Characteristics and Nitrogen Value of Stratified Bedded Pack Dairy Manure

"Compost" dairy barns are relatively new, and this manure, which we refer to as stratified bedded pack (SBP) dairy manure, has neither been characterized in detail nor defined in terms of its N supply. We measured physical characteristics, nutrient concentration, N mineralization, and N supply to corn (Zea mays L.) of SBP dairy manure from eight Minnesota farms. Concentrations of N, P, and K were generally higher than standard "book values" for solid dairy manure, were lower for P and K than typical solid dairy composts, and were highly variable within and among buildings. Average bulk density of SBP dairy manure was 58.2 lb/ft3. All SBP dairy manures produced nitrate during a 4-month-long incubation in soil, but the four with highest C:N ratios (19 to 21) immobilized N for 30 to 60 days. Infield fertilizer N equivalents to corn ranged from 1.4 to 12.1 lb/ton for quickly incorporated manure, but only -0.3 to 5.3 lb/ton when incorporation was delayed until spring. Guidelines for solid dairy manure were not reliable for predicting N availability from SBP dairy manure. Until validated prediction equations are available, we recommend farmers apply moderate rates of SBP dairy manure, incorporate it immediately to improve N supply, apply a basal rate of fertilizer N near planting time, and sidedress fertilizer N based on recommended soil or plant analysis for their region. Introduction "Compost" dairy barns generate a new manure source in the USA and other countries. Management of these manure packs, animal behavior, herd health, and milk production have been described in other publications [e.g., (1,3)]. Here, we emphasize the nutrient value of the manure pack to field corn and compare these results to solid dairy manure, but begin with a short description of manure pack management in these facilities. These barns have a retaining wall typically about 4 ft high that separates the feed alley from the manure pack. Bedding composed of sawdust, wood chips, or crop residues accumulates as additions are made to maintain a dry surface. Surface drying is promoted by a combination of forced air ventilation of the barn and twice daily stirring with an implement to incorporate dung into the pack and to mix wet bedding with dry. The result is a manure pack with two layers: (i) a loose surface several inches thick that gradually increases in manure content, moisture, and density until new bedding is added; and (ii) the compact layer, which increases in thickness as old surface layers are buried under new bedding, and which is partially composted. The total depth of the pack varies with the time since the last pack was removed, the time the cows spend on the pack, the area allotted per cow, and the frequency and amount of new bedding additions. Composting is restricted, presumably by lack of oxygen as the deeper layers are compressed, and active composting resumes when SBP manure is 17 July 2009 Crop Management moved to piles that are turned to provide aeration (12). In this paper, we introduce the term "stratified bedded pack" (SBP) dairy manure to avoid the implication that the material is fully composted and to indicate that this is a subset of solid (high dry matter content) dairy manure that is distinctly layered through active management of the loose surface layer. Little is known about the characteristics of SBP dairy manure, although two studies have presented some chemical analysis of the manure pack (2,5). Unlike other dairy manures, we anticipated that these managed deep manure packs could be sampled easily in advance of field application. In addition, the range of reported analyses for N and other nutrients is wide (3), so there is need for sampling recommendations. To our knowledge, there are no reports on N availability from SBP dairy manure. Such estimates are required for nutrient management planning, and it is not known whether estimates for solid dairy manure, such as those given by the MidWest Plan Service [MWPS; (7)], are relevant for SBP dairy manure. In an effort to address this lack of information, we conducted research on eight Minnesota dairy farms to determine: • physical characteristics and nutrient concentrations of SBP dairy manure in the pack; • spatial