Journal list menu

Volume 5, Issue 1 e20235
ORIGINAL RESEARCH ARTICLE
Open Access

Evaluation of temporal variability on tissue nutrient concentrations of canola

Vaughn Reed

Vaughn Reed

Dep. of Plant and Soil Sciences, Oklahoma State Univ., Stillwater, OK, USA

Contribution: Data curation, Writing - review & editing

Search for more papers by this author
Apurba K. Sutradhar

Apurba K. Sutradhar

Eastern Agricultural Research Center, Montana State Univ., Sidney, MT, USA

Contribution: Formal analysis, Writing - original draft

Search for more papers by this author
D. Brian Arnall

Corresponding Author

D. Brian Arnall

Dep. of Plant and Soil Sciences, Oklahoma State Univ., Stillwater, OK, USA

Correspondence

D. Brian Arnall, Dep. of Plant and Soil Sciences, Oklahoma State Univ., Stillwater, OK, USA.

Email: [email protected]

Contribution: Conceptualization, Data curation, Funding acquisition, Methodology, Project administration, Resources, Writing - review & editing

Search for more papers by this author
Josh Lofton

Josh Lofton

Dep. of Plant and Soil Sciences, Oklahoma State Univ., Stillwater, OK, USA

Search for more papers by this author
First published: 10 January 2022

Assigned to Associate Editor Yuanshuo Qu.

Abstract

The use of plant tissue analysis as a tool for diagnosing nutrient deficiency has steadily gained interest by producers and crop advisors in Oklahoma. A study was conducted during the 2012–2013 and 2015–2016 growing seasons at two locations near Stillwater and Perry, OK, to evaluate the variability of canola (Brassica napus L.) nutrient concentrations across various growth stages (STAGE), among several days within each growth stage (DAY), across several times during the day (TIME). Canola tissue samples were collected in the morning, noon, and evening for three consecutive days at rosette, postdormancy break, and prebolting stages. Samples were analyzed for macronutrients (N, P, K) and secondary (S, Ca, Mg) nutrient concentrations. Tissue nutrient concentrations were found to be extremely variable, as affected by main effects and interactions of sampling TIME, DAY, and STAGE. Nutrient concentration generally increased over time, for most nutrients. The analysis did not indicate any definitive pattern of tissue nutrient accumulation based on the fixed effects TIME and DAY. Although, the exact reasons for the different responses are not known, they may be related to soil physical and chemical property differences and variation in weather factors. While this work does not provide insight in the relationship between nutrient concentrations and crop yield, this does bring to light complications with using tissue analysis for management decisions. Further work must be done to investigate the relationships of tissue concentration and yield, and the viability of tissue sampling for nutrient management.

Abbreviations

  • ARS
  • Agronomy Research Station
  • DAY
  • days within each growth stage
  • ICP
  • inductively coupled plasma
  • LCB
  • Lake Carl Blackwell
  • STAGE
  • growth stages
  • TIME
  • times during the day
  • 1 INTRODUCTION

    Winter wheat (Triticum aestivum L.) production is the predominant cropping system in Oklahoma, with average yields of 2.43 Mg ha−1 in the last 5 yr (USDA NASS, 2021). In the region, wheat is primarily grown continually, with little rotation. This has led to increased production problems due to stagnate yields, increased disease incidence, and greater competition from problematic grassy weeds. Winter canola (Brassica napus L.) was brought into the southern Great Plains regions as a means to overcome these production challenges. However, to maximize both yields and production benefits, canola must be managed properly.

    In order for canola to be successfully grown and integrated into production systems, nutrient recommendations and appropriate fertility recommendations must be developed. At present, the most common way to determine the nutrients required for optimum growth is through soil sample analysis. However, for several years now, there has been an increased call for the use of plant tissue analysis as a management tool in development of fertilizer programs. While soil tests typically examine nutrient levels in the upper 15 cm of the soil profile, plant tissue analysis can integrate the nutrient pools present at the various soil depths (Silveira et al., 2019). Moreover, interest in plant tissue analysis has been fueled by increased use of field scouting, farmers and crop consultants’ higher level of technological sophistication, increased crop yields due to fertilization, and continued need for higher yields (Fageria, 2014; Schulte & Kelling, 1991).

