3 Wheat

Download this chapter in PDF | Download citation

3.6 WME6—spring wheat at high latitudes


WME6 comprises a vast area of high-latitude environments in the Northern Hemisphere where winters are too cold for autumn-sown winter wheat to consistently survive, especially in situations where protective snow cover is not assured. Thus spring wheat is planted in April–May and harvested after mid August. The crop is rainfed and, in most locations, often water-limited. At the southern limits of the WME6 region—where the January mean temperatures reach as high as −8 °C to −5 °C—winter wheat is often found alongside spring wheat and, if there is no winter killing, will substantially out-yield spring wheat.28

Major production in WME6 is found in North America and northern Asia, with northern Europe making a smaller contribution. Western Canada and the northern Great Plains states of the USA are the major producing areas in North America (Map 3.6). In northern Asia, major producing areas include half of the Russian Federation wheat area largely in Siberia, as well as northern Kazakhstan (Map 3.7), and Mongolia and north-east China.

WME6 is of special interest because global warming would likely permit a northern expansion of this mega-environment. Warming and other factors favouring winter wheat—such as greater genetic winter hardiness and better snow trapping with conservation tillage—may at the same time lead to contraction of the WME6 southern boundary.

Trethowan et al. (2006) summarised climatic conditions across the major regions of WME6 in the world and characterised leading wheat varieties from each, but did not discuss yield progress. Fortunately there are some recent detailed studies of progress from North America as well as Finland and Western Siberia.

Map 3.6

Map 3.6 Major spring and winter wheat regions of the USA and Canada. Source: Map developed from USDA CropScape and Monfreda et al. (2008)

Map 3.7

Map 3.7 Major spring and winter wheat regions of Europe and north-western Asia. Source: Crop areas hand-drawn after R.A. Fischer (personal observations) and Monfreda et al. (2008)

Wheat in North America—Saskatchewan, Canada

During the 2007–09 seasons, Canadian wheat area comprised 68% spring BW and 23% spring DW. About half of spring BW is found in the driest prairie province, Saskatchewan, which in 2004–08 (latest available data) grew 3.3 Mha of spring BW. Due to rotational diversification, the area of spring BW is now about half what it was 30 years ago. In contrast, canola area has expanded fourfold to reach 2.7 Mha, DW area has increased modestly to 1.7 Mha and broadleaf crops (including pulses, flaxseed and sunflower) now total ~2.3 Mha.

In Saskatchewan crops were traditionally grown on summer fallow (one crop every 2 years) but the proportion of those crops has declined to less than 15%, while the proportion under conservation tillage (stubble retention, chemical weed control during fallow, zero-till seeding, crop every year) has grown to more than 60% (Veeman and Gray 2010). This has happened in the past 20 years or so; thus there has been a true agronomic revolution in Saskatchewan. Although quite variable due to periodic droughts, spring bread wheat FY has registered significant progress of 19 kg/ha/yr (0.01 < P < 0.05) over the past 30 years (data sourced from STATCAN 2011). This is equivalent to a gain of 0.8% p.a. relative to the 2010 estimated FY of 2.25 t/ha.

Using spring BW registration trials (unprotected, mostly on summer fallow) at 13–15 locations across Saskatchewan, DePauw et al. (2007) showed rate of yield increase from breeding (1980–2003) to be 0.74% p.a. of the yield of the control cultivar Katepwa (released 1980). With data given in that paper, this can be converted to PYw progress of 25 kg/ha/yr, or 0.6% p.a. relative to the 2003 estimated PYw of 3.8 t/ha in this water-limited environment. This is an improvement over only 6.5 kg/ha/yr from trials covering 1947–92 (McCaig and DePauw 1995), equating to 0.2% p.a. relative to a 1992 PYw of 3.2 t/ha. These two numbers agree with a variety index—calculated from varieties observed to be grown by farmers and their relative PYw advantage (see Section 2.2)—which showed a 0.3% p.a. rate of progress between 1976 and 2006 (Veeman and Gray 2010). The increase in rate of breeding progress seen in DePauw et al. (2007) is notable given the stringent quality requirements for variety release (such as maintaining high protein content). Among other features, increased progress has been attributed to doubling population sizes in the breeding program.

