- Email: [email protected]
CO2 transport by urban plumes in the upper Spanish plateau
Science of the Total Environment 407 (2009) 4934–4938
Contents lists available at ScienceDirect
Science of the Total Environment j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s c i t o t e n v
CO2 transport by urban plumes in the upper Spanish plateau Isidro A. Pérez ⁎, M. Luisa Sánchez, M. Ángeles García, Beatriz de Torre Department of Applied Physics, Faculty of Sciences, University of Valladolid, c/Prado de la Magdalena s/n, 47071 Valladolid, Spain
a r t i c l e
i n f o
Article history: Received 12 March 2009 Received in revised form 19 May 2009 Accepted 21 May 2009 Available online 11 June 2009 Keywords: CO2 plume Richardson number RASS sodar Boundary layer
a b s t r a c t CO2 transport from two cities, Valladolid, over 20 km away and Palencia, over 40 km away from a rural site is analysed through three years of detrended CO2 concentrations obtained near the surface. Meteorological data were obtained from a RASS sodar. Directional analysis by histogram of concentrations above the 95th percentile revealed three differing sectors, one associated to a rural origin and two linked to both cities. Modes indicated anticyclonic turning during plume travel, confirmed by the daily evolution of the wind direction. At night, the Valladolid concentration median was 6 ppm above the Palencia median, which was 2 ppm higher than the rural sector median. Monthly evolution of daily maxima evidenced the Valladolid plume influence in spring and September, whereas the Palencia plume was noticeable in October and November. Skewness analysis showed almost symmetric distributions in the Valladolid plume and right skewed distributions in the Palencia and rural sectors. This result was attributed to the different mixing of both plumes. Vertical gradients of wind speed, direction and potential temperature were also calculated, and evidenced a stratified structure of the lower atmosphere at night and an almost uniform layer during the day. Finally, the median gradient Richardson number showed the highest values, occasionally above 0.8 for the Valladolid sector, implying lower mixing with the environment in the Valladolid plume than in the Palencia plume. © 2009 Elsevier B.V. All rights reserved.
1. Introduction High CO2 concentrations recorded at rural sites are critical in the life cycle of plants, since the timing of flowering is sensitive to this greenhouse gas (Springer and Ward, 2007). These high concentrations are the result of several factors such as sources and meteorological variables. In this sense, concentrations significantly increase during the night under stable stratification (Moriwaki et al., 2006). However, near-surface air temperature has little impact (Idso et al., 1999). Other factors are climatic conditions such as the warming effect on heterotrophic respiration (Kaufmann, 2007) and even the El Niño Southern Oscillation (Chamard et al., 2003). CO2 has been measured in urban environments, which are the major sources (Pataki et al., 2003; Henninger and Kuttler, 2007; Lai and Cheng, 2009). Directional analyses have also been performed to explore the impact of rural air masses on the urban environment (Idso et al., 2001; Nasrallah et al., 2003). Moreover, a comparison between urban and rural CO2 concentrations has occasionally been considered (Idso et al., 2002; George et al., 2007). However, analyzing the impact of CO2 urban plumes on rural environments is uncommon. It is known that in the absence of significant biological activity, CO2 has no major sinks and that its concentrations are thus influenced by source strength and atmospheric transport (Pataki et al., 2005), which is conditioned by the evolution of synoptic systems (Hurwitz et al., 2004).
⁎ Corresponding author. Tel.: +34 983 42 30 00x4187; fax: +34 983 42 30 13. E-mail address: [email protected] (I.A. Pérez). 0048-9697/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2009.05.037
This paper has a twofold aim. Firstly, CO2 origin in a rural environment is established with concentrations measured near the surface and meteorological data provided by a RASS sodar in a three-year campaign. This first objective is achieved through an original combination of directional and CO2 distribution analyses, which consider travel time from source and mixing processes with the environment. The dataset used is long and complete enough to successfully accomplish this objective. However, trend analysis is excluded since a longer observation series would be required (Haszpra et al., 2008). Exploring the impact of atmospheric stability on CO2 concentrations is the second objective of this paper, and is achieved by means of a RASS sodar. The ability of this device to describe atmospheric stability has already been proved (Pérez et al., 2009). However, further exploration of these data is necessary to gain a better insight into the evolution and structure of the lower atmosphere beyond the levels investigated by conventional devices.
2. Experimental description A three-year measuring period was used, commencing on 1 August 2002 at the Low Atmosphere Research Centre (CIBA), 41° 48′ 49″ N, 4° 55′ 59″ W, 24 km NW of Valladolid (Spain, 321 000 pp, 690 m above MSL) and 43 km SW of Palencia (Spain, 81 000 pp, 730 m above MSL). The location, Fig. 1, is a highly extensive plateau 840 m above MSL, with no relief elements, thus ensuring horizontal homogeneity. Non-irrigated crops and grass make up the surrounding vegetation, the roughness length thus being only a few centimetres.
I.A. Pérez et al. / Science of the Total Environment 407 (2009) 4934–4938
4935
Fig. 2. Wind direction 16-sector histogram for CO2 detrended concentrations above the 95th percentile, 20 ppm.
Fig. 1. Map showing the location of the measuring site, CIBA, and the cities of Valladolid and Palencia. Main level curves are also considered.
