Grazing and fertilizer, compost or manure application effects on a meadow bromegrass pasture on a thick black chernozem I. Productivity and sustainability

Abstract Short duration, intensive grazing management with high stocking rates may result in sufficient turn-over of nitrogen (N) to compensate for production-limiting soil-N deficiencies for grass pasture. In central Alberta a 0.5 ha block was seeded to “Fleet” meadow bromegrass (Bromus riparius Rehmann) in August 2002. Within this block, six fenced (9 m × 30 m) treatments were established in three replicates. They were (1) ungrazed—clip removal, (2) grazed—alone, (3) grazed—fertilizer, (4) grazed—fertilizer-compost, (5) grazed—hog manure, and (6) grazed—alfalfa (Medicago sativa L.) grass. Measurements were conducted over a 4-year period between 2003 and 2006 and grazing occurred at identical times as vegetative mass permitted. Biomass was harvested before and after grazing for calculation of dry matter (DM) yield and biomass consumed. Sub-samples were used for determination of N concentration and in vitro digestibility. Mean herbage N-yield for grazed treatments was 131% of ungrazed and greatest for grazed-fertilizer and grazed-fertilizer plus compost. Grazed paddocks with no added N produced similar DM yield to those with added N. Estimated nitrogen fixation contributed an annual average of 82 kg ha−1 to herbage-N yield from the alfalfa-grass paddocks. Barley (Hordeum vulgare L.) silage grown after termination of the grazed pastures produced 72% more herbage DM from grazed paddocks than ungrazed, but no significant (P < 0.05) differences occurred among amendments.


Introduction
The Alberta tame hay area decreased from 2.26 to 1.56 million ha and tame pasture from 2.46 to 2.19 million ha between 2010and 2019(Agriculture Statistics Factsheet 2011, 2019).The area of perennial tame pasture and hay decreased in western Canada in the Subhumid, Semiarid and Boreal Plains regions by 1.15, 0.18 and 0.87 million ha (total of 2.2 million ha), respectively, between 2006 and 2016 (Liang et al. 2020).This area is now occupied by cereal, pulse and oilseed production.To meet nutritive demands of the Canadian cow herd, pasture and hay productivity needs to be increased as the acres broken may not be easily replaced.A close positive relationship between beef cattle numbers and perennial land area has existed prior to and during this period (Liang et al. 2020).Beef cow numbers in Alberta declined from 2.0 million in 2006 to 1.44 million in 2021 (Alberta Cattle Feeders 2020;Statistics Canada 2021).Unfortunately, there is a 30-40year trend for decreasing hay yields in western Canada and across most Canadian provinces (Jefferson and Selles 2007).This puts pressure on the beef industry relative to feed supply in dry years and on expansion of the cow herd.
We understand that grassland productivity is N-limited (Malhi et al. 2004) and fertilizer-N application may increase yields but increasing fertilizer costs discourage use.Jefferson and Selles (2007) reported a long-term negative correlation between hay yield, and fertilizer price and change in average annual temperature.Positive responses for grass pasture productivity to fertilizer-N application have been observed (Elliot et al. 1961;Doran et al. 1963;Nuttal et al. 1980;Kopp et al. 2003) in studies conducted in the western Canadian Parkland.Because of increasing demand and higher costs of fossil fuels fertilizer-N prices are increasing.Therefore, methods must be developed to maintain or increase carrying capacity and beef production on pasture, while reducing dependency on fertilizer-N.
Some producers feel that intensive rotational grazing may eliminate all fertilizer-N requirements due to the return of animal excreta to the pasture (Baron et al. 2002).The animal ingests N in organic matter and returns it through urine and feces to pasture in a more mineralized and plant available form.Ruminants retain small amounts of ingested N with the remainder returned to pasture as excreta (Whitehead 1995;Mathews et al. 1996).However, fertilization may be necessary because excreta returned may cover only 20% or less of the pasture area (Haynes and Williams 1993;Russelle 1996) and losses from urine and dung can occur from volatilization (NH 3 ), denitrification (N 2 O) and leaching (NO 3 − ) (Haynes and Williams 1993;Whitehead 1995).Nuttal et al. (1980) found economic returns from mixed alfalfa-grass pastures maximized at 90 kg N ha −1 and 20 kg P ha −1 when the stocking rate was 3.7 head ha −1 .Herbage yields increased up to 185 kg N ha −1 , but nitrate accumulated in the 30-60 cm depth of the soil profile.
Using legumes and grass-legume mixtures (N 2 fixation) or manure may be less expensive routes to increasing soil-N supplies in pasture compared to application of fertilizer-N (Russelle 1992;Russelle 1996).Issah et al. (2020) found fixed N in alfalfa (200 kg N ha −1 ) > cicer milkvetch (128 kg N ha −1 ) > sainfoin (68 kg N ha −1 ).Approximately 70%-95% of alfalfa-N may be derived from N 2 fixation (West and Wedin 1985;Chen et al. 2004).In pastures much of the legume-N transfer to grass must occur below ground from root and nodule decay and from excreta from grazing livestock (Russelle 1996).Yield response of forages to manure-derived N or in combination with fertilizer-N reaches similar optima to fertilizer-N (Russelle 1996).Eghball et al. (2002) estimated that beef feedlot manure, composted beef manure and swine manure mineralize about 30, 18 and 40%, respectively, of their organic-N content during the year of application (Eghball et al. 2002).
Including forages as hay or pasture in a cereal rotation is usually promoted to maintain or increase soil organic carbon, but a legacy of N may also be left to following crops from previously grazed grasslands reducing need for fertilizer-N.Berg and Sims (2000), in Oklahoma, observed productivity gains due to residual or carry over effects of fertilizer application (0 to 102 kg N ha −1 , annually for 5 years) for 3 years after fertilizer-N application ceased.Carry-over effects on cereal yield following breaking of forage crops have lasted 2-7 years (Ferguson and Gorby 1971;Hoyt and Leitch 1983).Ferguson and Gorby (1971) observed NO 3 − -N pools 1 year after breaking of a grass stand ranged from 90 to 135 kg ha −1 compared to 35 kg ha −1 after grain-grain fallow.After alfalfa plow-down, 34% and 7.5% of residue-N may be available for uptake in the first and second year after breaking by the following crop in a rotation (Heichel 1987).
The hypotheses of the current research were (i) that intensive grazing per se does not provide sufficient nutrient supplies to support pasture productivity on a Black Chernozemic soil in a short-season subhumid environment comparable to fertilizer-N and manure-N application and (ii) while pastures fertilized with N may exhibit a carry-over response in subsequent crops after termination, pastures that are only grazed will not.
The objectives of this study were to determine if intensive rotational grazing per se could support pasture productivity comparable to swards with fertilizer or manure applications and to determine if grazed pastures with no amendments could provide residual benefits to a subsequent crop in rotation.

