CO2, CH4 and Particles Flux Measurements in Florence, Italy

2 downloads 87 Views 967KB Size Report
Eddy covariance (EC) flux measurements of CO2, CH4 and particles in densely urbanized area in. Florence are reported, and partitioned into emission ...
Available online at www.sciencedirect.com

ScienceDirect Energy Procedia 40 (2013) 537 – 544

European Geosciences Union General Assembly 2013, EGU Division Energy, Resources & the Environment, ERE

CO2, CH4 and particles flux measurements in Florence, Italy Beniamino Giolia,*, Piero Toscanoa, Alessandro Zaldeia, Gerardo Fratinib, Franco Migliettaa b

a CNR IBIMET, Via Caproni 8, 50145 Firenze, Italy LI-COR Biosciences GmbH, Siemensstr. 25A, 61352 Bad Homburg, Germany

Abstract

Eddy covariance (EC) flux measurements of CO2, CH4 and particles in densely urbanized area in Florence are reported, and partitioned into emission categories. CO2 fluxes are a net source to the atmosphere, with small inter-annual variability and high seasonality. CH4 fluxes represent about 8% of CO2-equivalent emissions, and do not exhibit significant seasonality. Heating and road traffic account for 68% and 32% of observed CO2 emissions, respectively, and for 14% of CH4 emissions, while gas network leakages are responsible for the residual. Particle fluxes exhibit a pronounced weekend decrease, highlighting that the main contribution to emissions comes from road traffic. © 2013The TheAuthors. Authors. Published by Elsevier © 2013 Published by Elsevier Ltd. Ltd. Selection and peer-review under responsibility of the GFZof German Research CentreResearch for Geosciences Selection and/or peer-review under responsibility the GFZ German Centre for Geosciences Keywords: eddy covariance; CH4 flux measurements; particles flux measurements;

1. Introduction Direct measurements of surface emissions with the eddy covariance (EC) micrometeorological technique might provide important information on spatial and temporal variability of surface fluxes, as demonstrated by the increasing number of measurement campaigns in urban and suburban areas in the recent years (Grimmond et al. [1], Velasco et al. [2], Matese et al. [3]). Since eddy covariance requires fast response gas analyzers to measure atmospheric concentrations, only carbon dioxide (CO2), water

* Corresponding author. Tel.: +39-055-3033750; fax: +39-055-308910. E-mail address: [email protected].

1876-6102 © 2013 The Authors. Published by Elsevier Ltd.

Selection and peer-review under responsibility of the GFZ German Research Centre for Geosciences doi:10.1016/j.egypro.2013.08.062

538

Beniamino Gioli et al. / Energy Procedia 40 (2013) 537 – 544

vapour and energy fluxes have commonly been measured, revealing that cities are an important source of the global carbon budget (Vogt et al. [4]). Recent technological developments include fast response instruments for atmospheric measurements of methane concentration (CH4) based on laser spectroscopy (Detto et al. [5]), that include open path devices especially designed for eddy covariance flux measurements (McDermitt et al. [6]); and particles optical counters capable of resolving a wide range of optical diameters maintaining a sufficient time response for flux measurements (Fratini et al. [7]). In this study, we deploy a long-term dataset of CO2 flux measurements, coupled to short-term campaigns of CH4 and particles flux measurements. The measurement site is located in a densely urbanized area in the central city area of Florence (Italy), where fluxes are entirely governed by anthropogenic emissions, considering the lack of green-space in the flux footprint. The importance of anthropogenic controls on observed emissions, such as road transportation, domestic and industrial heating, or gas-network leakages, is assessed through multi-variate statistical analysis using inventorial data and emission proxies such as traffic counters and gas network flow rates. This analysis allows the direct partitioning of CO2 and CH4 observed fluxes into their main drivers. 2. Material and Methods 2.1. Measurement site The measurement site is located in Florence, a 400.000 inhabitants city in central Italy about 80 km east of the sea. The orography of the site made by a closed basin and its continental climate favour episodes with high atmospheric stability and heavy pollution. The city centre of Firenze is characterized by small squares surrounded by a network of narrow streets linked to the ancient gates. There is almost no vegetation in the city-centre area. The meteorological and micro-meteorological variables are measured at the Osservatorio Ximeniano located in the city centre where a 3 m mast was erected on a typical tile roof 33 m above street level and 14 m above the average roof level of surrounding buildings (Fig. 1a). 2.2. Instrumentation Turbulent fluxes of CO2, momentum, and energy were collected using a sonic anemometer (Metek USA1) and an open-path CO2/H2O infrared gad analyzer (LI7500, LI-COR Bio-sciences, Lincoln, NE, USA). The distance between the two sensors was approximately 0.4 m. Raw data are collected at the frequency of 20 Hz. CH4 high frequency concentration was measured with an open-path methane analyzer (LI7700, LI-COR Bio-sciences, Lincoln, NE, USA) in conjunction with the CO2 fluxes system, sharing the same sonic anemometer (Fig. 1b). A new instrument called EOLO (Eddy cOvariance-based upLift Observation system) was installed to collect turbulent fluxes of dust. Size-segregated fluxes of mineral dust particles with aerodynamic diameters between 0.35 and 9.50 were measured. The system includes an Optical Particle Counter (OPC) (CI-3100 series, Climet Instruments Co., Redlands, CA, USA) and a Multi-Channel Analyzers (MCA8000, Amptek Inc., Bedford, MA, USA). Dust measurements are coupled to 3-D wind measurements provided by a dedicated sonic anemometer (Fig. 1b). An acquisition frequency of 10 Hz was adopted. The system is able to simultaneously record counts of particles falling in 18 size ranges. The number of particles per unit of sampled volume (number concentration) is than obtained. In this study, we report results related to diameters below 2.5 , referred to as PM2.5.

