flowering early;and Group IV with short plants flowering late (Chen 2014).However,the overall morphological classification of Chinese weedy rice remains incomplete.

The diversity of weedy rice types is related to their origin (Zhanget al.2015;Sudiantoet al.2016;Yang 2019).Historically,four models for the origins of weedy rice have been proposed (Chauhan 2021;Roma-Burgoset al.2021): (1) feralization of modern or landraces cultivated rice varieties into the weedy form by de-domestication;(2) emergence from intervarietal hybridization within domesticated rice;(3) emergence from crop-wildOryzaspecies hybrids;(4) direct descendants of wildOryzaspecies that have invaded agricultural habitats.According to the first two hypotheses,cultivated rice is the ancestor of weedy rice,whereas the latter two hypotheses both imply that wildOryzaplayed a role in the origin and evolution of weedy rice.The four hypotheses are not mutually exclusive,and recent research indicates that the origin of weedy rice may be multidirectional(Qiuet al.2014,2020;Huanget al.2017;De Leonet al.2019;Vigueiraet al.2019;Hoyoset al.2020).Recent advances in high-throughput sequencing technology have resulted in an increasing number of published studies supporting the first three hypotheses,but very few directly supporting the fourth (Qiuet al.2014,2020;Ziskaet al.2015;Vigueiraet al.2019).The ancestral cultivated rice may have been de-domesticated by hybridization (Reagonet al.2010),mutation (Caoet al.2006;Huanget al.2018)and other ways to originate weedy rice (Kane and Baack 2007).The de-domesticated weedy rice may have a coevolutionary relationship with the localjaponicaorindicacultivated rice (Kanapeckaset al.2016;Qiuet al.2017;Wuet al.2020).Chinese weedy rice has abundant genetic diversity,which is related to its geographical location and coexisting rice varieties (Huanget al.2017;Hanet al.2020).

Numerous investigations and studies on the occurrence and distribution of Chinese weedy rice have been conducted previously;however,these studies were aimed at local areas,such as Northeast China,Guangdong,and Jiangsu (Caoet al.2006;Zhanget al.2008;Liet al.2013).The overall infestation and distribution pattern of Chinese weedy rice is not clear,and the research on the diversity and classification of Chinese weedy rice based on morphological characters of the whole growth period is still incomplete.In the present study,weedy rice in 404 sampling sites was systematically investigated in China,and 287 weedy rice populations were collected.Forty-five morphological traits (IRRI 2013) of these 287 weedy rice populations were measured under common garden conditions in Nanjing,China.This study aimed to reveal the occurrence,morphological polymorphism,and distribution pattern of Chinese weedy rice.The combined descriptions of occurrence,distribution patterns and key morphological characteristics will aid in the identification of weedy rice groups and the development of appropriate strategies to prevent weedy rice spread and mitigate its damage,including periodic manual weeding at the appropriate time (easier to identify or operate due to morphological differences) to reduce seed bank buildup in the fields.

2.Materials and methods

2.1.lnvestigation on the occurrence of Chinese weedy rice

Weedy rice in 999 sampling sites (Appendix A) across China rice-growing regions (Meiet al.1988) was surveyed from 2009 to 2016 through visual scoring of weed infestation level on a seven-class scale during the heading stage of rice (Qiang 2005).The rice establishment methods of these sampling sites had both direct-directed rice and transplanted rice (their proportion in each province or autonomous region can be seen in Appendix B).The number of sampling sites in each province was decided based on their rice-planting area (set at least one sampling site every 100 000 ha),with the sites selected randomly selected in each city in a province.In the areas where weedy rice frequently occurred in previous reports (Lianget al.2009),the number of samples increased accordingly.Approximately 10 uniform units (10 fields,the area of each survey field was at least 667 m2) were sampled as only one sampling site,and their levels of infestation of weedy rice were scored visually on a seven-class scale(Appendix C) (Qiang 2005).According to the standards,the quantitative characteristics (coverage in the field area,relative height,abundance) of weedy rice at each sampling site were obtained.The overall weed infestation index (OWII) was calculated according to the formula:whereais the infestation level value,bis the number of fields corresponding to this infestation level,cis the number of quadrats per sampling site (which was 10),and 5 is the value of the highest infestation level (Qiang 2005).

2.2.Plant materials and growth conditions

Weedy rice populations were collected concurrently from each sampling site with weedy rice occurrence during the investigation.Among them,287 weedy rice populations(Appendix D) from 24 regions of China (including provinces,autonomous regions and municipalities) that can normally germinate were used for field experiments in 2017 and 2019.The experiments were conducted in the common gardens with the same cultivation conditions at Jiangpu Farm and Baima Experimental Base of Nanjing Agricultural University from May to November.These weedy rice possessed at least two or more of the following characteristics: dark hull,red pericarp,early maturity,and easy seed shattering.Standard cropping practices were applied to plant the materials (Ding 1961).Each weedy rice population was planted on a separate 3 m2plot in five rows (30 cm apart) at a distance of 20 cm between individuals.Each row was sown with one line with 10 individuals.Plots were separated from each other by 50 cm.Weedy rice populations were randomly assigned to plots;three seeds were planted per hole and thinned to one plant at 20 days after planting.Fertilization included the application of 225 kg ha–1N (75 kg ha–1each at planting,tillering and anthesis),112 kg ha–1P (at planting) and 56 kg ha–1K twice (at planting and tillering).Weeds were controlled by a single application of penoxsulam (22.5 g a.i.ha–1) plus cyhalofop butyl (112.5 g a.i.ha–1) using a commercial co-formulated mixture (Zhaoet al.2018).