CO2 measurements were carried out at nearly 2 m from surface using a MIR 9000 continuous analyser. Calibration of zero and span values was regularly performed using gas cylinders of ultra pure nitrogen and CO2 AIR LIQUIDE standards of 400 ppm with a precision of 0.8 ppm. All data were continuously recorded on a DASIBI 8001 datalogger (García et al., 2008). Meteorological variables such as wind speed, wind direction and virtual temperature were obtained from a DSDPA.90-24 sodar with RASS built by METEK GmbH (Bradley, 2008). RASS sodar data were continuously acquired, the only noticeable interruptions occurring for about 25 days (10 days in May 2003 and 15 days in June 2003). The minimum height proposed was 40 m and the maximum 500 m, measurements being limited to 20 m levels. The device generated tenminute wind speed and virtual temperature averages, together with other variables not considered in this paper. Each value was accompanied by its plausibility code, which represents the results of the plausibility test performed on the averaged power spectra, and was used to control data quality. Data were rejected according to the standard criteria considered by the manufacturer. Finally, CO2 and meteorological data were processed as semi-hourly mean values.
A 16-sector histogram based on the first selection of CO2 concentrations and wind direction at 40 m is presented in Fig. 2, which shows two sectors associated to frequent high concentrations surrounded by a wider sector where these observations proved much less frequent. Sectors linked to the highest frequencies were established according to the main sources: the cities of Palencia and Valladolid. 30% of these observations corresponded to the Palencia sector, 33% to the Valladolid sector and 37% to the remaining directions, henceforth referred to as the rural sector. Since the Palencia and Valladolid sectors have the same angular range, frequency by a unit angle for the rural sector was approximately one third that of the Valladolid sector. Discrepancies observed between Fig. 2 maxima and the locations of the two cities should be attributed to two reasons, the first being that the plume meanders due to the orography, since plumes are emitted at a level of more than 100 m below the measuring site. The second reason is the anticyclonic turning of wind, which is more evident in layers uncoupled from the surface. Turning for the Palencia plume was higher than for the Valladolid plume due to the higher travel time. Analysis of available wind data was also performed. Daily evolution of semi-hourly vector means of wind speed was observed at 40 m. Their median proved extremely low, 0.7 m s− 1. However, a noticeable contrast was obtained between day and night time. During the day, wind speed was highly variable, with a maximum of 1.6 m s− 1 at 15 GMT, accompanied by two minima, the first below 0.3 m s− 1 at the beginning of the diurnal period, 9.30 GMT, and a second below 0.4 m s− 1 at the end, 20 GMT. The opposite behaviour was observed during the night time, when wind speed remained nearly constant, about 0.7 m s− 1 from 22 to 8 GMT. Semi-hourly scalar means of wind speed were also calculated, their median being 5.2 m s− 1, higher than the median of vector means, although the range was similar,1.3 m s− 1. Consequently, wind persistence, the ratio of vector mean to scalar mean of wind speed, sometimes used to calculate the standard deviation of wind speed (George et al., 2008) proved low. The median of semi-hourly persistence was 0.14.
3. Results 3.1. Directional analysis CO2 concentrations evidenced a 379.5 ppm median and were lineally fitted, yielding an increase of 8 ppm over the whole measuring period (Pérez et al., in press). However, for our analytical purposes, they were detrended and, in order to consider only high concentrations, the 95th percentile corresponding to 20 ppm was selected. 2194 observations were above this concentration. Moreover, a second selection of these data was made with the same percentile. Only 97 observations above 77 ppm satisfied this assumption. These data occasionally appeared isolated although they frequently presented clusters that could be clearly considered as episodes. A detailed analysis of these episodes revealed that they always took place under anticyclonic conditions.
Fig. 3. Semi-hourly medians of detrended CO2 concentrations for the three sectors proposed.
4936
I.A. Pérez et al. / Science of the Total Environment 407 (2009) 4934–4938
Table 1 CO2 statistics corresponding to observations from 0 to 6 GMT (detrended concentrations in ppm). Sector
Number of observations
Median
1st quartile
3rd quartile
95th percentile
Valladolid Palencia Rural
1835 3014 6483
9.78 3.74 1.30
2.72 − 1.74 − 2.71
19.65 11.52 6.78
50.93 33.46 23.38
Anticyclonic turning was the main feature of wind direction. However, two kinds of periods were obtained. Changes were relevant when wind speed mean was low, but smoother during the day (with W wind, from 11 to 18 GMT), or night (with E wind, from 22.30 to 8.30 GMT), although at a nearly constant rate, about 0.7° h− 1. 3.2. Daily analysis Once directions associated to high CO2 concentrations were established, the daily evolution of all the concentrations was obtained by means of semi-hourly medians. Fig. 3 shows the result where the behaviour was nearly the same during the day for the three sectors proposed, due to the convection developing over this period that determined a noticeable dilution in the mixing layer. However, since stability prevailed at night time, higher concentrations were observed as well as significant differences among them. In order to quantify these differences, median concentrations from 0 to 6 GMT were calculated and presented in Table 1 accompanied by additional statistics. The Valladolid median was 6 ppm higher than the Palencia median, which was more than 2 ppm higher than the rural median, depending on the population of both cities. Finally, differences between sectors increased when higher percentiles were considered. As an example, the 95th percentile in the Valladolid sector was more than 17 ppm higher than this percentile in the Palencia sector, which in turn was over 10 ppm higher than the 95th percentile in the rural sector. To gain an insight into concentration behaviour, a time analysis was performed. Daily maxima were calculated and represented in Fig. 4 to obtain some information. As a general feature, the highest values were observed in spring followed by autumn. This behaviour must be attributed to vegetation and soil cycle which is linked to precipitations occurring in the two seasons. However, certain differences may be observed among the sectors considered. Median concentrations or interquartile range were higher in spring (median was 42 ppm in May) and September in the Valladolid sector. By contrast, the Palencia sector presented noticeable interquartile ranges in October and November (around 30 ppm) and the rural sector showed only a small contribution in autumn. Valladolid sector medians were 21 ppm higher than the rural sector in May and June, although the Palencia sector median was 5 ppm higher than the rural sector in October.
