Growth and agronomy of Miscanthus x giganteus for biomass production

Abstract

Miscanthus x giganteus is a highly productive, sterile, rhizomatous, C₄ perennial grass that should be considered as a feedstock for bioenergy production. Here, we review the current state of research and the future of M. x giganteus biomass production. Since the 1980s, the grass has been studied and used in Europe to produce heat and electricity via combustion. Since 2005, the US government has encouraged research using herbaceous feedstocks for conversion to ethanol for use as transportation fuel. M. x giganteus is being widely studied in the USA because of its potential to produce large quantities of biomass. This review examines the taxonomy and genetics, growth and development, physiology and agronomy, and modeled and potential ‘real-world’ yields of M. x giganteus in Europe and the USA. In addition, the invasiveness and the future perspective of M. x giganteus utility and research in the USA are also considered.

Figure 1.  Miscanthus giganteus planting using plugs near Dixon Springs, IL, USA (latitude 37.4535° N), 12 weeks after planting.

Figure 2.  Second-year Miscanthus x giganteus planting in October near Lexington, KY, USA (latitude 38.1290° N).

Note the distance between clumps when planted on 1 m² spacing.

Figure 3.  Third-year Miscanthus x giganteus planting in July near Lexington, KY, USA (latitude 38.1290° N).

Note the distance between clumps has closed compared to Figure 2.

Figure 4.  Mature Miscanthus x giganteus crop near Birmingham, UK (latitude 52.5878° N), ready for harvest in April.

Figure 5.  Clump of 2-year Miscanthus x giganteus rhizomes harvested in spring immediately prior to the onset of stem growth in Urbana, IL, USA (latitude 40.0425° N).

Figure 6.  Baling Miscanthus x giganteus in late-winter harvest in Urbana, IL, USA.

Ideal biomass energy crops efficiently use available resources, are perennial, store carbon in the soil, have high water-use efficiency, are not invasive and have low fertilizer requirements [1]. One grass that possesses all of these characteristics, as well as producing large amounts of biomass, is Miscanthus x giganteus[2]. In fact, it has been reported that in the USA, growing 11.8 million hectares (ha) of M. x giganteus would be required to produce 35 billion gallons of ethanol per year, while it would require 18.7 million ha of corn (grain plus stover) or 33.7 million ha of switchgrass to produce the same volume of ethanol [2]. Without doubt, a combination of crops will contribute to ethanol production in the USA, given the great range of growing environments and the optimization of crops to these varied conditions but, given its biomass yield potential, M. x giganteus is indeed impressive and worthy of development.

M. x giganteus is a highly productive, sterile, rhizomatous C₄ perennial grass that was collected in Yokahama, Japan in 1935 by Aksel Olsen. It was taken to Denmark where it was cultivated and spread throughout Europe and into North America for planting in horticultural settings [3–5]. Over time, it has been known as Miscanthus sinensis ‘Giganteus’, M. giganteus, Miscanthus ogiformis Honda, and Miscanthus sacchariflorus var. brevibarbis (Honda) Adati. [6]. Recent classification work at the Royal Botanic Gardens at Kew, England has designated it as M. x giganteus (Greef & Deuter ex Hodkinson & Renvoize) [7], a hybrid of M. sinensis Anderss. and M. sacchariflorus (Maxim.) Hack. [8,9]. M. sinensis, M. sacchariflorus and M. x giganteus and their varieties continue to be planted in horticultural settings, and Clifton-Brown et al. reported that M. sinensis is used in Japan for forage and thatching, whilst M. sacchariflorus is used in China in the cellulose industry [6]. In Europe, M. x giganteus has been widely studied since 1983 for combustion to produce heat and electricity [3]. In the USA, agricultural energy sources, including herbaceous feedstocks, are expected to contribute more than 800 million tons annually to the US biomass industry for the production of liquid transportation fuels by the year 2030 [10]. M. x giganteus will likely be an important component of this biomass owing to the productivity it has exhibited in early studies.

M. x giganteus taxonomy & genetics

Andersson first described Miscanthus in 1885 [5]. Grasses in this genus are perennial and native to southeastern Asia through China and Japan, and also to Polynesia and Africa [3,5,6,11,201]. Miscanthus is in the Poaceae family, Panicoideae subfamily and Andropogoneae tribe [11,12]. The Andropogoneae tribe also includes economically important grasses in the Saccharum, Sorghum and Zea genera [11]. In fact, Miscanthus is closely related to Saccharum and is known to hybridize with it and other grass species [12,13].

The number of species in the Miscanthus genus has been variably listed from 11 to 25 [3,5,6,11,201]. More recently, Clifton-Brown et al.[6] wrote that based on the genetic works of Chen and Renvoize [14], Ibaragi [15] and Ibaragi and Oshashi [16], there are 11–12 species in the genus, all with a basic chromosome number of 19.

The classification and identification of Miscanthus species and varieties have relied heavily on phenotypic characteristics, observed by those working in the nursery industry. Much of the grouping of taxa has been carried out by use of morphological information that has been collected over the years, as well as information obtained from its relatedness to Saccharum[12,13,17]. One type, M. x giganteus, has been identified for its potential as a biomass feedstock crop. This hybrid is the offspring of a diploid M. sinensis (2n = 38) and tetraploid M. sacchariflorus (4n = 76) [5,6,17,18]. The sterile hybrid, M. x giganteus, has been placed into its own taxonomic classification based on a combination of growth and developmental characteristics that other Miscanthus species are yet to exhibit [5]. M. x giganteus is an allotriploid with a chromosome number of 3n = 57 [18]. Currently, limited information is available about the genetic diversity within the genus, but as interest in its use as a biomass feedstock increases, so will the research focused on identifying the genetic similarities [9,19]. Some of the difficulty in characterizing the taxonomy of Miscanthus is due to its polyploid nature and its ability to hybridize with other species within the genus.

At the morphological level, studies have been conducted on Miscanthus to help identify and classify the species into different taxa. Most taxonomic studies only take into account the length of the inflorescence axis, length of the racemes, disposition of spikelets on the axis, nerves of glumes, dorsal hairs of glumes and the presence or absence of awns [5]. However, these studies fail to use all possible characteristics for assessing taxonomic relationships. A study conducted by the Trinity College Botanic Garden in Dublin, Ireland used 31 traits to help identify and classify 83 specimens [5]. In total there were seven species of Miscanthus used in the study and ten cultivars/hybrids. The end result of the study demonstrated that there was a clear separation between M. sinensis and M. sacchariflorus, while the other species that were included in the study fell into clusters closer to M. sinensis, but with unclear separation [5].