    In Oklahoma, there is growing interest in using nutrient concentration in plant tissue as a method for fertilizer recommendation. Part of the challenge with relying on tissue analysis for diagnosing nutrient deficiencies is that tissue nutrient concentrations vary by plant age, time of sampling, cultivar, and environmental conditions (Bates, 1971). Even when tissues of the same physiological age are selected, considerable variability still remains. Variations are mostly attributed to differences in soil nutrient availability, carbohydrate concentrations, and translocation of nutrients (Jarrell & Beverly, 1981; Marschner, 2002; Mills & Jones, 1996). Walworth and Kilby (2002) found that concentration of water-soluble nutrients, such as nitrogen (N), phosphorus (P), and potassium (K), tend to decrease as leaves age, while concentrations of other nutrients tend to increase with time. In a collaborative study in 2010, it was observed that N and magnesium (Mg) concentrations of corn (Zea mays L.) were lower later in the day compared with early morning, while other nutrient concentrations did not vary during the day (Mundorf et al., 2015). Furthermore, they reported that time of day effects occurred inconsistently across nutrients. Nitrogen concentrations were generally high early in the day, while P, K, calcium (Ca), Mg, and sulfur (S) concentrations varied inconsistently with time of day in both corn and soybean [Glycine max (L.) Merr.]. These results display that timing of sampling affects concentration of nutrients. This would indicate that plant tissue testing should be paired with preseason or in-season soil testing. However, under numerous scenarios, plant tissue analysis has contradicted soil test results and “rescue” foliar applications, or application occurring due to visual deficiency, have shown limited impact on yield (Kruger, 2001).

    Recommendations that integrate plant tissue analysis often utilize sufficiency ranges and have been developed for winter canola (Plank & Tucker, 2000). Within the specified range, the nutrient concentration would be considered adequate. Above or below this range, the nutrient concentration would be considered excessive or deficient. As opposed, a critical level is that concentration below which yields decrease or deficiency symptoms appear, and crop production could be improved by addition of nutrients (Jones & Eck, 1973). Since the exact concentration of a nutrient below which yields decline is difficult to precisely determine, others define the critical level as the nutrient concentration where 10% yield reduction would be expected (Campbell & Plank, 2000). In crop production, the objective is for the plant tissue results of a given crop to fall within the sufficiency range of a given nutrient. Yet, the use of sufficiency ranges has its limitations. Studies have shown that time of day in which tissue samples are collected impacts different nutrient concentrations inconsistently (Mundorf et al., 2015). Nutrient uptake can be different throughout the growth stages, which can impact concentrations found in plant tissue (de Oliviera Silva et al., 2021). Environmental effects that can change throughout the day have been shown to also impact nutrient concentrations (Andrews et al., 2001). Variations in plant nutrient concentration due to temporal fluctuations may necessitate more detailed protocols for the times at which tissue samples should be collected, as well as more detailed test calibration data.

    The objective of this work was to observe fluctuations of tissue concentrations of macronutrient and secondary nutrients temporally in canola. With this objective, temporal variability in the concentrations of nutrients in canola was investigated by sampling across growth stages (STAGE), among several days within each growth stage (DAY), across several times during the day (TIME). Our hypothesis was that element concentration in canola tissues may vary in TIME and DAY, not just overgrowth STAGE.

    Core Ideas

    • Canola tissue nutrient concentrations were affected by TIME, DAY, and STAGE samplings.
    • Stage of sampling produced the greatest changes in concentration.
    • The effects of TIME, DAY, and STAGE samplings were inconsistent across nutrients.

    2 MATERIALS AND METHODS

    Field experiments were conducted at the Oklahoma State University Agronomy Research Station (ARS), Stillwater, OK, in 2012–2013 and 2015–2016, and at Lake Carl Blackwell Research Station (LCB) near Perry, OK, in 2015–2016. Canola cultivar ‘DKW46-15’ (Dekalb, Monsanto Company) was planted at 19-cm row spacing at a target rate of 5.6 kg seeds ha−1. Agronomic practices for canola production followed the Oklahoma Agriculture Experiment Station and Oklahoma Cooperative Extension Service recommendations (Oklanola, 2015). The soil at ARS location was a Norge (fine-silty, mixed, thermic Udertic Paleustoll) loam, and at LCB location was a Pulaski (coarse-loamy, mixed, superactive, nonacid, thermic Udic Ustifluvent) fine sandy loam.

