PDF version
for printing

Impact
Statement

© 2009 Plant Management Network.
Accepted for publication 2 April 2009. Published 8 May 2009.

Using Forage Brassicas Under Irrigation in Mid-latitude, High-elevation Steppe/Desert Biomes

Leonard M. Lauriault, Forage Agronomist, Tucumcari Agricultural Science Center, New Mexico State University, Tucumcari, NM 88401; Steven J. Guldan, Superintendent, and Charles A. Martin, Agricultural Specialist, Alcalde Sustainable Agriculture Science Center, New Mexico State University, Alcalde, NM 87511; and Dawn M. VanLeeuwen, Agricultural Biometrician, Agricultural Biometrics Service, Agriculture Experiment Station, New Mexico State University, Las Cruces, NM 88003

Corresponding author: Leonard M. Lauriault. lmlaur@nmsu.edu

Lauriault, L. M., Guldan, S. J., Martin, C. A., and VanLeeuwen, D. M. 2009. Using forage brassicas under irrigation in mid-latitude, high-elevation steppe/desert biomes. Online. Forage and Grazinglands doi:10.1094/FG-2009-0508-01-RS.

Abstract

Brassicas may be useful for autumn forage in irrigated mid-latitude, high-elevation steppe/desert biomes. Aboveground dry matter yields and crude protein (CP) of kale (Brassica oleracea L.), rape (Brassica napus L.), and turnip (Brassica rapa L.) planted mid July and mid August were compared in northern New Mexico, USA, in multiple-location, three-cut [one 60 days after planting (DAP) and two 30-day regrowth periods] studies with four replications. Despite numerically lower first harvest yields for all species planted mid July, greater second harvest yields led to greater total yield for that planting (3005 kg/ha) compared to the August planting (2697 kg/ha). First harvest yields of all species were different (2105, 2899, and 2406 kg/ha for kale, rape, and turnip, respectively) and turnip had greater second cut yield than rape (314, 161, and 412 kg/ha for kale, rape, and turnip, respectively). Second cut yield by rape and third cut yields for all species were negligible. Kale total yields were lower than rape and turnip (2461, 3109, and 2983 kg/ha for kale, rape, and turnip, respectively). Species and planting date also affected CP. Rape has greatest yield potential for single-cut 60 DAP systems while turnip has better regrowth and sustained CP in these environments.

Introduction

Winter cereal pastures are widely used in the grassland steppe and high desert regions of the western USA during fall and winter when yield and quality of perennial species becomes limited (12). Forage brassicas, such as kale (Brassica oleracea L.), rape (Brassica napus L.), and turnip (Brassica rapa L.) also are frost tolerant (4) and often have forage dry matter (DM) yields equal to or higher than winter cereals.

Much of the previous research on forage brassicas has been conducted in regions having climates unlike the mid-latitude, high-elevation steppe and desert regions of the world, which includes much of the western USA. The value of Brassicas for supplementing perennial grass pastures in late fall in the eastern USA has been demonstrated (6). Under rainfed conditions in the southern Great Plains of the USA, Rao and Horn (12) concluded that maximum DM yields of kale and rape sown in September depended on the length of the growing period. Kalmbacher et al. (7) reported that multiple harvesting systems increased leaf yield compared to stockpiling in the subtropics. In the Atlantic region of Canada, delaying seeding from 25 July to 24 August reduced herbage DM yields of kale much more so than for rape or turnip when harvested 60 to 70 days after planting (8). In the northeastern USA, Wiedenhoeft (13) found a significant species × planting date interaction for DM yields in a study with rape, turnip, and a turnip hybrid. In addition, differences in fall DM yields among kale, rape, and turnip cultivars have been reported under rainfed conditions in the southern Great Plains (12) and northwestern Montana (11).