variation in nutrient concentration that would affect sampling protocol; and • fertilizer N equivalent (FNE) value of SBP dairy manure for corn grain production. SBP Dairy Barn and Manure Pack Management Eight cooperating farms were identified across Minnesota. Herd size in the sampled packs ranged from 38 to more than 200 lactating cows, with 60 to 97 ft2 of bedded pack area per cow (Table 1). On six farms, cows were confined to the building except during milking, but on the other two, cows had ad lib access to an outside lot and pasture. Some buildings were converted, whereas others were new construction. The amount of ventilation varied widely. Table 1. Physical characteristics of eight SBP dairy manure packs. x ‘Loose’ refers to the surface layer; ‘Compact’ to the remaining deeper part of the pack. y Jersey herd. All others were Holsteins or mixed large breeds. Farm Bedding Cow density (ft2/cow) Pack depth (inches) Pack density, as is (lb/ft3) Moisture, percent (w/w) loosex compact loose compact 1 Ground paperboard, beechwood, sunflower hulls 81 34 29.1 55.0 59 60 2 Ground particleboard, sawdust, small grain middlings 72 26 35.0 65.6 63 62 3 Sawdust, wood shavings, woodchips 64 27 41.0 55.3 74 70 4 Wood shavings, small grain straw 75 51 28.2 57.5 55 64 5 Very fine sawdust 97 29 28.6 60.6 55 61 6 Fine sawdust 89 48 49.4 58.9 60 64 7 Sawdust, woodchips 68 33 34.4 57.2 61 65 8 Fine sawdust 60y 30 41.0 56.7 59 64 17 July 2009 Crop Management All manure packs were based on woody materials with high C:N ratios (i.e., sawdust, wood shavings, or wood chips), but wood source varied from ground particleboard to fine hardwood sawdust. Some farmers had added other bedding on a trial basis, including sunflower (Helianthus annuus L.) hulls, small grain middlings or straw, and waste hay. All farmers stirred the pack twice daily with a skid steer or small tractor and frontor rear-mounted field cultivator, Danish S-shank cultivator, or multi-weeder. Bedding was added "as needed," but that definition varied among farmers. For example, one applied new bedding only when it became too wet for stirring. Characteristics of SBP Dairy Manure Each building, or discrete section of a partitioned building, was sampled and ancillary data were collected from the farmers in late September 2006. Two sampling devices were used: an Eijkelkamp Edelman auger for general chemical analysis of the entire compact pack depth and a custom corer to take samples from defined depths with known volumes. The Eijkelkamp auger was designed to sample wet soils and is available from many internet retail sources. We used it to collect samples that represented the entire compact pack depth. Samples of the loose surface bedding at each site were collected by hand. Six to nine locations in each manure pack were sampled with each device, following zig-zag transects to characterize areas near outside walls, the inside retaining wall, and the centerline of the pack, because we presumed these areas would differ due to cow behavior. Samples were stored in coolers with ice packs until they could be frozen. Samples from the auger were analyzed by a certified commercial laboratory for moisture, total N, organic N, ammonium-N (NH4-N), and total P and K. Subsamples from each sampler were ground with dry ice in a propeller-type mill (Steinlite Corp., Atchison, KS) and used for total N and C (combustion analyzer) and inorganic N analysis, and for N mineralization trials. Acetic acid was added to subsamples to retain NH4-N for total N analysis. PROC MIXED in SAS (SAS Institute Inc., Cary, NC) was used to determine effects of sampling location (fixed variable) within the barns and sampling depth (repeated measure), after transforming variables as needed to achieve more normal data distributions. Moisture concentration varied with depth at the time of sampling, averaging about 61% by weight for the surface and 64% for the compact layers (Table 1). These are slightly drier than typical solid dairy manure [67%, reference (8)] and wetter than dairy manure composts [55%, reference (11)]. Bulk density (as-is) of the compacted manure was similar among buildings, averaging 58.2 lb/ft3 (SD = 3.4 lb/ft3). Surface layer density varied, but contributed less than 10% of the total mass of manure at the time of