Journal list menu

Early View
ORIGINAL ARTICLE
Open Access

Barley residue biomass, nutrient content, and relationships with grain yield

Christopher W. Rogers

Corresponding Author

Christopher W. Rogers

USDA-ARS Northwest Irrigation and Soils Research Laboratory, Kimberly, Idaho, USA

Correspondence

Christopher W. Rogers, USDA-ARS Northwest Irrigation and Soils Research Laboratory, Kimberly, ID, USA. Email: [email protected]

Contribution: Conceptualization, Data curation, Formal analysis, Funding acquisition, ​Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing - original draft, Writing - review & editing

Search for more papers by this author
Curtis B. Adams

Curtis B. Adams

USDA-ARS Columbia Plateau Conservation Research Center, Pendleton, Oregon, USA

Contribution: Data curation, Formal analysis, ​Investigation, Methodology, Resources, Software, Validation, Visualization, Writing - review & editing

Search for more papers by this author
Juliet M. Marshall

Juliet M. Marshall

Plant Sciences Department, University of Idaho, Idaho Falls/Moscow, Idaho, USA

Contribution: Conceptualization, Funding acquisition, ​Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing - review & editing

Search for more papers by this author
Patrick Hatzenbuehler

Patrick Hatzenbuehler

Agricultural Economics and Rural Sociology Department, University of Idaho, Twin Falls, Idaho, USA

Contribution: Data curation, Formal analysis, ​Investigation, Methodology, Software, Validation, Visualization, Writing - original draft, Writing - review & editing

Search for more papers by this author
Garrett Thurgood

Garrett Thurgood

Plant Sciences Department, University of Idaho, Idaho Falls/Moscow, Idaho, USA

Contribution: Data curation, Formal analysis, ​Investigation, Methodology, Validation, Visualization, Writing - review & editing

Search for more papers by this author
Biswanath Dari

Biswanath Dari

Agriculture and Natural Resources, Cooperative Extension at North Carolina Agricultural and Technical State University, Greensboro, North Carolina, USA

Contribution: Data curation, ​Investigation, Methodology, Writing - review & editing

Search for more papers by this author
Grant Loomis

Grant Loomis

University of Idaho Extension, Blaine County, Hailey, Idaho, USA

Contribution: Data curation, ​Investigation, Methodology, Writing - review & editing

Search for more papers by this author
David D. Tarkalson

David D. Tarkalson

USDA-ARS Northwest Irrigation and Soils Research Laboratory, Kimberly, Idaho, USA

Contribution: Formal analysis, Validation, Writing - review & editing

Search for more papers by this author
First published: 15 May 2024

Assigned to Associate Editor Gabriela Abeledo.

Abstract

Determination of barley (Hordeum vulgare L.) nutrient uptake in residue biomass is important for agronomic, economic, and environmental decisions. Improved understanding of grain yield, residue biomass, nutrient uptake, and their relationships are needed. Research determined these factors in 2018 and 2019 from trials of four barley classes (spring animal feed, human food, and malt, as well as winter malt), using three common cultivars of each, at five locations in southern Idaho. Production environment created the largest difference in residue biomass and nutrient uptake. Barley harvest index ranged from 0.46 to 0.52 Mg Mg−1 across feed, food, and malt classes. Compared to previous estimates, nutrient concentrations from the combined dataset were greater than for N, less than for P, and greater than for K. Correlation of grain yields to nutrient uptake (excluding Cu and Fe) had r2 ranging from 0.68 to 0.89. At current prices, economic analysis indicated that fertilizer replacement costs for total residue biomass nutrients were greater than previous estimates and could greatly exceed current sale value. These relationships and value estimates can be used to improve prediction of barley residue biomass production and nutrient uptake to guide best management practices. The decision of how to utilize these metrics (on-farm, regional, etc.) should be considered based on known variation in measured nutrient and residue data and considered in relation to the proposed task.

Abbreviations

  • CV
  • coefficient of variation
  • HI
  • harvest index
  • R/G
  • residue biomass to grain
  • RUSLE2
  • Revised Universal Soil Loss Equation 2
  • SOM
  • soil organic matter
  • USDA
  • United States Department of Agriculture
  • 1 INTRODUCTION

    Barley production within the United States is concentrated in the Northern Plains and West in areas of high elevation and/or latitude due to favorable growing conditions and industry infrastructure (USDA-NASS, 2023a). Barley has a wide range of end-uses, but the majority is grown for grain, which can be generally classified into cultivars for animal feed, human food, and malting (Badea et al., 2018). Feed cultivars are used for animal rations and focus on protein, starch, and fiber content, where high protein content can be advantageous. Food barley has been largely developed for positive characteristics related to human nutrition and health, such as protein and fiber content (Hu et al., 2016; Obert et al., 2011). Malt barley predominates US production and is largely focused on grain quality for brewing, for example, low grain protein/nitrogen concentration that must stay below threshold levels (AMBA, 2021). Due to the typical end-uses of barley, grain yield and quality parameters are widely measured, but associated residue biomass and nutrient uptake values are generally not measured or reported.

    Barley produced in the United States is primarily direct harvested with a combine for grain, which is typically the portion of the crop of most interest. However, residue biomass removal for use in animal rations, animal bedding, mushroom production, and other end-uses has become a dominant practice across much of the western United States to enhance farm income (Tarkalson et al., 2011). Agronomic research for barley has largely focused on grain yield and quality, with little focus on residue biomass. Grain yield and grain N are routinely measured to estimate protein content in barley and are the focus of many studies (e.g., Bowman et al., 2001; Luo et al., 2019; Sparrow, 1979; Stark & Brown, 1987). Less research has been done to investigate nutrient uptake in barley residue biomass under varying environmental conditions, particularly when moving beyond N, P, and K (Barłóg et al., 2020; Beare et al., 2002; Malhi et al., 2006; J. J. Miller et al., 2009; Rogers, Dari, Hu et al., 2019; Tarkalson et al., 2009, 2011). There are two primary issues that arise when attempting to calculate residue biomass nutrient uptake in real-world conditions. First, barley is primarily produced for grain and while grain yield is nearly universally determined, residue biomass is not. Second, complete analysis of nutrient concentrations in the residue biomass are not routinely conducted leading to the need to use published estimates from limited sources. Thus, an assumption of both residue biomass and nutrient concentrations must be used, which both introduce variability when estimating residue biomass nutrient uptake.

    Prediction of residue biomass is an important factor in the widely used United States Department of Agriculture's (USDA) Revised Universal Soil Loss Equation 2 (RUSLE2) (Renard, 1997). The ratio of residue biomass to grain (R/G) index was used, but a specific ratio is not provided for barley, in contrast to wheat (Triticum aestivum L.) and oats (Avena sativa L.) that have ratios ranging from 1.3 to 2.0. McClellan et al. (2012) compared residue biomass estimates based on R/G index to linear relationships between grain yield and residue biomass for wheat. Regression relationships are presented for spring and winter barley from dryland production in Eastern Washington from data collected in the 1980s and 1990s but they do not calculate barley residue biomass based on R/G index, presumably as a specific value was not defined. Alternate approaches have relied on similar principles concerning the relationship between grain yield and residue biomass that were discussed in a review and analysis paper by Tarkalson et al. (2009). Their literature review resulted in an estimated harvest index (HI), the ratio of grain to total aboveground biomass, of 0.50, or an R/G index of 1.0, based on several studies conducted between 1987 and 2008. Dai et al. (2016) expanded this idea for wheat by determining wheat class specific HIs and estimates based on the following equations:
    Harvest index HI = Grain yield / Grain yield + residue biomass Residue biomass = 1 HI / HI × grain yield $$\begin{eqnarray*} {\mathrm{Harvest}}\,{\mathrm{index}}\,\left( {{\mathrm{HI}}} \right) &=& \left( {\mathrm{Grain}}\,{\mathrm{yield/}}\left( {\mathrm{Grain}}\,{\mathrm{yield}} \right.\right.\nonumber\\ &&\left.\left.+\, {\mathrm{residue}}\,{\mathrm{biomass}} \right) \right)\nonumber\\ &&\hspace*{-6.5pc}{{\mathrm{Residue}}\,{\mathrm{biomass}} = \left( {\left( {{\mathrm{1}} - {\mathrm{HI}}} \right){\mathrm{/ HI}}} \right) \times {\mathrm{grain}}\,{\mathrm{yield}}} \end{eqnarray*}$$

    Core Ideas

    • Limited research has focused on nutrient uptake and economic value of barley residue biomass (straw + chaff).
    • Production environment created the largest differences in residue biomass and nutrient uptake across barley classes.
    • Average harvest indices (HI) were 0.51, 0.47, 0.52, and 0.46 Mg Mg−1 for feed, food, spring malt, and winter malt barley.
    • N uptake combined across classes was greater than previous estimates, P was less, and K was greater.
    • Fertilizer replacement cost of nutrients removed in residue may far exceed the sale value.

