Vol.5 No. 1 May, 2012 FluxLetter The Newsletter of FLUXNET

In this issue of the FLUXLETTER, we profile a few of the sites that are using the covariance meth- od to determine the carbon, water and energy balance in urban across the globe. These sites document very different sources and sinks of carbon and energy than are found in natural systems. In this issue we profile sites in the UK, Poland and the U.S.

In This Issue: Flux measurements in urban ecosystems Flux measurements in urban Sue Grimmond and Andreas Christen ecosystems Sue Grimmond and Andreas Christen ……...….…….....Pages 1-8 The physical function- Land- ex- over an urban ing of a city – sometimes change in urban eco- (Figure 1c) is often domi- CO2, , and en- referred to as the ‘urban systems nated by fuel combustion ergy fluxes from ’ - requires a (C) from vehicles, industry vegetation and continuous intake of water, To formulate the ener- and buildings rather than Joe McFadden…..….Pages 9-13 food, fuel, materials and gy, water and carbon bal- the biological processes of power. The consumed re- ance equations over an ur- photosynthesis (P) and res- Urban flux measurements sources are converted into ban ecosystem, we include piration (Re). in Łódź, central Poland waste products, some of terms to describe the an- In addition to these Krzysztof Fortuniak, Włodzimierz which are injected into the thropogenic injections additional anthropogenic Pawlak, Mariusz Siedlecki, and urban atmosphere (i.e. (Figure 1). For example, terms, the partitioning of Mariusz Zieliński….Pages 14-20 waste heat, aerosols, pollu- the energy balance (Figure energy and water can be tants, greenhouse gases). 1a) contains a term to ac- affected dramatically by the Young Scientist Profile Many of the atmospheric count for the injection of specific construction mate- Simone Kotthaus….Pages 21-24 wastes affect health, weath- waste heat (anthropogenic rials, land cover and the Flux Measurements in er and climate at various heat flux QF). The urban three-dimensional form time and space scales – water balance (Figure 1b) (structure). For example, it Southern England, by hence they are important must include two additional is well documented that in KCL and CEH and must be incorporated terms - the water from irri- densely urbanized areas the Simone Kotthaus, Helen C. in weather, air gation / sprinkling (I) and heat storage in buildings Ward, Sue Grimmond and Jona- and climate models, often water vapour released by and fabrics can be sufficient than G. Evans….…Pages 25-31 with detailed spatial and fuel combustion (F). The to maintain an upward di- temporal resolution. carbon-dioxide exchange rected sensible heat flux Appendix throughout the night. This Sue Grimmond and Andreas Editor: Laurie Koteen, Producer: Dennis Baldocchi, can contribute to the gene- Christen……...….Pages 32-35 The FLUXNET Office at the University of California, Berkeley sis of the urban boundary Page 2

Flux measurements in urban ecosystems

Sue Grimmond and Andreas Christen

layer heat island and ena- bles better mixing of air pollutants. Another exam- ple is the presence of im- pervious surfaces and sew- er systems which increase the run-off term in the urban water balance rela- tive to the storage. Obvi- ously this affects storm water management and can cause more severe flooding if the limit of the system is Figure 1 - Conceptual representation of the urban (a) energy balance, (b) water balance, and (c) reached (higher peak dis- charge). CO budget for a balancing volume that reaches from the depth where no exchange with the sub- 2 Understanding how surface is found (z ) to the measurement height on a tower above the urban ecosystem (z ). The b t the partitioning of the ur- terms of the energy balance are - Q* : net all-wave radiation, QF : anthropogenic heat flux density ban energy and water bal- (electricity, fuel combustion, human metabolism), QH : sensible heat flux density, QE : latent heat ance equations is affected flux density, ΔQS : storage heat flux density (ground, buildings, air). The terms of the urban water by surface properties can balance are - p: Precipitation, F: water released by combustion, I : water injected by irrigation, E : help to mitigate some of evapotranspiration, ΔW : storage change in , Δr : infiltration, surface and subsurface run-off. the environmentally nega- tive effects of urbanization The terms of the urban CO2 budget are - FC : Net mass-flux of CO2 between urban surface and and we can use building atmosphere, ΔS : concentration change of CO in measurement volume, C : CO emitted by com- 2 2 and urban design practices bustion, Re : CO2 emitted by urban ecosystem respiration (soil, plants, humans), P : CO2 taken up to intentionally modify the by photosynthesis of urban vegetation (Modified from Feigenwinter et al. 2012). near-surface climates to our advantage – examples Table 1 - Distribution of the urban flux measurement sites recorded in the Urban Flux are albedo control, or the Network database based on the location of the cities in global climate zones. The bias to use of vegetation. sites to temperate mid latitude climates is evident. There is a clear gap on measurements on urban-atmosphere exchange in semi-arid and tropical climates. Why are urban flux towers needed? Fraction of No of urban flux population towers Similar to the overall aims of FLUXNET to Cities in tropical climates (A) 28% 3 quantify controls on land- Cities in arid and semiarid climates (B) 14.5% 3 atmosphere exchange of Cities in temperate climates (C) 44.6% 40 energy, water and carbon, flux towers in cities allow Cities in continental climates (D) 12.4% 15 us to infer the anthropo- Cities in polar climates (E) 0.3% 0 genic injections and/or the modified partitioning for a Total 100% 61 specific urban area. Unlike Page 3

Flux measurements in urban ecosystems

Sue Grimmond and Andreas Christen in non-urban areas, the and parameterize constants (Dabberdt and Davis 1978, rently 60 urban observa- determination of annual used in numerical models, Ching et al., 1983) and tion programmes from Net Ecosystem Exchange is such as the albedo, the studies in Vancouver (Yap different cities worldwide. rarely the goal, although a roughness length, or con- and Oke 1974, Yap et al. Current representation number of studies have stants to describe turbulent 1974). The first CO2 flux of urban ecosystems in quantified this for small transfer using similarity observations above urban FLUXNET components of the urban theory over urban surfaces. ecosystems were operated ecosystems (the specific Micrometeorological for only short periods Urban areas are very tower source area). The observation sites have been (Grimmond et al. 2002, varied. Land use objectives of researchers installed in cities across the Nemitz et al. 2002). While (residential, industrial, operating urban flux tow- world, (Table 1), howev- the number of sites have commercial) and the asso- ers include measuring sur- er, given the different ob- increased through time ciated land cover (surface face-atmospheric exchang- jectives compared to (Figure 2), there are still materials), structure es of energy, water and or grassland systems, only relatively few sites that (architectural styles), and greenhouse gases to evalu- a small number of sites have been operational for urban metabolism (e.g. ate or verify urban land have been operating for multiple-years or are on- traffic patterns, population surface models under multiple years and many going. Here we provide a densities, energy consump- known conditions. Other towers were operated for brief overview of urban tion) all vary between and studies have used flux data short periods (i.e. a few observation sites and draw within cities. Most obser- to evaluate emission inven- weeks). Among the earliest attention to the interna- vations have been conduct- tories (greenhouse gases, flux measurements of mo- tional Urban Flux Network ed in relatively dense pollutants, anthropogenic mentum and energy were database that contains ur- (compact) settings with heat flux). Further, da- those conducted in St Lou- ban flux towers. The Ur- mid-height attached or tasets from urban sites have is as part of the ban Flux Network lists detached buildings, or in been used to characterize METROMEX study information on the cur- open low-density areas with detached buildings (Figure 3). The simplest way to characterize the urban sur- face is the division between vegetated (pervious) and impervious (buildings and paved ground) plan areas. Figure 3 provides some indication of the range of urban settings in which urban flux measurements have been conducted in the past. The range of surface cover, structure, but also the urban metabolism Figure 2 - Number of active urban flux sites active per year (1990-2012) and measured (including cultural- turbulent fluxes. Source of data: Urban Flux Network database (May 2012). economic factors) result in Page 4

Flux measurements in urban ecosystems

Sue Grimmond and Andreas Christen a large range of energy partitioning, water availa- bility and greenhouse gas emissions between cities and within cities. One overarching goal of urban flux measurements is to link the partitioning and magnitude of fluxes not only to environmental and cultural controls but also to the nature of the sur- face.

CO2 exchange studies in urban ecosystems

By way of example, (Figure 4) shows ‘fingerprints’ of the meas- ured CO2 flux density for two contrasting urban eco- systems as a function of Figure 3 - The nature of the urban areas where flux measurements have been undertaken. The time of year and time of sites are characterized by the percentage of the plan area that is covered by building footprints, day. The left case shows impervious ground (streets, parking lots, sidewalks) and vegetation (including open water and observations conducted bare soil in some cases); the continent where the sites are located (symbol shape); and the Local over 6 years in suburban Climate Zone (colour). The text associated with each symbol indicates the city, the year of ob- Baltimore, USA, character- servations and sometimes the site within the city. For more details about the site codes see Ap- ized by extensive tree cov- pendix. Numbers underneath each legend entry refer to the total number of urban flux sites er and lawns (73%). The reported from this continent / Local Climate Zone (Note: due to lack of cover data, not all sites flux density of CO2 shows in the database are shown on the graph). Modified based on Christen et al. (2009) clear evidence of carbon uptake during the daytime from Basel, Switzerland. when space heating con- surprisingly, with increas- in the growing season and This is a site with little tributes to larger overall ing urban density, the ob- the fingerprint does not vegetation in the footprint emissions throughout the served emissions increase. substantially differ from (24%) and a clear domina- day (Vogt, 2009). Only the highly vegetated similar graphs over forest tion of transportation Figure 5 shows the low-density suburban areas ecosystems, although one emissions (morning and relationship between urban of Montreal (Mo08s), Bal- can note the overall higher evening rush hour). Small- density (expressed as plan timore (Bm02) and a park emissions in winter due to er fluxes during the day- area fraction of buildings) in Essen (Es07[p]) remain space heating (Crawford et time are more evident in and daily total summertime weak sinks, where the mid al, 2011). The right case the warm period/growing fluxes of CO2 measured on -summer GPP offsets any summarizes 5 years of data season than in the winter 20 urban tower sites. Un- local anthropogenic emis- Page 5

Flux measurements in urban ecosystems

Sue Grimmond and Andreas Christen

sions. Note that the rela- which typically cause CO2 -site comparisons are the tower (compare to Va08s tion in Figure 5 shows den- emissions outside the tow- fact that traffic emissions in [r] which shows the same sity vs. per-area fluxes of er source area. Only rarely the source area - by their dataset conditionally re- the urban ecosystems and is there a power plant in nature - come from mobile stricted to a sector does not imply that per- the source area of an urban vehicles which are not easi- without arterial roads). A capita emissions also in- flux tower (Sparks and ly attributable to local ac- corollary to this is the resi- crease with increasing den- Toumi 2010). Hence, the tivities vs. activities not dents within the tower sity. Also note that such actually measured CO2 linked to the area under source area leave the area inter-site comparisons are fluxes at a site are strongly investigation (e.g. through- (e.g. commuters). Thus, very limited, in particular affected by the mix of lo- traffic). For example, the the functional units of the when extended to the heat- cally emitting heating sys- ‘outlier’ Va08s[a] in Figure urban metabolism (‘carbon ing (winter) season. Part of tems (natural gas, coal, 5 is caused by the close footprint’ of a lifestyle / the energy required for wood) vs. externally emit- proximity of an arterial urban form) and the spatial space heating in buildings is ting heating systems through-road with 50,000 representation of emissions satisfied by electricity or (electricity, steam). An- vehicles per day that cross- (tower source area) are not district heating systems, other complication of inter es the source area of this coincident. Therefore the

analysis of CO2 flux data differs drastically from the ‘traditional’ forest and grassland sites in FLUXNET. Those compli- cations call for other forms of experimental control (e.g. comparison of week- end vs. weekday fluxes) or a spatial source area analy- sis linked to geographical information systems of local emission sources. Often in a city emission sources in the tower’s source area are geograph- ically well located and known, and fluxes meas- ured on urban flux towers can be a powerful tool to validate fine-scale emission models and infer emission factors (Velasco et al. 2005; Christen et al., 2011). Figure 4 - Carbon fluxes by time of day and time of year for a highly vegetated suburban Some cities have had ecosystem in Baltimore, USA (based on data from Crawford et al., 2011) and a dense ur- multiple CO2 measure- ban area in central Basel, Switzerland (based on data from Vogt et al., 2009). Page 6

