«A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Crop, Soil, and Environmental Sciences ...»
Westerman, P.W., L.M. Safiey, and J.C. Barker. 1987. Available nitrogen in broiler and turkey litter. ASAE Paper no. 87-4053. ASAE, St. Joseph, MI.
Xu, X., Y. Huang, and L. Zhou. 2001. Effect of dicyandiamide and hydroquinone on the transformation of urea-nitrogen-15 in soil cropped to wheat. Biol. Fertil. Soils 34:286-290.
Fig. 6.2. Nitrogen mineralization from dry ingredients and commercial fertilizers over a 112 d incubation study on Dewitt silt loam.
142 80 70 60
Fig. 6.3. Influence of DCD on N-fortified PL and BS granular fertilizer mineralization for a 112 d incubation study with Dewitt silt loam. Averaged over biosolids and binding agent treatments.
Fig. 6.4. Ammonium-N and NO3-N mineralization from dry ingredients and commercial fertilizers over a 112 d incubation study with Dewitt silt loam.
Fig. 6.5. Ammonium-N and NO3-N mineralization from N-fortified PL and BS granular fertilizers over a 112 d incubation study with Dewitt silt loam. Averaged over biosolids and binding agent treatments.
Arkansas producers raise over 1.2 billion broilers per year resulting in 1.7 million Mg of poultry litter (PL) waste excreta that must be disposed of in an environmentally friendly manner. Poultry litter granulation gives the opportunity to transform a highly variable low analysis fertilizer product into a uniform material than can be fortified to increase nutritional value. We initiated a rice (Oryza sativa) field study to test N fertilizer recovery efficiency (FRE) and agronomic efficiency for flood irrigated crops. We tested fresh PL, granulated PL fortified with urea (PLU), PLU fortified with dicyandiamide (DCD) (PLUDCD), and preflood applied urea at 67, 112,157, and 202 kg N ha"1 in a randomized complete block design in a 4 (N source) x 4 ( N rate) factorial arrangement. A 0-N control was also included. Poultry litter treatments were preplant incorporated and compared to conventional fertility guidelines of urea applications made immediately before the permanent flood was established. Soil samples were collected at early heading to test for inorganic N, total N and total C. Rice aboveground biomass was quantified at early heading and used to calculate total N fertilizer uptake and grain yield was determined at maturity. On average, rice plants assimilated 16, 23, 53, and 89% of applied total N for PL, PLU, PLUDCD, and preflood urea and produced 14, 17, 29, and 47 kg rice kg N"1, respectively. Fresh PL, PLU and PLUDCD had a 29, 37, and 62% N agronomic efficiency calculated on a preflood urea N application basis (100%) when applied preplant incorporated to flood irrigated rice systems.
DCD, dicyandiamide; epm, evolutions per minute; FRE, fertilizer recovery efficiency;
LSD, least significant difference; PL, poultry litter; PLU, PL + urea; PLUDCD, PLU + DCD.
Agricultural production in Arkansas is responsible for over $4.9 billion worth of sales each year (National Agricultural Statistics Service, 2006). Thirty three percent of Arkansas' agricultural sales are derived from row crop agriculture with the balance derived from animal production. In the United States, Arkansas ranks first in rice production (538 900 ha"1), fifth in cotton production (348 200 ha"1) and has significant hectarage of soybean, wheat and corn (1 145 700, 331 900 and 247 000 ha"1;
respectively) (National Agricultural Statistics Service, 2006). Extensive row crop agriculture equates to inorganic fertilizer needs that exceed 974 800 Mg per year (Arkansas State Plant Board, 2005).
Arkansas producers grow approximately 1.2 billion broilers (Gallus gallus domesticus) annually, second in production to Georgia (National Agricultural Statistics Service, 2006). High broiler production results in over 1.7 million Mg of PL waste (excreta plus bedding material) that contains appreciable inorganic and organic nutrients (Chamblee and Todd, 2002). Historically, poultry production has been concentrated in Northwest Arkansas and PL was land applied to nearby pastures resulting in a net accumulation of soil nutrients, especially P (Slaton et al., 2004). Excessive soil test P concentrations were shown to increase runoff P concentrations and may accelerate eutrophication in sensitive waterways resulting in regulation of on-farm disposal in
2006). Moving surplus PL nutrients from Northwest Arkansas to the vast row crop production areas in eastern Arkansas is an ideal scenario; however, it is not economically feasible to transport PL over 40 km based on fertilizer value in the current form without industry or government subsidies (Bosch and Napit, 1992; Govindasamy and Cochran, 1995).
