Measurements of coarse mode and primary biological aerosol transmission through a tropical forest canopy using a dual-channel fluorescence aerosol spectrometer

Abstract Aerosol in the size range 0.8–20 µ m was characterized according to optical equiva-lent diameter, D P , morphology and the presence of biological material, the latter de-termined by recording ﬂuorescence excited by ultraviolet light pulses at two diﬀerent 5 wavelengths. Single-particle measurements were performed within and subsequentlyabove a tropical rainforest in Borneo, Malaysia, in June and July 2008. In both loca-tions the aerosol number size distribution exhibited a primary biological aerosol (PBA)mode sized 2 µ m 2 µ m and possessed a wider morphological range than non-PBA.It also accounted for around 80% of the total number in the understory and 40% ofthe total number above canopy. Canopy transmission eﬃciencies for the total aerosolnumber and PBA are calculated to be 0.48±0.19 and 0.31±0.15 respectively, with theformer appearing to peak during the daytime because of a lack of PBA emission below


Introduction
A Primary Biological Aerosol (PBA) particle is defined as an aerosol particle that is discernibly all or part of a living organism.The classification includes pollen, animal and plant fragments, bacteria, viruses and fungal spores.The ubiquity of PBA, particularly fungal spores and pollen, in the atmosphere (e.g.Jaenicke and Matthias-Maser, 1993) has led to a renewed interest in aerobiology in the past 15 years.The fraction swept into cloud-forming regions of the atmosphere by updraughts can reside in the atmosphere for weeks and travel up to 1000 km.It is here that much of the current research interest is focussed, since pollen have been observed to initiate cloud droplet freezing at temperatures around 4 • C warmer than non-PBA in the laboratory (Diehl et al., 2001(Diehl et al., , 2002)).PBA number concentration in cloud-forming regions is ∼5 L −1 , which approximately matches that of ice nuclei (Matthias-Maser et al., 1999), yet their wider impact on cloud processes is not well quantified.Fungi thrive in dark, warm and damp environments with abundant soil nutrients or plant life.The lower strata of tropical forests are home to a high density of fungi (Lowman, 2004).Traditional biological sampling techniques used in and below tropical canopies (Gilbert, 2005) show the spore number concentration to be larger at the ground than in the upper canopy.The canopy itself can be regarded as a sink for vertically transported material, inhibiting its transfer to the free atmosphere.After nightfall, an upward heat flux typically occurs below the canopy as its top cools radiatively with warm air is trapped beneath (e.g.Kruijt et al., 2000).This suppresses the mixing of air through the canopy and therefore also the transport of aerosol.The net effect is that particles are channelled below the canopy until they reach a clearing large enough to allow their escape.
Tropical forests are estimated to cover 1090-1220 million hectares in total (Mayaux et al., 1998), representing a widespread PBA source.A first global estimate of the total fungal spore emission rate was placed at ∼50 Tg/yr by Elbert et al., 2007, so measuring the degree to which the canopy impedes PBA transport into the free atmosphere is a logical first step in bridging the gulf between the predicted and observed atmospheric PBA loadings.We present time-resolved measurements of a S.E.Asian tropical rainforest canopy's average particle transmission efficiency for PBA and total aerosol in an optical diameter range of 0.8-20 µm.Data was collected using a novel bio-aerosol spectrometer, the Wide Issue Bioaerosol Spectrometer, version 3 (WIBS-3), at the forest floor and above the canopy.Information about relative humidity (RH) and temperature (T ) at these positions is also included to illustrate the contrast between the wider area and the microclimate of the forest understory.(Marsh and Greer, 1992).The forest in Danum Valley is categorised as a tropical lowland evergreen rainforest and has an estimated average height of 35 m at the below-canopy measurement site.The valley receives an annual mean rainfall of 2778 mm (Fook, Malaysia Meteorological Dept., private communication, 2007) and April is the driest month (159 mm).Sunrise and sunset times were 06:00 and 18:15 LT respectively.
PBA and meteorological measurements were conducted in forest understory and above the canopy in two phases: 16 April-3 May and 19 June-23 July 2008, with 580 hours of data collected in total.Above-canopy WIBS-3 data was only collected in July, therefore this analysis and discussion is based solely on the second measurement phase in order to reduce uncertainty arising from month-to-month changes in conditions.The absolute vertical separation of the sites is ∼200 m.Data from the WIBS-3 was recorded for 14 full days below canopy and 2.2 days of continuous data were collected above canopy.Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version Interactive Discussion 3 Instrumentation used

