Sunshine duration across Australian regions and their coordinates .
Barley heading date is important in adapting barley genotypes to different environments. Important factors affecting heading date in barley are temperatures, photoperiod and sowing date. Sowing date is a management option to influence heading date. It is used to reduce climatic risks such as frosts and water stress at sensitive periods during crop development. Three major genes control heading date in barley. These genes regulate photoperiod (Ppd-H1 and Ppd-H2), vernalization (Vrn- H1, Vrn-H2 and Vrn-H3) and the duration of the vegetative phase (Eps). The Ppd-H1 locus on chromosome 2(2H) regulates flowering time under long days. Ppd-H2 on 2H regulates phenology under short day length. Vernalization is mainly controlled by three loci (VRN-H1, VRN-H2 and VRN-H3), which interact in an epistatic fashion to determine vernalization sensitivity. Appropriate physiological and simulation frameworks such as that of APSIM-Barley are required to complement breeding efforts in order to determine the location of the Eps genes and describe and quantify the effects of environment and management on gene expression and their impact on yields and quality in barley.
- earliness per se
World barley production is projected to reach 140 million metric tons (MMT) on 50 m hectares by 2016/2017  with its greatest economic impact being related to its use as feed, food and for malting. The demand is projected to reach 142 MMT by 2050 . It is therefore one of the major sources of income for the countries where it can be produced. Australia is the fourth largest barley grower in the world, producing about 8.7 MMT of barley in 2015 and contributes up to 30% of the world supply [2,3]. The supplies comprise 2.5 MMT of malting barley and 4.5 MMT of the feed barley [4, 5]. Currently, one‐third of the world production is used for malting . The grain is also widely used for human food and livestock feeds, starch production and chemical industries, while the straw is used for roofing huts and animal bedding. Grazing is sometimes performed after harvesting or when the crop is green [4, 7].
There are constraints facing barley-producing nations such as Australia; including transient, unpredictable and varying climatic conditions [8–10]. These environments are characterized by a lack of adequate water in spring and summer periods when evaporation and transpiration are rising rapidly when crops are in the later stages of development, which results in a terminal drought. There is also a problem of frost when the air temperature drops to 2°C or less. Damage to crops from frost may occur at any stage of development but is most damaging at and around flowering. These constraints result in a serious dilemma for growers who must decide whether to delay anthesis to avoid frost damage or flower as early as possible in order to escape the effects of terminal drought . Thus, it is important that barley cultivars demonstrate an adaptation with appropriate rates of development across the heterogeneous environments.
1.1. Barley phenology—its relationship with abiotic stresses, quality and yield
Plant phenology characterises the developmental life cycle events of plants and how these events are influenced by seasonal and inter‐annual variations in climate as well as habitat factors [12, 13]. In barley, different development stages, such as spikelet initiation and duration of grain development can seriously influence yield and quality. These stages are regulated by environmental factors such as temperature or growing degree days (GDD), duration and intensity of light, nutrition and husbandry techniques . Heading date is important in adapting barley genotypes to different stresses such as heat stress, waterlogging, salinity and drought. Heat stress can quickly deplete the available moisture through high rates of evapotranspiration and ultimately leading to terminal drought . Both heat and drought at late sowing may interrupt barley developmental processes usually from double ridge (DR) to maturity. The resultant effects of these stresses are reduction in plant height, dry matter accumulation and grain yield . On the other hand, low temperature at early growth stages (Zadoks GS10) may be required for vernalization especially for winter barleys to flower. Apart from the optimum conditions, poor biomass accumulation and significant yield losses are attributed to extreme conditions, high or low temperatures, drought or waterlogged anaerobiosis and other soil-related problems . Advances in crop phenology and modelling have helped with the understanding of how to assess biomass partitioning and effects of abiotic stresses in crops . Modelling has also helped understand the effects of different environments and sowing dates on growth and development of barley plants.