Materials and methods
The experimental site was located at Lacombe, Alberta, Canada (52 • 28 N, 113 • 45 W; 847 m) on an Orthic Black Chernozemic Ponoka soil (classified as an Udic Boroll in the USDA taxonomy) developed on glacio-fluvial lacustrine parent material of silt-loam to loam texture.The upper 10 cm of soil consisted of an average of 15% clay, 34% silt and 51% sand (Mapfumo et al. 2000).Soil pH using distilled water was 5.4.The soil organic matter content was approximately 9%-10% with a high mineralization potential.Initial means of NO 3 − , extractable P and K were 65, 109 and 481 μg g −1 in the surface 0-15 cm in May 2002.Soil nutrient content as affected by treatments in the current study will be shown in a subsequent article.Because of the coarse texture of the soil and undulating nature of the surrounding area this exact location was deemed suitable only for livestock production compared to crop production in research conducted in the 1950s and 1960s (Doran et al. 1963).

Application of amendments
Fertilizer, compost and hog manure were applied after the first grazing in June 2003, June 2004, June 2005and June 2006 on dates and at rates of equivalent total-N shown in Table 1.No additional fertilizer P and K were applied during the study as initial concentrations were high as indicated above.
Fertilizer-N was broadcast as ammonium nitrate at 112 kg N ha −1 (Table 1) using a dribble spreader.As stated previously Nuttal et al. (1980) found economic responses to nitrogen applications on pasture increased up to 90 kg N ha −1 .Kopp et al. (2003) increased beef cow carrying capacity of meadow bromegrass-alfalfa pastures by 64% by applying an average of 96 kg ha −1 fertilizer-N annually.
Compost was applied with a ground-driven compost spreader (Model 75D; Millcreek Turf Equipment, Lancaster PA) with a 1.5 m application width.The compost applicator was calibrated to deliver appropriate amounts of compost based on the dry matter (DM) percentage and preliminary total-N concentration, but when total applications were assessed, application was not as precise as desired, particularly in 2003 (Table 1).Ammonium nitrate was applied to the fertilizer plus compost paddocks at 66 kg N ha −1 prior to compost application as described for the fertilizer-N treatment.Beef  CN 2000, Leco Corp, St. Joseph, MI).Based on a calculated dry weight and total N concentration, amounts of compost were determined to apply the equivalent of 40 kg ha −1 total N in compost.
Hog manure was applied with the intent to deliver the same total N application rate as the fertilizer-N treatment.However, due to calibration difficulties with the applicator, actual rates of N varied from 73 to 146 kg N ha −1 (Table 1) with a mean of 96.Hog manure was stored for a year as slurry in a lagoon from a farrow to finish operation.Manure for the trial was obtained when the entire pit was being emptied.Samples of the slurry were taken during application and refrigerated at 4 • C until being sent to a commercial laboratory for complete analyses.The plot-sized, shop-built (Agricultural Technology Centre, Alberta Agriculture and Rural Development, Lethbridge, AB) applicator was designed to pump or inject liquid manure into the soil.It was mounted on a three-point hitch with manure delivered through hoses on to a splash plate into soil immediately behind 0.4 m diameter coulters spaced 0.6 m apart.

