Sheep Winter Diets Parameterized with Fecal DNA Metabarcoding and Forage Sampling Informs Mineral Nutrition Management

ABSTRACT Mineral nutrition deficiencies commonly occur in extensive sheep production systems, yet often coincide with critical production periods of breeding and gestation when quality and quantity of winter range are most limited. Ewe productivity may be limited by clinical and subclinical mineral deficiencies when grazing extensive rangelands during the winter months. The objectives of this study were to utilize fecal meta-barcoding DNA (f.DNA) to estimate dietary composition of ewes by plant morphological group (monocots and dicots), and quantified plant nutritional composition from clipped biomass to estimate adequacy of macro- and micro-minerals of the basal diet across extensively grazed sheep operations. Ewe fecal samples for f.DNA were collected from 19 extensive sheep operations across Wyoming and Colorado, USA, and analyzed at the plant morphological group level to estimate dietary composition as 1) Monocot (≥ 70%), 2) Dicot (≥ 70%), or 3) Mixed (monocot and dicot co-dominated, both proportions < 70%). Pooled forage species nutritional composition within morphological group (monocot vs. dicot) was then multiplied by the proportion of monocotyledonous species or dicotyledonous species and intake estimates of 2% of BW on an 80 kg ewe to estimate macro- and micro-mineral intake. Results from f.DNA indicated 36% of operations were categorized as Monocot (≥ 70% dietary component), 42% Dicot (≥ 70% dietary component), or 21% Mixed. A significant effect of plant morphological group category was observed for CP, Ca, K, Mg, and S (P < 0.05) where dicot dominated rangelands contained higher concentrations of these minerals compared to monocot or mixed monocot-dicot rangelands. Overall, dicot dominated rangelands provided greater macro- and micro-mineral content ultimately meeting more requirements for breeding and gestation than monocot or mixed monocot-dicot rangelands. Therefore, targeted supplementation must be considered for ewes on monocot or mixed monocot-dicot rangelands during winter months that coincide with critical production stages.