2.3.Morphological characters assessment

Fifteen weedy rice plants per population were randomly selected at 45,60 and 90 days after sowing (DAS).Their morphological characters (plant height,number of tillers,leaf number,leaf length and leaf width) were measured,and the leaf index and leaf area were calculated.The aboveground biomass,yield per plant,effective panicle number per plant,presence or absence of awns,panicle traits and the other traits were determined at maturity(Table 1).

Table 1 Forty-five morphological characters of weedy rice

2.4.Data processing and analysis

Weedy rice infestation and distribution pattern map were drawn using ArcGIS (10.6,Esri Inc.,CA,USA).Geographical-climatic information from 2006 to 2016 for locations where weedy rice was collected was obtained from the China Meteorological Network (http://data.cma.cn/site/index.html).The statistical analyses consisted of one-way analysis of variance (ANOVA);means werecompared using Duncan’s multiple range test (DMRT)(P＜0.05).The Canoco (5.0,Microcomputer Power Inc.,NY,USA) was used to perform canonical correlation analysis (CCA).The statistical package R was used for cluster analysis.Euclidean distance was used to calculate the distance between sampling sites,and the longest distance method was used for clustering.

The coefficient of variation (CV) was calculated by the following formula: CV=Standard deviation (SD)/Mean×100%.When calculating the diversity index,the qualitative traits were assigned values (presence or absence of awns as 1 and 0,yellow or black/brown hull as 1 and 2,red or white pericarp as 1 and 2,respectively),and the quantitative traits were graded.The 287 populations were divided into 10 levels,from the first level (Xi＜mean–2×SD) to the 10th level (Xi＞mean+2×SD),each 0.5×SD was a level,and the relative frequency of each level was used to calculate the diversity index.Shannon diversity index (H′) was calculated using eq.(1).Simpson’s diversity index (D) was calculated using eq.(2):where,Piwas the percentage ofi-level populations,and Ln was the natural logarithm.

3.Results

3.1.Occurrence and infestation pattern of Chinese weedy rice

According to the survey,weedy rice occurred in 387(39%) out of the 999 sampling sites in China (Fig.1).Although the number of sampling sites increased in some areas,their frequency of weedy rice did not change significantly with the number of sampling sites.Among the main rice-producing areas,Ningxia in Northwest China,Heilongjiang and Liaoning in Northeast China,Jiangsu in Central China,and Guangdong in South China were the hardest-hit areas infested by weedy rice,with an average frequency more than 50%.Weedy rice was also widely distributed in Anhui and Jiangxi in Central China,Guangxi and Hainan in South China,Yunnan and Guizhou in Southwest China,Jilin in Northeast China,and their average occurrence rate was 23–45%.Weedy rice infestation was limited in Fujian in South China,Sichuan and Chongqing in Southwest China,Hunan,Hubei and Zhejiang in Central China,with an average frequency less than 15%.

Fig.1 Occurrence and infestation of Chinese weedy rice.

A total of 85 sampling sites were seriously infested by weedy rice with OWII above 50%,representing 22%of the 387 sampling sites with weedy rice infestation.Among these 85 sampling sites,21 had an OWII of more than 75%,which almost led to rice production failure.About 61% of the seriously infested sampling sites were located in the Central China rice-growing region,mostly in Jiangsu.Besides,21% of the seriously infested sampling sites were distributed in Northwest China,mainly in Ningxia.About 24 and 44% of the sampling sites in Jiangsu and Ningxia were seriously infested by weedy rice,respectively.The remaining seriously infested sampling sites were located in South China (7%),Northeast China (7%),Southwest China (3%) and North China (1%) rice-growing regions.

A total of 114 sampling sites were moderately infested by weedy rice with OWII between 25 and 50%,accounting for 29% of the sampling sites with weedy rice infestation.Among these 114 sampling sites,56% were distributed in Central China,mostly in Jiangsu,about 33% of the sampling sites in Jiangsu were moderately infested by weedy rice.Besides,18% were found in Northwest China,mostly in Ningxia,Xinjiang and Gansu.The remaining moderate infestation sampling sites were in South China (11%),Northeast China (8%),Southwest China (4%) and North China (3%) rice-growing regions.

In addition,188 sampling sites were mildly infested by weedy rice,with OWII below 25%,accounting for 49% of the sampling sites with weedy rice infestation.These sampling sites with a relatively low infestation of weedy rice were distributed in almost all rice-growing regions.

3.2.Distribution pattern of Chinese weedy rice

The CCA of 15 geographical-climatic factors (Appendix D) and 45 morphological traits grouped the 287 weedy rice populations into three clusters.The first and second clusters were significantly influenced by precipitation of June,mean temperature of driest quarter,minimum temperature of June,July,August,September,mean temperature of the wettest quarter and warmest quarter,precipitation of the driest month and driest quarter.Most of those sampling sites were located in Central China,South China and Southwest China.The third cluster was determined by latitude,mean diurnal range and temperature annual range factors.Most of those sampling sites came from northeastern,northern and northwestern China (Fig.2).