Grazing management
Intermittent grazing on the fenced 9 m × 30 m paddocks was carried out at dates indicated in Table 1 in a manner described previously by Baron et al. (2002).Grazing occurred for all treatments (except for the ungrazed control) and replicates at the same time, at an average disk meter height (weighted disk method (Bransby et al. 1977) of 21 cm over years and grazing times.Heifers were removed from the paddocks at approximately 10 cm of residue remaining with grazing completed in no more than 3 days.However, the grazing period lasted only as long as it was required to graze down to 10 cm.From two to eight heifers were placed in a paddock at one time, depending on forage availability.On average this stocking density was 3.5 heifers per day in each paddock, with three grazing periods per year on each paddock, except in 2006 when two grazing periods occurred.On a full-scale intensively grazed pasture this would represent strip-grazing with stock density averaging 129 animals ha day −1 within a pasture strip.The rest period between grazing periods or return to the paddock averaged 48 days from 2003 to 2005, but was 115 days in 2006.Animals were provided with water in paddocks at all times.The ungrazed treatment was cut with a commercial farm swather at each grazing time and all material removed as in a hay management system.

Sampling procedure
Pasture sampling occurred before and after grazing.For randomization of sampling, a grid of 2 m × 2 m cells was superimposed on each paddock, with the perimeter of the grid 1 m from the outside of the paddock.Three locations on the grid were chosen at random for each sampling period, with the stipulation that each sample area must be at least one grid-square away from one chosen for a previous harvest in the same year, and two grid-squares away from one chosen for the same harvest.Within a grid-square, three stakes were placed at random to mark sampling areas.Prior to grazing, a rectangular frame 25 cm × 100 cm was placed with one corner against the marker stake, and the herbage inside the frame was cut to a height of 2.5 cm.The herbage harvested from each paddock was bulked and weighed fresh, then subsampled for determination of concentration of DM and forage nutritive value.After grazing, a similar procedure was followed with the frames placed diagonally opposite the pregrazing position at each marking stake.
Biomass sampled prior to grazing was assumed to be that available for grazing and was used in the calculation of DM yield summed over the season.Nitrogen yield was calculated as the sum of the product of DM yield and N concentration for each sampling time throughout the grazing season.Biomass sampled after grazing was assumed to be residual DM which would enter the residue pool of the pasture in an organic state, also summed over the season.The difference between available forage and residual material is that consumed by the grazing animals.

Chemical and biological composition
Sub-samples of before and after grazing fresh biomass (250 g) to be used for DM determination were weighed fresh and dried at 80 • C for 72 h then re-weighed and discarded.Similar sub-samples (250 g) taken before and after grazing to be used for determination of N concentration and in vitro true digestibility (IVTD) were dried at 50 • C for 72 h.The dried sub-samples were ground through a Wiley mill (Model No. 4; Arthur H. Thomas Co. Philadelphia, PA.) equipped with a 2mm screen.Concentration of total-N in both available forage and residual sub-samples was determined from dried, ground material as described above for compost.Concentration of IVTD was determined using a filter bag system (ANKOM Technology Corporation; Fairport, NY) similar to that described by Vogel et al. (1999).A 48-h stage 1 of the procedure described by Marten and Barnes (1980), then undigested residues were treated with neutral detergent fibre solution (Van Soest and Robertson 1980) in the determination.Concentrations of N and IVTD were determined for available forage and residue at each grazing but are expressed as the weighted concentrations for each season.Yields of IVTD were determined by summing the concentration times DM yields at each grazing for seasonal totals.
Sub-samples taken from each of the sampling grids in the alfalfa-grass treatment prior to grazing were hand separated into alfalfa and grass components to determine percentage alfalfa and grass on a DM basis.Plant density determinations were made three times during each growing season at three locations chosen randomly on three diagonal transects within each paddock.Quadrats (15 cm × 25 cm) were placed on the ground and meadow bromegrass, alfalfa and other species were counted on a plant basis.Averages of transects and times of year were analysed statistically as described for other data.
Derived variables West and Wedin (1985) concluded that in the absence of measurement (e.g., isotope dilution), N 2 fixation by alfalfa in orchardgrass-alfalfa stands could be assumed to be 90% of the N-yield attributed to the alfalfa component.While this value is in agreement with other research conducted in Western Canada (e.g., Chen et al. 2004) a lower N 2 fixation rate is possible due to effects of urine and feces (Russelle and Buziniky 1988) which could occur with the high stocking densities used in this study.In a review of a very large number of N 2 fixation studies in Europe, Anglade et al. (2015) found a median value of 70% and a range of 63%-80% of shoot N in alfalfa derived from atmospheric N. Therefore, for the sake of conservative approximation N 2 fixation was assumed to represent 80% of alfalfa-N content and the remaining 20% due to uptake from soil.
Excreted-N (kg ha −1 ) was determined as described in Baron et al. (2002).For grazed treatments excreted-N levels (dung and urine, kg ha −1 ) after animal consumption (kg ha −1 ) were estimated from the difference between available forage-N yield and residue-N yields before and after each grazing.Percent N retention by non-lactating grazing animals (Baron et al. 2002) was subtracted from 100% and multiplied as a fraction by N yield consumed after each grazing and summed over each season from 2003 to 2006.

