Improved efficiency of grazing dairies using complementary pasture species and irrigation scheduling

Objective 1. Screen warm and cool-season annual and perennial pasture species for seasonal forage productivity under grazing conditions.

Initial stands of cool season perennial legumes and grasses were excellent 3 months after establishment. However, intensity and frequency of grazing in the pasture-based dairy system resulted in stand failure of these species within 1 year of establishment. Because of the lag between time of establishment and utilization of these forages combined with the expense of establishment, we concluded that the short duration in which these forages could be utilized does not justify their inclusion into pasture systems for intensive grazing systems that are on these dairies. Warm season perennial grasses, on the other hand, were persistent whether they were harvested at 10, 20, or 30-day intervals. Yield of bahia grass was greater than bermudagrass when harvested at 10 day intervals, there were no differences among warm season grass yields when harvested at 20 day intervals, but 30 days between harvests favored bermudagrass production (Table 1). Monthly distribution of forage growth depended on forage species. The species with the greatest yields and best compliment of growth were cereal rye, annual ryegrass, and bermudagrass (Figure 1). It is important to note that yields of pasture mixtures that were complementary, not additive. In other words, the monthly dry matter production of two species grown together was less than the sum of the two species grown in pure stands. Therefore, when attempting to plan forage systems for producers, appropriate experimentation must be established to test the productivity and distribution of specific species combinations and not assume that monthly productivity can be estimated by the sums of yield data when grown in monocultures.

Because forage production of winter annuals was tested by no-till establishment of research plots within bermudagrass pastures, we predicted the production distribution of combined warm season and winter annual forage systems by summing the forage production of each. The forage species combination with the greatest yields was cereal rye and annual ryegrass overseeded into Tifton 85 bermudagrass. The production curves predict that using this system will provide little to no forage during the month of October because a) the bermudagrass becomes dormant with the onset of cool nighttime temperatures and short daylengths during October, and b) the winter annuals are in a phase of establishment (Figure 2). Peaks of forage production occur during late spring and mid-summer, suggesting excess forage could be harvested for hay and fed back to cows during periods of low forage production.

Metabolizable energy (ME) of bahiagrass did not change regardless of the interval between harvest (Figure 3). However, the ME of bermudagrass decreased as interval between harvests changed from 10 to 20 days in length. The ME of bermudagrass did not change as the interval between harvests changed from 20 to 30 days. The winter annual forages had greater average ME than did the warm season perennial grasses (Table 2). Surprisingly, the ME of annual ryegrass was the least of the forages tested while wheat had the greatest ME. Regardless the ME of all winter annual forages are suitable for lactation. Therefore, selection of winter annual species for pasture systems should be selected based upon how the forage production matches with the needs of the milking herd.

Objective 2. Determine moisture release curves for the major soil types found on cooperating grazing dairies and monitor soil moisture content from irrigation practices during periods of negative water balance.
Soils maps were used to delineate major soil types on each farm. Soil moisture sensors were placed into the major soil types within each of two center pivot irrigation systems on each farm. One pivot was designated as a ‘control’ pivot, in which the producer was asked to manage irrigation with their normal practices. Soil water data was downloaded to a computer (located in the manager’s office) every 15 minutes by telemetry, but the producer was not permitted to see the data and could, therefore, not adjust irrigation based upon soil moisture data. The second irrigation pivot was designated as a ‘treatment’ pivot, in which the telemetric soil moisture data was projected on a farm map. The data was presented in such a way to illustrate % field capacity and the amount of irrigation needed to bring the volumetric water up to 90% field capacity. Producers were instructed to irrigate when soil moisture reached 60% of field capacity.

Producers did not implement irrigation scheduling as instructed. Reasons for failure to do so varied depending upon farm. One manager had troubles operating the laptop computer in his office and did not know how to program the pivot to vary water application based upon soil moisture information on the map. The second farm manager did not prioritize irrigation scheduling because of other farm operations that took precedence.

Objective 3. Create a nitrogen budget to determine when, and by how much, the dairies are deficient in nitrogen supply.