    After residue biomass estimates are determined, they then require nutrient concentrations to scale to nutrient uptake. If not directly measured, nutrient concentration is largely calculated based on fixes. These estimates are important parts of nutrient management budgets for use by agencies such as the USDA-National Resource Conservation Service (USDA-NRCS, 2001). For barley specifically, Tarkalson et al. (2009) used county-level grain yield values reported by USDA. A HI of 0.50 was used to predict residue biomass using nutrient concentration values of 6.4 g N kg−1, 0.8 g P kg−1, and 16.5 g K kg−1 residue biomass.

    A few previous studies have used estimated nutrient uptake values to predict the economic value of these nutrients in small-grain residue biomass, primarily wheat, using various methods. Included is the study by Tarkalson et al. (2009) discussed above, which used estimates from several studies in the United States and a range of prices to estimate the associated total values of residue biomass nutrients, namely, N, P, and K, for wheat and barley. M. S. Reiter et al. (2015) implemented a similar analysis that also relied on nutrient concentration estimates from the literature, but only for wheat. M. S. Reiter et al. (2015) also included S in their calculations. These studies, especially in barley, are based on limited field trial data that could be improved by investigating current two-row barley cultivars in a range of environments. Specifically, this indicates an insufficiency of current data available on barley residue biomass and nutrient uptake, both regionally and in the western United States specifically. The lack of available data makes it challenging to predict residue biomass and determine the implications of residue biomass nutrient uptake especially when the residue biomass is baled and removed. Increased understanding of nutrient uptake of modern barley cultivars in diverse environments is also critical for optimally managing soil fertility and nutrient balances in relation to nutrient management budgets in barley agroecosystems.

    This research was designed to address the following objectives: (a) measure grain yield and residue biomass for multiple barley classes under varying environmental conditions, (b) determine the range of nutrient concentrations for (N, P, K, S, Ca, Mg, Fe, Zn, Mn, and Cu) for this residue biomass, (c) develop quantitative relationships between grain yield and residue biomass nutrient uptake using multiple approaches, and (d) estimate economic value of nutrients contained in barley residue biomass in terms of fertilizer replacement costs.

    2 METHODS AND MATERIALS

    2.1 Experimental methods

    Research was conducted in major barley production areas within the Snake River Plain of southern Idaho during the 2018 and 2019 growing seasons. Studies were conducted in grower fields (Ashton, Idaho Falls, Rupert, and Soda Springs [nonirrigated dryland location]), except the Aberdeen site, which was conducted at the University of Idaho Aberdeen Research and Extension Center, Aberdeen, ID (Figure 1). Irrigated locations (Aberdeen, Ashton, Idaho Falls, and Rupert) used sprinkler irrigation on a weekly basis through hand-lines and/or wheel-lines, with rates based on reported crop evapotranspiration (Agrimet Cooperative Agricultural Weather Network). With full evapotranspiration replacement, water stress is assumed to not occur at irrigated sites (USBR, 2016). Soil series and measured soil chemical properties are presented in Table 1. In brief, soil samples were collected prior to fertilization in the spring from a 0- to 60-cm depth, except P and K, which were collected from the 0- to 30-cm based on current production guidelines. Soil pH was measured potentiometrically, soil organic matter (SOM) by loss on ignition, inorganic N by 2 M KCl extraction and spectroscopy, and Olsen P, K, and calcium phosphate extractable S by inductively coupled-optical emission spectroscopy (ICP-OES) based on R. Miller et al. (2013).

    Details are in the caption following the image
    Experimental locations for research measuring grain yield, straw biomass, and nutrient uptake of modern feed, food, and malt barley cultivars within the Snake River Plain of southern Idaho in 2018 and 2019.
    TABLE 1. Preplant soil fertility conditions and fertilizer application rates (S, spring; W, winter) for southern Idaho barley trials within the Snake River Plain.
    Soil test values Fertilizer application rateb
    Location (planting date) Elevation (m) Soil series pH SOM (g kg−1) N Supplya (kg ha−1) P (mg ha−1) K (mg ha−1) S (mg ha−1) P (kg ha−1) K (kg ha−1) S (kg ha−1)

    Aberdeen

    (April 10, S)

    (Sept. 21, W)

    1342 Declo loam (Coarse-loamy, superactive, mesic Xeric Haplocalcid). 8.1 9.0

    310 (S)

    342 (W)

    21 340 63

    17 (S)

    24 (W)

    19

    120 (S)

    93 (W)

    Ashton

    (April 29)

    1603

    Marystown Robinlee Complex

    Marystown (Fine-silty, mixed, superactive, frigid Pachic Argixerolls).

    Robinlee (Coarse-silty, mixed, superactive, mesic Durinodic Haplocalcids).

    6.2 18.5 265 19 188 29 0 19 11

    Idaho Falls

    (April 23)

    1434 Pancheri SiL (Coarse-silty, mixed, superactive, frigid Xeric Haplocalcids). 7.7 15.0 205 22 184 40 15 0 20

    Rupert

    (April 17, S)

    (Sept. 29, W)

    1267 (2018) Portneuf SiL (Coarse-silty, mixed, superactive, mesic Durinodic Haplocalcids), (2019) Sluka SiL (Coarse-silty, mixed, superactive, mesic Xeric Haplodurids). 7.7 12.5

    303 (S)

    305 (W)

    25 240 38

    15 (S)

    25 (W)

    0

    34 (S)

    93 (W)

    Soda Springsc

    (May 7)

    1760 Foundem-Rexburg Complex (Foundem: Coarse-silty, mixed, superactive, frigid pachic haploxerolls: Rexburg: Coarse-silty, mixed, superactive, frigid Calcic Haploxerolls). 6.5 20.6 190 34 409 19 39 0 6
    • Note: Values are averages of 2018 and 2019 and represent the 0- to 60-cm depth, P and K are for the 0- to 30-cm depth.
    • Abbreviation: SOM, soil organic matter.
    • a N supply (kg ha−1) = inorganic N + applied fertilizer N, where inorganic N was calculated as kg ha−1 as NH4-N + NO3-N to a depth of 60 cm (Rogers et al., 2015).
    • b Applications of N as urea, P as monoammonium phosphate, K as potassium chloride, and S as elemental sulfur.
    • c 2018 Soil data only.

    Fertilizer application for each field was conducted in coordination with growers and based on soil inorganic N levels as measured via soil testing (Table 1). Fertilizer was applied in a single preplant application and incorporated with shallow tillage or knifed-in to a depth of approximately 10 cm to minimize N volatilization (Dari & Rogers, 2021) approximately 1 week prior to planting. Nitrogen rates were based on historical yield goals for each location and are reported as N supply (kg ha−1) = inorganic N + applied fertilizer N, where inorganic N is calculated as kg ha−1 NH4-N + NO3-N to a depth of 60 cm (Rogers et al., 2015). Sulfur was applied primarily for long-term maintenance of soil test levels. Potassium concentrations are routinely greater than the recommended soil test level due to the high native content in the region's soils that measured upward of 400 mg kg−1 in the 1960s but have decreased to <250 mg kg−1 in more recent studies (Rogers, Dari, & Schroeder, 2019; Tindall & Westermann, 1994). Small amounts of K were applied as part of fertilizer blends at the Aberdeen and Ashton locations (Table 1).

    Modern barley cultivars developed for high yield and quality were chosen for each class of barley to represent genetics grown over large areas of the study region (Table 2). Studies were separated by class of barley, including both spring (feed, food, and malt) and winter (malt only) crops. Winter malt barley was only grown in Aberdeen and Rupert, as this is the primary region where it can reliably be grown due to harsher winter conditions in higher elevation and/or latitude areas. Selected cultivars have proven to have good survivability, yield, and quality in milder climate regions along the Snake River Plain and are included in this study (Marshall et al., 2019, 2020). At each location and for each barley class, a single randomized complete block design experiment was conducted with four replications of each cultivar in each year of the study. Barley was planted with 18-cm row spacing using a cone-seeder drill with planting rates of 2,000,000 seeds ha−1 for irrigated barley and 1,500,000 seeds ha−1 for dryland barley. Plots dimensions at all winter locations, except for Aberdeen, were 1.5 m wide by 3 m long, plots in Aberdeen were 1.5 m wide by 2.8 m long, and spring locations were 1.5 m wide by 4.9 m long due to equipment variation among locations.