Fig.2 Canonical correlation analysis of 287 Chinese weedy rice populations.A,the relationships between sampling sites and geographical-climatic factors.The samples are roughly divided into three clusters.The first and second clusters were significantly influenced by precipitation of June,mean temperature of driest quarter,minimum temperature of June to September,mean temperature of the wettest quarter,mean temperature of warmest quarter,precipitation of the driest month,and precipitation of the driest quarter.Those sampling sites were collected from Central China,South China and Southwest China.The third cluster was determined by latitude,mean diurnal range and temperature annual range factors.They all collected from northern rice-growing regions such as Northeast China,North China and Northwest China.B,the relationships between morphological characters and geographical-climatic factors.Alt,altitude (in meters above sea level);Lon,longitude (E);Lat,latitude (N);Tmin 6–9,minimum temperature of June,July,August,and September (°C);Prec6,precipitation of June (mm);Bio2,mean diurnal range (°C);Bio7,temperature annual range (°C);Bio8,mean temperature of the wettest quarter (°C);Bio9,mean temperature of the driest quarter (°C);Bio10,mean temperature of the warmest quarter (°C);Bio14,precipitation of driest month (mm);Bio17,precipitation of driest quarter (mm);NPH,plant height measured from the base of the culm to the tip of the plant in the natural condition (cm);APH,plant height measured from the base of the culm to the tip of the plant in the vertical position (cm);TN,tiller number;LN,leaf number;LL,leaf length (cm);LW,leaf width (cm);LI,leaf index;LA,leaf area (cm2);Bio,biomass (g);YPP,yield per plant (g);EPN,effective panicle number;AWN,awn (1,having long awn;0,lacking or having short awn);PL,panicle length (cm);PBN,primary branch number;SBN,secondary branch number;PBD,primary branch density;BI,branch index;TNPP,total number of seeds per panicle;FNPP,filled seed number per panicle;ENPP,empty seed number per panicle;SD,seed density of panicle;ST,seed setting rate;HGW,100-seed weight (g);GL,seed length (mm);GW,seed width (mm);GTH,seed thickness (mm);GS,seed shape;HC,hull color;PC,pericarp color.

The species-environment correlations with the first and second axes were 0.828 and 0.677,respectively (Table 2).Although the first four axes accounted for only 41% of the variation,the Monte Carlo permutation test showed marked relationships between the environment and the morphological characters.Pearson correlation (Appendix E)showed that the geographical-climatic factors close to the positive axis of the first axis (such as the mean diurnal range and altitude) and the negative axis of the second axis (such as the temperature annual range and latitude)were significantly negatively related to most vegetative traits (plant height,tiller number,leaf number,leaf length,leaf width,leaf area) and yield traits (biomass,yield per plant,effective panicle number,panicle length,branch number,total seed number per panicle,seed density,seed setting rate).However,there was a significant positive correlation between the above geographical-climatic factors and seed traits such as seed width,seed thickness and 100-seed weight.The geographical-climatic factors close to the negative axis of the first axis (such as the minimum temperature of August and July,precipitation of the driest month,mean temperature of the wettest quarter,mean temperature of the driest quarter,mean temperature of the warmest quarter) and the positive axis of the second axis (such as the mean temperature of the driest quarter and precipitation of June) were significantly positively correlated with the above-mentioned vegetative traits and yield traits but negatively associated with the seed traits mentioned above.

3.3.Morphological types of Chinese weedy rice

Most Chinese weedy rice had seeds with red pericarp(95%) and yellow hull (94%),and were awnless (79%).The tiller number and leaf number of weedy rice at 45–90DAS ranged from 18–27 and 64–102,respectively,which were significantly higher than those of cultivated rice by 48–80%.In the early vegetative stage (45 DAS),weedy rice average plant height was 54 cm,which also was significantly higher than that of cultivated rice by 16%;their leaf length,leaf width and leaf area were 32,0.84 and 20.53 cm2,respectively,which were significantly greater than those of cultivated rice by 16,4 and 20%.After ripening,their total number of seeds per panicle(109) and hundred-seed weight (2.32 g) were less than cultivated rice by 18 and 35%.However,their effective panicle number and yield per plant were larger than cultivated rice by 45 and 14%.A single weedy rice plant produced approximately 19–74% more seed than cultivated rice (Table 3).

Table 2 The correlations of the geographical-climatic factors with the first four species axes from the canonical correspondence analysis of the morphological characters

Consistent with CCA,cluster analysis based on morphological characteristics assigned the 287 weedy rice populations in China to three groups (Fig.3);there were significant differences in morphological characters among these three groups (Table 3).

Fig.3 Cluster analysis of three groups of Chinese weedy rice based on 45 morphological traits.

Table 3 Morphological characteristics of three different weedy rice groups in China1)

The first group (six populations) was multi-tiller weedy rice with strong growth (vigor) (Fig.4).Tiller number and leaf number among members of this weedy rice group at 45–90 DAS ranged from 37–97 and 134–386,respectively,which were significantly higher than the other two groups by 0.79–6.33 times.In addition,their leaf width,biomass,yield per plant,effective panicle per plant,and seed density of panicle were significantly higher (0.11–2.92 times) than in the other two groups.The weedy rice in this group was awnless with yellow hulls and red pericarp.They were similar in morphology toindicarice and only distributed in Jiangsu in the Central China rice-growing region (Fig.5).