    Composite soil samples consisted of at least 10 soil cores collected from each location. Soil samples were taken by growth stage (rosette, postdormancy, prebolting) at the ARS in 2015–2016 season and at LCB. Only rosette stage samples were taken at ARS in 2012–2013 season. Soil samples were dried at 60 °C overnight and ground to pass a 2-mm sieve. All samples were analyzed for soil pH, NO3–N, plant available P and K index, soil organic C, and other macronutrients and micronutrients. Soil pH was measured by glass electrode pH meter in a 1:1 soil/water suspension and Sikora buffer solution, respectively (Sikora, 2006; Sims, 1996). Soil NO3–N was extracted with 1 M KCl solution and quantified by a Flow Injection Auto analyzer (LACHAT, 1994). Plant available P, K, Ca, and Mg were extracted using Mehlich 3 solution (Mehlich, 1984). Phosphorus, K, Ca, and Mg in the extract were then quantified by a Spectro CirOs inductively coupled plasma (ICP) spectrometer (Soltanpour et al., 1996). Soil organic C was determined using a LECO Truspec dry combustion carbon (C) analyzer (Nelson & Sommers, 1996). Soil sulfate (SO4–S) is extracted by 0.008 M Ca3(PO4)2 and analyzed by a Spectro CirOs ICP. Plant available Zn, Fe, Mn, and B are extracted by DTPA-sorbitol and quantified by ICP. Soil test reports are presented in Table 1.

    TABLE 1. Results of soil test analyses at different stages collected at the Agronomy Research Station (ARS) near Stillwater, OK, in 2012–2013 and 2015–2016 and Lake Carl Blackwell (LCB) near Perry, OK, in 2015–2016
    ARS 2012–2013b ARS 2015–2016 LCB 2015–2016
    Soil characteristicsa Rosette Rosette Postdormancy Prebolting Rosette Postdormancy Prebolting
    Organic matter, g kg−1 14.0 13.0 13.0 15.0 11.0 9.0 11.0
    pH 6.1 6.0 6.0 6.6 6.6 5.4 5.2
    NO3, mg kg−1 7.2 1.5 1.0 1.0 1.5 7.5 5.0
    P, mg kg−1 27.5 16.5 12.5 11.5 25.5 25.5 25.0
    K, mg kg−1 148 113 140 159 148 82 78
    SO4−2, mg kg−1 5.0 3.5 3.0 8.0 6.5 4.0 3.0
    Ca, mg kg−1 1,474 1,324 1,430 1,682 1,763 743 679
    Mg, mg kg−1 340 265 337 380 323 184 165
    Fe, mg kg−1 24.0 28.0 14.0 18.4 42.2 26.2 38.2
    Zn, mg kg−1 0.8 1.1 0.4 0.6 0.9 0.3 0.5
    B, mg kg−1 0.09 0.11 0.09 0.12 0.20 0.06 0.08
    Cu, mg kg−1 0.80 0.90 0.40 0.60 1.40 0.80 0.90
    • a Soil organic matter, dry combustion; pH, 1:1 soil/water; NO3, 1 M KCl; P, K, Ca, and Mg, Mehlich 3; SO4–S, 0.008 M Ca3(PO4)2; Fe, Zn, B, and Cu, DTPA-sorbitol.
    • b Data for postdormancy and prebolting stages are not available. Composite soil samples were collected from 0-to-15-cm depth.

    Prior to planting, the entire plot was fertilized via surface broadcast with 34 kg ha−1 urea (NH2–CO-NH2; 460–0–0 g kg−1 N–P–K), 56 kg ha−1 ammonium sulfate [(NH4)2SO4; 210–0–0–240 g kg−1 N–P–K–S] and 112 kg ha−1 di-ammonium phosphate [(NH4)2HPO4; 180–460–0 g kg−1 N–P–K].

    All plant tissue samples were randomly collected from an area of 9-by-4.5-m plot, with no sampling in the first 0.6 m of row to avoid field-edge effects. The experimental design was a randomized complete block with a 3 × 3 × 3 factorial arrangement (three growth STAGE, three DAY, and three TIME samplings), replicated three times.