Nutrient carryover
After grazing and measurements for pastures were completed in the fall of 2006, glyphosate (N-(phosphonomethyl) glycine) was applied at a rate of 2 L ha −1 via a conventional field crop sprayer to terminate all vegetative growth present.In the spring of 2007 "AC Lacombe" barley (Hordeum vulgare L.) was direct seeded at 100 kg ha −1 into all paddocks using a zero-tillage drill (John Deere Model 750: Deere & Co., Moline, IL).A whole-plant harvest was carried out at the soft dough stage using an automated plot harvester.A strip 1 m × 20 m was harvested at 5 cm above ground surface in each of the

Statistical analyses
The experimental design for each year of the study was a randomized complete block design.Because the treatments were perennial the randomization of treatments within reps was the same each year.The statistical analyses were done using SAS version 9.4 (SAS Institute Inc., Cary, NC).All variables were tested for homogeneity of variance using the conditional studentized residuals analyses from Proc GLIMMIX.Variables were declared to have a normal distribution, except % alfalfa composition, which exhibited a non-normal distribution.Year (2003Year ( , 2004Year ( , 2005Year ( and 2006) ) was included in the model for all variables, except barley whole-plant DM yield, nitrogen yield and nitrogen concentration, grown and harvested in 2007, only.Year was treated as a factor upon which measurements were repeated across.Therefore, a repeated measures analysis was conducted in accordance with the PROC MIXED procedure of SAS software (Littell et al. 1996;Piepho and Eckl 2013).The process of model fitting for PROC MIXED using the lowest Akaike Information Criterion (AIC) was identical to that described by Piepho and Eckl (2013) for perennial crops grown for consecutive years in a single location.The AR1 structure had the lowest AIC.Piepho and Eckl (2013) also found the AR1 structure lowered AIC for yield taken over multiple years at one location and a single-year (various production years) perennial trials across several locations.When an year × treatment interaction (P < 0.05), was significant (P < 0.05) variable mean responses were summarized for treatments within each year and for years.The variable alfalfa composition was analyzed in PROC GLIMMIX as a Beta error distribution with default logit link function.Percentage data (0-100) were transformed to proportions (0-1) prior to this analysis.The means shown by year are detransformed means for the sake of relevance, but differences shown were based on the comparisons of transformed means and a transformed LSD was used to determine mean differences.In all analyses mean difference were determined using LSD (P ≤ 0.05) when F statistics were significant at P ≤ 0.05.

Results
Mean monthly temperature in 2006 was 1.4 to 3.0 degrees above the long-term average from April to July.This was the only extended period of abnormal temperatures.Long-term mean monthly temperatures for the months of April to October are 3.7, 9.8, 13.6, 16.1, 14.9, 10.1 and 4.4 • C, respectively.Seasonal precipitation in 2002 (establishment year) and 2003 was only 68% and 76% of the respective long-term averages (Table 2).The very dry periods were May to July in 2002 and June to August in 2003.April-May precipitation was above average in 2003.The only abnormally wet year was 2006 when July-September precipitation was 147% of the long-term average.

Pasture herbage yield and nutritive value
Averaged over years, herbage yield of the ungrazed treatment was similar to that of the grazed-alone and alfalfa-grass treatments, but lower than treatments with fertilizer-N or manure added (Table 3).All of the grazed, including grazed alone and alfalfa-grass, pastures had similar DM yields.On average, the ungrazed treatment exported 10.0 t ha −1 of DM annually.Dm production declined from 13.4 t ha −1 in 2004 to 9.1 t ha −1 in 2006 in spite of greater precipitation in 2006.
The IVTD yields from all grazed treatments were higher than that from the ungrazed treatment (Table 3).In vitro true digestible yield followed the same pattern as herbage DM yield over years.The IVTD concentrations were similar among treatments.

Herbage N-yield
Averaged over years, the grazed fertilizer-N and the fertilizer plus compost treatments had numerically highest herbage-N concentrations and the ungrazed, the lowest (not significant; P > 0.05; Table 3).The high herbage N concentration in 2003 was attributed to the very high available N supply at the beginning of the trial and to the drought conditions of 2003 which restricted growth.
Herbage N-yield (Table 4) included N taken up from soil by meadow bromegrass in all treatments and N 2 fixed by the alfalfa in the grass-alfalfa mixture.Averaged over years ranking of treatments for N-yield followed that for herbage-N concentration, but ranking among treatments for N-yield var- x Within columns means not followed by the same letter are significantly different according to LSD (P < 0.05).
w SE = standard error of the mean.ied among years.There was no difference among treatments for 2003.From 2004 to 2006 the ungrazed treatment had the significantly lowest N-yield while the grazed-alone and alfalfa-grass treatments fluctuated yearly, with N-yields similar to the highest treatment or as low as the ungrazed.The fertilizer-N treatment was consistently among the highest in N-yield.By 2006, all grazed treatments had similar N-yields.
For the ungrazed control the herbage-N yield represented the amount of N exported (average, 181 kg N ha −1 year −1 ) each year.For all other (grazed) treatments, the N yield included N consumed or left in vegetative residue on the pasture.