The second group (196 populations) was large-leaf weedy rice (Fig.4).Most of the weedy rice in this group was awnless (84%),with yellow hulls (98%) and red pericarp (94%).Their leaf length at 60–90 DAS was 40–50 cm,leaf area at 60 DAS was 43.71 cm2,and leaf shape index at 60 DAS was 43.21,all of them were significantly higher than the other two groups by 19–98%.About 85% of this group was similar in morphology toindicarice,mostly distributed in South China and Central China rice-growing regions.Besides,15% of this group was similar in morphology tojaponicarice,primarily distributed in Southwest China and Central China ricegrowing regions,a few were found in North China and Northeast China rice-growing regions (Fig.5).

Fig.4 Morphological characters of three groups of Chinese weedy rice.A,Group 1.B,Group 2.C,Group 3.

Fig.5 Geographical distribution of three groups of Chinese weedy rice.

The third group (85 populations) was large-seed weedy rice with weak growth vigor (Fig.4).Weedy rice in this group had red pericarp,86% had a yellow hulls,and 34% had awns.Their hundred-seed weight was as high as 2.59 g,which was significantly higher than the other groups by 17–23%.In addition,their seed length and seed thickness were also 6–13% higher than the other two groups.Their tiller number and leaf number(45 and 90 DAS) ranged from 12 to 20 and 38 to 53,both significantly lower than the other two groups by 30–86%.In addition,their leaf length,leaf width,leaf area,biomass,yield per plant,number of effective panicles per plant,total seed number per panicle,filled seed number per panicle,seed density of panicle,panicle length,primary branch number,secondary branch number,the density of primary branch were also 13–75% lower than the other groups.A total of 78% of the weedy rice in this group wasjaponica-like type,mainly distributed in Northeast China,Northwest China and northern part of North China rice-growing regions;a few in Southwest China rice-growing region.Besides,22% of the weedy rice in this group wasindica-like type,mainly distributed in Central China rice-growing region and partly in Northeast China rice-growing region.

3.4.Morphological diversity of Chinese weedy rice

Chinese weedy rice had great variability of the 45 morphological characters assessed.The variation coefficient of 10 characters related to plant architecture,yield and panicle morphology was greater than 50%;the variation coefficient of 12 characters related to yield,leaf morphology and panicle morphology ranged from 25 to 50%;the remaining 23 traits related to seed morphology and leaf morphology had a variation coefficient less than 25% (Fig.6).The Shannon-Wiener index (H′) and Simpson’s diversity index (D)of the 45 morphological characters confirmed that Chinese weedy rice is strongly diverse in both vegetative and reproductive traits.H′and D of 20 traits was greater than 2 and 0.84 respectively;22 traits had H′ and D as 1.60 to 1.98,0.73 to 0.84 respectively;the remaining three traits had H′ and D less than 0.51 and 0.33,respectively (Fig.6).In the first group of weedy rice,the empty seed number per panicle had the largest variability and the most diversity,with H′,D and CV as 1.79,0.83 and 72%,respectively.In addition,there were two more traits with H′ greater than 1.50 and D greater than 0.70,and four more traits with CV greater than 25%.In the second group of weedy rice,the diversity of leaf shape index at 90 DAS was the most notorious,with H′ and D as 2.10 and 0.86,respectively;the CV of leaf number at 45 DAS was the largest,which was 53%.In addition,there were 39 more traits with H′greater than 1.50 and D greater than 0.70,and 14 more traits with CV greater than 25%.The third group of weedy rice,was the most diverse for seed thickness and plant height(60 DAS),with H′ and D as 2.00 and 0.85,respectively;the CV of leaf number (69%) at 45 DAS was the largest.In addition,there were 17 more traits with H′ greater than 1.50 and D greater than 0.70,and 18 more traits with CV greater than 25%.Among the three weedy rice groups,the first group had the smallest variation and diversity of morphological traits,the second group had the largest diversity of morphological traits,and the third group had the largest variability(Fig.6).

Fig.6 Morphological variation and diversity of different weedy rice groups.A,the coefficient of variation.B,Shannon-Wiener index.C,Simpson’s diversity index.All,the variation and diversity of all weedy rice populations;Group 1,the variation and diversity of the first group of weedy rice;Group 2,the variation and diversity of the second group of weedy rice;Group 3,the variation and diversity of the third group of weedy rice.NPH,plant height measured from the base of the culm to the tip of the plant in the natural condition (cm);APH,plant height measured from the base of the culm to the tip of the plant in the vertical position (cm);TN,tiller number;LN,leaf number;LL,leaf length (cm);LW,leaf width (cm);LI,leaf index;LA,leaf area (cm2);Bio,biomass (g);YPP,yield per plant (g);EPN,effective panicle number;AWN,awn (1,having long awn;0,lacking or having short awn);PL,panicle length(cm);PBN,primary branch number;SBN,secondary branch number;PBD,primary branch density;BI,branch index;TNPP,total number of seeds per panicle;FNPP,filled seed number per panicle;ENPP,empty seed number per panicle;SD,seed density of panicle;ST,seed setting rate;HGW,100-seed weight (g);GL,seed length (mm);GW,seed width (mm);GTH,seed thickness(mm);GS,seed shape;HC,hull color;PC,pericarp color.45,60,and 90 mean 45,60,and 90 days after sowing,respectively.

