aircraft measurements of nitrogen and phosphorus in and around the lake tahoe basin: implications...

Aircraft Measurements of Nitrogenand Phosphorus in and around theLake Tahoe Basin: Implications forPossible Sources of AtmosphericPollutants to Lake Tahoe

Q I Z H A N G , † J O H N J . C A R R O L L , *A L A N J . D I X O N , A N D C O R T A N A S T A S I O *

Atmospheric Science Program, Department of Land, Air andWater Resources, University of California, One Shields Avenue,Davis, California 95616-8627

Atmospheric deposition of nitrogen (N) and phosphorus(P) into Lake Tahoe appears to have been a major factorresponsible for the shifting of the lake’s nutrient responsefrom N-limited to P-limited. To characterize atmosphericN and P in and around the Lake Tahoe Basin during summer,samples were collected using an instrumented aircraftflown over three locations: the Sierra Nevada foothills eastof Sacramento (“low-Sierra”), further east and higher inthe Sierra (“mid-Sierra”), and in the Tahoe Basin.Measurements were also made within the smoke plumedownwind of an intense forest fire just outside the TahoeBasin. Samples were collected using a denuder-filterpack sampling system (DFP) and analyzed for gaseousand water-soluble particle components including HNO3/NO3

-, NH3 /NH4+, organic N (ON), total N, SRP (soluble

reactive phosphate) and total P. The average total gaseousand particulate N concentrations (( 1σ) measured overthe low- and mid-Sierra were 660 (( 270) and 630 (( 350)nmol N/m3-air, respectively. Total airborne N concentrationsin the Tahoe samples were one-half to one-fifth of thesevalues. The forest fire plume had the highest concentrationof atmospheric N (860 nmol N/m3-air) and a greatercontribution of organic N (ON) to the total N compared tononsmoky conditions. Airborne P was rarely observedover the low- and mid-Sierra but was present at lowconcentrations over Lake Tahoe, with average (( 1σ)concentrations of 2.3 ( 2.9 and 2.8 ( 0.8 nmol P/m3-airunder typical clear air and slightly smoky air conditions,respectively. Phosphorus in the forest fire plume was presentat concentrations ∼10 times greater than over theTahoe Basin. P in these samples included both fine andcoarse particulate phosphate as well as unidentified, possiblyorganic, gaseous P species. Overall, our results suggestthat out-of-basin emissions could be significant sources ofnitrogen to Lake Tahoe during the summer and thatforest fires could be important sources of both N and P.

1. IntroductionLake Tahoe, an ultraoligotrophic lake with exceptionaltransparency, is located in a relatively small bowl-shapedbasin near the crest of the Sierra Nevada (39°N, 120°W,elevation 1898 m; Figure 1) (1). The lake is broad and deep(surface area ) 512 km2, mean depth ) 313 m) and has arelatively small watershed (area ) 812 km2). The surroundingN-limited forested watershed (2) and the granitic geology ofthe basin yield relatively small amounts of nutrients in runoffinto the lake (3). These characteristics, in combination withthe low ratio of watershed area to lake surface area (1.6), giveLake Tahoe extremely low natural productivity and very highclarity. These characteristics also make the lake very sensitiveto direct atmospheric deposition. Recent estimates concludethat over half of the annual loading of nitrogen (N) and 25%of the phosphorus (P) (4-6) come from direct atmosphericdeposition. Delivery of N from atmospheric deposition andother sources has increased the lake’s level of fixed nitrogenand shifted its nutrient response from N-limited to pre-dominantly P-limited (3, 6, 7).

Airborne N and P compounds deposited to Lake Tahoecould originate either within the basin or be transportedfrom sources located outside of the basin. During theafternoon and evening hours of the warm season (ap-proximately May-September), low altitude winds in CentralCalifornia tend to be westerly, flowing from the coast throughvarious gaps in the western coastal mountains, then into theCentral Valley and continuing east up the slopes of the SierraNevada (8, 9). This pattern occurs on 72% of the warm seasondays (10). At night, the flow in the mountains is reversed,and the predominant flows are downslope on the westernside of the Sierra Nevada. The upslope, daytime flow of airis impacted by various agricultural, urban and transportationsources of air pollutants, resulting in significant concentra-tions of primary and secondary air pollutants being trans-ported to the Sierra Nevada (11, 12). Although the morepolluted Central Valley air should be diluted during transportup the mountain slopes, it is likely that significant levels ofpollutants from these out-of-basin sources reach the LakeTahoe Basin and contribute to pollutant deposition to thelake. To examine this issue, we have characterized theconcentrations of airborne nitrogen and phosphorus com-pounds upwind of, and over, the lake. To minimize effectsof near-ground local sources and to obtain regionallyrepresentative samples, measurements and sampling wereperformed using an instrumented light aircraft.

2. Experimental Methods2.1. Sample Collection. 2.1.1. Sampling Equipment. Sampleswere collected using a Cessna 182 aircraft which continuouslyrecorded position, air speed, altitude, temperature, relativehumidity, approximate wind speed and direction, concen-trations of ozone, NO and NOy, and number concentrationsof particles (13). NOy was measured by a Thermo Environ-mental Instruments 42C analyzer and is defined operationallyas the amount of NO resulting from passing the ambientsample through a molybdenum catalyst at 325 °C. Gaseousand particulate N & P species were collected using a URGdenuder-filter pack system (hereafter DFP). The first portionof this system consisted of a fabricated isokinetic Teflonnozzle inlet followed by a cyclone separator to remove coarseparticles. The average DFP flow rate in this study was 21slpm (standard liter per minute), corresponding to a cyclonecut point of ∼4.5 µm; average flow rates for individual samplesranged from 18 to 33 slpm, corresponding to cut points of3.4-5.0 µm. Thus particles collected in the cyclone will be

* Corresponding author phone: (530)754-6095; fax: (530)752-1552;e-mail: [email protected] (J.J.C.) and [email protected](C.A.).

