Tests to predict the potassium requirements of canola

Transcription

1 Tests to predict the potassium requirements of canola R.F. Brennan Department of Agriculture, 444 Albany Highway, Albany, WA 6330, Australia; ABSTRACT Significant increases in yields of dried shoots and grain of canola (Brassica napus L.) to fertiliser potassium (K) were obtained in 25% of the experiments on the predominantly sandy soils of south-western Australia (WA). The sandy soils have become K deficient for wheat production in recent years due to removal of K from the soils in grain and hay. This paper reports the results of a soil and tissue test study conducted to predict when fertiliser K was required by canola crops in WA. When percentage of the maximum (relative) grain yield responses to applied K were related to soil test K (Colwell procedure) values for the top 10 cm of soil about 50% of the variation was accounted for. No significant increases in grain yields to applied K were obtained at 75% of sites (42 of the 56 sites) for Colwell K >44 mg K/kg soil in the top 10 cm regardless of the soil test values below 10 cm. There was a grain yield increase to applied K at 9 sites (16% of sites) when Colwell K was <44 mg /kg to 30 cm. However, at 6 sites (11% of sites) there was no grain yield response to applied K even though soil test K was <44 mg/kg in the sandy top 10 cm of soil. At these 6 sites soil test K was >44 mg/kg below 10 cm so evidently adequate K was accessed by plant roots from the subsoil for shoot and grain production. The current soil test procedure (top 10 cm of soil only) did not identify that these 6 sites did not need fertiliser K. The concentration of K in dried shoots that was related to 10% grain yield response (critical prognostic K value) was about (g K/kg dried shoots) 45 at 7 weeks after seeding (WAS), 38 at 10 WAS, 27 at 13 WAS, and 20 at 18 WAS. Application of fertiliser K had no effect on either oil or K concentrations in grain. KEYWORDS: Critical K, Soil test K, % oil. INTRODUCTION There was adequate indigenous potassium (K) in soils used for cropping in south-western Australia (WA) so fertiliser K was not applied to the crops. However, due to removal of K from soil in grain and hay, most sandy soils used for cropping in WA have now become K deficient for grain production of wheat (Brennan et al. 2004). Fertiliser K now needs to be applied to these soils for profitable grain production of wheat. Since the mid 1990s, canola (Brassica napus L.) has been grown in rotation with wheat and lupin on most of these soils. The K requirements of canola on these soils are not known. The study reported here was undertaken to determine critical soil and tissue test values for canola grain production in WA. MATERIALS AND METHODS Field experiments A total of 56 experiments done from 1993 to Soils. The sites chosen for the experiments were selected so as to include a very wide range in soil K status in the top 30 cm of soil and included the major soils types used for canola production in WA. Field experiments. Each experiment comprised a completely randomised block of 3 or 4 amounts of fertiliser K replicated 4 times. Three amounts of K were used (0, 30 and 60 kg K/ha), and 4 amounts of K were used from (0, 15, 30 and 60 kg K/ha). Muriate of potash, (50%K) was used in all the experiments. The K treatments were applied to the soil surface of each plot 4 weeks after seeding. Basal fertilisers were applied to ensure that K was the only nutrient element to limit canola production. Basal fertilisers placed (drilled) with the seed while sowing were (kg/ha) 170 single superphosphate (9 % P, 10.5% S), 6 copper sulfate, 2 zinc oxide, and 0.25 sodium molybdate. Basal fertilisers applied by hand to the soil surface 4 weeks after sowing were 152 kg/ha urea