Residue
Residue DM and IVTD yields (Table 5) followed the same trend as pasture herbage yield (Table 3); that is, the ungrazed treatment was significantly lower than the others, which were similar.As noted previously, the ungrazed treatment was not an exact control for the grazed treatments as all biomass above the 2.5 cm cutting height was removed from the plot.The similarity of the grazed-alone treatment to those with amendments applied is of greater importance.IVTD concentration of residue followed a different pattern from the herbage.The fertilizer-N treatment had higher residue IVTD Table 5. Residue DM yield and in vitro true digestible and nitrogen yields and concentrations for meadow bromegrass pasture averaged over 4 years at Lacombe, AB.

Residue yield Concentration
Treatment DM z (mg ha −1 ) I V T D y (mg ha −1 ) N( k gh a −1 ) N( gk g −1 ) I V T D( gk g x Within columns, means not followed by the same letter are significantly different according to LSD (P < 0.05).
w SE = standard error of the mean.
concentration than the ungrazed, fertilizer + compost and alfalfa-grass treatments (Table 5).Grazed alone and hog manured treatments also had higher IVTD concentration than the alfalfa-grass The ungrazed treatment had lower residue-N yield than all of the grazed treatments which were similar within that criterion.(Table 5).Among grazed treatments alfalfa-grass and hog manured treatments had lower residue-N concentration than the N-fertilizer treatment (Table 5).Residue-N concentration was lowest in the ungrazed treatment.while the hog manured treatment was lower than the grazed alone treatment (Table 5).

Alfalfa-N
There were year to year variations in the percentage of alfalfa in the stand and these are related to yearly differences observed in N-yield and estimated fixed-N 2 (Tables 4 and 6).Averaged over years, N 2 -fixation contributed an estimated 82 kg ha −1 to the herbage-N of the grass-alfalfa mixture.Only in 2005 was N-yield from the fertilizer-N treatment greater than the alfalfa-grass pasture.In all other years, the N-yield from the grass-alfalfa was similar to the N-amendment treatments (Table 4).

Carryover effects
In 2007, following pasture termination, the whole-plant DM yields of barley grown without fertilizer-N or amendment on the previously grazed plots were 58%-92% higher than the ungrazed treatment (Table 7).The N-yield for the grazed treatments excepting the alfalfa-grass treatment were significantly higher than the ungrazed treatment.Although N con-Table 6. Alfalfa stand composition, N-yield, estimated N 2fixed and herbage N-uptake from soil for a mixed grassalfalfa pasture for 4 years at Lacombe, AB. z Data for stand composition were transformed to proportion (0-1) from % and analyzed as a beta distribution then means de-transformed; differences were determined with LSD at the transformed level.y Within columns, means not followed by the same letter are significantly different according to LSD (P < 0.05).
centration of the barley DM was not significantly different among treatments, the concentration for the amended treatments ranged from 16% higher than the ungrazed for the alfalfa-grass to 43% higher for the fertilizer-N.

