2009, Jung, M., Reichstein, M., Bondeau, A.

Global, spatially and temporally explicit estimates of carbon and water fluxes derived from empirical up-scaling eddy covariance measurements would constitute a new and possibly powerful data stream to study the variability of the global terrestrial carbon and water cycle. This paper introduces and validates a machine learning approach dedicated to the upscaling of observations from the current global network of eddy covariance towers (FLUXNET). We present a new model TRee Induction ALgorithm (TRIAL) that performs hierarchical stratification of the data set into units where particular multiple regressions for a target variable hold. We propose an ensemble approach (Evolving tRees with RandOm gRowth, ERROR) where the base learning algorithm is perturbed in order to gain a diverse sequence of different model trees which evolves over time. We evaluate the efficiency of the model tree ensemble (MTE) approach using an artificial data set derived from the Lund-Potsdam-Jena managed Land (LPJmL) biosphere model. We aim at reproducing global monthly gross primary production as simulated by LPJmL from 1998–2005 using only locations and months where high quality FLUXNET data exist for the training of the model trees. The model trees are trained with the LPJmL land cover and meteorological input data, climate data, and the fraction of absorbed photosynthetic active radiation simulated by LPJmL. Given that we know the "true result" in the form of global LPJmL simulations we can effectively study the performance of the MTE upscaling and associated problems of extrapolation capacity. We show that MTE is able to explain 92% of the variability of the global LPJmL GPP simulations. The mean spatial pattern and the seasonal variability of GPP that constitute the largest sources of variance are very well reproduced (96% and 94% of variance explained respectively) while the monthly interannual anomalies which occupy much less variance are less well matched (41% of variance explained). We demonstrate the substantially improved accuracy of MTE over individual model trees in particular for the monthly anomalies and for situations of extrapolation. We estimate that roughly one fifth of the domain is subject to extrapolation while MTE is still able to reproduce 73% of the LPJmL GPP variability here. This paper presents for the first time a benchmark for a global FLUXNET upscaling approach that will be employed in future studies. Although the real world FLUXNET upscaling is more complicated than for a noise free and reduced complexity biosphere model as presented here, our results show that an empirical upscaling from the current FLUXNET network with MTE is feasible and able to extract global patterns of carbon flux variability.

2009, Köppe, M., Hermann, M., Brenninkmeijer, C.A.M., Heintzenberg, J., Schlager, H., Schuck, T., Slemr, F., Sprung, D., van Velthoven, P.F.J., Wiedensohler, A., Zahn, A., Ziereis, H.

The origin of aerosol particles in the upper troposphere and lowermost stratosphere over the Eurasian continent was investigated by applying cluster analysis methods to in situ measured data. Number concentrations of submicrometer aerosol particles and trace gas mixing ratios derived by the CARIBIC (Civil Aircraft for Regular Investigation of the Atmosphere Based on an Instrument Container) measurement system on flights between Germany and South-East Asia were used for this analysis. Four cluster analysis methods were applied to a test data set and their capability of separating the data points into scientifically reasonable clusters was assessed. The best method was applied to seasonal data subsets for summer and winter resulting in five cluster or air mass types: stratosphere, tropopause, free troposphere, high clouds, and boundary layer influenced. Other source clusters, like aircraft emissions could not be resolved in the present data set with the used methods. While the cluster separation works satisfactory well for the summer data, in winter interpretation is more difficult, which is attributed to either different vertical transport pathways or different chemical lifetimes in both seasons. The geographical distribution of the clusters together with histograms for nucleation and Aitken mode particles within each cluster are presented. Aitken mode particle number concentrations show a clear vertical gradient with the lowest values in the lowermost stratosphere (750–2820 particles/cm3 STP, minimum of the two 25% – and maximum of the two 75%-percentiles of both seasons) and the highest values for the boundary-layer-influenced air (4290–22 760 particles/cm3 STP). Nucleation mode particles are also highest in the boundary-layer-influenced air (1260–29 500 particles/cm3 STP), but are lowest in the free troposphere (0–450 particles/cm3 STP). The given submicrometer particle number concentrations represent the first large-scale seasonal data sets for the upper troposphere and lowermost stratosphere over the Eurasian continent.

2009, Meier, J., Wehner, B., Massling, A., Birmili, W., Nowak, A., Gnauk, T., Brüggemann, E., Herrmann, H., Min, H., Wiedensohler, A.