Research has also continued at the molecular level to help delineate between the different cultivars and varieties, but many of the techniques have not been able to make a clear separation between the groups. For example, a study conducted by Hodkinson et al. using DNA sequences from internal transcribed spacers of ribosomal DNA [9] was able to distinguish interspecies relationships, but could not further identify differences within the species and cultivars. DNA markers have also been used for amplified fragment length polymorphism fingerprinting to detect differences in Miscanthus species [9]. Even though this study used a limited number of species for sampling, it was able to prove that markers could be used to detect differences within species and cultivars. One study that has shown promise used both amplified fragment length polymorphism and inter-simple sequence repeats to look into the genetic diversity that lies within the Miscanthus genome [9]. From this study it was seen that M. x giganteus was equidistant from M. sinensis and M. sacchariflorus. This study was also able to detect variations between the different M. sinensis varieties that were tested, whereas no variation was found within the different M. x giganteus specimens tested.

Some species that were within a given taxa were reclassified due to the lack of variation between them, which helps emphasize the fact that information based solely on phenotypic and morphological observations cannot clearly differentiate the species [9]. Human error is more prevalent in identifications based on visual recognition.

M. x giganteus growth & development

▪ M. x giganteus C₄ photosynthetic potential

M. x giganteus is a perennial grass that utilizes the C₄ photosynthetic pathway. However, it is unique among C₄ species that are typically susceptible to damage at cold temperatures [20], because it retains high photosynthetic activity at low temperatures and remains highly productive in cold climates [3,4,21,22]. Owing to its high productivity across a variety of conditions, M. x giganteus has been grown successfully from the Mediterranean climates of Spain to as far north as Scandinavia [23]. Naidu et al. found that while Zea mays showed depression in photosynthetic rates and reductions in pyruvate phosphate dikinase and the Rubisco large subunits when transferred to chilling conditions (14/10°C day/night), M. x giganteus was unaffected on both accounts [23]. Notably, M. x giganteus is able to develop photosynthetically active leaves at 8°C [23].

There are three C₄ photosynthesis subtypes – the phosphoenolpyruvate carboxykinase pathway, the NADP-malic enzyme (NADPME) pathway and the NAD-malic enzyme (NAD-ME) pathway [24]. As a member of the NADPME pathway group, the efficiency of photosynthetic CO₂ uptake in M. x giganteus is further increased. The NADPME pathway reduces CO₂ leakage from the bundle sheath to the mesophyll cells [25–27].

C₄ species characteristically demonstrate improved efficiency in nitrogen (N) and water-use [28,29]. Specifically, C₄ species can show N-use efficiencies twice those of C₃ species. Use of the NADPME pathway further improves the efficiency and productivity of M. x giganteus because photosynthetic N-use efficiency is favored owing to the higher kcat of Rubisco; that is, in species utilizing the NADPME cycle, a given Rubisco enzyme molecule operates more efficiently than in the NAD-ME and phosphoenolpyruvate carboxykinase cycles. Therefore, NADPME species require fewer Rubisco enzymes (and N) in order to achieve a given photosynthetic rate leading to a higher photosynthetic N-use efficiency of the overall system [30].

In terms of water-use efficiency, Beale et al. determined that M. x giganteus achieved a water-use efficiency – reported as the ratio of above-ground biomass to water loss – of 9.5 g kg^-1 under rain-fed conditions, a high value typical of C₄ species [31]. Furthermore, it has been demonstrated that M. x giganteus productivity increases with increased precipitation [32], even though it should be noted that total water-use rate is significant owing to high crop productivity. Into more agronomic terms, Pyter et al. determined that a minimum of 11mm precipitation for each ton per ha is necessary for successful biomass production [33]. Finally, Beale et al. showed that water-use efficiency of M. x giganteus, as measured by dry biomass production, was greater when grown in rain-fed conditions of greater soil water deficit conditions than when grown in irrigated settings with lesser soil water deficit [31].

▪ M. x giganteus photosynthetic productivity

Crop productivity is determined as the product of total solar radiation incident on an area of land, and the efficiencies of interception, conversion and partitioning of that sunlight energy into plant biomass [34]. Under field conditions, M. x giganteus achieves peak CO₂ assimilation levels of 40 µmol m^-2s^-1 [3,22].

Efficiency of light interception (ε_i) is largely dependent on canopy architecture, leaf-area index [35] and duration of the growing season [36,37]. M. x giganteus achieved an ε_i of 0.80 under midwestern US conditions, or 0.51–0.55 when calculated over the entire year [38]. At the same site, small plot trials achieved an ε_i of 0.72 [2], while trials in Europe achieved an ε_i of 0.32–0.83 when calculated over the growing season [36]. Heaton et al. obtained conversion efficiencies of photosynthetically active radiation (PAR) to above-ground material (ε_c,a) of 0.075 in the midwestern USA [2], while studies by Dohleman and Long [37] and Dohleman et al. observed an ε_c,a of 0.039–0.045 on large plot trials [38]. Beale and Long demonstrated in field trials in southeastern England that ε_c,a was 0.050–0.060, 39% above the maximum value observed in C₃ species [36]. Furthermore, when ε_c is calculated in terms of total (i.e., above-ground and below-ground) M. x giganteus biomass production (ε_c,t), it reaches 0.078 [36], which approaches theoretical maximum of 0.1. Studies performed in the midwestern USA by Heaton et al. reported a similar efficiency of intercepted PAR (0.075) [2]. Finally, early closure of the M. x giganteus canopy and an extended growing season allow for increased incident PAR [36,37] and, therefore, increased overall productivity.

▪ M. x giganteus growth cycle & longevity

As a perennial grass, M. x giganteus generally reaches maximum productivity in its third year of production [39]. During the establishment year, M. x giganteus can be planted from rhizomes or micro-propagated plantlets in April or May [40]. In the first year, the plant does not achieve maximum yield (Figure 1). After establishment, shoots emerge from the below-ground buds in April under midwestern US conditions [33]. By May, established plants reach nearly 2 m [33], and canopy closure occurs by late May to early June [2,37], enabling the grass to outcompete weeds and making weed control unnecessary [38]. Stand maturity is typically achieved after 2 or 3 years [41](Figures 2 & 3). Mature plants grow 3.5–4 m tall [23,40,42], with root structures reaching approximately 1.8 m deep [23]. In midwestern USA, maximum biomass is achieved in August [2], and the grass then flowers in mid-September to early October, although no viable seeds are produced [34]. Full plant senescence begins after a killing frost occurs as early as September through mid to late October, although senescence of the lower leaves coincides with full radiation interception in central Illinois, USA [33,40].