Introduction
Global rangelands are inherently diverse ecosystems comprised of variable abundances of native monocotyledonous plants (i.e., grasses and sedges) and dicotyledonous plants (i.e., forbs, shrubs, and trees) all of which may be utilized by livestock and wildlife.Critical to ruminant agriculture, extensive sheep production systems of the western USA (Wyoming, Utah, Nevada, Montana, Colorado, California, Washington, Oregon) make up approximetly 42% of all breeding sheep ( USDA-NASS 2023 ), and often rely on rangelands to provide their macro-and micro-nutrients for many months each year including winter months ( Julian et al. 2020 ).In Wyoming, USA, ranked third in breeding sheep numbers ( USDA-NASS 2023 ), plant communities exemplify this vegetation morphological diversity including alpine meadows, sagebrush steppe, shortgrass prairie and desert shrublands all of which are important for sheep grazing ( Knight et al. 2014 ).Such diversity can be a function of the relative dominance of monocotyledonous plants (including grasses and grasslike plants) or dicotyledonous plants (including forbs and shrubs particularly as dominant life forms on rangelands) which can an influence on nutrient availability to grazing animals ( Emanuele and Staples 1990 ).Such growth forms have been important differential predictors of mineral element concentrations ( Grings et al. 1996 ;Han et al. 2011 ) and forage quality parameters ( Julian et al. 2020 ;Stewart et al. 2021 ) with variable responses to temporal and plant-part trends ( Bumb et al. 2016 ).Additionally, rangelands are known to have plant productivity and climatic constraints that hinder accessibility, quality, and quantity of the natural grown forages ( Holechek et al. 1995 ;Fuhlendorf et al. 2017 ;Morris 2017 ).
Forage mineral elements naturally vary in concentration due to soil geochemistry, plant species, and plant seasonality ( Corona et al. 1998 ;Ganskopp and Bohnert 2003 ; USGS 2021 ) which creates a persistent obstacle for range sheep operations when implementing supplementation strategies.Surveys have found that 33% of producers in Montana, USA ( Page et al. 2018 ) and 42% of producers in Wyoming, USA ( Julian et al. 2020 ) were not providing adequate mineral supplementation to sheep.Critical production periods for ewes (considered when nutritional requirements relative to reproduction stages are elevated and include breeding, early gestation, late gestation, and early lactation) often coincide with when ewes are grazing rangelands during winter months and forages have lower accessibility, quality, and quantity.
Methods utilized for estimating intake have evolved through time ( Garnick et al. 2018 ).An emerging technique using deoxyribonucleic acid (DNA) barcoding of fecal samples to determine plant composition (henceforth f.DNA; Scasta et al. 2020 ;Craine 2021 ) has enhanced insights about post-ingestive diet botanical constituents.When f.DNA is combined with traditional techniques such as wet chemistry forage analysis based on ground-truthed sampling of the vegetation where grazing occurs ( Scasta et al. 2019b ;Stewart et al. 2021 ), enhanced quantification of nutritional shortfalls of ewes across many operations and conditions may be possible.Such integrated approaches could then provide precision trace-mineral management across the heterogeneous range-land vegetation features that are refined on a ranch-specific basis particularly with the integration of field-derived estimates with emerging technologies such as f.DNA which provide enhanced resolution on diet diversity in the context of ecosystem diversity ( Scasta et al. 2019b ;Scasta et al. 2020 ).Thus, our objectives were to utilize f.DNA barcoding to parameterize macro-and micromineral intake models from forage analysis at the plant morphological group level (e.g., monocots or dicots) to estimate mineral element shortfalls on extensively managed sheep operations during the winter months.We hypothesize that there will be differences in mineral content relative to dominant plant morphological groups, that range from monocot to dicot dominated, and that these insights can help to tailor recommendations for winter supplementation.
2 Threadleaf sedge is a grass-like species.
and representative of the individual producer's winter range grazing allotment.A minimum of 8 plants per species, of the dominant species, were sampled within each plot.Individual plant species across 3 plots were then composited and stored at −20 °C.Dry matter of ground grass and shrub species was calculated by drying ground material at 64 °C in a forced air oven for 24 h.Material was weighed and then dried again at 105 °C for 3 h.Nitrogen was analyzed (Method 990.03;AOAC 2006 ;Leco Corp., St. Joseph, MI) and crude protein (CP) was derived relative to N concentration.Forage mineral concentrations (P, Mg, K, Na, Fe, Zn, Cu, Mn, Mo, S) were analyzed using inductively coupled atomic plasma analysis ( Kovar, 2003 ).Selenium and Co were analyzed using AOAC 986.15 and 985.01, respectively.

f.DNA data
Out of the 28 operations that participated in the initial forage collection, 19 operations participated in the f.DNA component by providing one composite fecal sample from a grazing cohort of gestating ewes at each winter range grazing location ( Fig. 1 ).In brief, fresh fecal sub-samples were collected from the ground after following a minimum of 3 individual ewes at one sampling time point, and samples were composited into a single ranch-level sample ( n = 19).Fecal samples were stored at −18 °C and then sent for DNA meta-barcoding analysis at Jonah Ventures Laboratory (Boulder, CO, USA).The f.DNA process to extract plant DNA from fecal samples uses trnL c and h primers for PCR amplification.Clustered gene sequences are referenced to an Operational Taxonomic Unit (OTU) library.The BLAST nucleotide program was used to identify unknown sequences based on a minimum criterion of 97% base pair matching ( Craine et al. 2015 ;Craine et al. 2016 ).All results lacking OTU BLAST matches for domain, phylum, order, family, or genus were removed from the dataset.From there, all remaining OTUs were compared to georeferenced and verified occurrences of plant species in the USDA Plant database ( USDA-NRCS 2018 ) and/or Rocky Mountain Herbarium ( RMH 2008 ) for taxonomic and geospatial reconciliation ( Scasta et al. 2019a ).In some cases, if a genera made logical sense but a specific epithet (e.g., species) did not, then the plant was retained in the dataset by genus only.If no logical geospatial association could be determined at either genus or species, the read was removed from the dataset.Percentages for each plant species were then pooled and summed for two plant morphological groups, monotcotyledons (including grasses and sedges) and dicotyledons (including forbs, shrubs, shrubs, and trees) for each operation.