Flux measurements in urban ecosystems

Sue Grimmond and Andreas Christen ment systems and flux Atmospheric allow This was sufficient to im- (Fortuniak et al., this news- towers simultaneously op- the effects of vast changes pact the annual total car- letter); and comparison is erated to study linkages in the urban structure over bon flux (Liu et al. 2012). made across the gradient across scales. For example, time to be studied. Most In this issue of the from forest to suburb to Baltimore, USA as an ur- recently Song and Wang newsletter there are three megacity city centre ban LTER site has had on- (2012) have been able to articles on urban flux (Kotthaus et al. this newslet- going flux measurements analyze the impact of spe- measurements: the CO2 ter). At all sites there are a since 2001 (Crawford et cific interventions on CO2 exchange in urban ecosys- year plus of urban CO2 al. 2011). The urban area exchanges, notably by look tems with an emphasis on flux measurements (and has been used as an end -ing at the change in CO2 vegetation and soils often energy and water point on CO2 and tempera- fluxes during the Olympic (McFadden, this newsletter); balances) and at each, new ture transects for ecologi- Games in 2008 when vehi- long term measurements in avenues of understanding cal response studies (Ziska cle numbers and factory the core of a city in Poland are profiled. et al. 2007). Montreal, production were reduced. are demonstrated Canada had a year-round operation of three sites with a density transect from the rural surrounding into the denser suburban and urban residential areas (Bergeron and Strachan, 2010). In London, UK, observations of CO2 con- centration have been con- ducted in the outskirts for a number of years (see, for example Lowry et al., 2001). More recently there have been flux meas- urements undertaken at three sites in central Lon- don: on the Imperial Col- lege London campus (Sparks and Toumi 2010), on the BT tower (Helfter et al. 2010, Wood et al. Figure 5 - Summertime carbon-dioxide fluxes measured in different urban ecosystems as 2010) and at the King’s a function of urban density (expressed as plan area fraction of buildings). The text associat- College London Strand ed with each symbol indicates the city, the year of observations and sometimes the site Campus (Kotthaus and within the city. Two sites are split-up into ensemble averages for different wind sectors: Grimmond 2012). Es07[p] is a sector containing a large urban park, while Es07[u] is the urban sector of this In Beijing, long-term site. Va08s[r] is a residential sector with only local roads, whereas Va08s[a] is data from all observations on a tall urban wind sectors that contains busy arterial roads. For more details about the site codes see tower at the Institute of Table 4. Modified and expanded based on Christen et al. (2009). Page 7

Flux measurements in urban ecosystems

Sue Grimmond and Andreas Christen

In the next issue the Anderson DE, Taggart J (1978) Bound.-Layer Mete- Goldbach A, Kuttler W history of urban flux meas- (2002) Proc. 4th AMS Symp. or., 14, 105-121. (2012) International Journal urements from the 1970s Urban Environ., Norfolk, VA. Dahlkötter F et al. of Climatology, DOI: to today at one of the old- 10.1002/joc.3437. est micrometeorological Balogun AA et al. (2009) (2010Meteorologische tower sites in Vancouver, Boundary-Layer Meterorology, Zeitschrift 19, 565-575. Helfter C, et al. (2011) Canada will be reviewed 1-23. Feigenwinter C, et al. Atmospheric and (Christen et al., next news- Bergeron O, Strachan IB (2012) Physics, 11, 1913-1928. letter). (2011) Atmos. Env., 45, measurements over urban King T, Grimmond CSB 1564-1573. areas in Aubinet et al.. th Further resources (1997) 12 AMS Symp. on [Eds.] “Eddy Covariance - To learn more about Buckley SM, et al. (2011) Boundary Layers and Tur- A Practical Guide to Meas- all urban flux sites and re- Eos Trans. AGU, Fall Meet. bulence, 455–456. urement and Data Analy- cent publications (or if Suppl., Abstract B43B- sis”, Springer, 270 p. Kordowski K, Kuttler W your site is not listed in the 0295. (2010) Atmospheric Environ- Appendix – located at the Changhyoun P et al. (2010) Frey CM, et al. (2011) ment, 44, 2722-2730. end of the newsletter) then Atmos. Environ., 44, 2605- International Journal of Cli- visit the website http:// Lemonsu A et al. (2004). 2614. matology, 31, 218-231. urban-climate.com/wp3/ Journal of Applied Meteorolo- resources/the-urban-flux- Ching JKS et al. (1983) George K et al. (2007) gy, 43, 312–327. network. Recent synthesis Atmos. Environ., 41, 7654- Bound.-Layer Meteor. 25, Liu HZ, et al. (2012) papers on urban flux meas- 7665. 171-191. www.atmos-chem-phys- urements include Velasco Grimmond CSB et al. dicuss.net/ and Roth (2010), which Christen A et al. (2009) Eos Trans. AGU, 90(52), (1993) IAHS Publ., 212, 12 /7677 /2012/ discusses eddy covariance 165–174. Fall Meet. Suppl., Abstract Loridan T, Grimmond observations of CO2 fluxes in urban areas, and Loridan B31C-06. Grimmond CSB, et al. CSB (2012) J. Appl Me- teorol Climat, 51,219-241 and Grimmond (2012), Christen A et al. (2011) (1994) USDA FS General which provides an analysis Atmos. Environ., 45, 6057- Tech. Rep. NE-186, 41–61. Lowry D et al. (2001). J. of urban energy balance 6069. Grimmond, CSB, Oke TR Geophys. Res. , fluxes. A review of the (1995) J. Appl. Meteor., 34, 106, 7427-7448. techniques and challenges Coutts AM et al. (2007) 873–889. related to direct flux meas- Atmos. Environ., 41, 51-62. Masson V et al. (2008) urements in the urban en- and Atmospheric Crawford BR et al. (2009). Grimmond CSB et al. vironment can be found in Physics, 102, 135-157. Proc. of 8th Symp Urban (1996) Climate Res., 6, 45–57. the book “Eddy Covari- Environment, Phoenix, Matese A et al. (2009) J ance” edited by M. Aubi- Grimmond CSB Oke TR AZ. (1999J. Appl. Meteor., 38, Appl Meteorol Clim. 48, net, T. Vesala, and D. Pa- 1940-1947. pale in Chapter 16 Crawford BR et al. (2011) 922–940. (Feigenwinter et al., Atmos. Environ., 45. 896- Grimmond CSB et al. Moriwaki R, Kanda M 2012). 905. (2004) JGR - Atmospheres, (2004) “J. Appl. Meteorol. 109, D24101. 43, 1700-1710. References Dabberdt WF, Davis PA Page 8

Flux measurements in urban ecosystems

Sue Grimmond and Andreas Christen

Nemitz E et al. (2002) En- et Chaussées, 277, 5-18. King’s College London, viron. Sci. Technol., 36, Environmental Monitoring 3139–3146. Song T, Wang YS (2012) Atmos. Res., 106, 139–149, and Modelling Group, Newton T et al. (2007) 2012. Department of Geography, Geographic Annaler: 89 A London, United Kingdom (4) 331–347. Sparks N, Toumi R, (2010) ([email protected]) Atmos. Environ, 44, 5287- Andreas Christen Offerle B, et al. (2005) 5294 University of British Co- Journal of Climate, 18, 3983 lumbia, Department of Stewart I, Oke TR (2009) -3995. th Geography / Atmospheric 8 AMS Symposium on Science Program, Offerle B et al. (2006a) Urban Environment, Paper Vancouver, Canada Theor. Appl. Climatol., 84, J8.3. ([email protected]) 103-115. Velasco EVS et al. (2005) Offerle B, et al. (2006b) J Atmos. Envion. 39, 7433- Applied Meteorology, 45, 125 7446. -136. Velasco EVS, Roth M Oke TR et al. (1999) At- (2010) Geography Compass, mos. Environ., 33, 3919– 4/9, 1238–1259. 3930. Velasco E et al. (2011) Pawlak W et al. (2011) Theoretical And Applied Cli- International Journal of Cli- matology, 103, 501-517. matology, 31, 232-243. Vesala T et al. (2008) Tellus Peters EB, McFadden JP 60B: 188-199.

(2012), Journal of Geophysi- th cal Research-Biogeosciences, Vogt R (2009) 7 Int. Conf. In review. on Urban Climate, Yoko- hama, Japan, Ramamurthy P, Pardyjak ER (2011) Atmos. Environ, Vogt R et al. (2005). Theor 45, 73-84. Appl Climatol. 84,117-126 Rigby M et al. (2008) At- Yap D et al. (1974) J. Appl. mos. Environ. 42, 8943- Meteorol. 13, 40-45. 8953. Yap D, Oke TR (1974) J. Roth M et al.(2009). 8th Appl. Meteorol. 13, 880-890. AMS Symp on the Urban Environment, Phoenix, AZ Contact

Ruban V et al. (2011) Bul- Sue Grimmond letin du Laboratoire des Ponts Page 9

CO2, water vapor, and energy fluxes from urban vegetation and soils

Joe McFadden

Urban vegetation function that have relatively high sinks and sources in devel- ropolitan region. Our goal vegetation cover and that oped land are poorly is to develop a mechanistic The role of vegetation make up the largest area known, as are the their sea- understanding of the spatial in the micrometeorology of fraction of most metropoli- sonal dynamics and spatial and temporal drivers of urban areas, particularly its tan areas. For example, in patterns in different types these fluxes in the built aerodynamic, thermal, ra- Minneapolis–Saint Paul of cities and climates. Wa- environment and to provide diative, and evaporative where our study site is lo- ter vapor fluxes in urban the information needed to functions, has long been cated, 51% of the total land areas are of interest because improve the representation recognized by the urban area is residential while of their potential benefits of vegetation function in climate community (Oke, 12% is commercial, institu- for reducing storm water urban land surface schemes 1989). In contrast, studies tional, and other built-up runoff and the urban heat and biogeochemical mod- of urban CO2 fluxes began land use in the city center. island effect, and because of els. Here, I want to provide only during the past decade The net uptake of CO2 their potential costs associ- an overview and some key (Grimmond et al., 2002), by urban vegetation and ated with the use of irriga- results of a set of inter- and our understanding of soils is of interest not be- tion in arid climates, both linked studies over a subur- how different processes cause it can offset the fossil- of which affect energy use ban landscape at the (e.g., photosynthesis and fuel CO2 emissions of cities and, therefore, CO2 as KUOM tall tower. respiration by vegetation, (unfortunately, far from it), well. emissions from fossil fuel but rather because devel- My group is studying KUOM tower site combustion) contribute to oped land is a large, and how different components the net CO2 flux above ur- often rapidly growing, frac- of the urban vegetation and The study site is locat- ban areas lags behind the tion of the land use in most soil system contribute to ed in a suburban residential state-of-the-art for natural regions over which we CO2, water vapor, and en- area (Fig. 1) of the twin or agricultural environ- want to construct carbon ergy exchanges from the cities of Minneapolis–Saint ments. This is particularly budgets. Yet the magni- scale of an urban or subur- Paul, along the Mississippi important for residential tudes of vegetation carbon ban neighborhood to a met- River in east central Minne- and suburban landscapes