M. x giganteus should be harvested following full senescence, and harvestable dry-matter yield declines with the time left in the field beyond full senescence [43]. Heaton et al. determined that the majority of N translocation from above-ground biomass to below-ground biomass is achieved by late autumn [44], and M. x giganteus should be harvested at that point in order to avoid weather-related biomass losses over the winter [33]. In Europe, Lewandowski and Heinz [43] and Long and Beale [45] reported further N reductions in above-ground materials, and consequent increases in below-ground concentrations, by postponing harvest until late winter (Figure 4).

With senescence, M. x giganteus recycles N and other nutrients to the roots and rhizomes and the canes may be harvested at maximum dry matter, allowing for the primary harvest of carbon and hydrogen and sustained growth from the rhizome mass for future years [31,46,47]. Beale and Long present an in-depth analysis of N, phosphorous and potassium relations in a mature stand of M. x giganteus grown under European conditions [46]. All three nutrients demonstrated a similar pattern in above-ground material with values being highest early in the season, and lowering as they were diluted by additional carbon as the plant grew and, finally, declining drastically as the canopy senesced. In below-ground material, M. x giganteus showed less variance and an overall pattern of decline from emergence to mid-summer with a subsequent steady increase until February.

While a search of the literature was unable to find the ages of managed stands of M. x giganteus in Asia, Stewart et al. wrote that managed fields of M. sinensis persisted between 5 and 40 years when grazed in Japan [19], and Clifton-Brown et al. indicated that cut stands of M. sacchariflorus have been productive for 30 years when used by the cellulose industry [6]. The oldest European plantation of M. x giganteus is a 25-year stand in Denmark and the estimated lifetime of a plantation is 25 years [42]. In Illinois, the oldest stands were planted in 2002 [1] and crop monitoring is ongoing to gauge stand productivity over time.

M. x giganteus propagation & establishment

▪ Rhizome propagation

Owing to its sterility, M. x giganteus must be propagated vegetatively. The great challenge is to propagate it both efficiently and economically. Currently, most M. x giganteus is propagated by dividing the underground stems (rhizomes). These stems form nodes, internodes and buds similar to above-ground stems (Figure 5). M. x giganteus rhizomes also serve as an underground overwintering storage organ for the plant and the source of each year’s initial above-ground growth. Each spring, new shoots emerge from rhizome buds and use the reserves stored within the rhizome to initiate growth.

M. x giganteus rhizome propagation can be performed either in the late fall after plants have senesced and biomass removed, or more commonly, in the early spring prior to emergence [48]. However, propagation is not without challenges. The primary challenge is identifying practical and efficient methods of lifting and replanting M. x giganteus rhizomes. Early US research used rhizomes harvested by hand. In a study by Pyter et al., 50 g rhizomes yielded clumps of M. x giganteus rhizomes, which produced seven to ten harvestable rhizomes from 1-year-old plants and 25–30 rhizomes from 2-year old plants when harvested by hand [49]. Pyter et al. also demonstrated similar biomass yields when rhizomes of 20–25 g and rhizomes of 40 g were planted [50]. The age of the M. x giganteus mother plants also appears to be a factor in propagation and establishment success. There were significant differences in successful rhizome-clump production among 1-, 5-, and 9-year old mother M. x giganteus plants [51]. The 5-year-old plants fared much better, with 88% producing successful clumps, compared with 52 and 25% for the 9- and 1-year-old stands, respectively.

▪ Mechanically harvesting M. giganteus rhizomes

As hand harvesting is not a practical method for large-scale commercial propagation, various mechanical rhizome harvesters (e.g., Tomax Ltd., Portlaw Co. Waterford, Ireland) have been developed to produce an efficient process. The origins of these machines vary widely and include harvesters used traditionally in other crops, such as the potato or bermudagrass sod industries. Modern equipment can harvest rhizomes much more efficiently than by hand with the trade-off being that mechanically harvested rhizomes are typically smaller than those harvested by hand.

▪ Tissue culture

Micropropagation of M. x giganteus is a promising propagation option due to its ability to produce large numbers of plants in a relatively short amount of time. However, challenges remain. A detailed study by Lewandowski examined the differences in plant development between rhizome-propagated and microproagated (tissue culture) plants. Results demonstrated rhizome-propagated M. x giganteus had fewer but larger shoots, and thicker rhizome branches than M. x giganteus produced from tissue culture. The study also found that plants micropropagated through somatic embryogenesis developed faster than those propagated by in vitro tillering. However, all differences between rhizome-propagated and micropropagated plants diminished over time [48].

▪ Rhizome-derived plugs

Another propagation method that is gaining favor in the developing US feedstock industry is via rhizome-derived plugs. To date, no reviewed literature has analyzed this propagation and establishment method. In practice, small rhizome pieces are planted into pots approximately 3 cm in diameter and 15 cm deep. These small pots are placed in greenhouses, high-tunnels or other protected areas until the grass becomes well rooted and has developed adequate shoots to support in-field development. Following establishment, the plugs are transplanted into the field using mechanical transplanters similar to those used to plant small vegetables, nursery crops or tobacco plants. This method is proving to be a preferred method of planting because the grasses are already growing, reducing the risk of failed survival and/or establishment when planting rhizomes (Figure 2). These actively growing plants are vulnerable to dry weather, however, and irrigation should be available to ensure survival and establishment.

▪ M. x giganteus rhizome storage

M. x giganteus rhizomes can be successfully stored at 4°C for extended periods of up to 4 months [50]. However, other studies show decreased viability of up to 50% following storage [52]. Ongoing research is also evaluating the moisture levels in rhizomes necessary to support establishment following storage and planting. Preliminary results of this work indicate that a moisture threshold within the rhizome of approximately 50% or higher may be necessary for optimal planting success.

