Phenological Development of Spring Barley in a Short-Season Growing Area

livestock industry (CGC, 1996). Barley is also used for malt production and a small amount is used for human Understanding the phenological development of barley (Hordeum food. Understanding the development of barley in this vulgare L.) grown under field conditions in its major growing area in the northern Great Plains, the Canadian province of Alberta, is environment is crucial to continued advancement in its important for the development of a crop model in this area. Five breeding and to production systems and crop growth cultivars of barley registered for production on the Canadian Prairies simulation models for this crop. were grown at three locations in Alberta (Botha, Lacombe, and Olds) Phenology is the development of a plant through sucfrom 1993 to 1996. Measurements and estimates were made of 12 cessive growth stages. It is important for understanding growth stages, the final leaf number of the main culm, phyllochrons, biomass partitioning and stress assessment. Several and leaf area indices (LAIs). The average phyllochron was 69.1 growgrowth-staging systems have been developed for cereals ing degree days (GDD), and the final leaf number was 9.0. While and Landes and Porter (1989) compared 23 of them. location–year differences in GDD requirements to reach successive Commonly-used scales to describe plant growth, such stages could be related to environmental conditions, the genotypic as BBCH (Lancashire et al., 1991) and that of Haun effects and genotype 3 environment (g 3 e) interactions would require that specific genotypic coefficients be introduced into the model. (1973), are based on the external appearance of the For instance, ‘Manley’ required only 129 GDD to emerge and 493 plant. Other scales, such as that of Banerjee and GDD to reach Apex 1 but required 1495 GDD to reach physiological Wienhues [B&W] (1965), describe apical growth. maturity while ‘Tukwa’ required 133 GDD to emerge, 514 GDD to Although scales based on external appearance and reach Apex 1, and only 1431 GDD to reach physiological maturity. apical growth can be very precise, the relationships beDespite genotypic differences in reaching specific growth stages, all tween them are not always clear, and depending on why cultivars had very synchronous growth under the wide range of envistages are being measured, the selection of an approronments encountered in the 12 location-years of these tests. This priate scale is important (Landes and Porter, 1989). As should allow for the development of a crop model for barley that will well, the relationship of scales to environmental factors, accurately predict growth stages and the allocation of resources to temperature, photoperiod, and time are not always well the growth and maintenance of plant structures (leaves, stems, spikes, and kernels). understood. However, Rickman and Klepper (1995) developed an elegantly simple model relating the phenological development of the apex with the external development of the plant and thermal time. They illustrated T major growing area for barley in North Amerthe full-season synchronous initiation and development ica is in the northern Great Plains where the Canaof all vegetative and reproductive parts of the plant as dian province of Alberta produces approximately 6 Tg a function of average temperature sums. However, it on 2 million ha (AAFRD, 1999). About 80% of the may be that a nonlinear relationship that accounts for barley produced in Alberta supports its Can$3 billion temperature and daylength effects on growth, such as the beta function described by Jame et al. (1998), may P.E. Juskiw, Field Crop Dev. Cent., Alberta Agric., Food, and Rural better explain the relationship of growth to GDD. Dev., 5030 50th St., Lacombe, AB T4L 1W8 Canada; Y. Jame, Semiarid Prairie Agric. Res. Cent., Agric. and Agri-Food Canada, Box The phyllochron is the interval, either on a calendar or 1030, Swift Current, SK S9H 