sampling (excluding a barn with an 8-inchthick surface and a total pack depth of only 24 inches). Therefore, the total amount of manure to be applied can be estimated within about 10% by multiplying the measured volume of the manure pack (in cubic feet) times 58.2 lb/ft3. Concentrations of nutrients found within a manure pack were variable, often had non-normal distributions before data transformation, and differed among farms (Fig. 1). Total N concentration averaged about 1.12% and did not differ among areas in the building or between surface and compact layers. This average is similar to other reports on SBP manure (2,5), but higher than both solid dairy manure [0.55%, reference (8)] and dairy manure compost [0.85%, reference (11)]. Organic N did not differ with depth and averaged 0.95%. Average NH4-N concentration was lower in the surface layer (0.10%) than the compact layer (0.24%). 17 July 2009 Crop Management Fig. 1. Box plot of nutrient concentrations in manure packs from eight Minnesota SBP dairy barns. Each rectangle shows the 25th, 50th, and 75th percentiles of the data, while the whiskers show the 10th and 90th percentiles. Whiskers are missing when n < 9. Phosphate (P2O5) concentration in the pack averaged 0.27% in these barns and did not vary with manure layer or sampling location. This is equivalent to 5.4 lb P2O5 per ton and is greater than the "book" value of 3 lb/ton for solid dairy manure (7), but is similar to a more recent summary of solid dairy manure composition in the region (8). Dairy manure compost typically contains about twice this concentration of P2O5 (11), due to dry matter loss during composting. These SBP dairy manures contained 1.5 to 2 times the potash (K2O, about 13 lb/ton) as typical solid dairy manure (7,8), and about one-half as much K2O as typical dairy manure compost (11). Average K2O concentration was 0.72% in the surface layer of the pack and 0.64% in the compact layer, and did not differ by sampling location (data not shown). The lack of a defined spatial pattern in nutrient concentrations within SBP dairy manure packs indicates that these packs may be sampled randomly. The large differences among SBP dairy manures at application substantiate the need for chemical analysis on each farm (Table 2). In samples taken at discrete depths through the compact layer, however, concentrations of total and inorganic N changed with depth (data not shown). Consequently, the full depth of the pack must be sampled to obtain a representative analysis. Samples also could be taken from the face of the manure pack as it is being removed, or during field application. This last option will provide the best estimate of NH4-N for rapidly incorporated manure, because of likely ammonia losses during manure handling. The high pH of SBP dairy manure [average 8.6, reference (2)] exacerbates ammonia losses. Samples should be placed in tightly closed plastic bags and stored cold or frozen until analysis. We recommend that the final manure sample for analysis be composed of at least 10 samples from different areas of the barn or taken during application. 17 July 2009 Crop Management Table 2. Chemical characteristics and net NO3-N production pattern during incubation of SBP dairy manure sampled at application. x Net NO3-N production after 127 days of incubation. y Time lag following N immobilization before net NO3-N production equaled the amount in soil without manure. The range of C:N ratio for the entire pack depth in each barn (11.2 to 20.9) was similar to the range found by Janni et al. (5). At C:N ratios higher than about 15 in solid manure, net release of plant-available N is delayed (9). Based on these eight barns, it can be expected that N release from some SBP dairy manure would be limited or that N might be immobilized by soil microorganisms for the first several weeks of the growing season. Potential Net N Mineralization To test these concerns about N immobilization, three composite manure samples were generated for each building by combining ground subsamples from two or three sampling locations proportionally by depth (e.g., composite 1 combined locations 1, 2, and 3). Duplicate 10-g subsamples of each composite were mixed with 95 g loam soil in 120-mL containers, mixtures were brought to field capacity with tap