1. Introduction
Current North Dakota soybean fertility recommendations have been developed from data collected in the eastern third of North Dakota. This project evaluated selected soil fertility treatments aimed to improve fertilizer management for soybean producers in northwest and northcentral North Dakota [
1]. Northcentral and northwestern North Dakota’s climate is drier and cooler than the more typical North Dakota soybean growing regions [
2]. As a result of a relatively short growing season, climate, soybean growers are recommended to plant Group 0 maturity soybeans in Fargo, whereas 00 and 000 maturity group soybeans should be planted in northcentral and northwestern North Dakota [
3].
No soil fertility research had been conducted in northwest and north central North Dakota until these studies were initiated in 2016. Due to different soil types and climate, fertility research is needed in the new and non-traditional soybean growing areas to help improve grain quality and yield, and to maximize soybean farmer economic return on crop input investments [
4].
A large portion of the total nitrogen (N) required for soybean is often produced from the activity of the symbiotic N-fixing rhizobium (
Bradyrhizobium japonicum L.) [
5]. Consequently, N fertilizer is typically not recommended. Excess soil NO
3− can hinder rhizobia nodulation [
6]. Soil NO
3− levels greater than 90 kg ha
−1 have prevented nodulation and reduced yield. [
7]. The use of N fertilizer in soybean is most advantageous as a rescue application for when soybean nodules are not present [
4]. The use of nitrogen fertilizer can have little effect on soybean grain yield [
7] economic benefit to soybean farmers [
8].
Virgin soybean fields lack the rhizobia for symbiotic N fixation to occur. Without supplemental N or rhizobia, soybean yields are reduced [
9]. When soybean is grown in a field for the first time, a popular inoculation strategy is to use two different inoculum forms to better ensure successful inoculation [
10,
11,
12,
13]. Soybean seldom responds positively with inoculation if the field had been seeded to properly inoculated soybean previously [
13,
14,
15,
16]. If more than five years pass between soybean grown in a field, inoculation is suggested [
16].
Many have theorized [
17,
18,
19,
20,
21,
22] that an in-season N application could improve soybean yield and/or quality by increasing the sink of N. Dryland research suggests that there is little response from an in-season N application made during the soybean reproductive stage [
18]. However, a high yielding environment like irrigation may provide a yield response [
21]. Still, the cost of N inputs may not be agronomic [
20].
External soil and environmental factors influence symbiotic N fixation in soybean. Acidic soils with a pH less than 5.5 reduce soybean nodulation [
22]. Dry soil conditions reduce rhizobia movement to soybean root. The frequency of reported positive N fertilizer responses due to lack of adequate nodulation in soybean tends to increase under droughty conditions [
23].
Rhizobia require an oxygen (O
2) free environment to reduce N
2 gas into ammonium (NH
4+). Leg-hemoglobin is used by rhizobia to void the environment of O
2. Cobalt is needed to form leg-hemoglobin by synthesizing cobalamin, or more commonly known as vitamin B
12 (C
63H
88CoN
14P). Healthy and active nodules’ pinkish appearance is caused by leg-hemoglobin [
24]. Jayakumar and Jaleel [
25] increased soybean yield from 50 mg Co kg
−1 soil treatment compared to the untreated check. Cobalt applications greater than 100 mg Co kg
−1 soil may be toxic to soybean or the rhizobia [
25]. There are currently no scientifically accepted Co soil thresholds for rhizobia health [
26]. Soil testing for Co is difficult [
27].
Soybean grain removes approximately 14.5 kg P for each 1000 kg grain yield [
28]. The greatest P demand by soybean occurs during the reproductive stage [
29]. Despite the soybean yield gain of 0.5% per year [
30] and increase of soybean P content [
31], responses of soybean due to P fertilizer application have been inconsistent [
32,
33,
34,
35,
36,
37].
Results are mixed as to whether greater soybean yields can be obtained by broadcasting or banding P fertilizer, although the highest proportion of experiments indicates that broadcast P tends result in higher soybean yield compared to banded P. Several studies have recorded delayed soybean emergence and stand from P fertilizer application applied in furrow with no adverse yield impact [
32,
38,
39,
40]. Bullen et al. [
41] observed greater yields with banded P than broadcasted P. Others have observed no yield response when comparing broadcast or banded P [
32,
35,
42]. The mixed results could be from soybean being more efficient at acquiring soil P compared to other crops such as rape (
Brassica napis L.) or oat (
Avena sativa L.) [
43].