Figure 1. View to the northeast from the 150 m level of the KUOM tall tower in winter (left) and summer (right). Page 10

CO2, water vapor, and energy fluxes from urban vegetation and soils

Joe McFadden sota, USA (45°00′ N, 93° golf course, corresponding mate with a mean annual 150 m tall KUOM broad- 11′ W, 300 m a.s.l.). to Urban Climate Zone 5 (1970–2000) temperature cast tower. Eddy covari- This first-ring suburb is (Oke, 2004). It has ap- of 7.4°C (ranging from a ance systems were installed immediately outside the proximately 1000 inhabit- January mean of −10.5°C at 40 and 80 m on the tow- Saint Paul boundary and is ants km–2 and a housing to 22.9°C in July) and pre- er to permit investigation situated about midway be- density of 350 units km–2. cipitation of 747 mm. Re- of footprint effects, and a tween the city centers of The plan area fractional search at the site began in four-component radiome- Minneapolis and Saint Paul cover within the residential 2004 with an urban forest ter (CNR1) was installed at 150 m to obtain the largest possible footprint for radia- tive fluxes. The measure- ment heights were greater than twice the average 12- m height of the trees, which overtopped the roofs of the houses, so we could assume the measurements were obtained above the roughness sublayer. The

mean roughness length z0 was 1.2 m, a typical value for a suburban site. Each eddy covariance system consisted of a 3‐D sonic (Campbell CSAT3) and a closed‐path infrared gas analyzer (LI-Cor LI‐7000). Digital signals from the Figure 2. The typical configuration of streets, buildings, and vegetation at street level. sonic were transmitted to the tower (6 and 9 km away, respec- area is 43% tree canopy, inventory and biophysical base by fiber optic cables to tively). The study area ex- 34% turfgrass lawns (total measurement program, avoid RF interference from perienced rapid residential of 77% vegetated), and and flux measurements the broadcast tower. Air development in the 1950s 22% impervious surfaces, were active from 2005 to was sampled through 9.5 on land that had been used although the total impervi- 2009. mm I.D. pure FEP tubing previously for nurseries ous fraction is greater at a rate of 23 SLPM and a and small vegetable farms. when buildings and sealed Local-scale flux meas- bypass flow of 7.5 SLPM The current land use surfaces under the tree urements through 4 mm I.D. pure within 1.5 km of the flux canopy are considered PTFE tubing was delivered tower consists of 1 or 2 (total of 34%). Fluxes at the scale of to the gas analyzers using a story single‐family, de- Minneapolis–Saint Paul the suburban neighborhood system of needle valves, tached houses (Fig. 2) and a has a cold temperate cli- were measured from the mass‐flow meters, and two Page 11

CO2, water vapor, and energy fluxes from urban vegetation and soils

Joe McFadden rotary vane vacuum pumps The dominant species in

(lag time of ~9 s). The gas the lawn were the C3 cool- analyzers, calibration tanks, season turfgrasses Ken- and data acquisition sys- tucky bluegrass (Poa praten- tems were housed in two sis), tall fescue (Festuca insulated, thermostatically arundinacea), and perennial controlled, heated and ven- ryegrass (Lolium perenne). tilated rack‐mount enclo- The site was not irrigated sures at the tower base. and it received one applica- tion of fertilizer per year. Component flux meas- The grass was mowed urements weekly to a height of 70 Turfgrass lawns. We con- mm, and the clippings currently operated a porta- were left in place to de- ble eddy covariance system compose. The site was rep- over a 1.5 ha turfgrass field resentative of low- located within the tall tow- maintenance lawns such as er footprint. It included a those found in a majority of CSAT3 sonic anemometer residential yards in our and an open-path LI-Cor LI study area. -7500 gas analyzer installed Urban trees. During 2007 at 1.35 m above ground, and 2008, continuous sap along with a set of radiom- flow and leaf-level gas ex- eters, soil heat flux, tem- change measurements of perature, and volumetric trees in the urban area (Fig. moisture content . 3) were made by Emily Figure 3. Emily Peters making leaf-level gas exchange measurements on trees in the tall tower footprint.

Peters for her Ph.D. pro- Service Forest Inventory ject. The measured trees and Analysis (FIA) proto- comprised seven different col. Biophysical measure- genera (Fraxinus, Juglans, ments, including LAI, sur- Picea, Pinus, Quercus, Tilia, face radiative temperature, and Ulmus) and included soil temperature, and soil evergreen needleleaf and moisture were monitored deciduous broadleaf plant across the urban forest functional types. They sampling grid throughout Figure 4. A section of the sampling grid used for urban were representative of the the growing season (Fig. forest and biophysical measurements. The yellow circles dominant canopy species 4). represent 7.3 m radius (167 m–2) sampling plots (N = 1168 that had been determined plots in total) that were randomly stratified within 1/1000th from an urban forest inven- Fossil fuel combustion. size subsamples of an EMAP unit (yellow hexagons). tory following U.S. Forest We collected a variety of Page 12

CO2, water vapor, and energy fluxes from urban vegetation and soils

Joe McFadden

data on CO2 emissions budgets for the suburban roadways nearest the tall mates based on emissions from combustion sources neighborhood. Vehicular tower. A series of 1–2 factors of the passing vehi- and human respiration to traffic, the largest contrib- month campaigns with a cles. constrain our bottom-up utor to short-timescale fast-response vacuum UV Some key results on estimates of CO2 exchange variations in urban CO2 resonance fluorescence the relative contributions by vegetation and soils, and exchange, was measured analyzer of individual vegetation so that we could eventually every 15 minutes using (AeroLaser AL-5002) types to local scale evapo- produce complete CO2 loop detectors under the were used to obtain the transpiration and net CO2 CO:CO2 flux ratio, as a exchange are shown in tracer of combustion- (Fig. 5). There were 2–4

derived CO2 fluxes. In fold differences in both addition, I am one of the fluxes among vegetation PIs on an interdisciplinary types and marked seasonal team that used a survey, variations associated with utility billing records, par- their physiology, such as cel data, remote sensing, the contrast between ev- and field measurements to ergreen and deciduous quantify monthly carbon trees, and the mid- emissions from home en- summer decline of CO2 ergy use and human respi- uptake that is characteris- ration across the metro- tic of cool-season C3 politan region (Fissore et turfgrasses. The total al., 2011, 2012). growing season (Apr– Nov) net CO2 exchange of Key results vegetation and soils in the One of our first tasks residential area was –124 g C m–2, while the total was to assess CO2 ex- change over the turfgrass evapotranspiration was lawn, including the poten- 324 mm, or 61% of pre- tial effects of point sources cipitation at the site. such as traffic. This work When fluxes from each was carried out by Rebec- vegetation type were ca Hiller, a visiting re- scaled up using a footprint searcher from Switzerland model and a detailed land who has since gone on to cover data set, they agreed complete her Ph.D. at well with measurements ETH Zürich. Hiller et al. from the tall tower (2011) developed an em- (Peters et al., 2011; Pe- ters and McFadden, Figure 5: Mean daily fluxes in 2008 of (a) evapotranspira- pirical approach using a 2012). tion and (b) net CO exchange per unit area of canopy cover footprint model and meas- 2 of deciduous broadleaf trees and evergreen trees measured ured traffic volume, and by sap flow and leaf-level gas exchange, and non-irrigated found that it agreed well turfgrasses measured by a portable eddy covariance system. with “bottom-up” esti- Page 13

CO2, water vapor, and energy fluxes from urban vegetation and soils

Joe McFadden

Current work Acknowledgments Further reading tive meteorological observations at urban sites. IOM Report More recently, Olaf We thank more than Fissore, C., L. A. Baker, S. 81, World Meteorological Menzer has joined my lab 350 homeowners and the E. Hobbie, J. Y. King, J. P. Organization, Geneva. after completing his mas- managers of two golf McFadden, K. C. Nelson, and I. Jakobsdottir (2011), Peters, E. B., R. V. Hiller, ter’s thesis in the model- courses for granting us per- Carbon, nitrogen, and phos- and J. P. McFadden (2011), data integration group at mission to conduct this phorus fluxes in household Seasonal contributions of the Max Planck Institute research on their property; ecosystems in the Minneap- vegetation types to suburban for in Je- Larry Oberg, chief engi- olis–Saint Paul, Minnesota, evapotranspiration, Journal of na. For his Ph.D. project, neer of KUOM for access urban region, Ecological Ap- Geophysical Research- among other things, he is to the tall tower; Brian plications, 21(3), 619–639. Biogeosciences, 116, G01003. evaluating machine learning Horgan for access to the methods for gap-filling the turfgrass research site; Fissore, C., S. E. Hobbie, Peters, E. B., and J. P. tall tower flux time series, Duane Schilder of Ramsey J. Y. King, J. P. McFadden, McFadden (2010), Influence a complex problem that County for traffic count K. C. Nelson, and L. A. of seasonality and vegetation requires time-varying driv- data; the parks and recrea- Baker (2012), The residen- type on suburban microcli- ers of both combustion and tion departments of Minne- tial landscape: fluxes of mates, Urban Ecosystems, 13 ecological fluxes, as well as apolis and Saint Paul, Rain- elements and the role of (4), 443–460. spatial information about bow Tree Care, and S & S household decisions, Urban Ecosystems, 15(1), 1–18. Peters, E. B., and J. P. how the flux footprint Tree Specialists for provid- McFadden (2012), Continu- changes with each 30- ing aerial lift trucks and Grimmond, C. S. B., T. S. ous measurements of net minute observation. operators for tree sam- King, F. D. Cropley, D. J. CO2 exchange by vegetation In 2008 my lab moved pling. The following for- Nowak, and C. Souch and soils in a suburban land- from Minnesota to U.C. mer undergraduate stu- (2002), Local-scale fluxes of scape, Journal of Geophysical Santa Barbara where, along dents and researchers from in urban Research-Biogeosciences, In with continued modeling the lab contributed to the environments: methodolog- press. and synthesis of KUOM field work and instrument ical challenges and results data, we have begun new operation: Ahmed Ba- from Chicago, Environmental Peters, E. B., J. P. McFad- den, and R. A. Montgomery studies of vegetation func- logun, Ben Freeman, Pollution, 116, S243–S254. (2010), Biological and envi- tion in semi-arid urban Vikram Gowreesunker, Hiller, R. V., J. P. McFad- ronmental controls on tree areas focusing especially on Cheyne Hadley, Nicole den, and N. Kljun (2011), transpiration in a suburban water use. Overall, despite Ifill, Vicki Kalkirtz, Rebec- Interpreting CO2 fluxes landscape, Journal of Geophys- the complexity and logisti- ca Koetter, Mark McPhail, over a suburban lawn: the ical Research–Biogeosciences, cal challenges of urban Julia Rauchfuss, Scott Shat- influence of traffic emis- 115, G04006. sites, the results thus far to, and Yana Sorokin. I am sions, Boundary-Layer Mete- have been encouraging that grateful to Dean Anderson orology, 138(2), 215–230. Contact we are beginning to make and Sue Grimmond for sense of the role that vege- their advice and encourage- Oke, T. R. (1989), The micro- Joe McFadden [email protected] tation and soils play in de- ment on the urban tall meteorology of the urban for- Department of Geography termining land–atmosphere tower setup, and to NASA est, Philosophical Transactions of and Earth Research Institute fluxes in urban and subur- for a New Investigator the Royal Society of London, Series B, 324, 335–349. University of California ban areas. grant that provided the Santa Barbara, CA 93106– major funding for this pro- Oke, T.R. (2004), Initial 4060 ject (NNG04GN80G). guidance to obtain representa- Page 14