Forage nutritive value of brassicas also has been shown to be affected by seeding date, harvest date, and species. Wiedenhoeft and Barton (14) found that regrowth herbage was lower in fiber and higher in crude protein (CP) than the first growth herbage. While CP concentration is diluted as yield increases over time, digestibility remains high over extended periods making forage brassicas suitable for stockpiling (5,6). Guillard and Allinson (3) observed few differences in nutritive value between turnip cultivars; but, Wiedenhoeft and Barton (14) measured differences among species in the first harvest and regrowth. Jung et al. (5) concluded that differences among cultivars and species increase options to continue providing high quality forage as brassicas mature, whether sown in spring, summer, or fall.

Frost tolerance, high forage yield and nutritive value, and planting and harvest management flexibility of brassicas may make them particularly useful in the mid-latitude, high-elevation steppe and desert regions of the southwestern USA and climatically similar environments, which generally exhibit great variations in diurnal temperatures during autumn. Information is lacking, however, on the potential of forage brassicas in such environments. The objective of this study was to use a multiple-harvest fall production system to determine the effect of planting date on initial and regrowth herbage DM yield and CP of three Brassica species, including kale, rape, and turnip, the latter of which was represented by two cultivars.

Experimental Methods

This research was conducted in two summer-autumn growth periods (Years 1 and 2) at New Mexico State University's Alcalde Sustainable Agriculture Science Center (36.08°N, 106.05°W, elevation 1745 m) in a Fruitland fine sandy loam [coarse-loamy, mixed (calcareous), mesic Typic Torriorthent] and at Ghost Ranch, NM (36.33°N, 106.47°W, elevation 1970 m), in a Penistaja fine sandy loam (fine-loamy, mixed, mesic Ustollic Haplargids). The sites were approximately 38.5 km apart with a difference of over 200 m in elevation. Different fields were used each year at Alcalde, but the same field was used at Ghost Ranch each year. At Alcalde, the previous crop for Year 1 was blue corn (Zea mays L.) and for Year 2 the area had been in clean-tilled fallow for the previous 2 years. At Ghost Ranch the field had been used for vegetable production prior to the 2 years of the study. Averaged over locations and year, soil organic matter was generally less than 1%, pH was 7.77, N averaged 14.3 ppm, and P averaged 9.4 ppm. At each location, main plots were two planting dates (mid July and mid August), subplots were brassica species (‘Premier’ kale, ‘Sparta’ rape, and turnip cultivars ‘ForageStar’ and ‘Rondo’). The three repeated harvests and year were time factors. There were four randomized complete blocks each year at each site.

Weather data were collected from National Weather Service cooperative stations within 1 km of the study areas (Fig. 1). Consistent long-term temperature data from Ghost Ranch and some of the pre-growing season temperatures for Year 1 at Alcalde were not available. Temperatures did not vary greatly at either site across years; Alcalde was slightly warmer than Ghost Ranch each year (Fig. 1). Additionally, precipitation patterns were fairly consistent between sites within each year.

Fig. 1. Monthly mean air temperature and total precipitation for Alcalde and Ghost Ranch, NM, during 2 years of Brassica spp. research, and the long-term means.

Flood or sprinkler irrigation was used at Alcalde or Ghost Ranch, respectively. For each site-year, the test area was conventionally tilled for flat planting at Ghost Ranch or prepared for flood irrigation with borders surrounding the test at Alcalde. During planting each year at each site, plots received 50, 63, 0, and 44 kg/ha N, P₂O₅, K, and S, respectively. Brassicas were drilled in rows 18 cm apart in mid July and mid August each year at each site. Actual planting dates are given in Table 1. The seeding rates were 9, 9, and 4.5 kg/ha for kale, rape, and turnip, respectively. These seeding rates were twice or more the recommended rate for each species, but they were consistent to those used by other researchers (3,13). In the present study, the high seeding rate was used to assure establishment with the assumption that seedling mortality resulting from competition, pests, and the environment would reduce the plant densities and would have little effect on yield (5,13). Higher seeding rates also may reduce turnip root size and increase herbage production (13), which was the focus of this study. The study areas were irrigated throughout the summer-autumn growing season to supplement precipitation.

Table 1. Seeding and harvesting dates for Brassica spp. grown at two sites in the irrigated mid-latitude, high-elevation steppe/desert biome of the southern Rocky Mountains, USA.