3X2 Canada; and L. Kryzanowski, Agron. thermal unit basis, between the emergence of successive Unit, Alberta Agric., Food, and Rural Development, 905 O.S. Longman Building, Edmonton, AB T6H 4P2 Canada. Received 24 Mar. Abbreviations: B&W, Banerjee and Wienhues; g 3 e, genotype 3 2000. *Corresponding author ([email protected]). environment interaction; GDD, growing degree days; LAI, leaf area index; LAImax, maximum leaf area index. Published in Agron. J. 93:370–379 (2001). JUSKIW ET AL.: PHENOLOGICAL DEVELOPMENT OF SPRING BARLEY 371 MATERIALS AND METHODS leaves (Rickman and Klepper, 1995). For spring barley at a given temperature, the emergence of new leaves These tests were conducted from 1993 to 1996 at Botha was found by Cao and Moss (1989a) to be a linear [Daysland loam, 60% orthic Black Solod (coarse loamy, Typic function of time, and differences in the rate of appearArgiustoll with a natric horizon) and 40% thin orthic Black ance were found between temperatures and genotypes. Chernozem (coarse loamy, frigid Typic Haplustoll)], Lacombe [Penhold loam, orthic Black Chernozem (coarse loamy, frigid For spring barley as well, the emergence of new leaves Typic Haplustoll)], and Olds [Didsbury loam, orthic Black at a given daylength was found by Cao and Moss (1989b) Chernozem (coarse loamy, frigid Typic Haplustoll)]. These to be a linear function of time, and differences were soils are considered nutrient rich due to their high organic observed between daylengths and genotypes. Cao and matter contents of 7 to 13% and high N levels before seeding Moss (1989c) also pointed out that although phyllo(NO3 levels before seeding ranging from 32–111 mg kg2 and chrons for spring barley varied among genotypes and NH4 levels ranging from 8–39 mg kg2 in the top 0–15 cm combinations of temperature and daylength, they inof soil). creased as either temperature or daylength increased. Plot size from 1993 to 1995 was eight 6.10-m-long rows with Kirby (1995) reported that the rate of leaf appearance 0.14-m row spacing. Plots were trimmed to six 2.74-m-long rows with 0.14-m row spacing before final harvest. Plot size was set early in the life cycle. Frank and Bauer (1995) in 1996 at all three sites was eight 4.57-m-long rows with noted that the phyllochron has been widely accepted 0.14-m row spacing. Plots were twinned, two plots of the same by crop modelers for predicting plant development and treatment, so that biomass samples over the growing season by producers for determining the timing of management were taken from one plot and LAIs, phenological stages, and practices such as irrigation, fertilizer, and pesticide apthe yield at final harvest were taken from the adjacent plot. plications. Plots were trimmed to six 2.52-m-long rows with 0.14-m row Studies relating the morphological stages of barley to spacing before final harvest. thermal time have been made in Alaska (Dofing, 1992; Plots were seeded using a belt-type seeder in 1993 on 11 Dofing and Karlsson, 1993; Sharratt, 1999), Australia May at Botha, 14 May at Lacombe, and 12 May at Olds; in (López-Castañeda and Richards, 1994), North Dakota 1994 on 9 May at Botha, 5 May at Lacombe, and 12 May at Olds; in 1995 on 9 May at Botha, 9 May at Lacombe, and 11 (Bauer et al., 1993), Quebec (Ma and Smith, 1992), May at Olds; and in 1996 on 21 May at Botha, 13 May at Spain (Garcia del Moral et al., 1991), Syria (van OostLacombe, and 17 May at Olds. Seeds were treated with 2 mL erom and Acevedo, 1992), and under growth room conkg2 Vitavax single solution [a.i. carbathiin (5,6-dihydro-2ditions (Frank and Bauer, 1997). Phyllochrons for spring methyl-N-phenyl-1,4-oxathiin-3-carboxamide)]. The seeding barley have been determined in Alaska (Dofing and rate was approximately 200 seeds m2 based on kernel weights. Karlsson, 1993; Sharratt, 1999), Alberta (Jedel and The seeding depth was approximately 0.04 m. Helm, 1994b), North Dakota (Frank and Bauer, 1995), Fertilizer applications were based on soil tests done