Urban flux measurements in Łódź, central Poland

Krzysztof Fortuniak, Włodzimierz Pawlak, Mariusz Siedlecki, and Mariusz Zieliński

Introduction in Łódź since the 1960's, Long-term sites de- 45'45''N, 19°26'43''E, 204 but regular meteorological scription and instru- m a.s.l.). The building (17 The topographical set- measurements of the urban mentation m) is at least as high as sur- ting of the city, its size and -rural contrasts of the pa- rounding ones, so the building structure makes rameters started in 1992 Flux measurements in measurement height Ł ź Ł ź ód a good place for stud- (Kłysik & Fortuniak 1999, ód started in November (37 m) is close or just ies on modifications of local Fortuniak at al. 2006). Re- 2000 when Sue Grimmond above the blending height. climate by urbanization. It cently we have focused on and Brian Offerle (then of The site is located at the is located in the central the urban energy balance Indiana University) in- western edge of the old part of Poland on the big components and turbulent stalled the first open-path core of Łódź in dense de- European lowlands, and flux measurements in the eddy-covariance system at velopment, where build- with a population of about city. Hereafter we wish to the top of a 20 m mast, and ings are 7-13 m in height 750,000, it is among the share our experiences in mounted it on the roof of a and covers 15-40% of sur- biggest Polish cities. The this field and present se- university building at rounding area (Fig. 1). area is relatively flat, and Lipowa 81 str. (51° lected results from Łódź. Dark tarred roofs and as- only slightly inclined south- easterly (altitudes range from 180 m to 235 m a.s.l.). The absence of sig- nificant topographical fea- tures, such as lakes, rivers, valleys, mountains, or the sea allows urban effects to be discerned without inter- ference. In the city centre, buildings constructed about 100 years ago are mainly 15-20 m high and make up an extensive, fairly homog- enous and compact settle- ment of great density with clearly defined roof-level. The studies on urban cli- mate in Łódź have a long history, going back to the 1930's when a pair of urban -rural stations worked for a few years providing stand- ard meteorological data Figure 1: Area of investigation and location measurements sites. Solid lines surrounding needed for detecting of the the measurement point indicates source areas at p = 50, 75 and 90% calculated for turbu- singularities of urban cli- lent fluxes measured in unstable stratification (all available data selected for calculations). mate. After the war, the Black point at upper map indicates long-term measurements sites, blue dots – short exper- various elements of urban iments in years 2002-2003. Aerial photo source: Municipal Centre of Geodesics and Car- climate were investigated tographic Documentation of Łódź. Page 15

Urban flux measurements in Łódź, central Poland

Krzysztof Fortuniak, Włodzimierz Pawlak, Mariusz Siedlecki, and Mariusz Zieliński phalt roads dominate the east sector to 30-55% in In the first system, was measured by a thermo- artificial cover of the sur- the west (Fig. 2). The aero- turbulent fluxes were couple wire. The fast- rounding measurement dynamic roughness length measured by 3D sonic response data output was area, and result in low al- for momentum calculated anemometer (SWS- set as 10 Hz. Additional bedo in comparison to oth- from the logarithmic wind 211/3K Applied Technol- data included radiation bal- er cities (8%). Vegetation profile for close to neutral ogies, Inc.) and a krypton ance components cover depends on geo- stratification depends on (KH20 Camp- (measured by Kipp and graphic sector and varies flow direction, with an bell Scientific). Moreover, Zonen CNR1), tempera- from 10-20% in the north- average of about 2 m. fast-response temperature ture and , wind speed and direction and precipitation obtained from slow-response sensors. The eddy-covariance system worked up to September 2003 providing data on the urban energy balance (Offerle et al. 2005, 2006a,b), but standard meteorological measure- ments have been continued since that date. A new eddy -covariance system has been working at the same place since July 2006. The system is equipped with

Li7500 infrared CO2/H2O gas analyzer enabling esti- mation of the net exchange of carbon dioxide in paral- lel with measurement of the energy balance compo- nents. Wind components and temperature fluctua- tions are measured by RMYoung 81000 sonic anemometer. The second long-term eddy-covariance tower was established in June 2005 on the roof of the building of the Faculty of Geographical Figure 2: System maintenance and view from the top of mast at Lipowa site toward the Science at Narutowicza 88 str. (51o 46'24"N, city centre (eastward). Page 16

Urban flux measurements in Łódź, central Poland

Krzysztof Fortuniak, Włodzimierz Pawlak, Mariusz Siedlecki, and Mariusz Zieliński

Data management In the post-processing data quality assessment, we The pre- and post- focus on verification of the processing analysis as well stationarity postulate; the- as data management are oretical requirements of similar for both sites and the eddy-covariance meth- to that of other eddy- od that require time series covariance groups. Fluxes to be stationary in the aver- are calculated for 1 hour aging period. Three sta- periods by simple block tionarity tests are used to averaging. This relatively check this postulate: the extended interval is moti- test proposed by Foken and vated by the possible ex- Wichura (1996) with a istence of relatively large critical value of RNFW = eddies generated above 0.3; the non-stationarity the rough, warm urban ratio, NR, given by Mahrt surface. A comparison of (1998) with a critical value fluxes calculated for a 1-h of NR = 2; and the relative period with a mean value covariance stationarity cri- of four consecutive fluxes terion introduced by Du- based on 15-min periods taur et al. (1999) and mod- shows that the second ified by Nemitz et al. method underestimates (2002) with a critical value fluxes by about 5%, owing of the relative covariance to spectral losses. Before stationarity coefficient, Figure 3: Example of three stationarity tests used in post- flux calculations are im- RCS = 0.5 (Fig. 3). The processing data quality control of sensible heat flux, QH : plemented, spike detec- problem, or questions that (See text for further explanation.) tion is carried out. All flux remain, are if the data calculations are made in should be accepted as a natural wind coordinates 'good' when it passes all 19o28'52"E, 221 m a.s.l.). 10 Hz) and are located on a with double rotations. three tests or only one of This site is located 2.7 km horizontal arm at the top of Due to the possible influ- them. In the case of sensi- east from the first one in a the mast, giving a measure- ence of separation, ble heat flux, only about slightly less compact settle- ments height of almost covariance is maximized one-third of the data passes ment. The buildings 42 m, which should be just for the range ±2 s. Final all three tests. Rejection of around the site are slightly above the roughness sub- fluxes are corrected to so much of the data set sig- taller (average about layer. Similarly to the account for the influence nificantly reduces the 17 m). The tower itself is Lipowa site, the roughness of humidity on sonic tem- amount of data available for 25 m and it is mounted on length for momentum is perature, absorp- further analysis (like esti- the roof of a 16 m building. angular dependent and has tion over the krypton's mation of the diurnal evo- The fast-response sensors an average close to 2 m. bandwidth and density lution of the energy bal- (RMYoung 81000 and effects (WPL – Webb- ance, or cumulative flux- KH20 krypton hygrome- Pearman-Leuning correc- es). On the other hand, as ter, sample at a rate of tion). it is shown in Fig. 3, much Page 17

Urban flux measurements in Łódź, central Poland

Krzysztof Fortuniak, Włodzimierz Pawlak, Mariusz Siedlecki, and Mariusz Zieliński of the ‘good-looking’ data at night does not pass any of the tests. The other problem related to quality assess- ment is verification of the postulate of well developed turbulence. The similarity relations of integral turbu- lence characteristics, nor- malised standard variations Figure. 4: Normalized standard deviation of vertical wind component as a function of sta- of wind components and bility. Black dashed line – fit to the Łódź data, orange dotted line –function given by temperature, are common- ly used to test this postu- Foken and Wichura (1996). late. Our results suggests further considerations the direct influence of the from these directions must that this criterion must be might be needed to ensure mast (flux sensors are be used with caution. Addi- applied with caution. The high quality flux data for placed on the boom at a tional support for this hy- universal functions for ur- urban areas. For example, location about 1 m east pothesis comes from the ban areas might differ from an analysis of the angular from the mast). Two oth- wind rose for the site. The those suggested for flat dependence of normalised er peaks are more difficult gaps are observed for direc- rural surfaces, so the func- variance of vertical wind to explain, but they corre- tions of the street azimuth, tion’s parameters must be components in neutral spond to the azimuth of the whereas typical wind roses verified before application conditions at the Lipowa streets passing next to the for the region are more to avoid excluding potential- site shows three evident tower building. So, our uniformly distributed. For ly appropriate data (Fig. 4). peaks (Fig. 5). The first speculation is that the street the Lipowa site, the same, In addition to the around the 270o wind di- canyons modify flow. non-typical wind roses are standard procedures some rection can be attributed to Therefore, data obtained recorded by two different anemometers for both sides of the mast, so a possible influence of the mast itself seems not to be the reason (Fortuniak et al. 2012).