▪ Planting M. x giganteus rhizomes

Prior to planting, the soil is prepared by tilling or cultivating into a smooth, uniform planting bed. Equipment used to plant M. x giganteus rhizomes originated with planters used for other vegetatively propagated crops, such as potatoes or horseradish. Rhizomes are typically planted 10 cm deep [50]. Lewandowski et al. reported planting densities in Europe of one to four rhizomes per m² (10,000–40,000 rhizomes per ha) [3]. Pyter et al. indicated that planting densities of approximately 10,000–12,000 rhizomes per ha were successful in Illinois, USA [49]. Finally, Lewandowski et al. noted that irrigation during the planting year improved establishment rates [3].

After planting, not all rhizomes will sprout and develop into plants, and establishment efficiency can vary widely. One study showed a variation of between 50–95% emergence [52]. Pyter et al. reported 60–70% emergence among hand-harvested rhizomes. Increased planting densities can help alleviate problems with establishment [49].

▪ Cold tolerance

Once planted, survival of first-year M. x giganteus is highly dependent on the environment. In addition to competition from weeds and pests, cold tolerance and over-winter survival of first-year stands is also a concern in temperate areas with cold winters and little snow cover. Clifton-Brown and Lewandowski [53] and Clifton-Brown et al.[54] examined first-year cold tolerance, and their results indicate a major risk to viability when soil temperatures drop below -3°C at the 5-cm soil level, with lethal rates of up to 50% [53]. First-year plantings at the University of Illinois (USA) were severely damaged by a cold winter in 2008–2009. However, this appears to be primarily a problem with first-year plants, as Pyter et al. reported good survival of established M. x giganteus with winter air temperatures dropping as low as -29°C [49].

Agronomy of established M. x giganteus stands

▪ Fertility

The effect of fertilization on M. x giganteus yields varies widely based on location, and also varies from study to study. Almost all studies to date have focused on N fertilization. In three studies, N fertilization has had little or no effect on the yields of M. x giganteus, and in each case, this was attributed to fertile soils. In Austria, Schwarz et al. showed no yield increase in a third-year planting with abundant N fertilization (180 kg N ha^-1) [55]. Christian et al. found no response to N fertilization at England’s Rothamsted Research Farm (UK) after 14 years [56]. Finally, in West Germany, Himken et al. also found no effects from N fertilization in a fourth-year planting [57]. Conversely, in Italy, Ercoli et al. demonstrated significant increases in yield of up to 9.8 mg ha^-1, with 180 kg N ha^{- 1} averaged over 4 years in an irrigated planting [58]. Clifton-Brown et al. demonstrated increases in yield owing to N fertilization in some years, but not in others, indicating a potentially significant climatic effect [59]. These mixed results demonstrate the difficulty in recommending fertilizer rates and indicate a need for a more in-depth examination into the soil–plant relationship, and how nutrient supply affects yield. Whatever the effect, it is clear that any nutrient response is highly dependent on the soil type and planting location.

▪ Weed control in M. x giganteus

M. x giganteus competes poorly with weeds during the establishment phase, making weed control mandatory [3,60,61]. Yields of herbaceous perennial species can be reduced by weed growth through resource competition (water, nutrients, light and space), and also through the production of allelochemicals [62]. Mechanical, cultural and chemical weed-management practices are all options at various points during the establishment period.

Mechanical and cultural methods of weed control in M. x giganteus have included: using a rotary hoe between rows several times in the second year [55], cleaning rhizomes of loose soil before planting [63], cleaning tillage and planting equipment, timing planting to avoid emergence periods of problematic weeds, minimizing the weed-seed bank population through consistent weed control in prior years, and either banding fertilizer or foregoing fertilizer applications when planting and harvesting M. x giganteus only once each year at the recommended time [62]. After the second growing season, the canopy generally closes early in the season, reducing the effect of weed competition until the first killing frost.

In preparation for planting a perennial crop such as M. x giganteus, it is necessary to control existing weeds, especially perennials. This can be accomplished with systemic, nonselective herbicides such as glyphosate [64]. An application of paraquat, applied at first-shoot emergence, has been used effectively in mature M. x giganteus as its desiccated shoots will be quickly replaced [65].

In the USA, few herbicides are registered for use in ornamental plantings of Miscanthus spp., and no herbicides are currently registered for use in the biofuel planting of M. x giganteus. A summary of several herbicides safely used in M. x giganteus is presented in Table 1. Various preemergence (PRE) and postemergence (POST) herbicides have been used in the EU for weed control, and it is generally presumed that herbicides used in corn are safe on M. x giganteus[3,64]. Venturi et al. recommended herbicide applications during the first season and after each harvest [66]. Bullard et al. listed several herbicides successfully used in the EU in M. x giganteus production [65]. Huisman et al. suggested atrazine in the first year and glyphosate PRE the second year for effective weed control in M. x giganteus[64]. Prodiamine and aminopyralid were used experimentally in first and second year M. x giganteus plantings in Kentucky, USA with no phytotoxic response [David Williams, University of Kentucky, Pers. Comm.]. Anderson et al. found that several PRE and POST herbicides produced little or no phytotoxic response in M. x giganteus in a greenhouse environment at a 1 × application rate [67]. They also found that PRE herbicides, atrazine, isoxaflutole, pendimethalin and S-metolachlor and POST herbicides, bromoxynil, dicamba and mesotrione plus atrazine, were all safely applied to M. x giganteus under field conditions. Nonweeded check plot conditions significantly reduced the number of tillers per plant and above-ground biomass production, confirming the need for weed control during establishment.

▪ M. x giganteus pests & pathogens

Very few insect and other invertebrate pests have been found to infest M. x giganteus[68] and, to date, no reports of yield reduction have been cited. Christian et al. observed larvae of a moth, Mesapamea secalis L., feeding on M. x giganteus tissues in the spring, but final stem density did not appear to be affected [69]. Huggett et al. found that the oat or wheat aphid (Rhopalosiphum padi) could not complete its life cycle on M. x giganteus. However, Rhopalosiphum maidis, the corn leaf aphid, survived, was highly fecund and was able to transmit barley yellow dwarf virus (BYDV) [70]. This is especially troublesome because Miscanthus spp. can carry BYDV with or without showing symptoms. Hurej and Twardowski also observed negligible numbers of R. padi on first-year M. x giganteus plants [71]. Fall armyworm, Spodoptera frugiperda, was found to infest the whorls of M. x giganteus in field plots [68]. The yellow sugarcane aphid (Sipha flava) and corn leaf aphid were also found in M. x giganteus plantings in Illinois, Indiana, Kentucky and Nebraska, USA [72].