Unprocessed PL is a proven fertilizer source. Laboratory incubation studies by Bitzer and Sims (1988) and Hadas and coworkers (1983) estimated 500 to 666 g organic N kg total N"1 is mineralized during the growing season, but these results often differ from actual field studies. Cotton research by Sistani and coworkers (2004) reported adequate available N from preplant PL applied as supplemental inorganic N sources sidedressed to PL fertilized cotton crops did not increase yields over PL alone. Long-term PL evaluations in Alabama cited similar results with PL containing 92 to 113% of available N supplying power as ammonium nitrate on cotton and corn field studies (Mitchell and Tu, 2005). Pratt (1973) found similar N availability and proposed that 900 g organic N kg"1 was mineralized in the first year. However, other studies indicated lower efficiency of PL N compared to inorganic sources. For instance, Sims (1987) showed corn N efficiencies of 560 and 360 g fertilizer N kg total N"1 from ammonium nitrate and PL, respectively. Research in Arkansas on silt loam soils proposed that PL urea equivalents were 250 g N kg total N"1 in rice production systems (Golden et al., 2006). Poultry litter may not be as suitable for flood irrigated crops compared to upland crops since NO3-N will quickly undergo denitrification after the permanent flood is established.
Poultry litter can be chemically altered to increase its utility as a fertilizer. For
soluble P when amendments were added to PL during the poultry growing cycle.
Amendments effectively raised the N:P ratio and reduced potential P loss after field applications during runoff events. Another study added N during the pelletization process, effectively raising available N during an incubation and greenhouse study (Hamilton and Sims, 1995). Dicyandiamide, a nitrification inhibitor, was shown to reduce N loss in agronomic systems and increased plant N assimilation and yield (Cowie, 1918;
Norman et. al., 1988; Norman et. al. 1989; Reeves and Touchton, 1989). Norman and coworkers (1989) found that DCD increased rice plant N uptake of pre-plant applied urea-N by 28% compared to no DCD treatments. Nitrogen loss was decreased by reducing nitrification prior to flooding and thereby reducing denitrification.
The physical alteration of PL allows for production of a uniform product more acceptable to consumers than fresh PL. Agglomeration can efficiently alter organic products by combining powder constituents with liquid binding agents under centrifugal force to form spherical granules. Similarly, pelletization uses steam and pressure to combine powder products into cylindrical forms. However, PL pelletization was shown to increase N losses from denitrification compared to non-pelletized PL due to nitrification in the pellet and the resulting denitrification (Cabrera et al., 1993). We speculate that granulation versus pelletization may decrease microsites where nitrification occurs in PL products due to smaller fertilizer particle sizes. Cabrera and coworkers (1993) demonstrated that N volatilization and N mineralization rates were similar in both PL and pelletized litter sources. Golden and coworkers (2006) showed similar results with PL and pelletized PL having similar plant N uptake and rice yields.
fertilizers composed of PL and urea for rice production. A secondary objective was to evaluate DCD effectiveness when incorporated with PL products to see if N efficiency can be increased.
Research plots were established in 2004-2006 at the Rice Research and Extension Center near Stuttgart, AR (34°27'N; 91°33'W) to test N efficiency using PL and Nfortified PL granular fertilizers for flood irrigated rice. Plots were situated on a Dewitt silt loam (fine, smectitic, thermic Typic Albaqualfs) previously cropped with soybean [Glycine max (L.) Merr].
Fertilizer treatments included fresh PL, PLU, PLUDCD, and urea. Fresh PL was comprised of bedding material consisting of 50% rice hulls and 50% wood shavings, feces from 6 flocks of production broilers, and contained no litter treatment additives.
Poultry litter was ground to pass a 5.8 mm sieve, mixed with appropriate urea and/or DCD amounts in a rotary mixer, and granulated using a pin mixer (12D54L Pin Mixer, Mars Mineral, Mars, PA 16046). Exact agglomeration procedures, additive ratios and fertilizer descriptions were previously discussed and selected chemical properties are presented in Table 7.1. The PLU and PLUDCD granules were dried and sieved to pass a
4.75 mm sieve but remain on a 0.85 mm sieve.