Ammonia Volatilization
In general, the amount of N lost through NH3 volatilization in each sampling period was less than 2.4 kg N ha-1 at both farms (Fig. 4). Overall NH3 losses amounted to 1% of the applied N at the Wrens farm and 2% of the applied N at the Quitman farm. When expressed on an area basis, our NH3 losses are at or below 8 kg N ha-1 (Table 3). Our data indicate that pasture-based dairies are likely to have lower NH3 losses than published data for confined dairies.

The low NH3 emissions found on both farms were surprising considering the Wrens farm applied twice the amount of N than the Quitman farm. Because these farms were both under center pivot irrigation, it is likely that both the UAN applied through the irrigation and the broadcast urea moved into the soil by irrigation water soon after N applicatoin. It is well documented that supplemental irrigation can play an important role in the rate of NH3 volatilization from N fertilizers. This effect was probably due to leaching of urea into the soil, thereby reducing NH3 losses. Although information on irrigation timing was not available for our study, the small NH3 losses suggest that irrigation applications occurred after each urea application at the Wrens farm.

It is likely that not all NH3 volatilized from our pastures was derived from applied N fertilizers. Ammonia is also lost from N in cattle manure and urine deposited on the surface of the pasture, with urine being the larger of these two sources. The application of poultry litter at the Wrens farm may have also contributed to the measured NH3 losses. The methods used herein could not distinguish between sources of NH3 emission, but the results suggest that these pasture-based dairies emit low NH3 to the atmosphere even in high N input scenarios. We attribute the low emissions to irrigation moving excreta N as well as fertilizer N into the soil.

Nitrate Leaching
The maximum NO3 - concentration observed in the cup lysimeters at the Wrens farm was 4.8 mg N/L. This occurred in January, when water flux through a 1-m depth was small (3.14 cm). Large water fluxes were estimated in October/November and March/April, but NO3 - concentrations in those months were low resulting in low amounts of leached NO3 - (0 to 0.5 kg N/ha/month). Total NO3 - losses for the monitored period (June 2008 to August 2009) at the Wrens farm amounted to 2 kg N/ha (Table 3), which correspond to 0.3% of the applied N. Total water flux at 1-m depth in the monitored period was 55 cm, which corresponds to 33% of the total amount of water (precipitation + irrigation) received.

Nitrate concentrations and water fluxes at a 1-m depth from the main soil types at the Quitman farm (Tifton, Stilson, and Fuquay soils). In general, NO3 - concentrations and water fluxes were greater than those at the Wrens farm. Maximum concentrations ranged from 12 to 27 mg N/L with maximum water fluxes as high as 30 cm/month. As a result, leached NO3 - ranged from 1 to 16 kg N/ha month-1 at the Quitman farm. Total NO3 - leaching losses for the monitored period (August 2008 to August 2009) amounted to 40 kg N/ha, which corresponds to 12.5% of the applied N (Table 3). Total water flux at 1-m depth in the monitored period ranged from 88 to 95 cm in the different soils, which represented 44 to 49 % of the total amount of water (precipitation + irrigation) received.

The Farm Gate Nitrogen Budget

a. Managed Inputs and Outputs
Managed inputs included fertilizer, grain, corn silage, and hay used. Rates and total amounts of N fertilizer varied between farms. At the Wrens Farm, monthly N application rates ranged from 16 to 55 kg N/ha for a total yearly rate of 491 kg N/ha. At the Quitman Farm, monthly N applications ranged from 14 to 43 kg N/ha, for a comparatively smaller yearly rate of 297 kg N/ha. Consequently, fertilizer N as % of total N inputs represented 79% at the Wrens farm and 41% at the Quitman farm. The smaller rate of N fertilizer applied at the Quitman farm was in part counterbalanced by a larger rate of N
input in the form of grain feed (383 kg N/ha versus 97 kg N/ha at the Wrens farm). Grain feed as % of total N inputs represented 16% at the Wrens farm and 53% at the Quitman farm. Thus, grain feed and fertilizer N accounted for most of the N input on both farms (95% at the Wrens farm and 94% at the Quitman farm). Hay and silage made up the rest of the N inputs.

