Elevated Atmospheric CO2 & Temperature Across Urban-Rural Transect (2007) PDF

ARTICLE IN PRESS Atmospheric Environment 41 (2007) 7654–7665 www.elsevier.com/locate/atmosenv Short communication...

ARTICLE IN PRESS Atmospheric Environment 41 (2007) 7654–7665 www.elsevier.com/locate/atmosenv Short communication Elevated atmospheric CO2 concentration and temperature across an urban–rural transect K. Georgea,b,, L.H. Ziskaa, J.A. Buncea, B. Quebedeauxb a Crop Systems and Global Change Laboratory, USDA-ARS, Building 001, Room 342, 10300 Baltimore Avenue, Beltsville, MD 20705, USA b Department of Plant Sciences and Landscape Architecture, University of Maryland, College Park, MD 20742, USA Received 2 April 2007; received in revised form 23 July 2007; accepted 10 August 2007 Abstract The heat island effect and the high use of fossil fuels in large city centers are well documented, but by how much fossil fuel consumption is elevating atmospheric CO2 concentrations and whether elevations in both atmospheric CO2 and air temperature from rural to urban areas are consistently different from year to year are less well known. Our aim was to record atmospheric CO2 concentrations, air temperature and other environmental variables in an urban area and compare it to suburban and rural sites to see if urban sites are experiencing climates expected globally in the future with climate change. A transect was established from Baltimore city center (Urban site), to the outer suburbs of Baltimore (suburban site) and out to an organic farm (rural site). At each site a weather station was set-up to monitor environmental variables for 5 years. Atmospheric CO2 was consistently and significantly increased on average by 66 ppm from the rural to the urban site over the 5 years of the study. Air temperature was also consistently and significantly higher at the urban site (14.8 1C) compared to the suburban (13.6 1C) and rural (12.7 1C) sites. Relative humidity was not different between sites whereas the vapor pressure deficit (VPD) was significantly higher at the urban site compared to the suburban and rural sites. An increase in nitrogen deposition at the rural site of 0.6% and 1.0% compared to the suburban and urban sites was small enough not to affect soil nitrogen content. Dense urban areas with large populations and high vehicular traffic have significantly different microclimates compared to outlying suburban and rural areas. The increases in atmospheric CO2 and air temperature are similar to changes predicted in the short term with global climate change, therefore providing an environment suitable for studying future effects of climate change on terrestrial ecosystems. Published by Elsevier Ltd. Keywords: Microenvironment; Climate change; Urban ecology; Heat island; Long term 1. Introduction commercial buildings greatly affects the local air quality and energy balance. Cities are large con- The conversion of rural lands into urban areas sumers of fossil fuels, because in the US as in other with high traffic volumes and dense residential and countries, the majority of the population resides in Corresponding author. Crop Systems and Global Change urban areas (United Nations, 2004). This results in urban areas being responsible for the largest Laboratory, USDA-ARS, Building 001, Room 342, 10300 Baltimore Avenue, Beltsville, MD 20705, USA. proportion of anthropogenic emissions such as Tel.: +1 301 504 5527; fax: +1 301 504 5872. CO2 and nitrous oxides (Pataki et al., 2006). Human E-mail address: [email protected] (K. George). and vehicle activity has been found to contribute 1352-2310/$ - see front matter Published by Elsevier Ltd. doi:10.1016/j.atmosenv.2007.08.018  ARTICLE IN PRESS K. George et al. / Atmospheric Environment 41 (2007) 7654–7665 7655 more than 80% of the atmospheric CO2 in urban correlate to high traffic volume during workdays areas (Koerner and Klopatek, 2002). Also, the and is significantly reduced at weekends (Idso et al., conversion of natural land into roads and build- 1998, 2001, 2002; Nasrallah et al., 2003; Velasco ings changes the albedo and heat capacity of an et al., 2005). In Nottingham, UK an 8 month study area resulting in urban areas being significantly found a small difference in CO2 concentration of warmer than if it remained as a rural landscape only 5 ppm between a rural location and an urban (Oke, 1982). If urban environments are elevated city site (Berry and Colls, 1990), although the small in atmospheric CO2 and temperature for sustained difference between sites and large variability in CO2 periods this could provide a suitable system for concentrations are likely related to the close studying the effects of future global climate proximity of power stations to both sites which change on plant population and community dy- were only 15 km apart (Berry and Colls, 1990). In namics without the high cost of equipment