† Present address: Cooperative Institute for Research in Envi-ronmental Science (CIRES), 216 UCB, University of Colorado, Boulder,CO 80309-0216.

Environ. Sci. Technol. 2002, 36, 4981-4989

10.1021/es025658m CCC: $22.00 2002 American Chemical Society VOL. 36, NO. 23, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4981Published on Web 10/26/2002

considered “coarse”, while those collected on the Teflon filterwill be referred to as fine particles. Air was drawn throughthe DFP using an engine-driven vacuum pump, and the flowrate was measured between the end of the filter pack and thevacuum pump. Downstream of the cyclone were twoduplicate sets of denuder/filter packs (described below)arranged in parallel. Each set was activated separately,allowing two independent samples to be obtained per flight.The sampler intake was located outside the cabin well clearof the fuselage surface and engine exhaust.

Each DFP contained three denuders, coated to collect (inorder) gaseous HNO3, HNO2, and NH3. Prior to use, denuderswere coated either with 0.1% (w/v) NaCl in methanol (forHNO3 collection), 1% (w/v) Na2CO3 in 50:50 (v:v) methanol-H2O (for HNO2), or 1% (w/v) citric acid in methanol (forNH3). Denuders were prepared by adding 10 mL of the coatingsolution, shaking gently, pouring out the excess solution,and drying with purified air. To minimize contamination,denuders were prepared within 36 h of sampling. The filterpack of each DFP contained a Teflon filter (Zefluor, 2 µmpore size) to collect fine particles, followed by a Nylon filter(Nylasorb, 1 µm) and a citric acid impregnated Whatmanfilter to collect any HNO3 or NH3, respectively, that volatilizedfrom the upstream Teflon filter. All filters were 47 mmdiameter and were precleaned by repeatedly sonicating andshaking in Milli-Q water followed by rinsing with copiousMilli-Q. Whatman filters were prepared by soaking precleanedfilters in 50 mL of methanol solution containing 1.5% (w/v)citric acid and 1.5% (w/v) glycerin, decanting the solution,and drying in an ammonia-free vacuum desiccator. CoatedWhatman filters were kept individually in clean Petri dishes,sealed in clean plastic bags containing citric acid-coatedKimwipes, and stored in the dark at ∼4 °C for up to 3 weeks.After preparation, denuders and filter packs were cappedand kept in sealed plastic bags until immediately prior todeployment on the aircraft.

2.1.2. Sampling Times and Locations. Aircraft flight tracksduring DFP sample collection are shown in Figure 1 anddescribed in Table 1. Low- and mid-Sierra flights wereconducted on afternoons when the wind flow was predomi-nantly upslope. Lake Tahoe samples were collected twice

per day (one morning and one afternoon) for two consecutivedays during each sampling trip.

2.2. Sample Processing. As soon as possible after eachSierra flight, the intact DFP unit was brought to the laboratoryand the components were extracted immediately. For theTahoe Basin flights, samples from the first day were capped,sealed in plastic bags and stored in dry ice until they couldbe processed upon return to Davis the next evening.

Denuders were extracted with 6.0 mL of Milli-Q water(g18.2 MΩ-cm), while filters were extracted by shaking (3 hat ∼4 °C) in high-density polyethylene bottles containingMilli-Q. For the Teflon filters, one-half was wetted with 100µL of ethanol and extracted with 4.0 mL of Milli-Q water (forinorganic N and P analyses), while the other half (for organicN and P analyses) used the same procedure without ethanol(which interfered with the organic N analyses). Each filterhalf was extracted twice consecutively (see Section 2.4.).Nylon and Whatman filters were extracted with 10.0 mL ofMilli-Q. Filter extracts were not filtered since they containedno discernible particles. Significant amounts of coarseparticles were observed in the cyclone only for the forest fireplume samples. These particles were rinsed from the cycloneusing 2.0 mL of Milli-Q, and this solution was sonicated for60 min and filtered (0.22 µm Teflon) to remove insolubleparticles.

2.3. Sample Analysis. Concentrations of NH4+, NO3

-,NO2

-, SO42- and PO4

3- were analyzed using a Dionex DX-120Ion Chromatograph with conductivity detection. Organicnitrogen (ON) was determined as the difference in inorganicN concentrations in a given sample before and afteradjustment to pH ≈ 3 and illumination with 254 nm light (toconvert ON to inorganic forms) (14, 15). Since Teflon filterextracts were not filtered, reported concentrations of par-ticulate ON include both water-soluble and some portion ofthe less soluble species. The same UV-photooxidation methodwas used to measure total phosphorus (TP). However, sincethe photooxidation technique might liberate inorganicparticulate P as well as mineralize organic P, we refer to thedifference between TP and SRP values as “other phosphorus”(OP) rather than as organic P. Additional details of theanalytical procedures are given by Zhang and Anastasio (14).

FIGURE 1. Map of the study area. Dark solid lines represent flight paths for each sampling location. The approximate location and extentof the forest fire plume during sampling are also shown.

4982 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 23, 2002

Atmospheric concentrations of gaseous NH3, HNO2, andHNO3 were calculated based on the amounts of NH4

+, NO2-,

and NO3- collected on the citric acid, Na2CO3, and NaCl and

Na2CO3 coated denuders, respectively. Since a portion ofNO2

- might have been converted to NO3- on the Na2CO3

denuder during sampling (e.g., by reactions with O3) (16)reported values of HNO2 might be underestimated and HNO3

overestimated. Concentrations of gaseous organic N and TPwere calculated as the sum of the corresponding species onall three denuders. Because the denuder coatings used herelikely had low collection efficiencies for neutral or weaklyacidic/basic gases, reported concentrations of gaseousorganic N and TP might be underestimated. Atmosphericconcentrations of fine particulate species were calculatedbased on the concentrations on the Teflon filters. Concen-trations of N on the downstream Nylon and Whatman filterswere not significantly different from field blank values orwere below detection limits, indicating no apparent evapo-ration of particulate ammonium, nitrate, or organic nitrogen.