    Normal and healthy (no visual insect or disease damage) plant tissue samples were collected by hand clipping the whole plant 5 cm above the soil surface at three different stages; namely rosette, postdormancy break, and prebolting. The whole top method for sampling canola was used since whole tops are suitable for identifying major nutrient status as well as trace elements. At every stage, tissue samples were collected over a period of 3 d ( Table 2). For each day, samples were collected in the morning (8:00–9:00 a.m.), noon (12:00–1:00 p.m.), and evening (5:00–6:00 p.m.). Each sampling time consisted of three sets of 15 plants.

    TABLE 2. Date of canola sampling at different stages collected at the Agronomy Research Station (ARS) in 2012–2013 and 2015–2016, and Lake Carl Blackwell (LCB) near Stillwater, OK, in 2015–2016
    Location
    Growth stage Day ARS, 2012–2013 ARS, 2015–2016 LCB, 2015–2016
    Sampling date
    Rosette (fall) 1 11 Dec. 2012 8 Dec. 2015 15 Dec. 2015
    2 12 Dec. 2012 9 Dec. 2015 16 Dec. 2015
    3 13 Dec. 2012 10 Dec. 2015 17 Dec. 2015
    Postdormancy (early spring) 1 5 Mar. 2013 7 Mar. 2016 21 Mar. 2016
    2 6 Mar. 2013 8 Mar. 2016 22 Mar. 2016
    3 7 Mar. 2013 9 Mar. 2016 23 Mar. 2016
    Prebolting (late spring) 1 11 Apr. 2013 14 Mar. 2016 28 Mar. 2016
    2 11 Apr. 2013 14 Mar. 2016 28 Mar. 2016
    3 11 Apr. 2013 14 Mar. 2016 28 Mar. 2016

    Samples were sent to Oklahoma State University Soil, Water, and Forage Analytical Laboratory the morning after sampling. The samples were oven dried at 85 °C overnight and ground to pass 1-mm screen. Total N was determined using a dry combustion N Analyzer (LECO Truspec). The rest of the mineral contents were analyzed by a Spectro CirOs ICP Spectrometer (SPECTRO Analytic Instruments Inc.) following wet digestion (Undersander et al., 1993). An analysis of variance of treatment effects was conducted on tissue nutrient concentrations. Tissue nutrient concentration data were analyzed for each location after preliminary statistical analysis indicated a significant location effect. Data were analyzed according to a randomized complete block design using PROC GLIMMIX of SAS 9.4 (SAS Institute, 2011), assuming fixed treatment effects and random effects. Time of day (TIME), day (DAY), growth stage (STAGE), and their interactions were considered as fixed effects and blocks as random effects. Differences between the treatment means were assessed by using LINES option of the LSMEANS statement of SAS output pairwise comparisons of means only when the main effects were significant at p ≤ .05. SLICE function of SAS was used to perform a partitioned F-test LSMEANS means when interactions were significant.

    For interpretation of tissue nutrient concentration results, researchers used Plank and Tucker (2000) sufficiency ranges for canola, found in Table 3. Plank and Tucker sampled uppermost recently matured leaf blades prior to flowering to develop nutrient sufficiency ranges for canola. To facilitate observed nutrient concentration interpretation, nutrient concentrations were compared between samples at various growth stages with the lowest and highest suggested values (Plank & Tucker, 2000 ) at prebolting growth stage, as there were no other known available sufficiency ranges for rosette and postdormancy break. It is advised that the interpretation of nutrient sufficiency could be affected.

    TABLE 3. Critical sufficiency ranges for macronutrients in canola, adopted from Plank and Tucker (2000)
    N P K Ca Mg S
    %
    4.00–6.40 0.42–0.69 3.50–5.10 2.10–3.00 0.15–0.62 0.65–0.90

    3 RESULTS AND DISCUSSION

    3.1 Environmental condition

    Weather data for each experimental site during 2012–2013 and 2015–2016 growing seasons were obtained from Oklahoma Mesonet website and summarized in Figure 1 (http://www.mesonet.org/index.php).