Fig. 5. Daily evolution of wind speed gradient (s− 1) calculated between adjacent levels showing the stratified structure of the lower atmosphere during the night.
The distribution skewness proved another interesting feature, since the Yule-Kendal index (Wilks, 2006) was close to zero for the Valladolid sector (its monthly median was 0.06), and very similar for the Palencia and rural sectors (0.36 and 0.23 respectively). These values indicated that the distribution shape was closely linked to the wind sector. As source types, Valladolid and Palencia are similar, the Valladolid concentrations being higher due to the greater population (its emissions should be four times higher than Palencia emissions), and also due to the shorter time required by the plume to reach the measuring point, implying less mixing in the atmosphere. Consequently, the distribution shape for the Palencia sector should be similar to the rural sector since the plume dilution was higher than the Valladolid plume. The Palencia plume could only occasionally have a more direct impact on the measuring point. However, the Valladolid plume should retain source features. 3.3. Lower atmosphere structure The lower atmosphere structure was clearly revealed when wind speed, wind direction and potential temperature gradients were obtained. Calculations were made between adjacent levels and values obtained were attributed to the lower level. Potential temperature gradient was obtained from ∂Θ ∂T ≈ +Γ ∂z ∂z
ð1Þ
where Γ is the adiabatic lapse rate, which under dry conditions is 0.0098 K m− 1. Fig. 5 presents wind speed gradient medians considered in two-hour intervals to obtain smoother lines up to 300 m, since above this level data availability was lower. There was a sharp contrast between day and night. During the day, only a slight change was observed in the region investigated due to convection. However, during the night a stratified structure comprising several
Fig. 4. Box plot of daily maxima of detrended CO2 concentrations for the sectors used. The box corresponds to the interquartile range, and the line in the box is the median. Whiskers extend from 10th to 90th percentiles and outliers are 5th and 95th percentiles.
I.A. Pérez et al. / Science of the Total Environment 407 (2009) 4934–4938 Table 2 Coefficient medians of potential temperature and wind direction gradients, y, parameterised as a function of height, h, from 60 to 200 m and from 21 to 7 GMT. Equation
Coefficient Potential temperature gradient Wind direction gradient
y = ahb
a b r y = a + b 1n h a b r
0.237 −0.673 −0.981 0.044 −0.007 −0.965
5.734 − 0.855 − 0.953 0.620 − 0.106 − 0.904
r is the correlation coefficient.
canopies with a high wind speed gradient at 100, 160 and 240 m was in evidence. The same structure was obtained with the other variables, a nearly flat region during the daytime and a more simple stratified structure during the night time formed by enveloped layers up to about 200 m. This simple structure was parameterised with a logarithmic or power decrease from 60 to 200 m and from 21 to 7 GMT. Coefficient medians are presented in Table 2. Once potential temperature and wind speed gradients were calculated, both were used to obtain the gradient Richardson number, Ri, which is a static stability index and was calculated with the same procedure as gradients, by layers (King et al., 2008), with the expression ∂θ g ∂z Ri = 2 T ∂u
ð2Þ
∂z
where g is the gravity acceleration, T the temperature of the layer and u the wind speed. This number was obtained by sectors and their bihourly medians were presented in Fig. 6. The contrast between day and night also proved clear. However, sectors had similar Ri values during the day and differences were relevant during the night. The Valladolid sector presented the highest value, above 0.8 at 7 GMT, whereas the Palencia sector reached a value above 0.4, and the rural sector only above 0.3. Consequently, the highest concentrations obtained in the Valladolid sector were a result of not only the strength of the source, but also stronger stability in the lower atmosphere during the night which determined less mixing with the environment. In order to quantify the influence of stability on CO2 concentrations, the period from 0 to 6 GMT was considered, since a noticeable relationship between wind direction and concentration was previously described. Median concentrations were then calculated according to Ri intervals for the lowest level. These concentrations increased until Ri equaled 0.4. Above this value they remained nearly steady at around 9 ppm. This result enabled us to establish two Ri intervals. Above 0.4, stability was high and the 16-sector histogram of wind direction was similar to Fig. 2, i.e., directions of sources were relevant, though only during the period selected. With Ri below 0.4, stability was low and the 16-sector histogram was similar to the wind rose of the measuring site where prevailing directions were ENE and WSW. Consequently, CO2 sources did not appear for low stability. These two Ri intervals were used to calculate median concentrations in the sectors proposed. Change from low to high stability was associated with an increase in median concentrations of about 9, 8 and 6 ppm in Valladolid, Palencia and rural sectors, respectively.
4937
of Valladolid, some 20 km away; NNE, E and ESE, related to the city of Palencia, about 40 km away, and a third sector of rural influence. Episodes of extremely high concentrations above 77 ppm were linked to anticyclonic conditions. The daily evolution of the wind vector at 40 m showed extremely low values of wind speed with two minima at the beginning and end of the diurnal period. Wind direction turned anticyclonically, indicating some uncoupling between the lower atmosphere and surface. The median of Valladolid concentrations from 0 to 6 GMT was 6 ppm higher than the median of Palencia concentrations, which was 2 ppm higher than the rural concentrations. The analysis of daily maxima showed that the Valladolid plume impacted spring and September concentrations and the Palencia plume October and November concentrations. Additionally, concentrations in the Valladolid sector were nearly symmetric, whereas for the remaining sectors they were right skewed. This result was attributed to the different mixing degree of urban plumes in the rural environment. The Valladolid plume was only scarcely mixed since the travel time to the measuring site was shorter, whereas mixing with the environment was higher for the Palencia plume. Wind speed, wind direction and potential temperature gradients were calculated in 20-m layers, evidencing a stratified structure of the lower atmosphere during the night and a nearly uniform canopy during the day. In particular, wind direction gradient and potential temperature gradient may be observed as enveloped layers during the night up to 200 m, being parameterised from 60 to 200 m. The gradient Richardson number was calculated for the three sectors proposed. It was especially high during the night for the Valladolid sector, even above 0.8, evidencing lower dispersion and mixing. Finally, the method presented in this paper may prove useful to isolate CO2 urban sources and quantify their impact in less polluted environments.
