Grazing season length and stocking rate affect milk production and supplementary feed requirements of spring-calving dairy cows on marginal soils

The objective of this study was to investigate the effect of increasing stocking rate (SR) and extending grazing season (GS) length on pasture and animal productivity on a marginal, poorly draining soil type. The study was a multiyear (2017 to 2020, inclusive) whole farm systems evaluation with a 2 × 2 factorial experimental arrangement of treatments. The systems evaluated comprised 2 GS lengths, average (AGS; 205 d) and extended (EGS; 270 d), and the 2 whole farm stocking rates were medium (2.5 cows/ha) and high (2.9 cows/ha). We used this study design to create 4 grazing system intensities (500, 600, 700, and 800 cow grazing days per hectare per year). In 2017, cows were randomly allocated to 1 of the 4 whole farm systems precalving and remained on the same treatments for the duration of the study. We found no significant differences in total average annual pasture production [14,133 ± 538 kg of dry matter (DM) per hectare] or sward chemical composition between GS and SR treatments over the 4-yr period, with the exception of average crude protein content, which was lower for EGS (211 g/kg DM) compared with AGS (218 g/kg DM). Grazed pasture production was significantly increased in EGS treatments (+758 kg of DM/ha) compared with AGS (9,917 kg of DM/ha), whereas conserved silage DM production was greater for AGS (+716 kg of DM/ ha) compared with EGS (3,583 kg of DM/ha). Neither GS nor SR had a significant effect on daily or cumulative lactation milk and fat plus protein production per cow (5,039 and ±440 kg, respectively). Increasing SR resulted in increased milk fat plus protein yield per hectare based on increased grazed pasture utilization. These results add further credence to the important additive contributions of both extended grazing and SR intensification to achieve high levels of


INTRODUCTION
Dairy farming is widely acknowledged to be financially volatile, with an ever-changing landscape of milk and input prices, variable and fixed costs, milk yield, and other variables that affect farm financial returns (Ramsbottom et al., 2015). Among global dairy systems, pasture-based production, where the majority of animal feed requirements are derived from grazed or conserved pasture, are unique both in terms of prevalence and design (Horan and Roche, 2019). In Ireland, intensive dairy production is mainly aimed at maximizing milk production from grazed pasture during a season extending from mid-to late February to mid-to late November. The economic impacts of increased pasture production and utilization (t of DM/ha) on farm operating profitability (both per kilogram of product and per hectare) has been widely reported for such systems (Dillon et al., 2008;Ramsbottom et al., 2015;Neal and Roche, 2020), while there is growing awareness of the role of grazing systems in delivering ancillary benefits, such as enhanced farmland biodiversity and improved animal welfare (Delaby et al., 2020). On that basis, further improving the efficiency of pasture-based milk production systems is considered as a primary opportunity for the sustainable intensification of food production in Europe over the next decade (EU, 2019).
Intensive food production from grazed pasture is commonplace in temperate regions that have above average rainfall (Delaby et al., 2020). In such circumstances, soil conditions and inclement weather are among the biggest limitations affecting the extension of the grazing season (GS) to further increase the productivity of such systems (Creighton et al., 2011), Grazing season length and stocking rate affect milk production and supplementary feed requirements of spring-calving dairy cows on marginal soils particularly on more marginal and poorly draining soil types (Läpple et al., 2012;Ramsbottom et al., 2015). The Border, Midland, and Western Region of Ireland consists of 13 counties, including the 6 border counties with Northern Ireland, is one such region which is characterized by marginal poorly draining soil types resulting in a shorter GS, increased concentrate and forage supplementation, and reduced pasture utilization (Ramsbottom et al., 2015) and farm profitability (NFS, 2019), compared with national average values. Previous studies have concluded that extending the GS in such circumstances, could substantially improve farm profitability by decreasing overall direct costs by 10% (1.6 cents/L). Despite strong economic incentive, there remains a paucity of information on the impacts of extended grazing on cumulative pasture utilization on marginal soils. Although database studies such as Läpple et al. (2012) and Ramsbottom et al. (2015) and shorter term controlled-component experiments (Kennedy et al., 2006) can describe general responses based on estimated pasture utilization, full farmlet systems studies are required to measure system level responses and interactions among integrating components that cannot be captured elsewhere (Roche et al., 2018;Horan and Roche, 2019).
The objective of this study, therefore, was to investigate the effect of simultaneously increased GS length and system intensification (increasing stocking rate; SR) on pasture and animal productivity on a marginal poorly draining soil type where best practice grazing management practices are adopted to increase grazing days per hectare. The experimental hypothesis is that increasing GS length and system intensification, resulting in increased grazing days per hectare, would contribute to increased pasture productivity and animal performance while also substantially reducing the requirements for supplementary feeds on dairy farms on such soil types.