2.4. Quality Control. As a check on background con-centrations, we collected seven sets of field blanks (DFPassemblies deployed and treated in the same manner as realsamples except that no air was drawn through them). Averageconcentrations of N and P species on the field blanks weretypically ∼10% of the average sample values (Table 2). Limitsof quantification (LOQ) in samples were defined as threetimes the field blank value (14) or 2.5 times the methoddetection limit if the average blank concentration was belowthe detection limit (17). Concentrations of gaseous N species

reported in this study were usually well above the LOQ, andaverage concentrations of NO2

-, NO3-, NH4

+, and ON onsample denuders were at least 10 times higher than theaverage field blank (Table 2). SRP was never observed onblank (or sampled) denuders; when detected, OP in denudersamples was always above the LOQ. Analyte concentrationson Teflon filter field blanks were either below detection limit(SRP, TP and NO2

-) or were 10-20 times lower than averagesample values (NO3

-, NH4+ and organic N; Table 2).

Concentrations of all N and P species found on sampledTeflon filters were above LOQs.

Filter extraction efficiencies, determined from two suc-cessive extractions of a Teflon filter, were >95% and ∼90%for inorganic and organic N, respectively. Concentrations ofwater-soluble fine particulate ON were generally calculatedfrom amounts in the first extraction but also included thesecond extraction value if it was above the LOQ. Extractionefficiencies for P appeared to be ∼100% since concentrationsof phosphorus in the second extraction were always belowdetection limit. The photoconversion efficiency for organicnitrogen was monitored routinely using a mixture of stan-dards and was always quantitative, with recoveries rangingfrom 90 to 107%. The photoconversion efficiency for “other”P was not tested.

3. Results and Discussion3.1. Concentrations of Gaseous and Particulate Nitrogen.As shown in Table 1, we collected a total 20 samples from

TABLE 1. Sampling Flight Informationa

averages over sampling periodtime (PST)

date location start endsamplevol. (L)

air T(°C)

R.H.(%)

press.(mb)

NOynmol-N/m3

O3(ppbv)

N(>0.3)b

x106/m3N(>3.0)c

x104/m3

8-31-98 mid-Sierra 14:00 15:47 2263 18.2 21.0 785.9 205 78 14 42.45-24-01 low-Sierra 13:12 14:13 973 26.8 21.2 922.3 37 75 18 6.35-24-01 mid-Sierra 14:39 15:46 853 19.3 29.2 824.0 17 70 14 5.36-15-01 low-Sierra 12:40 13:43 1322 26.5 18.5 920.9 152 92 9 3.96-15-01 mid-Sierra 14:06 15:12 1169 19.8 24.8 822.4 68 87 8 3.17-05-01 low-Sierra 12:05 14.09 2880 29.0 24.4 917.8 263 87 19 35.07-17-01 Tahoe 9:13 11:09 2200 10.8 42.3 768.0 26 56 8 3.27-17-01 Tahoe 13:18 15:16 2424 12.8 43.4 769.2 55 60 11 3.17-18-01 Tahoe 8:49 10:48 2657 11.9 41.2 771.0 16 47 6 2.57-18-01 Tahoe 12:54 14:55 2292 14.9 39.2 770.3 48 55 6 3.38-02-01 Tahoe 8:45 10:45 2552 17.4 24.8 774.8 48 78 9 2.78-02-01 Tahoe 13:09 15:09 2687 19.9 23.7 773.6 51 71 9 2.88-03-01 Tahoe 8:31 10:31 2472 16.0 34.0 771.6 71 67 10 3.28-03-01 Tahoe 13:06 15:05 2487 17.2 33.2 770.3 73 71 10 3.48-15-01 Tahoe 8:48 10:45 2473 18.9 13.6 772.6 70 88 9 4.78-15-01 Tahoe 13:16 15:16 2312 22.2 15.4 771.9 79 80 9 3.68-16-01 Tahoe 8:26 10:26 1931 19.8 16.8 775.5 86 85 11 3.28-16-01 Tahoe 12:48 14:48 3960 23.2 13.7 772.5 85 84 14 7.69-05-01 forest fire 11:10 12:16 1871 11.0 24.9 724.5 298 69 55 32.19-05-01 forest fire 13:22 14:24 1447 11.5 22.3 724.7 270 71 57 28.9

a Data listed here were collected real-time (0.5 Hz) and were averaged over the period of each DFP collection period. The average mean elevations(above sea level) for sampling sites were: low-Sierra ∼ 400 m; mid-Sierra ∼ 770 m; Tahoe ∼ 1900 m; forest fire: ∼ 1900 m. The aircraft was flown300-350 m above ground level except in the forest fire plume (∼1000 m above the ground level). b Number of particles with diameter >0.3 µmas measured by an optical particle counter (OPC). c Number of particles with diameter >3.0 µm as measured by an OPC.

TABLE 2. Average Concentrations of Nitrogen and Phosphorus Species (( 1σ) on Field Blanks and the Average Sample/BlankRatios

field blank concentrationa average sample/blank ratiob

NO2- NO3

- NH4+ ONc SRP TP NO2

- NO3- NH4

+ ON SRP TP c

denuder 1.8 ( 1.1 10 ( 4.7 33 ( 31 14 ( 7.6 <0.6 3.5 ( 3.0 10 14 14 11 5.1filtersd <0.8 1.3 ( 1.3 1.9 ( 0.9 6.5 ( 5.7 <0.8 <0.8 18 20 11

a Units of nmol N per filter (or denuder) for NO2-, NO3

-, NH4+, and ON, and nmol P per filter (or denuder) for SRP and TP. Calculated using

0.5 times the method detection limit for any value below detection limit. b The ratio of the average concentration of a target species in the samplesover the average value of the corresponding species in the blanks. c Calculated using only sample values that were above detection limit. d Valueslisted only for Teflon filters since there appeared to be no significant evaporative loss of NH4NO3 or ON from Teflon filters (see Section 2.4).