4.Discussion

4.1.Reasons for the rapid spread and outbreak of Chinese weedy rice

The annual planting area of rice in China is about 30 million ha (Penget al.2008),and the yield is about 7 000 kg ha–1.Since 39% of the 999 sites sampled across China were infested with weedy rice (Fig.1),its annual occurrence was estimated at about 11.7 million ha (where weedy rice populations occur as weeds).Among the six rice-growing regions in China,Ningxia in Northwest China,Jiangsu in Central China,Liaoning in Northeast China and Guangdong in South China were the most seriously infested areas with weedy rice.In areas severely infested with weedy rice,crop yield can decrease by at least 20% (Liang and Qiang 2011).Among the rice fields infested by weedy rice,51% (5.97 million ha) were seriously infested (with OWII exceeding 25%),causing an estimated reduction of about 8.36 million tons of annual rice yield.About 5%(approximately 585 000 ha)of infested paddy fields had OWII more than 75%,which may lead to almost no effective rice yield.

The wide occurrence and outbreak of weedy rice are attributed to the following three factors in China.The large-scale application of direct seeding and simplified rice cultivation techniques have been considered one of the main reasons that aggravate weedy rice occurrence (Maet al.2008;Lianget al.2009;Liang and Qiang 2011).Our investigated results showed Ningxia in Northwest China,Jiangsu and Anhui in Central China,Heilongjiang and Liaoning in Northeast China,Guangdong and Hainan in South China where direct seeding rice cultivation was prevailed frequently infested by weedy rice,with an average incidence of more than 23% (Fig.1).The proportion of direct seeding rice in Ningxia could reach up to 90% (Chenet al.2019),and Guangdong and Hainan were 17–29% (Chenet al.2020),especially in Leizhou City,which may reach 80–85% (Liet al.2013).Weedy rice in direct seeding fields emerges earlier than cultivated rice and has more detrimental effects on the growth and yield of cultivated rice (Daiet al.2017).In the shallow-tillage fields,60–90% of the weedy seeds will be buried in the upper soil layer of about 5 cm (Nicholset al.2015),95% of the emergent weedy rice comes from the seeds in the 0–10 cm soil layer (Zhang Zet al.2019).Although the direct seeding area in Northeast China and Jiangsu was small at present,it used to be as high as 70–80% in the last century (Xieet al.2009;Chen 2013).The weedy rice in these direct seeding fields can produce a large number of seeds,and some of them can enter the deep soil seed bank for long-term dormancy.When conditions are suitable,dormancy will be broken quickly (Nicholset al.2015;Mavunganidzeet al.2020).Once the new germinated weedy rice is not cleared in time before seed shattering,its infestation will increase exponentially in the next few years,especially in rice–rice cropping systems such as double rice–cropping areas in South China,rice–wheat–rice rotation in Central China and single rice–cropping areas in North China.Therefore,until now,the infestation of weedy rice in Jiangsu and Northeast China is still severe (Fig.1).Rice cultivation in Sichuan,Chongqing,Guizhou Hunan,Hubei and Fujian were mainly transplanting (over 65%) (Chenet al.2020),so their OWII of weedy rice was mostly below 25% (Fig.1).

With the popularization of simplified cultivation techniques in China,chemical weed control and mechanized rice production have become the inevitable choice.Herbicides can effectively control sensitive weeds in paddy fields.However,they cannot effectively and selectively control weedy rice (Busiet al.2017),which is physiologically and biochemically similar to cultivated rice (Grosset al.2010).The socialization of agricultural machinery services between provinces and cities also significantly increases the risk of weedy rice spreading across regions (Gaoet al.2018).The mechanization level of rice cultivation and harvest in China was more than 70%,especially in Ningxia,Jiangsu and Northeast China(nearly 80%).Thus mechanization of rice production and the dependence on non-selective herbicides also exacerbated the outbreak of Chinese weedy rice in the absence of hand-weeding.

4.2.Factors affecting the morphological diversity and distribution of Chinese weedy rice

The variation in weedy rice is the main reason for its wide adaptation to various environments,which is also the reason for its management difficulties (Rathoreet al.2016).The H′ and D respectively represent the uncertainty of the trait level of the weedy rice population in random sampling and the probability that the traits of the two randomly sampled populations belong to different levels.The higher the H′ and D,the higher the diversity of this trait.The H′ and D of 42 traits in Chinese weedy rice were greater than 1.60 and 0.73,respectively (Fig.6).If two populations were randomly selected from the 287 weedy rice populations to determine the above 42 traits,they would have a high probability of belonging to different levels (above 73%).Chinese weedy rice is rich in morphological diversity,particularly in East China,similar to the genetic divergence reported by Liet al.(2015).The weedy rice in East China (the first two groups) had the largest diversity of morphological traits,40 traits with H′greater than 1.50 and D greater than 0.70 (Fig.6),and the first group of weedy rice was confined to Jiangsu (Fig.5).The diversity of Jiangsu weedy rice populations was greater than that from the other regions.The increased morphological diversity of weedy rice populations in Jiangsu is most likely attributed to the multiple origins or diverse sources of weedy rice individuals within or among the populations.The genetic differentiation among weedy rice populations in Jiangsu reached 39% and within the population reached 60% (Shaoet al.2011),which may be becauseindicarice varieties have been gradually replaced byjaponicarice varieties in Jiangsu for the past 10 years (Songet al.2015).The change ofindica–japonicarice varieties has gradually increased the source of weedy rice and promoted the differentiation of morphological types of weedy rice.Indica–japonicahybridization may increase the adaptability of newly formed weedy rice to the environment (Qiuet al.2014),enhance its competitiveness,and lead to the first group of weedy rice in Jiangsu rice fields.