MATERIALS AND METHODS
This study was carried out at Ballyhaise College (54° 051'N, 07° 031'W) in the Republic of Ireland over a 4-yr period from 2017 to 2020. All measures undertaken involving cows during the experiment were approved by the Teagasc Animal Ethics Committee. The experimental site comprises a variety of soil types including alluvial, brown earth, brown podzolic, and gley on a lower Silurian sandstone bedrock. The bedrock is overlain by drumlin-shaped mounds of stony, slowly permeable boulder clay and by gently sloping areas of glacial drift. The topography ranges from alluvial flatlands (along the Annalee River which transects the site) to various shaped, re-current drumlins with steep slopes (9°-8°) and intervening U-shaped valleys. The levees, which run contiguous to the river banks, are mostly brown earth soils and are moderately well drained. The area between the levees and the lower slopes of the drumlins are level and are mainly poorly drained. This poor drainage is caused by the high ground water table, very slow surface water run-off, and the slow permeability of the soil resulting from its poor structure and sticky consistency.

Experimental Whole Farm Systems
The 4 experimental groups were comprised of 2 GS lengths, average (AGS; 205 d from March 15 to October 20) and extended (EGS;270 d,February 15 to November 20), and 2 global SR treatments, medium (MSR; 2.5 cows/ha) and high (HSR; 2.9 cows/ha). The 2 GS treatments were chosen to represent the current average grazing season length in the Border, Midland, and Western Region of Ireland (205 d; Läpple et al., 2012) and best practice extended grazing on poorly drained wet soils (270 d; Patton et al., 2012). The MSR was designed as a low-cost pasture management system for a post European Union milk quota scenario based on high levels of pasture utilization and a low level of concentrate supplementation, while also providing adequate conserved forage for winter based on previous studies published at the same site (Patton et al., 2016). The HSR represented the highest overall SR allowable at farm level within the European Union Nitrates Directive (2.9 livestock units per hectare; Directive 91/676/EEC). Each experimental group has its own farmlet, which was managed separately depending on GS length and SR requirements. As the HSR system had increased daily pasture demand (kg DM/ha), less area within this treatment was conserved for pasture silage surpluses compared with the MSR treatment during the GS.
The experimental area was a permanent grassland site containing greater than 80% perennial ryegrass (Lolium perenne L.). At the beginning of the farmlet study, a total of 37 paddocks (of on average 0.87 ha) were grouped into 18 sets of 4 (balanced on location, pasture species, topography, drainage, distance from the milking parlor and soil type) and randomly assigned to each feed system (FS). All paddocks assigned to an individual FS at the beginning of the study remained in that FS for the duration of the study. The EGS treatments animals were turned out to grass by day and night as they calved, from mid-February, and remained at grass full-time until mid-November, with the exception of individual days where inclement weather conditions prevented animals from accessing pasture for even short access periods. In contrast, the AGS treatment remained indoors until mid-March and were rehoused in mid-October. A rotational-stocking system was practiced and on-off grazing (Kennedy et al., 2009) was used as a management tool to facilitate grazing for short access periods during inclement weather conditions. Due to the important impacts of weather conditions on pasture access in spring and autumn, the number of entire grazing days achieved for each treatment was documented. The number of entire grazing days per cow were the total number of days when each cow had access to pasture after both morning and afternoon milking (with each morning and afternoon allocation considered as 0.5 d). Separately, the total number of cow grazing days per hectare was also calculated as the product of the number of entire grazing days and the SR within each treatment. Grazing management was accomplished by weekly monitoring of farm grass supply within each SR treatment and all groups were managed similarly at grazing (in terms of target pregrazing yield, postgrazing compressed sward height and residency time in paddocks). The residency time within each paddock ranged from 1.5 to 2.5 d per subpaddock and was determined by achieving a target postgrazing compressed sward height of 4.0 to 4.5 cm. No mechanical topping of the swards took place for the duration of the experiment and all surplus pasture was conserved as winter feed. Annual chemical N fertilizer application was constant across all GS and SR treatments at 250 kg/ha with slurries, also evenly redistributed back across the experimental area based on GS length.
Concentrate supplementation commenced at 3 kg fresh weight per cow per day for the EGS cows post calving at grass, and was reduced and removed once grass growth was adequate to meet animal requirements. The AGS treatments commenced lactation on 5 kg of fresh weight concentrate per day and were supplemented with an additional high CP soybean-based concentrate while indoors on a predominantly grass silage diet. The ingredient composition of the main concentrate feed was 30% barley, 31% beet pulp, 28% maize distillers, 6% molasses, and 5% minerals plus vitamins. Mean concentrate feed quality was 135 g/kg of CP, 94 g/kg of crude fiber, 48 g/kg of ash, and 895 g/kg of OM. The ingredient composition of the high CP concentrate feed was 25% barley, 25% beet pulp, 12.5% maize meal, and 37.5% soybean. Mean high CP concentrate quality was 274 g/kg of CP, 87 g/kg of crude fiber, 49 g/kg of ash, and 865 g/kg of OM.