4. Conclusions A three-year analysis of semi-hourly CO2 detrended concentrations measured at 2 m in a rural site in the upper Spanish plateau was performed together with meteorological observations obtained from a RASS sodar. A 16-sector histogram of concentrations above the 95th percentile, 20 ppm, revealed three sectors: SE, SSE and S associated with the city
Fig. 6. Gradient Richardson number calculated for the three sectors proposed.
4938
I.A. Pérez et al. / Science of the Total Environment 407 (2009) 4934–4938
Acknowledgements The authors wish to acknowledge the financial support of the Interministerial Commission of Science and Technology and the Regional Government of Castile and Leon. References Bradley S. Atmospheric acoustic remote sensing. Boca Raton: CRC Press; 2008. 271 pp. Chamard P, Thiery F, Di Sarra A, Ciattaglia L, De Silvestry L, Grigioni P, et al. Interannual variability of atmospheric CO2 in the Mediterranean: measurements at the island of Lampedusa. Tellus Ser B Chem Phys Meteorol 2003;55:83–93. García MA, Sánchez ML, Pérez IA, de Torre B. Continuous carbon dioxide measurements in a rural area in the upper Spanish plateau. J Air Waste Manage Assoc 2008;58:940–6. George K, Ziska LH, Bunce JA, Quebedeaux B. Elevated atmospheric CO2 concentration and temperature across an urban–rural transect. Atmos Environ 2007;41:7654–65. George KV, Verma P, Devotta S. Locating air quality monitoring station using wind impact area diagram. Environ Monit Assess 2008;145:113–8. Haszpra L, Barcza Z, Hidy D, Szilágyi I, Dlugokencky E, Tans P. Trends and temporal variations of major greenhouse gases at a rural site in Central Europe. Atmos Environ 2008;42:8707–16. Henninger S, Kuttler W. Methodology for mobile measurements of carbon dioxide within the urban canopy layer. Clim Res 2007;34:161–7. Hurwitz MD, Ricciuto DM, Bakwin PS, Davis KJ, Wang W, Yi C, et al. Transport of carbon dioxide in the presence of storm systems over a Northern Wisconsin forest. J Atmos Sci 2004;61:607–18. Idso CD, Idso SB, Balling Jr RC. The relationship between near-surface air temperature over land and the annual amplitude of the atmosphere's seasonal CO2 cycle. Environ Exp Bot 1999;41:31–7.
Idso CD, Idso SB, Balling Jr RC. An intensive two-week study of an urban CO2 dome in Phoenix, Arizona, USA. Atmos Environ 2001;35:995-1000. Idso SB, Idso CD, Balling Jr RC. Seasonal and diurnal variations of near-surface atmospheric CO2 concentration within a residential sector of the urban CO2 dome of Phoenix, AZ, USA. Atmos Environ 2002;36:1655–60. Kaufmann RK. Monthly changes in atmospheric carbon dioxide at Mauna Loa: possible asymmetric effects of warming. Earth Interact 2007;11. doi:10.1175/EI224.1. King JC, Lachlan-Cope TA, Ladkin RS, Weiss A. Airborne measurements in the stable boundary layer over the Larsen Ice Shelf, Antarctica. Bound Layer Meteorol 2008;127:413–28. Lai LW, Cheng WL. Air quality influenced by urban heat island coupled with synoptic weather patterns. Sci Total Environ 2009;407:2724–33. Moriwaki R, Kanda M, Nitta H. Carbon dioxide build-up within a suburban canopy layer in winter night. Atmos Environ 2006;40:1394–407. Nasrallah HA, Balling Jr RC, Madi SM, Al-Ansari L. Temporal variations in atmospheric CO2 concentrations in Kuwait City, Kuwait with comparisons to Phoenix, Arizona, USA. Environ Pollut 2003;121:301–5. Pataki DE, Bowling DR, Ehleringer JR. Seasonal cycle of carbon dioxide and its isotopic composition in an urban atmosphere: anthropogenic and biogenic effects. J Geophys Res 2003;108(D23):4735. doi:10.1029/2003JD003865. Pataki DE, Tyler BJ, Peterson RE, Nair AP, Steenburgh WJ, Pardyjak ER. Can carbon dioxide be used as a tracer of urban atmospheric transport? J Geophys Res 2005;110:D15102. doi:10.1029/2004JD005723. Pérez IA, Sánchez ML, García MA, de Torre B. Boundary layer structure and stability classification validated with CO2 concentrations over the Northern Spanish Plateau. Ann Geophys 2009;27:339–49. Pérez IA, Sánchez ML, García MA, de Torre B. Daily and annual cycle of CO2 concentration near the surface depending on boundary layer structure at a rural site in Spain. Theor Appl Climatol in press. doi:10.1007/s00704-009-0112-2. Springer CJ, Ward JK. Flowering time and elevated atmospheric CO2. New Phytol 2007;176:243–55. Wilks DS. Statistical methods in the atmospheric sciences. Amsterdam, Boston: Academic Press; 2006. 627 pp.