VOL. 36, NO. 23, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4983

four separate locations (low- and mid-Sierra, Lake Tahoe,forest fire plume) during May-September. Because thenumber of samples from any location was relatively small,our data represent “snapshots” of N and P concentrationsand distributions rather than longer term mean values. Inaddition, because samples from different locations weregenerally not collected on the same day (with the exceptionof two of the low- and mid-Sierra pairs), it should be notedthat the comparisons made below between different locationsalso include variability because of temporal differences.

3.1.1. Low- and Mid-Sierra. Concentrations of totalnitrogen (TN) over the low- and mid-Sierra were among thehighest measured in this study, with average values (( 1σ)of 660 ( 270 and 630 ( 350 nmol N/m3-air, respectively(Figure 2). At both the low- and mid-Sierra transects, gaseouscompounds were much more important than fine particlesas sources of N, accounting for approximately 85% of thetotal measured N (Figure 2). NHx (i.e., NH3 + NH4

+) was thedominant component of the N budget with average con-centrations of 460 (( 220) nmol N/m3-air and 450 (( 280)nmol N/m3-air at the low- and mid-Sierra, respectively. Thesehigh concentrations of NHx are likely a result of agriculturalemissions from the Central Valley (18) as well as contributionsfrom forests and livestock at the lower elevations of themountains. As with TN and NHx, the average concentrationsof O3 and airborne particles also showed little differencebetween the two transects (Table 1).

In contrast, the average concentration of HNO3/NO3- over

the low-Sierra transect was roughly 50% higher than that atthe mid-Sierra (120 ( 5 nmol and 79 ( 20 nmol N/m3-air,respectively). Similarly, NOy concentrations along the low-Sierra transect were approximately 2.2 times higher than thevalues measured at the mid-Sierra transect on the same day(Table 1 and Figure A, Supporting Information). Conversely,the concentration of organic N was somewhat higher at themid-Sierra compared to the low-Sierra (Figure 2). The averageconcentration of gaseous ON over the mid-Sierra was roughly50% higher than at the low-Sierra, while the averageconcentration of fine particulate ON was roughly 30 nmolN/m3-air at both transects (Figure 2). The fine particulateON over the Sierra sites is comparable to the averageconcentration of water-soluble organic N in PM2.5 at Davis(15) approximately 65 km west of the low-Sierra transect.

Given the lengths of the sampling transects, horizontaldiffusion should have had only a minor dilution effectbetween the low- and mid-Sierra transects, and any signifi-cant dilution would be the result of increasing mixed layerdepth as the air moved up slope. However, the observedmixed layer depths varied by less than 100 m between thedaily paired low- and mid-Sierra transects. Most of thepollutants measured did not exhibit significant gradientsbetween these transects, indicating either that the air massesmoved up the slope to at least the mid-Sierra transect withminimal vertical diffusion or that sources along the flow pathoffset diffusional dilution effects. The former seems morelikely. However, it is interesting to note that the loss in HNO3/NO3

- between the low- and mid-Sierra is partially offset byan increase in organic N.

3.1.2. Tahoe Basin. On four of the sampling days, air inthe Tahoe Basin was clear and concentrations of atmosphericN were approximately 2 to 4.5 times lower than concentra-tions at the lower elevation Sierra sites to the west (Figure2). During 2 days in mid-August, however, the air in the TahoeBasin exhibited a blue haze (i.e., was slightly smoky) andaverage particle counts were about 28% higher for the >0.3µm and 75% higher for the >3.0 µm sizes than on the cleardays (Table 1 and Figure A, Supporting Information). Thehaze was distributed throughout the region and was pre-sumably due to distant forest fires. As shown in Figure 2, theaverage concentration of total N under these slightly smokyconditions was more than twice as high as under clearconditions (310 ( 57 and 140 ( 33 nmol N/m3-air, respec-tively). While average concentrations of NHx and HNO3/NO3

-

were essentially the same in clear and slightly smoky air, theconcentration of organic N was ∼10 times higher under theslightly smoky conditions. Both gaseous and particulateorganic nitrogen were enriched in the slightly smoky air:gaseous ON by a factor of ∼4 (63 ( 9 in slightly smoky airversus 15 ( 8 nmol N/m3-air in clear conditions) andparticulate ON by a factor of ∼25 (125 ( 45 versus 5 ( 3 nmolN/m3-air) (Figure 2).

3.1.3. Forest Fire Plume. The large increase in organicnitrogen under the slightly smoky conditions described abovesuggested that forest fires were a large source of atmosphericorganic nitrogen. We were able to test this hypothesis severalweeks later by sampling at the edge of a forest fire plume,

FIGURE 2. Average concentrations of nitrogen species in the gas phase (bottom, solid portion of each bar) and in fine particles (top, hatchedportion of each bar) at different locations or under different conditions. Each bar color represents a particular sample location/condition,as described in the legend. Error bars represent 1 standard deviation. ON, organic nitrogen; TN, total nitrogen.


∼35 km downwind of an intense wild fire (Figure 1). As shownin Figure 2, the average N concentration in the forest firesample was the highest measured in this study, 860 (( 120)nmol N/m3-air. Of course, given the strong spatial gradientswithin the smoke, concentrations of N and other specieswere likely much higher deeper within the plume.

Figure 3 illustrates the large differences between the Tahoeclear air, Tahoe slightly smoky air, and the forest fire sample.Concentrations of any given N compound were very similarthroughout the eight clear air samples, and the slightly smokyair samples had similar values with the notable exception ofthe increase in organic nitrogen. In contrast, the forest firesamples were enriched in all of the measured N species.Compared to the clear air conditions, the forest fire samplescontained, on average, approximately 5 times more NHx, 4times more HNO3/NO3

- and 10 times more organic N (Figures2 and 3). Interestingly, while concentrations of inorganic Nwere 4-5 times higher in the forest fire samples comparedto the slightly smoky Tahoe air, concentrations of ON wereapproximately equal in the two situations (Figures 2 and 3).One possible reason for this difference is that inorganic Ngases might have reacted with organic or elemental carbonparticles to produce particulate organic nitrogen in the moreaged, slightly smoky air; reactions such as these have beenreported in a few laboratory studies (19-21).