Animals
The data presented was collected from 108 high Economic Breeding Index (EBI; Berry, 2015) animals of Holstein-Friesian or Jersey Holstein-Friesian breeds. The average EBI of the animals was €175 during the 4-yr period of the study (ICBF, 2020). In year 1, the experimental animals were assigned to 1 of the 4 whole farm systems before calving based on breed, parity, calving date, previous milk yield, BCS, BW, and genetic merit (EBI). All multiparous animals were subsequently retained on the same feeding system for the duration of the study. Primiparous animals entering the study were randomly assigned to feed treatment based on EBI, breed, calving date, and precalving BW and BCS.

Pasture Measurements
Grazing details were collected on all paddocks during each grazing rotation for each of the 4 yr of the study (2017-2020). Pregrazing pasture mass (>3.5 cm horizon) was determined before grazing on each paddock for each of the treatments by harvesting 5 quadrants (0.5 m × 0.5 m) of grass using Gardena hand shears (Accu 60, Gardena International GmbH). The 5 quadrants were spaced equally along the diagonal of each paddock. All mown pasture from each quadrant was collected and weighed, and a subsample was taken and dried for 16 h at 90°C for DM determination. The average paddock pregrazing pasture mass above a cutting height of 3.5 cm was then calculated. Pregrazing and postgrazing sward heights were also determined on each paddock before and after grazing by taking between 30 and 50 measurements across the diagonal of the paddock using a folding pasture plate meter with a steel plate (Jenquip).
Pasture removed was also calculated as proposed by (Delaby et al., 1998) based on the following formula: Pasture removed = (pregrazing height − postgrazing height) × sward density, kg of DM/cm per ha.
Daily pasture allowance and daily pasture disappearance were then calculated based on the residency time within each paddock. Grazing data were analyzed for demand. Silage intakes were measured on a weekly feed budget basis where daily silage allocated to dry cows and milking cows were weighed separately and recorded for each animal in each group for each week.
Pasture samples were collected from each paddock for each FS and were dried at 40°C for 48 h and milled through a 1-mm sieve. Samples were bulked by FS by week and analyzed for DM, ash, ADF, NDF (Soest, 1963), CP (Leco FP-428; Leco Australia Pty Ltd.), and OM digestibility (Morgan et al., 1989).