in the and Syria (van Oosterom and Acevedo, 1992). Leaf fall of the year previous to seeding; 112 kg ha2 of a premix development rates in barley have been studied by Cao blend of ammonium phosphate (NH4H2PO4) and potassium and Moss (1989a, 1989b). Final leaf numbers in barley chloride (KCL) was incorporated with the seed. Weeds were controlled by hand weeding and the application of herbicides. have been determined in Alaska (Dofing, 1992), Alberta In 1993 at Lacombe, Lorsban [a.i. chlorpyrifos (O,O-diethyl (Jedel and Helm, 1994a), Australia (López-Castañeda O-(3,5,6-trichloro-2-pyridinyl) phosphorothioate)] was apand Richards, 1994), Britain (Gallagher, 1979), North plied at 1200 mL ha2 to control cutworms. Dakota (Bauer et al., 1993), Quebec (Ma and Smith, The five barley cultivars used in these tests were ‘Brier’, 1992), and Syria (van Oosterom and Acevedo, 1992). ‘Duel’, Manley, Tukwa and ‘Seebe’. Brier, Duel, and Tukwa Variation found in the literature of values (GDD) for are six-row cultivars while Manley and Seebe are two-row phenological traits are, in part, due to differences in cultivars. Duel and Manley are malt types while the others when the measurements started (sowing vs. emergence). are classed as feeds. Tukwa and Duel mature 1 d earlier than As well, they may reflect location differences in temperthe cultivar Harrington while Brier is similar maturing and atures and daylengths (Jame et al., 1998). As Cao and Manley and Seebe mature 3 to 4 d later (AAFRD, 1997). Brier, Duel, and Seebe are tall cultivars at 83, 90, and 84 cm, Moss (1986b) found that the phyllochron decreased as respectively, while Manley is a standard-height cultivar at 78 daylength increased, so the smaller phyllochrons found cm and Tukwa is a semidwarf cultivar at 72 cm (AAFRD, in Alaska vs. those in Syria could be explained on the 1997). Seebe has good resistance to scald [Rhyncosporium basis of daylength differences. The variation found in secalis (Oud.) J. Davis], which is generally the most prevalent the literature illustrates why it is important that data be leaf disease in Alberta, while Brier and Tukwa have intermedicollected specifically for the development and validation ate resistance and Duel and Manley are susceptible (AAFRD, of a barley crop model in the environment where it will 1997). Tukwa and Brier have intermediate resistance to the be used unless all of the variability can be accounted spot form of net blotch (Pyrenophora teres f. maculata for by coefficients and equations. Drechs.), also a common leaf disease in Alberta, while the other three cultivars are susceptible (AAFRD, 1997). The objective of the research reported in this paper was to study the phenological development of several barley cultivars grown under field conditions. These Weather Measurements data, in conjunction with biomass data that will be reWeather data were collected from on-site weather stations ported later, were used to develop and validate a crop located within 2 km of the plots. Minimum, maximum, and model for barley (SPARC-Barley, developed by Y. mean temperatures were collected daily along with solar radiaJame), especially for use in the Alberta portion of the tion and precipitation data using a LI-1000 DataLogger (LICOR, Lincoln, NE) at all location-years except Olds in 1993 Great Plains. 372 AGRONOMY JOURNAL, VOL. 93, MARCH–APRIL 2001 when a Campbell system [Campbell Scientific (Canada) Corp. chron in GDD. Final leaf numbers were determined as the number of leaves on the main culm. Leaf areas were measured Edmonton, AB] was used. When data were missing, missing values were replaced by data obtained from Agriculture and three to five times throughout the growing season using a LICOR LAI-2000 Plant Canopy Analyzer (LI-COR, Lincoln, Agri-Food Canada for the Lacombe site (approximately 3 km from the site); Olds College for the Olds site (approximately NE). Due to lodging at Olds in 1995, LAIs could not be measured using this method. 