Contents lists available at ScienceDirect
Science of the Total Environment j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s c i t o t e n v
CO2 transport by urban plumes in the upper Spanish plateau Isidro A. Pérez ⁎, M. Luisa Sánchez, M. Ángeles García, Beatriz de Torre Department of Applied Physics, Faculty of Sciences, University of Valladolid, c/Prado de la Magdalena s/n, 47071 Valladolid, Spain
a r t i c l e
i n f o
Article history: Received 12 March 2009 Received in revised form 19 May 2009 Accepted 21 May 2009 Available online 11 June 2009 Keywords: CO2 plume Richardson number RASS sodar Boundary layer
a b s t r a c t CO2 transport from two cities, Valladolid, over 20 km away and Palencia, over 40 km away from a rural site is analysed through three years of detrended CO2 concentrations obtained near the surface. Meteorological data were obtained from a RASS sodar. Directional analysis by histogram of concentrations above the 95th percentile revealed three differing sectors, one associated to a rural origin and two linked to both cities. Modes indicated anticyclonic turning during plume travel, confirmed by the daily evolution of the wind direction. At night, the Valladolid concentration median was 6 ppm above the Palencia median, which was 2 ppm higher than the rural sector median. Monthly evolution of daily maxima evidenced the Valladolid plume influence in spring and September, whereas the Palencia plume was noticeable in October and November. Skewness analysis showed almost symmetric distributions in the Valladolid plume and right skewed distributions in the Palencia and rural sectors. This result was attributed to the different mixing of both plumes. Vertical gradients of wind speed, direction and potential temperature were also calculated, and evidenced a stratified structure of the lower atmosphere at night and an almost uniform layer during the day. Finally, the median gradient Richardson number showed the highest values, occasionally above 0.8 for the Valladolid sector, implying lower mixing with the environment in the Valladolid plume than in the Palencia plume. © 2009 Elsevier B.V. All rights reserved.
1. Introduction High CO2 concentrations recorded at rural sites are critical in the life cycle of plants, since the timing of flowering is sensitive to this greenhouse gas (Springer and Ward, 2007). These high concentrations are the result of several factors such as sources and meteorological variables. In this sense, concentrations significantly increase during the night under stable stratification (Moriwaki et al., 2006). However, near-surface air temperature has little impact (Idso et al., 1999). Other factors are climatic conditions such as the warming effect on heterotrophic respiration (Kaufmann, 2007) and even the El Niño Southern Oscillation (Chamard et al., 2003). CO2 has been measured in urban environments, which are the major sources (Pataki et al., 2003; Henninger and Kuttler, 2007; Lai and Cheng, 2009). Directional analyses have also been performed to explore the impact of rural air masses on the urban environment (Idso et al., 2001; Nasrallah et al., 2003). Moreover, a comparison between urban and rural CO2 concentrations has occasionally been considered (Idso et al., 2002; George et al., 2007). However, analyzing the impact of CO2 urban plumes on rural environments is uncommon. It is known that in the absence of significant biological activity, CO2 has no major sinks and that its concentrations are thus influenced by source strength and atmospheric transport (Pataki et al., 2005), which is conditioned by the evolution of synoptic systems (Hurwitz et al., 2004).
⁎ Corresponding author. Tel.: +34 983 42 30 00x4187; fax: +34 983 42 30 13. E-mail address: [email protected] (I.A. Pérez). 0048-9697/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2009.05.037
This paper has a twofold aim. Firstly, CO2 origin in a rural environment is established with concentrations measured near the surface and meteorological data provided by a RASS sodar in a three-year campaign. This first objective is achieved through an original combination of directional and CO2 distribution analyses, which consider travel time from source and mixing processes with the environment. The dataset used is long and complete enough to successfully accomplish this objective. However, trend analysis is excluded since a longer observation series would be required (Haszpra et al., 2008). Exploring the impact of atmospheric stability on CO2 concentrations is the second objective of this paper, and is achieved by means of a RASS sodar. The ability of this device to describe atmospheric stability has already been proved (Pérez et al., 2009). However, further exploration of these data is necessary to gain a better insight into the evolution and structure of the lower atmosphere beyond the levels investigated by conventional devices.
2. Experimental description A three-year measuring period was used, commencing on 1 August 2002 at the Low Atmosphere Research Centre (CIBA), 41° 48′ 49″ N, 4° 55′ 59″ W, 24 km NW of Valladolid (Spain, 321 000 pp, 690 m above MSL) and 43 km SW of Palencia (Spain, 81 000 pp, 730 m above MSL). The location, Fig. 1, is a highly extensive plateau 840 m above MSL, with no relief elements, thus ensuring horizontal homogeneity. Non-irrigated crops and grass make up the surrounding vegetation, the roughness length thus being only a few centimetres.
I.A. Pérez et al. / Science of the Total Environment 407 (2009) 4934–4938
4935
Fig. 2. Wind direction 16-sector histogram for CO2 detrended concentrations above the 95th percentile, 20 ppm.
Fig. 1. Map showing the location of the measuring site, CIBA, and the cities of Valladolid and Palencia. Main level curves are also considered.