Many scoring systems for plant growth stages have been developed to describe phenology in cereals . The most widely used scales are Feekes scale  and Zadoks scale . The majority of the scales described only the morphological traits , while very few describe the apical developmental processes especially on barley . The vital developmental stages that have significant effects on yield and quality are DR (Zadoks GS30), construction phase which includes stem elongation (31), heading and anthesis (51&61 as in barley) and grain filling stage [19, 20]. DR and terminal spikelet (TS) can only be detected through destructive examination (Figure 1). There have been no consistent reports on the use of correlated traits for the determination of the flower/floral initiation stage (GS30), although in some cases surrogate traits such as number of leaves on the main stem have been used to determine this stage .
Developmental stages drive appearance of the main traits in cereals and are sensitive to climate and management [22, 23]. DR stage is an elementary step for predicting flowering date and crop yields and has been an important trait for improving crop productivity and adaptation . Short construction phase (e.g. from Zadoks GS31 to 65) or short grain filling periods often lead to low yields . The increase in the rate of grain filling is a positively correlated with grain weight . A significant increase in grain yield in the high rainfall zones of Australia (usually along the coast with over 550 mm rainfall) was shown to be due to a longer duration between result from an increased time from GS31 to GS65 (i.e. stem elongation to anthesis) in wheat .
1.2. Factors affecting phenological development in barley
Barley development has three main stages: germination/emergence to double ridge/stem elongation (GS10 to GS30/31), stem elongation to heading/anthesis (GS30/31 to GS51/61) and heading to physiological maturity (GS51 to GS 94) [15, 28, 29]. The first stage, the basic vegetative phase, is the stage for the production of phyllochrons, roots and tillers. The agronomic significance of this stage is the generation of enough biomass for livestock feed especially for dual‐purpose genotypes . The second stage involves the termination of vegetative growth at GS29. This stage signifies reproductive growth, spikelet initiation and the onset of stem elongation  and is required for the production of a higher number of spikelets which directly link to grain yield . The third stage is grain filling, influencing grain size and weight . This stage is essential for yield increase as well as quality. All the three stages are regulated mainly by genetic, environment as well as management.
Important environmental factors affecting barley developmental stages include temperatures and photoperiod [15, 34, 35], both of which vary simultaneously in field conditions. This variability affects developmental events that determine flowering time and consequently yield . Temperature is very important for all plant physiological processes , especially for the variation in days to spikelet initiation , days to heading and days to flowering (anthesis) in cereals [38, 39]. This observation is supported by Hay et al.  and Ellis et al.  who reported that the rate of primordia initiation especially spikelet in barley has a linear relationship with average daily temperatures. The rate of initiation of organs in barley also increases linearly with temperature; the optimum temperature for organ development is 25–30°C . However, low temperature is required in some cereals to stimulate flowering (vernalization); this term has
Growing degree days (GDD) are often used for measuring developmental events in barley. GDD is the accumulation of the mean daily temperature above a base temperature (0°C in the case of barley) [34, 42]. Below 0°C, the development of the crop will cease while above 0°C the growth will increase linearly with temperature [34, 36, 43]. Using zero as the base temperature (Tb = 0°C) in wheat, Acuna et al.  identified an ideotype which may develop sufficient tiller numbers at 650°Cd to harvest about 400–500 heads/m2. The same ideotype had a construction phase duration (CPD) within 800–1200°Cd to be able to escape frost and partitioned more assimilate to developing grain. Miralles et al.  reported the GDD that ranges between 950–2000°Cd and 1300–2100°Cd from sowing to flowering in both barley and wheat, respectively . This result reflects the high variability in the GDD for flowering time in barley, indicating that yield might be manipulated using this characteristics.