In the forest fire sample we also quantified nitrogen incoarse particles collected in the cyclone. Concentrations ofNO2

-, NO3-, NH4

+ and ON in the coarse particles were 4.9,18, 32 and 28 nmol N/m3-air, respectively, yielding a totalcoarse particulate N content of 83 nmol N/m3-air. Addingthis amount to the average amount of nitrogen recovered inthe gas phase and fine particles (780 nmol N/m3-air) yieldsa total concentration of nitrogen in the forest fire samplesof 860 nmol N/m3-air.

3.2. Nitrogen Distributions. 3.2.1. Sierra and Tahoe Basin.The distributions of atmospheric nitrogen species over thelow- and mid-Sierra were very similar: 1) NHx (primarily asNH3(g)) was the single largest component of the N budget,accounting for ∼70% of total N; 2) concentrations of organicnitrogen and HNO3/NO3

- were approximately equal, eachcontributing ∼15% of the total N; 3) HNO2/NO2

- accountedfor <1% of the total N, and 4) most (g66%) of any given Nspecies (or class) was present in the gas phase (Figure 4a,b).Note that gaseous organic N concentrations reported hereare lower bounds and might be significantly underestimated,since the denuder coatings employed likely have relativelylow collection efficiencies for both neutral gases as well asweakly acidic or basic gases (see Section 2.3.).

Interestingly, although absolute N concentrations weremuch lower, the pattern of nitrogen distribution during clear

air conditions in the Tahoe Basin was very similar to thatdescribed above for the more polluted low- and mid-Sierratransects (Figure 4c). Nitrogen speciation in the slightly smokyair in the Tahoe basin was quite different, however, and wasdominated by fine particulate organic nitrogen, whichaccounted for 41% of the TN pool (Figure 4d). The concen-tration of gaseous ON was also much higher in the slightlysmoky air, accounting for 21% of the total N budget, similarto the amount of gaseous NH3. Gaseous HNO3 and fineparticulate NO3

- and NH4+ accounted for the remaining

∼15% of the TN pool in these samples (Figure 4d). In addition,there were roughly equal amounts of fine particulate andgaseous N in the slightly smoky air at Tahoe, in contrast tothe low- and mid-Sierra and clear Tahoe conditions wheregaseous N forms (especially NH3) dominated.

In all of these cases the concentration of fine particulateNH4

+ was consistently higher than that of NO3-, often by a

factor of 2 (Figure 4). This “excess” ammonium is balancedby sulfate, which was generally present at molar concentra-tions similar to those of NO3

- (Table A, Supporting Informa-tion). The average mole ratios of NH4

+: NO3-: SO4

2- in fineparticles were 1:0.5:0.4 at the low-Sierra, 1:0.8:0.6 at the mid-Sierra, 1:0.5:0.6 in the Tahoe Basin under clear air conditions,and 1:0.4:0.6 at Tahoe when the air was slightly smoky. Thus,in addition to NH4NO3, ammonium in fine particles at thesesites is also bound in forms such as ammonium sulfate.

3.2.2. Forest Fire Plume. Most of the airborne N from theforest fire was in the gas phase while the least was in coarseparticles; the ratio of gaseous, fine and coarse particulate Nwas approximately 8:3:1 (Figure 5). As seen in all previouscases except the slightly smoky air at Tahoe, gaseous NH3

dominated the N pool in the forest fire samples, representing∼50% of the total N (Figure 5). Organic nitrogen was the nextmost abundant nitrogen class in the forest fire plume,accounting for ∼25% of the total airborne N (Figure 5). Thusthe relative abundance of organic N in the forest fire plumewas higher than that in clear air (13% of TN) but lower thanthat in the slightly smoky air (62% of TN) at the Tahoe Basin(Figures 4c,d and 5).

Unlike the situation in the Tahoe Basin and at the lowerelevations of the Sierra Nevada mountains, fine particles inthe forest fire plume contained considerably more NO3

- thanNH4

+ (Figure 5). The ratio of NH4+:NO3

-:SO42- in the forest

fire plume was approximately 1:2.6:1.3, suggesting that a largefraction of the fine particulate NO3

- and SO42- was balanced

by cations such as potassium and calcium, which arecommon in particles from biomass burning (22, 23). In coarseparticles, the ratio of NH4

+:NO3-:SO4

2- was approximately1:0.7:0.5, suggesting that NH4NO3 and (NH4)2SO4 were themajor inorganic N compounds (Figure 5).

FIGURE 3. Concentrations of N species in the gas phase (bottom, solid portion of bar) and fine particles (top, hatched portion) over theTahoe Basin and at the edge of a forest fire plume. Each bar represents a different N species or class, as described in the legend.


3.3. Measurements of Phosphorus. Atmospheric phos-phorus levels were always below detection limits duringsummertime flights over the lower and middle elevations ofthe Sierra (<0.8 nmol P/m-3-air). In contrast, phosphoruswas commonly found in samples collected over the TahoeBasin, with 67% of flights finding measurable levels. Theaverage (( 1σ) concentration of total phosphorus (TP) duringtypical clear air conditions in the Tahoe Basin was 2.3 ( 2.9

nmol P/m3-air, while that in the slightly smoky Tahoe airwas slightly higher at 2.8 ( 0.8 nmol P/m3-air. Phosphorusin the Tahoe air was always found as fine particulate SRP,except in one flight during the slightly smoky conditionswhere gaseous “other” P (i.e., not SRP) represented ∼30% ofthe TP. The average phosphorus concentration (gaseous +fine PM + coarse PM) at the edge of the forest fire plume (26( 8.7 nmol P/m3-air) was ∼10 times greater than the Tahoevalues (Figure 6). Furthermore, air from the forest fire plumewas most enriched in P relative to N. As shown in Figure 6,the average TN:TP ratio in the forest fire plume was 22 ( 11,while values in the Tahoe Basin were 42 ( 36 (clear air) and59 ( 10 (slightly smoky air). However, since coarse particleswere not collected in samples from over Lake Tahoe andfrom the low- and mid-Sierra transects, P concentrations atthese sites might be underestimated relative to the forest fireplume samples. In the forest fire air, the coarse particleswere much more enriched in P relative to N compared to thefine particles; TN:TP ratios were approximately 6:1 and 36:1,respectively.