Monthly N exported in milk ranged from 574 to 1,526 kg N for the Wrens farm, and from 405 to 956 kg N for the Quitman farm (data not shown). Total exported N in milk as a percentage of total N inputs was 15.8% for the Wrens farm and 17.1% for the Quitman farm
Thus, both farms used similar proportions of the input N to produce milk

Outputs in Environmental Losses

a. N2O Emissions
In the field study carried out at the Wrens farm, N2O emissions followed distinct patterns for each of the treatments. Emissions were greater from chambers containing urine and manure than from control chambers, with the larger rates of emissions occurring within the first 2 days after application. Although urine and manure emission rates peaked between 4 and 5 mg N2O m2/hr, overall emission rates remained at a lower rate for the majority of the study period. Carbon dioxide emissions were greater for the manure treatment than from the urine or control treatments for two days following application, but were not different from the control treatments thereafter. Between days 7 and 10, CO2 emission was smaller from the urine treatment than from the manure or control treatment. Estimating the total loss of N2O-N over two weeks for a grazed paddock resulted in 1.78 and 0.78 kg N/ha for Studies 1 and 2, respectively. In both studies, total N2O emission was greater from urine and manure treatments than from control treatments. Furthermore, in the second study, emission from urine was greater than emission from manure. This may have been due to more anaerobic conditions created by the manure application, as indicated by a greater emission of CO2.

The second study in this experiment had lower overall emissions, but a higher WFPS, which may have been due to the fact that under wetter conditions a larger proportion of the emitted N is converted to N2 gas rather than to N2O. In addition, an increase in WFPS would decrease nitrification, decreasing N2O emissions. Because the studies were located in different paddocks, it is possible that different soil characteristics, such as pH, bulk density, etc. could have played a role in the different results of both studies. Of the total inputs of N on the Wrens farm, N2O represented a 2.65% loss from the system (3.36% of the total applied fertilizer), with estimated monthly losses ranging from 0.52 to 3.13 kg N/ha (Table 4). This represented a loss of 16.5 kg N/ha for the year. These results are consistent with other studies conducted under similar conditions. Compared to other outputs in the N balance, N2O was second to least of the outputs from the system.

b. NH3 Volatilization
Monthly ammonia losses ranged from 0.5 to 4.1 kg/N ha at the Wrens farm, and from 0.1 to 1.5 kg N/ha at the Quitman farm (Table 4). When expressed as percentage of the total N inputs, total ammonia loss was 3.0% at the Wrens farm and 0.9% at the Quitman farm. Greater losses at the Wrens farm were likely due to the use of surface applied granular urea, in contrast to the Quitman farm, which used UAN solution through the irrigation system.

c. NO3 Leaching Losses
Monthly nitrate losses ranged from 0 to 0.5 kg N/ha for the Wrens farm and from 0 to 1.6 kg N/ha for the Quitman farm. Annual losses were estimated at 1.6 kg N/ha for the Wrens farm and at 38.3 kg N/ha for the Quitman farm (Table 4), which correspond to 0.3% and 5.3% of total N inputs, respectively. Larger water fluxes from the soil profiles, paired with significant NO3 concentrations were the driving factors for NO3 losses on both farms. Nitrate concentrations were highest on both farms between December and April. Although water fluxes during those months were not consistently high, losses from these months were still the main contributors to NO3 loss, with March and April being the highest. These elevated concentrations of NO3 during the winter were expected and is explained by slower plant growth during the winter, and lower relative efficiency of N uptake by plants. Typically cool season forages have shallower root depths compared to warm season forages. This coupled with higher rainfall during these months may leach NO3 past the rooting zone. The Quitman farm also received 40 cm (irrigation and precipitation) more than the Wrens Farm during the year. It is interesting to note that although N fertilization at the Wrens farm was not reduced during winter, NO3 leaching losses were low during winter.