and Phoenix, USA CO2 concentration was monitored other resources currently needed to carry out such for nearly a year and values ranged from a daily studies. minimum of 390 ppm rising to a daily maximum of The elevation of air temperatures in city centers 491 ppm, although a maximum value of 619 ppm compared to less built-up areas is a phenomenon was attained (Idso et al., 2002). Over a two week that has been recorded since1833 (Oke, 1982). The period CO2 concentration varied significantly from main factors contributing to the urban heat island day to day with the highest peak of 650 ppm are the high thermal conductivity of buildings and which was 76% higher than the low of 369 ppm other man-made structures, the low albedo and (Idso et al., 2001). Whereas Day et al. (2002) found geometry of city surfaces and low evapotranspira- the urban area was elevated by 19 ppm compared tion (discussed in Taha et al., 1991). Heat released to a suburban area, but their study area was a directly from building ventilation and vehicular distance from major streets and less influenced by traffic also contribute to city heating and can vary vehicle emissions. Few studies have concurrently annually and diurnally (Fan and Sailor, 2005). The compared urban to rural CO2 concentrations to degree of heating within a city is very variable and is determine the amount by which CO2 concentrations unique to each city based on its location, building are elevated by urbanization and whether any layout and traffic (Oke and Maxwell, 1975). A city increases are sustained and consistent from year to wide examination of ground level air temperature year. in Baltimore, USA found that minimum tempera- Urban areas are affecting the microclimate, but tures are closely related to population and the few studies have recorded these changes for long difference between urban and rural minimum periods of time to ensure the consistency of data temperatures has been increasing as population and suitability for investigating effects on plant increases (Brazel et al., 2000). The densely built-up biological systems, nor monitored other global areas in the center of Baltimore had ground level climate change variables concurrently. The aim of air temperatures 5–10 1C warmer than residential this study was to investigate whether a high or forested and agricultural areas (Brazel et al., population city center has a climate similar to that 2000), making it a good model to study climate predicted in the short term (50–100 years) with differences between urban centers and adjacent global climate change. Baltimore was selected as it is rural areas. one of the largest cities in the USA with dense Near-surface CO2 concentrations have been residential and commercial buildings and high documented in several cities across the world traffic volumes in the city center. The outskirts of (Vancouver, Canada; Kuwait City, Kuwait; Mexico Baltimore become more suburban with green areas City, Mexico; Basel, Switzerland; Nottingham, UK; on the outskirts of the city and becoming rural Phoenix, USA) to evaluate the dynamics of atmo- dominated by agricultural land. This location is spheric CO2 over short periods of time (Berry and ideal for comparing microclimate changes from an Colls, 1990; Reid and Steyn, 1997; Idso et al., 2001; urban city center transitioning to a more suburban Nasrallah et al., 2003; Velasco et al., 2005; Vogt and rural areas. The objectives of this study were to et al., 2006). The majority of these studies analyzed characterize the microenvironment associated daily and diurnal fluctuations in CO2 concentra- with an urban location relative to a suburban and tions and concluded that the major source of CO2 is rural location. A secondary objective was to from vehicular traffic as peak CO2 concentrations compare the microenvironmental characteristics to  ARTICLE IN PRESS 7656 K. George et al. / Atmospheric Environment 41 (2007) 7654–7665 climatic conditions predicted with global climate State Highway Administration). The rural site is change. located on an organic farm (Fig. 1A), which predominantly grows alfalfa and orchard grass for animal feed. The area is dominated by agricultural 2. Methods land mainly for grazing with a few residences scattered across the landscape. The nearest roads 2.1. Site description to the site are 1.1 km away and the annual average daily traffic is 5298 vehicles (2002–2006, A transect was established running west from Maryland State Highway Administration). At downtown Baltimore city to a rural agricultural each site four plots were established 2  2 m2 in area in western Maryland. Three sites were selected size, each containing the same uniform fallow along the urban–rural gradient: an organic farm agricultural soil to a depth of 1.1 m and extant seed near Buckeystown, Maryland (391180 N 771260 W, bank. The plant communities in each plot were