The distribution of P in the forest fire samples was verydifferent from that of N: coarse particles accounted for ≈50%of total phosphorus, while the remaining P was splitapproximately evenly between fine particles and the gas phase(Figure 7). Average TP concentrations in the gas phase, fineparticles, and coarse particles were 5.7 ( 1.9, 6.7 ( 2.2, and13.9 nmol P/m3-air, respectively, leading to a roughly 1:1:2distribution of P. Surprisingly, SRP (i.e., phosphate that

FIGURE 4. Average N distributions in the gas phase (g) and fine particles (p) at the low- and mid-Sierra transects and within the TahoeBasin during clear and slightly smoky air. The average total N (TN) concentration (( 1σ) for each site/condition is listed in parentheses.

FIGURE 5. Average N distribution in the gas phase (g), fine particles(fine), and coarse particles (coarse) at the edge of a forest fireplume.


dissolved during our aqueous extraction) accounted for only≈40% of total P, while other formssincluding phosphorusspecies collected on the citric-acid coated denudersaccounted for ≈60% of TP (Figure 7). While we cannot ruleout the possibility that very small P-containing particles werecollected on the denuder, the denuder results suggest thatsignificant concentrations of gaseous P are emitted fromforest fires. While we do not know the identities of thesegas-phase phosphorus compounds, based on the denudercoatings we used the species were likely somewhat acidic orbasic. In addition, they were likely to be organic since theyhad high volatilities (i.e., were collected as gases). The “otherP” measured in the coarse particles might also includeorganophosphorus species as well as particulate phosphatethat did not dissolve during the initial aqueous extraction.

3.4. Comparison with Previous Measurements at LakeTahoe. It should be noted that our measured atmosphericconcentrations cannot by themselves be used to inferdepositional fluxes of atmospheric N and P to the lake surface.Calculating these deposition rates based on our measuredatmospheric concentrations would require numerical mod-eling with a good deal of (unknown) meteorological inputand was beyond the scope of this project. Alternatively,deposition rates could be estimated from empirical deposi-tion velocities determined in past studies. However, giventhe uncertainties in deposition velocities in general, and theirdubious applicability to concentration measurements ob-tained ∼300 m above the lake surface, we did not attemptthis estimation. Although we have not calculated depositionrates, our data do allow useful comparisons of the distribu-

tions of N and P species between atmospheric samples anddeposition samples.

For example, during our aircraft measurements in theTahoe Basin the ratio of NHx to total N(V) (i.e., HNO3 andNO3

-) was approximately 3:1 under both clear and slightlysmoky conditions (Figure 4c,d). This ratio is much higherthan the previously reported value of ≈1:1 for both wet anddry deposition collected by buckets placed on the surface ofthe lake (6) as well as in air measurements made at groundlevel in the Tahoe Basin (24). The apparent enrichment inN(V) at lake level compared to in air 300-350 m above thelake suggests either that atmospheric HNO3/NO3

- depositsmuch more quickly than NHx or that ground level emissionsof HNO3/NO3

- precursors (e.g., NOx from combustion) aremuch greater than those of NHx. Since HNO3 and NH3 havesimilar deposition velocities to water surfaces (25, 26), andvehicles emit much more NOx than NH3 (27) a vehicularground-level, in-basin source of NOx might explain some ofthe difference between aircraft and ground-based ratios ofNHx to N(V). Wood combustion is another possible reasonthat ground-based samples are more enriched in HNO3/NO3

-, although the N(V):NHx ratio in wood smoke (22) istypically smaller than from vehicles (27). Another possibilityis that the deposition samples collected by buckets locatedon the Lake surface (6) contained greater contributions fromvery coarse particles, which were not analyzed in our aircraftsamples, and that these particles contributed significantly tothe total N budget in deposition and were greatly enrichedin nitrate compared to ammonium.

While we measured considerable amounts of airborneorganic N over Lake Tahoe, the contribution of ON to the TNbudget varied widely, from ≈13% during clear air conditionsto ≈60% during slightly smoky conditions (Figure 4). Inprevious deposition measurements made at the lake surface,organic nitrogen was typically a larger component than inour aircraft measurements, accounting for ∼25% of the totalN in wet deposition and ∼75% in dry deposition (6). Thegreater contribution of ON in dry deposition might be dueto collection of very coarse particles enriched in organicnitrogen, such as soil particles, pollens or plant debris, thatwould not have been collected by the aircraft sampling train.In addition, aircraft measurements of atmospheric ONconcentrations might be underestimated since our samplingtechnique likely has low collection efficiencies for neutral orweakly acidic/basic ON compounds in the gas phase (seeSection 2.3.). While the identities of individual organic Ncompounds were not determined in this study, a previousstudy found that free and combined forms of amino acidsaccounted for approximately 20% of particulate ON in Davis,CA (15). In addition to these bioavailable species, the ONpool measured here likely contains a number of toxiccompounds such as nitroaromatics, nitrogen heterocyclesand nitrosoamines (21).

In terms of P, particulate water-soluble reactive phosphate(SRP) was almost always the only form found in the Tahoeair, while other forms of P (defined as OP ) TP - SRP)dominated in the forest fire plume, accounting for ≈60% ofthe total P (Figure 7). In contrast, atmospheric depositionmeasured at lake level contains both SRP and “other P” (with“other P” often dominating) during both smoky and non-smoky periods (UC Davis, Tahoe Research Group, unpub-lished data).