The hygroscopic properties of atmospheric aerosols are highly relevant for the quantification of radiative effects in the atmosphere, but also of interest for the assessment of particle health effects upon inhalation. This article reports measurements of aerosol particle hygroscopicity in the highly polluted urban atmosphere of Beijing, China in January 2005. The meteorological conditions corresponded to a relatively cold and dry atmosphere. Three different methods were used: 1) A combination of Humidifying Differential Mobility Particle Sizer (H-DMPS) and Twin Differential Mobility Particle Sizer (TDMPS) measurements, 2) A Hygroscopic Tandem Differential Mobility Analyzer (H-TDMA), and 3) A simplistic solubility model fed by chemical particle composition determined from Micro Orifice Uniform Deposit Impactor (MOUDI) samples. From the H-DMPS and TDMPS particle number size distributions, a size-resolved descriptive hygroscopic growth factor (DHGF) was determined for the relative humidities (RH) 55%, 77% and 90%, and particle diameters between 30 and 400 nm. In Beijing, the highest DHGFs were observed for accumulation mode particles, 1.40 (±0.03) at 90% RH. DHGF decreased significantly with particle size, reaching 1.04 (±0.15) at 30 nm. H-TDMA data also suggest a decrease in growth factor towards the biggest particles investigated (350 nm), associated with an increasing fraction of nearly hydrophobic particles. The agreement between the H-DMPS/TDMPS and H-TDMA methods was satisfactory in the accumulation mode size range (100–400 nm). In the Aitken mode range (<100 nm), the H-DMPS/TDMPS method yielded growth factors lower by up to 0.1 at 90% RH. The application of the solubility model based on measured chemical composition clearly reproduced the size-dependent trend in hygroscopic particle growth observed by the other methods. In the case of aerosol dominated by inorganic ions, the composition-derived growth factors tended to agree (± 0.05) or underestimate (up to 0.1) the values measured by the other two methods. In the case of aerosol dominated by organics, the reverse was true, with an overestimation of up to 0.2. The results shed light on the experimental and methodological uncertainties that are still connected with the determination of hygroscopic growth factors.

2009, Reutter, P., Su, H., Trentmann, J., Simmel, M., Rose, D., Gunthe, S.S., Wernli, H., Andreae, M.O., Pöschl, U.

We have investigated the formation of cloud droplets under pyro-convective conditions using a cloud parcel model with detailed spectral microphysics and with the κ-Köhler model approach for efficient and realistic description of the cloud condensation nucleus (CCN) activity of aerosol particles. Assuming a typical biomass burning aerosol size distribution (accumulation mode centred at 120 nm), we have calculated initial cloud droplet number concentrations (NCD) for a wide range of updraft velocities (w=0.25–20 m s−1) and aerosol particle number concentrations (NCN=200–105 cm−3) at the cloud base. Depending on the ratio between updraft velocity and particle number concentration (w/NCN), we found three distinctly different regimes of CCN activation and cloud droplet formation: (1) An aerosol-limited regime that is characterized by high w/NCN ratios (>≈10−3 m s−1 cm3), high maximum values of water vapour supersaturation (Smax>≈0.5%), and high activated fractions of aerosol particles (NCN/NCN>≈90%). In this regime NCD is directly proportional to NCN and practically independent of w. (2) An updraft-limited regime that is characterized by low w/NCN ratios (<≈10−4 m s−1 cm3), low maximum values of water vapour supersaturation (Smax<≈0.2%), and low activated fractions of aerosol particles (NCD/NCN<≈20%). In this regime NCD is directly proportional to w and practically independent of NCN. (3) An aerosol- and updraft-sensitive regime (transitional regime), which is characterized by parameter values in between the two other regimes and covers most of the conditions relevant for pyro-convection. In this regime NCD depends non-linearly on both NCN and w. In sensitivity studies we have tested the influence of aerosol particle size distribution and hygroscopicity on NCD. Within the range of effective hygroscopicity parameters that is characteristic for continental atmospheric aerosols (κ≈0.05–0.6), we found that NCD depends rather weakly on the actual value of κ. A compensation of changes in κ and Smax leads to an effective buffering of NCD. Only for aerosols with very low hygroscopicity (κ<0.05) and also in the updraft-limited regime for aerosols with higher than average hygroscopicity (κ>0.3) did the relative sensitivities ∂lnNCD/∂lnκ≈ (ΔNCD/NCD)/(Δκ/κ) exceed values of ~0.2, indicating that a 50% difference in κ would change NCD by more than 10%. The influence of changing size distribution parameters was stronger than that of particle hygroscopicity. Nevertheless, similar regimes of CCN activation were observed in simulations with varying types of size distributions (polluted and pristine continental and marine aerosols with different proportions of nucleation, Aitken, accumulation, and coarse mode particles). In general, the different regimes can be discriminated with regard to the relative sensitivities of NCD against w and NCN (∂lnNCD/∂lnw and ∂lnNCD/∂lnNCN). We propose to separate the different regimes by relative sensitivity ratios, (∂lnNCD/∂lnw)/(∂lnNCD/∂lnNCN) of 4:1 and 1:4, respectively. The results of this and related studies suggest that the variability of initial cloud droplet number concentration in convective clouds is mostly dominated by the variability of updraft velocity and aerosol particle number concentration in the accumulation and Aitken mode. Coarse mode particles and the variability of particle composition and hygroscopicity appear to play major roles only at low supersaturation in the updraft-limited regime of CCN activation (Smax<0.2%).