Animal Measurements
Cows were milked at 0700 and 1530 h daily throughout lactation in all 4 yr of the study. Weekly milk production was derived from individual milk yields (kg) recorded at each milking. Milk fat, protein, and lactose concentrations were determined weekly from 1 successive evening and morning milking sample from each cow using a Milkoscan 203 (DK-3400, Foss Electric). Weekly fat plus protein (milk solids; MIS) were also calculated. Milk, fat, protein, lactose, and MIS yield per hectare were calculated by multiplying total lactation milk and MIS production per cow by the respective SR of each treatment.
Individual cow BW was recorded weekly upon exit from the milking parlor using an electronic scale (Tru-Test Ltd.), whereas individual cow BCS was assessed every 3 wk by the same individual throughout the study on a scale of 1 to 5 in increments of 0.25, as outlined by (Edmonson et al., 1989). The BW and BCS variables analyzed were mean BW and BCS, in addition to BW and BCS at calving, at nadir, and at the end of lactation.

Statistical Analyses
The effects of GS and SR on total lactation yields for milk, MIS, milk composition, BW, and BCS were determined using mixed models (Proc Mixed; SAS Version 9.4, SAS Institute Inc.), with cow included as a repeated effect to account for the repeated lactations per cow; additionally, a compound symmetry covariance structure with heterogeneous variances provided the best fit to the data. Initial models included the effects for GS, SR, season (spring, mid-lactation, and autumn), parity (1, 2, 3, and >3), year (2017,2018,2019,2020), breed (Holstein-Friesian and Jersey Holstein-Friesian), calving day of year, and interactions of GS, SR, season, parity, and breed. The effects of GS, SR, season, and year on grazing characteristics were determined using mixed models (Proc Mixed, SAS) with rotation number included as a repeated effect. A compound symmetry covariance structure among records within paddock provided the best fit to the data.

Climatological Conditions During the Study Period
Monthly rainfall and mean soil temperature (10 cm) data for the 4-yr period and for the previous 10-yr average period are presented in Figure 1. Average annual rainfall over the 4 yr (1,042 mm) was similar to the 10-yr mean (1,028 mm) and ranged from a minimum of 937 mm (2017) to 1,162 mm (2020) during the study period. Although monthly rainfall amounts were similar to long-term averages, seasonal rainfall varied significantly, particularly during the extremities of the GS in spring and autumn. Cumulative spring rainfall amounts were 82, 130, and 157% of the 10-yr average (206 mm) in 2017, 2019, and 2020, respectively, whereas autumn levels ranged from 70% of 10-yr average values (273 mm) in 2018 to 150% in 2019. In contrast, mean soil temperatures were consistent with long-term average values during the study period.