First harvests were taken approximately 60 DAP followed by harvests of regrowth at 30-day intervals when regrowth was sufficient. Actual harvest dates occurred as weather permitted and are given in Table 1. Because planting dates were 30 days apart, some harvest dates were on the same date but represented different cuttings. For each harvest, a 0.836-m² quadrat placed in the center of a 1.5-m² area was hand-clipped from each plot leaving 5-cm stubble. The corners of the sampling areas were staked at both sites each year so that samples were always collected from the same areas in the plots for the duration of the study. Each sample was placed in a separate bag and dried for 48 h at 65°C to determine DM yield and then ground to pass through a 1-mm sieve for nutritive value analysis for CP. After each harvest, the entire 1.5-m² area was clipped to 5 cm and the residue removed to negate any border effects. Laboratory analysis for CP was conducted using the micro-Kjehldahl protocol (1).

The test was analyzed as a multi-location, multi-year trial. A randomized complete block split-plot design with repeated measurements was executed within each location and year. Planting date was the whole plot treatment and brassica cultivar was the split-plot treatment with harvest as a repeated factor. For DM yield, which had no missing cells, fixed effects included the main effects of site, year, planting date, cultivar, and harvest and all possible interactions. Rather than writing contrasts to assess whether significant cultivar differences were due to among or within species differences, a second analysis partitioned cultivar into terms corresponding to species and cultivar within species. Blocks and a term corresponding to the whole plot error were included as random effects. A repeated statement specifying the sub-plot as the subject and an unstructured covariance accounted for both correlations among the harvests from the same plot and unequal variance among the harvests (9). The analysis was conducted using SAS PROC MIXED (version 9.1.3, SAS Institute Inc., Cary, NC) and used the ddfm = KR option to estimate denominator degrees of freedom and adjust standard errors.

For CP, harvest data were missing for several plots. This led to a design with missing cells; the usual type III tests of fixed effects are not valid when there are missing cells (10). Consequently, the analysis used only the first harvest data with a model as described above but without the repeated factor, harvest, and any factor effects involving harvest. A limited analysis including data from both harvests included only one fixed effect with levels corresponding to every combination of year × site × planting date × harvest × cultivar present in the data. This analysis used ESTIMATE statements to compare harvests for each cultivar while controlling for the variables site, year and planting date. This analysis accounted for correlations between the harvests but did so by replacing the repeated statement with a random effect for the split-plot error term. It also included random effects for block and the whole-plot error.

For all analyses, for significant effects, an LSMEANS statement with the PDIFF option and in some cases the SLICE option was used to determine where differences occurred. Significance was defined for (P < 0.05). Unless otherwise indicated, differences were at that level.

Year and site were generally not significant as main effects for any measured variable; however, several interactions including these effects were significant. Because field history, soil types, and irrigation techniques differed between sites and years and climatic conditions varied across years, year, site, and replicate effects were combined to form the block effect.

Dry Matter Yield

Temperature throughout the summer/autumn growing season was likely a factor in the planting date × harvest interaction (P < 0.01, Table 2). Though not significant (P < 0.13), first harvest yield of the mid-July planting was numerically less than the mid-August planting (Table 2), possibly due to high temperature during the growth period prior to that harvest (Fig. 1). Wiedenhoeft (13) attributed reduced yield in summer plantings to reduced precipitation as well as air temperatures that exceeded 20°C. Irrigation was used to supplement precipitation throughout the summer-autumn growth period each year at each site. Consequently, variable precipitation likely had no effect on DM yield for any harvest in any year or site (Table 2, Fig. 1). Mean monthly air temperatures did approach or exceed 20°C at both sites during July and August of both years (Fig. 1), likely leading to the yield difference for the first harvest (Table 2).

Table 2. Dry matter yield (kg/ha) of Brassica spp. by harvest as influenced by planting date in the irrigated mid-latitude, high-elevation steppe/desert biome of the southern Rocky Mountains, USA.

 ^x Data are the least squares means of two sites (Alcalde & Ghost Ranch, NM), two years, four Brassica entries (kale, rape, and two turnip cultivars), and four replicates.