5 km from the site); and the Conservation and Development Branch of Alberta Agriculture, Food, and Rural Development for the Botha site (approximately 25 km from the site). GrowStatistical Analysis ing Degree Days were calculated as the accumulation of daily The experimental design was a randomized complete block mean temperatures with five cultivars and four replicates at each location. Analyo(Tm . 08C) [1] ses were conducted using PROC GLM of SAS (SAS Inst., 1988). Locations and cultivars were treated as fixed effects where Tm is the mean temperature. When Tm was unavailable, while years and times of measurements were treated as ranbut the maximum and minimum temperatures were available, dom effects. Errors appropriate to mixed models were used GDD was calculated as to test for the significance of the effects (Steel and Torrie, o[(Tmax 1 Tmin)/2 . 08C] [2] 1980). Phyllochrons were determined as the slope of the linear equations of weekly leaf numbers regressed against GDD where Tmax is the maximum temperature and Tmin is the miniaccumulated since emergence using PROC GLM of SAS (SAS mum temperature. Inst., 1988). For the B&W data, stage was regressed as a linear function of GDD from emergence to sampling date using Soil Measurements PROC GLM. Leaf numbers at sampling for B&W staging In 1993 and 1994, soils were sampled in the fall previous were also regressed against GDD at sampling to predict the to planting to determine nutrient status. Before seeding and leaf number at double ridge (B&W 4) and glume initials after harvest in 1994, soils were sampled using a hand corer (B&W 7). Quadratic equations were fitted to the LAI data to determine moisture content to depths of 30 or 45 cm. Soil by year, location, and cultivar to determine maximum LAI nutrient analyses were done by Norwest Labs, Edmonton, AB. (LAImax) and time of LAImax in GDD. In 1995 and 1996, deep soil core samples were collected before seeding and after harvest to depths of 0.9 to 1.5 m. RESULTS AND DISCUSSION Deep core samples were analyzed for moisture, N, and bulk densities. Organic matter, sand, silt, and clay content were The Environments determined on a selected subsample. Phosphorous, K, pH, Accumulated precipitation during the growing season and electroconductivity were determined for the 0to 15-cm depth. Soil nutrient analyses were done by Soil and Crop was highest at Olds in 1993, 1995, and 1996 with initial Diagnostic Centre, Edmonton, AB. soil moisture contents of 0.366 and 0.287 kg kg21 in 1995 and 1996, respectively (initial soil moisture not Plant Measurements determined in 1993). However, in 1994, Olds had the lowest accumulated precipitation (Fig. 1) and a low iniPhenological stages were monitored throughout the growing season and based on 50% of the plot at each stage: emertial soil moisture content of 0.151 kg kg21 for the top gence (the first true leaf visible, BBCH 10); Apex 1 (the first 0 to 15 cm. Accumulated precipitation for Botha and node 0.01 m or above the soil surface, BBCH 30); flag leaf Lacombe were very similar in 1994 and 1995 (Fig. 1). emerged (the ligules of the flag leaf visible, BBCH 39); flag While Botha had a low initial soil moisture content in leaf fully elongated (fully elongated flag leaf, no BBCH stage); 1994 (0.179 kg kg21 for the top 0–15 cm), initial soil anthesis (the central florets of the head having shed their moisture content at Lacombe was higher (0.246 kg kg21 pollen, BBCH 61); heading (heads fully emerged from the for the top 0–15 cm). In 1995, initial soil moisture conleaf sheath of the flag leaf, BBCH 55); peduncle fully elongated tents were good at both locations (0.279 kg kg21 at Botha (the last internode or peduncle fully elongated, no BBCH stage); milk (kernels with milky endosperms, BBCH 75); soft and 0.387 kg kg21 at Lacombe). Lacombe had higher dough (kernels with soft, doughy endosperms, BBCH 85); accumulated precipitation in 1993 and lower accumuphysiological maturity (loss of green color from the peduncle, lated precipitation