    TABLE 2. Class and cultivar information for barley studies conducted over two-growing seasons (2018 and 2019) in southern Idaho.
    Barley class Cultivar Release organization
    Spring food Goldenhart (hull-less) USDA-ARS/Idaho Agricultural Experiment Station
    Kardia USDA-ARS/Idaho Agricultural Experiment Station
    Transit (hull-less) USDA-ARS/Idaho Agricultural Experiment Station
    Spring feed Champion Westbred, LLC
    Claymore Highland Specialty Grains
    Lenetah USDA-ARS/Idaho Agricultural Experiment Station
    Spring malt Gemcraft USDA-ARS/Idaho Agricultural Experiment Station
    LCS Odyssey Limagrain Cereal Seeds
    Voyager Aneheuser Busch InBev
    Winter malt Charles USDA-ARS/Idaho Agricultural Experiment Station
    Endeavor USDA-ARS/Idaho Agricultural Experiment Station
    WintMalt KWS Lochow Seeds

    At physiological maturity, whole barley plants were cut near the soil surface using a hand sickle from the second row of the plot from a 0.11 m2 row area and collected in paper bags for transport to driers. Samples were then dried at 55°C until a constant weight was achieved. Barley grain was removed from whole plants by hand partitioning the plant into grain and residue biomass (i.e., including all remaining components after grain separation, including grain head chaff). Samples were then weighed for biomass and subsequently ground to 1 mm using a Wiley Mill (Thomas Scientific). HI was calculated as barley grain mass divided by the total biomass (grain + residue biomass). Ground samples were used for all subsequent chemical analyses. Total C and N were determined by high-temperature combustion based on the principles of the Dumas method (Variomax CN, Elementar Americas). Nutrients (P, K, Ca, Mg, S, Fe, Zn, Mn, and Cu) were extracted by nitric acid digestion (Jones & Case, 1990) and measured via ICP-OES. Nutrient uptake in residue biomass was calculated based on measured weight of each component and measured nutrient concentrations.

    2.2 Statistical analysis

    Data were statistically analyzed using the SAS 9.4 software package (SAS Institute Inc.). Analysis of variance (ANOVA) was performed using the MIXED procedure, individually for each class of barley. Before conducting an ANOVA, the data were checked to ensure they satisfied the assumptions of normality and equal variances using histograms, QQ Plots, and plots of residuals. Location, barley cultivar, and the interaction of location and cultivar were fixed effects in the statistical model. Year (season) and the interaction of block and year were random effects in the model. Tukey's method was used for pairwise mean comparisons. All treatment effects were considered significant at p < 0.05. The MEANS procedure was used to compute means and coefficient of variation (CV) values by location and overall. SigmaPlot 15.0 software (Systat, Inpixon) was used to create graphs and conduct regression analyses. Linear regression between grain yield (Mg ha−1) and residue biomass (Mg ha−1) for each barley class and for combined data was conducted. Residue biomass was calculated based on grain yield using several methods (class-specific HI, single HI, unconstrained regression, and regression through the origin [RTO]) and compared to measured values by regression. For grain yield and nutrient uptake relationships (kg ha−1), a single-step calculation procedure combined across classes was used to allow estimation of nutrient uptake based on measured grain yield and nutrient uptake measured in the current study, which is similar to approaches using a fixed HI (or R/G index) and fixed nutrient concentrations. This single-factor RTO used a simple first-order linear regression model (y = mx + b, where m = slope and b = intercept) (b = 0). Adjusted r2, 95% confidence intervals, and standard error of the estimate (SEE) values were computed by SigmaPlot.

    2.3 Economic value of barley residue biomass nutrients

    The economic analysis focused on the commonplace choice of barley producers to either bail residue biomass, for either on-farm use or commercial sale, or leave it in the field. Regarding the choice of whether to remove barley residue biomass or leave it in the field, the main variables of consideration are the potential price that can be received for sale of bailed barley residue biomass, the costs of mowing, raking, baling, and transporting the residue biomass, the value of the nutrients in the residue biomass that would need replacement if removed. Additional benefits of retaining residue biomass are acknowledged (e.g., stabilizing soil from erosion and enriching the SOM) but were not readily quantifiable by economic analysis.

    Estimated relationships between barley grain yield and nutrient uptake from the field trial data form the basis for the economic analysis associated with four levels of barley grain yields that are representative of low to very high yields—3, 6, 9, and 12 Mg ha−1. Yields of 12 Mg ha−1 are above those typically seen in commercial fields but are presented since achieving this yield is possible and some small-plot yields in the current study were measured in this range. Nutrient uptake was estimated from grain yield based on class-specific HI and nutrient concentrations and reported for each class based on the equation.
    Nutrient uptake = Class specific HI × grain yield × class specific nutrient concentration $$\begin{eqnarray*}{\mathrm{Nutrient}}\,{\mathrm{uptake}} &=& \left( {{\mathrm{Class - specific}}\,{\mathrm{HI}} \times {\mathrm{grain}}\,{\mathrm{yield}}} \right)\nonumber\\ &&\times {\mathrm{class}}\,\,{\mathrm{specific}}\,{\mathrm{nutrient}}\,{\mathrm{concentration}}\end{eqnarray*}$$

    Additionally, the mean of all classes was calculated, and this average is reported. Also, the single-factor regression estimates for the relationships between barley grain yields and residue biomass nutrient uptake were used to calculate nutrient uptake. Historical prices for each nutrient were obtained from the fertilizer cost portions of University of Idaho Crop Enterprise Budgets (University of Idaho, 2019). The crop budgets for 2001–2019 irrigated potatoes in Eastern Idaho were used since these included prices of all nutrients (N, P, K, and S). These are nominal prices and so are representative of prices producers paid rather than relative prices (e.g., adjusted for general inflation). The historical prices were used to identify a range, from low to high, that have been observed in the study region over the past several decades. The crop budgets report in fertilizer equivalent units, N, P2O5, K2O, and S prices, and were converted to elemental values where needed (P and K) using the conversion factors in M. Reiter (2020) (Table 3). The final portion of the analysis combined the ranges of residue biomass nutrient quantities (kg ha−1) associated with the several levels of grain yields with the nutrient prices ($ kg−1) to obtain residue nutrient values ($ ha−1).

    TABLE 3. Estimates of fertilizer costs for four key nutrients (N, P, K, and S) accumulated by barley residue biomass used to calculate residue nutrient value from southern Idaho barley trials within the Snake River Plain during the 2018 and 2019 growing seasons.
    Nutrient pricesa
    Nutrient Low price ($ kg−1) Average price ($ kg−1) High price ($ kg−1)
    N 0.66 1.00 1.46
    Pb 0.18 0.36 0.55
    K 0.27 0.63 1.26
    S 0.26 0.40 0.55
    • a Prices ranges are based on historical prices in University of Idaho Enterprise Budgets for Potatoes from 2001 to 2019.
    • b P2O5 and K2O prices were converted to P and K values, respectively, based on conversion factors in Reiter (2020).

    3 RESULTS AND DISCUSSION

    3.1 Grain yield and residue biomass

    Grain yield differed (p < 0.05) by location for all barley classes (Tables 4–7). There were no grain yield differences by cultivar, though differences were trending toward significance (p ∼ 0.06) for spring food and malt barleys. For spring feed barley, grain yields averaged 6.88 Mg ha−1 where the Aberdeen and Rupert locations had greater average yields (8.15 Mg ha−1) compared to the lowest yielding Soda Springs dryland location (5.5 Mg ha−1) (Table 4). Residue biomass for spring feed barley followed a similar trend and ranged from 4.29 to 8.62 Mg ha−1. Grain yields of spring food barley averaged 6.72 Mg ha−1 and ranged from 6.01 Mg ha−1 to 7.53 Mg ha−1 (Table 5). Aberdeen and Rupert had the greatest food barley grain yields averaging 7.44 Mg ha−1. Residue biomass of food barley averaged 7.78 Mg ha−1 with a site maximum of 9.59 Mg ha−1. Spring malt barley yield averaged 6.99 Mg ha−1 across locations, with Aberdeen and Rupert averaging 8.10 Mg ha−1 (Table 6). Residue biomass for spring malt barley averaged 6.63 Mg ha−1 with a maximum of 8.60 Mg ha−1. Winter malt barley is only grown in more optimal areas in the region due to winter kill that occurs in colder production locations (Marshall et al., 2019). Winter malt grain yield averaged 7.90 Mg ha−1 and can be quite large as exhibited by the Aberdeen location, where grain yields averaged 10.5 Mg ha−1 (Table 7). Along with increased grain yield, residue biomass of winter malt barley was large with a mean of 9.22 Mg ha−1. For each class of barley, including all treatment factors, the overall CV for grain yield ranged from 25% to 33% and the CV for residue biomass ranged from 28% to 37% (Tables 4–7). Observations among production sites, seasons, and crop cultivars were within the range seen in the region and include individual plot values significantly lower and higher than the averages presented in the tables and discussed here (Marshall et al., 2019; Rogers, Dari, Hu et al., 2019).