The above PYw results suggest that the yield gap for spring BW in Saskatchewan in 2003 was ~75% of FY, with PYw and FY showing similar relative rates of progress. The gap is somewhat surprising given the widespread inclusion of canola and pulses in the crop rotation, the adoption of conservation tillage and the apparent appropriate use of nitrogen fertiliser (DePauw et al. 2010). The gap estimate could have been biased somewhat due to inflated values of PYw because of the tendency to grow the trials on summer fallow and to plot edge effects in wet seasons. For example, the breeders’ practice of planting plot borders to winter wheat (R. DePauw, pers. comm. 2011) may balance underground moisture competition, but in good seasons is unlikely to compensate for the benefits of extra light intercepted by plot edges.

For spring BW in Saskatchewan, McCaig and DePauw (1995) noted that grains per square metre (GN) increased with year of release, but that time to maturity, plant height, grain weight (GW) and hectolitre weight (weight per hectolitre, an important measure of grain plumpness and predictor of milling yield) had not changed. More recent studies show increases to GW and grain-filling rate (Wang et al. 2002), and nitrogen HI (the proportion of total nitrogen uptake in grains at maturity), with faster and more complete mobilisation of nitrogen to the grain (Wang et al. 2003). Interestingly—given the emphasis placed on deep rooting by an early Saskatchewan wheat breeder, E. A. Hurd—recent studies show no change in soil water extraction to 120 cm depth at maturity (Wang et al. 2007), which was also the case with DW wheat varieties (see next paragraph).

This analysis of spring BW in Saskatchewan is strengthened by a report on spring DW progress (Clarke et al. 2010). These authors showed that over the 1964–2009 period, FY increased at a rate of 16 kg/ha/yr (P  <  0.01), or 0.7% p.a. of the estimated 2009 FY of 2.2 t/ha. From 6–12 trials annually across the main durum area—unprotected, largely disease-free, all on summer fallow (C. Pozniak, pers. comm. 2011)—rate of breeding progress was strongly linear at 0.7% p.a. of the yield of control cultivar Hercules (released 1969). Quality parameters (protein and pigment concentration, and gluten strength) steadily increased, while disease resistance was maintained. The average PYw of Hercules was 2.5 t/ha (C. Pozniak, pers. comm. 2011) and the estimated 2009 PYw was 145.5% of the Hercules PYw value. Thus the relative rate of PYw progress converts to 17 kg/ha/yr, or 0.5% of the estimated 2009 PYw of 3.6 t/ha, and the yield gap is 65%. As for BW varieties mentioned above, Veeman and Gray (2010) calculated a variety index for DW varieties grown by farmers, which increased at 0.4% p.a. between 1976 and 2006. Thus the yield gain for DW wheat in Saskatchewan is fairly similar to spring BW, with the same concerns regarding possible inflation of the PYw estimate.

Given the agronomic revolution in Saskatchewan, it is surprising that FY progress for both bread and durum wheats has not exceeded breeding progress as reflected in PYw progress. This trend is most likely explained by increased cropping intensity with the decline of summer fallow, a phenomenon that is likely to occur throughout the lower rainfall parts of WME6 with the adoption of conservation tillage.

Wheat in North America—North Dakota, USA

Thirty-one per cent of the harvested wheat area in the USA is in WME6. North Dakota to the immediate south-east of Saskatchewan and in the northern Great Plains is the major spring BW producing state, with a steady annual area of ~2.6 Mha of spring BW. This rainfed crop is grown in rotation with a large diversity of broadleaf crops including canola, sunflower, pulses and flaxseed (total area ~1.3 Mha), plus soybean (1.7 Mha). Only ~0.6 Mha of spring DW and minor amounts of winter wheat are grown. FY rate of progress for spring BW wheat is 1.0% p.a. of the estimated 2010 FY of 2.5 t/ha (Figure 3.7).

Underdahl et al. (2008) reported breeding progress across 33 varieties (1968–2008) grown unprotected at five representative sites over two representative years (2004 and 2005). While genetic resistance to leaf rust and Fusarium head blight (FHB) improved with time, both were evident in most tests and influenced yield. It is possible to correct progress in PYw for change in genetic disease resistance by fitting a multiple linear regression with year of release and disease as independents. For varieties released after 1985 there was a good fit for yields (R2 = 0.66), with significant but small coefficients for leaf rust index (Roelfs et al. 1992) and FHB score (Underdahl et al. 2008), and a highly significant coefficient for year of release (28 kg/ha/yr).