Particle measurements
Coarse mode aerosol number concentration, morphology and fluorescence (used to indicate PBA) were sampled using the WIBS-3: a single particle, dual-wavelength UV fluorescence spectrometer, (Kaye et al., 2005).The instrument operates by exciting single particles sequentially, using two ultra-violet pulses at 280 nm and 370 nm, selected to excite primarily the co-enzyme NADH and the amino acid Tryptophan.Two detectors, tuned to the emission bands of Tryptophan and NADH (310-400 nm and 400-600 nm respectively), record the intensity of any fluorescence induced by the UV pulses.
Particles with optical equivalent diameter (D P ) 0.5-20 µm are sized using Mie scattering.Each particle passes through a laser beam and the elastically scattered light is focussed onto a quadrant PMT that samples its intensity at four points.The summed intensity is used to calculate optical size and the fractional standard deviation between the quadrants parameterises particle morphology; a low standard deviation indicates a more symmetrical scattering pattern, normally produced by a spherical particle.This morphological parameter is known as the Asymmetry Factor, AF, whose distribution is intrinsically log-normal.
The particle number concentration measured by WIBS-3 was confirmed in-situ by comparison against a commercial optical particle counter (GRIMM Dust Monitor Model 1.108) co-located with the WIBS-3 in the understory.

Meteorological measurements
Absolute temperature (T ) and relative humidity (RH) were measured in the understory using a warmed probe for improved accuracy in near condensing environments (Vaisala Model HMT 337 consisting of a PT100 RTD sensor, range −70 to +180 • C, accuracy ±0.2 • C at 20 RH, accuracy ±1.0% from 0 to 90% and ±1.7% from 90 to 100%).

Methods
It was assumed that PBA are the only fluorescent aerosol in the coarse mode and that to be recorded as such, their fluorescence must exceed an instrument-defined minimum threshold in either channel.The measured PBA concentration is therefore a subset of the total measured concentration.A similar protocol has been employed with other UV bioaerosol spectrometers (e.g.Agranovski and Ristovski, 2005).Random noise from UV pulse intensity fluctuations, component fluorescence and electrical noise is estimated to cause an over-estimate of PBA number that is 2-3% of the total measured number because of non-fluorescent particles being misclassified as fluorescent in either channel.
To calculate diurnal averaged time series, data of 5-min time resolution was separated according to the half-hour of the day in which it occurred and averaged across the whole dataset.Simultaneous monitoring above and below the canopy was not possible, so the canopy transfer efficiencies were calculated by taking the continuous above-canopy data and dividing it by a time series generated by replicating the diurnal average of the below-canopy data.The resulting time series of transfer efficiencies was diurnally averaged to illustrate its apparent dependence on time of day. a thermal insulator at night.The result is a highly consistent diurnal cycle of both variables.Above the canopy both exhibit a larger dynamic range and RH is usually lower because of less inhibited turbulent exchange outside the canopy.

Particle concentration below the canopy
A strong diurnal cycle of coarse aerosol number is observed in the understory.Fig- ure 3 shows the total WIBS-3 number concentration for 0.8 µm<D P <20 µm, N T W , for one characteristic week.PBA number, N PBA , clearly dominates the variation of N T W and initially peaks at 15:00 LT.A sustained series of bursts maintains a concentration of ∼1500-2000 L −1 following sunset, ceasing before sunrise and giving rise to concentrations ∼300 L −1 .Aerosol removal by rain washout appears to be mitigated by interception and channelling of rainwater by the canopy.
The average diurnal cycle of N T W is displayed in Fig. 4(i), which shows N T W to be consistently smaller in the daytime than it is between 15:00 LT and sunrise.Comparing the diurnally averaged N PBA number, printed in Fig. 4(ii), with the corresponding RH plot in Figure 1 (lower panel) suggests that PBA release and the recovery in RH both occur at 15:00 LT, signifying a possible link.The nature of the emission process is thought to be fungal spore release triggered by raised relative humidity and this will be discussed in detail in a forthcoming paper.

Particle concentration above canopy
Above the canopy, N PBA is reduced to around one-quarter of its mean below-canopy value, from 1200 L −1 to 300 L −1 , and there is no evidence of the understory's diurnal cycle occurring here.difference between the plots in Fig. 5 represents the non-biological aerosol component, N NON, which takes around 2 h to reassert itself compared to the rapid recovery in N PBA .