Selected results

The relatively long term nature of our flux measurements allows for detailed studies of energy σ Figure 5: Normalised variance for vertical wind component, w/u­­*, by wind direction balance components, and a for neutral conditions (|ζ| < 0.05) (left), and wind frequency distribution (15o intervals) few examples are presented classified according to different wind speed (1 h average) (right) for Lipowa site (after: For- here. The most evident tuniak et al. 2012). feature of the energy bal- Page 18

Urban flux measurements in Łódź, central Poland

Krzysztof Fortuniak, Włodzimierz Pawlak, Mariusz Siedlecki, and Mariusz Zieliński ance in this and other urban thermal properties of a hours a day, the sum of the time, even during night areas is the relatively large city, e.g. the large heat the turbulent fluxes re- hours. Negative values of portion of energy partition- storage capacity. As in mains primarily positive. QE are extremely rare. ing devoted to sensible heat many other urban sites, Moreover, in Łódź we The long term meas- flux, QH (Fig. 6). In sum- during winter when Q* is observe positive latent urements of the carbon mer, at noon, this flux only positive for a few heat flux values most of dioxide flux, FCO2, reach about 150 W m-2 on demonstrate that the city is average at both sites, a significant source of this whereas latent heat, QE, gas. The annual course of -2 remains at 80-100 W m . FCO2 is the reverse of the Similar values of QH are temperature evolution. observed in spring, but the The monthly totals in win- lack of transpiration by ter are on the level of -3 -1 vegetation causes QE to be 1200 g m month . Such significantly lower. The high fluxes of CO2 in win- highest values of QH are tertime are a result of an- more than 300 W m-2 on thropogenic emissions from summer days. Conversely, domestic heating sources a few examples of negative and the high density of ur- values of QH occur ban traffic. In summer, throughout the day during both anthropogenic sources warm advection after a cold are reduced (no need for period. Extremes of QE are heating and smaller car hardly detectable. Under traffic during vacation), and fine weather, values can some amount of CO2 is reach about 150 W m-2. taken up by vegetation. As The maximum values are a consequence, the summer expected immediately fol- totals are about 500- lowing summer showers 700 g m-3 month-1. when rain falls on the hot The seasonal changes urban surface. However, determined by different open-path eddy-covariance anthropogenic (heating, systems do not provide traffic) and natural factors reliable data in such situa- (vegetation age, length of tions. Other characteristic the day) are also clear in results unique to built-up the diurnal pattern of FCO2 sites, are the positive value (Fig. 7). During winter, a of QH in late afternoon/ clear maximum of FCO2 is evening when the radiation observed between 9 am balance, Q*, turns nega- and 6 pm due to high emis- tive. This phenomena can sions of CO2 all day long. be attributed to the heat Figure 6: Average diurnal courses of energy balance com- This maximum vanishes in release from the city, and ponents at two measurement sites in Łódź calculated on the summer when only a associated with the altered the base of all available data from the period 2005-2010. small increase in emissions Page 19

Urban flux measurements in Łódź, central Poland

Krzysztof Fortuniak, Włodzimierz Pawlak, Mariusz Siedlecki, and Mariusz Zieliński

Figure 7: Mean diurnal course of FCO2 in seasons at Lipowa site calculated separately for full weeks, weekdays and week- ends for the period July 2006 to May 2011. is observed in morning rush long term towers, short observed in other studies. was observed above the hours. A separation of the observations were carried Sensible heat fluxes have a grassland, reaching week-ends and working out in the years 2002-2003 positive relationship with ‑2.9 g m-3 day-1 in the sum- days gives additional infor- (Offerle et al. 2006b, Paw- the extant of impervious mer of 2003 (the same mation about the net ex- lak et al. 2011). Three sites surface cover, and Bowen time mean flux at Lipowa ‑3 -1 change of CO2 in urban were located in the post- ratios show an inverse was 20.5 g m day ). In areas. On summer week- industrial district, residen- relationship with increas- residential areas, daily to- - ends, a minimum of FCO2­ tial area, and suburban air- ing vegetation cover. Net tals were about 12 g m 3 -1 is observed during sunlight port grassland. The results CO2 flux is lower at all day lower than at hours, which indicates a show the close adherence three sites than at Lipowa Lipowa and in post- significant CO2 uptake by of flux partitioning to the station. As expected, a industrial areas, about urban vegetation through surface characteristics, as negative (downward) flux 10 g m-3 day-1 lower. photosynthesis. The magni- Additional information on tude of carbon uptake al- turbulence and sensible most compensates for the heat flux over the centre of anthropogenic emission at Łódź is gained from the the sites, and therefore, the large aperture scintillome- net flux of CO2 drops to 0. ter (BLS 900, Scientec). Spring and Autumn are The emitter is located on characterised by higher the mast of the Lipowa site values of FCO2 during the 5 m below the top, and the day, even on the weekends. receiver is mounted on the In these seasons the reduc- roof of a high University tion of net flux on non- building 3.2 km east from working days is pro- Figure 8: Diurnal evolution of sensible heat fluxes calcu- the emitter, nearby nounced. lated from scintillometer (free convection assumption – Narutowicza site (Fig. 1). Other activity red line), measured at Lipowa (green), and Narutowicza Thus a measurement path (blue) sites. Mean value +/- standard deviation for the passes over the city centre. In addition to the two period 11.06.2010 -5.11.2010. We started the measure- Page 20

Urban flux measurements in Łódź, central Poland

Krzysztof Fortuniak, Włodzimierz Pawlak, Mariusz Siedlecki, and Mariusz Zieliński ments in August 2009, but central European city cen- K, Oke TR (2006a) Tem- there are many gaps in the Acknowledgements tre. Boundary-Layer Mete- poral variations in heat records because of different orol, submitted fluxes over a central Euro- technical problems (the Many thanks to Sue pean city centre. Theor building with receiver was Grimmond and Brian Of- Fortuniak K, Kłysik K, App Climatol 84: 103–115 renovated in 2010 and ferle who organised the Wibig J (2006) Urban– 2011), and measurement first flux measurements in rural contrasts of meteor- Offerle B, Grimmond CSB, limitations. The instrument Łódź and who have super- ological parameters in Fortuniak K, Pawlak W specifications for the BLS vised all experiments since Łódź. Theor App Clima- (2006b) Intra-urban differ- 900 indicate it can be used 2003. We are also very tol 84: 91–101. ences of surface energy for distances longer than grateful to Sue Grimmond fluxes in a central European our path, but it seems to be and Laurie Koteen for their Kłysik K, Fortuniak K, city. Journal of Applied too long for polluted urban assistance in text improve- (1999) Temporal and spa- Meteorology and Climatol- air. Especially during the ment. Funding for this re- tial characteristics of the ogy 45: 125–136 search was provided by the urban heat island of Łódź, winter, when air is close to Pawlak W, Fortuniak K, saturation, the low visibil- Polish Ministry of Science Poland. Atmospheric En- and Higher Education vironment 33: 3885–3895 Siedlecki M (2011) Carbon ity in the city centre reduc- dioxide flux in the centre es the signal below the (State Committee for Sci- Mahrt L (1998) Flux sam- of Łódź, Poland – analysis threshold. In other seasons, entific Research) under pling errors for aircraft of a 2–year eddy covariance scintillometry gives compa- grants 2P04E 041 28 (2005 and towers. Journal of measurement data set. Int J rable results to the eddy- -2007), N N306 276935 Atmospheric and Oceanic Climatol 31: 232–243, covariance method (Fig. 8), (2008-2010) and N N306 Technology 15: 416–429 DOI: 10.1002/joc.2247. even with the free convec- 519638 (2010-2011). tion assumption. References Nemitz E, Hargreaves KJ, Contact: In the near future we McDonald AG, Dorsey JR, intend to initiate measure- Dutaur L, Cieslik S, Carra- Fowler D (2002) Micro- Krzysztof Fortuniak ments of fluxes ra A, Lopez A (1999) The meteorological measure- [email protected] (undergoing purchase pro- detection of nonstationarity ments of the urban heat Włodzimierz Pawlak cedure of Li7700) at the in the determination of budget and CO2 emissions [email protected] Lipowa site. In addition, deposition fluxes. Proceed- on a city scale. Environ- since last autumn we have ings of EUROTRAC Sym- mental Science and Tech- Mariusz Siedlecki been conducting open-path posium ‘98, vol. 2. WIT Press: nology 36: 3139–3146. [email protected] eddy-covariance system Southampton, 171–176 Offerle B, Grimmond (including CO2) measure- Mariusz Zieliński ments on a typical Polish Foken T, Wichura B CSB, Fortuniak K (2005) mariusz.r.zielinski@ farmland, about 60 km east (1996) Tools for quality Heat storage and anthro- gmail.com from Łódź. The data from assessment of surface-based pogenic heat flux in rela- this system could be use as flux measurements. Agri- tion to the energy balance Department of Meteorolo- a reference to the urban cultural and Forest Meteor- of a central European city gy and Climatology, Uni- stations. This summer we ology 78: 83–105 centre. Int J Climatol 25: versity of Łódź, Poland 1405–1419 will start measurements Fortuniak K, Pawlak W, (energy balance, CO2, Siedlecki M (2012) Integral Offerle B, Grimmond CH ) at Biebrza wetland in 4 turbulence statistics over a CSB, Fortuniak K, Kłysik eastern Poland. Page 21

Young Scientist Profile: Simone Kotthaus

Where do I work? My cur- The measurement of rent research interests surface fluxes using eddy covariance towers is inher- Since 2009 I have been ently challenging in a high- part of the Micrometeoro- ly urbanized environment. logical research group of Urban-specific aspects of Prof Sue Grimmond at surface-atmosphere ex- changes, such as heteroge- King’s College London neous and complex surface (KCL), UK. I am now in cover and anthropogenic the final year of my PhD. activities, affect EC obser- My work uses two differ- vations at various scales. ent approaches to observe Micro-scale emissions of Simone Kotthaus the exchange of sensible heat, moisture and exhaust heat in urban areas. First, I gases are very common, especially from non- the building scale. An ex- am also contributing to the measure sensible heat flux residential buildings, and ample is presented here for latest development of the using the eddy covariance can have a distinct impact one of our sites, where LUCY model (Allen et al. (EC) method, with obser- on the observed turbulent micro-scale anthropogenic 2010) for anthropogenic vations of two flux towers surface fluxes. I developed sources are located east of heat flux estimates at vari- the tower (Figure 1). The situated in Central London a filter in order to identify ous scales (Lindberg et al. these emissions, which can impact of building scale 2012, in preparation). An- (see pages 25-31 of this anthropogenic heat flux on other challenge for the newsletter). Second, I use show up as explicit signals in high frequency EC time the sensible heat exchange analysis and interpretation surface remote sensing series (Kotthaus and Grim- is highest during daytime of EC measurements is the techniques. At the flux mond 2012). This new when the building is in use determination of the local sites, we are also collecting and values can exceed 70 scale footprint; the latter technique is used to ex- 2 long-term datasets of radia- clude the influence of mi- W m . This order of mag- being sensitive to surface tion and turbulent fluxes of cro-scale anthropogenic nitude is similar to results roughness/zero plane dis- from other studies (e.g. placement height. For the latent heat and carbon di- sources from the high- Iamarino et al. 2011) and urban surface, various pa- oxide (Kotthaus et al. frequency observations so that the resulting turbulent demonstrates the im- rameterizations have been 2012, in preparation). My fluxes can be analysed with portance of anthropogenic proposed which allow for PhD is supervised by Prof respect to their local scale emissions in the urban sur- the calculation of these Grimmond and Prof source area. We further face energy balance. parameters based on build- Wooster, with whom I am apply this filter to quantify Despite their relevance ing geometries and infor- working on the remote the anthropogenic contri- to urban climate, anthro- mation which can be ex- pogenic heat flux is still tracted from digital eleva- sensing analysis. bution to heat, moisture and carbon dioxide fluxes difficult to quantify, and tion models (Grimmond EC flux measurements at the building scale. This is modelling approaches are and Oke 1999). For our often used to address their sites in central London, I in the urban environ- one of the first attempts to directly observe the an- extreme spatial and tem- implemented a procedure ment and the role of poral variability. Besides which allows us to calcu- anthropogenic activi- thropogenic component of turbulent surface fluxes at our observational studies, I late roughness parameters ties: Page 22