    TABLE 4. Spring feed barley grain yield, residue biomass, harvest index, and nutrient uptake (C, N, P, etc.) averaged across three cultivars and two-growing seasons (2018 and 2019) in southern Idaho.
    Treatments Grain yield Residue biomass Harvest index C N P K
    Mg ha−1 %CV Mg ha−1 %CV Mg Mg−1 %CV Mg ha−1 %CV kg ha−1 %CV kg ha−1 %CV kg ha−1 %CV
    Location
    Aberdeen 7.82aa 27 8.46a 23 0.48b 10 3.40a 22 70.4a 41 4.81b 37 231a 32
    Ashton 5.70b 27 4.83b 23 0.54a 10 2.59b 23 32.8bc 37 2.72c 68 130b 43
    Idaho Falls 6.89ab 29 7.32a 19 0.48b 13 3.09a 20 40.3b 45 7.38a 60 137b 25
    Rupert 8.47a 22 8.62a 27 0.50b 9.5 3.53a 27 62.7a 44 6.02ab 33 203a 45
    Soda Springs 5.52b 42 4.29b 41 0.56a 6.2 1.88b 41 24.7c 57 2.09c 45 79.7c 64
    Overall 6.88 33 6.71 37 0.51 12 2.79 36 46.2 59 4.60 68 156 53
    p-values
    Location (L) <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001
    Cultivar (C) 0.4743 0.5683 0.0279 0.3602 0.7799 0.4334 0.0815
    L × C 0.6721 0.1405 0.2955 0.1885 0.0159 0.4622 0.0208
    Ca Mg S Fe Zn Mn Cu
    kg ha−1 %CV kg ha−1 %CV kg ha−1 %CV g ha−1 %CV g ha−1 %CV g ha−1 %CV g ha−1 %CV
    Location
    Aberdeen 61.0a 29 13.1a 27 19.0a 31 90.5 196 148a 62 275a 32 8.68 116
    Ashton 23.9c 33 2.86c 40 10.2c 24 107 108 87.6bc 87 266ab 45 5.34 151
    Idaho Falls 41.8b 35 6.22b 34 13.4b 35 114 152 113ab 48 149c 38 7.72 85
    Rupert 53.1a 29 12.2a 29 14.0b 30 354 123 113ab 59 203bc 31 11.2 122
    Soda Springs 20.2c 59 3.24c 46 5.13c 50 82.7 150 39.6c 71 143c 73 4.15 108
    Overall 40.0 53 7.52 67 12.4 50 150 170 100 75 207 50 7.41 125
    p-values
    L <0.0001 <0.0001 <0.0001 b <0.0001 <0.0001
    C <0.0001 0.5165 0.0470 0.2195 0.9872
    L × C 0.0535 0.0046 0.0017 0.1801 0.5572
    • Abbreviation: CV, coefficient of variation.
    • a Within a parameter, means, between locations not sharing a letter are significantly different based on Tukey's test (< 0.05).
    • b Missing p-values, indicated by –, were excluded due to lack of normality in the data.
    TABLE 5. Spring food barley grain yield, residue biomass, harvest index, and nutrient uptake (C, N, P, etc.) averaged across three cultivars and two-growing seasons (2018 and 2019) in southern Idaho.
    Treatments Grain yield Residue biomass Harvest index C N P K
    Mg ha−1 %CV Mg ha−1 %CV Mg Mg−1 %CV Mg ha−1 %CV Mg ha−1 %CV Mg ha−1 %CV Mg Mg−1 %CV
    Location
    Aberdeen 7.34aba 21 9.59a 18 0.43b 8.5 3.91a 18 78.9a 30 5.36bc 29 230a 21
    Ashton 6.01c 28 5.32c 31 0.53a 8.6 2.26c 30 36.6bc 46 3.01cd 73 137b 49
    Idaho Falls 6.13bc 26 7.80b 20 0.44b 16 3.24b 19 49.3b 51 9.86a 68 153b 29
    Rupert 7.53a 22 9.48a 28 0.45b 12 3.87a 28 74.2a 55 7.18b 47 224a 50
    Soda Springs 6.45abc 20 5.64c 24 0.54a 5.2 2.56c 24 32.9c 36 1.95d 33 114c 36
    Overall 6.72 25 7.78 33 0.47 14 3.23 32 56.8 57 5.86 79 178 46
    p-values
    Location (L) 0.0026 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001
    Cultivar (C) 0.0600 0.1349 0.0040 0.2086 0.0961 0.1131 0.2372
    L × C 0.8684 0.8181 0.7260 0.7803 0.5944 0.0610 0.7465
    Ca Mg S Fe Zn Mn Cu
    kg ha−1 %CV kg ha−1 %CV kg ha−1 %CV g ha−1 %CV g ha−1 %CV g ha−1 %CV g ha−1 %CV
    Location
    Aberdeen 57.7a 31 13.3a 26 18.0a 19 113 170 205a 59 309a 52 13.6 144
    Ashton 25.2c 48 3.79c 59 11.7b 44 125 121 95.3c 60 255ab 47 4.93 122
    Idaho Falls 43.1b 28 7.53b 36 15.9a 36 126 130 160ab 49 164c 42 10.2 88
    Rupert 51.6ab 29 13.1a 28 15.5a 33 308 112 147bc 38 214bc 31 29.9 74
    Soda Springs 30.3c 29 4.07c 32 8.29c 36 137 23.3d 42 245abc 31 0.596 346
    Overall 42.8 43 8.82 57 14.5 39 150 154 138 68 236 50 13.1 135
    p-values
    L <0.0001 <0.0001 <0.0001 b <0.0001 0.0002
    C <0.0001 0.1200 0.0258 0.7043 0.0218
    L × C 0.7049 0.7181 0.5357 0.9022 0.4254
    • Abbreviation: CV, coefficient of variation.
    • a Within a parameter, means, between locations not sharing a letter are significantly different based on Tukey's test (< 0.05).
    • b Missing p-values, indicated by –, were excluded due to lack of normality in the data.
    TABLE 6. Spring malt barley grain yield, residue biomass, harvest index, and nutrient uptake (C, N, P, etc.) averaged across three cultivars and two-growing seasons (2018 and 2019) in southern Idaho.
    Treatments Grain yield Residue biomass Harvest index C N P K
    Mg ha−1 %CV Mg ha−1 %CV Mg Mg−1 %CV Mg ha−1 Mg ha−1 %CV Mg ha−1 %CV Mg Mg−1 %CV
    Location
    Aberdeen 7.80aba 22 8.60a 23 0.48b 9.0 3.50a 23 63.3a 35 4.39a 37 210a 28
    Ashton 5.89c 21 4.33c 23 0.58a 7.3 1.84c 23 31.2bc 37 1.50b 60 116c 44
    Idaho Falls 6.88bc 26 7.19b 20 0.49b 16 3.03b 20 41.8b 51 6.20a 71 142c 27
    Rupert 8.40a 22 8.18ab 28 0.51b 8.7 3.37ab 28 56.4a 52 4.35a 38 172b 44
    Soda Springs 5.81c 36 4.81c 26 0.54a 9.4 2.12c 25 24.9c 27 1.69b 42 77.9d 44
    Overall 6.99 28 6.63 37 0.52 12 2.77 35 43.9 56 3.60 79 145 48
    P-values
    Location (L) <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001
    Cultivar (C) 0.0609 0.6844 0.0007 0.5458 0.2925 0.2606 0.0802
    L × C 0.8062 0.7157 0.5849 0.6251 0.7031 0.1350 0.3573
    Ca Mg S Fe Zn Mn Cu
    kg ha−1 %CV kg ha−1 %CV kg ha−1 %CV g ha−1 %CV g ha−1 %CV g ha−1 %CV g ha−1 %CV
    Location
    Aberdeen 60.7a 27 12.5a 28 18.2a 31 316 106 193a 30 259a 29 10.1 110
    Ashton 22.8c 33 3.11c 46 10.1c 31 210 113 89.1b 40 259a 52 4.13 110
    Idaho Falls 42.3b 30 6.21b 42 15.6b 38 213 97 127b 45 142b 45 8.39 85
    Rupert 49.4b 34 11.9a 35 12.8b 30 395 114 102b 47 192b 37 14.3 62
    Soda Springs 24.3c 44 5.29bc 36 5.70d 46 155 137 42.2c 99 143b 57 4.87 120
    Overall 40.0 49 7.84 61 12.5 49 260 121 111 62 196 49 8.39 103
    p-values
    L <0.0001 <0.0001 <0.0001 b <0.0001 <0.0001
    C <0.0001 <0.0001 0.2075 0.0015 0.0071
    L × C 0.1592 0.1208 0.9456 0.0722 0.7351
    • Abbreviation: CV, coefficient of variation.
    • a Within a parameter, means, between locations not sharing a letter are significantly different based on Tukey's test (< 0.05).
    • b Missing p-values, indicated by –, were excluded due to lack of normality in the data.
    TABLE 7. Winter malt barley grain yield, residue biomass, harvest index, and nutrient uptake (C, N, P, etc.) averaged across three cultivars and two-growing seasons (2018 and 2019) in southern Idaho.
    Treatments Grain yield Residue biomass Harvest index C N P K
    Mg ha−1 %CV Mg ha−1 %CV Mg Mg−1 %CV Mg ha−1 Mg ha−1 %CV Mg ha−1 %CV Mg Mg−1 %CV
    Location
    Aberdeen 9.65aa 16 10.5a 18 0.47a 5.7 4.24a 19 64.2a 32 5.32a 31 293a 24
    Rupert 6.08b 27 7.92b 33 0.44b 12 3.31b 33 40.8b 38 6.85a 61 184b 35
    Overall 7.90 30 9.22 28 0.46 9.9 3.77 28 52.5 41 6.08 53 239 37
    p-values
    Location (L) <0.0001 0.0005 0.0010 0.0014 0.0004 0.1228 <0.0001
    Cultivar (C) 0.2856 0.3451 0.9692 0.2531 0.5619 0.8127 0.7456
    L × C 0.1323 0.9542 0.0930 0.9180 0.5550 0.9381 0.5648
    Ca Mg S Fe Zn Mn Cu
    kg ha−1 %CV kg ha−1 %CV kg ha−1 %CV g ha−1 %CV g ha−1 %CV g ha−1 %CV g ha−1 %CV
    Location
    Aberdeen 66.5a 25 13.4a 28 21.7 27 38.7 160 156a 55 265a 45 6.93 99
    Rupert 32.5b 31 7.19b 31 11.8 28 54.3 240 111b 56 155b 37 7.39 77
    Overall 49.5 44 10 43 16.7 41 46.5 217 134 58 210 51 7.16 87
    p-values
    L <0.0001 <0.0001 <0.0001 b 0.0655 <0.0001
    C 0.2346 0.7083 0.4334 0.3887 0.1040
    L × C 0.9727 0.5621 0.5159 0.9010 0.9097
    • Abbreviation: CV, coefficient of variation.
    • a Within a parameter, means, between locations not sharing a letter are significantly different based on Tukey's test (< 0.05).
    • b Missing p-values, indicated by –, were excluded due to lack of normality in the data.