Yields (PYw) corrected for leaf rust and FHB are shown in Figure 3.7. The estimated 2008 PYw in Figure 3.7 is 4.0 t/ha, giving a rate of 0.7% p.a. PYw progress. Without correction for disease, there was a biased and higher linear slope for year of release (39 kg/ha/yr) because older varieties had become more susceptible to disease. Assuming that PYw is representative in space and time, the yield gap in North Dakota therefore is ~1.5 t/ha or 60% of FY (Figure 3.7). The yield gap in Figure 3.7 appears to be closing only gradually, perhaps for the same reasons as given for Saskatchewan, and despite the fact that springs are becoming warmer and spring wheat planting dates earlier in the North Dakota region (Lanning et al. 2010), something that should favour FY, other things equal.

Underdahl et al. (2008) found that breeding between 1968 and 2008 did not change plant height or days to heading, but did increase GN, GW and hectolitre weight.

Grain yield in tonnes per hectare is plotted against year from 1981 to 2010. Water-limited PY data are shown at higher grain yield than FY data. A regression line fitted to PY data shows a slope of 28 kg/ha/yr. A regression line fitted to FY data shows a slope of 25 kg/ha/yr.

***P < 0.01

Figure 3.7 Change in farm yield (FY), plotted against year, and corrected (see text) water-limited potential yield (PYw), plotted against year of variety release, for spring bread wheat from 1981 to 2010 in North Dakota, USA. Source: NASS (2012) for FY; Underdahl et al. (2008) for PYw

Wheat in Finland

At very high latitudes (60 °N and 65 °N) and under mostly humid conditions, Finland now grows 0.20 Mha of wheat, 95% of which is spring bread wheat. The area has doubled since Finland joined the European Union (EU) in 1995. FY increased between 1980 and 1996, then dropped away sharply in 1995 (from 4.0 to 2.0 t/ha) as farmers adjusted to the new regime, before recovering to a long-term trendline showing steady progress of 35 kg/ha/yr (0.01 < P <  0.05) over the past nine years (equating a rate of 1.0% p.a. FY progress), with an estimated FY of 3.7 t/ha in 2008.

Progress has been analysed by Peltonen-Sainio et al. (2009a) using the Official Variety Trials conducted during 1976–2006 at 28 locations across the country; trials appeared to follow good farmer practice, but were unprotected. A mixed-model statistical technique determined the fixed effect of variety, which was then plotted against year of variety introduction into the trials. Breeding progress was linear throughout (36 kg/ha/yr) but average yield of these trials fell away after 1995 before recovering to be closely parallel to FY, although a few hundred kilograms per hectare above FY (Peltonen-Sainio et al. 2009a).

Peltonen-Sainio et al. (2009a) ascribe levelling-off of yields to the EU agricultural policy and changes in markets (for example, poorer farmer terms of trade as support shifted from grain price support to direct income payments). Nitrogen use has been restricted, so fertiliser rates have declined and practices have become less intensive. The Peltonen-Sainio et al. (2009a) data suggest that PY (with the best agronomy) was at least 4.8 t/ha in 2006; PY was somewhat greater in the south than in the north, in proportion to the change in growing-season length (P. Peltonen-Sainio, pers. comm. 2011). From these data, the rate of PY progress in 1976–2006 was 0.8% p.a. and the 2006 yield gap is ~30% of FY.

PY increase in Finland in spring bread wheat has been associated with increased GN rather than increased GW (Peltonen-Sainio et al. 2007). Harvest index (HI), currently only 0.40, is seen by Peltonen-Sainio et al. (2009a) to offer scope for further PY increase through breeding. Climate change effects have been minor to date, although warming has permitted earlier spring cereal sowing (by 1–3 days per decade since 1980) (Kaukoranta and Hakala 2008). Future warming is expected to increase wheat yield through earlier planting and a longer growing season, as well as increasing the cultivable area for wheat (Peltonen-Sainio et al. 2009b).