3.5. Implications for Possible Sources of AtmosphericN and P to the Tahoe Basin. Although our atmosphericconcentration measurements cannot be used to derivequantitative rates of N and P deposition to Lake Tahoe, theycan be used to qualitatively suggest some potential sourcesof these nutrients to the lake. For example, as shown in Figure2, there was a large gradient in concentrations of atmosphericN between the low- and mid-Sierra (where N concentrations

FIGURE 6. Concentrations of phosphorus (left plot) and the TN: TPratio (right plot) in the Tahoe Basin and at the edge of a forest fireplume. Tahoe Basin values include gaseous and fine particulatespecies, while the forest fire plume includes these as well ascoarse particulate species.

FIGURE 7. Distribution of P at the edge of the forest fire plume inthe gas phase (g), fine particles (fine), and coarse particles (coarse).SRP ) soluble reactive phosphate; OP ) “other” phosphorus.


are high) and the Lake Tahoe Basin (where N concentrationsare lower, especially under clear conditions). In addition,the N distribution at Tahoe under clear air conditions wasvery similar to those at the low- and mid-Sierra transects(Figure 4). These results suggest that emissions from theCentral Valley, and from lower elevations on the westernslopes of the Sierra, could be a significant out-of-basin sourceof N to Lake Tahoe, subject to control by regional and localmeteorological conditions. In the summer, the prevailingwesterly, upslope winds could transport appreciable amountsof N to the high Sierra and Lake Tahoe. Indeed, similar typesof transport have been demonstrated previously for airpollutants in the southern Sierra (6, 11, 12, 28-30). In thewinter, however, low boundary layer heights in the CentralValley, reduced westerly flows, and frequent inversions overthe Tahoe Basin likely make transport of material from theValley unimportant compared to in-basin sources. Transportof P from the Central Valley and Sierra foothills to Lake Tahoeappears to be generally negligible, even in the summer, giventhat phosphorus was never detected in the low- and mid-Sierra summer sampling transects.

In addition to the Central Valley as a potential source ofN, our measurements clearly indicate that forest fires, andperhaps biomass burning in general, could be significantsources of both nitrogen and phosphorus to downwindlocations. Indeed, during the period when the sampled forestfire was burning (Figure 1) the average bulk deposition ratesof total N and P to Lake Tahoe were ∼10 and ∼5 times higher,respectively, than the rates during nonsmoky periods in thepreceding or following month (UC Davis, Tahoe ResearchGroup, unpublished data). The link between forest fires andlake clarity has been studied before, most notably in a studyby Goldman and co-workers where algal production in thelake was found to increase significantly in response to firesin the watershed (31).

Past studies have shown that biomass combustion emitslarge amounts of NHx, HNO3/NO3

- and organic nitrogen intothe atmosphere as a result of the release of fuel nitrogen (22,23, 32-35). Furthermore, strong emissions of NH3 areassociated with smoldering biomass burning (36), a fact thatcoincides with our finding of high concentrations of NH3 inthe forest fire plume. As far as we are aware, there are nopublished reports of phosphorus emissions from forest firesand only very limited data for P emissions from woodcombustion. However, Schauer et al. (37) recently reportedthat P accounts for a minor but significant portion (0.004-0.007%) of the total fine particle mass emitted from fireplacecombustion of pine and oak. Our measurements clearlyindicate that forest fires can be large sources of both gaseousand particulate P to downwind locations. Our finding ofapparently nonphosphate forms of P in the gas phase andcoarse particles is surprising and suggests that organic andother P forms might be significant products of combustion.Sources of P to Lake Tahoe are especially important for lakeclarity and health since algal growth in the lake is currentlyprimarily P-limited (3, 6, 7). In addition to forest fires,residential combustion of wood in and around the TahoeBasin also might result in significant loadings of N and P tothe lake. Combustion during the winter months should beespecially effective at delivering nutrients to the lake sincethe basin in winter is often capped with inversions, whichtend to trap in-basin emissions. Since the ash from woodcombustion contains significant amounts of N and P (22, 38,39), disposal of this material is another possible combustion-derived source of nutrients to the lake.

AcknowledgmentsWe thank Ingrid George, Jeff Chan and Steve Zelinka forassistance with sample collection and analysis and Dr. JohnReuter, Scott Hackley, and Dr. Charles Goldman for providing

Lake Tahoe deposition data and relevant information. Thiswork was supported by the U.S. EPA (R819658 and R825433)Center for Ecological Health Research at the University ofCalifornia at Davis. Although the information in this docu-ment has been funded in part by the United States Envi-ronmental Protection Agency, it may not necessarily reflectthe views of the Agency and no official endorsement shouldbe inferred. Additional funding for this work was providedby a graduate fellowship from the University of CaliforniaToxic Substances Research and Teaching Program (Ecotoxi-cology component) and by a Jastro-Shields Graduate Re-search Award from the University of California at Davis.

Supporting Information AvailableTable A (concentrations of atmospheric gaseous and fineparticulate N species at different locations during summer)and Figure A (average concentrations of air pollutants atdifferent locations or under different conditions). Thismaterial is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited(1) Hatch, L. K.; Reuter, J. E.; Goldman, C. R. Can. J. Fish. Aquat.

Sci. 1999, 56, 2331-2339.(2) Goldman, C. R. Limnol. Oceanogr. 1988, 33, 1321-1333.(3) Jassby, A. D.; Goldman, C. R.; Reuter, J. E. Arch. Hydrobiol. 1995,

135, 1-21.(4) Reuter, J. E.; Miller, W. W. In The Lake Tahoe Watershed

Assessment; Murphy, D. D., Knopp, C. M., Eds.; USFS PSW-GTR-175. 1: Albany, CA, 2001; pp 215-399.

(5) Reuter, J. E.; Heyvaert, A. C.; Luck, M.; Hackley, S. H.; Dogrul,E. C.; Kavvas, M. L.; Askoy, H. Investigations of stormwatermonitoring, modeling and BMP effectiveness in the Lake TahoeBasin; John Muir Institute for the Environment, University ofCalifornia, Davis, 2001.

(6) Jassby, A. D.; Reuter, J. E.; Axler, R. P.; Goldman, C. R.; Hackley,S. H. Water Resour. Res. 1994, 30, 2207-2216.