Young Scientist Profile: Simone Kotthaus

is clearly evident in this against a high quality black figure. - I am planning to body system. During both, further explore the role of day- and night-time flights, shadows in urban areas in many volunteers were in- controlling surface temper- volved in ground truth ob- ature variability, which servations which I arranged again influences energy to support the airborne exchange via radiation and data collection. turbulent sensible heat Multi- and hyper- flux. I am studying the ef- spectral data collected fect of thermal anisotropy from these flights are used (Voogt and Oke 1998, i.e. to create land cover maps the diversity of surface with high spatial resolution temperature found in the and detailed information urban environment) based on surface material. Fur- on thermal imagery col- ther observations are con- lected with hand-held de- ducted to characterize Figure 1: Anthropogenic contribution to sensible heat flux vices. These are also used emissivity of urban materi-

QH at the building scale as observed by Kotthaus and Grim- in conjunction with ther- als using Fourier Transform mond (2012) at KSS site in Central London, by time of day mal data from airborne Infrared Spectroscopy and wind direction. measurements. Various (FTIR). These maps then observations were made serve as the basis for sur- and source area locations perature estimates to be during two airborne cam- face temperature calcula- iteratively with the analyti- used for sensible heat cal- paigns in central London tions from thermal infrared cal footprint model of Kor- culations. Due to the im- 2008 and 2011, in collabo- images. Further, I am mann and Meixner (2001). mense spatial heterogeneity ration with the Airborne working towards the pa- of the urban surface and its Research & Survey Facility rameterization of the aero- Surface remote sensing complex, three- (ARSF) of the British Na- dynamic resistance for heat in the urban environ- dimensional structure, we tional Environmental Re- in the dense urban area ment: need to consider the effects search Council (NERC). I around our flux sites. In on surface temperature as had the chance to specifi- conjunction with air tem- In addition to flux influenced by material cally design the observa- perature observations, I measurement and pro- properties, orientations of tions performed in 2011 to will be able to create maps cessing, I use various re- the surface facets (walls, match my research require- of sensible heat flux and mote sensing techniques in roofs) and shadows from ments, with ground truth then to compare the results my PhD research. This buildings and vegetation, observations supported by to EC observations. enables me to observe spa- among others. An example NERC’s Field Spectrosco- tially distributed sensible of the latter is illustrated by py Facility (FSF). The fan- How did I get here? heat flux based on the aero- a transect of surface tem- tastic cooperation with My way to urban cli- dynamic resistance method peratures across a road in both facilities allowed me mate research (Voogt and Grimmond Central London (Figure 2). to create different flight 2000). Currently, I am The cooling effect of vege- line scenarios, to adjust I came to London for working on the retrieval of tation and by the shadow sensor settings, and even to two reasons, to discover representative surface tem- cast by trees and buildings do a calibration of the air- the intriguing field of urban borne thermal imager climate, and also because of Page 23

Young Scientist Profile: Simone Kotthaus

the availability of a variety extensive observational of measurement techniques network in Benin and Ni- involved in my PhD. A ger, which involved a series combination of EC flux of field trips for instrument observations and creative maintenance and data col- application of remote sens- lection. This gave me the ing techniques was just chance to actually experi- what I was looking for. ence the different weather I studied Meteorology conditions we were aiming (Diploma degree, BSc + to understand. It was here MSc) at the Institute of that I started to work with and Meteorolo- a flux tower for the first gy at the University of Co- time, which then became logne, Germany. During one core aspect of my final this diverse course of Diploma thesis. My work study, I was exposed to the explored the response of various different aspects of surface energy exchange as and observed at a tree-savannah chemistry. Aside from my flux site to synoptic condi- formal studies, I had my tions in the year 2006. first practical encounter During this time I learned a with boundary layer mete- lot about the usage of ob- orology when I started to servational data, especially work as a student research EC processing. In addition, assistant at Forschung- I used model data szentrum Jülich, Germany, (ECMWF re-analysis), sat- where I analysed data from Figure 2: Thermal and visible image (TESTO hand-held ellite remote sensing imag- airborne air quality obser- thermal imager) of road in Central London (03/06/2011 es (Meteosat) and radio- vations. Soon I became at 13BST) with transect of temperatures along cross sec- sonde data for the analysis interested in the work of tion: indicated by black line in thermal image. Radiative of meso scale weather pat- the Tropical and Subtropi- temperatures in °C, not corrected for emissivity. terns and clouds. Our re- cal Meteorology research search group was responsi- group of Prof Andreas the dynamics of the West hurricanes hitting the US ble for the radiosonde data Fink. As a student research African climate and its in- East coast have their origin collection in Benin during assistant in his group, my terdependency with global in circulation patterns an extensive AMMA field work contributed to two circulation patterns is cru- (African Easterly Waves) campaign in 2006 (Parker projects concerned with cial for the improvement of forming here. At that et al. 2008). The experi- the West African Monsoon forecast abilities in an area time, some of my work ence gathered in West Af- (AMMA, www.amma- so highly dependent on was concerned with the rica laid the cornerstone international.org) and the scarce rainwater availabil- forecast potential of West for my interest in field climate of Benin in particu- ity. However, the impact African rainfall for Atlantic work, the practical use of lar (IMPETUS, of synoptic phenomena in hurricane activity (Fink et instruments and the analy- www.impetus.uni- this region is even more far al. 2010). Also, I was part sis of observational data. koeln.de). Understanding -reaching, given that some of the team operating an Finally, the opportunity to Page 24

Young Scientist Profile: Simone Kotthaus

learn more about remote Kormann, R. and F.X. ture. Int. J. Remote Sensing, sensing techniques and to Meixner, 2001: An Analyt- 19, 895-920. combine it with my ical Footprint Model for knowledge on flux meas- Non-Neutral Stratification. Contact urements brought me to Bound. Lay. Met., 99, 207– Simone Kotthaus London where I am now 224. [email protected] using both in the applica- Environmental Monitoring Kotthaus, S. and C.S.B. tion of urban climate re- and Modelling Group, Grimmond, 2012: Identifi- search. Department of Geography cation of Micro-scale An- King’s College, London, UK References thropogenic CO2, Heat and Moisture sources – Pro- Allen, L., F. Lindberg, and cessing Eddy Covariance C.S.B Grimmond, 2011: Fluxes for a Dense Urban Global to city scale urban Environment. Atmos. anthropogenic heat flux: Env., 10.1016/ model and variability, Int. j.atmosenv.2012.04.024 J. Climatol., 31, 1990– 2005. DOI: 10.1002/ Parker, D.J., Fink, A., joc.2210 Janicot, S., Ngamini, J.-B., Douglas, M., Afiesimama, Fink, A. H., J. M. Schrage E., Agusti-Panareda, A., and S. Kotthaus, 2010: On Beljaars, A., Dide, F., the Potential Causes of the Diedhiou, A., Lebel, T., Nonstationary Correlations Polcher, J., Redelsperger, between West African Pre- J.-L., Thorncroft, C.D, cipitation and Atlantic Hur- Wilson, G.A. 2008: The ricane Activity. J. Climate, AMMA radiosonde pro- 23, 5437–5456. gram and its implications for the future of atmospher- Grimmond, C. and T. ic monitoring over Africa, Oke, 1999: Aerodynamic BAMS. 89, 1015-1027. properties of urban areas derived from analysis of Voogt, J. and C.S.B. Grim- surface form. J. Appl. Me- mond, 2000: Modeling teor., 38, 1262-1292. surface sensible heat flux using surface radiative tem- Iamarino, M., S. Beevers, peratures in a simple urban and C.S.B. Grimmond, area. J. Appl. Meteor., 1681, 2011: High-resolution 1679-1669. (space, time) anthropogen- ic heat emissions: London Voogt, J. and T. Oke, 1970-2025. Int. J. Climatol. 1998: Effects of urban sur- doi: 10.1002/joc.2390 face geometry on remotely -sensed surface tempera- Page 25

Urban Flux Measurements in Southern England, by KCL and CEH

Simone Kotthaus, Helen C. Ward, Sue Grimmond and Jonathan G. Evans

The micrometeorology research group at King's College London (KCL) and the Centre for and (CEH) are op- erating flux sites in differ- ent urban environments in southern England. While the KCL sites are located in a highly urbanized area in Central London, CEH fo- cuses on suburban charac- teristics in the town of Swindon, situated 120 km west of London (Figure 1). The population density of southern England is high, particularly in the south Figures 1 & 2: Locations of Swindon, Wytham Woods and Greater London in the Unit- east (population 8,523,074 ed Kingdom (1) and flux towers in London, KSK & KSS (2). in mid-2010, excluding London, Office for Nation- al Statistics, ONS), and still convective activity, the exceeding 7.8 million in cover is clearly dominated rapid population increase is area is still relatively dry 2010 (ONS) in an area of by anthropogenic materi- anticipated. (compared to the rest of 1572 km2. The KCL flux als. Due to the proximity The climate is charac- the UK) with the lowest sites are located at the KCL to the River Thames, 14% terized as temperate ma- amounts of precipitation Strand Campus (Table 1), of the surface is open wa- rine, with the influence of occurring in the Thames just north of the River ter, leaving only about 5% mid-latitude cyclones Valley and the London area. Thames, in a very dense to vegetation (grass and crossing the UK from West On average, annual rainfall urban environment in Cen- street trees). to East. Passages of the in central and South East tral London. Loridan et al. Two measurement frontal systems account for England is 777 mm. The (2012) classified the sur- towers are installed above precipitation, clouds and mean minimum tempera- rounding area as High Den- roofs (Figure 2). First, ob- increased wind speeds; the ture is 1.2°C in February, sity UZE (Urban Zone for servations started at the prevailing wind direction is and the mean maximum Energy partitioning, Lori- KSK site in October 2008 from the south west. How- temperature is 21.7°C in dan and Grimmond, 2012) using a single tube mast ever, the easternmost part July. and according to an image- (triangular tower, Clark of the British Isles, is rela- based classification ap- Masts CSQ T97/HP). In tively sheltered diminishing KCL flux sites in Cen- proach (Stewart and Oke, November 2009 the second the impact of low pressure tral London 2009), it can be described site (KSS), approximately systems and since it is close as compact midrise LCZ 60 m to the north of the to the European mainland, The UK capital is a (Local Climate Zone). first site, became opera- continental weather condi- growing metropolitan area With 43% impervious tional. Here instruments tions can extend to this with the population of ground and 38% buildings, are mounted on top of an area. Despite enhanced Greater London already the surrounding surface extendable triangular tow- Page 26