Grass Production, Grazing Characteristics, and Feed Inputs
Total average annual pasture production over the 4-yr period was 14,133 (SEM = 538) kg of DM per hectare (Table 1) and was unaffected by GS length, SR, or their interaction. Year had a significant impact on pasture characteristics due to variation in climatic conditions, and we found no significant interactions between year, GS length, and SR. Despite the overall similarity in DM production, grazed pasture production was greater (P < 0.05) in EGS (+758 kg of DM/ha) compared with AGS (9,917 kg of DM/ha), whereas conserved silage DM production was greater (P < 0.05) for AGS (+716 kg of DM/ha) compared with EGS (3,583 kg of DM/ha). Similarly, grazed pasture production was greater (P < 0.05) in HSR (+754 kg of DM/ha), whereas less pasture was conserved as pasture silage (−726 kg of DM/ha) compared with MSR (9,919 and 4,304 kg of DM/ha, respectively). The number of entire grazing days per cow and cow grazing days per hectare was greater for EGS (+30 d/cow and +79 d/ha, respectively), whereas only cow grazing days per hectare were greater for HSR (+88 d) compared with MSR (540 d/ha). On average, EGS achieved an additional +15 grazing days per cow and +38 cow grazing days per hectare in spring, and an additional +16 d per cow and +44 cow grazing days per hectare during autumn. In contrast, HSR had minimal effect on cow grazing days per hectare during spring Cahill et al.: EFFECTS OF GRAZING SEASON AND STOCKING RATE ON PRODUCTIVITY and autumn (+ 5 and +8 d, respectively) with the vast bulk of additional cow grazing days achieved during mid-season (+75 d/ha).
The effect of GS, SR, and season on grazing characteristics is presented in Table 2 for the entire grazing season and separately for spring (turn out to pasture until the end of the first rotation in April), mid-season (from mid-April until late July), and autumn (from August until the end of the grazing season in mid-November). Season had a significant effect on grazing characteristics and we also found GS by season interactions (P < 0.001) for both pregrazing pasture mass and paddock residency time. Postgrazing residual sward height was greatest during mid-season (46 mm), least  The SR had no significant impact on any of the grazing characteristics measured. Significant interactions between GS and season were observed for both pregrazing pasture mass and paddock residency time. The interactions arose due to the reduced pregrazing pasture mass (−344 kg of DM/ha) and residency time (−0.6 d) for EGS in spring compared with AGS (1,302 kg of DM/ ha and 3.6 d, respectively) and increased pregrazing pasture mass (+460 kg of DM/ha) and residency time (+0.8 d) for EGS during autumn compared with AGS (1,451 kg of DM/ha and 2.7 d, respectively). The effects of GS, SR, and season on sward chemical composition is presented in Table 3. We found no significant differences in sward chemical composition between GS and SR treatments with the exception of CP content, which was lower (P < 0.01) for EGS (211 g/kg) compared with AGS (218 g/kg).
Concentrate and silage supplementation varied significantly between GS, SR, and S both within lactation and based on the combined lactation and nonlactating periods ( Table 4). The AGS treatment received 646 kg of DM of parlor concentrate during lactation compared with 597 kg for EGS over the 4 yr of the study with the majority of the additional concentrate fed while indoors during early spring (+20 kg of DM/cow) and early midseason (+29 kg of DM/cow). In addition, AGS also received an additional 47 kg of DM of high CP supplement during the spring indoor period to supplement the lower CP content of the indoor pasture silage diet. Pasture silage supplementation during lactation was also greater (P < 0.001) for AGS (896 kg of DM/cow) compared with EGS (555 kg of DM/cow) with the difference evenly split between spring and autumn. Total concentrate supplementation was greater (P < 0.001) for AGS (702 kg of DM/cow and 1,858 kg of DM/ha) than EGS (598 kg of DM/cow and 1,437 kg of DM/ ha, respectively). Similarly, when lactation and nonlactation pasture silage supplementation is combined, total pasture silage requirements were also greater (P < 0.001) for AGS (1,859 kg of DM per cow and 5,025 kg of DM/ha) compared with EGS (1,436 kg of DM/cow and 3,873 kg of DM/ha, respectively). Stocking rate had no significant effect on supplementary feed characteristics per cow; however, both total concentrate (+865 kg of DM/ha; +116%) and pasture silage (+2,295 kg of DM/ ha; +125%) requirements per hectare were greater (P < 0.001) for HSR compared with MSR (747 and 1,839 kg of DM/ha, respectively). Finally, we found a significant GS by SR interaction evident for total silage requirements per hectare as the additional silage required with HSR was greater (P < 0.01) with AGS (+869 kg of DM/ha; +19%) compared with EGS (+392 kg of DM/ha; +11%). No significant difference was found in total pasture growth (14,224 kg of DM/ha) between SR treatments; however, more herbage was harvested as silage within the MSR (4,304 vs. 3,579 kg of DM/ ha/yr for HSR). When silage requirements (Table 4) are considered against silage conserved (Table 1)      ments whereas the HSR system only produced 82 and 70% of silage requirements for EGS and AGS, respectively.