Potassium (K) is needed for the activation of several enzymes, water uptake, and it aids in moderating turgor pressure [
44]. Soybean has a high K demand and removes more than twice the amount of K that a regional corn grain harvest removes [
29]. Over time a soil can be depleted of K as a result of soybean production. One Mg of soybean grain removes about 22 kg of K
2O and removes more K
2O than corn > wheat = oat > barley (
Hordeum vulgare L.) [
28].
Available soil K in North Dakota is analyzed using the 1-N ammonium acetate method (NH4OAc) [
45]. When the K soil test indicates a high probability of soybean yield response soil test guidelines are used to determine fertilizer rates [
22,
46]. Some have observed that the 1-N ammonium acetate method has not predicted K response of soybeans consistently. Soybean may more effective than other crops at mining soil K [
47]. The reliability of soil K analysis may be caused by pH, clay mineralogy, soil moisture, redox potential, weathering, and different K pools. Potassium in the soil is mainly unavailable and comprised as a primary mineral [
48]. Breker [
46] evaluated various K soil testing methods and found the NH4OAc method [
45] to be most accurate. However, it’s been found that that when the clay mineralogy is accounted for, K soil testing thresholds can be shifted to better reflect a fertilizer response based off of the NH4OAc K soil test [
49]. As a result, current North Dakota K fertilizer guidelines have been adjusted to reflect the impact of K retention influenced by difference clay type minerals [
1].
Foliar feeding offers possible practical benefits to farmers who might add a fertilizer to a post-emergence herbicide or other pest management treatment to save trips over a field. To minimize leaf damage, foliar feedings should occur in the morning when temperatures are relatively cool and the plant is turgid [
50]. Soybean yield improvement due to foliar fertilizer application tend to be infrequent and small. Increase of oil or protein content are unlikely or insignificant [
37,
50,
51].
Plants require iron (Fe) in chlorophyll production and function [
52]. Soybean is susceptible to Fe deficiency chlorosis (IDC). Iron deficiency chlorosis typically occurs during V1 through V5 growth stages [
53]. Soybeans improve Fe uptake by secreting organic acids that acidifies, chelates, and enables Fe reduction the rhizosphere, converting Fe
3+ to the much more soluble Fe
2+ [
54]. Planting an IDC tolerant soybean variety is the most practical, effective, and agronomically viable IDC management tool than using an iron chelate [
55,
56]. The application of soil applied Fe o-o-EDDHA chelate may reduce IDC severity [
54,
56,
57,
58]. However, foliar applications of o-o-EDDHA had little impact on soybean yields versus seed applied o-o-EDDHA [
57]. Improvement of IDC has been observed under severe conditions from oat companion crops, due to uptake of soil NO
3− and reduced soil moisture by the oat [
56].
A majority of the soils in the Midwest US can provide soybean with a sufficient amount of micronutrients. Soybean rarely responds positively to the use of the micronutrients [
59,
60]. Research conducted in North Dakota [
61,
62] and Minnesota [
63] did not observe a consistent positive yield and quality response from the use of S fertilizer. A recent meta-analysis [
64] did not observe a soybean grain yield or composition improvement due to sulfur fertilization when exposed to drought stress. Soil organic matter can impact sulfur fertilization as the meta-analysis did observe a significant response when the soil organic matter levels were between 25 and 32 g kg
−1.
Soybean does not require supplemental N fertilizer if
Bradyrhizobium japonicum is present in the soil in sufficient numbers through residual bacteria populations or through inoculation of the seed. Soybean requires P and K; however, a yield or quality response will not necessarily result from application even if the soil test P and K analysis is identified as ‘low’. Foliar fertilizers have tended to provide little benefit to soybean yield or quality in previous trials [
51]. Iron deficiency chlorosis is best managed by cultivar selection [
55]. Liming soils is important to soybean production if the soil is too acidic (pH is less than 5.5) as nodulation is greatly reduced [
4].