 ^y Least squares means followed by the same letter are not significantly different (P ≤ 0.05) based on the results of the pdiff analysis in SAS for the interaction.

 ^z Dates are approximate and represent harvests taken 60 days after planting for Harvest 1 and regrowths at 30-day intervals (Harvests 2 and 3) from for the mid-July and mid-August planting dates.

Generally, second harvest yields of the August planting and all third harvest yields were negligible, likely due to low temperatures (Table 2, Fig. 1). Prestbye and Welty (11) stated that brassicas stopped growth when temperatures fell below 2.2°C. Daily minimum temperatures did not fall below 2.2°C prior to the September harvest at either site in either year. They did fall below that level several times between mid September and mid October at both sites and, although temperatures rose above 2.2°C during the day, growth rates were still likely reduced. Only regrowth from the July planting in Year 2 was measurable for a third harvest taken in November. The third harvest for the August plantings would have been in mid December, but regrowth was immeasurable. Daily minimum temperatures had consistently dropped below 2.2°C by early November of both years and growth of the brassicas had basically ceased, preventing any December harvest (11). Temperatures were low enough by late October or early November in Year 1, and by late November in Year 2, to cause freeze damage (6).

The year × planting date interaction was significant for total yield. Despite an apparent reduction in yields for the mid-July planting due to high temperatures, that planting had higher total yields than the mid-August planting in Year 1 (3184 and 2536 for the mid-July and mid-August plantings, respectively) while there was no difference between yields in Year 2, (2826 and 2857 kg/ha for the mid-July and mid-August plantings, respectively). Year 2 yields were intermediate to Year 1 yields. Total forage yield was greater for the mid-July planting due to greater regrowth for that planting date (3005 vs. 2697 kg/ha for total yield of the July and August planting dates, respectively; Table 2). This phenomenon and the effects of the shortened growing season in Year 1 are similar to the findings of Rao and Horn (12) regarding yield of summer-sown brassicas. A longer regrowth period (i.e., one 60-day period rather than two 30-day periods) may lead to greater yield making earlier planting more feasible; especially if late summer temperatures are cooler and temperatures do not drop below 2.2°C too early in the autumn (Fig. 1, Table 2).

Differences existed within the species × harvest interaction (P < 0.01) and for species total yield (P < 0.01) (Table 3). Although initial turnip yields were lower than initial rape yields, regrowth by turnip was greater leading to no difference between the two species in total yield. These yield differences among species for both initial and total yield are consistent with those measured by Guillard and Allinson (3), although they used a single-cut system approximately 90 DAP. Kale and rape require longer growth periods to maximize productivity and decreasing temperatures and daylength will affect their yield more than turnip (3). Initial yield of turnip in this study (Table 3) was approximately three times that measured for 60-day growth of August plantings in Pennsylvania (6). The more southerly location (but higher elevation) and less change in daylength may enhance later-season growth potential compared to the mid-Atlantic and northeastern USA where Jung and Shaffer (6) and Guillard and Allinson (3) did their work. Regrowth across species was approximately 12% of initial growth (calculated from Table 3), which is consistent to 35-day regrowth of summer grown kale measured by Fraser et al. (2).

Table 3. Dry matter yield (kg/ha) of Brassica spp. by harvest and total yield in the irrigated mid-latitude, high-elevation steppe/desert biome of the southern Rocky Mountains, USA.

 ^x Harvests were taken 60 days after planting for Harvest 1 and regrowths at 30-day intervals (Harvests 2 and 3).

 ^y Data are the least squares means of two sites (Alcalde & Ghost Ranch, NM), two planting dates (mid July and mid August), two years, and four replicates and, for turnips, two cultivars (ForageStar and Rondo).

 ^z Harvest least squares means for the interaction or total yield least squares means followed by the same letter are not significantly different (P ≤ 0.05) based on the results of the pdiff analysis in SAS.