Photoperiod is also a key environmental factor that affects the development of barley especially in temperate countries. The duration of day-length has a predictable pattern that drives evolutionary plant responses. This forms the basis of classifying barley as a long day, short day and day‐neutral (DN) plant . Photoperiod can significantly influence the duration of both vegetative spikelet initiation and stem elongation periods in wheat and barley . The long day‐length of higher northern and lower southern latitudes causes both photoperiod‐sensitive and photoperiod‐insensitive cultivars to flower early. The winter type responds strongly to long days, while spring types vary depending on the selection criterion . For example, in Western Europe and parts of North America, short days increase the duration of vegetative growth of spring‐sown barley. This lengthens the time for biomass accumulation and hence increases yield  under temperate growing conditions. Australia has a unique environment that differs from other barley growing regions in higher latitudes , where daylength is much shorter in winter, but considerably longer in summer (Table 1). However, there is variation in the extent of sensitivity of Australian genotypes to photoperiod. The baseline for the sensitivity ranges from 8 to 10 h of exposure, below which no flowering initiation occurs, while the upper limit ranges from 13 to 18 h . The sensitivity of both vernalization and photoperiod starts immediately after plant emergence .
|Region||Coordinate||Seasonal sunshine hour duration|
|NT, Katherine||14.465°S, 132.26° E||11–12||12–13||12–13||11–12|
|Victoria, Mallee||35.11°S, 142.36°E||9–11||11–14||12–14||10–12|
|Ravensthorpe, WA||33.32°S, 119.82°E||9–11||11–14||12–14||10–12|
|SA, Wimmera||37.82°S, 140.78°E||9–11||11–15||13–14||10–12|
|Brisbane, NSW||27.47°S,153.02° E||10–12||11–15||12–14||10–13|
|NSW, Queensland||20.92°S, 142.70° E||9–12||11–14||12–14||12–14|
|Launceston, Tas||41.44°S, 147.14°E||9–11||11–15||13–15||9–13|
Management is also an important factor that affects phenology in barley; important factors include sowing date, fertilizer application, irrigation and other management practices. Matching the phenology with an appropriate sowing window allows growers to better manage climate risks that are particularly pronounced in Australia, where early sowing may expose the heading of spring barley to frost, while late flowering and terminal drought can curtail grain filling and hence reduce yield . Plant growth and development are affected by high temperatures and water stress in late sowing of spring barley [15, 47], high temperatures at early sowing in Punjab of India  and low temperatures with early sowing in Russia . In the same experiment with early sown crops, low tillering capacity as a result of low temperatures caused a significant reduction in grain yield . Similar results were reported by Ram et al. , in that very early sowing of spring barley may affect tillering capacity, although in this case due to high temperatures and reduction in biomass accumulation for late sowing. In addition, sowing date was found to have the most significant effects on the phenological development of cereals, especially during the GS31 (stem elongation phase) in wheat .
2. Physiological and molecular mechanisms for heading date and their effects on grain yield and quality
Heading date (spikelet emergence, Zadoks GS51) is a complex trait in barley that has direct impact on grain yield and quality and also forms the basis of evolutional adaptation to the changing climate. The mechanisms of regulating this trait are so complicated that there has been no final conclusion on the specific number of genes that are involved and their interactions . Although the expression of this trait is governed by complex factors such as genetics, physiology and environment , neither the plant breeder nor the physiologist can clearly explain their interactions. For example, physiologists have not answered some unresolved developmental issues such as the regulation of the developmental rate within an environment and the cause of the transition between one growth stage to another. Equally, crop breeders need to account for the gene functions, the number of the genes and their interactions that are involved in expression of heading date .
