# «University of California Division of Agriculture and Natural Resources Committee of Experts on Dairy Manure Management September 2003 February 2004, ...»

Nitrate leaching in cropland soils is the result of dynamic processes that involve organic N mineralization, plant uptake of N, irrigation scheduling and plant stresses due to water and/or N deficiencies. The dynamic interrelations between N inputs, N uptake, and N leaching can be captured in model simulations. Here we used the computer model ENVIRO-GRO (Pang and Letey, 2000) to illustrate nitrogen leaching losses in forage crops grown on manure fertilizer only. In the simulations, manure nutrients were applied at specified nutrient levels throughout the crop growing season at bi-weekly intervals.

Half of the applied N was assumed to be in organic nitrogen form, half was assumed to be in the ammonia form. Details of the simulations can be found in Appendix E, section 2.

If the N inputs from manure equal the 100% maximum N crop uptake potential, a 100% relative yield would only be achieved if leaching and other nitrate losses were eliminated.

In reality some leaching and other losses occur. Hence, the crop will be stressed, relative yields and evapotranspiration decline, crop N uptake will be lower, and even more nitrate will be available for leaching (Figure C-1). The N inputs exhibit the most significant impact on nitrate leaching. Leachable nitrate levels in the soil profile initially increase over time and approach a steady-state condition within approximately five years (Figures C-2 to C-4). The simulations demonstrate that 100% relative crop yields may be achieved only if the N input levels are significantly greater than the maximum crop N uptake potential (i.e., 117, 163, and 231% of the maximum crop N uptake potentials considered in the simulations).

When N input is at 117% of the maximum crop N uptake potential, the nitrate leaching potential of the soils is well controlled and approaches a steady-state condition within 2-3 years. At the low leaching fraction (i.e. LF = 0.07), N is expected to accumulate in the soil profile as nitrate leaching remains low and steady over time. Using a realistic range for leaching fractions (i.e. LF = 0.18 to 0.29), essentially all excess N should be expected to be mineralized and is potentially leachable. At the higher N input levels (163 and 231% of maximum crop N uptake potentials), the excess N is expected to be released as nitrate. The time required to reach the steady-state equilibrium is dependent on the leaching fractions. At the low leaching fraction (LF = 0.07), the nitrate leaching potential rises slowly and is expected to take a relatively long time to reach steady-state. However, the nitrate leaching potential will rise rapidly over time and reach steady-state conditions in approximately three simulated years.

On fields that have been receiving the same amount of manure annually for five or more years, the soil organic N level is likely to be in a steady-state. Nitrate is generated from the organic N pool at an approximately constant rate as illustrated in the model simulations. In such a situation, and if inorganic N applications are properly timed, the C-1 percent of nitrate leached is likely to closely track the irrigation leaching percentage.

Hanson et al. (1999) found that a leaching fraction of 15% to 30% is practically achievable in border check and furrow systems. Hence, we can expect that the nitrate leaching, under the best of field conditions, is approximately on the order of 10% to 15% of the N applied.

Some field studies have shown that in irrigated farming, the mass of nitrate leached beyond the crop root zone is related to the volume of water that percolates beyond the root zone. Rible et al. (1979) took deep soil cores at 58 sites in California under a wide range of cropped and non-cropped systems. Drainage volume at the sites was estimated to range from 1 to 41 inches per year. The authors determined an approximate relationship between mass of N leached and volume of deep percolation, with 3.4 lb nitrate-N ac-1 yr-1 (3.8 kg nitrate N ha-1 yr-1) leached for each inch (2.54 cm) of water draining below the root zone in the same time period. This corresponds to a concentration of nitrate-N in the drainage water of approximately 15 mg L-1. Rible et al. (1979) observed no significant relationship between the mass of nitrate-N leached and the amount of N applied to the land, probably due to the wide range of land use, vegetation, and management across the 58 sites. Drainage volume was a far more important controlling factor.

1.00 300 0.90

Figure C-1: Simulated relative yields (RYN), relative evapotranspiration (RET), and nitrate leaching (NL) at various irrigation amounts. N input is equal to 100% of the maximum crop N uptake potential.