Total forage yields of kale in this study (Table 3) were similar to concurrent companion studies using a single-cut system approximately 120 DAP (S. J. Guldan, unpublished data), suggesting that these would be expected yields for that species under most harvest management regimes. Overall, because of low first-harvest yield, kale was not as productive as rape or turnip (Table 3). Lower yield by kale was consistent with the findings of Kunelius et al. (8). Contrary to the findings of Kunelius et al. (8) and Wiedenhoeft (13), no species × planting date interaction was observed in the present study, likely due to the lower latitude and less impact of changes in daylength. The site × cultivar within species effect was significant (P < 0.01) because ForageStar yielded better at Ghost Ranch while Rondo yielded better at Alcalde (data not shown).

Crude Protein

Due to zero or negligible yields for some harvests in both years, only forage from the first and second harvests was analyzed for CP concentration. The only effect that could be tested using all available data had a unique level for every combination of year × site × planting date × harvest × cultivar, which was significant (P < 0.01). Because of the missing values, harvests had to be analyzed separately. Consequently, discussion emphasizes the first harvest and only minimal comparisons involving harvests were made.

For the first harvest, there was no significant difference in CP among species (142, 137, and 135 g/kg for kale, rape, and turnip, respectively, P > 0.42); however, planting date and its interactions with year and site were all significant at P < 0.01 (data not shown). The planting date × species interaction also was significant within the first harvest (Table 4). Delaying planting increased CP in the first harvest for all species; however, the increase was greater for turnip because CP levels were lower than the other species for the July planting. Additionally, the numeric increase for rape was not significant (Table 4). There also was a difference between turnip cultivars across planting dates because the increase in CP of Rondo was greater than that of Forage Star (data not shown).

Table 4. Crude protein concentration (g/kg) 60 days after planting of Brassica spp. by planting date in the irrigated mid-latitude, high-elevation steppe/desert biome of the southern Rocky Mountains, USA.

 ^x Data are the lsmeans of two sites (Alcalde & Ghost Ranch, NM), two years, and four replicates and, for turnips, two cultivars (ForageStar and Rondo).

 ^y Least squares means of the interaction followed by the same letter are not significantly different (P ≤ 0.05) based on the results of the pdiff analysis in SAS for the interaction.

Overall, forage CP increased from the first to second harvest (Table 5). For each species, in a comparison using all available second harvest data but selecting first harvest means so as to both average across and also control for site, year, and planting date, second harvest CP estimates were consistently higher than first harvest estimates. These results are consistent with the findings of Wiedenhoeft and Barton (14), although the actual values for the first harvest (presented in the text above) are somewhat lower than those measured by several others (6,12,14) and may have been due to dilution by high yield (5). Apparent differences among species in the second harvest yields also are consistent with the results of Wiedenhoeft (13). As with differences among species previously described for the species × harvest interaction for yield (Table 3), slower growth rates due to cooler temperatures (Fig. 1) affecting maturity stage at harvest likely had a role in the increase in CP from the first to second harvest. This effect was likely magnified for kale and rape and is evinced by the declining yield across harvests shown in Table 3. For rape, regrowth yield was likely too low for the higher CP content to be of much value (Tables 3 and 5).

Table 5. Crude protein concentration (g/kg) of Brassica spp. by harvest date in the irrigated mid-latitude, high-elevation steppe/desert biome of the southern Rocky Mountains, USA.

 ^x Harvest 1 was taken 60 days after planting and Harvest 2 was taken after 30 days of regrowth. Due to missing cells for the second harvest, data are means of all available second harvest data but selecting first harvest means so as to both average across and also control for site (Alcalde and Ghost Ranch, NM), year (1 and 2), planting date (mid July and mid August), and, for turnip, cultivars (ForageStar and Rondo).

Crude protein concentrations in the present study may have been slightly lower than those measured by others (6,12,14), but the yields reported here were higher in the first harvest than those measured elsewhere (Table 3) (6). Hence, the lower CP may have been due to dilution. Additionally, it is not likely that N-deficiency reduced yields of regrowth cuttings because the CP concentration increased (Table 4) and no deficiency symptoms were observed. Nitrogen uptake above that applied (50 kg/ha applied vs. 52, 65, and 57 kg/ha in two harvests of kale, rape, and turnip forage, respectively, as calculated from Tables 3 and 4) would likely have been residual soil N, which was approximately 30 kg/ha. This phenomenon of greater N uptake than N applied by forage brassicas is consistent with findings of other researchers as calculated from their published results (2,5,6,8,12).