Beniamino Gioli et al. / Energy Procedia 40 (2013) 537 – 544

2.3. Eddy covariance flux measurements The eddy covariance technique was applied to retrieve halff hourly fluxes. Quality-control procedures included despiking, instrument failure detection in rainy conditions, and stationarity analysis (Foken and Wichura [8]). Single point storage correction was applied as described in Papale et al [9]. A threshold friction velocity equal to 0.13 m s 1 was derived by analyzing the dependence of night time CO 2 flux on u on selected periods and detecting the value beyond which the flux leveled off. The storage correction was applied before the u correction to avoid the double counting effect. Data that failed quality criteria were gap-filled with a procedure based on a mean diurnal variation (Falge et al. [10]), where a missing observation at a certain time is replaced by the mean for that time based on adjacent days. Path averaging and sensor separation corrections were applied using the Moore [11] first order transfer functions. For methane, correction terms related to air density fluctuations affecting both spectroscopic response and mass density retrieval. Particle flux data were retrieved as described in Fratini et al. [7]. In addition, two quality filters were applied: a despiking procedure, eliminating individual flux values that show too large differences with the respect to the preceding and the following values in the time series; an uncertainty estimation, as in Buzorius et al. [12], resulting in individual flux values being eliminated if relative uncertainty was larger than 40%. This led to elimination of most fluxes of small intensity, notably night-time fluxes. Available data cover overall about 60% of the whole study period (Toscano et al. [13]). 2.4. Experimental campaigns EC flux measurements of CO2 are made long-term since seven years, while short-term campaigns have been aimed at measuring CH4 and particles fluxes. W 2 includes Particle fluxes for a period extends from June to December 2010, while the second refers to CH4 flux measurements, spanning 3 months from March to June 2011.

Fig. 1. (a) View of the measurement site of "Osservatorio Ximeniano" in central Florence, Italy; (b) detail of the eddy covariance mast, equipped with two sonic anemometers (Metek and Gill), CO2 and H2O open path analyzer (Li7000), CH4 open path analyzer (Li7700), and optical particle counter (EOLO).

539

540

Beniamino Gioli et al. / Energy Procedia 40 (2013) 537 – 544

2.5. Inventorial data Inventorial variables used in this study include number of circulating vehicles measured at several check-points in the city area, used as a proxy for road traffic intensity, and the natural gas flow-rates in the city network measured at the two main distribution points that serve the city area, used as a proxy for domestic heating intensity. A linear multi-regressive framework was adopted to partition observed CO2 and CH4 fluxes into the components related to road traffic and domestic heating, respectively. Refer to Gioli et al. [14] for a complete description of the methodology. 3. Results and Discussion CO2 long-term fluxes computed across the long-term study period (2005-2012) are always a net source throughout the year (Fig. 2a). Inter-annual variability is rather small within 10% of mean values, and associated with a high seasonality (Fig. 2a). Fluxes range from 39 to 172% of the mean annual value in the summer and winter respectively. Air temperature is not correlated with emissions in warm seasons (Tair air < 288 K°) (Fig. 2b). Seasonality is therefore likely controlled by domestic heating in the cold season, while fluxes are more stable in the warm season (Fig. 2a).

Fig. 2. (a) yearly course of CO2 fluxes at weekly resolution across entire study period 2005-2012; error bars represent 95% confidence intervals (b) relation between CO2 fluxes and air temperature at weekly resolution; red line is the linear fit for Tair < 288°K while blue line is the linear fit for Tair 288°K.