The climatic regions in China are divided into 12 temperature zones,24 arid or humid regions,and 56 climatic zones (Zhenget al.2013).The differentiation of the geographical distribution of weedy rice groups based on morphological characteristics is affected by the climatic environment.The morphological diversity of Chinese weedy rice groups is also closely related to the diversity of the climatic environment in China (Fig.2).The results of CCA demonstrated the relationship between morphological characters and geographicalclimatic factors.The first axis extracted by analysis was positively correlated with the mean diurnal range(0.592) and altitude (0.513),but negatively correlated with the minimum temperature of August (–0.630) and July(–0.612).The following traits close to the first axis were easily affected by the minimum temperature of August and July,mean diurnal range and altitude: leaf type index,panicle length,primary branch density,leaf width (90 DAS),biomass,yield per plant,leaf number,tiller number(Fig.2).The second axis extracted by analysis was positively correlated with the mean temperature of the driest quarter (0.551) and precipitation of June (0.460),but negatively correlated with the temperature annual range (–0.596) and latitude (–0.594).The following traits close to the second axis were easily affected by the mean temperature of the driest quarter,precipitation of June temperature annual range and latitude: total number of seeds per panicle,filled seed number per panicle,empty seed number per panicle,seed density of panicle,leaf length (60 DAS),leaf width (90 DAS),leaf area (60 DAS and 90 DAS),primary branch number,effective panicle number (Fig.2).In addition to the geographical-climatic factors mentioned above,the first and second axes were somewhat associated with other environmental factors,but the correlation is weak.

In the south of the Qinling Mountains–Huai River line(double cropping rice region in South China,southern part of rotation rice region in Central China,single and double cropping rice region in Southwest China),the average temperature in the rice-growing season is higher;indicarice is dominant.This study found that weedy rice in this area had evolved into the first two weedy rice groups with strong growth vigor to compete with cultivated rice for sunlight,water and nutrients.Their plants are taller,with more tillers,more leaves and larger leaf area (Fig.4).The diversity of leaf traits,plant height,tiller number and panicle number of weedy rice in this area was also the most abundant with H′ greater than 1.90 (Fig.6).In the north of the Qinling Mountains–Huai River line (single cropping rice region in North China,Northeast China and Northwest China),the mean diurnal range of temperature is large,and the precipitation is less and mainly concentrated in the wettest season.The third group of weedy rice in this study was dominant in this area,its aboveground biomass and yield per plant relatively decreased compared with the other two groups,but the 100-seed weight increased (Table 3).This is conducive to improving the survival rate of seeds in unfavorable environments (such as drought and cold) to make up for the disadvantage caused by the insufficient amount of seeds.The diversity of seed thickness,100-seed weight and seed setting rate of weedy rice in this area was also rich,and its H′ was greater than 1.95 (Fig.6).Under the common garden in Nanjing,weedy rice from high latitudes has a shorter growth period,a lower plant height,and a decrease in tillers,leaves,and leaf areas,while weedy rice from low latitudes shows the opposite change.The climatic environment in high altitude regions is similar to that in high latitude,so their morphological traits are also similar,where the third group of weedy rice is also dominant.So there is some specialization or regional adaptation of weedy rice that determines its performance under a common garden situation.

Agricultural production also affects the distribution of weedy rice groups (Wongtameeet al.2017).Chinese weedy rice mostly evolved by de-domestication of cultivated rice ancestors (Qiuet al.2017;Sunet al.2019).Complex climatic conditions determine the diversification of suitable growing areas and rice varieties (Meiet al.1988),which may also affect the diversity of weedy rice (Fig.2).Genetic differences exist among weedy rice populations in China related to their coexisting rice varieties (Caoet al.2006;Songet al.2015;Hanet al.2020).Cheng’s index is a common method for evaluatingindica–japonicatypes in rice and weedy rice based on morphological characteristics such as seed length to width ratio and leaf hairs (Chen 2014).In this study,most of the first two groups of weedy rice were distributed in South China (Fig.5),and they had the following phenotypic characteristics similar to those of southernindicarice:taller plants,more tillers,thicker stems,wider leaves,and slender seeds (Fig.4).The third group of weedy rice,mostly in North China (Fig.5),was just the opposite of the first two groups,showing the phenotypic characteristics similar to northernjaponicarice.

Although rice is mainly self-pollinated,bi-directional gene flow between cultivated rice and weedy rice does occur (Zhanget al.2018).Some weedy rice genomes contain both temperatejaponicaandindicagenomic components;at least part of the weedy rice originated fromindica–japonicahybridization,which could increase the environmental adaptability of their offspring (Qiuet al.2014).The coexistence of cultivated rice and weedy rice increases the possibility of gene flow between them.The morphological distribution of weedy rice was affected by cultivated rice,but its morphological characteristics are different from associated cultivated rice.Weedy rice needs to adapt to the local environment and evolve its survival advantage over cultivated rice.This survival advantage is the product of its adaptation to the environment and parallel evolution with its associated cultivated rice (Xieet al.2020).In this study,most of the second group of weedy rice wasindica-like type,but there was still 15%japonica-like type (mostly in Southwest China,Central China,North China and Northeast China rice-growing regions).Similarly,in the third group,which wasjaponica-like type,there was still 23%indica-like type (mostly in Central China and Northeast China ricegrowing regions) (Fig.5).It implies that the weedy rice types are still in dynamic change.The differentiation ofindica–japonicatypes of Chinese weedy rice may need further in-depth and detailed studies.