    Details are in the caption following the image
    Bars in the figures represent precipitation while the lines are temperatures. Precipitation observed are monthly averages and normal precipitation is the average monthly total precipitation for 30 yr from November to April. Tmax and Tmin observed are the monthly average maximum temperatures while Tmax normal and Tmin normal monthly average of 20-yr maximum and minimum temperatures

    Differences of daily maximum and minimum temperatures at all three sampling STAGEs of canola at the study sites are presented in Table 4. At ARS in 2012–2013, early growing months (seedling germination to 4–8 leaf stage) were slightly warmer as compared with the 20-yr normal (Figure 1). Monthly average maximum (Tmax) and minimum (Tmin) temperatures were near the 20-yr normal from November 2012 to April 2013. November and December of 2012 were slightly drier months than the 30-yr normal but was generally wetter from February to April 2013 except in March. In general, monthly average temperatures and total precipitation indicated that 2012–2013 season was a typical year for growing winter canola.

    TABLE 4. Differences of daily maximum and minimum temperatures at rosette and prebolting stages of canola in Agronomy Research Station (ARS), near Stillwater, OK, from 2012–2013 and 2015–2016, and Lake Carl Blackwell (LCB), near Perry, OK, from 2015–2016
    Location
    Growth stage Day ARS, 2012–2013 ARS, 2015–2016 LCB, 2015–2016
    Temperature
    °C
    Rosette (fall) 1 13 10 13
    2 14 18 10
    3 19 15 9
    Postdormancy (early spring) 1 4 18 11
    2 3 15 18
    3 11 13 20
    Prebolting (late spring) 1 12 19 21
    2 10 13 12
    3 14 13 13

    In 2015–2016, maximum and minimum temperatures from November 2015 to April 2016 were 3–4 °C higher than the 20-yr normal, except in January 2016, when temperatures were slightly lower (Figure 1). Total monthly precipitation was higher (38–51 mm) than normal in the fall but was slightly lower or near normal from January to March 2016. This site had 50 mm more precipitation in April 2016 than normal.

    At LCB in 2015–2016, monthly temperatures and total precipitation were similar to ARS 2015–2016 (Figure 1). Maximum and minimum temperatures from November 2015 to April 2016 were 1 to 5 °C higher than the 10-yr normal except in January 2016 when maximum temperature was 2 °C lower than the average. Total monthly precipitation was above average in the months of November and December but was 25–38 mm below average in the months of January–March.

    3.2 Tissue nutrient sufficiency range

    In this study, tissue N concentrations were within the sufficiency range all stages and locations, except for rosette stage at ARS 2015–2016 and LCB 2015–2016 (Figure 2). Studies have shown that spring canola cultivars generally accumulate greater N in tissues at first flower growth stages, and then decreased toward maturity regardless availability (Chamorro et al., 2007), which is supported by our results.

    Details are in the caption following the image
    Mean tissue macro- and secondary- nutrients concentrations as affected by sampling TIME (M = morning, N = noon, and E = evening), DAY, and growth STAGE. Error bars indicate standard error of the mean. Horizontal dotted lines represent sufficiency ranges for each nutrient, adopted from Plank and Tucker (2000)

    Mean tissue P concentration was below the lower limit of the sufficiency range at prebolting at all three sites. Based on soil test results ( Table 1) and current Oklahoma State University interpretations, soil-test P values were below the lower limit of P index, which was considered to be deficient for canola production (Zhang et al., 2017). Diammonium phosphate (180–460–0 g kg−1 N–P–K) was applied at the rate of 112 kg ha−1 to offset this deficiency. Fertilizer-P application slightly increased tissue P concentration later in the season at the ARS studies but never achieved values within the sufficiency range. This is an indication that P was not readily available for canola especially at LCB and may have resulted in lower tissue P concentration at postdormancy break and prebolting. This slower absorption of P at early growth stage was similar to the results reported by Yang et al. (2010).

    Tissue K concentration was below the lower limit of the sufficiency range at all days, times, and stages at ARS in 2015–2016, and at LCB. Tissue K concentration was only found to be within sufficiency range at ARS in 2012–2013 at the prebolting stage. Based on soil test results (Table 1) and current Oklahoma State University interpretations, soil-test K was below sufficiency at the ARS in 2015–2016 at the Rosette stage, and at LCB at both the postdormancy and prebolting stages. Soil-test K was above sufficiency levels at the ARS in 2012–2013 season at the rosette stage.

    Tissue S concentrations were below the lower limit of sufficiency range at rosette and postdormancy stage at ARS in 2012–2013. Sulfur uptake increased to its sufficiency range at prebolting. In 2015–2016 study, mean concentration of S was within the sufficiency range at all growth stages. At LCB, tissue S concentration was below the lower limit of sufficiency range at all sampling points. Tissue concentration of Ca was below the lower limit of sufficiency range at ARS 2012–2013 and at LCB. Tissue Ca concentration indicated that Ca was not deficient at ARS in 2015–2016 study. Tissue Mg concentrations were within the sufficiency range at all site-years regardless of sampling TIME, DAY, and STAGE.

