3506B24 Final Report
Linking Air Quality Observations to Prescribed Burning
Download 11.49 Mb.
|
- Navigate this page:
- CO vs NOy Regression at OLC Slope
- 0.745 Mar03 10.3
- 0.653 June03 16.9
- Table 26 and Figure 36
- Mil Base Surr Reg
- Jan-03 6,008 5,154
- Feb03 4,366 8,380
- Mar03 6,946 11,333
- Apr03 9,551 4,946
6.5 Linking Air Quality Observations to Prescribed Burning 6.5.1 Meteorology, Trace Gas and PM2.5 Mass Relationships at OLC Figures 33, 34 and 35 provide an overview of the temporal development of monthly averaged pollutants concentration patterns observed at the OLC site for the period December 2002 until June 2003. The averaged diurnal cycles of the major meteorological parameters and pollutants concentrations depicted in Figure 33 indicate the expected warming trend under increasing solar radiation. Average wind speeds show a clear diurnal variation with daytime maxima and nighttime minima indicating convective influences. Ozone shows typical diurnal profiles with daytime maxima however, the highest averages appeared during the month of April and not as expected in June. This may be in part a climatologic response to the abnormally high amounts of precipitation and cloud coverage that dominated the general weather conditions in the south-eastern United States most of spring and early summer of 2003 except for April (see Fig. 7), which is also reflected in the trend of the relative humidity averages and standard deviations of those (see vertical bars) in Figure 33 top panel. CO and NOy show a bimodal character with maxima in the mornings and evenings. Both compounds are products of incomplete combustion and are emitted from biomass burning but primarily also from automobile engines, hence, the morning peak is indicative of increased emissions during traffic rush hour. It is important to note here, that more than 15,000 people commute between Fort Benning and the cities of Columbus, GA and Phenix, AL every day [P. Gustafson, EMB Fort Benning, Personal Communication], and that OLC is only about 2 km away from the installation’s main post and the major roads leading to it. The rapid decrease to midday minima of CO and NOy is a combined effect of reduced emissions and increased atmospheric mixing, as the BL becomes deeper and convectively mixed by increased vertical turbulence. The average daytime maxima in wind speed were largest in January and February, when solar radiation was low, suggesting larger movements of weather fronts in systems of denser pressure gradients compared to later months. These conditions lead to lowest daytime minima in CO, NOy and PM2.5 concentrations. Figure 34 depicts wind rose plots for the wind direction frequency, wind speed, PM2.5, CO, and NOy, with the red traces representing daytime conditions (1100-1800 EST), and blue for “nighttime” (1800-1100 EST). Apparently at night, the measurement site is characterized by weak easterly component flow, which in turn seems to systematically switch to a stronger westerly component flow during the day. Weak south-easterly flow seems to carry pollutant emissions from prescribed burning at the Fort Benning military installation during the active burn period, especially from January to April, however, for June, when no burns were conducted on the Fort, no distinct increase of average pollutants concentrations is observed in air masses coming from SE. Considering the acres burned on the Fort each month between December 2002 and May 2003 with burn sites mainly E and NE from OLC except in May (see Table A2), progressing from roughly 1,500 acres in December to 6,000 (Jan), 4,500 (Feb), 7,000 (Mar), peaking at 9,500 in April, rapidly declining to ~500 acres in May, and no burns at all in June, the wind rose plots clearly show that the meteorological forces of horizontal advection and vertical turbulent mixing play an important role in shaping the ambient concentrations encountered at OLC. The sensitivity to meteorological conditions, i.e. the dilution of emissions and the influence of other potential sources becomes evident in April, when strong pollutant impact occurred from NNE directions, although most burns were conducted NE and SE from OLC, which, considering the wind roses’ scale in fact are the directions with the highest nighttime pollutant concentrations of all months. 
Figure 33: Monthly average diurnal patterns of major meteorological (top) and pollutant concentrations measured at the OLC site ~1 mile outside the Fort Benning, GA military installation, calculated for the months December 2002 through June 2003.
speed (black), air temperature (red), and photo- synthetically active radiation (PAR, pink).
orange), sum of nitrogen oxides (NOy, purple), nitrogen oxide (NO, green), and fine PM (PM2.5, black).  WD (%) WS (m/s) PM2.5 (μg/m3) CO (ppbv) NOy (ppbv)
Dec’02
Feb’03 Mar’03
May’03
CO versus NOy measured at OLC, for months December 2002 through June 2003 scaled in size by the NO/NOy ratio. Top: Level of correlation including standard errors of the slope and intercept; associated wind directions are scaled by color (right) with calm conditions in grey.
of day. The solid line represents the slope (8) and intercept (140 ppbv) of a typical urban environment with mainly mobile sources influence, serving for comparison [Parrish et al., 2002]. Figure 35 shows linear correlations of the 30 minute averaged CO versus NOy concentrations measured at OLC, for each month between December 2002 and June 2003. The symbols’ size is scaled by the level of NO/NOy, which is an indicator for the age of the air mass or the freshness of the emission. The slopes and intercepts are summarized in Table 25. The line with a slope of 8 and intercept of 140 ppbv, representing typical urban conditions where mobile source emissions dominate [Harley et al., 2001; Parrish et al., 2002], is plotted in the bottom panel of Figure 35 for comparison of the conditions encountered each month at OLC. While the CO/NOy emission ratios of biomass burning sources have been observed to range between 25 and 35 [derived e.g. from Lobert et al., 1991], the monthly averaged ratios at OLC seem to be clearly lower than that. However, in many instances, the OLC ratios are also clearly higher than what would indicate mobile source emissions, thus indicating a mostly mixed influence of sources, which is associated with the above mentioned finding, that the morning traffic rush hour coincides with a still shallow BL from the previous night, and predominantly (although weak) easterly component flow. Table 25: Correlation coefficients (R2), averages and standard errors of slopes and intercepts of linear regressions of CO vs. NOy 30-min data from OLC between December ‘02 and June ‘03.