2009, Lehsten, V., Tansey, K., Balzter, H., Thonicke, K., Spessa, A., Weber, U., Smith, B., Arneth, A.

We developed a technique for studying seasonal and interannual variation in pyrogenic carbon emissions from Africa using a modelling approach that scales burned area estimates from L3JRC, a map recently generated from remote sensing of burn scars instead of active fires. Carbon fluxes were calculated by the novel fire model SPITFIRE embedded within the dynamic vegetation model framework LPJ-GUESS, using daily climate input. For the time period from 2001 to 2005 an average area of 195.5±24×104 km2 was burned annually, releasing an average of 723±70 Tg C to the atmosphere; these estimates for the biomass burned are within the range of previously published estimates. Despite the fact that the majority of wildfires are ignited by humans, strong relationships between climatic conditions (particularly precipitation), net primary productivity and overall biomass burnt emerged. Our investigation of the relationships between burnt area and carbon emissions and their potential drivers available litter and precipitation revealed uni-modal responses to annual precipitation, with a maximum around 1000 mm for burned area and emissions, or 1200 mm for litter availability. Similar response patterns identified in savannahs worldwide point to precipitation as a chief determinant for short-term variation in fire regime. A considerable variability that cannot be explained by fire-precipitation relationships alone indicates the existence of additional factors that must be taken into account.

2009, Müller, C., Iinuma, Y., Karstensen, J., van Pinxteren, D., Lehmann, S., Gnauk, T., Herrmann, H.

Monomethylamine (MA), dimethylamine (DMA) and diethylamine (DEA) were detected at non-negligible concentrations in sub-micrometer particles at the Cap Verde Atmospheric Observatory (CVAO) located on the island of São Vicente in Cape Verde during algal blooms in 2007. The concentrations of these amines in five stage impactor samples ranged from 0–30 pg m−3 for MA, 130–360 pg m−3 for DMA and 5–110 pg m−3 for DEA during the spring bloom in May 2007 and 2–520 pg m−3 for MA, 100–1400 pg m−3 for DMA and 90–760 pg m−3 for DEA during an unexpected winter algal bloom in December 2007. Anomalously high Saharan dust deposition and intensive ocean layer deepening were found at the Atmospheric Observatory and the associated Ocean Observatory during algal bloom periods. The highest amine concentrations in fine particles (impactor stage 2, 0.14–0.42 μm) indicate that amines are likely taken up from the gas phase into the acidic sub-micrometer particles. The contribution of amines to the organic carbon (OC) content ranged from 0.2–2.5% C in the winter months, indicating the importance of this class of compounds to the carbon cycle in the marine environment. Furthermore, aliphatic amines originating from marine biological sources likely contribute significantly to the nitrogen content in the marine atmosphere. The average contribution of the amines to the detected nitrogen species in sub-micrometer particles can be non-negligible, especially in the winter months (0.1% N–1.5% N in the sum of nitrate, ammonium and amines). This indicates that these smaller aliphatic amines can be important for the carbon and the nitrogen cycles in the remote marine environment.

2009, Birmili, W., Alaviippola, B., Hinneburg, D., Knoth, O., Tuch, T., Borken-Kleefeld, J., Schacht, A.

Atmospheric particle number size distributions of airborne particles (diameter range 10–500 nm) were collected over ten weeks at three sites in the vicinity of the A100 urban motorway in Berlin, Germany. The A100 carries about 180 000 vehicles on a weekday. The roadside particle distributions showed a number maximum between 20 and 60 nm clearly related to the motorway emissions. The average total number concentration at roadside was 28 000 cm−3 with a total range of 1200–168 000 cm−3. At distances of 80 and 400 m from the motorway the concentrations decreased to mean levels of 11 000 and 9000 cm−3, respectively. An obstacle-resolving dispersion model was applied to simulate the 3-D flow field and traffic tracer transport in the urban environment around the motorway. By inverse modelling, vehicle emission factors were derived that are representative of a fleet with a relative share of 6% lorry-like vehicles, and driving at a speed of 80 km h−1. Three different calculation approaches were compared, which differ in the choice of the experimental winds driving the flow simulation. The average emission factor per vehicle was 2.1 (±0.2) · 1014 km−1 for particle number and 0.077 (±0.01) · 1014 cm3 km−1 for particle volume. Regression analysis suggested that lorry-like vehicles emit 123 (±28) times more particle number than passenger car-like vehicles, and lorry-like vehicles account for about 91% of particulate number emissions on weekdays. Our work highlights the increasing applicability of 3-D flow models in urban microscale environments and their usefulness for determining traffic emission factors.