CO2 measurements were carried out at nearly 2 m from surface using a MIR 9000 continuous analyser. Calibration of zero and span values was regularly performed using gas cylinders of ultra pure nitrogen and CO2 AIR LIQUIDE standards of 400 ppm with a precision of 0.8 ppm. All data were continuously recorded on a DASIBI 8001 datalogger (García et al., 2008). Meteorological variables such as wind speed, wind direction and virtual temperature were obtained from a DSDPA.90-24 sodar with RASS built by METEK GmbH (Bradley, 2008). RASS sodar data were continuously acquired, the only noticeable interruptions occurring for about 25 days (10 days in May 2003 and 15 days in June 2003). The minimum height proposed was 40 m and the maximum 500 m, measurements being limited to 20 m levels. The device generated tenminute wind speed and virtual temperature averages, together with other variables not considered in this paper. Each value was accompanied by its plausibility code, which represents the results of the plausibility test performed on the averaged power spectra, and was used to control data quality. Data were rejected according to the standard criteria considered by the manufacturer. Finally, CO2 and meteorological data were processed as semi-hourly mean values.
A 16-sector histogram based on the first selection of CO2 concentrations and wind direction at 40 m is presented in Fig. 2, which shows two sectors associated to frequent high concentrations surrounded by a wider sector where these observations proved much less frequent. Sectors linked to the highest frequencies were established according to the main sources: the cities of Palencia and Valladolid. 30% of these observations corresponded to the Palencia sector, 33% to the Valladolid sector and 37% to the remaining directions, henceforth referred to as the rural sector. Since the Palencia and Valladolid sectors have the same angular range, frequency by a unit angle for the rural sector was approximately one third that of the Valladolid sector. Discrepancies observed between Fig. 2 maxima and the locations of the two cities should be attributed to two reasons, the first being that the plume meanders due to the orography, since plumes are emitted at a level of more than 100 m below the measuring site. The second reason is the anticyclonic turning of wind, which is more evident in layers uncoupled from the surface. Turning for the Palencia plume was higher than for the Valladolid plume due to the higher travel time. Analysis of available wind data was also performed. Daily evolution of semi-hourly vector means of wind speed was observed at 40 m. Their median proved extremely low, 0.7 m s− 1. However, a noticeable contrast was obtained between day and night time. During the day, wind speed was highly variable, with a maximum of 1.6 m s− 1 at 15 GMT, accompanied by two minima, the first below 0.3 m s− 1 at the beginning of the diurnal period, 9.30 GMT, and a second below 0.4 m s− 1 at the end, 20 GMT. The opposite behaviour was observed during the night time, when wind speed remained nearly constant, about 0.7 m s− 1 from 22 to 8 GMT. Semi-hourly scalar means of wind speed were also calculated, their median being 5.2 m s− 1, higher than the median of vector means, although the range was similar,1.3 m s− 1. Consequently, wind persistence, the ratio of vector mean to scalar mean of wind speed, sometimes used to calculate the standard deviation of wind speed (George et al., 2008) proved low. The median of semi-hourly persistence was 0.14.
3. Results 3.1. Directional analysis CO2 concentrations evidenced a 379.5 ppm median and were lineally fitted, yielding an increase of 8 ppm over the whole measuring period (Pérez et al., in press). However, for our analytical purposes, they were detrended and, in order to consider only high concentrations, the 95th percentile corresponding to 20 ppm was selected. 2194 observations were above this concentration. Moreover, a second selection of these data was made with the same percentile. Only 97 observations above 77 ppm satisfied this assumption. These data occasionally appeared isolated although they frequently presented clusters that could be clearly considered as episodes. A detailed analysis of these episodes revealed that they always took place under anticyclonic conditions.
Fig. 3. Semi-hourly medians of detrended CO2 concentrations for the three sectors proposed.
4936
I.A. Pérez et al. / Science of the Total Environment 407 (2009) 4934–4938
Table 1 CO2 statistics corresponding to observations from 0 to 6 GMT (detrended concentrations in ppm). Sector
Number of observations
Median
1st quartile
3rd quartile
95th percentile
Valladolid Palencia Rural
1835 3014 6483
9.78 3.74 1.30
2.72 − 1.74 − 2.71
19.65 11.52 6.78
50.93 33.46 23.38
Anticyclonic turning was the main feature of wind direction. However, two kinds of periods were obtained. Changes were relevant when wind speed mean was low, but smoother during the day (with W wind, from 11 to 18 GMT), or night (with E wind, from 22.30 to 8.30 GMT), although at a nearly constant rate, about 0.7° h− 1. 3.2. Daily analysis Once directions associated to high CO2 concentrations were established, the daily evolution of all the concentrations was obtained by means of semi-hourly medians. Fig. 3 shows the result where the behaviour was nearly the same during the day for the three sectors proposed, due to the convection developing over this period that determined a noticeable dilution in the mixing layer. However, since stability prevailed at night time, higher concentrations were observed as well as significant differences among them. In order to quantify these differences, median concentrations from 0 to 6 GMT were calculated and presented in Table 1 accompanied by additional statistics. The Valladolid median was 6 ppm higher than the Palencia median, which was more than 2 ppm higher than the rural median, depending on the population of both cities. Finally, differences between sectors increased when higher percentiles were considered. As an example, the 95th percentile in the Valladolid sector was more than 17 ppm higher than this percentile in the Palencia sector, which in turn was over 10 ppm higher than the 95th percentile in the rural sector. To gain an insight into concentration behaviour, a time analysis was performed. Daily maxima were calculated and represented in Fig. 4 to obtain some information. As a general feature, the highest values were observed in spring followed by autumn. This behaviour must be attributed to vegetation and soil cycle which is linked to precipitations occurring in the two seasons. However, certain differences may be observed among the sectors considered. Median concentrations or interquartile range were higher in spring (median was 42 ppm in May) and September in the Valladolid sector. By contrast, the Palencia sector presented noticeable interquartile ranges in October and November (around 30 ppm) and the rural sector showed only a small contribution in autumn. Valladolid sector medians were 21 ppm higher than the rural sector in May and June, although the Palencia sector median was 5 ppm higher than the rural sector in October.