Wheat in Western Siberia, Russian Federation

The Russian Federation grows ~14 Mha of spring wheat (2006–08 average; ROSSTAT 2012), approximately as shown in Map 3.7. The western Siberian region, lying between latitude 54 °N and 56 °N and longitude 60 °E and 90 °E, grows about one-half of the 14 Mha of spring wheat grown in the Russian Federation (Morgounov et al. 2010). Focusing on the centrally located Omsk district, annual precipitation is low (325 mm) and wheat is generally grown on summer fallow (one crop every two years). FY is quite variable (as might be expected in a low-rainfall region), but during 1980–2011 has nevertheless grown significantly (0.01 < P  <  0.05) at a rate of 1.1% p.a. relative to the estimated 2011 FY of 1.6 t/ha (data supplied Y. Zelenskiy, CIMMYT, pers. comm. 2012).

At the Siberian Institute of Agriculture (SRIA) at Omsk, vintage trials over seven years (2002–08) compared 47 varieties grown in the region between 1900 and 2000 (Morgounov et al. 2010). Plots were unprotected and received no fertiliser, but the Chernozem soil was very fertile with 6–7% organic matter. Since leaf rust levels and heading date were recorded and had effects on yield, a multiple regression was again made but using grain yield of only the most recent 26 varieties (1976–97) along with the independent variables: leaf rust score, days from seedling emergence to heading and selection year. The regression coefficient for leaf rust was significant (–5.5 kg/ha/%; P = 0.03), while both that for days to heading (84 kg/ha/d; P = 0.008) and year of release (56 kg/ha/yr; P  <  0.001) were very highly significant (overall R2 = 0.702). For an average number of days from emergence to heading (44 days) and zero rust, the predicted PYw was 4.4 t/ha for a variety selected in 1997. Thus the rate of increase in PYw of 56 kg/ha/yr amounts to an impressive 1.3% p.a. of the 1997 variety PYw. The yield gap was large, at ~175%. There is, however, insufficient information to ensure that the soil and climate at SRIA are representative of the region, and thus to ascribe this gap solely to management.

The strong breeding progress seen at SRIA, Omsk, may explain why varieties from western Siberian performed best across all WME6 environments in the study of Trethowan et al. (2006). The Siberian varieties tend to be tall (average 106 cm, no major dwarfing genes used) and late flowering because of their photoperiod sensitivity. Neither of these traits has been changed consistently by breeding at Omsk and both are probably critical for high-latitude adaptation in spring wheats (Morgounov et al. 2010). Breeding progress was, however, highly significantly associated with both increased GN and GW, the former driving about three-quarters of the PYw yield increase seen at Omsk.

Immediately south of the western Siberian region are the vast rainfed spring wheat lands of northern Kazakhstan. Under drier conditions than in Western Siberia, currently 12 Mha of spring BW is grown for a FY of about 1.2 t/ha (Petrick et al. 2013). Data on progress are currently unavailable, but Omsk varieties are among those grown in the region. Northern Kazakhstan is adopting conservation tillage rapidly, with very positive results.

Conclusion for spring wheat at high latitudes

Progress in FY (0.7–1.0%) appears quite respectable across the WME6 examples studied, as does PY and PYw progress (0.6–1.3%); the latter progress is largely from breeding and has been achieved in the face of strong emphasis on grain quality in the North American situations. The adoption of conservation tillage has been a major agronomic innovation in the past 20 years in North America and Kazakhstan; however, its contribution to PYw is not clear because it has reduced the proportion of wheat grown after summer fallow, especially in western Canada. Yield gaps range from 30% to 85% of FY, with the exception of a likely larger yield gap in Western Siberia that needs further investigation. Not enough information was available on farm-level constraints anywhere, except that a less yield-positive European Union (EU) policy environment is reported from Finland. WME6 in North America is unique for the generally high degree of rotation of wheat with broadleaf grain crops (canola, pulses, sunflower and flaxseed). Such rotation is also facilitated by conservation tillage, but when compared to wheat, most of these rotation crops have the disadvantage of leaving little residue for snow trapping. The plant phenotype for best performance under spring sowing at high latitude appears to be unique, as exemplified by the most recent tall day-length-sensitive Siberian varieties.

28Wheat will be killed by extreme Tmin occurrences in the absence of snow cover (Tmin < approximately 20 °C), so average Tmean can be taken as only an approximate guide.

Download this chapter in PDF | Download citation | Back to top