than Botha in 1996 (Fig. 1). However, corresponding to approximately 30% kernel moisture content, both locations had good initial soil moisture contents BBCH 87). Emergence was determined as GDD from sowing in 1996 (0.301 kg kg21 at Botha and 0.283 kg kg21 at while all other stages were determined as GDD from Lacombe). The long-term average precipitation (May– emergence. Sept.) is 303 mm for Botha, 311 mm for Lacombe, and To assess early development, dissections were made in 1994 335 mm for Olds (Jedel and Salmon, 1993). While 1993 and 1995 to determine the apical development of the main and 1995 were near-average years at Olds, 1994 and culm. Stages were assigned according to the scale of B&W (1965). Leaf numbers were counted on the main culm be1996 were dry. At Botha and Lacombe in all 4 yr, rainfall fore dissections. was below average. Leaf counts based on the Haun (1973) scale were made Mean daily temperatures during the growing season once a week from emergence to flag leaf on the main culm. remained .08C and ,258C at all locations (Fig. 2), Five plants were selected per plot by tagging shortly after which is well within the growing range for barley (Jame emergence. Using fine multicolored wire, tillers were distinet al., 1998). The only exception was a cold period in guished from the main culm. We marked the fifth leaf of the early May 1996. Temperature patterns within years were main culm by cutting approximately 1 cm from the end of the very similar at the three locations, except at Botha in leaf to give a blunt end that was easily visible as the stem 1993. Temperatures were slightly higher at Botha and elongated and lower leaves senesced. These data were regressed against GDD, and the slope was taken as the phyllolower at Olds compared with Lacombe, reflecting the JUSKIW ET AL.: PHENOLOGICAL DEVELOPMENT OF SPRING BARLEY 373 Fig. 1. Total precipitation (mm) from sowing at three locations in Alberta, Canada (Botha, Lacombe, and Olds) in 1993, 1994, 1995, and 1996. long-term averages from May to September of 13.78C from sowing to emergence of 91 to 192 in the environments studied (Table 2) was greater than the 63 GDD at Botha, 12.98C at Lacombe, and 12.58C at Olds (Jedel reported by Bauer et al. (1993) for spring barley grown and Salmon, 1993). in North Dakota, but it does encompass the 125 GDD reported by Kirby (1995) for barley grown in Britain. Plant Growth The relationship of emergence to genotypic and enviEmergence ronmental effects was complex, being influenced by the The year 3 location 3 cultivar interaction effect for soil moisture availability at seeding, precipitation after seeding, soil and air temperatures, and genotype. While emergence was significant (Table 1). The range in GDD Fig. 2. Mean daily temperature (8C) from sowing at three locations in Alberta, Canada (Botha, Lacombe, and Olds) in 1993, 1994, 1995, and 1996. 374 AGRONOMY JOURNAL, VOL. 93, MARCH–APRIL 2001 Table 1. Mean squares (MS) from the analysis of variance for growing degree days (GDD) from sowing to emergence and from emergence to nine phenological stages of five barley cultivars grown from 1993 to 1996 at three locations in Alberta, Canada (Botha, Lacombe, and Olds). Source of Flag leaf Heads fully Kernels at variation df Emergence† Apex 1 emerged Anthesis emerged soft dough Year (y) 3 5 247.92 (ns) 28 081.9 (ns) 9 219.3 (ns) 47 245.1 (ns) 286 572.8** 84 101.6 (ns) Location (1) 2 426.50 (ns) 120 418.8* 69 520.9 (ns) 112 679.2 (ns) 239 969.2* 217 218.1** y 3 1 6 2 347.50** 20 529.3** 23 492.0** 27 023.7** 28 535.1** 19 184.4** Error a 36 296.90 512.1 275.2 330.4 277.5 621.0 Cultivar (c) 4 259.22 (ns) 3 123.1* 11 528.3** 21 657.3** 38 834.0** 12 547.0** y 3 c 12 446.99** 818.3* 618.7 (ns) 1 327.9* 1 312.8 (ns) 484.7 (ns) l 3 c 8 192.09 (ns) 442.5 (ns) 255.2 (ns) 644.6 (ns) 856.7 (ns) 235.5 (ns) y 3 l 3 c 24 127.24* 347.4 (ns) 287.4 (ns) 495.6** 859.1** 602.1** Error b 144 78.1 256.3 196.0 149.8 145.4 245.8 Source of Flag leaf fully Peduncle