2.1. Genetic regulation of heading (flowering) date
Barley improvement dates back to its domestication period. However, significant yield improvements began only in the 1950s as a result of the application of more advanced plant gene technologies . Studies have been conducted on the genetic improvement, which include the determination of genotypic and phenotypic variability in days to flowering and growth phases, as well as comparing different sets of cultivars in barley [29, 50] and wheat . Recent advances in the area of molecular breeding using highly polymorphic molecular makers such as simple sequence repeat (SSR) and single nucleotide polymorphisms (SNPs) have also led to significant progress in the improvement of yield and quality in barley. These markers are being used to tag genes or quantitative trait loci (QTLs) that are of economic importance, offering promises in their use in marker‐assisted selection (MAS) . Among all the markers, SNPs and SSRs are unanimously believed to be best suited for the use in marker‐assisted breeding . They have been used to assess most of the genes in barley and other cereals through cDNAs, expressed sequence tags (ESTs) and sequenced PCR amplicons that provide use of SNPs in protein encoding transcribed genes.
A genetic study on early versus late heading in barley was first reported in 1907 . Exploration of the effects of environment on heading date was initiated following the reports of Garner et al. . This opinion was supported by other studies that sources of variation in flowering date among different genotypes of barley were due to the effects of seasonal variation, location and sowing date . As a result, genotypic differential response to photoperiod, vernalization and other environmental conditions have been conducted. Three groups of genes are responsible for variations in heading date. These include Ppd (photoperiod) [20, 44, 56, 57], Vrn (vernalization) [58, 59] and earliness
2.1.1. Photoperiod genes (Ppd)
The photoperiod pathway is generally classified into two components: the circadian clock and the photoperiod clock regulators . The clock is the receptor of light stimuli perceived by phytochromes (phyA to phyE) and cryptochromes (cry1 and cry2) which are red and far red receptors and blue light receptors, respectively . Temperate cereals, including barley, are quantitative long‐day crops ; however, some varieties differ in their response to photoperiod , as mentioned above. Photoperiod sensitivity can significantly influence the duration of both vegetative and stem elongation periods in wheat and barley . Longer day‐length causes both photoperiod‐sensitive and photoperiod‐insensitive barley cultivars to flower early. There is evidence that specific stages in floral development in wheat and barley may also be sensitive to light intensity  and heavy shading during the later stages of ear development may result in the infertility of the spike . As a result, flowering in photosensitive plants like barley may be entirely inhibited if the light intensity is reduced sufficiently during long periods because of the low level of availability of carbohydrates within shaded plants . Therefore, genotypes vary in the photoperiodic threshold below in which flowering initiation will not take place.
In barley, two photoperiod genes influencing flowering time are
Many other QTLs have been identified from different populations. Ren et al.  also detected a major QTL under 18‐h photoperiod in glasshouse experiments and mapped the QTL to the Xp12m50B199–Xp13m47B399 interval of flanking markers on chromosome 4H which accounted for 77.48 and 37.81% of phenotypic variation for long photoperiod response in Australia and China, respectively. The gene,
2.1.2. The vernalization genes (VRN)
Vernalization is the requirement for prolonged low temperature to advance flowering in cereals and depends on the growth habit (such as spring or winter types). The winter types of barley require cold exposure before flower initiation, typically below 10°C for a period between 4 and 6 weeks , depending on base temperature as defined above. In contrast, the spring types have minimal low temperature dependency and are usually insensitive to vernalization and Short day photoperiod . This behaviour is characteristic of many temperate cereals like barley [87, 88] and associated with the capacity of a genotype to survive the cold winter during vegetative stages [87, 89, 90]. Barley is an excellent model for genetic analysis of low‐temperature tolerance in fall‐sown cereals . Its responses to vernalization have been observed to vary greatly among genotypes and between growth phases [45, 90].