Western corn rootworm (WCR; Diabrotica virgifera virgifera LeConte) is a major pest in maize. WCR is able to complete its life cycle on M. x giganteus, although it tends to favor maize with respect to rate of development, adult dry weight and proportional emergence rates [73]. However, it is unclear whether these findings support the likelihood of the potential benefit of M. x giganteus being considered a refuge in insect resistance management or the potential problem of it being a reservoir for larger numbers of WCR that would, in turn, infect maize crops.

In addition to potential insect pests, two species of Xiphinema (Xiphinema americanum and Xiphinema rivesi) and one species of Longidorus (Longidorus breviannulatus) nematodes were detected in soils surrounding M. x giganteus roots at several sampling sites in midwest USA. [74]. High numbers of L. breviannulatus appeared to destroy fibrous roots and stunt lateral roots.

Although no diseases have been found that greatly affect M. x giganteus production [60], certain pathogens such as Fusarium spp. [3], BYDV [75,76], Helminthosporium spp. and Drechslera spp. [77], and Miscanthus blight (Leptosphaeria spp.) [78] have been found on Miscanthus spp., including M. x giganteus. Ahonsi et al. observed leaf blight caused by Pithomyces chartarum on 100% of newly established and 2-year old M. x giganteus in research plots in Kentucky, USA [79]. The blight killed some leaves and tillers, and the researchers suggested that the observed level of disease severity could affect M. x giganteus production [79]. Agindotan et al. have isolated sugarcane mosaic virus on M. x giganteus with a new method for detecting RNA viruses [80]. Tsukiboshi et al. identified Ephelis japonica as the fungal pathogen that causes black choke disease known to infect M. floridulus (Labill.) and M. tinctorius Hack. (Eulalia), although it is not known if M. x giganteus is a host for this disease [81].

▪ Biomass harvesting

Harvesting technology for M. x giganteus is currently an active area of research, but very little work has been published to date. Typically, modern hay-harvesting equipment, including cutters, conditioners and balers, is suitable for the harvest of M. x giganteus in many situations (Figure 6). However, the equipment must be operated more slowly than in hay crops owing to the density and toughness of the M. x giganteus stems. Biomass is typically harvested in late fall or winter, after the biomass dries in the field. Heaton reported that biomass moisture levels decreased from near 50% in October to below 10% in February; bales of biomass stored under cover can remain in good condition for at least 3 years [Heaton EA. The comparative agronomic potential of M. x giganteus and Panicum virgatum as energy crops in Illinois. PhD Dissertation. University of Illinois at Urbana-Champaign, USA (2006)].

▪ Removing M. x giganteus

Currently, there are no published experiments investigating M. x giganteus removal. M. x giganteus is susceptible to herbicides such as glyphosate [82] and fluazifop-P [64], and rhizomes are easily destroyed with plowing [83]. Speller suggested that several applications of glyphosate or fluazifop-P followed by fall tillage would control M. x giganteus[63]. Fall applications followed by spring applications of glyphosate significantly reduced M. x giganteus shoot number and above-ground biomass the following summer [Anderson EK. Herbicide phytotoxicity response and eradication studies in Miscanthus × giganteus. M.S. Thesis. University of Illinois at Urbana-Champaign, USA (2010)]. Anderson also found that spring tillage with at least one glyphosate application significantly reduced shoot number and aerial biomass in the same season. However, significant regrowth occurred following both experiments, indicating the need for control measures to be employed over more than one year [Anderson EK. Herbicide phytotoxicity response and eradication studies in Miscanthus × giganteus. M.S. Thesis. University of Illinois at Urbana-Champaign, USA (2010)]. In a separate study, treatments involving a spring application of glyphosate (2.25 kg acid equivalent ha^-1) followed by mowing caused the greatest reduction in M. x giganteus shoot number (92%) when measured in the fall of the same year, followed by weekly mowing (73%) [E Anderson, Unpublished Data].

M. x giganteus modeled & observed yields

The majority of the literature reporting dry biomass yield for M. x giganteus originates from European studies. Ceiling peak biomass yields in established stands of M. x giganteus have approached 40 t dry matter (DM) ha^-1 in some European locations, although it may take 3–5 years to achieve these ceiling yields [84]. Across Europe, harvestable yields of up to 25 t DM ha^-1 from established stands of M. x giganteus have been reported in areas between central Germany and southern Italy, while peak yields in central and northern Europe have ranged between 10–25 t DM ha^-1, and in excess of 30 t DM ha^-1 in southern Europe [3]. A quantitative review of established M. x giganteus stands across Europe reported a mean peak biomass yield of 22 t DM ha^-1, averaged across N rates and precipitation levels [1].

▪ M. x giganteus growth & yield models

The development of crop models is a means to synthesize research, as it relates to genetics, the environment and physiology; and can aid in policy-making decisions in the areas of emphasis projected by the model, such as potential yield across the landscape. However, crop growth and productivity models have limitations depending on the simplicity or complexity of the model [85]. Empirical models predominantly rely on observed experimental data, while mechanistic models rely heavily on processes necessary to describe a particular system. Mechanistic, or process-based models, inevitably require more detail for parameter estimation, yet can provide valuable insight into particular aspects of crop growth, such as predictions of carbon assimilation, growth and yield [86].

Several productivity models have been developed for predicting and simulating growth and yield of M. x giganteus. The majority of the predictions and validations of these models have been developed for regions and environments across Europe. Table 2 shows predicted dry biomass yields (ton DM ha^-1) simulated from these models and observed yields (ton DM ha^-1) obtained from these studies.

The first of these models, developed and described in Clifton-Brown et al., is an empirical model parameterized for M. x giganteus. This model was developed by European researchers to predict above-ground dry matter production, which assumes water and nutrient supplies to be nonlimiting [87]. We refer to this model as the Clifton-Brown-2000 model to distinguish it from an improved development of the model, MISCANMOD, which predicts yield under rain-fed, or otherwise potentially water-limiting conditions. The model developed by Clifton-Brown-2000 has been used for predictions in other studies [1] and expanded to contain improvements [88].

Clifton-Brown et al. explain improvements made to the Clifton-Brown-2000 model, which was modified to predict dry biomass yields of various genotypes of Miscanthus, including M. x giganteus, under rain-fed conditions across Europe. This development, MISCANMOD, has been used in several studies to predict M. x giganteus growth and yield across Europe and in the USA [88–90].