    Grain yield was regressed against residue biomass for each barley class, as well as combined across classes (Figure 2, Table 8). Due to the predominately irrigated high-input environment of the studies, little data were collected at grain yields less than 2 Mg ha−1, with winter malt plot-level yields rarely falling below 4 Mg ha−1. All regression relationships were significant with r2 from 0.43 to 0.65, with slopes ranging from 0.87 to 1.00 and predominately falling near a 1:1 linear relationship. Only the winter malt class had an intercept that differed from zero (p < 0.05), where RTO would result in slopes (m) of 0.96, 1.14, 0.97, 1.17, and 1.04 for feed, food, spring malt, winter malt, and combined data. Food and winter malt barley slopes were similar and would largely predict more residue biomass at a given grain yield compared to feed and spring malt. However, removal of winter malt from the combined regression resulted in relatively small changes in the combined data relationships indicating the winter malt data fit about the axis was not overwhelmingly driving the combined regression results as shown in Table 8. Regression slopes were steeper than those from McClellan et al. (2012) who compiled data from lower yielding nonirrigated sites in Eastern Washington near the Idaho border. Usage of the McClellan et al. (2012) equation to predict residue biomass from grain yield, while reasonable, would generally be underestimated in the current dataset, especially at high grain yields, as evidenced by their slopes of 0.83 and 0.70 and intercepts of 1.75 and 1.30 Mg ha−1 for winter and spring barley, respectively.

    Details are in the caption following the image
    Barley residue biomass (feed, upper left panel; food, upper right panel; spring malt (S), middle left panel, and winter malt (W), middle right panel, and combined data, down left panel) and harvest index (combined data, down right panel) as a function of grain yield (Regression) and also as a regressionthrough the origin (RTO). Data were collected from five locations in southern Idaho in the 2018 and 2019 growing seasons. Regression equations are listed in Table 8.
    TABLE 8. Linear regression results (Figure 2) for relationships between barley (feed, food, malt, and combined data) grain yield and residue biomass from studies conducted over two-growing seasons (2018 and 2019) in southern Idaho.
    Regression Regression through the origin (RTO)
    Data Independent variable Dependent variable ma b Adj. r2 SEE ma b Adj. r2 SEE n
    Feed Grain yield (Mg ha−1) Residue biomass (Mg ha−1) 0.87 0.7 0.63 1.5 0.96 0 0.95 1.5 120
    Food Grain yield (Mg ha−1) Residue biomass (Mg ha−1) 1.00 1.1 0.43 1.9 1.14 0 0.94 2.0 108
    Spring malt Grain yield (Mg ha−1) Residue biomass (Mg ha−1) 0.91 0.3 0.55 1.6 0.97 0 0.93 1.9 150
    Winter malt (W) Grain yield (Mg ha−1) Residue biomass (Mg ha−1) 0.88 2.3 0.65 1.6 1.17 0 0.95 2.1 44
    Mean Grain yield (Mg ha−1) Residue biomass (Mg ha−1) 0.93 0.8 0.53 1.8 1.04 0 0.94 2.0 422
    Mean (W) Grain yield (Mg ha−1) Residue biomass (Mg ha−1) 0.90 0.8 0.50 1.8 1.00 0 0.94 1.8 378
    • a In the linear regression equation (y = mx + b): m = slope and b = intercept. Measures of error are also given, including adjusted r2 (Adj. r2) and standard error of the estimate (SEE).

    HI describes the ratio of grain to total aboveground biomass, is commonly used as an indicator of crop efficiency, and has been used to estimate residue biomass based on grain yield across the United States (Tarkalson et al., 2009). HI differed by location for each barley class, where environmental differences were likely a contributing factor, but mechanistic reasons are not known or speculated at this time (Figures 4-6). Previous work in wheat to predict residue biomass has proposed defining HI based on class, a method which to our knowledge, has not been utilized for the range of barley grown in the current study (Dai et al., 2016). Average HI for individual barley classes were 0.51, 0.47, 0.52, and 0.46 Mg Mg−1 for feed, food, spring malt, and winter malt, respectively. Across cultivars and locations, feed and spring malt HI were relatively similar, as were food barley and winter malt (Tables 4–8). Feed and spring malt often have genetic similarities and are thus more phenotypically similar than they are to either food or winter malt cultivars (Marshall et al., 2019). Food varieties bred for high fiber content (Hu et al., 2016; Obert et al., 2011), as in the study, are often taller with lower yields (Marshall et al., 2019; Rogers et al., 2017). These results indicate that the HI of 0.50 used by Tarkalson et al. (2009) is quite reasonable but, as with McClellan et al. (2012), could be a source of error, causing both overestimation and underestimation of residue biomass depending on the class of barley grown. HI values of 0.33–0.43 from RUSLE2 for wheat and oat are lower than those measured for barley in the study, and thus would overestimate barley residue biomass based on grain yield. Similar to Porker et al. (2020), we found relationships between grain yield and residue biomass but did not find a consistent relationship between grain yield and HI (Figure 2).

    Four approaches were investigated for calculating residue biomass from measured grain yield. First, and similar to Dai et al. (2016), estimates of residue biomass in relation to grain yield were calculated based on measured grain yields using separate HI for each barley class (Figure 3). Second, a single fixed HI of 0.50, similar to combining classes and the previous estimates of Tarkalson et al. (2009), was used to estimate residue biomass from grain yield. Third, an unconstrained regression model of grain yield to residue biomass was determined, similar to McClellan et al. (2012). Finally, RTO between grain yield and residue biomass was investigated. Each of the four methods was compared to measured residue biomass at the individual plot level. Methods were relatively similar in fit about a 1:1 line as well as in their R2 values indicating that all relationships would be similar with minor variation. All models indicated that estimated residue biomass was generally slightly lower than measured residue biomass. The single fixed HI resulted in the poorest relationship with an r= 0.94 and a slope of 0.92, while estimation by barley class-specific HI resulted in the closest slope to the 1:1 line (m = 0.97), followed by the RTO regression (m = 0.96) and the unconstrained regression (m = 0.95). Overall, estimates would be relatively similar, particularly across the average range of grain yields in the region. While all methods were reasonable, further analyses will focus on residue biomass calculations using the HI by class approach to refine estimates corresponding with methods previously used by Tarkalson et al. (2009) for the region and recommend them for use for varying wheat classes by Dai et al. (2016).