Vernalization in cultivated barley is mainly controlled by three major Vrn genes , Vrn1, Vrn2 and Vrn3 , or HvVrn1, HvVrn2, HvVrn3 , or Vrn‐H1, Vrn‐H2 and Vrn‐H3 . The Vrn‐H1 (also named as Sgh2 or Sh2) is located in the middle of the long arms of 5H [67, 88]. The Vrn‐H2 (Sgh or Sh) is found on chromosome 4H , while the Vrn‐H3 (Sgh3 or Sh3) is on 1H . Vrn‐H1 translates the fruit‐like MADS‐box transcription factor which is an ortholog
The second locus is the Vrn‐H2 (Sgh2 or Sh2) which encodes for the zinc finger‐CCT (ZCCT‐H) transcription factor  and is also vernalization dependent. A partial or total deletion of part of this locus has been shown to cause a non‐functional mutation of the gene and a recessive form is responsible for the spring growth habit in both barley and wheat [65, 92, 95]. However, it is necessary to understand that the effects of Vrn‐H2 under field conditions can only be verified using a variety of sowing dates . The authors further stated that the gene does not affect heading date when crops were autumn sown.
The third is the Vrn‐H3 (Sgh3 or
Since there is a form of homogeneous genetic system for all the cereals with a high degree of synteny (physical co‐localization of genetic loci on the same chromosome in an individual or species), the results of one species are frequently applicable to other members of the cereal family [87, 97, 98], including barley. Consequently, the cloning of the candidate genes in diploid wheat (
2.1.3. Basic vegetative phase BVP (Earliness per se, Eps)
Barley has the potential to grow and produce economically viable yields under a wide range of diverse environment-types. Early growth plasticity is determined during the vegetative phase , which has extensive genetic diversity. One of these genes is the
2.2. Effects of Eps gene on developmental phases
The genetic and physiological processes that are linked to the adaptation of barley are due to broad differences in the developmental phases. These phases include both the time and duration from spikelet initiation (GS30) and up to grain filling (Zadoks GS70). Most previous research conducted has centred on the effects of Eps gene on the flowering time [61, 65, 67] with fewer focussing on the variations in the duration of each of developmental stages [61, 106, 107]. For example, an study conducted by Lewis et al.  using single seed decent and near isogenic lines (NILs) observed a significant interaction between the Eps gene and the timing and duration from vegetative to flowering phase especially from double ridge to terminal spikelet stage in diploid wheat. The interaction showed that the NIL genotypes with the early allele, Eps‐e, had the transition to DR stage 35 days earlier (67% less) than the genotypes with the Eps‐l alleles. The SSD genotypes had highly significant differences (P < 0.0001) in both heading time and number of spikelets per spike between Eps alleles (eps‐e and eps‐l). The genotypes with the late allele Eps‐l flowered 61 days later than those with eps‐e alleles (with 76% across temperatures) and produced a mean of 8.7 more spikelets for each spike which was a 56% increase across temperatures . However, results of Valárik et al.  and Zikhali et al.  showed only a few days of differences (from 1 to 5 days) in flowering time between a pair of near isolines (NILs) and their recombinant inbred lines (RL) in both wheat and rice. In addition, no significant interaction (P = 0.67) was observed between Eps genes and the stem elongation stage , which is the beginning of construction phase. Also temperature had no significant effects on the gene determining spikelets number per spike . Contrary to this opinion, Slafer et al.  observed that lengthening the duration of the stem elongation phase, without modifying total time to anthesis, could increase the number of grains/m2 and consequently the number of grains per unit land area . In general, variations in both flowering time and spikelet number per spike could be due to pleiotropic effects of a single gene or to the effect of tightly linked multiple genes with additive effects [71, 105].