A shallow nocturnal BL is especially formed under clear skies and effective radiative cooling of the ground at night. Smoldering can occur well into the next morning and sometimes even into the second day (e.g. as observed on 03/28, see VOC analysis above). Also, clear nights are potentially associated with weak catabatic flows along the Upatoi and Ochille Creeks that flow across the Fort Benning installation from the east into the Chattahoochee River at about 1 km to the south of OLC. Considering those possibilities, nighttime weak easterly flows can potentially carry pollutants emitted from still smoldering burn units to the OLC site where they “meet” and mix with the emissions from early morning rush hour traffic. Hence, based on 30 min averages, high CO/NOy ratios appear in the above scatter plots on an individual and more isolated basis. A detailed summary of the average meteorological and air quality conditions observed at OLC before, during and after each of the prescribed burns conducted at Fort Benning between February and May is given in Tables A6 of the Appendix section. It is evident, that the morning periods were not solely influenced by traffic rush hour emissions, indicated by individual highly correlated CO/NOy ratios ranging between 6 in February and 22 in May, with a general trend of larger ratios later in the season. The NO/NOy fraction of these air masses was typically high and NO was most significantly correlated with NOy. The highest NO fractions coincided with the lowest CO/NOy ratios during these morning traffic rush hour periods, indicating relatively fresh emissions from mobile sources. The significant increase of the CO/NOy regressions’ slopes with time progressing into early summer correlates with the increasing amount of acres burned, although, this relationship may also be an effect of the decreasing lifetime of NOy. Nitric acid (HNO3) is a major NOy compound, and is formed photo-chemically, therefore increasing its NOy fraction as time progresses into warmer seasons. Since HNO3 deposits rapidly to surfaces, NOy lifetime is significantly shorter e.g. in June than in February or March. Since no burns were conducted in the vicinity of OLC in June, i.e. on the near-by installation, the arriving air masses were likely photo-chemically aged and well-diluted, showing some of the smallest absolute CO and NOy concentrations. This allowed significant loss of NOy via HNO3 surface deposition, resulting in the highest CO/NOy ratio of 16.9 ±0.3 of all seven months. Prescribed burn emissions still have potentially contributed to this effect, since significant amounts of prescribed burns were conducted in Georgia, as shown in the following section. 6.5.2 Influences on Regional CO Background Due to the long-lived character of most tracers emitted by biomass burning, and due to the specific emission ratio of CO to NOy, the more regional effect of biomass burning emissions can be investigated by comparing the intercepts of CO and NOy regressions from different monitoring sites located in the greater region: Given continuing funding for the operation of the FAQS monitoring network, CO and NOy measurements will continue at OLC and near Griffin, GA, a site approx. 90 km to the NNE of OLC. It is important to note that biomass is not only burned at Fort Benning, instead, prescribed burning is practiced also by other organizations and private land owners in the surrounding region. The burning activity and intensity (in areas burned) is recorded by the Georgia Forestry Commission (GFC), who issues permits for those individual burns. The GFC collects and keeps these records as monthly totals for each county, allowing to compare the burn activities on Fort Benning with those in the surrounding counties of Chattahoochee, Harris, Marion, Muscogee, Quitman, Schley, Stewart, Talbot and Webster (see Table 26). Furthermore, we relate those activities with the above intercepts from the CO-NOy regressions from the OLC data, which are a reasonable representation of the average atmospheric CO background level of the greater region. Both Table 26 and Figure 36 below illustrate, that those activities seem to have a direct effect on regional background levels of CO, as well as CO/NOy ratios. Table 26: Prescribed burns (in acres) at Forts Benning and Gordon during fiscal year 2003 in comparison with burns/fires occurring at the surrounding regions and rest of Georgia. Note, that the counties of Chattahoochee, Harris, Marion, Muscogee, Quitman, Schley, Stewart, Talbot and Webster are considered surrounding region of Fort Benning, as are Richmond, Burke, Columbia, Glascock, Jefferson, Lincoln, McDuffie and Warren for Fort Gordon. The county level data was kindly provided by Mr. Daniel Chan from the Georgia Forestry Commission (GFC).
Figure 36: Monthly average CO background levels (± 1-σ standard errors) derived from CO/NOy regressions at OLC (left) compared with prescribed burn areas at Forts Benning and Gordon, their surrounding region according to Table 26 (with only 10% of “Rest of GA” acres plotted!). The atmospheric lifetime of CO is about 2 months in the free troposphere with its most important sink being the reaction with the hydroxyl radical (OH), while continental soil and the oceans contribute relatively little to the CO removal on a global scale. Background levels of CO are being monitored routinely by the global network of the NOAA Climate Monitoring and Diagnostics Laboratory (CMDL). Most of the CO background monitoring is conducted at remote islands or at coastal stations in order to minimize continental influence from anthropogenic emissions and biomass burning; therefore the CMDL data mainly characterize the marine BL over oceans. Only a few mountain observatories sample free tropospheric air over continents. Figure 37 compares trends of CO and OH background levels for the Northern Hemisphere, adapted from Novelli et al. [1998]. An important feature of atmospheric CO is its seasonal cycle, gradually accumulating during the dark period of the year with low OH concentrations between early fall and late spring, and being rapidly consumed from April to June due to intensive OH removal. The average OH cycle depicted in the figure is the result from a three-dimensional chemical tracer model [Spivakovsky et al., 1990], confirming its strong influence on background CO. The year-to-year variability in the [CO] trend reflects a corresponding inter-annual variability in [OH] (not depicted). 
Download 11.49 Mb. Share with your friends:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