    Details are in the caption following the image
    Comparisons of actual barley residue biomass (feed, food, spring malt [S], and winter malt [W]) to predict residue biomass, using several approaches (class-specific harvest index (HI), upper left panel; a single HI = 0.50, upper right panel; and two grain by residue regression equations, down panels, Table 8). Data were collected from five locations in southern Idaho in the 2018 and 2019 growing seasons. The solid lines are regression fit through the origin (RTO), the short-dashed lines are 95% confidence intervals, and the dotted lines are the 1:1 relationship. SEE, standard error of the estimate.

    3.2 Measured residue biomass nutrient concentrations and uptake

    Nutrient concentrations were averaged across locations and cultivars and are reported for each barley class (feed, food, spring malt, and winter malt), as well as averaged across classes (Table 9). Variation in the range of the values is indicated by the first and third quartile values that contribute to variability when scaling up based on estimated nutrient concentrations. This variation was expected since nutrients (particularly minerals) are considered the factor with the most variability in feed ingredients and nutrient availability can vary substantially across production environments (Al-Dalain et al., 2020). Residue biomass N concentrations medians ranged from 5.9 to 6.7 g kg−1, with food barley residue biomass N concentrations on the higher end with a median of 6.7 g kg−1 and winter malt barley on the lower end with a median of 5.9 g kg−1. Nitrogen concentrations from India (Narolia et al., 2013) in a malt barley study where yields were generally half or less than in the current study reported substantially lower N estimates of residue biomass N concentration of 4.4 g kg−1, but the data are similar to the compiled average from North America of 6.4 g kg−1 reported by Tarkalson et al. (2009).

    TABLE 9. Concentration data for nutrients in barley residue biomass, summarized by quartiles, for four market classes of barley (spring feed, spring food, spring malt, winter malt, and combined data).
    Nutrient Unit Q1 Median Q3 Q1 Median Q3 Q1 Median Q3 Q1 Median Q3 Q1 Median Q3
    Spring feed barley Spring food barley Spring malt barley Winter malt barley Combined
    C g kg−1 408 419 431 407 416 425 411 418 431 404 414 418 409 417 428
    N g kg−1 5.0 6.2 8.2 5.4 6.7 8.3 5.1 6.2 7.5 4.7 5.9 6.8 5.1 6.2 7.8
    P g kg−1 0.43 0.60 0.84 0.43 0.61 0.97 0.29 0.45 0.65 0.45 0.58 0.91 0.37 0.55 0.74
    K g kg−1 17 21 27 17 22 26 17 20 25 21 26 29 18 21 26
    Ca g kg−1 4.5 5.8 6.6 4.6 5.4 6.1 4.8 5.8 6.7 4.0 5.3 6.3 4.6 5.6 6.6
    Mg g kg−1 0.71 1.0 1.4 0.82 1.1 1.4 0.83 1.1 1.4 0.93 1.0 1.3 0.79 1.1 1.4
    S g kg−1 1.4 1.7 2.2 1.6 1.8 2.1 1.5 1.8 2.3 1.5 1.8 2.1 1.5 1.8 2.2
    Fe mg kg−1
    Zn mg kg−1 7.0 10 24 14 16 25 12 17 22 9.6 13 18 8.5 15 23
    Mn mg kg−1 19 28 39 22 27 38 20 27 36 17 21 28 19 26 36
    Cu mg kg−1
    • Note: The combined data are pooled across all locations, cultivars, and years (2018 and 2019) for studies conducted in southern Idaho. Q1: first quartile; median: second quartile; Q3: third quartile.

    Spring malt barley had a relatively low median residue biomass P concentration of 0.45 g kg−1 compared to other barley classes where the median was near 0.60 g kg−1 (Table 9). Tarkalson et al. (2009) reported greater P concentrations of 0.80 g kg−1 likely due to differences in cultivars and environment. Potassium was relatively elevated, but consistent across barley classes with medians ranging from 20 to 22 g kg−1 with winter barley having a median of 26 g kg−1. These values are on the higher end or greater than most other studies, where Tarkalson et al. (2009) used 16.5 g kg−1. Particularly interesting is that our median value (21.0 g kg−1) is the maximum reported value in a widely used feed ration database based on a number of studies (Feedipedia, 2015). These higher residue biomass K concentrations are likely due to luxury consumption of K due to the high soil K availability in the study region (Table 1). In the 1960s, available soil K concentrations in the region measured upward of 400 mg kg−1 but have decreased to less than 200 mg kg−1 in a study published nearly 30 years ago (Tindall & Westermann, 1994). It was determined in a recent study that a range of soils in Snake River Plain had an average of 251 mg kg−1 K (Rogers, Dari, & Schroeder, 2019). While soil K is well above sufficient levels in many areas of the region, continued production and removal of forage crops and the entirety of small-grain production (grain and residue biomass) will necessitate fertilizer additions in the future. The micronutrients Fe and Cu had a particularly large amount of variation, due to a large number of samples that were under detection limit in the 2019 dataset and seemingly high outliers. Increased understanding of residue biomass nutrient concentrations, including concentration ranges, will be beneficial not only for estimating uptake, but also by reporting the ranges that can occur both between and within classes to provide an indication of variability that can be expected when scaling up from single nutrient concentration estimates. Increased understanding of residue biomass nutrient concentrations, including concentration ranges, will be beneficial not only for estimating uptake but also for improving estimates of when residue biomass is part of feed rations as sources such as the National Research Council (2007) largely report a single value.

    Measured nutrient uptakes in residue biomass were calculated by multiplying the measured nutrient concentrations (Table 9) by the residue biomass per hectare at the individual plot level. Spring feed barley differed based on location for all measured nutrient uptake parameters (excluding Cu and Fe), cultivar for Ca and S, and the interaction was significant for N, K, Mg, and S (Table 4). Nitrogen uptake had a wide range of uptake from 24.7 to 70.4 kg ha−1. Other nutrients followed relatively similar trends with variation occurring among locations. Spring food barley differed by location for all parameters, by cultivar for Ca, S, and Mn, and higher level interactions did not occur (Table 5). Food barley residue biomass N uptake was on the higher end at 56.8 kg ha−1. Spring malt barley differed based on location for all measured parameters (excluding Cu, and Fe), and by cultivar for Ca, Mg, Zn, and Mn (Table 6). Phosphorus in spring malt barley was low as influenced by the low concentrations of P where uptake was calculated as 3.60 P kg ha−1. Large nutrient uptake occurred at times for winter barley with K being representative of this at 239 kg ha−1.

    A single-factor RTO allowing simple conversion from grain yield to nutrient uptake based on measured residue biomass and nutrient concentrations was calculated, where variation from both residue biomass and nutrient estimates contributes to the model variability (Figure 4). RTO was conducted for consistency with residue biomass and grain yield relationships discussed above, and we acknowledge that certain elements would have poorer fits with an unconstrained regression (Table 10). The 95% confidence intervals were generally small about the regression line, but the SEE was at times large. The grain yield to N uptake RTO regression had an r2 of 0.80, where each Mg of grain production resulted in 6.90 kg N uptake in the residue biomass. Phosphorus uptake in residue biomass is in lesser quantities than in the grain but is an important parameter to create farm-level uptake and removal estimates for nutrient management budgets (Leytem et al., 2017). Potassium is concentrated in barley residue biomass at harvest, where each Mg of grain produced resulted in a residue biomass uptake of 23.85 kg K. Potassium uptake in barley residue biomass was greater than any other nutrient and the high levels of K in residue biomass likely were a result of luxury consumption (Bartholomew & Jannsen, 1929), particularly because of the high levels of available soil K in the study region (Table 1). Recommendations on micronutrients in residue biomass are limited. The grain yield-residue biomass nutrient uptake relationships developed in this research provide a valuable, though variable, means to estimate potential removal of nutrients in barley residue biomass, if residue biomass is exported, even when data on residue biomass and nutrient concentrations are lacking. These findings indicate that the predicted mean value for the entire population of barley is likely to be well represented by these regressions, but individual point estimates would prove more variable.