Urban Flux Measurements in Southern England, by KCL and CEH

Simone Kotthaus, Helen C. Ward, Sue Grimmond and Jonathan G. Evans er (Aluma T45-H). The equipped with eddy covari- gap-filled (Kotthaus and tests comprise a new latter was moved to the ance (EC) systems (see list Grimmond, 2012). Eddy despiking approach and an west end of the building in of instrumentation in Table covariance fluxes are calcu- automatic Identification of March 2012, forming the 1), allowing for the obser- lated using ECpack (van Microscale Anthropogenic third KCL flux site in Lon- vation of turbulent ex- Dijk et al., 2004) with a Sources (IMAS, Kotthaus don referred to as KSSW change of sensible heat, series of pre- and post- and Grimmond, 2012). The (Figure 3). Measurement latent heat and carbon di- processing steps which are latter is used to filter effects height a.g.l. is 39 m at oxide. EC measurements implemented in order to of micro-scale emissions so KSK, 49 m at KSS and 50.3 are sampled at a frequency improve the quality of the that the calculated fluxes m at KSSW, which is of 10 Hz and fluxes are results and to maximise are representative for the equivalent to a ratio of 1.9, calculated based on 30 min data availability. A detailed local scale source area. 2.2 and 2.3 compared to block averages. Meteoro- description of the data pro- mean building height zH, logical observations from cessing procedure is pre- CEH flux measure- respectively. the and the sented by Kotthaus and ments in Swindon All three sites are rain gauge are cleaned and Grimmond (2012). These A key aim of the Swindon field campaign is to provide Swindon KSS (KSK/KSSW), energy, water and carbon Central London fluxes for a suburban envi- Location 51°35' N 1°48' W 51°30' N, 0°7' W ronment within the UK. Suburban Central business district Datasets are fairly rare for Climate conditions Temperate marine (mid-latitude) urban environments in gen- Period of operation May 2011 – present Oct 2008 – present eral and UK campaigns Observed fluxes Q*, QH, QE, FC Q*, QH, QE, FC have tended to focus on Instrumenta- Sonic R3, Gill Instruments CSAT3, Campbell Scientific heavily built areas. Never- tion Gas analyser Li7500, LiCOR Biosciences Li7500/Li7500A, LiCOR theless, suburban areas are Biosciences an important land use and Radiometer NR01, Hukseflux CNR1/CNR4, Kipp&Zonen house over 80% of the UK Weather station WXT 510, Vaisala WXT510/WXT520, population (Gwilliam et Vaisala al., 1998). The mosaic na- Rain gauge Tipping Bucket, Casella ARG100, Campbell Scientific ture of surface cover found UCZ (Local Climate Zone) open lowrise compact midrise in these areas, comprising Surface cover Impervious ground 40% 43% green space, buildings, a Buildings 19% 38% mixture of pervious and Vegetation 36% 5% impervious surfaces cou- Open water 0% 14% pled with human behaviour patterns and seasonal varia- Other 4% (gravel, bare soil) 0% tions in vegetation makes Mean building height zH (m) 5.5 21.9 the suburban environment Zero-plane displacement height zd (m) ~ 3.9 (0.7 zH) 14.2 a complex location to Roughness length z0 (m) ~ 0.5 (0.1 zH) 1.9 study, where many differ-

Sensor height zS (m) 12.5 49 (39, 50.3) ent natural and anthropo- genic factors combine. Table 1: Site characteristics and instruments used at the two urban sites Through observation and Page 27

Urban Flux Measurements in Southern England, by KCL and CEH

Simone Kotthaus, Helen C. Ward, Sue Grimmond and Jonathan G. Evans analysis of these factors and building height 5.5 m) and periods than EC. The EC presented (Figure 5) in the the resulting fluxes, im- vegetation forms a signifi- instrumentation is mount- form of median diurnal provements in understand- cant proportion of the sur- ed at 12.5 m above cycles (shading represents ing the physical process can face cover (37%), with ground, which is 2.3 times the inter-quartile range be made. In turn these can major contributions from mean building height (IQR)) to show how sur- help to support more in- gardens, green corridors (Figure 4, mast location face energy exchange varies formed decisions about and roadside grass or trees. marked in orange). Details with season in the two ur- future development in Being urbanised, a large of additional instruments ban areas in Southern Eng- terms of , the proportion of the ground is that appear at these sites land. Further, the compari- environment and human covered with impervious are presented in Table 1. son of the two locations health. materials, including public (London, KSS and Swin- Measurements are lo- roads, parking and play- Urban Surface Energy don) reveals how a differ- cated in the town of Swin- grounds as well as paved Balance ent degree of urbanization don (population 175 000), driveways and patios. influences the surface flux- one of the fastest growing In addition to the flux- Three components of es. This becomes especially towns in Europe. The land es measured using eddy the surface energy balance apparent in the residual of use within Swindon is typi- covariance (results present- are observed at both loca- the three energy fluxes cal of UK suburbia, com- ed here), there are three tions, London and Swin- (RES = Q*-QH -QE). The prising residential and scintillometers installed don, namely net all-wave interpretation of the resid- commercial areas, over Swindon, enabling radiation (Q*) and the tur- ual, however, should be transport links and green sensible and latent heat flux bulent fluxes of sensible done carefully because it space. The study area is estimates to be made that heat (QH) and latent heat includes all uncertainties, located north of the town are representative of a larg- (QE). Data collected in errors and differences be- centre; buildings are gener- er scale (2-5 km2) and po- summer 2011 (JJA) and tween the processing. All ally 1-2 storeys (mean tentially for shorter time winter 2011/12 (DJF) are four components exhibit a

100 m

Figures 3 & 4: Aerial photo (NERC ARSF) of KCL Strand Campus with locations of flux towers (3), and Aerial photo- graph (©GeoPerspectives) of Northern Swindon showing the location of the flux mast (orange star), and pictured right (4). Page 28

Urban Flux Measurements in Southern England, by KCL and CEH

Simone Kotthaus, Helen C. Ward, Sue Grimmond and Jonathan G. Evans clear diurnal pattern in the heat within the urban cano- responds to Q* with up- ment of Central London. two seasons. At both sites, py layer that are produced ward fluxes during daytime Here, turbulent sensible radiation is the main source by buildings, traffic and and downward heat heat flux is transporting of energy during summer human metabolism. Even transport during the night. heat upwards even during when median Q* reaches though it is still difficult to The latter is weaker in the the night, and the median up to about 400 W m-2 quantify, its impact is summer when the temper- values never drop below 50 during daytime. During clearly evident when look- ature gradient between W m2. Positive night time winter months net all-wave radiation is found to be higher in Swindon (daytime max. median 117 W m2) compared to London (80 W m2), which also holds true for night time values. This can be explained by larger outgoing long wave radiation fluxes in London. A higher degree of urbani- zation (expressed by for example,. a larger fraction of impervious surfaces with high heat capacities and deeper street canyons) fa- vours surface heat storage during daytime. The heat is then transported upwards, partly via radiative fluxes during night time (Christen and Vogt 2004). The small- er heat storage capacity of the suburban environment in Swindon further leads to Figure 5: Median diurnal pattern of surface energy balance components for summer (JJA) a weaker urban heat island and winter (DJF) at London (KSS) and Swindon: net all-wave radiation Q*, turbulent sen- effect so that the relatively sible heat flux QH, turbulent latent heat flux QE and residual RES. cooler surface exhibits less radiative heat loss, even ing at the turbulent fluxes surface and overlying air is sensible heat fluxes in win- during daytime. of sensible and latent heat. smaller. Here, anthropo- ter have been observed in In addition to radiative Daytime sensible heat flux- genic heat flux does not other cities (e.g. Helsinki, input and surface heat stor- es are higher during sum- seem to play a major role Järvi et al., 2009). Heat age, the anthropogenic heat mer than winter at both in determining the diurnal from anthropogenic emis- flux Qf, is an important sites, however pronounced patterns. In contrast, Qf sions maintains wintertime component of the urban differences appear in the significantly alters the sur- sensible heat fluxes larger surface energy balance. It London vs. Swindon com- face energy balance in the than the net all-wave radia- includes latent and sensible parison. In Swindon QH very dense urban environ- tion, even during daytime. Page 29

Urban Flux Measurements in Southern England, by KCL and CEH

Simone Kotthaus, Helen C. Ward, Sue Grimmond and Jonathan G. Evans

This persistently strong seasons. During winter, ment. Also in Swindon heat physical processes govern- sensible heat flux inhibits latent heat flux in Swindon is stored in the urban cano- ing the carbon flux (FC) in stable atmospheric stratifi- rises to the same order of py, but the anthropogenic each case (Figure 6). Each cation for most of the time. magnitude as the sensible heat flux would be ex- site has a visible diurnal In terms of energy parti- heat flux. The residual cal- pected to be of smaller pattern in both summer tioning, both the dense culated from the three en- magnitude. and winter; overall FC is urban and the suburban ergy balance components lower in summer for all environment show domi- summarizes the discussed Carbon Fluxes across sites. Looking at the three nant sensible heat flux dur- analysis. In London the different land uses sites together, it is possible to pick out similarities and differences that result from the relative importance of biogenic and anthropogenic activity. Wytham is essen- tially a natural environ- ment: an ancient woodland close to the River Thames near Oxford (51o46’ N 1o20’ W, see Thomas et al. (2011) for full details) where any impact of hu- man activity is negligible in comparison to the vegeta- tion. The woodland acts as a substantial carbon sink, assimilating 0.75 mol C m-2 day-1 for the summer peri- od shown here. In compar- ison, London acts a source of similar magnitude (0.68 Figure 6: Carbon fluxes for three case study days in summer and winter for the city cen- mol C m-2 day-1), whilst tre (London), suburban neighbourhood (Swindon) and woodland (Wytham). Swindon lies in between the two extremes and is a small ing summer month. How- sum of the turbulent fluxes source, emitting 0.06 mol ever, the impact of the exceeds energy supplied by Case studies of a few C m-2 day-1 for the three larger vegetation fraction radiation during night time example days comparing a days in summer shown (36%) in Swindon is obvi- in general and also during heavily urbanised city cen- here. For both seasons the ous where turbulent flux of daytime in winter; the re- tre (London), suburban release of CO2 increases as latent heat is about twice as sulting negative residual residential (Swindon) and the level of urbanization high as in London where includes the anthropogenic deciduous woodland increases and vegetative restricted moisture availa- heat flux and surface stor- (Wytham) in southern fraction decreases. bility favours sensible heat age, both of importance in England provide some in- Similar trends can be flux instead, during both this dense urban environ- sight into the relevant seen in the diurnal cycles. Page 30

Urban Flux Measurements in Southern England, by KCL and CEH

Simone Kotthaus, Helen C. Ward, Sue Grimmond and Jonathan G. Evans

London is always a source sources of CO2 during November 2000 (Nemitz sources of CO2 and roads of CO2, although much night time. For Wytham, et al., 2002), possibly due that are busy throughout smaller in summer (peak the night time release is to colder temperatures the day. values around 15 μmol C due to respiration (mainly leading to increased heating These case studies m-2 s-1) than in winter stem (Fenn et al., 2010)) demand in Edinburgh, im- demonstrate seasonal (peak values over 40 μmol and is generally larger in proved fuel efficiency and changes, highlight interest- C m-2 s-1). Swindon shows summer than winter due to different characteristics of ing features of the diurnal some carbon uptake during warmer temperatures. In the source area. In London cycles and exemplify differ- summer daytimes – evi- Swindon there is likely to the CO2 flux is highest dur- ences between natural, dence of significant photo- be a smaller contribution ing winter daytimes, when suburban and city centre synthetic activity by subur- from plant respiration (due human activity is at a maxi- environments. Overall the ban vegetation, accounting to the reduced surface cov- mum. In Swindon, the urbanised areas act as CO2 for 36% of surface cover. er) but a significant contri- night time fluxes are not sources, whilst the wood- During summer months, bution from human respi- very different between the land is a sink, and London when solar radiation and ration, which will be larger seasons but during winter emits far more CO2 per leaf area index are at a still in London. Human there is no uptake of CO2 – unit area during winter- maximum, the uptake by exhalation was found to in the middle of the day time than Wytham can vegetation almost balances account for 38% of the emissions from human ac- assimilate during summer. the anthropogenic and res- summer time CO2 flux in tivities cannot be offset by The presence of vegetation piration source terms here, Tokyo (Moriwaki and Kan- minimal photosynthetic and lower population den- but has a much smaller da, 2004), although more activity. sity in Swindon in the sum- effect in London where the densely populated than There is a clear traffic mer mean it exhibits trends vegetation fraction is small London. signal observable in the more comparable to rural