fully Physiological variation df elongated df elongated Kernels at milk df maturity Year (y) 1 43 455.8 (ns) 2 138 642.1 (ns) 50 602.4 (ns) 3 269 172.2 (ns) Location (l) 2 103 301.5 (ns) 2 203 864.5 (ns) 248 651.6* 2 60 753.5 (ns) y 3 l 2 16 451.2** 4 67 679.9** 15 971.4** 4 43 146.6** Error a 18 172.6 27 1 979.5 326.0 30 1 370.0 Cultivar (c) 4 13 462.2** 4 43 252.4** 6 384.0** 4 22 979.7** y 3 c 4 414.7 (ns) 8 2 700.8 (ns) 434.1 (ns) 12 3 281.4 (ns) l 3 c 8 337.9 (ns) 8 4 135.7 (ns) 495.0 (ns) 8 2 228.4 (ns) y 3 l 3 c 8 219.2 (ns) 16 2 400.3* 537.2* 16 1 472.2** Error b 72 158.1 108 1 193.3 266.5 120 479.4 * Significant at the 0.05 level. ** Significant at the 0.01 level. † Where year was tested by [MS(y 3 l) 1 MS(y 3 c) 2 MS(y 3 l 3 c)]; y 3 l was tested by [MS(error a) 1 MS(y 3 l 3 c) 2 error b]; and other effects were tested by errors appropriate to a mixed model. the malt types may have been expected to have faster 4), which is consistent with the model of Rickman and emergence than the feed types—as rapid, uniform gerKlepper (1995). As the highest-order significant interacmination is a desirable trait in the malt house—no differtion was for year 3 cultivars (Table 5), only these linear ences in emergence were found in the six-row cultivars relationships were determined (Table 4). In the 2 yr in (Table 3). For the two-row cultivars, Manley, the malt which B&W growth stages were assessed, there was type, was earlier than Seebe in 8 of 12 location-years little difference between cultivars, with higher slopes in (significantly so in four, data not shown) while Seebe 1994 than 1995, suggesting more rapid development in was significantly earlier than Manley in two. It is this 1994 vs. 1995 (Table 4). The low initial soil moistures type of g 3 e interaction that a crop model must be able in 1994 may have stimulated the apex to initiate succesto predict. While emergence may have been expected to sive stages of apical development earlier. The late cultibe most rapid in the warm, moist soils of 1995, it was var, Manley, tended to have fewer GDD requirements in fact most rapid in the cool, moist soils of 1996 (data to reach double ridge than the other cultivars while not shown). the other late cultivar, Seebe, had the greatest GDD requirement (Table 4). These results are contrary to the Apical Growth Stages results of Garcia del Moral et al. (1991), who found that the days to double ridge were fewer in early maturing In the current study, the B&W growth stages were cultivars than in cultivars of medium maturity for spring found to be a linear function of GDD for barley (Table barley grown in Spain. They also indicate the importance of determining the genotypic coefficients for all Table 2. Means, range, and standard error (SE) for phenological growth stages rather than relying on only late stages. stages, phyllochrons, and final leaf number of five barley cultiAverage GDD to double ridge was 158, ranging from vars grown from 1993 to 1996 at three locations in Alberta, 123 to 184 (Table 4); this was fewer than the 215 GDD Canada (Botha, Lacombe, and Olds). found by Bauer et al. (1993) for barley grown in North Range Dakota. The greater GDD required in 1994 did not Trait Mean SE Min. Max. appear to be due to warm weather although under conGDD (08 C base) trolled conditions, Frank and Bauer (1997) found that Sowing to emergence 133 1.1 91 192 barley required more GDD to reach double ridge under Emergence to: hot (268C) conditions than cool (188C) conditions. AverApex 1 505 3.1 410 601 age GDD to glume initials was 312, ranging from 273 Flag leaf emerged 681 2.7 565 809 Flag leaf fully elongated 755 5.0 648 864 to 369 (Table 4). This range reflects the findings of Anthesis 813 3.4 705 966 Frank and Bauer (1997) for barley grown under cool Heading 916 5.5 756 1109 Peduncle fully elongated 991 6.6 827 1204 (188C) conditions. Kernels at milk 1036 4.8 913 1169 During the time from double ridge to glume initials, Kernels at soft dough 1242 4.