Milk Production and Composition, BW, and BCS
The effect of GS, SR, and their interaction on mean daily milk production during the study is shown in Table 5 and Figures 2 and 3. Average lactation length was similar for all GS and SR groups (278 d). We found no significant effect of GS length or SR on mean daily milk yield (17.0 kg), fat content (5.15%), and fat yield (0.87 kg/cow) during lactation, whereas GS also had no significant impact on lactose content (4.06%), and the persistency of lactation was similar for both GS and SR treatments ( Figure 2). The EGS treatment achieved a greater (P < 0.001) milk protein content (3.93%) than AGS (3.88%) from spring and maintained superior milk protein content over the entire lactation. Similarly, SR had a significant (P < 0.05) impact on both milk protein and lactose contents, which was lower for HSR (−0.03% in each case) compared with MSR (3.92 and 4.83%, respectively). We found a significant (P < 0.05) GS by SR interaction evident for both milk fat and protein contents due to the superior composition of MSR (+0.07 and +0.06%, respectively) compared with HSR within AGS, whereas HSR achieved superior milk fat and protein contents compared with MSR within the EGS treatment (+0.18 and +0.01%, respectively). Season had a significant (P < 0.001) impact on all daily milk production parameters. Mean daily milk, fat, and protein yields were greatest (P < 0.001) in spring (21.6, 1.12, and 0.66 kg/d, respectively), least in autumn (11.6, 0.65, and 0.50 kg/d, respectively), and intermediate during summer (19.4, 0.98, and 0.75 kg/d, respectively). Milk fat content were greatest in autumn (5.51%), least in mid-lactation (5.03%), and intermediate in spring (5.21%). In contrast, milk protein content were greatest in autumn (4.28%), least in spring (3.52%), and intermediate in mid-lactation (3.85%). Finally, milk lactose content was greatest in spring (4.93%), least in autumn (4.69%), and intermediate in mid-season (4.84%).
The effects of GS and SR on total lactation milk production are shown in Table 6. Neither GS nor SR had a significant impact on cumulative lactation milk and fat plus protein production per cow (5,039 and 440 kg, respectively). Although GS also had no significant impact on milk and fat plus protein production per hectare (13,755 and 1,188 kg, respectively), HSR produced significantly (P < 0.001) more milk and fat plus protein per hectare (+1,853 and +190 kg, respectively) compared with MSR (12,828 and 1,093 kg, respectively). In addition, the seasonal profile of milk and fat plus protein production per hectare varied significantly, with the majority of the additional milk (+66%) produced during mid-lactation ( Figure 4). Finally, there was no significant effect of GS length or SR on BW and BCS during lactation.