Fig. 5. Daily evolution of wind speed gradient (s− 1) calculated between adjacent levels showing the stratified structure of the lower atmosphere during the night.
The distribution skewness proved another interesting feature, since the Yule-Kendal index (Wilks, 2006) was close to zero for the Valladolid sector (its monthly median was 0.06), and very similar for the Palencia and rural sectors (0.36 and 0.23 respectively). These values indicated that the distribution shape was closely linked to the wind sector. As source types, Valladolid and Palencia are similar, the Valladolid concentrations being higher due to the greater population (its emissions should be four times higher than Palencia emissions), and also due to the shorter time required by the plume to reach the measuring point, implying less mixing in the atmosphere. Consequently, the distribution shape for the Palencia sector should be similar to the rural sector since the plume dilution was higher than the Valladolid plume. The Palencia plume could only occasionally have a more direct impact on the measuring point. However, the Valladolid plume should retain source features. 3.3. Lower atmosphere structure The lower atmosphere structure was clearly revealed when wind speed, wind direction and potential temperature gradients were obtained. Calculations were made between adjacent levels and values obtained were attributed to the lower level. Potential temperature gradient was obtained from ∂Θ ∂T ≈ +Γ ∂z ∂z
ð1Þ
where Γ is the adiabatic lapse rate, which under dry conditions is 0.0098 K m− 1. Fig. 5 presents wind speed gradient medians considered in two-hour intervals to obtain smoother lines up to 300 m, since above this level data availability was lower. There was a sharp contrast between day and night. During the day, only a slight change was observed in the region investigated due to convection. However, during the night a stratified structure comprising several
Fig. 4. Box plot of daily maxima of detrended CO2 concentrations for the sectors used. The box corresponds to the interquartile range, and the line in the box is the median. Whiskers extend from 10th to 90th percentiles and outliers are 5th and 95th percentiles.
I.A. Pérez et al. / Science of the Total Environment 407 (2009) 4934–4938 Table 2 Coefficient medians of potential temperature and wind direction gradients, y, parameterised as a function of height, h, from 60 to 200 m and from 21 to 7 GMT. Equation
Coefficient Potential temperature gradient Wind direction gradient
y = ahb
a b r y = a + b 1n h a b r
0.237 −0.673 −0.981 0.044 −0.007 −0.965
5.734 − 0.855 − 0.953 0.620 − 0.106 − 0.904
r is the correlation coefficient.
canopies with a high wind speed gradient at 100, 160 and 240 m was in evidence. The same structure was obtained with the other variables, a nearly flat region during the daytime and a more simple stratified structure during the night time formed by enveloped layers up to about 200 m. This simple structure was parameterised with a logarithmic or power decrease from 60 to 200 m and from 21 to 7 GMT. Coefficient medians are presented in Table 2. Once potential temperature and wind speed gradients were calculated, both were used to obtain the gradient Richardson number, Ri, which is a static stability index and was calculated with the same procedure as gradients, by layers (King et al., 2008), with the expression ∂θ g ∂z Ri = 2 T ∂u
ð2Þ
∂z
where g is the gravity acceleration, T the temperature of the layer and u the wind speed. This number was obtained by sectors and their bihourly medians were presented in Fig. 6. The contrast between day and night also proved clear. However, sectors had similar Ri values during the day and differences were relevant during the night. The Valladolid sector presented the highest value, above 0.8 at 7 GMT, whereas the Palencia sector reached a value above 0.4, and the rural sector only above 0.3. Consequently, the highest concentrations obtained in the Valladolid sector were a result of not only the strength of the source, but also stronger stability in the lower atmosphere during the night which determined less mixing with the environment. In order to quantify the influence of stability on CO2 concentrations, the period from 0 to 6 GMT was considered, since a noticeable relationship between wind direction and concentration was previously described. Median concentrations were then calculated according to Ri intervals for the lowest level. These concentrations increased until Ri equaled 0.4. Above this value they remained nearly steady at around 9 ppm. This result enabled us to establish two Ri intervals. Above 0.4, stability was high and the 16-sector histogram of wind direction was similar to Fig. 2, i.e., directions of sources were relevant, though only during the period selected. With Ri below 0.4, stability was low and the 16-sector histogram was similar to the wind rose of the measuring site where prevailing directions were ENE and WSW. Consequently, CO2 sources did not appear for low stability. These two Ri intervals were used to calculate median concentrations in the sectors proposed. Change from low to high stability was associated with an increase in median concentrations of about 9, 8 and 6 ppm in Valladolid, Palencia and rural sectors, respectively.
4937
of Valladolid, some 20 km away; NNE, E and ESE, related to the city of Palencia, about 40 km away, and a third sector of rural influence. Episodes of extremely high concentrations above 77 ppm were linked to anticyclonic conditions. The daily evolution of the wind vector at 40 m showed extremely low values of wind speed with two minima at the beginning and end of the diurnal period. Wind direction turned anticyclonically, indicating some uncoupling between the lower atmosphere and surface. The median of Valladolid concentrations from 0 to 6 GMT was 6 ppm higher than the median of Palencia concentrations, which was 2 ppm higher than the rural concentrations. The analysis of daily maxima showed that the Valladolid plume impacted spring and September concentrations and the Palencia plume October and November concentrations. Additionally, concentrations in the Valladolid sector were nearly symmetric, whereas for the remaining sectors they were right skewed. This result was attributed to the different mixing degree of urban plumes in the rural environment. The Valladolid plume was only scarcely mixed since the travel time to the measuring site was shorter, whereas mixing with the environment was higher for the Palencia plume. Wind speed, wind direction and potential temperature gradients were calculated in 20-m layers, evidencing a stratified structure of the lower atmosphere during the night and a nearly uniform canopy during the day. In particular, wind direction gradient and potential temperature gradient may be observed as enveloped layers during the night up to 200 m, being parameterised from 60 to 200 m. The gradient Richardson number was calculated for the three sectors proposed. It was especially high during the night for the Valladolid sector, even above 0.8, evidencing lower dispersion and mixing. Finally, the method presented in this paper may prove useful to isolate CO2 urban sources and quantify their impact in less polluted environments.