f.DNA intake modeling
Proportional composition of plant morphological groups (monocots vs. dicots) obtained from f.DNA at each ranch were used to estimate diet composition at the plant morphological group level.To accomplish this, the macro-and micro-mineral concentrations quantified from biomass clippings at each operation ( n = 19) were categorized by plant morphological group and then multiplied by the operation's associated morphological group proportion (% monocot or % dicot).The estimated amounts of each mineral for monocotyledonous and dicotyledonous plant species were then summed to provide a% of dietary Dry Matter (DM) for macrominerals Ca, P, K, Mg, S, Na CP, or mg/kg for micro-minerals Zn, Mn, Cu, Fe, Mo, Se, Co.To estimate the adequacy of f.DNA proportional dietary composition in meeting daily requirements (g/day or mg/d), intake estimates of 2% of BW of a hypothetical 80 kg ewe (1.64 kg daily DMI) were utilized from the Nutrient Requirements of Small Ruminants ( NASEM 2007 ).The 80 kg ewe BW es-timate was also confirmed from producer questionnaire data from Julian et al. (2020) .Because management of ewes on winter range in the western USA coincides with critical production periods of breeding through late gestation, percent of daily requirements met was also simulated to account for production stage (breeding, early gestation, late gestation) and litter type (singe or twin) according to NASEM (2007) .Operations were categorized by the proportion of monocotyledonous species relative to dicotyledonous species in the diet as follows: Monocot operations ( n = 7) where the proportion of monocotyledonous species (i.e., graminoids including grasses and sedges) was ≥ 70%, Dicot operations ( n = 8) where the proportion of dicotyledonous species (i.e., forbs, shrubs, and trees) was ≥ 70%, and Mixed monocot-dicot operations ( n = 4) where both dicotyledonous species and monocotyledonous species proportions were < 70%, respectively.

Statistical analyses
The Means procedure in SAS (v.9.4; SAS Inst.Inc., Carry, NC) was used to calculate means ± standard deviation (SD) grouped by consumption type described above and then by each individual operation for CP, P, Mg, K, Na, Fe, Zn, Cu, Mn, Mo, S, and Co.The Frequency procedure was used to determine the proportion of monocots vs. dicots per operation, based on the proportion of reads from the f.DNA data.To determine the effect of plant morphological type on nutrient and mineral element concentrations, all macro-and micro-mineral element concentrations found in forages were transformed to g and mg, respectively.Forage species within each plant morphological type was the experimental unit, and were analyzed in the GLIMMIX procedure with the fixed effect of ranch plant morphological type.

Results
Crude protein, macro-and micro-mineral concentrations of monocotyledonous species and dicotyledonous species collected from sheep winter grazing locations are summarized in Tables 1 and 2 .A comparison of each mineral concentration by the 3 plant morphological groups are presented in Table 3 .Operations 1 Monocot = proportion of grass in sheep diets was ≥ 70%.
2 Mixed Dicot-Monocot = proportion of grass and shrub was co-dominated and neither proportion was > 70%.
3 Dicot = proportion of shrubs and forbs in sheep diets was ≥ 70%. 4 Calculated concentration in diets estimated based on NASEM (2007) for a 80.0 kg ewe consuming 1.6 kg DM during production stages of breeding, early gestation (gestating twins), and late gestation (gestating twins).categorized as, dicot and mixed monocot-dicot dominated operations had significantly greater estimated dietary concentrations of CP, Ca, K, Cu when compared to monocot dominated operations ( P < 0.05; Table 3 ).Estimated dietary concentrations of P, Mg, S in dicot dominated operations were greater than those categorized as monocot dominated, whereas Na, Zn, Fe, Mo, Se, and Co were similar across dietary categorizations ( P > 0.05; Table 3 ).
Estimated sheep dietary composition (e.g., proportion monocots to dicots) per operation ( n = 19) and categorized from f.DNA analysis are presented in Fig. 2 .Of the 19 sheep operations with f.DNA data seven operations (1-7) were categorized as monocot dominated operations, four operations (8-11) were categorized as mixed monocot-dicot co-dominated operations, eight operations (12-19) were categorized as dicot dominated operations.Across operations, the mean monocot percent and SD was 44 ± 31.5% and the mean dicot percent and SD was 56 ± 31.5%.Over half (53%; 10 of 19) of the sampled operations were found to have higher consumption of dicot species compared to monocot species.Two operations (13%; 2 of 19) contained 100% dicot diet composition as estimated from f.DNA.
Estimated nutrient intake based on plant morphological type and percent of operations meeting requirements for breeding, early gestation (single and twin), and late gestation (single and twin) are summarized in Tables 4 and 6 .Monocot operations ( n = 7; Table 4 ) met Fe, Mo, and Se requirements (100%; 7 of 7), but did not meet CP, P, K, S, and Na requirements (0%; 0 of 7) for breeding, early gestation and late gestation, whereas percent of operations meeting requirements generally decreased for Ca, Mg, Zn, Mn, Cu, and Co with advancing stage of production.Mixed monocot-dicot operations ( n = 4; Table 5 ) met Mn, Fe, Mo, and Se requirements (100%; 4 of 4) but did not meet S requirements (0%; 0 of 4) for breeding, early gestation, and late gestation.Percentage of Mixed monocotdicot operations meeting requirements for Ca, K, Mg, Na was generally greater than CP, P, Zn, and Cu where a greater proportion of these ranches were deficient in as stage of production advanced beyond breeding.
Dicot categorized operations ( n = 8; Table 6 ) met K, Mn, Fe, and Mo requirements (100%; 8 of 8) for breeding, early gestation, and late gestation.Operations meeting requirements was variable for CP, Ca, P, Mg, S, Na, Zn, Cu, Se, and Co, still nutrient shortfalls across operations was greatest for CP, P, Zn and Cu as stage of production advanced.