Productivity and grazing
This study showed that on a Black Chernozemic sandy soil with relatively high soil organic matter the effects of intensive rotational grazing on a pure stand of meadow bromegrass produced similar, but slightly lower (94%), DM yields than those fertilized with N, manure or compost additions or grown as an alfalfa-grass mixture.Pasture herbage DM yields for grazed alone were similar to grazed with fertilizer-N added at more than 100 kg N ha −1 (Table 3).Soils with high organic matter contents such as in this study generally have high potential to release plant available-N, especially in the surface 0-15 cm (Malhi et al. 1992).
Generally, the grazed alone-treatment was more balanced with respect to N-dynamics on a system basis than the nongrazed treatment, which resembled hay management.In other research under hay management, with nutrient removal, DM yield increased with N application up to 200 kg N ha −1 at Lacombe and Eckville, AB (Malhi et al. 2002).The high soil NO 3 concentration prior to commencement of the study may have contributed to the lack of differences between the grazed alone and N amended treatments.The N taken up by herbage in the ungrazed and grazed-alone treatments originated entirely from available soil mineral N at the initiation of the study and mineralization of soil organic matter.For the grazed alone treatment, most of the N taken up and consumed was returned to pasture.A small amount of the N consumed (5 to 25%) is retained by the animal and the rest excreted in urine and feces (Whitehead 1995;Mathews et al. 1996), which is subject to loss (Russelle 1996).As N content of the sward increases a greater percentage of ingested-N is returned as urine compared to feces; urine-N is then susceptible to loss or plant uptake within a short time (Haynes and Williams 1993;Whitehead 1995).Because non-lactating animals were used in grazing of the paddocks it was assumed that 90% of ingested N was returned to all the grazed pastures (Whitehead 1995).An estimation of N consumed by cattle averaged over the grazed treatments in this study is 209 kg N ha −1 (134-343 kg N ha −1 ) when residue-N is subtracted from average herbage-N yield.Although some losses from urine and dung must have occurred, the calculated average of N returned in excreta for the grazed-alone treatment (166 kg N ha −1 ) was 81% of the excreta-N from grazed plus fertilizer-N and 85% of the average of grazed treatments that included amendments or alfalfa (Table 4).The amount of or-ganic residue-N returned to pasture in the grazed-alone treatment (135 kg N ha −1 ) was almost identical to that of the average of all other grazed treatments (Table 5) with N added from outside sources, but was 1.64 times the ungrazed treatment.
Quantification of N-losses is beyond the cope of this study, but some loss of N deposited is likely.It is difficult to quantify losses because the size of losses and the proportion among the loss processes are dependent on soil moisture, air temperature, soil texture, compaction, N content of urine and capacity for herbage growth to take up soil-N in the pasture (Whitehead 1995;Soussana and Lemaire 2014;Selbie et al. 2015).Based on averages of European, Australian and New Zealand information Selbie et al. (2015) summarized the disposition of N deposited in urine patches, which includes losses, as 13% ammonia volatilization, 2% nitrous oxide, 20% leaching, 41% uptake and 26% gross immobilization.Climatic conditions and density of urine patches play a role in overall area based losses.Losses due to volatilization may range from 25% to 30% in temperate areas such as in the current study (Russelle 1992), although may be higher in semi arid regions under high temperatures.Under the dry conditions in this study leaching losses were likely much lower than indicated by Selbie et al. (2015).We estimated uptake of applied fertilizer-N to be approximately 41%, which agrees with uptake of N from urine patches estimated by Selbie et al. (2015).
N-export or removal from ungrazed treatment averaged 181 kg N ha −1 annually.Using the ungrazed treatment as a reference, the grazed-alone treatment contributed 57 kg N ha −1 more to its average annual herbage uptake of N (Table 4).In the absence of nutrient replacement, DM yields of hay crops decrease over time as observed with the ungrazed treatment in the current study.The ungrazed treatment was not a perfect control for the grazed treatments as most nutrients contained in above ground vegetation were removed by clipping and harvest.This would include removed P and K, as well as N, which may also cause nutrient deficits that limit DM yield.In the grazed treatments these other consumed nutrients would pass through the animal as some form in the excreta (Haynes and Williams 1993).So other nutrient imbalances and deficits were more likely in the ungrazed than grazed treatments.
The variation among treatments for residue IVTD was unexpected and hard to explain.For grazed treatments the residue consisted of a mixture of stems and leaves, whereas the ungrazed treatment residue would have consisted of stubble or stem bases alone.Animals prefer to graze leaves over stems and leaves have higher digestibility than stems (Nelson and Moser 1994;Baron et al. 2000).The proportion of leaves and stems grazed in treatments may have differed.Thus those grazed treatments with higher stem content left in the residue may have had lower IVTD.For example, in the alfalfagrass treatment alfalfa stems were visible in the standing residue.Alfalfa stem bases may be 50% less digestible than grass stems due to high lignification resulting in low cell wall digestibility (Buxton and Fales 1994).Averaged over three stages of plant development meadow bromegrass leaves and stems exhibited in vitro organic matter digestibilities of 714 and 686 g kg −1 , respectively, in 1987 and 736 and 726 g kg −1 , respectively, in 1989 (Baron et al. 2000).

Imported sources of N
Nutrients from fertilizer-N, compost and hog manure were added to a system that was already close to a nutrient balance (i.e., grazed alone).It is acknowledged that N-yield for treatments with fertilizer-N applied was higher than grazed-alone, although not always significantly.This would indicate that the meadow bromegrass attained a threshold mass but continued to take up and assimilate more N if available.Ferguson and Gorby (1971) found that N concentrations in wheat following breaking from grass, grass-alfalfa or alfalfa continued to increase if available soil-N was present even when there was no further DM yield increase.
In this intensive system an estimated N-uptake for the Namended treatments averaged 46 kg ha −1 more than the grazed-alone, but did not add substantially to herbage DM yield.This is an N-use efficiency of 41% based on 112 kg ha −1 added N for the grazed-fertilzer treatment.In central and north-central Alberta and Saskatchewan, spring wheat had N-use efficiency of added N ranging from 30% to 55% with N fertilizer (urea) rates from 25 to 100 kg N ha −1 (Haderlein et al. 2001) with higher efficiency at lower N rates.Herbage yield of the fertilizer and fertilizer plus compost treatments increased by 18% and 20% respectively, compared to the ungrazed control, but only 6% and 8% when compared to the grazed-alone treatment, respectively (Table 3).It may be that other forage pasture species have greater yield potential and possess higher nutrient demands, but this is not likely.
The quantity of mineral-N provided by the fertilizer-N treatment at time of application may not have been necessary on this soil.In an adjacent pasture area Baron et al. (2001) found that nitrate accumulations in the 0-60-cm depth ranged from 100 to 200 kg ha −1 depending on grazing intensity, after 4 years of grazing and annual applications of 100 kg N ha −1 as ammonium nitrate fertilizer to perennial grass and cerealbased pastures.The fertilizer-compost treatment provided a total of 264 kg N ha −1 of inorganic fertilizer and 215 kg N ha −1 in composted manure over 4 years (479 kg N ha −1 total).
The hog manure added 385 kg N ha −1 in organic and mineralized forms of N and the fertilizer treatment added 448 kg N ha −1 in inorganic-N.While the grazed-fertilizer-compost treatment had the most total-N applied, it produced DM yield as high as any treatment and actually reduced the amount of inorganic fertilizer applied by 184 kg N ha −1 or 41% over 4 years.Because mineral N is susceptible to leaching or denitrification losses it is environmentally advantageous to supply N in organic forms if it is equally effective in supplying crop N requirements.About 90% of organic and mineralizable-N is available from hog manure within the year of application, compared to 20% from composted beef manure (Eghball et al. 2002).It may be that a substantial amount of hog manure-N was lost due to NH 3 volatilization at application.

