Effect of Nitrogen Source and Management on Ammonia Volatilization Losses from Flooded Rice—Soil System

Ammonia volatilization was studied by equipping capped greenhouse pots with a forced-draft system with external acid trap or by placement of open pots in a closed gas-lysuneter (allowing plant growth) with internal acid traps. In both systems air turbulence was optimized to simulate undisturbed open systems. Flooded soils were fertilized with approximately 50 or 100 kg N/ha of granular urea (GU), ammonium sulfate (AS), and two modified urea products—sulfur-coated urea (SCU) and urea supergranule (USG). The first three materials were broadcast and incorporated, whereas the last was placed at a depth of 8 cm. Ammonia volatilization from urea proceeded rapidly following hydrolysis of urea in the floodwater, leading to losses of up to 50% of the applied urea within 2 to 3 weeks. Ammonia loss from (NH4)2SO, occurred to a lesser extent due to a lack of alkalinity and occurred at a nearly constant rate, accumulating to ~15% loss in 3 weeks. Ammonia losses from the modified urea materials were negligible. Soil pH had little effect on the pH of the floodwater and, thus, on the ammonia volatilization process. However, ammonia volatilization losses were generally reduced by {actors that reduced the level of ammoniacal N in the floodwater, such as increasing soil CEC and reduced N application. Daily ammonia volatilization losses correlated well (r = 0.92) with the NH3(aq) concentration of the floodwater sampled between 1000 and 1100 hours each day. This observation holds promise for the development of a simple technique for assessing ammonia volatization losses from flooded soils based on simple physical and chemical parameters of the floodwater. Additional Index Words: fertilizer efficiency, gas lysimeter, algae immobilization. Vlek, P. L. G., and E. T. Craswell. 1979. Effect of nitrogen source and management on ammonia volatilization losses from flooded rice-soil systems. Soil Sci. Soc. Am. J. 43:352-358. R ECOVERY OF FERTILIZER N by the rice plant is notoriously low, particularly if applied early in the growing season (De Datta et al., 1969). This poor efficiency of utilization is generally attributed to the susceptibility of N to leaching, ammonia volatilization, and denitrification, particularly early in the season when plants are too small to compete efficiently with these loss mechanisms. The importance of ammonia volatilization as a loss mechanism, is poorly understood and generally underestimated, mainly as a result of inadequate measuring techniques used in the past (Vlek and Stumpe, 1978). Techniques to measure ammonia volatilization in a more representative way are presently being developed (Denmead et al., 1977; Lemon, 1977; Kissel et al., 1977). These techniques all attempt to prevent restrictions to free exchange of ammonia between the volatilizing surface and the atmosphere, which could otherwise critically reduce NH3 loss (Vlek and Stumpe, 1978; Kissel et al., 1977). In a recent attempt to assess ammonia volatilization losses under paddy field conlitions in unrestricted systems, Bouldin and Alimagno found losses of NH3 to be more extensive than reported previously. Some of the factors responsible for this discrepancy were studied by Vlek and Stumpe (1978). In this communication we report on two techniques to realistically estimate ammonia volatilization from flooded soils under greenhouse or growth chamber conditions in the absence or presence of rice plants. These techniques were used to assess the extent and rate of ammonia volatilization from several soils fertilized with granular urea, ammonium sulfate, sulfurcoated urea, or deep-placed supergranules of urea. MATERIALS AND METHODS Ammonia Volatilization Measurement Techniques Forced-Draft System—Ammonia volatilization from fallow, flooded soils was measured in the greenhouse by using covered pots which were swept with air, and the air then scrubbed of NH3 by an acid trap. Plastic greenhouse pots (7.5-liter, 25-cm diameter were fitted with clear plexiglass covers (0.6-cm thickContribution from the Agro-Economic Division of the International Fertilizer Development Center (IFDC), Muscle Shoals, AL 35660. Received 31 July 1978. Approved 11 Oct. 1978. "Soil Scientists, respectively. "D. R. Bouldin, and B. V. Alimagno. 