Discussion
Estimations of dietary composition and nutrient intake of grazing ruminants has been a persistent challenge for animal scientists, range scientists, and managers ( Garnick et al. 2018 ).An array of methodologies exists to lend important insights but each has unique advantages and limitations ( Mayes and Dove 20 0 0 ; Garnick et al. 2018 ) that can inhibit application for specific study objectives desired.Fecal meta-barcoding DNA, notwithstanding it's limitations ( Scasta et al. 2019a ), provides unique and reliable insights that can be adjusted to appropriate taxonomic, functional, or morphological plant group levels for diverse scientific investigations ( Craine 2021 ).Additionally, contextualizing operational production practices, such as potential deficiencies identified in the participating sheep producers in this study ( Julian et al. 2020 ) provides data-driven recommendations to improve supplementation programs across heterogeneous rangelands.More specifically, data generated at each sheep operation (forage nutritional composition and f.DNA) provided operational-level insights in near real-time for sheep producers to make management adaptations as necessary while establishing generalizable recommendations.

Table 4
Dietary nutrient composition, calculated intake (g/d or mg/d) and% requirements met across production stages from f.DNA "Monocot" ( ≥ 70% monocotyledonous plant species in diet) categorized operations ( n = 7).Estimated intake f.DNA "Monocot" category 4 Daily requirements list the range of requirements from breeding, early, and late gestation (twins).

Plant morphological group nutritional composition differences
Sheep operations that have a high dicot component of the available rangeland vegetation community during winter months are more likely to meet mineral requirements for breeding, early gestation, and late gestation ( Tables 3 and 6 ).Corona et al. (1998) found mineral element concentrations to decrease as plants matured due to increased lignin, cellulose, and hemicellulose.However, differences in mineral element concentrations between monocot and dicot species suggest that dicot species are not directly correlated with plant maturity and may not follow the same deterioration of forage quality ( Emanuele and Staples 1990 ). Furthermore, previous studies describe similar patterns when comparing forage quality of monocots to dicots and suggest variance is significant function of plant growth form ( Bumb et al. 2016 ).For example, Han et al. (2011) and Grings et al. (1996) observed forbs were richest in minerals compared to grass foliage, while Julian et al. ( 2020) and Stewart et al. (2021) also describe similar findings when quantifying dormant grass and shrub quality.Therefore, precision tracemineral supplementation considerations should be made for sheep operations that do not have a high dicot plant component available.

Morphological group dietary composition
Results from f.DNA indicated 36% of operations were categorized as Monocot ( ≥ 70% dietary component), 42% Dicot ( ≥ 70% dietary component), or 21% Mixed (Monocot and Dicot codominated).Therefore, when dicotyledonous plants are available on winter range they may be consumed more readily given the optimal nutritional resources that they provide.While studies suggest sheep preference for monocots or dicots may be driven by their respective and relative availability ( Reppert 1960 ;Bartolome et al. 1998 ), Royer et al. (2005) found sheep preference for shrubs specif- ically was lower than the proportion available during the summer months, but found sheep to spend more time browsing shrubs in the winter months.Additionally, Cook and Harris (1950) describe sheep diets on Utah winter range contained 70% browse and 30% grass.