Figure C-2: Simulated potential N leaching loss at four different irrigation regimes for year round forage production that receives N applications equivalent to 117% of the maximum crop N uptake of 600 kg ha-1 yr-1. The leaching fraction (LF) of the four irrigation regimes range from 7% to 43%.

Figure C-3: Simulated potential N leaching loss at four different irrigation regimes for year round forage production that receives N applications equivalent to 162% of the maximum crop N uptake of 600 kg ha-1 yr-1. The leaching fraction (LF) of the four irrigation regimes range from 7% to 43%.

Figure C-4: Simulated potential N leaching loss at four different irrigation regimes for year round forage production that receives N applications equivalent to 231% of the maximum crop N uptake of 600 kg ha-1 yr-1. The leaching fractions (LF) of the four irrigation regimes range from 7% to 43%.

1. Hanson, B., L. Schwankl, and A. Fulton. 1999. Scheduling irrigations: When and how much water to apply. Pub. 3396. Division of Agric. & Natural Resources, University of California. Oakland, CA.

2. Pang, X. P. and J. Letey. 1998. Development and Evaluation of ENVIRO-GRO,

**an integrated water, salinity, and nitrogen model. Soil Sci. Soc. Am. J. 62(5):**

1418-1427.

3. Rible, J.M., P.F. Pratt, L. J. Lund, and K.M. Hotzclaw. 1979. Nitrates in the unsaturated zone of freely drained fields. P. 297-320. In Pratt, P.F. Nitrate in effluents from irrigated lands. Final report to the National Science Foundation.

PB-300582, US Dept. if Commerce, National Technical Information Service, Springfield, VA.

C-6

**APPENDIX D:**

Mineralization – Conceptual Model and Field Studies Andrew Chang Perhaps the most significant difference between the use of commercial fertilizer and manure in nutrient management is the significant amount of organic nitrogen present in the manure. Organic nitrogen that is applied to soils during land application is not immediately available for plant uptake. Rather, it is subject to natural biochemical degradation processes that are generally referred to as “mineralization”. Ultimately the degradation processes yield nitrogen in inorganic form (ammonium). Only in the inorganic form is nitrogen available to plants. For purposes of nutrient management, the amount of organic nitrogen that ultimately mineralizes and the rate at which organic nitrogen is mineralized and, hence, available for plant uptake, is important. Here we provide an overview of the conceptual models for mineralization and a brief review of the literature on mineralization of organic nitrogen applied with manure.

Mathematical Expression. Salter and Green (1933) demonstrated that soil organic matter and nitrogen transformations could be described by first-order reaction kinetics.

**Stanford and Smith (1972) expressed the organic N mineralization with the equation:**

where Nm(t) (mg kg-1) is the mineralized N at time t (days), No (mg kg-1) is the mineralization potential (total mineralizeable organic N), and (d-1) is the first-order reaction N mineralization rate constant. Parameter estimation has been made by a nonlinear least squares method (Smith et al. 1986; Deans et al., 1986). Representative values for net mineralization rates and first-order rate constants for some ecosystems have been published by Smith and Paul (1990).

Deans et al. (1986) proposed two models for N mineralization; the first model took the

**mathematical form of Equation D–1 and the second used a double exponential equation:**

where S and (1-S) represent labile and recalcitrant organic N fractions decomposing at rates h and k, respectively. Their work indicated that the double exponential equation resulted in a better fit of experimental data on N mineralization. Cabrera and Kissel (1988) showed that the rate constants k and h and the apparent size of the two nitrogen pools, S and (1-S) were significantly affected by the length of the incubation time.

Paustain et al. (1992) modeled as many as five pools of decomposing organic matter, which they describe as plant residue broken into structural and metabolic pools, and soil organic matter (SOM) separated into three pools -- active, slow and passive. Others may have created as many as seven separate organic nitrogen pools to model the decomposition of SOM (Norton, 2000). With limited data points, the divisions of SOM D-1 and the associated nitrogen into multiple pools resulted in better regression models. The subdivisions reflected little reality in terms of the nature of SOM and many of the pools were not physically definable.