1976. NH3 volatilization losses from IRRI paddies following broadcast applications of fertilizer nitrogen. Internal report, IRRI, Philippines. VLEK & CRASWELL: EFFECT OF N SOURCE AND MANAGEMENT ON AMMONIA VOLATILIZATION LOSSES 353 ness) which were clamped tightly to the rim of the pots. A flexible rubber gasket prevented air leakage under considerable pressure. The air inlet was a 13-mm diam. copper tube with holes in it which was shaped in a semicircle around the inside wall of the pot to create air movement across the water surface of the pot. Untreated house air was piped via a manifold and tubing to each of the copper tubes. A single 1.5-cm diameter air outlet pipe was installed in the cover opposite the air inlet. The inlet and outlet did not extend more than 2 cm below the cover, and the floodwater was maintained at between 3 and 4 cm from the cover. The acid traps were 80-cm lengths of PVC pipe (5.4-cm ID) sealed at the bottom and filled with 0.5JV H2SO4 to a depth of approximately 25 cm (500 ml). A coarse air dispenser of PVC pipe (90-cm long, 2.8-cm ID) was connected to the outlet pipes of the covered pots with tygon tubing. Scrubbed air was allowed to escape freely from the open tops of the acid traps. The system accommodated flows of up to 20 liters/rnin without air leaks or acid splashing. The acid trap effectively captured volatilized NH3 when checked with another trap in series. The air flow rate was calibrated to lose NH3 from 100 ppm N (NH4)2CO3 solutions at the same rate as from open pots set side by side to the covered systems in the ventilated greenhouse. This rate was determined to be approximately 15 liters/ min. Temperature of the floodwater was monitored by thermometers extending through the covers into the water. Gas-Lysimeter System—Ammonia volatilization from flooded soil planted with rice was measured in clear (plexiglass) gaslysimeters 120 by 340 by 360 cm in size, which could accommodate a 7.5-liter plastic greenhouse pot on a stand 8 cm above the lysimeter bottom. Next to the pot, a whispering electric fan with an 11-cm diameter blade provided upward movement of air by drawing it over a 1.5 by 15.5 by 25.0 cm acid trap placed under the pot and fan. A second fan, installed 17.5 cm above the top of the pot and 5 cm from the lysimeter wall, provided air movement across the water surface. A stainless steel cooling coil 3 m long controlled the temperature and humidity. Condensed water from the lysimeter atmosphere was directed into the floodwater or back to the acid trap by a flotation device which maintained a set water level in the pot, after passing through a Dowex 50 filled funnel to retrieve dissolved ammoniacal nitrogen (AN). A diagram of the gas-lysimeter is presented in Fig. 1. Six gas-lysimeters were built and installed in a growth chamber providing 12 hours of light (~760 /tEm^sec-) and 12 hours of darkness. Inside temperature and floodwater temperature were monitored by regular thermometers. Ammonia volatilization in the gas-lysimeter was calibrated against volatilization at approximately the same temperature in the open and was found to be the same when using 0.5 liters of 0.2N H2SO4 for an acid trap, and turning off the top fan during the night. Experimental Procedures Ammonia volatilization losses were primarily studied on two soils used in earlier fertilizer evaluation experiments (Craswell and Vlek, 1978). The Decatur silt loam is a Rhodic Paleudult— pH 5.0, total-N 0.14%, CEC 24.8 meq/100 g; Crowley silt loam is a typic Albaqualf-pH 6.2, total-N 0.13%, CEC 20.5 meq/100 g. In addition, a heavy textured soil—Sharkey clay, a Vertic Haplaquept with a pH of 5.8, total-N of 0.19, and CEC 62.3 meq/ 100 g, was included for comparison. Urease activities of these soils as determined by the method of May and Douglas (1976) were 52, 11.1, and 8.4 /jg NH,-N/g soil/hour for Crowley, Decautr, and Sharkey, respectively. The soils were fertilized with one of the following fertilizers: granular urea (GU), ammonium sulfate (AS), sulfur-coated urea-21% S (SCU), applied to soil with 1-cm standing water and incorporated to 5 cm, or urea supergranule—1 g (USG) placed at 8-cm soil depth. If fertilized with P and K, soils received 100 ppm of each element as Ca.(HsPOt):!-H.f) and K2SO4, 1 day before N application, and incorporated to a 10-cm soil depth. In a first series of two experiments with the forced-draft system, 30 greenhouse pots (7.5 liter) containing 7.5 kg of airdry Decatur or Crowley soil were flooded with demineralized water, puddled, and allowed to preincubate for 1 week with 1 cm of standing water. Soils were then fertilized with P and K and, subsequently, six pots each were treated with GU, AS, SCU, or USG, leaving four pots to serve as controls. Nitrogen was applied at a rate of 460 mg of N/pot or 95 kg of N/ha. Whisper Fan