Nutrient intake estimations and management implications
Nutrient intake estimations in the current study do not account for the nuance in forage biomass restrictions, adverse weather constraints, or other grazing intake limitations common in extensive sheep production systems yet still provide an important baseline for what standing forage provides.Sheep operations surveyed in a companion study reported managing ewes on winter rangelands for a majority of the production cycle (73% of operations > 131 d on winter range; Julian et al. 2020 ) highlighting the reliance on the grazing diet to provide the majority of macro-and micro-nutrients, especially when supplements are not provided.
Seasonal changes in forage quality and quantity necessitate additional protein supplementation on range sheep operations and strategies to meet these demands have been widely investigated and reviewed ( Thomas and Kott 1995 ;Schauer et al. 2010 ;Bohnert and Stephenson 2016 ).Not surprisingly, CP intake estimates in the current study are limiting across locations irrespective of dietary morphological group categorization, especially throughout the periods of gestation and for ewes gestating multiple lambs.Crude protein requirements increase 27% from breeding to early gestation, and 25% from early gestation to late gestation.Similarly, CP requirements between a twin-carrying ewe is 25 g/d more than a single-carrying ewe.Assumptions were based on twin-bearing 80 kg ewes and thus single-bearing ewes would not experience as dramatic a restriction in CP intake.Still, historical increases in ewe prolificacy indicate that greater proportion of ewe management cohort would be carrying multiple lambs (30-70%; Stewart et al. 2020 ).
Macro-and micro-mineral supplementation strategies are challenging on extensive sheep operations due to variability in mineral supplementation intake (e.g., over-and under-consumption of minerals; Stewart et al. 2021 ).An impetus to the current study was driven by concerns for macro-and micro-mineral shortfalls in the basal diet, particularly, in extensively managed flocks across Wyoming and surrounding states.Studies quantifying the number of sheep operations providing a complete trace mineral supplement to sheep on range varied from 47% of Wyoming producers ( n = 19; Julian et al. 2020 ) to 67% of Montana producers ( n = 21; Page et al. 2018 ).In general, Monocot operations failed to meet requirements for CP, P, K, Mg, S, Na, and Cu from breeding to late gestation ( Table 4 ), Mixed monocot-dicot operations failed to meet requirements for CP, P, S, and Cu ( Table 5 ), and Dicot operations failed to meet requirements for CP, P, and Cu ( Table 6 ).Thematically study results indicate, increasing proportions of dicots resulted in greater adequacy of macro-and micro-nutrients than monocot predominated winter grazing sites.The estimated available macro-and micro-nutrients in forage resources across sheep operations in the current study indicate the ability of sheep to utilize and thrive on grazing lands with a greater degree of plant heterogeneity.
The competitiveness of extensive sheep production systems in the western USA and internationally is the ability of most sheep breeds to thrive under low-input management conditions utilizing available natural resources.Their ability to utilize a diverse array of plant morphological types in the current study shows their dietary plasticity as grazing resources vary through time in quality and quantity ( Scasta et al. 2016 ).The beneficial implications of this greater selectivity based on certain conditions in the current study would be a decreased reliance on supplemental sources of CP, Ca, P, K, Mg, S, and Cu.In contrast, results in the current study suggest monocot dominated winter grazing sites do not possess adequate basal concentrations of macro-and micro-minerals for breeding and gestation and sheep managed on similar sites should be managed accordingly.Not withstanding the unique dietary flexibility of sheep findings suggest that a greater proportion of ranches will require supplemental interventions beyond those of the basal diet quantified in the current study.
Sheep operations surveyed in a companion study reported a variety of supplementation management strategies representative of the ewes grazing winter range in the current study, e.g., 84% provide supplemental protein or energy, 16% do not provide any supplement ( Julian et al. 2020 ).Accounting for the macro-and micronutrients of additional protein or energy supplements combined with estimations of the f.DNA informed diet composition offer important insights to precision supplementation strategies.
Phosphorous is a commonly deficient macro-mineral for livestock grazing rangelands ( Corona et al. 1998 ;Ganskopp and Bohnert 2003 ;Sprinkle et al. 2018 ) and is evident by our data irrespective of dietary plant morphological group categorization ( Tables 4  and 6 ).Mechanistically, low dietary P is a result of low P concentrations in soils and plants across the USA ( Underwood 1981 ;Smith et al. 2014 ) but was most pronounced in monocot categorized operations.
None of the Monocot operations (0%, Table 4 ) met K requirements for breeding, early gestation, and late gestation.This sug-gests the diverse forage resources across participating ranches in this study area contribute to different dietary K concentrations between monocot and dicot categorized operations ( Tables 3 ).Previous research also shows shrubs containing higher K concentrations compared to grasses ( Sprinkle et al. 2018 ;Belesky et al. 2020 ;Julian et al. 2020 ).Potassium is not readily stored in tissues and therefore daily supplementation is necessary ( NASEM 2007 ).Production losses such as decreased growth and feed intake are associated with K deficiencies ( NASEM 2007 ).
No Monocot or Dicot operations met S requirements for breeding, early, and late gestation ( Tables 4 and 5 ).Sulfur is an antagonistic mineral and interferes with Cu and Se absorption and utilization, resulting in a maximum tolerable level of 0.50% for high forage diets ( NASEM 2007 ).The current study shows the highest concentration of S found in Gardner's saltbush at the maximum level (0.50%, Table 1 ); however, based on the dietary plant morphological group Monocot (0.09%, Table 4 ), Mixed (0.17%, Table 5 ), and Dicot (0.23%, Table 6 ) operations do not reach this maximum tolerable level and further suggests minimal antagonistic interference with Cu and Se absorption.Sulfur concentration in the diet is further complicated by water sources which are known to contain high levels of S ( Suttle 2010 ).Page et al. (2018) found high concentrations of sulfates in water sources could lead to possible deficiencies of Se, Cu, Mo, and Zn due to its antagonistic role with these minerals.While Stewart et al. (2021) suggests S antagonism of Se is unlikely to show clinical Se deficiency when Se and status of the animal and diet are adequate, which is observed in the current study ( Tables 4 and 6 ), with the exception of 88% of Dicot operations meeting early and late gestation Se requirements.
No Monocot operations (0%, Table 4 ) met Na breeding, early gestation, and late gestation requirements, while 50% and 75% of Mixed ( Table 5 ) and Dicot ( Table 6 ) operations, respectively, met breeding, early, and late gestation requirements.This highlights the higher concentration of Na found in shrubs to meet requirements (12.81 g vs. 0.75 g; Table 3 ).Specifically, Atriplex species that were sampled (Gardner's saltbush, shadscale saltbush; Table 1 ) are known to be Na accumulators ( Sagers et al. 2017 ) further explaining why Monocot operations did not meet Na requirements and why only 37% of producers in Wyoming supplemented white salt during winter months ( Julian et al. 2020 ).
Copper is an important, and yet often excluded, supplemental micro-nutrient in sheep production systems due to issues related to Cu toxicity (15 mg/kg DM; NASEM 2007 ).Still, 75% of Mixed operations and 100% of Dicot operations did meet Cu breeding requirements, 0% of Monocot operations met breeding, and 0% of all operations regardless of plant morphological group dominance met early and late gestation requirements.Copper deficiency in sheep may be attributable to high dietary Mo, low dietary Cu:Mo ratios ( < 2:1), low Cu forage concentration, dietary S-Fe-Mo antagonism, and variability in Cu forage concentrations ( Suttle 2010 ;Stewart et al. 2021 ).Clinical Cu deficiencies are observed when ewes do not consume more than 5 to 6 mg of Cu/kg of DM during pregnancy which may adversely affect the conceptus ( NASEM 2007 ;Stewart et al. 2021 ).While all operations regardless of plant morphological group dominance did provide slightly more than the 5 to 6 mg (6.25 mg, 7.12 mg, 7.43 mg, respectively), the other factors previously mentioned may lead to Cu shortfalls from the basal grazing diet.
Other important macro-and micro-minerals that operations met more consistently but are worthy of careful monitoring in extensive sheep production systems include Ca and Zn.Ewes on Dicot operations received twice as much Ca compared to ewes on Monocot operations (14.75 g vs. 7.40 g; Table 3 ) which suggests Ca concentrations are heavily influenced by the diverse plant communities observed and ewes grazing on Dicot operations are not likely to experience a shortfall in Ca requirements due to forages sam-pled providing adequate Ca.With a rapid increase in Ca requirements (74% from breeding to late gestation; NASEM 2007 ) and low availability of Ca in forages on Mixed and Monocot operations, Ca deficiencies are likely to arise.Metabolic dysfuntion related to Ca deficiency or hypocalcemia "milk fever" occurs during the periparturient period and can result in significant ewe and lamb mortality, especially in multiple bearing ewes ( NASEM 2007 ;Suttle 2010 ).Across all operations ( n = 19) 38% or less met Zn requirements for breeding, early gestation, and late gestation ( Tables 4 and 6 ).Since Zn is not adequately stored throughout the body it is important to continually supplement ewes in order to maintain adequate physiological concentrations ( NASEM 2007 ).While clinical deficiencies of Zn are easily noticed such as decreased wool growth, appetite, and overall growth ( Suttle1988 ; Page et al. 2020 ), subclinical deficiencies such as decreased fertility ( Hostetler et al. 2003 ) are likely unnoticed and may contribute to unneccesary production losses.