(7) Jassby, A. D.; Goldman, C. R.; Reuter, J. E.; Richards, R. C. Limnol.Oceanogr. 1999, 44, 282-294.

(8) Zaremba, L. L.; Carroll, J. J. J. Appl. Meteor. 1999, 38, 1463-1473.

(9) Fosberg, M. A.; Schroeder, M. J. J. Appl. Meteor. 1966, 5, 573-589.

(10) Hayes, T. P.; Kinney, J. J.; Wheeler, N. J. California surface windclimatology; California Air Resources Board, Sacramento, CA,1984.

(11) Van Ooy, D. J.; Carroll, J. J. Atmos. Environ. 1995, 29, 1319-1330.

(12) Bytnerowicz, A.; Carroll, J. J.; Takemoto, B.; Miller, P. R.; Fenn,M. E.; Musselman, R. C. In Integrated Assessment of EcosystemHealth; Scow, K. M., Ed.; Lewis Publishers: Boca Raton FL, 2000;pp 93-118.

(13) Carroll, J. J.; Dixon, A. J. Aircraft Measurements of Meteorologicaland Air Pollution Profiles in the Lake Tahoe Basin; CaliforniaAir Resources Board, Sacramento, CA, 2000.

(14) Zhang, Q.; Anastasio, C. Atmos. Environ. 2001, 35, 5629-5643.(15) Zhang, Q.; Anastasio, C.; Jimenez-Cruz, M. J. Geophys. Res.-

Atmos. 2002, 107(D11), 10.1029/2001JD000870.(16) Allegrini, I.; De Santis, F.; Di Palo, V.; Febo, A.; Perrino, C.;

Possanzini, M. Sci. Total Environ. 1987, 67, 1-16.(17) Greenberg, A. E.; Eaton, A. D.; Clesceri, L. S.; American Public

Health Association; American Water Works Association; WaterPollution Control Federation, Standard Methods for the Ex-amination of Water and Wastewater: 1991 Supplement to the17th ed.; American Public Health Association: Washington, DC,1991.

(18) Coe, D. L.; Chinkin, L. R.; Loomis, C.; Wilkinson, J.; Zwicker, J.;Fitz, D.; Pankratz, D.; Ringler, E. Technical support study 15:Evaluation and improvement of methods for determiningammonia emissions in the San Joaquin Valley, final report STI-95310-1759-FR; Prepared for the California Air Resources Board,1998.

(19) Barletta, B.; Bolzacchini, E.; Meinardi, S.; Orlandi, M.; Rindone,B. Environ. Sci. Technol. 2000, 34, 2224-2230.

(20) Chang, S. G.; Novakov, T. Atmos. Environ. 1975, 9, 495-504.(21) Graedel, T. E.; Hawkins, D. T.; Claxton, L. D. Atmospheric

Chemical Compounds: Sources, Occurrence, and Bioassay;Academic Press: Orlando, 1986.


(22) Andreae, M. O.; Merlet, P. Global Biogeochem. Cycle 2001, 15,955-966.

(23) McDonald, J. D.; Zielinska, B.; Fujita, E. M.; Sagebiel, J. C.; Chow,J. C.; Watson, J. G. Environ. Sci. Technol. 2000, 34, 2080-2091.

(24) Tarnay, L.; Gertler, A. W.; Blank, R. R.; Taylor, G. E. Environ.Pollut. 2001, 113, 145-153.

(25) Lee, D. S.; Halliwell, C.; Garland, J. A.; Dollard, G. J.; Kingdon,R. D. Atmos. Environ. 1998, 32, 431-439.

(26) Hertel, O.; Christensen, J.; Runge, E. H.; Asman, W. A. H.;Berkowicz, R.; Hovmand, M. F.; Hov, O. Atmos. Environ. 1995,29, 1267-1290.

(27) Durbin, T. D.; Wilson, R. D.; Norbeck, J. M.; Miller, J. W.; Huai,T.; Rhee, S. H. Atmos. Environ. 2002, 36, 1475-1482.

(28) LeNoir, J. S.; McConnell, L. L.; Fellers, G. M.; Cahill, T. M.; Seiber,J. N. Environ. Toxicol. Chem. 1999, 18, 2715-2722.

(29) Datta, S.; McConnell, L. L.; Baker, J. E.; Lenoir, J.; Seiber, J. N.Environ. Sci. Technol. 1998, 32, 1378-1385.

(30) Zabik, J. M.; Seiber, J. N. J. Environ. Qual. 1993, 22, 80-90.(31) Goldman, C. R.; Jassby, A. D.; Amezaga, E. D. Verh. - Int. Ver.

Theor. Angew. Limnol. 1990, 24, 499-503.

(32) Hurst, D. F.; Griffith, D. W. T.; Cook, G. D. J. Geophys. Res.-Atmos. 1994, 99, 16441-16456.

(33) Lee, D. S.; Atkins, D. H. F. Geophys. Res. Lett. 1994, 21, 281-284.(34) Crutzen, P. J.; Andreae, M. O. Science 1990, 250, 1669-1678.(35) Lobert, J. M.; Scharffe, D. H.; Hao, W. M.; Crutzen, P. J. Nature

1990, 346, 552-554.(36) Griffith, D. W. T.; Mankin, W. G.; Coffey, M. T.; Ward, D. E.;

Riebau, A. In Global Biomass Burning: Atmospheric, Climatic,and Biospheric Implications; Levine, J. S., Ed.; MIT Press:Cambridge, Mass, 1991; p 569.

(37) Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. T.Environ. Sci. Technol. 2001, 35, 1716-1728.

(38) Erich, M. S. J. Environ. Qual. 1991, 20, 576-581.(39) Erich, M. S.; Ohno, T. Water, Air, Soil Pollut. 1992, 64, 475-485.

Received for review March 19, 2002. Revised manuscriptreceived August 16, 2002. Accepted September 11, 2002.

ES025658M


aircraft measurements of nitrogen and phosphorus in and around the lake tahoe basin: implications...

Documents