elevation 109.8 m) approximately 87 km west of allowed to establish naturally, further site descrip- Baltimore (rural site), a nature center approximately tions and results are described in Ziska et al. (2003, 11 km west of Baltimore (391180 N 761410 W, eleva- 2004). tion 98.9 m) on the outer edge of the city (suburban site), and Baltimore city center (391160 N 761360 W, elevation 6.8 m; urban site). Each site is surrounded 2.2. Site microenvironmental measurements by grass, which is mowed frequently through the growing season. The urban site is surrounded by At each site a weather station was established that large commercial and residential buildings and is monitored the following variables with a 15 min very close to a large body of water within the city averaging interval using a CR10X data logger center (Fig. 1C). The area is bordered by busy city (Campbell Scientific, USA): air temperature and roads less than 0.1 km in distance from the site with relative humidity (CS500, Campbell Scientific, annual average daily traffic of 58,083 vehicles (only USA), atmospheric CO2 concentration (S151, 2005 and 2006 data available, Maryland State Quibit, Canada), soil temperature (Model 107, Highway Administration). The suburban site is Campbell Scientific, USA) and moisture (Echo surrounded by trees as it is part of the Carrie EC-20, Decagon Devices, USA), wind speed and Murray nature center within the Gywnn Falls Park direction (Model 03001, R. M. Young Company, (Fig. 1B). The park area is approximately 480 ha USA), and total and diffuse radiation (Sunshine and is surrounded by housing and lawns. The sensor BF3, Delta-T Devices, UK) and precipita- nearest roads are 1.5 km with annual average daily tion (Tipping bucket rain gage TE525, Texas traffic of 33,716 vehicles (2002–2006, Maryland Instruments, USA). The sensors had a sky view Fig. 1. Aerial view of the three sites: (A) rural; (B) suburban; and (C) urban. Images were obtained from Google Earth Beta (v4.1.7076.4458).  ARTICLE IN PRESS K. George et al. / Atmospheric Environment 41 (2007) 7654–7665 7657 factor of 0.93, 0.83 and 0.87 at the rural, suburban Additionally ozone at a height of approximately 1 m and urban sites respectively, indicating the sensors was monitored periodically through the summer had an open sky and were not subjected to long months of 2003 and 2004 and more extensively in periods of shading by trees and buildings. All 2005 and 2006 using chemically sensitive filter sensors were factory calibrated before their installa- paper. Wet deposition of nitrate and nitrite in rain tion in the field. The CO2 analyzers were calibrated water and dry deposition of nitrate from chemically every two weeks with known CO2 concentrations. sensitive filter paper, changed weekly, were mon- Air temperature sensors were not aspirated. Atmo- itored throughout the year of 2005. Soil nitrogen spheric CO2, air temperature, wind speed and content was measured at the start of the growing direction and radiation variables were measured season to estimate the potential impact of the 1.5–2.0 m off the ground. Tipping rain buckets for input of nitrogen to a site from deposition. To precipitation were located approximately 1 m off the remove variability associated with water availabil- ground. Soil temperature and moisture were mea- ity, evaporation at each site was monitored through sured at a depth of 10 cm below the soil surface. the growing season using an evaporation pan 600 Rural 2002 2003 Suburban 550 Urban 500 450 400 350 2004 2005 CO2 concentration (ppm) 550 500 450 400 350 2006 Average all years 550 500 450 400 350 0 4 8 12 16 20 0 4 8 12 16 20 24 Time of Day (hours) Fig. 2. Near surface atmospheric CO2 concentration averaged over a 24 h period for each site along the transect for the 5 years of the study. Data were recorded every 15 min resulting in 96 values over a 24 h period consequently data points and error bars are not included but the data are connected by straight lines. The error bars shown in each graph is the maximum standard deviation from all sites.  ARTICLE IN PRESS 7658 K. George et al. / Atmospheric Environment 41 (2007) 7654–7665 (EP180, Global Water Instrumentation, USA) and using analysis of covariance. The covariate was time any soil moisture deficit at a site compared to the of day and the independent factors were site and others was eliminated by watering. This was to year. Air temperature, soil temperature, precipita- remove any variation in water availability to plant tion, RH and VPD were also analyzed using communities growing at each site for another analysis of covariance but the covariate was day experiment being conducted along this transect. of year. Differences in ozone and soil nitrogen The growing season was defined as the period content between sites were analyzed using analysis between the last frost preceding winter and the first of variance. Data were transformed where appro- frost prior to the onset of the following winter. priate to meet the assumptions of normality and Frost was defined as temperatures reaching below equality of variances for ANOVA. Total and diffuse 0 1C for at least an hour. radiation at the urban and rural sites was analyzed using the nonparametric Mann–Whitney U test as the data were not normally distributed. The 2.3. Statistics variance of soil moisture variables was not equal so the nonparametric Kruskall Wallis test was Atmospheric CO2 was examined to see if there performed. All statistics were performed using were diurnal differences between sites and years Statview (SAS Institute, USA). 