    Details are in the caption following the image
    Barley residue biomass (feed, food, spring malt [S], and winter malt [W]) nutrient uptake as a function of grain yield. Data were collected from five locations in southern Idaho in the 2018 and 2019 growing seasons. The solid lines are regression results, the short-dashed lines are 95% confidence intervals, and the long-dashed lines are the regression results plus and minus the standard error of the estimate (SEE). No statistical results were generated for Fe or Cu due to analytical issues. Regression equations are listed in Table 10.
    TABLE 10. Linear regression results (Figure 4) for relationships between barley grain yield and nutrients in residue biomass.
    Independent variable Dependent variable ma b Adj. r2 SEE n
    Residue nutrients
    Grain yield (Mg ha−1) Residue biomass N (kg ha−1) 6.901 0 0.80 25.0 422
    Grain yield (Mg ha−1) Residue biomass P (kg ha−1) 0.612 0 0.68 3.09 418b
    Grain yield (Mg ha−1) Residue biomass K (kg ha−1) 23.85 0 0.87 68.2 422
    Grain yield (Mg ha−1) Residue biomass Ca (kg ha−1) 6.019 0 0.89 15.6 422
    Grain yield (Mg ha−1) Residue biomass Mg (kg ha−1) 2.000 0 0.82 4.09 422
    Grain yield (Mg ha−1) Residue biomass S (kg ha−1) 1.902 0 0.87 5.30 422
    Grain yield (Mg ha−1) Residue biomass Fe (g ha−1) c
    Grain yield (Mg ha−1) Residue biomass Zn (g ha−1) 0.542 0 0.68 80.85 422
    Grain yield (Mg ha−1) Residue biomass Mn (g ha−1) 0.662 0 0.82 98.60 419
    Grain yield (Mg ha−1) Residue biomass Cu (g ha−1)
    • Note: The analyses integrate all treatment factors included in the study, including data collected from five locations in Southern Idaho from four market classes of barley over 2 years (2018 and 2019). Each analysis was constrained to pass through the origin (b = 0).
    • a In the linear regression equation (y = mx + b): m = slope and b = intercept. Measures of error are also given, including adjusted r2 (Adj. r2) and standard error of the estimate (SEE).
    • b Analytical issues for a small number of samples reduced (n) for P and Mn.
    • c Missing values, indicated by –, were excluded due to lack of normality in the data.

    3.3 Prediction of residue biomass nutrient uptake

    Four grain yield targets (3, 6, 9, and 12 Mg ha−1) were used to estimate residue biomass nutrient uptake of N, P, K, and S for the four barley classes in line with Dai et al. (2016) and based on previous estimates for the region by Tarkalson et al. (2009) (Figure 5). We recognize other nutrients can represent a substantial cost when they are needed but have not quantified them as they are less commonly applied, are applied at lower application rates, have limited recommendations, and do not have robust resources available for accurate cost estimation. The mean of the four barley classes (feed, food, spring malt, and winter malt) was calculated and is also presented in Figure 5. Additionally, the measured nutrient concentrations and residue biomass (single-factor regressions) were used to calculate nutrient uptake at these grain yields (Figures 4 and 5). The HI of 0.50 and concentrations from Tarkalson et al. (2009) were multiplied by the target grain yields to allow comparison between previous estimates for N, P, and K and are indicated by horizontal lines at each grain yield level in Figure 5. Thus, data are compared across classes, based on mean values, and compared to previous estimates for the region where they are available.

    Details are in the caption following the image
    Barley residue biomass nutrient uptake (N, P, K, and S) for four grain yield levels predicted using multiple approaches. Predictions were made using barley class-specific harvest index (HI) and nutrient concentration (feed, food, spring malt [S], and winter malt [W], a mean estimate averaging across classes, and a single-factor prediction (SF-RTO) across classes based on regression of grain yield to nutrient uptake from data collected during two-growing seasons (2018 and 2019) in southern Idaho. Horizontal lines for N, P, and K are based on previous nutrient uptake estimates for the region based on HI and nutrient concentration data reported by Tarkalson et al. (2009) and average price form Table 3.

    Nitrogen uptake estimates were both greater than and less than those previously calculated for the region, where food N uptake and the measured regression were well above previous estimates at all grain yield levels (Figure 5). Hull-less high-fiber food barley has only recently been introduced into production and its varied genetics from previously researched cultivars likely play a role in this difference. Phosphorus uptake was less than previous estimates where food barley and winter malt barley had relatively large uptakes compared to spring malt and feed. Again, the recent introduction and lower number of developed cultivars of hull-less high-fiber food and winter malt in the western United States result in less time in development and, thus, less refinement in available cultivars from less efficient historical germplasm (Obert et al., 2006, 2011; Stockinger, 2021). Further, the semiarid high pH soil environment differs from most regions and may result in nutrient binding in non-accessible compounds. Similar to P, K was found in an especially high quantity with winter malt and food barley having the highest uptake. Soils in the region are quite high in K and as established earlier, luxury consumption can be a major factor. Clearly, the estimates indicate substantially greater K than previous estimates. Sulfur uptake did not vary greatly, but spring food and winter malt still produced large amounts relative to other classes.

    3.4 Economic value of barley residue biomass nutrients

    Figure 5 presents the total residue biomass nutrient quantities for N, P, K, and S at four yield levels (3, 6, 9, and 12 Mg ha−1). For comparison to previous estimates, Tarkalson et al. (2009) nutrient uptake values were used but fertilizer cost estimates were adjusted to match prices within the current study dataset, as their estimates were well over a decade old (Table 3). Values from Figure 5 were multiplied by a single average cost estimate from Table 3 for N, P, K, and S ($1.00, 0.36, 0.63, and 0.40 kg−1, respectively) and thus, result in the same trends as discussed in the residue biomass uptake section presented in terms of monetary value (Figure 6). The total N, P, K, and S (S was not included by Tarkalson et al., 2009 but is a relatively small cost and thus was retained throughout) estimates indicates that estimates from the current dataset result in generally greater nutrient removal for N and K but lower for P. Despite P being greater in Tarkalson et al. (2009), the relatively small quantity of P in residue biomass (most P is translocated to barley grain) did not greatly sway cost estimates. Thus, fertilizer replacement value of nutrients estimated from Tarkalson et al. (2009) was both greater than and slightly less than for N, greater for P, and less for K. In the region, spring malt barley represents the largest class of production and thus, the lower concentration estimates would need to be considered when scaling up based on production amounts. The large removal, especially in winter malt and spring food, would influence estimates when grown in greater amounts. Total nutrient value (N, P, K, and S) for the four barley classes based on average prices ranged from $58.59 to $81.74, $117.17 to $163.47, $175.76 to $245.21, and $234.35 to $326.94 ha−1 for grain yields of 3, 6, 9, and 12 Mg ha−1, respectively (Figure 6). Feed and spring malt resulted in lower dollar value estimates than food and winter malt due to less nutrient uptake. In comparison, previous estimates from Tarkalson et al. (2009) would have resulted in estimates of $51.25, $102.50, $153.75, and $205.00 ha−1 for yield of 3, 6, 9, and 12 Mg ha−1, respectively.

    Details are in the caption following the image
    Barley residue biomass nutrient value (N, P, K, and S; $ ha−1) for four grain yield levels, predicted using multiple approaches. Predictions were made using barley class-specific HI and nutrient concentration (feed, food, spring malt (S), and winter malt (W), a mean estimate averaging across classes), and a single-factor prediction (SF-RTO) across classes based on regression of grain yield to nutrient uptake from data collected during two-growing seasons (2018 and 2019) in southern Idaho. Horizontal lines for N, P, and K are based on previous nutrient uptake estimates for the region based on HI, nutrient concentration, and nutrient cost data reported by Tarkalson et al. (2009).

    Differences in residue biomass and nutrient uptake in the current study compared to Tarkalson et al. (2009) are likely due to variation in HI to estimate residue biomass from grain yield and greater nutrient concentrations in some instances. These differences likely occur due to numerous reasons. Spring malt and feed barley are the primary classes grown in the region as well as in more northern and higher elevation areas with two-row high-fiber food and viable winter malts only available relatively recently (Hu et al., 2016; Obert et al., 2006, 2011). Food and winter malt barley are grown on lower acreage due to the recent introduction and known limitations of each, including winter kill in colder regions for winter malt and lower yields for high-fiber food barley (Marshall et al., 2019). Limiting comparison of our data to spring feed and malt results in closer estimates to Tarkalson et al. (2009) for N and K.