Stampfl et al. used MISCANMOD to describe Miscanthus growth under several EU land-use scenarios in order to make predictions of electricity generation [88]. In this study, European monthly climate data were incorporated into the model to predict peak yield (Fall yield), and harvestable yield, calculated as a 33% reduction from peak yield [91]. This reduction is described as the percentage of biomass that is lost or reduced from the time that peak yield is achieved in autumn or at the end of the growing season, to the time that the crop is actually harvested, generally that same winter or the following spring. Modeled harvestable yields in Table 2 for this study are a combination of data under two different land-use scenarios.

Heaton et al. applied the Clifton-Brown-2000 model to predict yields across varying locations in Illinois, USA. In this application of the Clifton-Brown-2000 model, Heaton et al. used climate data from representative sites across the state to model peak biomass yields [1]. At the time of this study, no published data from Illinois M. x giganteus plantings were available for yield validation. However, published data from field studies in Illinois have since shown M. x giganteus mean peak and harvestable yields across the state to be 38.2 and 29.6 t DM ha^-1, respectively [2]. Dohleman and Long also reported peak biomass yields reaching 29.5 and 30.3 t DM ha^-1[37].

Yields projected by MISCANMOD in Illinois (USA) vary across the state: 30–34 t DM ha^-1 in northern Illinois, 33–36 t DM ha^-1 in central Illinois, and 37–42 t DM ha^-1 in southern Illinois [88], which are similar to observed data for Illinois regions reported in Heaton et al.[2]. These projections are considered under the constraint of rain-fed production, which is typical for Illinois crop production. Interestingly, these yield predictions lie within the range of yield predicted by Heaton et al., using the Clifton-Brown-2000 model, which assumes nonwater- and non-nutrient-limiting growing conditions [1].

A long-term study in Ireland showed that harvestable yield can change over time as the crop ages [59]. The authors described three phases of growth (peak Fall yields) as: phase 1 – ‘yield building’, years 2–4 (8.3 ± 0.6 t DW ha^-1); phase 2 – ‘stable yield’, years 5–11 (17.2 ± 0.46 t DW ha^-1); and phase 3 – ‘reduced yield’, years 12–16 (11.4 ± 0.38 t DW ha^-1). MISCANMOD predictions were similar to the calculated observed yields during the ‘stable yield’ phase, but it overpredicted yield during the ‘reduced yield’ stage, as it was unable to account for the factor or factors that caused the stand to lose vigor. Observed harvestable yield, calculated peak yield, and modeled peak yields following establishment years have been digitized from this study and are shown in Table 2.

Hastings et al. describe MISCANFOR, an improved version of MISCANMOD [92,93]. The improvements allow the application of the model to be extended to different environmental conditions, such as predictions of drought or frost conditions that may affect or kill existing stands of Miscanthus. Hastings et al. reported that the improvements implemented into MISCANFOR significantly reduced the variation between modeled and observed predictions previously obtained from MISCANMOD. They further suggested that robust predictions of M. x giganteus across Europe can be obtained using MISCANFOR. Combining yield estimates from MISCANFOR with estimates of soil organic carbon, M. x giganteus shows the potential to sequester 2–3 t C ha^-1 year^-1 in Ireland [94].

Another model (referred to as Ritcher-2008) and developed in the UK describes the impact of water supply during the growing season on M. x giganteus yield, which can vary greatly across sites and regions [95]. In their study, yields from established stands of M. x giganteus at 14 sites averaged 12.8 t DW ha^-1. They concluded that yield variability can largely be explained by soil-available water, which for estimating purposes requires regional soil and geological information.

Most of the previously described models have been driven by light interception and leaf area, radiation-use efficiency and temperature thresholds based upon principles established by Monteith [34]. A semi-mechanistic model has been developed [85], and parameterized for M. x giganteus and is based upon a previous plant production model, Windows Intuitive Model of Vegetation Response to Atmospheric and Climate Change [96]. This model is referred to here as the Miguez-2009 model, which considers important physiological processes combined with climate data to predict plant growth. Yields predicted from the model at several locations were adequate in their agreement with observed data, while other predictions differed from observed data for proposed reasons described by Miguez et al.[86]. In general, this model possesses great potential for simulating growth and yield and for estimating photosynthetic rates and CO₂ assimilation. However, because this model is process based, it requires more detail than previously described models and partitioning coefficients must be modified for different environments [86]. Improvements in this model will benefit greatly from additional field research, so that parameter estimates can be improved and fine-tuned to adequately predict growth and yield in a wide range of environments in Europe and the USA.

▪ M. x giganteus yields in the USA

In the USA, published field research related to M. x giganteus growth and yield is limited to Illinois [2,37,50] and Kansas [97,98]. Table 3 summarizes biomass yield data from these states. Current research is in progress and is being expanded to locations throughout the USA, including the Midwest, Great Plains and Atlantic Coast regions. Data from these studies will provide valuable information that can be used to improve current models. Such data include information about responses to new environments, growth data across (i.e., planting to establishment) and within seasons (emergence to harvest) and response to N fertilizer. These data will also provide information regarding the adaptation region for growing M. x giganteus in the USA. One concern is the ability of this crop to withstand harsh winter environments (low and fluctuating winter temperatures) and areas of low precipitation or prolonged periods of drought. M. x giganteus possesses winter hardiness traits obtained from M. sinensis[54]. However, first-year plantings of M. x giganteus in east-central Illinois and west-central Indiana resulted in poor winter survival, while established stands in that same region were largely unaffected [Unpublished Data]. The risk of growing this crop in some environments may be confined primarily to the establishment years, providing that ample time is available to develop deep root and rhizomes systems for storing carbohydrate reserves that can contribute to winter survival. In northern Europe, artificial testing has shown that M. x giganteus rhizomes are severely affected by temperatures below -3.4°C [53]. In addition, M. x giganteus has little tolerance to drought or capacity to survive under low precipitation environments, and it has been shown that it will begin to senesce under water-stressed conditions [99].

Potential for M. x giganteus invasiveness

Many have asserted that since M. x giganteus is a sterile triploid hybrid that does not produce viable seed, it poses no risk as an invasive weed [100,201]. M. x giganteus has been grown in experimental and commercial plantings in the EU for over 20 years with no evidence of invasiveness [201]. Possible sources of propagules for M. x giganteus as a weed would originate in past production fields or borders of production fields [63], accidental dispersal during transport, equipment transmittance, or extreme weather events. Harvey and Hutchens noted that M. x giganteus is not likely to be a weed problem and that plowing around M. x giganteus fields would be an effective control [82]. Although other Miscanthus spp. are known to have self-seeding invasive potential, M. x giganteus was identified as a plant that may be a good horticultural selection owing to its non-invasive nature, as 0% of its seeds germinated [101]. The European Miscanthus Improvement project has recommended that any new genotype of Miscanthus spp. be sterile (i.e., triploid) to avoid potential invasiveness issues [3].