Conclusions

Planting brassicas in mid July in the irrigated mid-latitude, high-elevation steppe and desert environments similar to the southern Rocky Mountains of the USA provides earlier availability of forage than planting in mid August, although the yields of the first harvest might be reduced by high summer temperatures. If the purpose is to extend the grazing season, producers would still benefit more by planting earlier to avoid low temperatures in autumn and waiting 60 DAP or longer to graze. Rape has greatest potential for use in 60 DAP, single grazing systems in these environments while turnips had lower initial yield but greater potential for regrowth and sustained nutritive value. Yield potential for kale during this study was too low to be considered of much value for multiple-cut systems in the irrigated mid-latitude, high-elevation steppe and desert environments similar to the southern Rocky Mountains of the USA.

Acknowledgments

We gratefully acknowledge the technical assistance of David J. Archuleta and Val S. Archuleta; office assistance from Phyllis Moya, Dora Valdez, Augusta Archuleta, and Patricia Lopez; and our coworkers at the NMSU Library Document Delivery Service.

Literature Cited

1. AOAC, 1990. Official Methods of Analysis, 15th Edn. Association of Official Analytical Chemists, Arlington, VA.

2. Fraser, M. D., Winters, A., Fychan, R., Davies, D. R., and Jones, R. 2001. The effect of harvest date and inoculation on the yield, fermentation characteristics and feeding value of kale silage. Grass Forage Sci. 56:151-161.

3. Guillard, K., and Allinson, D. W. 1988. Yield and nutrient content of summer- and fall-grown forage Brassica crops. Can. J. Plant Sci. 68:721-731.

4. Guldan, S. J., Martin, C. A., and Daniel, D. L. 1998. Interseeding forage Brassicas into sweet corn: Forage productivity and sweet corn yield. J. Sust. Agric. 11:51-58.

5. Jung, G. A., Byers, R. A., Panciera, M. T., and Shaffer, J. A. 1986. Forage dry matter accumulation and quality of turnip, swede, rape, Chinese cabbage hybrids, and kale in the Eastern USA. Agron. J. 78:245-253.

6. Jung, G. A., and Shaffer, J. A. 1993. Planting date and seeding rate effects on morphological development and yield of turnip. Crop Sci. 33:1329-1334.

7. Kalmbacher, R. S., Everett, P. H., Martin, F. G., and Jung, G. A. 1982. The management of Brassica for winter forage in the sub-tropics. Grass Forage Sci. 37:219-225.

8. Kunelius, H. T., Sanderson, J. B., and Narasimhalu, P. R. 1987. Effect of seeding date on yields and quality of green forage crops. Can. J. Plant Sci. 67:1045-1050.

9. Littell, R. C., Milliken, G. A., Stroup, W. W., and Wolfinger, R. D. 1996. SAS System for Mixed Models. SAS Institute Inc., Cary, NC.

10. Milliken, G. A., and Johnson, D. E. 1992. Analysis of Messy Data, Volume I: Designed Experiments. Chapman & Hall, New York, NY.

11. Prestbye, L. S., and Welty, L. E. 1993. Evaluation of winter brassica varieties for forage production. Montana AgRes. 10:11-14.

12. Rao, S. C., and Horn, F. P. 1986. Planting season and harvest date effects on dry matter production and nutritional value of Brassica spp. in the Southern Great Plains. Agron. J. 78:327-333.

13. Wiedenhoeft, M. H. 1993. Management and environment effects on dry matter yields of three Brassica species. Agron. J. 85:549-553.

14. Wiedenhoeft, M. H., and Barton, B. A. 1994. Management and environment effects on Brassica forage quality. Agron. J. 86:227-232.

A contribution of the New Mexico Agricultural Experiment Station, New Mexico State University, Las Cruces, NM.