2009, Tuch, T.M., Haudek, A., Müller, T., Nowak, A., Wex, H., Wiedensohler, A.

Sizes of aerosol particles depend on the relative humidity of their carrier gas. Most monitoring networks require therefore that the aerosol is dried to a relative humidity below 50% r.H. to ensure comparability of measurements at different sites. Commercially available aerosol dryers are often not suitable for this purpose at remote monitoring sites. Adsorption dryers need to be regenerated frequently and maintenance-free single column Nafion dryers are not designed for high aerosol flow rates. We therefore developed an automatic regenerating adsorption aerosol dryer with a design flow rate of 1 m3/h. Particle transmission efficiency of this dryer has been determined during a 3 week experiment. The lower 50% cut-off was found to be smaller than 3 nm at the design flow rate of the instrument. Measured transmission efficiencies are in good agreement with theoretical calculations. One dryer has been successfully deployed in the Amazon river basin. We present data from this monitoring site for the first 6 months of measurements (February 2008–August 2008). Apart from one unscheduled service, this dryer did not require any maintenance during this time period. The average relative humidity of the dried aerosol was 27.1+/−7.5% r.H. compared to an average ambient relative humidity of nearly 80% and temperatures around 30°C. This initial deployment demonstrated that these dryers are well suitable for continuous operation at remote monitoring sites under adverse ambient conditions.

2009, Wex, H., Petters, M.D., Carrico, C.M., Hallbauer, E., Massling, A., McMeeking, G.R., Poulain, L., Wu, Z., Kreidenweis, S.M., Stratmann, F.

Secondary Organic Aerosols (SOA) studied in previous laboratory experiments generally showed only slight hygroscopic growth, but a much better activity as a CCN (Cloud Condensation Nucleus) than indicated by the hygroscopic growth. This discrepancy was examined at LACIS (Leipzig Aerosol Cloud Interaction Simulator), using a portable generator that produced SOA particles from the ozonolysis of α-pinene, and adding butanol or butanol and water vapor during some of the experiments. The light scattering signal of dry SOA-particles was measured by the LACIS optical particle spectrometer and was used to derive a refractive index for SOA of 1.45. LACIS also measured the hygroscopic growth of SOA particles up to 99.6% relative humidity (RH), and a CCN counter was used to measure the particle activation. SOA-particles were CCN active with critical diameters of e.g. 100 nm and 55 nm at super-saturations of 0.4% and 1.1%, respectively. But only slight hygroscopic growth with hygroscopic growth factors ≤1.05 was observed at RH<98% RH. At RH>98%, the hygroscopic growth increased stronger than would be expected if a constant hygroscopicity parameter for the particle/droplet solution was assumed. An increase of the hygroscopicity parameter by a factor of 4–6 was observed in the RH-range from below 90% to 99.6%, and this increase continued for increasingly diluted particle solutions for activating particles. This explains an observation already made in the past: that the relation between critical super-saturation and dry diameter for activation is steeper than what would be expected for a constant value of the hygroscopicity. Combining measurements of hygroscopic growth and activation, it was found that the surface tension that has to be assumed to interpret the measurements consistently is greater than 55 mN/m, possibly close to that of pure water, depending on the different SOA-types produced, and therefore only in part accounts for the discrepancy between hygroscopic growth and CCN activity observed for SOA particles in the past.

2009, Kulmala, M., Asmi, A., Lappalainen, H.K., Carslaw, K.S., Pöschl, U., Baltensperger, U., Hov, Ø., Brenquier, J.-L., Pandis, S.N., Facchini, M.C., Hansson, H.-C., Wiedensohler, A., O'Dowd, C.D.

The European Aerosol Cloud Climate and Air Quality Interactions project EUCAARI is an EU Research Framework 6 integrated project focusing on understanding the interactions of climate and air pollution. EUCAARI works in an integrative and multidisciplinary way from nano- to global scale. EUCAARI brings together several leading European research groups, state-of-the-art infrastructure and some key scientists from third countries to investigate the role of aerosol on climate and air quality. Altogether 48 partners from 25 countries are participating in EUCAARI. During the first 16 months EUCAARI has built operational systems, e.g. established pan-European measurement network for Lagrangian studies and four stations in developing countries. Also an improved understanding of nanoscale processes (like nucleation) has been implemented in global models. Here we present the research methods, organisation, operations and first results of EUCAARI.