However, there are many reasons to be concerned about introducing new plant species into the USA for bioenergy plantings. Most attributes of an ‘ideal’ energy crop are unfortunately also common to many invasive weed species [102,103]. The sterile cordgrass hybrid, Spartina townsendii, underwent faulty mitosis to produce the fertile and invasive allopolyploid S. anglica, showing that present sterility is no guarantee of continued sterility [202]. Giant reed (Arundo donax) has also become invasive in California waterways via rhizome fragmentation [104].

Before introducing a species into a new habitat, an analysis should first be made of the likelihood of that species becoming invasive. The Australian-based Weed Risk Assessment designed for Australia and New Zealand by Pheloung et al.[105] has been modified, used and validated around the world [102,106,107] for predicting invasiveness. Even though M. x giganteus was rated an unlikely invader by the Weed Risk Assessment, Barney and DiTomaso recommended that it should still go through a pre-introduction screening through ecological analyses [102].

In summary, although M. x giganteus is not likely to become invasive in noncrop areas, herbicides are available that will likely control its growth. Glyphosate and tillage employed at various timings will eradicate M. x giganteus in a production field. However, applications will need to be repeated throughout a growing season and over more than one year to fully control the root and rhizome structures.

Future perspective of M. x giganteus use in the USA

At present in the USA, there is much ongoing research focused on developing renewable energy sources. One source of renewable energy will be herbaceous feedstocks, and current studies have been triggered by the need to produce large amounts of renewable energy from these herbaceous feedstocks in the most environmentally positive manner whilst using the fewest land resources possible. Since past studies of M. x giganteus have produced promising outcomes, improving our understanding of M. x giganteus genetics, growth and development; and interactions with the environment and management schemes are important areas of study.

M. x giganteus has only been seriously studied in the USA since 2000. As there are several herbaceous and woody feedstocks being considered, and one US goal for producing renewable energy using feedstocks is to not interfere with food or livestock production, it is important to determine the environments in which M. x giganteus is best suited to being produced, and the management required to achieve optimal yields. While regional studies have increased greatly since 2000 and can be found in more than 11 states at present, there is still a need for more studies that will evaluate optimal areas for growing specific biomass crops. Moreover, because agronomic research has sometimes produced conflicting results (e.g., N-fertility requirements), it appears that local studies will be important to identify the optimal management schemes for growing locales.

Other agronomic needs for M. x giganteus, lie in better handling of its establishment, harvest and pest control. Propagating and establishing M. x giganteus has not reached the level of success of most widely planted agricultural crops. Developing reliable plant materials, establishment methods and planting equipment will be necessary if M. x giganteus is to become a commercial crop. While harvesting M. x giganteus has been accomplished using traditional forage equipment in the current small-scale plantings, dedicated machines will be necessary for the hundreds of thousands of hectares that will be raised to meet the energy goals. Finally, while there have been pests identified on US M. x giganteus, it remains to be seen if they will become a commercial problem. Research in pest-management areas is necessary in order to be prepared for handling and understanding weed, insect and disease pest problems as they arise.

It is likely that improved types of Miscanthus will be available in the future given the short time that geneticists and breeders have studied members of the genus. Currently, geneticists and breeders from both US universities and industry are involved in selecting and producing Miscanthus that are improvements of the horticultural M. x giganteus. Developing high-yielding types that are adapted to various US conditions and are easily established and harvested, without the environmental concerns of invasiveness or large demands for fertility or pest-control inputs, is most likely to be achievable.

In the USA, it is likely that a number of grasses will contribute to the production of liquid transportation fuels; for example, given the success of ethanol production in Brazil using sugarcane (Saccharum spp.) [108], this grass might be the optimal crop in the few US locations where it can be commercially produced. Furthermore, at present there is great interest in another Saccharum species, energy cane, for cellulosic ethanol production in southeastern USA, because of high biomass yields and greater cold tolerance than sugarcane [109]. While both of these canes will potentially produce more ethanol per ha, the future remains bright for M. x giganteus and other types of Miscanthus for biomass production. Grasses in the genus produce high biomass yields and can be successfully produced in much larger proportions of eastern USA than either sugarcane or energy cane. Moreover, there is much research being conducted that will likely improve our ability to efficiently manage and utilize M. x giganteus and other types of Miscanthus. As long as the USA places emphasis on ethanol production for liquid transportation fuels originating from cellulosic sources, these grasses will be studied and utilized.

Table 1.  List of several herbicides safely used on Miscanthus x giganteus, their mode/site of action and the highest rates tested without phytotoxic effects.

Herbicide	Mode/site of action	Highest rate tested without phytotoxic effects (kg ai ha^-1)	Ref.
Pre-emergence
Acetochlor	Cell division inhibitor	9.79	[68]
Atrazine	Photosystem II inhibitor	6.74	[68]
Imazethapyr	ALS inhibitor	0.07	[68]
Isoxaflutole	HPPD inhibitor	0.17	[68]
Pendimethalin	Microtubule assembly disruptor	6.40	[68]
Prodiamine	Microtubule assembly disruptor	0.73	[D Williams, University of Kentucky, Pers. Comm.]
S-metolachlor	Cell division inhibitor	7.14	[68]
Saflufenacil	PPO inhibitor	0.10	[68]
Saflufenacil + dimethenamid-p	PPO inhibitor, cell division inhibitor	0.12 + 1.10	[68]
Post-emergence
Aminopyralid	Synthetic auxin	0.27	[D Williams, University of Kentucky, Pers. Comm. ]
Atrazine	Photosystem II inhibitor	4.49	[68]
Bromoxynil	Photosystem II inhibitor	1.12	[68]
Clopyralid	Synthetic auxin	0.48	[65]
2,4-D ester	Synthetic auxin	1.06	[68]
Dicamba	Synthetic auxin	1.12	[68]
Dichlorprop	Synthetic auxin	3.33	[66]
Fluroxypyr	Synthetic auxin	0.36	[66]
Foramsulfuron	ALS inhibitor	0.02	[68]
Halosulfuron	ALS inhibitor	0.07	[68]
Isoproturon	Photosystem II inhibitor	2.00	[66]
MCPA	Synthetic auxin	3.75	[66]
Mecoprop-P	Synthetic auxin	3.60	[66]
Mesotrione	HPPD inhibitor	0.42	[68]
Metsulfuron-methyl	ALS inhibitor	0.006	[66]
Nicosulfuron	ALS inhibitor	0.02	[68]
Primisulfuron	ALS inhibitor	0.08	[68]
Tembotrione	HPPD inhibitor	0.04	[68]
Triclopyr + clopyralid	Synthetic auxin, synthetic auxin	0.31 + 0.10	[D Williams, University of Kentucky, Pers. Comm.]

ai: Active ingredient; ALS: Acetolactate synthase; ha: hectare; HPPD: Hydroxyphenylpyruvate dioxygenase, MCPA: 2-methyl-4-chlorophenoxyacetic acid; PPO: Protoporphyrinogen oxidase.