35 7 Daily Daily 30 Urban 6 Suburban Urban-rural 25 Urban-suburban 5 Rural Suburban -rural 4 20 3 15 2 10 1 5 0 0 -1 -5 -2 Air temperature difference (°C) 30 Day Day 6 5 Air temperature (°C) 25 4 20 3 15 2 10 1 5 0 0 -1 -5 -2 30 Night Night 6 25 5 4 20 3 15 2 10 1 5 0 0 -1 -5 -2 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Fig. 3. Daily (24 h), daytime and nighttime air temperature for each site along the transect averaged over the 5 years of the study. The error bars for the graphs on the left side are the maximum standard deviation. The graphs on the right represent the deviations in temperature between sites, the error bars are one standard deviation.  ARTICLE IN PRESS K. George et al. / Atmospheric Environment 41 (2007) 7654–7665 7659 3. Results 16% from 2002 to 2006, which is low compared to 2000 and 2001 where CO2 concentration was 3.1. Atmospheric CO2 elevated by 21% (Ziska et al., 2004) and 31% (Ziska et al., 2003), respectively. This variation is Atmospheric CO2 concentration was significantly consistent with our finding that CO2 concentration different between the three sites (Po0.01; Fig. 2). differs significantly between years (P ¼ 0.01), The highest concentration on average across the 5 although the average range in CO2 concentra- years of the study was at the urban site (488 ppm) tions between years for this study was small the lowest at the rural site (422 ppm) and the 443–459 ppm. Time of day, as expected, also suburban site intermediate to the other two sites significantly affected CO2 concentration (Po0.01) (442 ppm). CO2 concentration at the urban site with the lowest CO2 concentration in the early compared to the rural was increased on average by afternoon and peaking in the early hours of the 30 Urban Suburban 25 Rural Soil temperature (°C) 20 15 10 5 0 Urban-suburban Suburban-rural Urban-rural Soil temperature difference (°C) 4 2 0 -2 Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec Fig. 4. Annual soil temperature (10 cm depth) averaged over the 5 years of the study and differences between sites on a monthly basis. The error bars are one standard deviation.  ARTICLE IN PRESS 7660 K. George et al. / Atmospheric Environment 41 (2007) 7654–7665 morning (Fig. 2). There was no difference between difference in soil temperature was between the urban CO2 concentrations on the weekend days (448.6 ppm) and rural sites (0.7 1C) and the urban and suburban compared to the week days (448.7 ppm) at any site. sites (0.5 1C) there was no significant difference in soil temperature between the rural and suburban site 3.2. Air and soil temperature (0.1 1C). The major difference between sites occurred in the last 4 months of the year (Fig. 4) as the urban Daily air temperature differed significantly be- site remained warmer longer while temperatures tween the sites (P ¼ 0.01; Fig. 3) with the highest air dropped quicker at the other two sites. temperature at the urban site (14.8 1C), the lowest at the rural site (12.7 1C) and the suburban site falling 3.3. Moisture variables in between the two extremes (13.6 1C). Across the State of Maryland average annual air temperature Precipitation was not significantly different between from 1901 to 2000 was 12.1 1C (http://www.ncdc. sites (P ¼ 0.17) or between years (P ¼ 0.54), although noaa.gov/oa/climate/research/cag3/md.html). Daily there was considerable variation in precipitation over air temperature was also significantly different the 5 years of the study (Table 2). The highest between years (Po0.01) with the highest average precipitation was in 2003 (1196 mm), which was also temperature in 2006 (14.3 1C) and the lowest in 2003 the wettest year on record for the State of Maryland. (13.2 1C). The same significant differences were This resulted in relative humidity being significantly apparent when temperature was calculated for night greater (Po0.01) and vapor pressure deficit (VPD) to time and day time (Po0.01; Fig. 3). The tempera- be significantly reduced (Po0.01) in 2003 compared ture differences at night time were: urban (13.1 1C), to the other years (Table 2). Relative humidity was suburban (10.9 1C) and rural (10.4 1C) and at day not significantly different between