2.3. Effects of temperature on the Eps genes
Temperature is the major environmental factor affecting Eps genes in barley and wheat [61, 101]. Differences exist among genotypes carrying Eps‐e for early heading and Eps‐l alleles for late heading in wheat . There are significant interactions between the Eps gene and temperature [60, 61]. The genotypes with Eps‐l alleles had no interaction with temperatures (21°Cd difference), while lines carrying the Eps‐e allele had a shorter thermal time to heading at 16°C than at 23°C (336°Cd difference) . Lewis et al.  further showed that the thermal time to flowering for the genotypes with the Eps‐e gene was approximately 1557°Cd. These are 1118°Cd less than the thermal time for the late genotypes (2675°Cd) with Eps‐l gene. Slafer et al.  used four wheat varieties and six differential temperatures (10–25°C) to study the effect of temperature on growth stages. They showed that the developmental phases of individual genotypes were most sensitive to temperature from sowing to anthesis. This variation can be attributed to the allelic diversity at Eps locus in the lines studied. Hence, an in‐depth research of genetic variability of the earliness
2.4. Effect of sowing date on the Eps genes
In order to maximize yield potential in any environment, cultivars must have an appropriate flowering window and life cycle duration in the target environment [28, 60]. Sowing date is an important factor that governs flowering period, the timing of which needs to escape biotic and abiotic stress. Out of these three major genes, Ppd, Vrn and Eps, Eps genes do not respond to the differential to vernalization or photoperiod and still control timing and duration of flowering independent of these stimuli .
2.5. Effect of the Eps gene on grain quality
Yield and quality are important complex traits in any breeding programme. Improvement of these traits is very difficult to achieve due to their genetic, physiological and physical complexity. Grain quality could either be physical, such as size, hardness and lustre, or nutritional such as malting quality. The relationship of Eps gene on heading, spikelets number and number of grains per spike has been shown to be correlated with yield . However, grain quality can also be improved by manipulating the Eps gene loci  to improve the physical traits such grain weight or hardness. Experiments conducted by Herndl et al.  indicated that with a shorter pre‐anthesis period, the relationship between yield and protein is always negative. Crop breeders often focus on increasing yield with little attention on quality traits; thus, there is less information on the effects of the Eps gene on quality traits.
Heading date is a polygenic trait, controlled by Ppd [111, 112], vernalization [37, 38, 41, 111] and Eps genes [111, 113]. These genes interact in an additive nature (cumulative effects of non‐ allelic genes to a quantitative trait) . The genetic analysis of Eps showed that it can be simply inherited as a Mendelian inheritance, but molecular analysis has not been able to identify an appropriate molecular marker to determine its location .
Crop modelling in agriculture has been used as a physiological framework to undertake simulation of dynamic crop phenology that support crop improvement programmes . Physiologically sound simulation tools will provide quantitative assessments of crop development and yield relative to the genotype, climate, soil and management in sustainable farming systems . These tools should provide ex-ante impact assessments of research outcomes across a wide range of environments . This is particularly true for Australia where a highly variable climate poses challenges for production and crop improvement . With respect to climate change, temperature has increased by 0.9 °C since 1910 per annum and severe heat and drought spells are occurring more frequently in Southern Australia. Total annual rainfall and frost events may also increase in some of the temperate regions such as Tasmania . Other reports indicate a 2–5% reduction in rainfall across most parts of the country . Harrison et al.  emphasizes the need for serious attention on the impacts of climate stress on plant phenology. Globally, climate change will further increase temperatures, modify the amount and distribution of rainfall and consequently reduce the probability of reliable food and forage production, thereby causing a significant threat to food security and improved livelihood . More than 30% barley yield loss will be attributed to climate change as a result of drought and heat stress by 2050 . Despite this, there has been little research work to determine how the climate change will broadly affect whole farm systems [119, 122], including systems farming barley.
Mathematical functions are being used as tools to simulate crop phenology to predict the effects of climate events and changing environments on yield and quality . These tools help explain the interaction of some of the complex traits related to development and growth and their interaction with the environment. For example, ecophysiological quantitative equations were used by Yin et al.  to describe the response of flowering to photoperiod and temperature to predict days to heading and yield in diverse conditions. The equations were used as empirical and mechanistic models to provide important framework for simulating a number of events in crop growth, especially predicting heading date [24, 123]. The empirical models are often based on accumulation of growth degree days adjusted by vernalization and photoperiod [123, 124], while mechanistic models are based on the production of leaves and floral primordia at the apexes .