Implications
This study evaluated forage nutrient and mineral element concentrations for varying winter rangeland types, as characterized by their dominant plant morphological groups and supported by advanced fecal DNA metabarcoding technologies, across 19 extensive operations in Wyoming and Colorado USA.Nutritional modeling methods used in the current study combined proximate analysis of forages combined with f.DNA to elucidate proportion of monocots and dicots at a regional scale.The results revealed that increasing forage species heterogeneity, particularly the relative abundance of dicotyledons, on sheep winter range can improve the adequacy of macro-and micro-nutrients in sheep winter diets particularly given the dietary selectivity of sheep which can readily utilize forbs and shrubs ( Scasta et al. 2016 ).Accordingly, sheep may be more adept at optimizing the nutritional resources provided by dicot dominanted rangelands.Moreover on western rangelands avoiding the elimination of native dicotyledonous species, or restoring them, has been emphasized for biodiversity ( Black et al. 2011 ;Tilley et al. 2022 ) but could also have positive benefits for small ruminant enterprises.Essential to such enterprises, supplementation strategies can be guided by understanding plant heterogeneity or homogeneity at winter grazing sites.Managers and providers of sheep supplements need to prioritize supplements that contain adequate CP, P, K, S, Na, and Cu especially as physiological demands increase from breeding through gestation.Supplemental interventions, although required to a lesser extent on dicot dominated rangelands should precede the physiologically demanding stages of breeding and gestation in extensive sheep productions systems.