sites (P ¼ 0.12) but time: urban (16.5 1C), suburban (16.2 1C) and rural VPD was significantly greater at the urban site (14.8 1C). The biggest difference in temperature (Po0.01) as air temperature was also higher. occurring at night between urban and rural sites Soil moisture was significantly different between (2.7 1C). Although these differences in temperature site and year. The wettest site on average was the between sites are small, compared to the annual urban site this was because soil moisture in 2004 at variation in temperature, they are consistent through this site was higher than previous years (Table 1). out the year (Fig. 3). The increase in air temperature This may have been because of additional watering at the urban site resulted in the growing season being to meet evaporative demand. Apart from the high longer compared to the suburban and rural sites. On soil moisture at the urban site in 2004 all other sites average the growing season was 258 days at the and years had very similar average soil moisture. urban site and 215 and 210 days, respectively, at the suburban and rural sites (Table 1). 3.4. Solar radiation Soil temperature (10 cm depth) was significantly different between sites (Po0.01; Fig. 4). The highest Total and diffuse radiation was measured at the temperature was at the urban site (14.8 1C) and the urban and rural sites during 2005 and 2006. On a lowest at the rural site (14.1 1C). The greatest daily basis there was no significant difference Table 1 Growing season length in days at each site over 5 years based on the last day that a frost occurred after winter and the first day that a frost appeared before winter Year Rural Suburban Urban Last frost First frost day Growing Last frost First frost day Growing Last frost First frost day Growing day season day season day season length length length 2002 4 April 2 November 213 8 April 1 November 208 24 March 27 November 249 2003 16 March 23 October 222 2 April 23 October 205 15 March 2 December 263 2004 8 April 10 November 217 8 April 10 November 217 25 March 18 December 269 2005 16 April 11 November 210 22 March 17 November 241 15 March 18 November 249 2006 9 April 15 October 190 23 March 13 October 205 21 March 5 December 260  ARTICLE IN PRESS K. George et al. / Atmospheric Environment 41 (2007) 7654–7665 7661 Table 2 Moisture variables from each site across the transect Variable Site 2002 2003 2004 2005 2006 Average Precipitation (mm) Rural 1484 1288 906 978 767 1112 (4) Suburban 778 1150 970 1065 1052 1027 (4) Urban 785 1151 818 679 809 867 (3) RH (%) Rural 61.5 (20.7) 77.1 (14.9) 72.9 (15.2) 70.9 (13.9) 69.1 (14.5) 70.3 (8.9) Suburban 63.8 (18.1) 78.6 (16.1) 69.5 (15.7) 68.6 (14.2) 68.1 (15.9) 68.9 (10.2) Urban 60.2 (15.6) 68.4 (15.8) 66.3 (15.5) 64.9 (15.8) 60.7 (16.1) 64.1 (7.5) VPD (kPa) Rural 0.62 (0.38) 0.35 (0.25) 0.42 (0.26) 0.47 (0.31) 0.50 (0.30) 0.47 (0.20) Suburban 0.60 (0.38) 0.37 (0.26) 0.48 (0.28) 0.52 (0.32) 0.55 (0.35) 0.51 (0.22) Urban 0.72 (0.47) 0.54 (0.38) 0.59 (0.36) 0.67 (0.46) 0.72 (0.46) 0.65 (0.33) Soil moisture (%) Rural 13.8 (1.8) 10.1 (5.3) 9.7 (4.7) 12.0 (3.4) Suburban 10.1 (1.4) 9.3 (3.0) 12.5 (2.3) 10.8 (1.6) Urban 19.3 (2.2) 8.7 (3.9) 12.7 (3.5) 14.4 (2.7) Precipitation is summed over the year. RH, VPD and soil moisture are a daily average over the year. Average is the over all 5 years and the values in brackets are the daily standard deviations. Soil moisture was not recorded in 2002 and 2003. between total radiation (P ¼ 0.08) at the urban between sites (P ¼ 0.59). The addition of wet and (average 347.9 mmol m2 s1) and rural (average dry deposition to soil contributed only 2.1%, 1.8% 398.5 mmol m2 s1) sites. Diffuse radiation and 1.2% nitrogen annually to the rural, suburban although more than 50% lower than total radiation, and urban sites, respectively. was not significantly different (P ¼ 0.19) between the urban (average 159.2 mmol m2 s1) and rural 4. Discussion (average 189.3 mmol m2 s1) sites. Although total radiation is slightly reduced in the urban compared Across the transect, atmospheric CO2 and tem- to the rural site, diffuse radiation is also lower perature were elevated at the urban site and gradually suggesting that the potential for increased air decreased out to the suburban and rural sites. This is pollution in the urban environment is not increasing consistent with other studies that have found air diffuse radiation. temperature and atmospheric CO2 are closely related to population and associated high traffic volume in 3.5. Ozone and nitrogen deposition urban city centers (Idso et al., 1998, 2001, 2002; Brazel et al., 2000; Nasrallah et al., 2003; Velasco Ozone concentrations did not differ significantly et al., 2005). Increased air temperature at the urban between the sites along the transect (P ¼ 0.10; site significantly increased the VPD. Nitrogen deposi- Fig. 5). There was a