Fresh poultry litter, PLU and PLUDCD treatments were applied to dry soil and incorporated using a rotary tiller in late spring (Table 7.2). 'Wells' rice was planted at 112 kg seed ha"1 immediately after fertilizer incorporation. In 2005 and 2006, preflood
10-cm deep flood was established within 24 h to reduce chances of N loss via NH3 volatilization (Norman et al., 1999) (Table 7.2). Urea treatments in 2004 were applied to moist soil due to above normal rainfall. All N sources were applied on a total N basis at 67, 112, 157, and 202 kg N ha"1. A no-fertilizer control was also included. Other rice production practices, including P and K blanket applications, were followed as outlined bySlaton(2001).
Aboveground rice plant samples from 1-m of row were taken at early heading for total N uptake (Guindo et al., 1994) (Table 7.2). Plant samples were oven dried at 55°C until a constant weight was achieved. Dried samples were weighed and ground to pass a 1-mm sieve and analyzed for total N by dry combustion using a LECO FP-428 (LECO Corp., St. Joseph, MI 49085) (Colombo and Giazzi, 1982). Total N uptake was calculated by multiplying aboveground biomass by the plant total N concentration. Grain yield was determined by harvesting the middle 7 rows from each plot with a plot combine (Table 7.2). Grain yields were adjusted to 135 g water kg rice grain" prior to statistical analysis (Yoshida, 1981; Fan et. al, 1998).
Soil samples were taken from all treatments at early heading to a 10-cm depth for total N, total C, and inorganic N determination and placed in air-tight bags (Table 7.2).
Soil samples were placed on ice in the field and frozen at -20°C within 6 h of sampling until inorganic N was determined (Nelson and Bremner, 1972). Inorganic N was extracted by adding 30 mL 2 MKC1 to 3 g dry equivalent soil and shaking for 1 h at 200 evolutions per minute (epm) (Mulvaney, 1996). The suspension was allowed to settle for
(Whatman pic, Middlesex, UK). Ammonium-N and NO3-N were determined using continuous flow analysis on a Sanplus System (SKALAR Inc., Norcross, GA 30071). Fifty g of moist soil was dried at 50°C until a constant weight was achieved for moisture determination (Gardner, 1986). Total C and N were analyzed using dried soil with combustion techniques on a Vario MAX CNS (Elementar Analysensysteme GmbH, Hanau, Germany) (Colombo and Giazzi, 1982).
The rice field experiment was arranged in a factorial arrangement of 4 (N sources) x 5 (N rates) using a randomized complete block design with 4 replications. The experiment was repeated over 3 years, effectively giving a split plot model with year serving as the main plot treatment and N treatments serving as sub-plots (Mcintosh, 1983). Inorganic N, total N and total C was analyzed using general linear model procedures (PROC GLM) with SAS v. 9.1 (SAS, 2003). Year and N treatments were analyzed as fixed effects while replication was a random effect. Least significant difference (LSD) was used to separate means (Gomez and Gomez, 1984). Rice grain yield and total rice aerial biomass N uptake was analyzed using simple linear and nonlinear (2-degree polynomial) regression procedures (PROC REG) using SAS v. 9.1 (SAS, 2003). Regression equations for the highest significant model were presented. Regression equations for each source were separated using contrast statements when similar degrees were present (linear to linear and quadratic to quadratic) (Gomez and Gomez, 1984).
Confidence intervals were used to compare linear to non-linear relationships at specific points of interest. Linear slopes were used to estimate plant response per unit of N (kg
derivative was used to locate peak yield and N uptake for non-linear models. For 2degree polynomial relationships, the N and N values were solved using the peak N rate, a confidence interval established and the lowest similar response was used to develop N agronomic efficiency. An alpha level of 0.10 was used for all statistical analysis and was chosen a priori.
Rice N uptake varied by year and each year is presented separately. Rice N uptake reacted in a linear response to all N sources in all years and N sources were compared according to their N FRE measured by slope from N uptake response models.
In 2004, preflood urea treatments had the lowest N uptake out of the 3 years (58% N FRE) (Fig. 7.1 and Table 7.3), but was similar to plant recovery values reported by Wilson and coworkers (1989) and Guindo and coworkers (1994) (63 and 62% N FRE, respectively). The PLU and PLUDCD granular fertilizers had similar N FRE as preflood applied urea (46 and 48% vs. 58%, respectively). Poultry litter had the lowest N FRE (22%) out of all N sources.