Comparison of total and PBA size distributions in each location
Mean number size distributions (d N/d logDp) were calculated for (i) all particles and (ii) the biological and non-biological components separately.The respective distributions are shown in Figure 6(i) and (ii), in above and below canopy pairs.Each of the three plots exhibits similar size modes above and below the canopy, albeit with different absolute concentrations.The mode at around 1 µm, present only in Fig. 6(i), is enhanced from 300 to 500 L −1 above the canopy and the 2 µm<D P <5 µm peak concentration reduces from ∼80 L −1 below canopy to ∼200 L −1 above it.The PBA and non-PBA distributions shown in Fig. 6 (ii) exclusively contain different size modes, suggesting the ∼1 µm mode is a different species and, because of its strengthened presence above the canopy, it may originate outside the forest.

Comparison of asymmetry factor distributions in each location
In order to compare the morphological properties of the two size modes, AF data was split into several size bands.This process was performed for above and belowcanopy data and their frequency distributions, each normalised to its own maximum, are plotted in Fig. 7(i) and (ii).In the understory distributions, the AF mode lies at 5 for D P ≤2 µm and for bigger particles a systematic shift to an AF mode at 11 takes place.Above canopy the same modes appear but the transition to the high-AF regime is not complete until the 4-6 µm size range is reached.The gradual transition takes place because the PBA and non-PBA size distributions overlap at 2<D P <4 µm but neither species dominates the total number in this size range above the canopy.Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version Interactive Discussion

Canopy transmission
Diurnal average canopy transmission efficiencies at 30-min time resolution were calculated using the WIBS data for (i) all particles sized 0.8-20 µm and (ii) those suspected to be PBA and these are shown in Fig. 8(i) and (ii).The error bars show standard deviation, a characteristic of the day-to-day variation in above-canopy number concentration, which as a proportion of the total concentration is far larger than below the canopy where a more consistent diurnal cycle occurs.Mean daytime, night-time and whole day transmission efficiencies for PBA, E PBA , and total number, E T W , are displayed in Table 1.
In both Fig. 8 (i) and (ii), there is a clear minimum at 14:00 and a maximum at 09:00 h, indicating the transition between the daytime and night-time regimes.The overall mean E T W exceeds the corresponding E PBA because of the reduced N PBA and enhanced N NON above the canopy.

Conclusions
The recorded AF, PBA and total number size spectra all contain the same features above and below the canopy, suggesting that the coarse aerosol sampled at each height is from the same source.Above the canopy the significantly reduced abundance of PBA, which dominates the total number of particles larger than 2 µm, is consistent with production below and transmission through the canopy with a finite efficiency.The 1 µm d N/d logD P mode has a negligible fluorescent component, likely attributable to it being a different species to the larger size mode rather than a lack of fluorescence sensitivity to smaller fluorescent particles.This conclusion is reinforced by the contrasting AF distributions at these sizes, which show modally lower AF than for larger particles.These findings are contrary to laboratory tests using the WIBS-3 using polystyrene latex (PSL) spheres.Modal AF actually rose by 2-3 (arbitrary units) when 3 µm PSL spheres were substituted with 1 µm PSL spheres, ruling out the possibility that instru-Introduction

Conclusions References
Tables Figures

Back Close
Full The above-canopy time series of concentration according to the WIBS-3 features nothing to suggest sustained periods of exchange at a particular time of day.The diurnal cycles of derived canopy transmission efficiency, printed in Fig. 8, are primarily influenced by the changes in concentration in the understory.What on first inspection appears to be a period of enhanced particle transmission around local midday is in fact the period during which the understory concentration is at its minimum, although enhanced transport could also conceivably occur at this time of day, when PBA emissions are at their weakest.The mid-afternoon minima in the transmission efficiency plots appear because of the initiation of the below-canopy PBA emission mechanism re-starting, thereby increasing the relative number below canopy.Since a channelling of aerosol beneath the canopy is expected, it is unclear whether the large differences between above-canopy concentrations during the two-day dataset are dictated by direct transmission through the canopy or the cumulative effect of forest clearings in the surrounding area.Introduction

Conclusions Introduction
Full

Figures 1
Figures 1 and 2 display the average diurnal cycles of RH and T above and below the canopy.Relatively small standard deviations (bars) in the below-canopy plots indicate the variation of T and RH is suppressed because the canopy shields the site from incoming solar radiation, prevents efficient exchange of moist and dry air and acts as 18970 Fig. 5 depicts the time series of N T W and N PBA .N PBA does not appear to exhibit any clear diurnal pattern above the canopy.Non-PBA concentration (N NON = N T W −N PBA ) is somewhat enhanced compared to the understory.Short-term fluctuations are driven by rain washout, as observed around 18:00 LT on July 19.
the tropical AF measurements.