2. Materials and Methods
Prospective sites were screened each spring to establish experiments in one high pH (pH > 7.2), one low pH (pH < 6.2), and low Olsen P (P < 8 g kg
−1). To screen, six cores were collected at the 0–15 cm and 15–60 cm depth by a 19 mm diameter hand probe. The cores were composited and sent to the North Dakota State University Soil Testing Laboratory, Fargo, ND for analysis (
Table 1). pH was determined with a 1:1 soil:deionized water slurry [
65] Soil nitrate was measured by the cadmium reduction method [
66]. Sodium bicarbonate extractant (Olsen) was used to determine P [
67]. Soil K levels were assessed by the NH4OAc method [
46]. Zinc, iron, and copper were measured by DTPA extraction [
68]. Sulfur was determined with a monocalcium phosphate extractant [
69]. Soil salinity was assessed with a 1:1 soil:distilled water slurry [
70]. Moisture of sugar beet waste lime (SBWL) was determined by comparing mass before and after a sample was oven dried for 24 h at 105 °C. Calcium carbonate equivalent was determined by hydrochloric acid dissolution [
71].
The Noonan and Columbus site soil types were Williams loams (fine-loamy mixed superactive frigid Typic Argiustolls). The Minot site soil type was Aastad loams (fine-loamy mixed supractive frigid Pachic Argiudolls). The Riverdale site soil type was a Wilton silt loam (fine-silty mixed superactive frigid Pachic Haplustoll) [
72]. Initial soil tests and site locations are shown on
Table 1 and
Figure 1 respectively. Soybean cultivars were chosen based on maturity (00 maturity or shorter) and IDC rating. Soybeans planted at the Columbus sites were rated IDC tolerant (
Table 2) [
4,
73,
74,
75].
Experimental units were 3.04 m wide and 9.14 m long. Experimental sites were managed using best management practices as described by Kandel [
3]. Soybeans were seeded at a rate of 370,500 pure live seeds ha
−1 with a custom-built single disk opener cone plot planter. Planter row spacing was 16.8 cm. The experimental design was a randomized complete block with twelve treatments and four replications. The fertilizer treatments were: check, inoculation (
Bradyrhizobium japonicum L.), hand applied urea (46-0-0) (55 kg ha
−1), hand applied monammonium phosphate (11-52-0) (110 kg ha
−1), hand applied waste sugarbeet (
Beta Vulgaris L.) lime (4400 and 8800 kg ha
−1). Nutrient contents of fertilizer treatments and sugarbeet waste lime are reported in
Table 3 and
Table 4 respectively. The urea was treated with N-(n-butyl)-thiophosphoric triamide (Agrotain Advance 1.0
®) to reduce ammonia volatilization [
77]. Waste sugarbeet lime treatments were applied only at the low pH Minot and Riverdale sites (
Table 1) [
4].
Ammonium polyphosphate (10-34-0), orthophosphate-polyphosphate (6-24-6), Fe ortho-ortho-EDDHA [
78], and naked ortho-ortho-EDDHA [
79] fertilizers were applied as a liquid in furrow at 7.1 L ha
−1 (
Table 4) The ortho-ortho-EDDHA treatments with and without Fe were applied only at the Columbus and Noonan sites (
Table 2) [
4]. Foliar treatments (
Table 2 and
Table 3) were applied with a hand boom sprayer at the V5 (@V5) and R2 (@R2) growth stage [
80]. The foliar treatments were a mixture of anhydrous ammonia-phosphoric acid-potassium hydroxide based liquid fertilizer (3-18-18) and 3-18-18 with ammonium sulfate (+AMS) (1.1 kg ha
−1) applied at a rate of (28 L ha
−1). Sites 5 and 6 had an additional treatment of in-furrow cobalt-sulfate applied at a rate of 2.9 kg ha
−1 [
4].
Table 5.
Mean and range of soybean grain yield of each site.
Table 5.
Mean and range of soybean grain yield of each site.