The short-term campaign of CH4 flux measurements was focused on the transition between winter and spring, spanning the progressive decrease in natural gas usage intensity. Natural gas usage is in fact maximum in the winter due to domestic heating, while it drops to a minimum value in spring and summer. Measured CH4 fluxes to the atmosphere are relevant, they in fact represent about 8% of CO2equivalent GHG emissions (Fig. 3). The temporal pattern is rather different between CH 4 and CO2 over the same study period: while CO2 exhibits a clear transition form high to low emission intensity, CH 4 fluxes do not exhibit any statistically significant seasonality (Fig. 3). This behaviour suggests that domestic heating, by transforming CH4 into CO2 with combustion, is not the direct driver of CH4 emissions to the atmosphere.

Beniamino Gioli et al. / Energy Procedia 40 (2013) 537 – 544

Fig. 3. seasonal course of CO2 (red, left axis) and CH4 (black, right axis) fluxes during winter-spring transition in 2011; error bars represent 95% confidence intervals.

Dynamics at hourly time scales is reported for CO2, CH4, PM2.5, and inventorial proxies for road traffic and domestic heating intensities (Fig. 4). Since CH4 and particle fluxes were measured during two different short-term campaigns, these figures are not related to the same period of the year. Nevertheless, they provide insight on the hourly distribution of fluxes, and the incidence of morning and afternoon peaks.

Fig. 4. hourly resolution daily courses of CO2 fluxes (a), CH4 fluxes (b), PM 2.5 particle fluxes (c), road traffic amounts ((d), left axis in red), natural gas flow rates ((d), right axis in black); for CO2 and CH4 fluxes, daily courses are referred to all study period, warm season and cold season; error bars represent 95% confidence intervals.

541

542

Beniamino Gioli et al. / Energy Procedia 40 (2013) 537 – 544

CO2 and CH4 fluxes exhibit an enhanced morning peak, and a lower peak in the afternoon (Fig. 4a and 4b). A larger peak in the morning is likely related to nocturnal storage that occurs below the measurement point, and that is measured at a later stage during the early morning when turbulent mixing is enhanced. PM2.5 fluxes instead, exhibit two similar peaks in the midday and in the late evening (Fig. 4c). This difference acts as a time shift of particulate fluxes with respect to scalar fluxes, suggesting that the overall diurnal dynamic of particles is regulates both by turbulent conditions and by the presence of deposition. During the night and early morning in fact PM2.5 fluxes are negative, and the surface acts as a sink of particles since turbulent intensity is not sufficient to sustain particle mixing (Fig 4c). This translates also on a time shift of few hours of morning and afternoon flux peaks, with respect scalar fluxes and inventorial proxies (Fig. 4d). A similar analysis was made investigating the patterns of fluxes and inventorial quantities as a function of the day of the week (Fig. 5). Both road traffic and natural gas usage are reduced at weekends (Fig 5d), and the reduction is larger for road traffic (32%). CO2 fluxes reflect this with an average reduction of 24% (Fig. 5a). PM2.5 fluxes exhibit the most pronounced weekend decrease that was computed at -39% (Fig. 5c), while CH4 do not show any statistically significant reduction (Fig. 5b). These figures confirm that CH4 is not directly related to road traffic and domestic heating, and suggest that the contribution of heating to particle emissions is relatively small compared to road traffic.

(b

(a

(c

(d

Fig. 5. weekly courses of CO2 fluxes (a), CH4 fluxes (b), PM 2.5 particle fluxes (c), road traffic amounts ((d), left axis and black circles), natural gas flow rates ((d), right axis and red circles); error bars represent 95% confidence intervals.

Numerically, relative contributions of road traffic and domestic heating to observed CO2 and CH4 emissions was estimated through multi-variate analysis combined with inventorial data and emission proxies such as traffic counters and gas network flow rates (Gioli et al. [14]), revealing that domestic

Beniamino Gioli et al. / Energy Procedia 40 (2013) 537 – 544

heating accounts for more than 80% of observed CO2 fluxes in the winter, and for about half of the fluxes in summer (Table 1). Those two source categories are instead responsible for only 14% of observed CH4 fluxes (Table 1). The major residual part, responsible for 86% of observed emissions, is likely composed of gas network leakages. Such leaks are in fact a constant term that does not depend on season and on effective network usage intensity (Gioli et al. [14]). Table 1. Source partitioning of observed CO2 and CH4 fluxes into components related to road traffic, domestic heating, or other components. CO2