significant difference between tion although highest at the rural site was not great years (Po0.01); on average the ozone concentration enough to increase soil nitrogen content compared to for 2005 was 45 ppb721 (one standard deviation) the other sites. Along the transect changes in the and for 2006 35 ppb713 which were not high microclimate (IPCC, 2007) and deposition of nutri- enough to affect plant physiology and growth. ents (Denman et al., 2007) are consistent with Ozone was also measured for short time periods in predictions of modifications in the environment 2003 (day 248–273, September) and 2004 (day expected with global climate change. It appears that 78–174, March–June) and the average ozone densely populated urban areas could provide a setting concentrations respectively were 24 ppb77 and that is suitable for studying the effects of future 34 ppb711, which during these short time periods global climate change on terrestrial ecosystems. is similar to the ozone concentrations shown for Globally averaged surface atmospheric CO2 2005 and 2006 (Fig. 5). concentrations are 379 ppm (2005; IPCC, 2007) Wet and dry deposition was highest at the rural and are estimated to increase by 50–100 ppm by and lowest at the urban site (Table 3). Soil total 2100 (Friedlingstein et al., 2006). Along the transect nitrogen content was not significantly different on average the lowest CO2 concentration was at the  ARTICLE IN PRESS 7662 K. George et al. / Atmospheric Environment 41 (2007) 7654–7665 140 2005 120 100 80 60 40 Tropospheric ozone (ppb) 20 0 140 Rural 120 2006 Suburban Urban 100 80 60 40 20 0 0 30 60 90 120 150 180 210 240 270 300 330 360 Day of year Fig. 5. Tropospheric ozone measured at each site through the growing season of 2005 and 2006. The error bars are one standard deviation. rural site, 422 ppm, which increased by 66 ppm at Air temperature globally varies greatly but on the urban site. This difference in CO2 concentration average it is predicted to increase by 1.1–6.4 1C by between sites was maintained over the 5 year period 2100 and this increase will be greatest on land in the with the average annual CO2 concentration at the northern hemisphere (IPCC, 2007). Along the rural site ranging from 395–439 ppm and the urban transect temperature was significantly different site 448–537 ppm. This difference between the urban between sites with an average daily difference and rural sites is similar to the increase in CO2 between the urban and rural sites of 2.1 1C. This concentration found in Phoenix, USA, where an difference was greatest in September and at night increase of 111–185 ppm was reported from a and is likely a consequence of anthropogenic pristine rural site to the city center (Idso et al., heating, such as building and industrial energy 1998, 2001). The sustained increase in CO2 con- consumption, and vehicle fuel combustion, which centration over 5 years between an urban and rural can contribute significantly to the urban heat island site, is within the range expected with global climate in winter (Fan and Sailor, 2005). This resulted in the change predictions (IPCC, 2007), and is expected to growing season at the urban site being 36–70 days significantly impact biological systems. longer over 5 years than the other two sites as  ARTICLE IN PRESS K. George et al. / Atmospheric Environment 41 (2007) 7654–7665 7663 Table 3 Dry and wet deposition nitrogen measured in 2005 at each site and compared to US-EPA data (http://cfpub.epa.gov/gdm/index.cfm) Rural Suburban Urban US-EPA, Beltsville, MD Dry deposition (mg m3) Nitric acid 3.96 3.27 5.24 2.04 Nitrate 1.33 Ammonium 1.43 Wet deposition (kg ha1) Nitrate 7.12 5.08 4.01 2.82 Nitrite 0.19 0.12 0.08 Ammonium 1.76 Soil N (kg ha1) 771.8 735.8 783.9 Total wet deposition (kg ha1) 11.75 8.37 6.59 4.38 Total dry deposition (kg ha1) 6.11 4.35 3.43 2.38 Soil nitrogen content was quantified at each site and addition from total wet and dry deposition was estimated across the transect. Total wet deposition is nitrate and nitrite summed from each site and ammonium estimated as 38% of nitrate based on US-EPA values. Total dry deposition is estimated as 52% of wet deposition based on US-EPA data (http://cfpub.epa.gov/gdm/index.cfm). freezing temperatures did not occur until later in the and VPD significantly reduced. This is consistent with year. Soil temperature was also higher by 0.7 1C at a previous study that found air temperature modified the urban compared to the rural site. The soil moisture variables more than relative humidity except temperature difference between the sites was largest during wet periods (Barradas, 1991). VPD directly in the last four months of the year, similar to the air affects plant physiology (Aphalo and Jarvis, 1991), temperature patterns. Air temperature was consis- influencing gas exchange and growth rates of plants, tently higher at the urban