Declaration of competing interest
Authors A .A .M. Julian, J. D. Scasta, W.C. Stewart declare no conflicts of interest associated with manuscript "Sheep winter diets parameterized with fecal DNA metabarcoding and forage sampling informs mineral nutrition management."

Fig. 2 .
Fig. 2. Estimated f.DNA dietary proportion of monocotyledons and dicotyledons of ewes grazing dormant forages across 19 operations collected from December to January 2018-2020.

Table 1
CP, macro-and micro-mineral concentrations found in common dicotyledonous shrub and forb plant species collected on winter range ( n = 28 operations). 1

Table 3
Least-squares means ( ± SE) for the main effect of plant morphological dietary categorization based on ewe daily CP, macro, and micro mineral intake (g/d or mg/d) across 19 winter range grazing locations.
a-b Values within a row with different superscripts differ significantly at P < 0.05, according to a LSD.
Calculated concentration in diets estimated based on NASEM (2007) for a 80.0 kg ewe consuming 1.6 kg DM during production stages of breeding, early gestation (gestating twins), and late gestation (gestating twins).
1 Average forage macro (%) and micro (mg/kg) minerals found in Monocot operations. 2 CP, Ca, P, K, Mg, S, and Na are measured in g/d DM.Calculated concentration in diets based on NASEM (2007) for a 80.0 kg ewe consuming 1.6 kg DM for breeding, early gestation (gestating twins), and late gestation (gestating twins).3Zn, Mn, Cu, Fe, Mo, Se, and Co are measured in mg/d DM.