4.3.Strategies for controlling Chinese weedy rice

In this study,we found that the infestation of Chinese weedy rice is very severe.More than 34% of paddy fields have been infested by weedy rice,and OWII of weedy rice in some areas exceeds 75% (Fig.1).Chinese weedy rice in different regions was diverse in types and morphologies (Fig.6),complex and multidimensional in phenology and ecology distribution (Fig.5).Therefore,according to the occurrence and infestation degree of Chinese weedy rice in different regions,it is imperative to adopt different control strategies,including cultivation and tillage management,harvest method optimization,and herbicide control (Deloucheet al.2007;Chauhan 2013;Olajumokeet al.2016).

Comprehensive strategies such as cultivation,tillage,mechanical and chemical measures should be taken to control weedy rice in regions where weedy rice was seriously infested,such as Ningxia,Liaoning,Jiangsu,and Guangdong (OWII above 25%) (Fig.1).It is best to avoid direct seeding and use transplanting in these regions.Transplanting rice (including mechanized transplanting)has large seedlings and is more competitive with weedy rice seedlings,and combining the establishment of a flooding water layer or deep tillage (which buries seeds under 10 cm in the soil) can suppress the germination emergence of weedy rice (Deloucheet al.2007;Chauhan 2013;Zhang Zet al.2019).The infestation of weedy rice in Central China,such as Hunan,Hubei and Henan,was relatively milder,and the main reason for this may be that the main cultivar almost is hybrid rice,which has a competitive advantage over weedy rice at the early stage in these regions.In addition,the cultivation of hybrid rice is mainly transplanting,conferring cultivated rice to more advantages.For direct seeding,it is better to sow rice seeds in rows or holes so that the weedy rice between rows can be easily identified.Weedy rice between rows can be removed in time with non-selective contact herbicides,but those within rows need to be removed by hand.Increasing the seeding rate in the direct seeding system may beneficially suppress weedy rice (Estorninoset al.2005;Chauhan and Johnson 2011).

Chemical herbicides are an important part of modern weed control.Before sowing,inducing the weedy rice seeds in the soil seed bank to germinate then killing them by non-selective herbicides can reduce or avoid their occurrence and infestation (Deloucheet al.2007).Intercepting seeds at the water entrance and removing floating seeds with a net during irrigation before rice planting could significantly reduce or even deplete the seed bank (Zhanget al.2021).Weedy rice is more recognizable as they enter the reproductive stage because of their “wild” characteristics and appearance(tall plants,early maturity,shattering and red pericarp).Once the weedy rice plant is identified,the whole plant should be removed by hand-weeding,chemical or mechanical weeding as soon as possible (at least before seed shattering) to prevent it from producing more seeds (Deloucheet al.2007).Non-selective contact herbicides such as glufosinate-ammonium could also be applied with the smear method (do not smear on crops) to control the weedy rice between rows (Chauhan 2013).Since most weedy rice has earlier tillering and heading and has a height difference (mostly higher than that of cultivated rice by 16%) from cultivated rice,this method can also be used for such weedy rice types in the tillering or heading stage.Although there is no selective herbicide currently targeting all types of weedy rice,some herbicides have better control effects on weedy rice under specific conditions.For example,alachlor,dimethenamid,acetochlor,oxadiazon and metolachlor can be used to control weedy rice in transplanting fields(Eleftherohorinos and Dhima 2002;Chauhan 2013;Yang 2015).Benzobicyclon,a post-emergence herbicide,can controlindica-like weedy rice injaponica-cultivated rice fields.Hence,it may be applied toindica-like weedy rice regions wherejaponicarice is cultivated,such as Jiangsu and Shandong and some regions of Zhejiang,Liaoning and Ningxia.The development of herbicideresistant rice (such as imidazolinone herbicide-resistant rice) provides an option for the chemical control of weedy rice in paddy fields (Olofsdotteret al.2000;Schiocchetet al.2015),although there is a risk that resistance genes can drift to weedy rice through pollen-mediated gene flow bidirectionally (Kumaret al.2008;Goulartet al.2012;Sunet al.2015;Merottoet al.2016;Zhanget al.2018).

In areas with minor or no weedy rice infestation (OWII below 25%) (Fig.1),strict preventive measures must be taken to block importation.In addition to the visible mixing and contamination of weedy rice seeds in cultivated rice,the pollen-mediated gene flow of weedy rice to cultivated rice also leads to cryptic weedy rice among cultivated rice seeds (Zhanget al.2015,2018).It is imperative to implement planned quarantine on rice seeds based on the weedy rice limit standard for the purity test of cultivated rice seeds as soon as possible to prevent weedy rice disposal through the commercial transaction of cultivar seeds.For example,in Latin America,where weedy rice is severely infested,Uruguay has a minor infestation of weedy rice due to its strictest rice seed-testing system for mixing weedy rice seeds (Deloucheet al.2007).The pericarp color of dominant weedy rice is mainly red,which can be used as a trait indicator to identify visible weedy rice contaminants.However,in this experiment,a small amount of weedy rice had no red pericarp,but their plant was very high,and it had the characteristics of early maturity and strong shattering.Therefore for these invisible weedy rice,developing high-throughput molecular detection technology is necessary as the technical basis for implementing strict seed quarantine.Combine harvesters operating across regions are an important way for the spread of weedy rice (Gaoet al.2018).Some of thejaponica-like weedy rice in Group 2 and theindica-like weedy rice in Group 3 may invade these areas through seed transmission.Machines used for rice production,such as land preparation,sowing,harvesting,and threshing,should be clean and without weedy rice seeds before moving from one field to another;the same is for irrigation water.

