We describe results obtained during a summer-long campaign focussing on the particulate matter exchange between the atmosphere and a northern hardwood forest and its contribution to forest nitrogen deposition. Due to the large uncertainties around particle fluxes, our approach involves both direct measurements (of the particle chemical composition and of particle number fluxes) and a size-resolved mechanistic model to deduce nitrogen fluxes.
The site is located in Haliburton Forest and Wildlife Reserve in central Ontario (45.28Â°N, 78.55Â°W). Land morphology is relatively heterogeneous with an undulating topography on the granitic Canadian Shield. The average annual precipitation for the area is 1050 mm and the mean annual temperature is 5Â°C (Environment Canada). The site is an uneven-aged managed forest dominated by sugar maple (Acer saccharum Marsh), with American beech (Fagus grandifolia Ehrh.), eastern hemlock (Tsuga canadensis L.) and yellow birch (Betula alleghaniensis Britt.). The average projected leaf area index is ~ 6, while the canopy height and crown base height are estimated as 22 m and 8 m. Based on Raupach (1994) and Harman and Finnigan (2007), the displacement height has been estimated to be 16.5 m and z0 to be 1.1 m. Regional estimates of nitrogen deposition are among the highest in North America, and recent observations suggest that Haliburton Forest has transitioned to phosphorus limitation (Gradowski and Thomas, 2006, 2008).
Half-hourly fluxes of momentum, sensible heat and latent heat were measured using an open path system consisting of a CSAT3 sonic anemometer (Campbell Scientific), a LI-7500 infrared H2O gas analyzer (LI-COR), and an HMP45C temperature and humidity probe. Size-resolved particle number concentrations were measured using two high-time-resolution continuous particle sizing systems: a Ultra High Sensitivity Aerosol Spectrometer UHSAS (DMT), counting particles between 0.06 and 0.6 microns at 10Hz, and an Aerosol Particle Sizer APS (TSI), counting particles between 0.5 and 20 microns at 1Hz. These measurements were carried out for 57 days between Aug 16th and Oct 10th of 2011. Direct particle number fluxes have been calculated by eddy covariance and are compared with a detailed mechanistic model of deposition (Petroff et al., 2009). This model accounts for the particle characteristics (size distribution and density), as well as the forest morphology (vertical profile of the leaf area density, statistical distributions of the leaf size and orientation) and the meteorology (temperature, friction velocity and Monin-Obhukhov length). These parameters were measured on site.
To estimate the influence of particulate species on nitrogen deposition, a low-pressure Micro-Orifice Uniform Deposition Impactor (MSP) was used on 15 occasions throughout the summer and early fall to collect 24-hours integrated size-resolved (11 stages with cut points between 0.056 and 18 Âµm) particle samples later analysed by ion chromatography for water-soluble compounds. These detailed measurements of the size-resolved mass loadings of N-containing species allow us to deduce an average deposition flux to the forest. To do so, we rely on the model, described above, that is evaluated using the size-resolved fluxes of particle number.
Here we present preliminary results of the ion composition of the atmospheric aerosols and their relative contribution to the nitrogen deposition to the forest.
In this remote area, the mass loadings were low (total PM2.5 and ions PM1.8 respectively of 4.9 Âµg.m-3 and 1.5 Âµg.m-3 during the study) and the aerosol composition is characterized by an accumulation mode (between 0.1 and 2.5 Âµm) dominated by ammonium sulfate and a coarse mode (larger than 2.5 Âµm) by calcium nitrate and calcium sulfate (see Fig. 1 upper panel, left axis). In most cases, the accumulation mode was neutralized, while the coarse mode displayed an excess of cations, likely related to the inability of ion chromatography to quantify carbonate.
The model predictions of the deposition velocity (the ratio of flux and concentration) are low for such a broadleaf forest where the characteristic obstacle size is taken as 3 cm. It equals to less than about 1 mm.s-1 in the accumulation mode, in agreement with direct flux measures of Pryor et al., 2006, and re-analized by Petroff et al., 2009) and between some mm.s-1 and a few cm.s-1 for larger particles (see Fig. 1 upper panel, right axis).
Fig1. Upper panel: Median concentrations of the main ionic species, distributed per particle size, together with the deposition velocity predicted by the model. Lower panel: Median fluxes of the main ionic species, distributed per particle size.
With help of this model, we estimate mass fluxes of nitrogen species associated with the different size bins (see Fig. 1 lower panel). Ammonium represents about 80% of the nitrogen mass, mostly located in the accumulation mode (average 16 nmol.m-3), but only accounts for about 31% of the nitrogen deposition (average 0.87 Âµmol.m-2.d-1). Meanwhile, nitrate and nitrite represent about 10% of the nitrogen mass, mostly located in the coarse mode (average 2 nmol.m -3 each) but account respectively for 40 and 29% of the nitrogen deposition (1.12 and 0.80 Âµmol.m -2.d -1). The species contribution to the total nitrogen deposition is thus not only driven by its mass abundance but also by their presence in the efficiently deposited coarse mode.
These measurements highlight the need to account for the coarse mode contribution when evaluating the nitrogen load to terrestrial ecosystems. Relying on techniques that only account for the particles smaller than 1 micron (such as AMS) or 2.5 micron (bulk filter measurements of PM2.5) is likely to underestimate nitrogen inputs from particle deposition.