water, and all mixtures were incubated with four nonamended soil samples at 95°F for 18 weeks. Soil moisture was determined gravimetrically several times during the incubation and water was added as needed to maintain the samples near field capacity. Three-gram subsamples were removed at intervals and extracted with 2 M KCl (about 10:1 v/w) and analyzed for NH4-N and NO2plus NO3-N (hereafter, NO3-N) by flow injection colorimetry (4,6). Net amounts of inorganic N were calculated after subtracting the inorganic N present in the non-amended soil. During the first two weeks of incubation, manure-derived NH4-N declined to near zero and thereafter remained below 0.04 lb N per ton in all cases. After 3 months of incubation, all manures produced more NO3-N than soil alone, but there were two different patterns of net NO3-N production (Fig. 2). Three manures maintained a positive amount of NO3-N (above that produced by soil alone) throughout the 4-month-long incubation. In contrast, five manures resulted in periods of NO3-N immobilization that in four cases (Farms 3, 4, 5, and 7) reduced NO3-N availability to the same as or less than that produced by non-amended soil. These four manures had C:N ratios of 19 to 21, whereas the others ranged from 11 to 13 C:N (Fig. 1). Farm Manure rate (ton/acre) Chemical characteristics (lb/ton) Net NO3-N mineralization Total N Organic N NH4-N P2O5 K2O Concentrationx (lb N per ton) Lagy (days) 1 6.2 32 27 5 5 15 2.9 0 2 16.3 24 22 3 9 19 4.1 0 3 21.0 14 11 3 4 15 1.8 31 4 17.0 11 6 5 4 12 1.0 50 5 13.4 18 15 3 6 11 0.9 63 6 9.3 25 20 5 4 10 2.0 0 7 14.4 16 10 6 4 14 0.7 47 8 8.4 29 23 6 5 10 3.1 0 17 July 2009 Crop Management Fig. 2. Net nitrate concentration in soil incubated under controlled conditions with manure from eight SBP dairy barns in Minnesota. Error bars indicate SE for Day 127. The four manures that resulted in net NO3-N concentrations near or below zero might be expected to cause temporary chlorosis in field crops that receive manure without additional fertilizer N. Even with immediate incorporation in the incubation trial, NO3-N production in the first week ranged from only 10 to 71% of the initial NH4-N. This low apparent nitrification efficiency was likely due to both ammonia losses and microbial immobilization, but this remains to be verified. Field Trials to Measure N Supply Fields selected for the experiment had soybean [Glycine max (L.) Merr.] as the previous crop, no other commercial fertilizer N application during the trial, no manure application in the previous three years, and no alfalfa (Medicago sativa L.) in the previous five years. Topsoil pH ranged from 6.1 to 8.2. Buildings were emptied in late October to early November 2006, the manure was applied to the field either immediately or piled for as long as 6 days. Each farmer applied manure to strips 60 ft long and 30 to 42 ft wide, depending on the width of the spreading pattern. Manure rates ranged from 6.2 to 21 tons/acre, with the goal of applying about 100 lb available N per acre, based on solid dairy manure guidelines (7). Samples of the manure were taken on the day of application and stored in closed containers in coolers with commercial ice packs until the samples could be frozen. As soon as was practical (2 to 48 h after application), one strip in each of three replicates was tilled with a disk, field cultivator, chisel plow, or disk ripper to incorporate the manure. The other manured strip and the nonmanured (control) strip were not tilled until the following spring. Fertilizer P and K were applied according to University of Minnesota recommendations (10) to alleviate potential deficiencies indentified in 7 of the 8 fields. Corn was planted by the farmers perpendicular to the manure strips. Fertilizer rate plots were 15 ft wide (6 rows of corn at 30-inch row spacing) by the width of the manure strip. After planting, urea was applied at 0, 30, 60, or 120 lb N per acre in the manured strips and at 0, 30, 90, and 180 lb N per acre in the nonmanured strips. An additional field trial was conducted at the University of Minnesota Southern Research and Outreach Center (SROC) near Waseca, using manure from Farm 6. Two manure rates (6.5 and 9.3 tons/acre) were applied on the same day as on Farm 6 and were either incorporated immediately