4. Conclusions A three-year analysis of semi-hourly CO2 detrended concentrations measured at 2 m in a rural site in the upper Spanish plateau was performed together with meteorological observations obtained from a RASS sodar. A 16-sector histogram of concentrations above the 95th percentile, 20 ppm, revealed three sectors: SE, SSE and S associated with the city
Fig. 6. Gradient Richardson number calculated for the three sectors proposed.
4938
I.A. Pérez et al. / Science of the Total Environment 407 (2009) 4934–4938
Acknowledgements The authors wish to acknowledge the financial support of the Interministerial Commission of Science and Technology and the Regional Government of Castile and Leon. References Bradley S. Atmospheric acoustic remote sensing. Boca Raton: CRC Press; 2008. 271 pp. Chamard P, Thiery F, Di Sarra A, Ciattaglia L, De Silvestry L, Grigioni P, et al. Interannual variability of atmospheric CO2 in the Mediterranean: measurements at the island of Lampedusa. Tellus Ser B Chem Phys Meteorol 2003;55:83–93. García MA, Sánchez ML, Pérez IA, de Torre B. Continuous carbon dioxide measurements in a rural area in the upper Spanish plateau. J Air Waste Manage Assoc 2008;58:940–6. George K, Ziska LH, Bunce JA, Quebedeaux B. Elevated atmospheric CO2 concentration and temperature across an urban–rural transect. Atmos Environ 2007;41:7654–65. George KV, Verma P, Devotta S. Locating air quality monitoring station using wind impact area diagram. Environ Monit Assess 2008;145:113–8. Haszpra L, Barcza Z, Hidy D, Szilágyi I, Dlugokencky E, Tans P. Trends and temporal variations of major greenhouse gases at a rural site in Central Europe. Atmos Environ 2008;42:8707–16. Henninger S, Kuttler W. Methodology for mobile measurements of carbon dioxide within the urban canopy layer. Clim Res 2007;34:161–7. Hurwitz MD, Ricciuto DM, Bakwin PS, Davis KJ, Wang W, Yi C, et al. Transport of carbon dioxide in the presence of storm systems over a Northern Wisconsin forest. J Atmos Sci 2004;61:607–18. Idso CD, Idso SB, Balling Jr RC. The relationship between near-surface air temperature over land and the annual amplitude of the atmosphere's seasonal CO2 cycle. Environ Exp Bot 1999;41:31–7.
Idso CD, Idso SB, Balling Jr RC. An intensive two-week study of an urban CO2 dome in Phoenix, Arizona, USA. Atmos Environ 2001;35:995-1000. Idso SB, Idso CD, Balling Jr RC. Seasonal and diurnal variations of near-surface atmospheric CO2 concentration within a residential sector of the urban CO2 dome of Phoenix, AZ, USA. Atmos Environ 2002;36:1655–60. Kaufmann RK. Monthly changes in atmospheric carbon dioxide at Mauna Loa: possible asymmetric effects of warming. Earth Interact 2007;11. doi:10.1175/EI224.1. King JC, Lachlan-Cope TA, Ladkin RS, Weiss A. Airborne measurements in the stable boundary layer over the Larsen Ice Shelf, Antarctica. Bound Layer Meteorol 2008;127:413–28. Lai LW, Cheng WL. Air quality influenced by urban heat island coupled with synoptic weather patterns. Sci Total Environ 2009;407:2724–33. Moriwaki R, Kanda M, Nitta H. Carbon dioxide build-up within a suburban canopy layer in winter night. Atmos Environ 2006;40:1394–407. Nasrallah HA, Balling Jr RC, Madi SM, Al-Ansari L. Temporal variations in atmospheric CO2 concentrations in Kuwait City, Kuwait with comparisons to Phoenix, Arizona, USA. Environ Pollut 2003;121:301–5. Pataki DE, Bowling DR, Ehleringer JR. Seasonal cycle of carbon dioxide and its isotopic composition in an urban atmosphere: anthropogenic and biogenic effects. J Geophys Res 2003;108(D23):4735. doi:10.1029/2003JD003865. Pataki DE, Tyler BJ, Peterson RE, Nair AP, Steenburgh WJ, Pardyjak ER. Can carbon dioxide be used as a tracer of urban atmospheric transport? J Geophys Res 2005;110:D15102. doi:10.1029/2004JD005723. Pérez IA, Sánchez ML, García MA, de Torre B. Boundary layer structure and stability classification validated with CO2 concentrations over the Northern Spanish Plateau. Ann Geophys 2009;27:339–49. Pérez IA, Sánchez ML, García MA, de Torre B. Daily and annual cycle of CO2 concentration near the surface depending on boundary layer structure at a rural site in Spain. Theor Appl Climatol in press. doi:10.1007/s00704-009-0112-2. Springer CJ, Ward JK. Flowering time and elevated atmospheric CO2. New Phytol 2007;176:243–55. Wilks DS. Statistical methods in the atmospheric sciences. Amsterdam, Boston: Academic Press; 2006. 627 pp.