site and similar to which can impact the urban climate. predictions of global climate change resulted in Urban environments can impact air quality variables fewer frost days and warmer night time tempera- other than atmospheric CO2 such as tropospheric tures (IPCC, 2007). ozone and nitrogen deposition. In 2006 Baltimore Precipitation was above average in Maryland experienced 17 days where ozone levels on average for 4 of the 5 years of the study and 2003 was the were above 100 ppb for 8 h and western Maryland wettest year recorded over more than 100 years experienced 2 days (considered unhealthy for sensitive (http://www.ncdc.noaa.gov/oa/climate/research/cag3/ groups), all occurring between May and August md.html). Although precipitation measured in our (http://www.mde.state.md.us/Programs/AirPrograms/ study was not significantly different between sites Monitoring/aqsummaries/index.asp). Our measure- and years, there is a great amount of variation ments at each site indicated peak values between June across the region and between years (Table 2). and September but on average across the year ozone Studies have found that precipitation increases over concentration was below levels that would affect urban areas (Burian and Shepherd, 2005) and is human or plant physiology (McKee, 1994). Wet and predicted to increase in North America with climate dry deposition added a small percent of nitrogen change (Diffenbaugh et al., 2005). Although we compared to soil nitrogen content. Nitrogen deposition saw no differences in precipitation in our study, appeared to be highest at the rural site compared to the in the past decade the State of Maryland as a urban although soil nitrogen values were not different whole is seeing more extreme and higher annual between sites. Total solar radiation appears to be precipitation than experienced in the last 100 years slightly lower at the urban compared to the rural site, (http://www.ncdc.noaa.gov/oa/climate/research/cag3/ which is likely indicative of the urban site experiencing md.html). A consequence of urbanization on some shading from surrounding buildings during part moisture variables observed in our study, was the of the day. Diffuse radiation, which can be increased by increase in air temperature at the urban site particulate matter such as from vehicular emissions, significantly increased VPD, whereas, during wet was not different between the sites and on average was years relative humidity was significantly increased lower at the urban compared to the rural site. It is  ARTICLE IN PRESS 7664 K. George et al. / Atmospheric Environment 41 (2007) 7654–7665 evident from our data that urban environments provide Silva Dias, P.L., Wofsy, S.C., Zhang, X., 2007. Couplings a microclimate that is representative of changes between changes in the climate system and biogeochemistry. predicted in the future with global climate change In: Solomon, S., et al. (Eds.), Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to (IPCC, 2007), consequently vegetation within urban the Fourth Assessment Report of the Intergovernmental areas are currently experiencing elevated atmospheric Panel on Climate Change. Cambridge University Press, CO2 and temperature levels that can significantly affect Cambridge, UK and New York, USA, pp. 500–587. plant growth compared to rural areas. Diffenbaugh, N.S., Pal, J.S., Trapp, R.J., Giorgi, F., 2005. Fine- Over the 5 year period of this study it appears scale processes regulate the response of extreme events to global climate change. Proceedings of the National Academy that a densely populated, urban center, is affecting of Sciences 102 (44), 15774–15778. the microclimate so that it is similar to climate Fan, H., Sailor, D.J., 2005. Modeling the impacts of anthro- predicted globally in the next 100 years. Over the 5 pogenic heating on the urban climate of Philadelphia: a years of the study CO2 concentrations and air comparison of implementations in the two PBL schemes. temperature were consistently higher at the urban Atmospheric Environment 39, 73–84. Friedlingstein, P., Cox, P., Betts, R., Bopp, L., von Bloh, W., site compared to the suburban and rural sites. In the Brovkin, V., Cadule, P., Doney, S., Eby, M., Bala, I., John, J., USA 80% of the population resides in urban areas Jones, C., Joos, F., Kato, T., Kawamiya, M., Knorr, W., and urbanization is continuing to increase (United Lindsay, K., Matthews, H.D., Rayner, T., Reick, C., Nations, 2004). Increasing land use change from Roeckner, E., Schnitzler, K.