or in spring. No N fertilizer was applied to manured plots; nonmanured plots received 0, 30, 60, 90, 120, or 150 lb N per acre. Grain yield was determined by hand-harvesting 40 ft of row, drying and shelling the ears, and weighing the grain. Grain yields were not collected at Farm 2, because fertilizer N was mistakenly sidedressed by the farmer to the 17 July 2009 Crop Management entire field. PROC REG in SAS was used to fit linear responses and PROC NLIN was used to fit the quadratic-plateau model to corn grain yield response to fertilizer N (Fig. 3). From these curves, the rate of N at which the marginal profit was maximized (Maximum Return To N, MRTN) was calculated, assuming a 0.15 ratio of fertilizer N cost ($/lb N) to corn grain value ($/bu). Fig. 3. Response of corn grain yield to fertilizer N at eight Minnesota locations, and estimates of the FNE for incorporated (Inc) and nonincorporated (Non) SBP dairy manure. FNE was estimated by shifting manure data along the x-axis to the fertilizer N rate at which they best fit the fertilizer-only regression equation [shifts indicated by arrows from original data (open symbols) for Farm 1]. Yields are expressed with 15.5% moisture. 17 July 2009 Crop Management During this study, growing season conditions were near normal at most farms, but Farms 4 and 6 were drier than normal in mid-season. Two farms (3 and 6) showed no response to fertilizer or manure N. SBP dairy manure did not enhance corn grain yield when adequate N was available (i.e., there was no significant yield difference for manure vs. no manure at high N rates). Three of the farms (1, 7, and 8) apparently did not reach maximum yield with fertilizer N alone, whereas corn on Farms 4 and 5 and the SROC attained yield plateaus with sufficient fertilizer N. The MRTN ranged from 0 to 153 lb N per acre and occurred at yields from 157 to 221 bu/acre (Table 3). Table 3. Corn grain yield response to fertilizer N and estimated FNE of SBP dairy manures. w MRTN = Maximum Return To N fertilizer, assuming a ratio of 0.15 for fertilizer N cost ($/lb N) to corn price ($/bu). x NA indicates that the value was not applicable or not estimable. y Italicized FNE values are the smallest solution to the least squares fit, are located at the beginning of the yield plateau, and may be underestimates. z SROC = Southern Research and Outreach Center, Waseca, MN. FNE was not affected by manure rate. Fertilizer N equivalence The FNE of SBP dairy manure was estimated by a least squares technique. Manure data were shifted along the x-axis (by adding or subtracting N from all four actual fertilizer N rates) until the difference was minimized between the yield response curve for nonmanured plots and the measured yields of the four N treatments within each manure management (see example in Fig. 3, Farm 1). We assumed no interaction between manure and fertilizer N rate. The sums of squares of the residuals were calculated at 1 lb N per acre increments along the fertilizer response curve, maintaining the original difference among N rates for the manured plots. For example, the sums of squares of the differences between measured yields and corresponding points on the fertilizer response curve were found for 0, 30, 60, and 120 lb N per acre, for 1, 31, 61, and 121 lb N per acre, and so on. For nonincorporated manure, the process also included negative increments of fertilizer N. These FNE estimates were then standardized to pounds of N per ton of manure applied. Based on our field trials, the actual FNE ranged from 1.4 to 12.1 lb N per ton for rapidly incorporated SBP dairy manure, and from -0.3 to 5.3 lb N per ton for fall-applied SBP dairy manure that was not incorporated until spring (Table 3). Using the MWPS guideline for solid dairy manure with SBP dairy manure resulted in FNE estimates varying from 5.5 to 11.3 lb N per ton for rapid incorporation and 1.5 to 6.8 lb N per ton for no incorporation until spring. Many of the measured FNE values were much smaller than the N supply predicted by the MWPS for solid dairy manure, resulting in overestimates of N supply ranging up to 86 lb/acre at the manure rates we used (Fig. 4). In Farm Corn response without manure FNE (lb N per ton) N rate at MRTNw (lb N per acre) Yield at MRTN (bu/acre) Incorporated Nonincorporated