    3.3 Effect of time, day, and growth stage on tissue nutrient concentration

    A significant main effect was present for sampling TIME, DAY, and STAGE, and their interactions for measured nutrient concentrations for 3 site-years are presented in Table 5. Least significant means for the main effects for each location are presented in Figure 2.

    TABLE 5. Significance of F values for fixed sources of variation from statistical analyses on macronutrients
    Sources of variation Nutrient concentration
    N P K S Ca Mg
    P > F
    Agronomy Research Station, 2012–2013
    Time <.01 .03 .01 <.01 .76 .67
    Day .20 .07 .16 .91 .07 .41
    Stage <.01 <.01 <.01 <.01 <.01 <.01
    Day × time .39 .76 .40 .07 .02 .01
    Stage × time .11 .69 .02 .07 .03 .13
    Stage × day .13 .05 .28 .14 .82 .68
    Stage × day × time .17 .71 .07 .03 .15 .10
    Agronomy Research Station, 2015–2016
    Time .11 1.00 .42 .30 .42 .06
    Day .22 .84 .79 .02 <.01 .02
    Stage <.01 <.01 <.01 <.01 <.01 <.01
    Day × time .55 .04 .34 .77 .39 .11
    Stage × time .31 .61 .36 .77 .13 .56
    Stage × day .87 .09 .22 <.01 .31 .76
    Stage × day × time .82 <.01 .77 .14 .02 .23
    Lake Carl Blackwell, 2015–2016
    Time .59 .05 .14 .64 .83 .28
    Day .42 .28 .06 .31 .66 .44
    Stage <.01 <.01 <.01 <.01 <.01 <.01
    Day × time .07 .42 <.01 .46 .19 <.01
    Stage × time .46 .34 .10 .10 .03 .93
    Stage × day .92 .34 .70 .71 <.01 .78
    Stage × day × time .60 .17 .04 .09 <.02 .09

    Effect of TIME and DAY on tissue nutrient concentration varied among locations. Effect of STAGE was also significant for all nutrients but present across all locations. Additionally, interactions involving STAGE were found and will be discussed later in the publication.

    Nitrogen concentrations was significantly influenced by TIME at the ARS 2012–2013 and STAGE at all locations but was never significantly affected by DAY. At ARS 2012–2013, greater concentrations were found at both morning and noon compared with evening sampling. At all locations, tissue N concentration was greater at postdormancy break and prebolting stage compared with rosette (α = .05, p < .001), but no difference could be found between postdormancy and prebolting. The soil test analysis values taken from different growth stages at ARS 2015–2016 and LCB did not reflect any increase in plant available N that would lead to such an increase in concentration in the tissue. No interactions between TIME, DAY, and STAGE were found across all locations for N concentration.

    Tissue P concentration was affected by main effect only for TIME and STAGE at LCB, and time at ARS 2012–2013. At ARS 2012–2013, morning provided the highest P concentration, significantly different from evening but not different from noon. At LCB, noon had greater P concentration than morning but did not differ with evening (p < .001). As for STAGE, both rosette and postdormancy stages had greater P concentration than prebolting but were not statistically different from each other. Interactions between the main effects were found at the other locations. ARS 2012–2013 had an interaction between STAGE and DAY, while ARS 2015–2016 had a three-way interaction between all main effects, STAGE, DAY, and TIME. When ARS 2012–2013 was sliced by STAGE, variability between days only occurs at the prebolting stage, while sliced by DAY provided variability across all days. At ARS 2015–2016, where interaction between all main effects was found, when sliced across STAGE, significant variability was found at the rosette stage, while other stages, variability was not significant. When sliced across TIME and DAY, variability was found across all slices.

    Tissue K concentration was affected by main effect only for STAGE at ARS 2015–2016. As growth stage progressed at ARS 2015–2016, K concentration increased, with each stage being significantly different from each other, and rosette stage having the greatest K concentration (p < .001). At ARS 2012–2013, an interaction was significant between the STAGE and TIME. When sliced by STAGE at this location, variability was found in the prebolting stage, while no variability was found between the postdormancy and rosette stages. When sliced by TIME, variability was found across all times. A three-way interaction was found at LCB, and when sliced by DAY or TIME, variability was found across all temporal variables. However, sliced by STAGE, prebolt and rosette stages were the source of variability, with postdormancy not having significant variability.