-G., Schnur, R., Strassmann, rural agricultural areas to dense populated urban K., Weaver, A.J., Yoshikawa Zeng, C., 2006. Climate–carbon cycle feedback analysis: results from the C4MIP model areas will have significant impacts on carbon intercomparison. Journal of Climate 19, 3337–3353. emissions and cycling (Imhoff et al., 2004). Idso, C.D., Idso, S.B., Balling, R.C., 1998. The urban CO2 dome of Phoenix, Arizona. Physical Geography 19, 95–108. Acknowledgments Idso, C.D., Idso, S.B., Balling, R.C., 2001. An intensive two-week study of an urban CO2 dome in Phoenix, Arizona, USA. Atmospheric Environment 35, 995–1000. We would like to thank Ernie Goins, Danielle Idso, S.B., Idso, C.D., Balling, R.C., 2002. Seasonal and diurnal Reid and Jonathan Clark for assistance with data variations of near-surface atmospheric CO2 concentration collection and site maintenance. This research was within a residential sector of the urban CO2 dome of Phoenix, supported by the Biological and Environmental AZ, USA. Atmospheric Environment 36, 1655–1660. Research Program (BER), US Department of Imhoff, M.L., Bounoua, L., DeFries, R., Lawrence, W.T., Stutzer, D., Tucker, C.J., Ricketts, T., 2004. The conse- Energy, Interagency Agreement no. DE-AI02- quences of urban land transformation on net primary 02ER63360 to L.H.Z. and J.A.B. productivity in the United States. Remote Sensing of Environment 89, 434–443. IPCC, 2007. Summary for Policymakers. In: Solomon, S., Qin, References D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., H.L. Miller, M. (Eds.), Climate Change 2007: Aphalo, P.J., Jarvis, P.G., 1991. Do stomata respond to relative The Physical Science Basis. Contribution of Working Group I humidity? Plant, Cell and Environment 14, 127–132. to the Fourth Assessment Report of the Intergovernmental Barradas, V.L., 1991. Air temperature and humidity and human Panel on Climate Change. Cambridge University Press, comfort index of some city parks of Mexico City. Interna- Cambridge, United Kingdom and New York, USA, pp. 18. tional Journal of Biometeorology 35, 24–28. Koerner, B., Klopatek, J., 2002. Anthropogenic and natural CO2 Berry, R.D., Colls, J.J., 1990. Atmospheric carbon dioxide and emission sources in an arid urban environment. Environ- sulphur dioxide on an urban/rural transect-I. Continuous mental Pollution 116, S45–S51. measurements at the transect ends. Atmospheric Environment McKee, D.J., 1994. Tropospheric Ozone: Human Health and 24A, 2681–2688. Agricultural Impacts. Lewis Publishers, Boca Raton, pp. 333. Brazel, A., Selover, N., Vose, R., Heisler, G., 2000. The tale of Nasrallah, H.A., Balling, R.C., Madi, S.M., Al-Ansari, L., 2003. two climates–Baltimore and Phoenix urban LTER sites. Temporal variations in atmospheric CO2 concentrations in Climate Research 15, 123–135. Kuwait City, Kuwait with comparisons to Phoenix, Arizona, Burian, S.J., Shepherd, J.M., 2005. Effect of urbanization on the USA. Environmental Pollution 121, 301–305. diurnal rainfall pattern in Houston. Hydrological Processes Oke, T.R., 1982. The energetic basis of the urban heat island. 19, 1089–1103. Quarterly Journal of the Royal Meteorological Society 108, 1–24. Day, T.A., Gober, P., Xiong, F.S., Wentz, E.A., 2002. Temporal Oke, T.R., Maxwell, G.B., 1975. Urban heat island dynamics in patterns in near-surface CO2 concentrations over contrasting Montreal and Vancouver. Atmospheric Environment 9, 191–200. vegetation types in the Phoenix metropolitan area. Agricul- Pataki, D.E., Alig, R.J., Fung, A.S., Golubiewski, N.E., tural and Forest Meteorology 110, 229–245. Kennedy, C.A., McPherson, E.G., Nowak, D.J., Pouyat, Denman, K.L., Brasseur, G., Chidthaisong, A., Ciais, P., Cox, R.V., Romero Lankao, P., 2006. Urban ecosystems and the P.M., Dickinson, R.E., Hauglustaine, D., Heinze, C., Hol- North American carbon cycle. Global Change Biology 12, land, E., Jacob, D., Lohmann, U., Ramachandran, S., da 2092–2102.  ARTICLE IN PRESS K. George et al. / Atmospheric Environment 41 (2007) 7654–7665 7665 Reid, K.H., Steyn, D.G., 1997. Diurnal variations of boundary layer Vogt, R., Christen, A., Rotach, M.W., Roth, M., Satyanarayana, carbon dioxide in a coastal city–observations and comparison A.N.V., 2006. Temporal dynamics of CO2 fluxes and profiles with model. Atmospheric Environment 31, 3101–3114. over a central European city. Theoretical and Applied Taha, H., Akbari, H., Rosenfeld, A., 1991. Heat Island and Oasis Climatology 84, 117–126. effects of vegetative canopies: micro-meteorological field- Ziska, L.H., Gebhard, D.E., Frenz, D.A., Faulkner, S., Singer, measurements 44, 123–138. B.D., Straka, J.G., 2003. Cities as harbingers of climate United Nations, 2004. World Urbanization Prospects: The 2003 change: Common ragweed, urbanization and public health. Revision. United Nations: New York. Journal of Allergy and Clinical Immunology 111, 290–295. Velasco, E., Pressley, S., Allwine, E., Westberg, H., Lamb, B., Ziska, L.H., Bunce, J.A., Goins, E.W., 2004. Characterization of 2005. Measurements of CO2 fluxes from the Mexico an urban/rural CO2/temperature gradient and associated City urban landscape. Atmospheric Environment 39, changes in initial plant productivity during secondary 7433–7446. succession. Oecologia 139, 454–458.