JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 97, NO. D13, PAGES 14,645-14,649, SEPTEMBER 20, 1992 Nitrous Oxide from Coal-Fired Power Plants: in the Plumes Experiments M. A. K. KHALIL AND R. A. RASMUSSEN Global Change Research Center, Dept. of Environmental Science and Engineering; Oregon Graduate Institute, Beaverton, Oregon Wereport studies of N20 concentrations in the ambient plumes of power plants. The results show that there isvery little excess N20 within 100m of thestack. There is some evidence that N20 is formed in theplumes farther downwind. Themagnitude of the effect cannot be quantitatively evaluated with the present data. Based onour results and published studies, weestimate that the global direct emissions of N20 from power plants are about 0.05 Tg/yr out of a total anthropogenic source of about 7 Tg/yr. 1. INTRODUCTION The concentration of nitrous oxide isincreasing. Because of its atmospheric characteristics, N20 is regarded asone of the main tracegases that can add to future globalwarming and contribute to stratospheric ozonedepletion. Nitrous oxideis produced by natural processes in soilsand oceans; perhaps most is releasedto the atmosphere. Until recently it was thought that coal-fired power plants were the major anthropogenic source causing much of the observed increase in atmospheric concentrations. Studies of power planteffluent, taken from within the stacks showed large amounts of N20 [Pierotti andRasmussen, 1976; Weiss and Craig,1976; tiao et al., 1987]. These findings were extrapolated to estimate that some 5 Tg/yr of N20 may come from power plants and perhaps 7 Tg/yr from all coalburning [Hao et al., 1987](1 Tg = 10 ;2grams). Such amounts were sufficient to explain the present trends and the entirechange of N20 concentration of about285 partsper billion by volume(ppbv) in preindustrial timesto the 310 ppbvobserved at present. Recently, Muzio andKramlich [1988]discovered that when mixtures of flue gases, suchas in power plant effluent, are stored,reactions occur in the presence of SO2, NO,o and HzO(v ) that can produce substantial amounts of N20 in any container. All pastwork on N20 from coalcombustion had relied on stored samples and thus became suspect. Subsequent research showed that indeed when stack samples are stored, N20 is formed, and when SO 2 and H20 are eliminated, there is very little N20 in the samples, implying that coal combustion may not be a significant global source [Linak et al., 1990;Sloanand Laird, 1990;Yokoyama et al., 1991]. In this paper we address twopoints.First, we reporta study in which samples werecollected in the plumes of a power plant rather than in the stacks. Processes that lead to the formation of N20 in containers may occur also in the plume. Second, we Copyright 1992 by the American Geophysical Union Paper number92JD00844. 0148-0227/92/923D-00844502.00 estimatethe limits of global N20 production from power plants based on the currentstudies. 2. THE EXPERIMENT IN TIlE PLUME Samples were collected on three flights of a Cessna-206 aircraftcapable of flyingat slowspeeds. The background samples were taken mostly upwind of the power plant, and samples in the plumewere takenat various short distances from the stack. We tried to collect the samples as close to the end of the stack as practicalso as to obtain measurable excesses of N20 concentrations. Thisrequired flying at slow speeds of about 80 km/hrandat lowaltitudes of 250-400 m. Each experiment consisted of simultaneous collection of two samples. Air was pumped (usingtwo Metal Bellows Corp. MB158 pumps) into 0.8 L, internally electropolished stainless steel containers. Air was split into two streams: one went directly to one sampling container and the other stream went through an Ascarite• trap before going to a second container. The Ascarite© trap (which is NaOH on a base) removes water vapor,SO2, NOx and COz, and possibly other tracespecies. Samples so collected are not susceptible to the artifact, discussed earlier, that leads to production of N20 in containers. In all, 30 such experiments were conducted constituting 60 samples. There were 16 experiments that intercepted the plume as evidenced by visualobservations, perceived turbulence, andmeasured CO 2concentrations. The remaining 14 generally were of ambientair, but in the final analysis only 8 were determined to be unequivocally background air around the powerplant. The experiment wasconducted aroundthe Centralia Coal- fired Power Plant (Centralia, Washington)close to our laboratory (September 4 and 18 and October 24, 1991). The Centralia planthastwo combustors (and two stacks) together capable of producing a maximum of 1280MW of power; it is one of the largest power plantson the West Coast of the United States. It uses locally minedsubbituminous coal. The nitrogen content of this coal is estimated to be between 0.8% and 1.8% [Smith,1911], whichis similar to mostcoalusedin the United States. 14,645