    Tissue S concentration was affected by main effect only for STAGE at LCB. The prebolting and postdormancy stages had the greatest S concentration, significantly different from rosette stage. Two-way interaction between STAGE and DAY was present at ARS 2015–2016. When sliced by STAGE, all three stages had significant variability. Sliced by DAY, only DAY 3 did not present significant variability. A three-way interaction was present at ARS 2012–2013 for S, and both prebolting and rosette stages had significant variability present when sliced by STAGE. When sliced by DAY or TIME, variability was significant across all iterations of temporal variables.

    Tissue Ca concentration was never affected by main effect only across all locations, with some permutation of interaction being present. At ARS 2012–2013, both DAY ´ TIME and STAGE ´ TIME interactions were present, yet no three-way interaction was found. When sliced by STAGE, the rosette stage was the only growth stage that had significant variability. When sliced by DAY and TIME, variability was significant across all iterations. Three-way interaction was significant at both ARS 2015–2016 and LCB locations. For both locations, all slices by DAY and TIME displayed significant variability across all iterations. When sliced by STAGE, both postdormancy and prebolting stages had significant variability at ARS 2015–2016, while rosette and prebolting stages had significant variability at LCB.

    Tissue Mg concentration was affected by main effects only at all locations. STAGE at ARS 2012–2013 was a significant effector of Mg concentration, as each stage was significantly different from each other, with rosette having the greatest concentration, and postdormancy had the lowest concentration. Both STAGE and DAY were significant effectors at ARS in 2015–2016. As stage progressed, there was a significant increase of Mg concentration, with prebolt stage having the greatest concentration, while rosette had the lowest concentration. Observing the DAY variable, across all stages and time at this location, Day 3 had the highest concentration but was only different from Day 1. STAGE was a significant effector of Mg concentration at LCB, following the same pattern as ARS 2015–2016, with concentrations increasing as stages progressed. There was a two-way interaction at ARS 2012–2013, between DAY and TIME. When sliced by DAY, Day 1 was the only day with significant variability, however when sliced by TIME, no time was found to have significant variability (α = .05), but trends were found at the morning and noon timings (p = .0668, p = .0668). A similar interaction was found at LCB as well, between DAY and TIME. Sliced by DAY, Day 2 was the variable that had significant variability, and sliced by TIME, both morning and evening had significant variability.

    Canola tissue nutrient concentrations were affected by TIME, DAY, and STAGE samplings, but the effects were inconsistent across nutrients. Most nutrient concentrations remained within if already within acceptable ranges, or outside if already outside of acceptable ranges on DAY or TIME intervals. STAGE produced the greatest changes in concentration, most often increasing the concentrations into sufficiency ranges yet across all temporal intervals, variation remained relatively large. Seasonal fluctuations of nutrient concentrations have been reported prior and contributed to the vegetative growth functions occurring during each stage, which supports our findings of STAGE variation (Stateras & Moustakas, 2017 ).

    Difference in concentration in the morning can be expected, in that remobilization or increased transport of nutrients due to growth demands and transpiration are not yet in full force (low flow rate of water and nutrients); but by noon to early evening, high water flux rates occur leading to low nutrient content (Mundorf et al., 2015; Schurr, 1998). However, not all nutrients are affected similarly. Plant concentrations of more mobile nutrients, compared with less mobile nutrients, would be expected to be more affected by TIME sampling. Yet our results provide that nutrients considered immobile in the plant have DAY and TIME variation, such as Ca and S (Figure 2).

    Large differences of daily high and low temperatures on each sampling day at rosette and prebolting stages may have contributed to the low nutrient tissue concentration in canola (Table 4). Large differences in temperatures can lead to high utilization of nutrient in the plants, leading to a lower nutrient concentration in plant tissues (Taiz & Zieger, 2006). Many studies have demonstrated the strong influence of temperature on uptake of different nutrients (Asare & Harlin, 1983; Chopin et al., 1995; Duke et al., 1989; Engels & Marscher, 1996). Temperature is not an isolated issue; nutrient concentration depends on other environmental factors. Further studies are required to determine if and how different environmental factors affect nutrient tissue concentration in canola.