Pratt et al. (1973) presented an approach for calculating the annual rates of N mineralization expressed as a series of fractional proportions of any given application of manure, hereinafter referred to as a decay series. For fresh bovine manure with 3.5% N dry weight, they used a 0.75, 0.15, 0.10, 0.05 decay series, meaning that if 100 kg of organic N were applied in this form, 75% of the N would be mineralized the first year (75 kg); 15% of the remaining 25 kg would mineralize the second year (4 kg); 10% of the remaining 21 kg would mineralize the third year (2 kg); and so on. Subsequent releases will be negligible. The amount of N available in the first year is much larger because it includes the N already in mineral form, and organic N that is readily mineralizeable.

Conceptually, the decay series depicts a first-order reaction kinetics. Fitting this decay series into the mathematical form of Equation D-1 results in = 3.8 x 10-3 day-1 and an organic N mineralization half-life of 182 days for the manure application.

Pratt et al. (1973) determined that much less N would mineralize the first year if the manure had been exposed over time to the weather while deposited and dried outdoors before collection for disposal. Consequently, the total N contents of the dried and weathered dairy manure are considerably lower. Nitrogen decay series for these manure types vary with the total N contents and their calculated and half-life values are given in Table D-1.

N Transformations of Dairy Manure. We obtained data from field studies where dairy manure had been applied, crops grown, and the mineralization of N was either determined or could be estimated from crop yields such that k values for Eq. D-1 could be calculated. Studies which include all of the above factors are very rare. There seems to be abundant data for N mineralization in natural ecosystems, but studies where dairy manure was applied are limited.

Pratt et al. (1976ab) report a study where dairy manure and liquid steer manure were applied at various rates over a four-year period at the Moreno Field Station of the University of California, Riverside. The four-year field trial tracked the yields of Sudan grass and barley; quantities of N accumulated in the crop dry matter and in the soil organic matter; losses due to nitrate leaching; and annual amounts of N applied. Results indicated that 40 to 50% of applied N mineralized during the first year and 10 to 20% of

Talarczyk et al. (1996) report a 1992 to 1995 study of corn grown with dairy manure in Wisconsin. Fresh dairy manure tested for N, P, K, and sulfur was applied at a uniform rate of 35 tons per acre (78.6 mT ha-1) on separate plots in November of the previous year, January and March without incorporation until April, and on plots for a fourth treatment in April with incorporation 4 to12 days after application. The results of these treatments were compared with a check plot and plots that received pre-plant urea fertilizer at rates of 75, 125, and 175 pounds N per acre (84, 150, 196 kg ha-1). All plots received a starter fertilizer of 8-32-17 applied in a band at planting. Each treatment was replicated three times in randomized blocks.

As the November, January, and March applications remained on the soil surface for a long time prior to incorporation, there were opportunities for N to be lost prior to the incorporation. This would cause uncertainties in assessing the N transformations. There may also have been leaching losses that have not been accounted for. Assuming that the quantity of N available to the crop from the manure came from mineralization and that there were no losses of mineralized N via denitrification or leaching, we may calculate a lower bound for k, the first order reaction rate constant, accordingly. First, the yields of corn for grain using the four-urea application rates were plotted and equations obtained for yield vs. N application. Using these regression equations and the measured yield for the manure plots, the mineralized N was calculated. Since the total N in the 35 tons of manure was known and the mineralized N was estimated from the yields, Equation 1 could be applied to calculate k. The N release of the April manure application may be used to assess the N mineralization rate for the summer. The N releases of November and April manure applications could then be compared to determine the mineralization rate for the winter.

Although the authors state that no manure was applied to the plots for two years prior to the beginning of this experiment in 1992, they do not seem to account for N that might be mineralized in the second, third and fourth years of the experiment. It is understood that 35 tons per acre (78.6 mT ha-1) were applied each year. While the data was not perfect, it offered an opportunity to compare the N mineralization of cold (November to April) vs.