DISCUSSION
In the context of the recent intensification of milk production systems in Ireland (Kelly et al., 2020;Cahill et al.: EFFECTS OF GRAZING SEASON AND STOCKING RATE ON PRODUCTIVITY  Ramsbottom et al., 2020), the current study was designed to evaluate the potential for extended grazing and increased SR to improve both pasture utilization and animal performance and reduce supplementary feed requirements within intensive grazing systems on a marginal soil type. Previous studies have questioned the viability of milk production on heavy clay soils in Ireland due to lower pasture production and shorter GS (Brereton and Hope-Cawdery, 1988;Shalloo et al., 2004). The current results are in contrast to these previous conclusions, as all treatments achieved high levels of pasture production and utilization efficiency in comparison with current average dairy industry estimates (Horan and Roche, 2019). Moreover, the current study has demonstrated that large quantities of high-quality pasture can be grown and efficiently used on such soil types when appropriate grazing management practices are adopted. The effects of early turn out to pasture in spring or extended grazing in late autumn have been investigated in many previous studies, albeit that the benefits of extending the GS in 1 period is typically achieved by deferring grazing at other times (McFeely and MacCarthy, 1981;Kennedy et al., 2006;Claffey et al., 2019). This is the first study to evaluate the combined impacts of extended grazing in both early spring and late autumn on a marginal soil type with poor drainage characteristics where the combined impacts of impeded land drainage, reduced soil temperatures, latitude, and topology make grazing conditions considerably more challenging at the study site.
The results of the study highlight the potential of extended grazing in early spring and late autumn to maintain high levels of milk productivity from pasture with a significantly reduced requirement for both concentrate and silage supplementation. Indeed, the significant reduction in supplementary concentrate (−15%) and pasture silage (−23%) requirements for EGS represents the primary benefit of extending GS length within the current analysis similar to Kennedy et al. (2006). In addition, the provision of high-quality spring pasture within the early lactation diet of EGS resulted in superior milk protein content in spring, which subsequently persisted throughout the entire GS (Table 5; Figure 3). Although O'Donovan et al. (2004) reported improved pasture quality due to reduced pasture mass and increased leaf content in early grazed Cahill et al.: EFFECTS OF GRAZING SEASON AND STOCKING RATE ON PRODUCTIVITY swards, no such subsequent effect was observed in this study, which is similar to the findings of Kennedy et al. (2006). The chemical composition of swards previously grazed in early spring, reported in these earlier studies was similar to that reported across all treatments in this study, indicating a high proportion of digestible leaf material within the sward in all treatments. The overall similarity in sward chemical composition between treatments is unsurprising as EGS, although grazed earlier in spring and later in autumn, required comparably increased spring and autumn rotation lengths and pasture masses to extend the grazing season (Table 2). In addition, the high pasture mass of AGS treatments in spring due to the extended winter housing period was quickly corrected via increased silage conservation in early summer thereby minimizing any potential impact on mid-season and subsequent pasture quality. In addition to the effects of the timing of initial defoliation, the optimum SR that should be imposed with extended grazing treatments has received little attention. Stocking rate impacts animal performance through its influence on daily pasture allowance per cow with a decrease in individual cow performance as SR increases (McCarthy et al., 2011(McCarthy et al., , 2016. Although the effects of SR on milk productivity within grazing systems have been widely reported, there have been conflicting reports on the effect on pasture production and utilization, particularly on marginal soil types where pasture growth distribution is more seasonal with lower spring and autumn production (Patton et al., 2016). Moreover, previous studies have shown that GS length is reduced as SR increases (McCarthy et al., 2011). Similar to both McCarthy et al. (2012) and Coffey et al. (2018) on free draining southern Irish soils, we found no significant effect of SR on total pasture production within the current study. At the same time, as grazing days per cow remained constant, increasing SR resulted in a significant increase in grazing days per hectare (+88 d), and grazed pasture used (+754 kg of DM/ha), culminating in additional fat plus protein production (+190 kg/ha). As additional supplementary feeds and reduced silage conservation were used to  support the increased demand of HSR, overall grazing intensity (measured via postgrazing height) was similar for both SR treatments. Moreover, as only MSR conserved sufficient silage to meet animal requirements, the increased milk production of HSR treatments must be considered in the context of inadequate silage conservation and a consequent requirement to purchase silage deficits elsewhere outside the system. Grazing season length had a significant impact on silage self-sufficiency with extended grazing producing 82% of silage requirements for HSR in comparison to just 70% for the AGS treatment. Ultimately, as the high SR resulted in both increased grazed pasture utilization and an increased deficit of winter forage production, the impacts of increased SR within the study reflect the impacts of overall system intensification (defined as increased SR and imported feed requirements) rather than just the impacts of increased SR. On that basis, and irrespective of the level of intensification, the results of the study indicate that extended grazing can yield increased grazed pasture utilization and minimize supplemental feed requirements on poorly draining and marginal soil types.
A unique feature of the design of this study was to permit the evaluation of the combined impacts of extended grazing and increased SR culminating in increasing cow grazing days per hectare from 501 d in AGS MSR to 669 d within EGS HSR. The overall GS length achieved for EGS was 39 d below target (270 d) due to a combination of 22 missed grazings (equivalent to 11 entire days) during the GS when ground conditions and rainfall necessitated that animal remained indoors to avoid pasture damage, in addition to an early housing of EGS during autumn due to limited pasture availability. Although the rehousing of grazing dairy cows is unavoidable during periods of inclement conditions, the significant impact of pasture shortages during autumn suggest that building increased pasture supplies in early autumn can facilitate a further extension to GS length and requires further evaluation on poorly draining soils. Without considering the ancillary benefits of extended grazing days on labor and housing related costs, the combined annual feed cost savings (€95/ha) predominantly from extended grazing are considerable. In addition, Ramsbottom et al. (2015) has previously observed that the true differential in production costs between systems differing in supplementary feed use is frequently +160% of the difference in feed costs alone, and so a complete financial evaluation of the economic differences between treatments is merited. Moreover, these results add further credence to the important additive contributions of extended grazing across a range of intensification levels to maximize pasture and land resource utilization by increasing grazed pasture utilization in dairy systems on marginal soils.

CONCLUSIONS
Results from this experiment on a marginal soil type indicate that extending GS length can maintain high  levels of milk production per hectare across a range of system intensification levels while significantly reducing the reliance on supplementary feeds within spring-calving grazing systems. In addition, the effects of extended grazing of grazed pasture utilization were consistent at both levels of SR intensification evaluated within the study. The results emphasize the importance of additional efforts to increase pasture availability and utilization and animal performance on poorly draining and marginal soil types.