4.4.Future research needs and outlook

The development of sustainable integrated management of weedy rice is always urgently needed as urbanization drives labor-saving rice cultivation (Ziskaet al.2015;Julianoet al.2020;Svizzero 2021).Deep studies on biology,ecology and genetics will provide a theoretical basis with new control techniques and utilization for weedy rice in the future (Ziskaet al.2015;Nadiret al.2017;Julianoet al.2020).The first scientific issue is how to originate for weedy rice,i.e.,evolutionary mechanism of biological characteristics related to weediness including strong competitive ability,early maturity,shattering and dormancy.Although numerous studies have been addressed these weedy characteristics,the evolutionary mechanism largely remains unknown (Thurberet al.2013;Nadiret al.2017;Xieet al.2020;Svizzero 2021).Weedy rice is co-species with cultivated rice but becomes a severe noxious weed through de-domestication of the latter (Liet al.2017;Qiuet al.2017,2020).Comparative whole-genome analysis has revealed the origin of weedy rice from cultivated rice,but the genetic study seems to have answered how to evolve into weedy traits.We infer that epigenetic mechanisms may play a key role in the evolution of the above weediness characteristics of weedy rice (Xieet al.2020).Hence,we propose that an extensive methylome analysis on weedy rice may elucidate the epigenetic mechanism of weediness evolution.

Artificial intelligence and information technology can be introduced to the IWM of weedy rice on the aspects of prediction occurrence pattern and range and identification techniques for weedy rice based on either generative adversarial network architecture or hyperspectral imaging techniques (Zhang Yet al.2019;Petchsod and Sucontphunt 2021).Moreover,global changes such as climate change (such as rising CO2concentration,extreme temperature or precipitation),urbanization-driven cultivation shift and genetically-modified rice will impact weedy rice and its control.Because weedy rice with greater adaptability may have increased competitiveness in a changing climate;and climate change may also lead to alterations in rice establishment methods and varieties,affecting the gene flow between weedy rice and cultivated rice,thereby affecting the distribution,spread,and infestation degree of weedy rice directly or indirectly(Ziskaet al.2012,2014,2015).This may also be another worthy scientific issue to be studied further in the future.

Direct seeding is labor and costs savings and easy mechanization for rice cultivation,but direct seeding rice suffers from enormous weed infestation pressure,particularly weedy rice (Kumar and Ladha 2011).Direct seeding tolerant rice cultivars may be bred to make them germinate rapidly and highly competitive at the early establishment stage.The rice varieties are supposed to be more suitable for direct seeding because they can suppress the growth of weedy rice through competitive advantages in the early stages.Weedy rice may be a rich genetic resource material for special direct seeding cultivars (Nadiret al.2017).Weedy rice with stronger competitiveness is more adaptable to the natural environment and climate change,and a variety of new genes or QTLs for resistance to biotic and abiotic stresses have been identified in weedy rice (Jia and Gealy 2018;Fogliattoet al.2020).With tools such as genomewide association studies (GWAS) and re-sequencing approaches (Liet al.2017;Qiuet al.2017,2020),we can dissect the inheritance of various excellent morphological,physiological and developmental traits in weedy rice,such as higher stress resistance and disease resistance,and stronger competition.These genetic resources of weedy rice can be used in future breeding programs to develop more competitive cultivars with improved quality and yield(Hanafiahet al.2020;Jeonget al.2020).

5.Conclusion

Weedy rice widely occurred in all rice-growing regions in China,with an incidence of 39%.The sampling sites with 50% or more OWII were mainly found in Central China(Jiangsu),Northwest China (Ningxia),Northeast China(Heilongjiang and Liaoning) and South China (Guangdong and Hainan) rice-growing regions.Weedy rice also occurred sporadically in the North China and Southwest China rice-growing regions with a milder infestation.A total of 287 Chinese weedy rice populations were divided into three morphotypes: Group 1,strong tiller type with strong growth vigor,only distributed in Jiangsu;Group 2,large-leaf type with strong growth vigor and weak seed,mainly distributed in the south of Central China,East China,South China and Southwest China;Group 3,large-seed type with weak growth vigor,mainly distributed in Northeast China,North China and Northwest China.Group 2 had the largest morphological diversity.These three weedy rice groups had significant differences in vegetative and reproductive traits such as leaf traits,plant architecture,yield,seed traits and panicle traits.With the Qinling Mountains–Huai River Line as the boundary,the morphotypes of Chinese weedy rice showed a climatedependent geographic differentiation.Their morphological and distribution characteristics were closely related to the geographical-climatic factors and cultivated rice types,suggesting that weedy rice may co-evolve with local rice varieties.A better understanding of the dynamics of weedy rice populations will result in more effective advice for weedy rice identification and control.Studying the morphological characteristics and type distribution related to competitiveness and adaptability of different weedy rice populations can early warn their spread and dispersal in different regions and offer new ideas for rice breeding.

Acknowledgements

This research was supported by the Major Scientific and Technological Project of Hainan Province,China(ZDKJ202002),the China Transgenic Organism Research and Commercialization Project (2016ZX08011-001),the National Key Research and Development Program of China (2016YFD0200805).The authors thank Prof.Bernal E.Valverde (Research and Development in Tropical Agriculture,Costa Rica) for editing the article.

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