Table 2.  Miscanthus x giganteus modeled and observed yields of established stands of M. x giganteus obtained from studies using different models developed for predicting dry biomass yields and based on several different models.

Model	Region	Modeled yield (t DW ha^-1)				Observed yield (t DW ha^-1)		Ref.
		Peak (nonlimited)^†	Peak^‡	Harvestable^§		Peak^¶	Harvestable^#
Clifton-Brown-2000	Ireland, EU UK, EU	16–26 20–27				28		[88]
Clifton-Brown-2000	Illinois, USA	27–44						[1]
MISCANMOD	Austria, EU Belgium, EU Belgium, EU France, EU Germany, EU Greece, EU Ireland, EU Italy, EU The Netherlands, EU Portugal, EU Spain, EU Sweden, EU UK, EU	32.9–34.3 27.1 21.6 32.9–41.9 24.4–31.1 49.6–53.9 18 36.1–50.4 22.3–25.3 48.4–48.9 40.8 23 21.3–23.8	17.6–23.3 24.9 21.8 18.8–20.5 20.9–27.4 11.4–16.4 18.2 11–27.3 22.6–23 15.2–16.4 17.6 10.9 21.6–23.4			17–30 17 42–49 17–30 18 25 39 22 18–28	22 16 10 30 10–20 20–26 14 15–27 13–17 26–30 14 13 11–17	[90]
MISCANMOD	Denmark, EU France, EU Latvia, EU Lithuania, EU Poland, EU			10.6 14.8–21.1 13.8–14.4 16–16.7 17.6–19.9				[91]
MISCANMOD	Ireland, EU		8–23			9-20^††	7–16	[60]
MISCANMOD	Illinois, USA		30–42			32.5–50.8		[90]
MISCANFOR	Europe		16.3-18					[94]
MISCANFOR	Europe		15–15.5					[95]
Ritcher-2008	UK, EU			9.6			12.8	[97]
Miguez-2009	Austria, EU Denmark, EU Greece, EU Ireland, EU Italy, EU The Netherlands, EU		28.5 30.7 32.6–36.5 21.9–25.2 41.9 30.8			29.6 27.7 22.9–42.7 11.5 –18.6 42.3 21.5		[89]

Data are expressed as a range of predicted or mean observed values, or individual mean values of dry biomass yield, and have been obtained directly or extracted from the cited references using digitizing software.

^†Modeled irrigated or nonwater-limiting peak yield.

^‡Modeled rain-fed peak or end of season yield.

^§Modeled harvestable yield, harvested in winter or spring, or at least 2 months after end of growing season.

>^¶Observed peak yield, at or near end of growing season.

^#Observed harvestable yield, harvested in winter or spring, or at least 2 months after end of growing season.

^††Calculated observed peak yield, based upon normalized observed harvestable yield.

DW: Dry weight; ha: Hectare; t: Ton.

Table 3.  Dry biomass yields obtained from studies in Illinois and Kansas, USA.

Location (USA)	Peak yield *t DW ha^-1)		Harvestable yield (t DW ha^-1)			Notes	Ref.
	Seasons ≥ 3		Season 1	Season 2	Seasons ≥ 3
Shabbona, IL (northern) Urbana, IL (central) Simpson, IL (southern)	31.2 45.5 42.3				20.9 33.4 34.6	Average of seasons 3–5. Established in winter 2002	[2]
Champaign, IL (central)	29.9				17.9	Average of seasons 3 and 4. Established in 2005	[38]
Troy, Kansas (northeastern) Manhattan, Kansas (northeastern)			4.0 2.7	13.7 11.8		Established in 2002	[99]

Data are expressed as mean peak (yield at or near end of growing season) and mean harvestable yield (yield obtained during winter or once the plants became dormant).

DW: Dry weight; ha: hectare; t: Ton.

Miscanthus x giganteus:

Sterile hybrid of Miscanthus sinensis and Miscanthus sacchariflorus exhibiting great potential as a biomass feedstock.

Miscanthus sinensis:

Grass species with a number of distinct varieties and a parent of M. x giganteus.

Miscanthus sacchariflorus:

Vigorously rhizoma Miscanthus sinensis: Grass species with a number of distinct varieties and a parent of M. x giganteus. tous grass species and a parent of M. x giganteus.

Miscanthus:

Genus of approximately 11–12 grass species native to southeastern Asia through China and Japan, and also to Polynesia and Africa

Rhizome:

Underground, horizontal stem used to asexually propagate M. x giganteus.

Executive summary

	▪ Originally used in landscape plantings, Miscanthus x giganteus is now being studied owing to its ability to produce large amounts of herbaceous biomass.
	▪ M. x giganteus is a sterile triploid resulting from the cross of Miscanthus sinensis and Miscanthus sacchariflorus.
	▪ M. x giganteus is likely to be planted in temperate areas of the eastern USA because it can photosynthesize effectively at relatively low temperatures, requires relatively large amounts of moisture to be productive, and is harvested when dormant during winter.
	▪ Ongoing research is addressing agronomic questions regarding the propagation, establishment, culture and harvest of M. x giganteus, but additional study is needed to ensure sustainable, commercial-scale production of M. x giganteus as a bioenergy crop.
	▪ Models of M. x giganteus growth and potential yields exist for both Europe and the USA, but additional field-based data is necessary to improve the model accuracy.

Financial & competing interests disclosure

This work was primarily supported by the College of ACES Experiment Station, Department of Crop Sciences and the Energy Biosciences Institute at the University of Illinois, Urbana-Champaign, IL, USA. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.