d. Nitrogen Surplus
The N balances from both farms resulted in a surplus of unaccounted for N. Surplus N found on the Wrens farm was 375 kg N/ha/yr or 78% of the total input N. On the Quitman farm, surplus N was 559 kg N/ha or 77% of total N inputs (Table 4). Nitrogen surplus may include N in animals exported from the farm (culled cows and bull calves), increases in soil inorganic N, N losses in surface runoff, and N immobilized in soil organic fractions. Nitrogen amounts in animals exported from the farm and N losses in runoff were not estimated, but were probably small. Increases in inorganic N could not account for N surpluses because in both farms the amount of inorganic N in the soil decreased during the study. At the Wrens farm, the area-weighted inorganic N changed from 98 kg N/ha in the initial sampling to 55 kg N ha-1 in the final sampling, for an average loss of 43 kg N/ha. Similarly, at the Quitman farm, the area-weighted inorganic N changed from 65 kg N/ha in the initial sampling to 56 kg N/ha in the final sampling, for an average loss of 9 kg N/ha. Therefore, immobilized N was the main suspected source of unaccounted N.

Objective 4. Develop educational and outreach information for training producers to be more efficient at, or transition into, pasture-based dairying.

Educational materials were developed that included information developed from this research project, but as important was information and techniques that were gleaned during interactions with producers. For example, pasture management was not an objective of this research project, but while conducting research on-farm we observed various philosophies and strategies on pasture utilization. Use of rising plate meters to quantify forage mass is an integral component of the pasture-based farms, so we purchased electronic rising plate meters and developed calibration curves for each of the forage species tested. We have used the rising plate meters in pasture trials, observed pasture responses to management based upon RPM data, and have modified the way we manage pastures because of the utility of the RPM. Use of RPMs was not an objective in this project, but the technology we have developed using RPMs is a direct spin-off from this project and we have now included use of RPMs into our outreach programs. Meantime, we observed pasture management methods used by experienced pasture-based dairy managers, those transitioning into pasture-based systems, and those who are new to the pasture-based dairying paradigm. We have learned from successes and mistakes from our producer clientele and have modified our outreach efforts accordingly. We have also acquired a better understanding of animal management and the importance of a good animal genetics program to minimize the nutritional and environmental stresses associated with pasture-based dairying in the Coastal Plains of Georgia.

One other aspect of pasture-based dairy systems for which we have gained an appreciation are some of the associated intangibles. In particular is the lifestyle of the producers and how they view their systems. We have included producers into the outreach agendas to provide first-hand testimonials as to the family value of the pasture-based system. Producers who we have interacted with have given testimonials that a)” the pasture based system has provided me with an opportunity to employ my children and provide them with a future in the dairy industry”, b) mission statements such as the following: “Our mission statement is to provide the highest quality product we can, with the least cost, in a socially and environmentally benign system, that leaves something for the next generation”, and c) we have time to take the grandchildren on extended trips in our RV. While these statements are from seasoned veteran pasture-based dairy owners, they are indicative of the mindset of the system and illustrate the value of personal time.

• Table 3. Summary of total fertilizer N applied and losses associated with the sampling period at Wrens (1 Jun 2008-31 Aug 2009) and Quitman (1 Aug 2008-31 Aug 2009) Farms.
• Table 4. Farm-gate nitrogen balance of two pasture-based dairy farms located on the Georgia Coastal Plain.
• Table 1. Yield of Tifquick bahiagrass, Coastcross II (CCII) and Tifton 85 bermudagrasses at different harvest frequencies (2 year means)
• Table 2. Mean metabolizable energy for winter annual species when overseeded in bermudagrass pastures.
• Figure 2. Monthly forage dry matter distribution of a hypothetical pasture mixture of cereal rye, annual ryegrass, and bermudagrass grown in the Georgia Coastal Plain.
• Figure 3. Metabolizable energy content of Tifquick bahiagrass, Coastcross II, and Tifton 85 bermudagrasses harvested at different intervals.
• Figure 4. Population trends of dairy cows on pasture vs. total cow numbers in Georgia. Note: Pasture-based numbers for 2011 are based upon current permit applications for pasture-based dairy operations.
• Figure 1. Growth distribution of species or species mixtures that had best persistence and productivity from small plot research trials located on pasture-based dairies.