2 (46% N) and 120 kg/ha gypsum (17% S). Canola cv. Narendra was sown in After 1996 TT canola cultivars Karoo ( ), Pinnacle ( ) and Surpass 501 ( ) were sown. For each cultivar, 4-5 kg seed/ha was sown about 2 cm deep. Measurements Soil test K. Sites for the experiments were selected in February -March to ensure sites were uniform for soil type to 30 cm depth. Soil samples were collected to 30 cm depth at 25 random locations, and separated into the following profile depths: 0-10, 10-20, cm. Soil test K was measured as described by Colwell and Esdaile (1968). Shoot yields and concentration of K in dried shoots. Yields of dried shoots were measured about 7 weeks after sowing (WAS) at 8 sites, 10 WAS at 11 sites, 13 WAS at 12 sites, and about 18 WAS (mid flowering) at 12 sites. Yields were measured by cutting canola shoots at ground level within 5 random 40 cm by 100 cm quadrats per plot. The canola shoots were dried for 3 days in a 70 C forced-draught oven before weighing. Subsamples of the dried shoots were used to measure concentration of K in the shoots. The dried samples were ground and digested in sulfuric acid and hydrogen peroxide and the concentration of K in the digest was measured by flame photometry. Grain yields and concentration of K and oil in grain. Grain yields were measured by machine-harvesting the middle 6 rows of each plot and weighing the grain. Subsamples of the harvested grain were used to measure concentrations of K in grain, using the same method to measure concentrations of K in dried shoots, and concentration of oil in grain. Grain yields, and concentration of oil in grain, were corrected to 8.5 % moisture content of grain. Analysis of data Relating yield responses to soil test K. Previous unpublished data, showed that when >30 kg K/ha was applied canola shoot and grain yields were on the maximum yield plateau of the relationships between yield and the amount of K applied. In this study we therefore assumed the 60 kg K/ha treatment was on the maximum yield plateau for the relationship between yield and the amount of K applied. Percentage of the maximum (relative) grain yield responses to applied K were calculated as follows: y = [(Y K60 - Y K0) /Y K60 ]*100 (1) where y was the relative grain yield response (%), Y K60 was the grain yield produced when 60 kg K/ha was applied, and Y K0 was the grain yield produced when no K fertilizer was applied. Data for the relationship between relative grain yield response to applied K and Colwell soil test K were fitted to an exponential equation: y = a + bexp(-cx) (2) where y was the relative grain yield response (%), x was the soil test K (mg K/kg soil), and a, b and c were coefficients. The soil test K value that was related to a yield response of 10% was calculated using the fitted equation, and was defined as the critical soil test K value below which grain yield responses to applied fertiliser K are likely and vice versa. Estimating critical K concentrations in dried shoots. For each time of sampling dried shoots, data for the relationship between relative dried shoot yield response to applied K and the concentration of K in the dried shoots were fitted to equation 2, where y was the relative shoot yield response (%) to applied K, x was the concentration of K in dried shoots (g K/kg dried shoots), and a, b and c were coefficients. The fitted equations were used to determine the K concentration in dried shoots that was related to a relative shoot yield response of 10% to provide critical diagnostic tissue test values for shoots at each time of sampling (7, 10, 13 and 18 WAS). Likewise, the relationship between relative grain yield response to applied K and the concentration of K in dried shoots, were fitted to equation (2), where y was the relative grain yield response to applied K (%), and x was the concentration in dried shoots (g K/kg dried shoots). This was done separately for shoots collected at 7, 10, 13 and 18 DAS. The K concentration that was related to a grain yield response of 10% was determined from the fitted equations, to provide critical prognostic tissue test values at each sampling time. For each time

3 of sampling, tissue test values below the critical prognostic value indicate likely decreases in grain yield due to K deficiency in the shoots, and vice versa. RESULTS AND DISCUSSION Dry shoots yield and grain yield responses to applied K Significant yield responses of dried shoots and grain to applied K occurred in sandy soils at 14 of the 56 experiments (about 25% of the experiments). Soil test K Soil test K values and soil depth. Colwell soil test K values were largest in the top 10 cm of soil for 43 out of the 56 sites (78% of sites). For the remaining 13 sites Colwell soil test K values were either about similar in the top 30 cm (5 sites) or the values increased down to the 30 cm depth (8 sites). The 5 sites had soil test K values that ranged from 42 to 48 mg/kg in the top 10 cm. The 8 sites had soil test K values in the top 10 cm that ranged from 42 to 56 mg/kg. Relating grain yield response to soil test K values Fig. 1. Relationship between relative canola grain yield and Colwell K values for top 10 cm of soil. Fitted relationship for y = exp(-0.08x), R 2 = Symbols [Colwell K values] ( ) high and decreasing, ( ) high and increasing, ( ) low and uniform, ( ) low and increasing, ( ) low and decreasing, and ( ) low with poor grain yields. When relative grain yield responses of canola to applications of fertiliser K were related to Colwell soil test K values measured in 0-10, or cm profile depths, for all 3 profile depths, about 50 per cent of the variation was accounted for (Fig. 2). Therefore soil test K of the top 10 cm adequately reflected the soil K status of the soils for canola grain production. Critical soil test K values (the soil test value that was related to 10% grain yield response to applied K) was 44 mg/kg in the top 10 cm of soil. Only sandy soils showed significant grain yield responses to applied K in our study. The non-responsive sites had a diverse range of soils, including sandy soils, that had >44 mg/kg to 30 cm. However, there were a total of 6 sites that had soil test K values <44 mg/kg in the top 10 cm of soil that showed no significant grain yield responses to applied fertiliser K (Fig. 2). At all 6 sites the soil test K values were >44 mg/kg in the cm soil horizon. Evidently at these 6 sites, when canola roots grew deeper into the soil, they accessed adequate K for grain production of canola, so that despite the inadequate K status of the top 10 cm of soil, no grain yield responses to applied K occurred at the 6 sites.

