1Semiarid Prairie Agricultural Research Centre, Swift Current; ²Semiarid Prairie Agricultural Research Centre, Indian Head; ³Saskatoon Research Centre, Scott; ⁴Saskatoon Research Centre, Melfort; ⁵Saskatchewan Centre for Soil Research, U of S

The atmospheric concentrations of nitrous oxide (N₂O), carbon dioxide (CO₂) and methane (CH₄) have been increasing rapidly. We are concerned about this because all three gases absorb long-wave radiation thereby increasing the amount of heat trapped in the earth’s atmosphere. In addition, N₂O is involved in the destruction of stratospheric ozone, which shields the earth from biologically harmful ultra-violet radiation. Chlorofluorocarbons (CFCs), used in refrigeration systems, have been the major contributors to ozone destruction. Their recent phase out has not resulted in as rapid a recovery of the ozone layer as had been anticipated. Simulation studies now suggest that the rapid increase in stratospheric N₂O is slowing the recovery, and could eventually begin to reverse some of the gains already made.

Canada has responded to these concerns by committing to a significant reduction in greenhouse gas emissions within the next decade. Current estimates suggest that agricultural activities are responsible for 10 to 15 % of Canada’s human-induced greenhouse gas emissions (Neitzert et al. 1999). While CO₂ is the major concern for most industries, the main greenhouse gas emitted by the agricultural sector is N₂O. Nitrous oxide is emitted directly and indirectly from crop and livestock production systems. Approximately 40% of the direct emissions from crop production systems are associated with the use of nitrogen (N) fertilizers. These estimates are calculated using a methodology derived by the Intergovernmental Panel on Climate Change (IPCC). The IPCC methodology assumes, as a default value, that 1.25 % of applied N is lost directly as N₂O (IPCC 1997). No accommodations are made for differing soil or environmental conditions, which are known to strongly influence N₂O emissions. In addition, the timing, placement and formulation of N fertilizer may have a significant influence on the amount of N₂O emitted. Nitrous oxide emissions from anhydrous ammonia (AA), for example, have been reported to be substantially higher than other N formulations (Eichner 1990). The proposed loss coefficients need to be tested for western Canadian soil and climatic conditions and modified if necessary.

An extensive research project has been undertaken in Saskatchewan to address a number of the major questions regarding N₂O loss from N fertilizer use. This presentation will discuss some preliminary results pertaining to four specific questions:

1. How much N from fertilizers is lost directly as N₂O under western Canadian conditions?
2. Does fertilizer N source (AA vs. urea) influence direct losses of N₂O?
3. Does application time (spring vs. fall) influence direct losses of N₂O?
4. Does fertilizer placement (side-row vs. mid-row band) influence direct losses of N₂O?

The data presented in this paper is taken from an ongoing study investigating the influence of timing, placement and form of N fertilizer on N₂O emissions. Briefly, this study includes field experiments at four locations, Swift Current, Scott, Indian Head and Star City. There are three test crops, canola, flax and wheat, at each site although only measurements made on the wheat crops are presented in this paper. Treatments that will be referred to here consist of urea and anhydrous ammonia (AA) banded in the fall or spring. Spring banded N was placed either to the side of the seed row (side-row band) or midway between alternate crop rows (mid-row band).

Nitrogen fertilizer was applied at 80 kg N ha^-1 at Star City and Indian Head, and 60 kg N ha^-1 at Swift Current and Scott. All sites received applications of P fertilizer (11-51-0), at rates of 17 kg P₂O₅ ha^-1 at Scott and Swift Current and 23 kg P₂O₅ ha^-1 at Star City and Indian Head. Phosphorus was placed with the seed except for the side-banded urea treatments where both N and P were placed in the side-row band. A blanket treatment of K₂SO₄ was broadcast prior to seeding at all sites to ensure sufficiency of these nutrients.

Trace gas samples were collected using vented soil chambers as described by Lessard et al. (1994). Sample collection and analysis followed accepted protocols similar to those described by Lemke et al. (1998). Annual estimates of N₂O loss are the sum of N₂O lost during the growing season plus the following spring thaw period.

Indian Head experienced unusually cool temperatures and above average precipitation during May and June of 2000, and somewhat drier than normal conditions for the balance of the season. Star City, Swift Current, and Scott experienced moderately dry conditions during early May, then above average precipitation from mid-May through to late July. Conditions at all sites were dry during the fall of 2000 and the spring of 2001. Overall, the growing season and fall of 2001 were very dry with most locations receiving 65 % or less of their mean long-term precipitation. Similarly, the spring of 2002 was very dry.

Previous studies in western Canada have shown that N fertilizer application generally increases N₂O emissions when considered on a "per acre" basis (e.g. Corre et al. 1999, Lemke et al. 1999). Preliminary results from our study confirm these findings, but in our study the differences were frequently not statistically significant. Nitrogen fertilizer increased N₂O emissions significantly only at one of four sites (Table 1) in the 2000/2001 season, and at two of three sites in the 2001/2002 season (Table 2). Our results also indicate very little difference in total N₂O emissions from the N management systems we compared. Emissions from plots receiving N as AA tended to be slightly higher than those receiving N as urea, however the differences were not statistically significant. Similarly, emissions tended to be higher from mid-row compared to side-row banded N, but statistically significant differences were noted in only two of the six site-years presented. Applying N in the fall seems to increase the risk of N₂O emissions during the following spring thaw, but on an annual basis N₂O emissions were generally not significantly affected. It should be noted that snow cover during the study period was limited at most sites, which in turn reduced the risk of N₂O production during spring thaw. As a consequence, we have limited confidence in the results of the spring vs. fall applied N comparison.

Table 1. Estimated annual N₂O loss from selected treatments at four sites in Saskatchewan for the period May 2000 to April 2001.