Alfalfa-grass pasture
Actual average herbage yield for alfalfa-grass was almost identical to the grazed only which was pure meadow bromegrass.While the benefits of a legume component in the stand are acknowledged, there was no advantage over a grazed pure meadow bromegrass stand from a productivity perspective in this study or in N-uptake by the following barley crop (Table 7).In hay management (ungrazed) legume content may have been advantageous in replacing exported N, although removal of other nutrients may have been limiting also.Most of the N taken up by the grass in the alfalfagrass stand would have to be taken up from the soil solution after mineralization from original soil organic matter, or from decomposing plant residues including legume fixed-N 2 or legume material that was consumed and excreted as feces and urine (Russelle 1992;1996).Theoretically, the grazed alfalfa-grass should have been a tighter N-balance with less N ha −1 derived from original soil organic matter and the estimated N 2 fixation.However, this is speculation within the scope of the study.The approximate demand by the grass portion for soil-N for the grass-alfalfa treatment was about 68% of the pure grass stands in the amended treatments as determined by the average quantity of grass in the stands and its N-concentration.Thus, the demand for N from soil per se is smaller for the grass-alfalfa stand than from the pure grass stands, because the alfalfa portion of the stand should be largely self-sufficient (West and Wedin 1985).
On average estimated N 2 fixed in the grass-alfalfa treatment was 82 kg ha −1 year −1 .Yang et al. (2010) indicated average fixation for alfalfa-bromegrass hay (% alfalfa unspecified) was 60 kg ha −1 , and was linearly related to legume biomass yield.Mean pasture herbage yield in the current study (11.5 t ha −1 (Table 3)) was about three times the base yield in their work (3.7 t ha −1 ).Therefore, we might have expected as much as 180 kg ha −1 fixed N. In 2004, fixed N 2 was estimated at 152 kg N ha −1 when the alfalfa composition was 41% (Table 6).Averaged over years, alfalfa composition was 29%, which would likely result in considerably less fixation averaged over years.The decreasing percentage of alfalfa over years (Table 6) had more to do with increasing density of grass than reduction in alfalfa plant number (data not shown), but the dry weight of alfalfa in the herbage decreased as did estimated N 2 fixation.The reduction in alfalfa content may have been an impact of increasing competition from grass and frequent defoliation on alfalfa plant size or density (Pearen and Baron 1996).In bromegrass-alfalfa mixtures cut up to four times per year alfalfa yield in the mixtures declined by 43%-70% and stem density declined by 22%-58% in the first cut over 3 years when grown at Lacombe and Vegreville, AB.Alfalfa stem density at cut four in their study (Pearen and Baron 1996) declined by 44%-67%.Appreciable amounts of legume-N may be transferred to the grass in a mixture, but may be dependent on legume-grass species mix and climate.For grasslegume mixtures [alfalfa-orchardgrass (West and Wedin 1985) and alfalfa-meadow bromegrass (Chen et al. 2004)] under pasture management N 2 fixed ranged from 68 to 70 kg ha −1 .In Minnesota, Heichel and Henjum (1991) observed between 6% and 47% of grass uptake was derived from transfer from the legume; in Texas Gebhart et al. (1993) found less than 5% legume-N was transferred between alfalfa and crested Wheatgrass, but 20% of legume-N was transferred from sweet clover to crested wheatgrass.
Nutrient carry over DM productivity and N-yield of whole-plant barley after termination of the perennial treatments supports the conclusion that the intensive pasture management (grazed-alone) had the greater capacity to retain soil-N within the pasture system than forage removal without nutrient replacement in the non-grazed treatment.When N-uptake of the ungrazed treatment was subtracted from the average of Nuptakes from the other five treatments (Table 7) they contributed 78 kg ha −1 more N to barley uptake following stand termination.The grazed-alone treatment increased the following crop, whole-plant barley yield by 77% and N-uptake by 131% over the ungrazed control.Therefore, grazing increased or maintained productivity and contributed 83 kg N ha −1 more to the whole-plant barley crop compared to the ungrazed control.The alfalfa-grass treatment increased whole-plant barley yield by 66% and N-uptake by 95% and contributed 60 kg N ha −1 more to barley uptake compared to the ungrazed control.By comparison the fertilizer-compost treatment increased whole-plant barley yield by 95% and Nuptake by 141% over the ungrazed control and contributed 97 kg N ha −1 more to the barley crop than the ungrazed control.Thus, the fertilizer-compost treatment was very effective in storing N within the system and making it available to the subsequent crop after perennial stand termination.
Intensive grazing was more influential on N-uptake of the following barley crop than the methods that involved N-imports, including the alfalfa-grass treatment.Although there was no significant difference among grazed treatments for herbage-N uptake in barley, when the N-uptake of the grazed alone treatment is subtracted from the others, only the grazed fertilizer-compost treatment contributed mathematically more N (14 kg ha −1 ); the lowest contribution of N in barley was the alfalfa-grass treatment (−23 kg ha −1 ).Implications Perrillat et al. (2004) concluded that use of fertilizer-N on pastures increases productivity and carrying capacity and thereby reduces cost per animal, mitigating financial risk of pasture backgrounding and finishing operations in Saskatchewan.In 1999, fertilizer costs at rates applied (112 kg N ha −1 ) represented 25 to 30% ($107 ha −1 ) of pasture operating costs.The current (2021) fertilizer-N cost of $900-1050 per tonne of 46-0-0 (Richardson Pioneer Ltd, Camrose, AB, April-September 2021) would be $88-111 ha −1 more and would account for about 50% of pasture operating costs.Compared to extensive grazing a benefit of intensive rotational grazing should be reduced fertilizer-N requirement due to greater turnover of soil nutrients through the grazing animal and more even distribution of excreta (Moore 1999).At this time producers are unwilling to invest in fertilizer application on either hay or pasture due to slim margins in beef calf production.Our results indicate that on black Chernozemic soils intensive rotational grazing may not require fertilizer application to produce herbage yields similar to approximately 100 kg N ha −1 , annually or at least fertilizer-N rates lower than expected in hay production.In hay production fertilizer application, including N, P and K are likely required due to nutrient removal.Where soil mineralization potential is lower than in the soils used here, the advantage for intensive rotational grazing may not be the same as in the current study.
In a hay clipping study Malhi et al. (2002) estimated that seeding mixed alfalfa-smooth bromegrass stands provided savings in fertilizer-N relative to pure grass stands equivalent to about 100 kg N ha −1 .Based on results of our study the benefit of intensive grazing per se of the pure stand of meadow bromegrass may be equivalent to application of 112 kg N ha −1 , although grazing alone contributed an average of only 57 kg N ha −1 with reference to the ungrazed treatment during the grazing phase of the experiment from 2003 to 2006 (Table 4).In the following barley crop the grazed alone treatment supplied 83 kg N ha −1 more than the ungrazed and similar amounts as the grazed fertilizer treatment (Table 7).
In the current study DM production of the alfalfa-grass treatment was not superior to the grazed alone treatment averaged over 4 years and it would be difficult to differentiate the legume impact from the grazing impact.Further, the barley crop following grass-alfalfa did not take up more N than the grazed alone treatment.Thus in terms of DM production and N-uptake by the following crop, on this black soil the grazing impact dominated the legume impact and was similar to the grazing plus fertilizer-N application.However, in an annual system, using field pea as a legume-N source, Beckie et al. (1997)