    While we have shown mean data at times, prediction using class-specific HI and nutrient concentrations would be reasonable when production class is known. Another major factor is that southern Idaho is an optimal environment for barley production with generally low humidity, sufficient irrigation, and favorable temperatures. Additionally, many areas are unsuitable for other crops due to growing season length and thus, barley competes less for a place in the rotation. As a result of this, Idaho produces the largest amount of barley (33%) in the United States with irrigated state yields that are some of the highest in the country approaching 6.0 Mg ha−1 (USDA-NASS, 2023b). Other crops grown in rotations in the region, particularly potato (Solanum tuberosum L.) and sugar beet (Beta vulgaris L.), retain high nutrient concentrations in the aboveground biomass that is recycled back into the soil in subsequent growing seasons (Leytem et al., 2023). Soil nutrient content can also be elevated due to the low-rainfall semiarid environment which limits runoff, soil formation characteristics associated with the Snake River Plain, and residuals from previous nutrient applications (Rogers, Dari, & Schroeder, 2019). The current study fills an important gap in estimating nutrient uptake and economic value of barley in high-input systems where water and nutrients are typically available at rates that largely minimize stress to the plant.

    In addition to class comparisons, low (minimum), average, and high (maximum) prices (in $ kg−1) for each nutrient, in terms of fertilizer cost, were evaluated over the period of 2001–2019. For this analysis, we used the HI-specific estimate averages (mean) of nutrient uptake from Figure 5 to explore fertilizer price ranges and their effect on economic value. As above, low and high fertilizer price estimates (Table 3) are a constant for each nutrient, and thus, similar trends as were seen with average nutrient uptake are found if extrapolating based on fertilizer price ranges. The average prices per kg of N and K were higher than those of S and P (Table 2). Nitrogen and K also had the widest ranges between the highest and lowest prices ($0.79 and $0.99 kg−1 for N and K, respectively). The calculations of the residue biomass nutrient economic value ranges, which combine the residue biomass nutrient uptake and nutrient prices for the combined data, are included in Table 11. The highest residue biomass nutrient economic values were those for K and N, for which the nutrient contents and fertilizer nutrient prices were also the highest. Likewise, the nutrient values were lowest for P and S, which reflect both their having the lowest quantities in residue biomass and relatively low fertilizer prices. The total value of nutrients (N, P, K, and S) in residue biomass, based on historical average prices, ranged from $68.34 ha−1 for 3 Mg ha−1 grain yields to $273.36 ha−1 for 12 Mg ha−1 grain yields. Additionally, nearly $500 ha−1 in total nutrients is estimated under maximum yield and price scenarios. Thus, the naturally high K content in the soils in this region has substantial value, particularly if fertilizer costs continue to increase in the future. With continued production and associated extraction of K, testing of changes in soil K levels will have greater importance as replacement needs increase over time, especially where total aboveground biomass removal is a common practice, such as with small grains with residue biomass harvest, and in systems that include alfalfa (Medicago sativa L.), and silage corn (Zea mays L.).

    TABLE 11. Estimates of fertilizer replacement cost for four key nutrients (N, P, K, and S) accumulated in barley residue biomass based on the average of class-specific harvest index (HI) and nutrient concentrations averaged across classes, locations, cultivars, and years (2018 and 2019) derived from studies conducted in southern Idaho.
    Residue nutrient valuesa
    Grain yield (Mg ha−1) N ($ ha−1) P ($ ha−1) K ($ ha−1) S ($ ha−1) Total ($ ha−1)
    Class-specific HI mean 3 20.08 (13.25, 29.31) 0.65 (0.32, 0.99) 45.33 (19.43, 90.66) 2.28 (1.48, 3.14) 68.34 (34.49, 124.10)
    6 40.16 (26.50, 58.63) 1.30 (0.65, 1.98) 90.66 (38.85, 181.31) 4.57 (2.97, 6.28) 136.68 (68.97, 248.20)
    9 60.23 (39.75, 87.94) 1.95 (0.97, 2.98) 135.98 (58.28, 271.97) 6.85 (4.45, 9.42) 205.02 (103.46, 372.31)
    12 80.31 (53.00, 117.25) 2.60 (1.30, 3.97) 181.31 (77.71, 362.62) 9.14 (5.94, 12.56) 273.36 (137.95, 496.41)
    • Note: Bold residue biomass nutrient values were calculated with average prices and the ranges below them in parentheses are based on low and high prices.
    • a Residue nutrient quantities and nutrient prices were based on values in Table 3.
    • [Correction added on 8 June 2024, after first online publication: In table 11 – the 98.97 value was mis entered into the table and is changed to 68.97.]

    A main takeaway message from this economic analysis is that the value of the nutrients in barley residue biomass in any given growing season is dependent on both the yields (grain and residue biomass) and prices of the nutrients. Additionally, the expected nutrient prices for the next growing season are particularly important to consider when assessing whether to remove the residue biomass or leave it in the field. In growing years with higher yields and nutrient prices, the price that a farmer receives for the removed residue biomass should be adjusted upward to reflect its higher value or retention within field considered based on overall farm economics. Other factors to consider when making the choice of whether to gather and sell barley residue biomass or leave it in the field include those that influence the cost of mowing, raking, and baling the hay, including fuel for utilized equipment and labor, as well as transportation costs. If an outside party were to implement the removal and sale, as is relatively common in the study region, then a producer would at least want to obtain a price for the baled residue biomass that covers a portion of the expected costs of replacing the nutrients in the removed residue biomass via fertilizer purchases in preparation for the subsequent growing season. Thus, the multi-step breakdown allows specific measured residue biomass and/or grain nutrient concentrations and economic prices to be used to refine the estimate.

    4 CONCLUSIONS

    Barley is widely produced in the semiarid western United States. Production nearly always includes harvest of grain and the associated nutrients, but the residue biomass is also commonly removed from the field for various end-uses that contribute to on-farm revenue. The data from the current study provide a broader and more robust range for potential nutrient uptake and removal through residue biomass harvest than was previously available for barley. Specifically, the range of HI and nutrient concentrations specific to feed, food, spring malt, and winter malt could be used to refine estimates based on these specific production classes. The multi-step breakdown allows specific measured residue biomass nutrient concentrations and economic prices to be used to refine the estimate when available. In addition to the current monetary value of nutrients contained in barley residue biomass, producers should consider the long-term economic sustainability of their operations when evaluating the value of the residue biomass they are selling. From this perspective, barley residue biomass nutrient removal should be considered in relation to the potential value of the nutrients in the residue biomass being removed and the implications for the fertilizer replacement cost now and in the future. Economic analysis indicated that barley residue biomass nutrient values are highly dependent on the levels of grain yields (and corresponding residue biomass) and fertilizer prices, especially those for N and K. However, other factors need to be considered by producers when making the choice of whether to harvest barley residue biomass or leave it in the field. These factors include soil nutrient concentrations available for crop use, alternate non-fertilizer sources of replacement nutrients (e.g., irrigation water, atmospheric depostion, etc.), and residue biomass management for upcoming crops. The non-nutrient values of residue biomass, such as erosion control and enrichment of SOM, should likewise be considered in the context of the agronomic characteristics of a given field. Finally, the decision of what level to utilize these metrics (on-farm, regional, etc.) to calculate nutrient uptake should be considered based on known variation in relation to the proposed task.

    AUTHOR CONTRIBUTIONS

    Christopher W. Rogers: Conceptualization; data curation; formal analysis; funding acquisition; investigation; methodology; project administration; resources; software; supervision; validation; visualization; writing—original draft; writing—review and editing. Curtis B. Adams: Data curation; formal analysis; investigation; methodology; resources; software; validation; visualization; writing—review and editing. Juliet M. Marshall: Conceptualization; funding acquisition; investigation; methodology; project administration; resources; software; supervision; validation; visualization; writing—review and editing. Patrick Hatzenbuehler: Data curation; formal analysis; investigation; methodology; software; validation; visualization; writing—original draft; writing—review and editing. Garrett Thurgood: Data curation; formal analysis; investigation; methodology; validation; visualization; writing—review and editing. Biswanath Dari: Data curation; investigation; methodology; writing—review and editing. Grant Loomis: Data curation; investigation; methodology; writing—review and editing. David D. Tarkalson: Formal analysis; validation; writing—review and editing.

    ACKNOWLEDGMENTS

    Support for this research was provided by the Idaho Barley and Wheat Commissions. The authors thank Scott Pristupa, Dr. Lauren Vitko, Erin Thurgood, Irene Shackleford, and Rebecca Caldera for laboratory assistance and Chad Jackson and Tod Shelman for assistance in the field.

      CONFLICT OF INTEREST STATEMENT

      The authors declare no conflicts of interest.