    Differences in day to day climatic conditions can also impact daily nutrient concentrations as well, for similar reasons as described for TIME. Differences in concentration due to DAY among growth stages found in this study can be compared with a winter wheat study reported by Roth et al. (1989). In this study, nutrient concentrations were significantly different in 1985 and 1986, on consecutive days on three and eight occasions, respectively.

    The variation in nutrient concentration within TIME and among DAY samplings raise concern about reliance on plant tissue analysis as a diagnostic tool for fertilizer recommendation for canola production. The effect of TIME on tissue nutrient concentration can probably be eliminated by adjusting sampling time until nutrient remobilization or transport of nutrients are more stable within a day (Mundorf et al., 2015), but the effect of DAY sampling may be more of a concern as this sampling event is, again, more likely affected by daily climatic conditions.

    Tissue nutrient concentrations also varied widely as toward maturity. No visual deficiency symptoms were observed in canola plants, despite several nutrient concentrations were below the lower limit of the sufficiency range. The absence of deficiency symptoms in the canola may suggest that there is lower requirement for of those nutrients, or the proposed critical tissue nutrient concentrations may be high for canola grown in Oklahoma. A study in Arkansas displayed that the addition of fertilizer boron (B) did not increase tissue B concentration enough to reach the sufficiency level, nor did it affect canola yield, suggesting that the proposed critical tissue concentration may be high for canola grown in that area (Slaton et al., 2009).While significant variation of nutrient concentration within these short-term sampling parameters call into question the validity of plant tissue testing, nutrient concentration shifting from above the critical concentration or within the sufficiency range to below these values also create a major concern. These changes would create a discrepancy of whether to make an application or not based on TIME or DAY sampled. Additionally, many nutrients did not increase above the lower limit of the sufficiency range. Questions remained as to whether or not these sufficiency ranges can be used for winter canola production. These differences outline the high temporal variability.

    Nutrient accumulation dynamics are complex and take into account several external factors not investigated in this study but can still be influenced by different times of the day/week/stage. Soil solution electrical conductivity, which can change in a growing season due to producer inputs and precipitation, has been shown to impact mineral concentrations in Brassicas (Ding et al., 2018). Nutrient movement within a plant is related to the rate of transpiration at any given time as affected by the cumulative strength of other sinks, such as rapidly growing leaves or reproductive organs (Mundorf et al., 2015). Nitrate uptake models depict that light cycles and solar radiation, and soil temperature can impact nitrate accumulation as well (Malagoli et al., 2004). The concentration of certain nutrients can impact the accumulation of other nutrients, such as those that assist in nutrient movement and diffusion (Maillard et al., 2016). These factors can make utilizing nutrient concentrations as a method for determining fertilizer recommendations challenging.

    The data summarized in this study emphasizes that, while tissue analysis is a very useful tool for diagnosing nutrient deficiency, soil-test results must be used in making nutrient management decisions.

    4 CONCLUSIONS

    Managing nutrient inputs for crop production can be a difficult activity when one considers all of the factors affecting nutrient supply from the soil and nutrient demand of the crop. Most agronomists can easily discern spatial patterns in these factors across a landscape but addressing the issue of temporal fluctuations is a challenge. This study investigated the effects of temporal intervals on nutrient concentration in winter canola, specifically TIME, DAY, and growth STAGE. Our results displayed that nutrient concentration of macronutrients (N, P, K) and secondary (S, Ca, Mg) nutrients had variability across all temporal samplings, yet no patterns could be found across locations, even across same nutrients. While this work does not provide insight in the relationship between nutrient concentrations and crop yield, this does bring to light complications with using tissue analysis for management decisions. Further work must be done to investigate the relationships of tissue concentration and yield, and the viability of tissue sampling for nutrient management.

    AUTHOR CONTRIBUTIONS

    Vaughn Reed: Data curation; Writing – review & editing. Apurba K. Sutradhar: Formal analysis; Writing – original draft. Daryl B. Brian Arnall: Conceptualization; Data curation; Funding acquisition; Methodology; Project administration; Resources; Writing – review & editing. Josh Lofton: Writing – review & editing.

    CONFLICT OF INTEREST

    The authors declare no conflicts of interest.