Conclusion
The hypotheses were that intensive grazing per se does not provide nutrient supplies and pasture productivity comparable to fertilizer-N and manure-N application and while pastures fertilized with N may exhibit a carry-over response in subsequent crops after termination, pastures that are only grazed will not.In contrast to those hypotheses, our conclusion is that on a black Chernozemic soil, intensive rotational grazing without imported N addition can provide sufficient nutrients to provide equivalent productivity to pastures where supplemental N is applied during years of grazing.The carryover affect of sufficient available N to meet the requirement of a succeeding annual crop was equally evident for the grazed alone treatment as for those which received N amendments.for at least 1 year after stand termination.
observed a benefit of 25 kg N ha −1 in the moist dark brown soil zone and Rasmussen et al. 2012 concluded that white clover and alfalfa grown in mixtures with grasses could replace between 88 and 220 kg N ha −1 of fertilizer-N in a succeeding barley crop.

Table 1 .
Description of grazing and amendment treatments imposed on meadow bromegrass pastures from 2003 to 2006 at Lacombe, AB.

Table 2 .
Monthly precipitation and total precipitation from April until October inclusive from 2002 to 2006 and long-term average monthly and seasonal total precipitation at Lacombe, AB.

Table 3 .
Pasture herbage DM yield, in vitro true digestible yield and in vitro true digestible concentration for above ground meadow bromegrass pasture (available forage) averaged over 4 years at Lacombe, AB.
z DM = dry matter.y IVTD = in vitro true digestibility.

Table 4 .
Herbage nitrogen yield as affected by grazing and amendment treatments from 2003 to 2006 at Lacombe, AB.
z Within years, means not followed by the same letter, a to c, are significantly different according to LSD (P < 0.05).

Table 7 .
Whole-plant barley DM yield, nitrogen yield and nitrogen concentration in the year following pasture termination at Lacombe, AB.Within columns, means not followed by the same letter, a to b, are significantly different according to LSD (P < 0.05).
z DM = dry matter.y