Site | Parameter | Yield |
---|
| | Mg ha−1 | p-Value | Variance |
---|
1 | Mean | 2.50 | 0.269 | 70.05 |
Range | 2.04–3.19 |
2 | Mean | 2.07 | 0.868 | 19.496 |
Range | 1.98–2.12 |
3 | Mean | 1.74 | 0.971 | 26.823 |
Range | 1.50–1.86 |
4 | Mean | 1.86 | 0.609 | 6.641 |
Range | 1.72–1.99 |
5 | Mean | 1.24 | 0.400 | 8.827 |
Range | 1.07–1.39 |
6 | Mean | 1.46 | 0.545 | 8.138 |
Range | 1.25–1.63 |
A lack of soil moisture was an issue during these experiments. Precipitation during the 2016, 2017, and 2018 growing seasons were below the long-term regional average [
81,
82,
83]. The 2016 growing season was slightly warmer than the 30-year climatic mean [
81]. Whereas, the 2017 and 2018 growing season temperatures were near normal [
82,
83].
Soybeans were harvested using a Kincaid 8-xp small plot combine [
84] and cleaned by a vacuum-type seed cleaner [
85] before yields, protein, and oil content were determined. Grain yield was determined by dividing the cleaned harvested seed mass by the plot area. Oil and protein content were measured using a DA 7200 NIR analyzer [
86] and was corrected for 13% moisture. Individual site analysis of variance was performed using the PROC GLM procedure of SAS software version 9.4 [
4,
87].
4. Discussion
The analysis of variance procedure determined that all fertilizer treatments across all environments did not impact soybean grain yield, protein content, and oil content at the 0.05 significance level (
Table 5,
Table 6 and
Table 7). This agrees with several others [
7,
8,
13,
14,
15,
16,
18,
33,
34,
35,
36,
37,
38,
39,
51,
54,
59,
60,
61,
62,
63,
64].
However, the lack of precipitation could have impacted soybean yield and may have contributed to the lack of fertilizer response. Water stress shortens the period of seed fill which reduces grain yield [
88] and can lessen protein and oil content Water stress tends to cause seed abortions and increase seed size. Proteins are more likely to accumulate in the larger source to sink ratio of fewer and larger seeds [
89].
Adequate soil moisture is important to soybean production during the R1 to R6 growth stages [
80] and leads to more complete pod development and seed-fill [
90]. Rainfall in our experiments during the 2016, 2017, and 2018 growing seasons were below the long-term regional average [
81,
82,
83] and less than the 406 mm required by a North Dakota soybean crop [
91]. Additionally, rainfall was particularly low during the important R1 to R6 growth stages [
80]. Ideal rainfall has been defined as approximately 70 mm at the R3 to R4 growth stages [
92]. No-till cropping predominates Northwest and Northcentral North Dakota which reduces stress from dry soils and may thereby aid late season soybean water demands. No-till improves soil available water. Advantages of residue in no-till fields increases snowfall catch and diminishes evaporation [
93]. No-till further improves dry soil conditions because water infiltration is greater than that under conventional tillage [
94]. These factors increase soil profile moisture that can be used by soybean when dry surface soil conditions are present. Brun et al. [
95] observed a 516 kg ha
−1 yield advantage of no-till soybean over conventional till soybean. This was a result of 67.8 mm more soil water during the growing season [
4].
Inoculum treatments did not impact soybean yield, oil, and protein content (
Table 5,
Table 6 and
Table 7). The mean yield, oil, and protein content were 1770 kg ha
−1, 335 g kg
−1, and 155 g kg
−1 respectively. Though not reported, various plants were pulled from the soil along experimental edges at various treatments and growth stages during the integrated pest management scouting portion of this project. All soybean plants observed appeared to have healthy and active nodules. Fields where soybean is a common crop grown within a rotation, will not likely benefit from the inoculum treatment [
8]. This indicates the previous soybean crop grown within the past four years have effectively cultured northwest and northcentral North Dakota soils with rhizobia and the confirms with current North Dakota soybean fertilizer guidelines [
1].
The urea treatment did not impact soybean yield, oil, and protein content (
Table 5,
Table 6 and
Table 7). The mean yield, oil, and protein content were 1960 kg ha
−1, 334 g kg
−1, and 154 g kg
−1 respectively. Positive nitrogen responses under dryland conditions have been found to be rare and offer little economic benefit [
7]. Unless nitrogen is used as a rescue treatment [
4].
Phosphorus treatments did not impact soybean yield, oil, and protein content (
Table 5,
Table 6 and
Table 7). The range of yield, oil, and protein content for the 11-52-0, 10-34-0, and 6-24-6 treatments were 1823–1850 kg ha
−1, 334–336 g kg
−1, and 154–155 g kg
−1 respectively. A yield response was expected from the P applications due to sites 1, 2, 3, 4, and 5 Olsen P soil tests (
Table 1) were 8 ppm or less and are considered in the “low” fertility category [
1].