CH4

Summer

Winter

Road traffic

52 %

18 %

12 %

Domestic heating

48 %

82 %

2%

Other

--

--

86 %

4. Conclusions We have demonstrated that the eddy covariance method can provide high temporal resolution urban GHG and particles fluxes, allowing a direct assessment of emission magnitude. Seasonal, weekly and diurnal emission patterns may be used for developing temporal disaggregation schemes of inventorial gross estimates, and as validation data for emission modeling studies. With such an observational framework, that is easily deployable and replicable in other cities, principal emission source categories can be assessed at the urban landscape scale, providing valuable data to verify and validate inventorial and indirect estimates, and also to monitor the gas network efficiency. With the growing emphasis on the impact of non-CO2 gases and dust emissions on climate change, the methodologies presented here can actually help assessing the environmental impact of a city area. While further research is needed to assess also additional GHG contributions like N2O emission from road traffic, our results can provide a baseline to evaluate the impact of future energy policies that may be applied, like restrictions of road traffic, increase in building energy efficiency and improvement in the maintenance of gas distribution networks. References [1] Grimmond CSB, Salmond JA, Oke TR, Offerle B, Lemonsu A. Flux and turbulence measurements at a densely built-up site in Marseille: Heat, mass (water and carbon dioxide), and momentum. J. Geophys. Res. Atmos. 2004; 109, D24. [2] Velasco E, Pressley S, Allwine E, Westberg H, Lamb B. Measurements of landscape. Atmos. Environ. 2005; 39: 7433 -7446. [3] Matese A, Gioli B, Vaccari FP, Zaldei A, Miglietta F. Carbon dioxide emissions of the city center of Firenze, Italy: measurement, evaluation, and source partitioning. J. Appl. Meteorol. Clim. 2009; 48: 1940-1947. [4] Vogt R, Christen A, Rotach MW, Roth M, Satyanarayana ANV. Temporal dynamics of CO2 fluxes and profiles over a central European city. Theor. Appl. Climatol. 2006; 84: 117-126. [5] Detto M, Verfaillie J, Anderson F, Xu L, Baldocchi D. Comparing laser-based open and closed-path gas analyzers to measure methane fluxes using the eddy covariance method. Agric. For. Meteorol. 2011; 151: 1312 1324

543

544

Beniamino Gioli et al. / Energy Procedia 40 (2013) 537 – 544

[6] McDermitt D, Burba G, Xu L, Anderson T, Komissarov A, Riensche B, Schedlbauer J, Starr G, Zona D, Oechel W, Oberbauer S, Hastings S. A new lowcovariance. Appl. Phys. B. 2011; 102: 391-405. [7] Fratini G, Ciccioli P, Febo A, Forgione A, Valentini R. Size segregated fluxes of mineral dust from a desert area of northern China by eddy covariance. Atmos. Chem. Phys. 2007; 7: 2839-2854 [8] Foken T, Wichura B. Tools for quality assessment of surface-based flux measurements. Agric. For. Meteorol. 1996; 78: 83105. [9] Papale D, Reichstein M, Aubinet M, Canfora E, Bernhofer C, Kutsch W, Longdoz B, Rambal S, Valentini R, Vesala T, Yakir D. Towards a standardized processing of net ecosystem exchange measured with eddy covariance technique: algorithms and uncertainty estimation. BG 2006; 3: 571-583 [10] Falge E, Baldocchi D, Olson R, Anthoni P, Aubinet M, Bernhofer C, Burba G, Ceulemans R, Clement R, Dolman H, Granier A, Gross P, Grünwald T, Hollinger D, Jensen NO, Katul G, Keronen P, Kowalski A, Lai CT, Law BE, Meyers T, Moncrieff J, Moors E, Munger JW, Pilegaard K, Rannik Ü, Rebmann C, Suyker A, Tenhunen J, Tu K, Verma S, Vesala T, Wilson K, Wofsy S. Gap filling strategies for defensible annual sums of net ecosystem exchange. Agric. For. Meteorol. 2001; 107: 43-69. [11] Moore C.J. Frequency response corrections for eddy correlation systems. Bound.-Layer Meteor. 1986; 37: 17-35. [12] Buzorius G, Rannik Ü, Nilsson ED, Vesala T, Kulmala M. Analysis of measurement techniques to determine dry deposition velocities of aerosol particles with diameters less than 100 nm. J. Aerosol Sci. 2003; 34: 747-764. [13] Toscano P, Gioli B, Fratini G, Zaldei A, Matese A, Vaccari FP, Miglietta F. CO2 and Aerosol Particle Fluxes Measured by Eddy Covariance in Florence, Italy. International Workshop on Urban Weather and Climate: Observation and Modeling, Beijing, China; 2011, O-25. [14] Gioli B, Toscano P, Lugato E, Matese A, Miglietta F, Zaldei A, Vaccari FP. Methane and carbon dioxide fluxes and source partitioning in urban areas: The case study of Florence, Italy. Environ. Pollut. 2012; 164: 125-131