4 At 75% of sites soil test K was >44 mg/kg in the top 10 cm of soil and no grain yield responses to applied K occurred regardless of the soil test K value below 10 cm. Also at most sites (50 of the 56 sites, or 89% of sites) in our study the top 10 cm of soil adequately estimated the need for fertiliser K for canola grain production. This was because the soil test either identified adequate K in the topsoil so testing the subsoil was not important, or the inadequate K identified in the topsoil coincided with there being inadequate K to 30 cm (or deeper) for canola grain production. Only at 6 sites was K inadequate in the topsoil but adequate in the subsoil and the current soil test procedure failed to detect the subsoil K because only the topsoil was collected to measure soil test K. We therefore recommend that fertiliser K be applied to canola crops if a Colwell soil test K value of <44 mg/kg is obtained for the top 10 cm of soil. However this does mean that fertiliser K will be applied to some soils with >44 mg/kg Colwell soil test K in the subsoil requiring no fertiliser K for canola grain production, as occurred at 6 of the 56 sites (11% of sites) in our study. We recommend soil testing be done before sowing either wheat or canola on sandy soils in WA, and to apply fertiliser K to wheat and canola crops for soil test K <50 mg/kg. Critical K values for lupin crops are not known and can not be determined because to date lupin crops have shown no responses to applied fertiliser K in WA. The need to apply fertiliser K to wheat and canola crops on some K deficient sandy soils in WA may also maintain an adequate K status in these soils for lupin crops grown in rotation with wheat and canola. Concentrations of K in dried shoots Applications of fertiliser K increased concentrations of K in dried shoots of canola (data not shown). For experiments in which yield of dried shoots was measured several times during the year, the diagnostic and prognostic concentration of K in dried shoots tended to be similar for each time of sampling. Diagnostic and prognostic K levels decreased as the growing season progressed, a well established trend for concentration of most nutrient elements in plant tissue. Relative yield response of dried shoots (%) K c o n c e n t r a t io n in d r ie d s h o o t s ( g / k g ) Fig.2. Relationship for 36 experiments between K concentrations in dried canola shoots at 7WAS, 10WAS and 13WAS and relative grain yield response (prognostic tissue test). Symbols ( ) 7WAS, ( ) 10WAS and ( )10WAS and grain ( ) 13WAS. Concentrations of K in grain Concentrations of K in the grain (g /kg) ranged from (median value 6.5) for the nil- K treatments, compared with (median value 6.6) when 60 kg K/ha was applied.

5 Consequently, application of the largest amount of K in our study had a negligible effect on the K concentration in grain, with increases ranging from 2-5%, most increases not being significant. Our finding support results of Holmes and Ainsley (1977, 1978) that applications of K fertiliser had little effect on the K concentration of canola grain. Effect of applied K on oil concentrations in canola grain Oil concentrations were measured in 46 of the 56 experiments. Applications of K had no significant effect on the concentration of oil in grain in all experiments, except at 2 sites. Application of fertiliser K at 1 site increased the %oil by about 3.3 %, and at the other it decreased the %oil by about 2.3%. These results agree with Holmes (1980) that applications of fertiliser K had no effect on oil concentration of canola grain. DISCUSSION When first cleared for agriculture the soils in WA had adequate K for crops. However, due to removal of K from soil in grain and hay, recent research has shown that many sandy soils in WA are now deficient in K for grain production. These sandy soils are now also deficient in K for grain production of canola. The concentration of K that was related to 10% of the relative grain yield response (critical prognostic K) was about (g K/kg) 44 for shoots collected at 7WAS, 38 for samples collected at 10WAS, 27 for samples collected 12WAS, and 21 for shoots collected at 18 WAS. These prognostic concentrations agree favourably with g/kg with previous results (no growth stage indicated, Holmes 1980). About 37 g/kg was found to be the critical diagnostic concentration in 42-day old canola plants in a glasshouse study (Brennan and Bolland 2004). We could find no other data for critical K levels in whole shoots of canola. There are published data for either youngest mature leaves or mature leaves for which g/kg is regarded as adequate (Weir and Cresswell 1994). Tissue test values of K at flowering (18 WAS) do not allow sufficient time to apply fertiliser K to that crop for grain production. Therefore we suggest plant samples need to be collected earlier, such as at about 7 WAS. ACKNOWLEDGMENTS Funds were provided by the Grain Research and Development Corporation (DAW 0075) and by the Western Australian Department of Agriculture. Messrs. J Majewski, FM O Donnell, TD Hilder and RJ Lunt provided technical assistance. REFERENCES Brennan, R.F, M.D.A. Bolland, and J.W. Bowden 2004: Potassium deficiency, and molybdenum deficiency and aluminium toxicity due to soil acidification, have become problems for cropping sandy soils in south-western Australia. Aust. J. Exp. Agric 44, (). Colwell J.D. and R.J. Esdaile 1968: The calibration, interpretation, and evaluation of tests for the phosphorus fertiliser requirements of wheat in northern New South Wales. Aust. J. Soil Res. 6, Holmes M.R.J. 1980: Nutrition of the oilseed rape crop. Applied Science Publishers: London. Holmes M.R.J., and Ainsley A.M. 1977: Fertiliser requirements of spring oilseed rape. J. Sci. Food Agric. 28: Holmes MRJ, and A.M. Ainsley (1978) Seedbed fertiliser requirements of winter oilseed rape. J. Sci. Food Agric. 29, Weir R.G., and G.C. Cresswell 1994: Plant nutrient disorders. 4. Pastures and field crops. Inkata; Melbourne.