Four important phenology models: 3s‐beta‐model, 3‐plane‐linear‐model, modified‐rice‐clock‐ model (m‐RCM) and a logistic model were developed and evaluated in rice [39, 126]. All models were able to predict the flowering time in varying environments although with varying degree of precisions. Model parameter values from reciprocal transfer experiments also resulted in realistic differences in flowering time across all the genotypes in different environments. The models were able to partition variation due to environment and that of the genotypes. Thus, ecophysiological model could be very important for dissecting the relationship between genotype and phenotype . Chapman et al.  also developed mechanistic model called QSUN to estimate growth, development and yield of a diverse range of genotypes of sunflower under varied environments. Their model was able to account for leaf area index (r2 = 0.65), total biomass (r2 = 0.96) and grain yield (r2 = 0.93) when tested against actual phenological data. QSUN was also used to analyse the production risk of sunflower grown in highly variable subtropical environments in order to undertake decisions such as the choice of an adapted cultivar and appropriate sowing window in order to obtain higher yields . Another dynamic model was used to investigate the causes and impact of climate change on peanut production in Northern Australia . The model was used in conjunction with the information of district yield to offer an in‐depth study of long‐ term production risk. The study indicated that the stabilization of the above‐average yield, which was due to stable summer rainfalls, was responsible for the rapid expansion of peanut industries at that time. Such studies assist in gaining better understanding of complex GxExM and the identification of traits required to manage crops in variable and changing environments . Therefore, choice of an appropriate simulation models for predicting phenology are essential for choosing the best‐adapted cultivar for a specific production environment and for helping with timely planning of strategic or tactical management [9, 123].
3.1. The agricultural production systems simulator (APSIM)
The Agricultural Production Systems Simulator (APSIM) is a cropping systems simulation model that combines several decision-support tools. APSIM may be used for accurate predictions of how traits like heading dates impact on grain yield and biomass of different crop genotypes in alternative environment and management conditions and also to consider the long‐term consequences of cropping systems on soil conditions . APSIM may also be used to increase the understanding factors influencing heading date of barley when grown under field conditions [12, 115]. The tools within APSIM can also be used to describe genetic parameters regulating phenology with the function of daily temperature and photoperiod to predict flowering time and consequently yield and quality . The major challenges facing barley production are water stress coupled with heat stress during spring and summer and frost events in winter and early spring. District yield records showed about 85% yield loss due to frost events in Australia . A later study with more information was conducted to gauge the impact of frost on grain production in Australia . In the study, APSIM simulated the effects of frost on wheat production areas across Australia. The model predicted increased frequency of frost events in the Australian wheat belt (the main barley production regions in Australia) and also an increase in the mean temperatures with significant yield loss. Zheng et al.  concluded that breeding for frost tolerance could give about 20% yield advantage. As for most other modules, the barley module of APSIM (APSIM‐Barley) simulates the phenology in a daily time step. The module uses inputs of weather such as radiation and temperature and initial soil nitrogen . The module has 11 growth stages, from sowing to harvest (GS0–GS100) . Manschadi et al.  took advantage of APSIM’s scientific basis by assessing barley growing patterns under different environment and management. The model was able to account 91% and 82% of the variation for biomass accumulation at maturity and grain yield. Although negative correlation between yield and quality traits  and between life cycle and quality traits  were reported, maintenance of quality traits is critical in order to alleviate malnutrition . The majority of crop models are constrained in predicting both the physical and cryptic (nutritional) grain quality such as grain size and grain‐N content , although the model has been used to account for both above and below ground biomass, growth, water, N uptake and leaching . The same model was used to explain some quality parameters such as grain size and grain protein concentration . As a result of the challenges due to frost and drought events faced by barley production, a simulation model, QBAR, was developed to identify appropriate management options such as sowing date in order to increase yield . Thus, QBAR can simulate phenology, soil water, leaf area, biomass accumulation and yield of barley . A more detailed analysis of extreme terminal drought effects and frost risk should be conducted that should include several sowing dates and varieties. QBAR was later modified to (APSIM‐Barley) , accounted for 91 and 82% variances for biomass accumulation and grain yield, respectively . Further improvement on the QBAR model has been the integration of crop nitrogen balance and grain quality module . It was later used to account for the effects of extreme climatic events, frost and terminal drought on yield and yield components, of which paddock‐based crop models could not explain . The authors proposed that QBAR can be used to determine the best management decisions such as sowing date to obtain highest grain yield even in the events of frost and water stress.