Current North Dakota P
2O
5 recommendations suggest those sites should receive 11 to 60 kg ha
−1 of P
2O
5 under the soil P conditions which were present at the sites [
1]. The broadcast 11-52-0 treatment applied nearly met or exceeded the recommendation depending on the site. The row-placed starter P fertilizer treatments (
Table 4) delivered fertilizer P at rates that were less than current P fertilizer recommendations [
1]. Soybean has been reported as an efficient scavenger of soil P [
43] which might explain the lack of a fertilizer P response in our experiments.
Regional research suggests that broadcast P fertilizer applications are superior to banded applications [
42,
96]. In-furrow P fertilizer applications can reduce and/or delay germination without negatively impacting yield [
32,
38,
39]. The in-furrow 6-24-6 and 10-34-0 fertilizers did not impact yield and were less than the rate that resulted in stand/yield reduction others [
39,
42,
96].
Sugar beet waste lime treatments did not impact soybean yield, oil, or protein content (
Table 5,
Table 6 and
Table 7). The mean yield, oil, and protein content were 1873 kg ha
−1, 351 g kg
−1, and 151 g kg
−1 respectively. Soil pH less than 5.5 has previously shown to reduce soybean nodulation [
23] though all environments’ in this experiment had a soil pH was greater than 5.5. Soybean root observations indicate that the nodules were not impacted by the acidic environment.
Foliar fertilizers applied at the V5 and R2 growth stage [
80] did not impact yield, protein, or oil content (
Table 5,
Table 6 and
Table 7). The range of yield, oil, and protein content for the foliar treatments were 1748–1777 kg ha
−1, 307–312 g kg
−1, and 153–155 g kg
−1 respectively. These data agree with Mallarion and Haq [
36] and Haq and Mallarino [
97] for the V5 growth stage timing and Haq and Mallarino [
98] at the R2 growth stage. That foliar fertilizer research [
36,
97,
98] occurred on soils where P and K were at sufficient or greater levels. However, our studies had Olsen P test values categorized in the “low” to “very low” (
Table 1) range causing us to anticipate a positive response to foliar fertilizer [
1].
Soygreen and Levesol treatments did not impact soybean grain yield, oil, or protein content (
Table 5,
Table 6 and
Table 7). The mean Soygreen yield, oil, and protein content were 1683 kg ha
−1, 320 g kg
−1, and 157 g kg
−1 respectively. Whereas the Levesol yield, oil, and protein content were 1778 kg ha
−1, 320 g kg
−1, and 156 g kg
−1 respectively. The alkaline environments (
Table 1) were chemically susceptible to IDC because CO
32− were present [
72]. Carbonates can neutralize acids secreted by soybean roots that aid in the reduction of Fe
3+ to Fe
2+ [
56]. Excessive moisture can increase IDC by solubilizing more acid neutralizing CO
32− [
99]. During these experiments, soil moisture was not excessive [
82,
83,
84]. Soil pH was less than 8 (
Table 1) where IDC is more likely to occur [
100]. Additionally, soybean cultivars that were determined to be IDC tolerant [
74,
75,
76] were planted in the alkaline sites, which would decrease the potential for IDC impairing soybean growth [
55].
The cobalt treatment was added in year three of this experiment to sites 5 and 6. Cobalt did not impact soybean yield, oil, or protein content (
Table 5,
Table 6 and
Table 7). The mean yield, oil, and protein content were 1345 kg ha
−1, 334 g kg
−1, and 160 g kg
−1 respectively. Cobalt was applied to serve as a vitamin for the rhizobia and assist with the synthesis of leg-hemoglobin needed for nitrogen fixation [
24]. Greenhouse experiments have shown a positive impact from small amounts (50 mg Co kg
−1 soil treatment) of cobalt [
25].
The mean yield, oil, and protein content of the untreated check was 1770 kg ha−1, 335 g kg−1, and 155 g kg−1 respectively. Though no significant statistical difference was observed, the untreated check was the ninth highest yield, sixth greatest oil content, and fifth highest protein content.