3.2. Genotype by environment by management (G × E × M)
Information on the target environment for which crop cultivars are to be improved is vital to plant breeders . This is because a higher genetic (G) diversity for flowering time has been reported in diverse environments (E) of barley growing regions where frost and drought events limit crop growth as well as inappropriate agronomic management (M) to improve crop growth and yield . This concept can be extended to include both abiotic factors such as soil, water stress and waterlogging as well as induced stresses due to abiotic (salinity) and biotic factors (pests, weeds and diseases). Environmental factors can be classified into two types: (1) micro‐environmental factors such as year‐to‐year variation in rainfall, drought conditions, pest incidence and (2) macro‐environmental factors which include soil type and management practices [140, 141]. The association between the environment and the genotype to produce a specific phenotype is termed as the G × E interaction . Hence, the G × E interaction determines the adaptability and suitability of a specific genotype to a range of environments. Environment could also be a time boundary, such as a year (annuals) . Hence, matching heading date to diverse environment may give a large G × E interaction . Variation in the developmental stages usually from DR to grain filling stages in barley is influenced by G × E . For highly variable environment types like those found in Australia, there is a need for specific adaptation (genotype response and better performance in a specific environment) arising from G × E interaction . Löffler et al.  used a crop model index approach to account for the G × E interaction effect in US maize breeding trials. Factorial regression (FR) has been used to describe crop interaction with their environment and help understand G × E . A linear generic model was used to analyze the interaction of 96 genotypes to different environment . The model was able to explain 81% of the total variation in heading date across the environments. The introduction of molecular markers has aided our understanding of the effect of individual gene or QTL effects rather than the cultivar . Yin et al.  used a four‐parameter ecophysiological model to predict grain yield when QTL‐based data inputs were used. The model when used together with the QTL map was able to sufficiently predict days to flowering in barley , suggesting that the model could be used to help breeders in Australia to adapt new varieties.
Recent advances in plant breeding combined with dynamic models are now allowing partitioning of the effect of management in the G × E approaches . Higher genetic gain (GA) in yield has been attributed to better understanding of G × M interaction effects in maize crops, where the progressive yield increase in the US has been associated with superior genotypes being grown at higher density . Another example is the choice of a combination of non‐tillering genotypes (G) and row spacing (M) in drought prone land can help realize sustainable production and additional value in obtaining moderate yield instead of complete crop failure due to limited availability of water [72, 148]. Another study was conducted to check the performance chickpea genotypes under two different managements, irrigation and rain‐ fed management systems. The study showed highly significant yield differences among genotypes and between the two management practices for all the important traits. The study also revealed that both yield and yield components were improved by an average of 48% increase in the number of pods per plant, 36% in total dry weight and 17% in grain yield in the management involving irrigation .
It is therefore important to note that the use of models capable of accounting for G × E × M interaction in breeding and agricultural systems can be a powerful tool to better understand environment-specific, complex gene expressions. APSIM‐Barley has been used to describe broad adaptation of barley genotypes in anticipation frost or water stress across Australia. In another experiment, leaf area and yield of Baudin, Flagship, Buloke and Capstan, barley cultivars were assessed. The model also reasonably explained the relationship between the leaf area duration and yield as influenced by weather . However, the model has not been used to explore the potential agronomic benefits of exploiting G × M interaction in a specific environment. Matching specific genotypic traits to management option within a target (specific) environment will assist breeders in trait selection and the design of their breeding programs.