(5%) and population densi- It is during winter that wintertime CO2 flux from environments than those ty is much higher. At this patterns in human behav- Swindon. Peaks corre- seen in heavily built-up time of year the im- iour can be most clearly sponding to morning and London or Edinburgh, portance of gardens and seen for the urban loca- evening rush hour are pro- whereas during winter the urban green spaces for off- tions. The influence of veg- nounced; this is particular- largely dormant vegetation setting urban emissions can etation is minimal, as many ly evident for 23/01/12 plus increased emissions be seen on a daily basis in species are dormant and (Monday) but is much less from home heating and the Swindon. However the solar radiation is low marked at the weekends. A influence of traffic results presence of other (non- (illustrated by low daytime strong traffic signal would in fluxes dominated by vegetative) land uses and uptake / night time release be expected from this resi- anthropogenic activities. emissions from human ac- at Wytham), so that other dential suburban site, Comparison between these tivity (traffic, water heat- activities dominate the ob- where the roads can get three sites offers infor- ing, air conditioning, respi- served urban fluxes. Lon- busy with many people mation about the impact of ration) mean the uptake is don is a major source of travelling to and from human behaviour, the role much smaller than for the CO2, with building and work or school within the of urban vegetation and the woodland site (peak values water heating, traffic emis- rush hour periods. The response of the atmosphere less than 10 μmol C m-2 s-1 sions and respiration all rush hour signal is much to land use change and ur- in Swindon compared to contributing. For these weaker in London, usually banization. almost 30 μmol C m-2 s-1 study days, peak emissions only discernible in the for Wytham). in London are lower than morning, reflecting the All three sites are were found in Edinburgh in importance of other Page 31

Urban Flux Measurements in Southern England, by KCL and CEH

Simone Kotthaus, Helen C. Ward, Sue Grimmond and Jonathan G. Evans

Acknowledgements P. Aalto, R. Hillamo, T. nal Fluxes of Radiation, King’s College, London, UK Mäkelä, P. Keronen, E. Heat, Water Vapor, and [email protected] Thanks to Oxford Uni- Siivola, T. Vesala, and M. Carbon Dioxide over a versity for their collabora- Kulmala (2009) The urban Suburban Area. J Appl Me- Helen C. Ward tion with CEH in the oper- measurement station teorol. 43(11):1700-1710. King’s College London, ation of the Wytham flux SMEAR III: Continuous Environmental Monitoring tower site, and to the own- monitoring of air pollution Nemitz E, Hargreaves KJ, and Modelling Group, De- ers of the Swindon proper- and surface–atmosphere McDonald AG, Dorsey JR partment of Geography, ty who very kindly agreed interactions in Helsinki, and Fowler D (2002) Me- London, UK, and Centre to have the flux mast in- Finland. Boreal Env. Res., teorological Measurements for Ecology and Hydrology stalled in their garden. 14 (suppli. A), 86-109. of the Urban Heat Budget (CEH), Wallingford, UK Funding from NERC and Co2 Emissions on a [email protected] ClearfLO, NERC/ARSF, Kotthaus S and CSB Grim- City Scale. Environ Sci NERC and EUF7 BRIDGE mond (2012) Identification Technol. 36(14):3139- Sue Grimmond are gratefully acknowl- of Micro-scale Anthropo- 3146. King’s College London, genic CO2, Heat and Mois- Environmental Monitoring edged Stewart, I.D., and T.R. ture Sources - Processing and Modelling Group, De- Oke (2009) Newly devel- Eddy Covariance Fluxes for partment of Geography, References oped “thermal climate a Dense Urban Environ- London, UK zones” for defining and Christen, A. and R. Vogt ment Atmospheric Envi- [email protected] ronment 10.1016/ measuring urban heat is- (2004) Energy and Radia- land magnitude in the can- tion Balance of a Central j.atmosenv.2012.04.024 Jon Evans opy layer. Preprints, T.R. Centre for Ecology and European City. Int. J. Cli- Loridan T & CSB Grim- Oke Symposium & Eighth matol., 24,1395-1421. Hydrology (CEH), Wall- mond (2012) Characteriza- Symposium on Urban Envi- ingford, UK Fenn K, Malhi Y, More- tion of energy flux parti- ronment, January 11–15, [email protected] croft M, Lloyd C and tioning in urban environ- Phoenix, AZ. Thomas M (2010) Com- ments: links with surface seasonal properties Journal Thomas MV, Malhi Y, prehensive Description of Fenn KM, Fisher JB, More- the Carbon Cycle of an of Applied Meteorology and Climatology 51,219- croft MD, Lloyd CR, Tay- Ancient Temperate Broad- lor ME and McNeil DD leaved Woodland. Biogeo- 241 doi: 10.1175/JAMC- D-11-038.1 (2011) Carbon Dioxide sciences Discuss. 7(3):3735 Fluxes over an Ancient -3763. Loridan, T., F. Lindberg, Broadleaved Deciduous Gwilliam M, Bourne C, O. Jorba, S. Kotthaus, and Woodland in Southern Swain C and Pratt A C.S.B. Grimmond (2011): England. Biogeosciences. 8 (1998) Sustainable Renew- High resolution simulation (6):1595-1613. of surface heat flux varia- al of Suburban Areas. Jo- Contact seph Rowntree Founda- bility across central London tion, York: pp. with UZE urban zones, Simone Kotthaus submitted. Environmental Monitoring Järvi, J., H. Hannuniemi, and Modelling Group, De- Moriwaki R and Kanda M T. Hussein, H. Junninen, partment of Geography (2004) Seasonal and Diur- Page 32

Appendix: Flux measurements in urban ecosystems

Sue Grimmond and Andreas Christen

(a) Code City, Country LCZ Land- zH zm E C A O Study peri- Operated by Primary Reference Site-Name use (b) od on flux measure- (m) ments Bm02 Baltimore, Open R 5.6 41 ● ● 2001 - * Indiana Universi- Crawford et al. (2011) USA lowrise ty / UMBC/USFS (Cub Hill) Ba09 Basel, Switzer- Compact R C 16.7 40 ● ● 2009 - * University of land midrise Basel (Aeschenplatz)

Ba02s1 Basel, Switzer- Open R 13 16 ● 06/2002 - ETH Zürich / Christen and Vogt land lowrise 07/2002 University of (2004) (Allschwil) Basel

Ba04 Basel, Switzer- Compact R S O 17 39 ● ● 01/2004 - * University of Vogt (2009) land (Klin- midrise Basel gelbergstrasse) Ba02u2 Basel, Switzer- Compact R C 13 38 ● 08/1994 - University of Christen and Vogt land midrise 09/2002 Basel (2004) (Spalenring) Ba02u1 Basel, Switzer- Compact R C 15 31 ● ● 06/2002 - University of Vogt et al. (2006) land midrise 07/2002 Basel (Sperrstrasse) Bj08 Beijing, China Compact R C S O 30 47 ● ● ● ● 01/1991 - * Chinese Academy Song and Wang (IAP 325m highrise CO2 since of Sciences (2012) tower) 2006 Be05 Berlin, Germa- Compact R C N/A 125 ● 05/2005 - * Technical Univer- ny (Steglitzer midrise sity Berlin Kreisel) Ch92 Chicago, USA Open R 6.7 N/A ● Indiana University Grimmond et al. (Suburban) lowrise (1994) Ch95 Chicago, USA Open R 6.3 27 ● 06/1995 - Indiana University Grimmond et al. (Dunning) lowrise 08/1995 (2002) Dn01 Denver, USA Large R C I S O 7 60 ● ● ● 05/2001 - U.S. Geological Anderson and Taggart (South Denver) lowrise 09/2007; Survey (2003) 01/2011-* Db09 Dublin, Ireland Compact R C N/A N/A ● ● 08/2009 - * Nat. Univ. of (Marrowbone midrise Ireland Maynooth Lane) & UCD Ed00 Edinburgh, UK Mixed R C I 10.8 65 ● 10/2000 - CEH Edinburgh Nemitz et al. (2002) (Nelson Monu- 11/2000 ment) Es07 Essen, Germany Mixed R C 15 26 ● 09/2006 - University of Kordowski and Kuttler (Grugapark) 11/2007 Duisburg-Essen (2010) Page 33

Appendix: Flux measurements in urban ecosystems

Sue Grimmond and Andreas Christen

(a) Code City, Country LCZ Land- zH zm E C A O Study period Operated by Primary Reference Site-Name use (b) (m) on flux measurements Fl05 Florence, Italy Compact R C 25 33 ● ● ● ● 2005 - * Italian National Matese et al. (2009) (Ximeniano) midrise Research Council Gz07 Giza, Egypt Open S O 22 35 ● ● 11/2007- University of Frey et al. (2011) (Cairo University) highrise 2/2008 Basel Gl10 Gliwice, Poland Compact S O 20 28 ● ● 04/2010- University of (SUT) midrise 03/2011 Basel He05 Helsinki, Finland Mixed R I 20 31 ● ● ● 12/2005 - * University of Vesala et al. (2008) (Kumpula) Helsinki He10f Helsinki, Finland Compact C 22 42 ● ● 06/2010 - University of (Fire Station) midrise 01/2011 Helsinki He10h Helsinki, Finland Compact C 24 60 ● ● 09/2010 - * University of (Hotel Torni) midrise Helsinki Ho07 Houston, USA Open R C I 5 60 ● ● ● 07/2007 - * Texas A&M Uni- Changhyoun et al. (Northside Village) lowrise versity (2010) Kc04 Kansas City, USA, Open R 6.9 35 ● 08/2004 University of Balogun et al. (2009) (Greenwood) lowrise Missouri Lo06 Lodz, Poland Compact R S O 11 37 ● ● 2000-2003 / University of Pawlak et al. (2011) (Lipowa) midrise 2006 - * Lodz and this newsletter Lo05 Lodz, Poland Compact R S O 16 42 ● 06/2005 - * University of Offerle et al. (2006a,b) (Narutowicza) midrise Lodz Ld06 London, UK Compact R C S O 8.8 19 ● 10/2006 - * CEH Edinburgh Helfter et al. (2011) (BT Tower) midrise 0 Ld08 London, UK Compact S O 22.8 48 ● ● 10/2008 - * King`s College Kotthaus & Grimmond (KSK) midrise London (2012) Ld09 London, UK Compact S O 22.8 60 ● ● 10/2009 - * King`s College Kotthaus & Grimmond (KSS) midrise London (2012),& this newsletter Sc91u Sacramento, Open R 4.8 N/ ● 08/1991 Indiana University / Grimmond et al. USA, (Urban) lowrise A UBC (1993) Sg94 Los Angeles, Compact R 4.7 18 ● 07/1994 Indiana University Grimmond et al. USA, (San Gabriel) lowrise (1996) Ar94 Los Angeles, Open R 5.2 33 ● ● 1993-1995 Indiana University Grimmond et al. USA, (Arcadia) lowrise (1996) Ar93 Los Angeles, Large R 5.2 31 ● 07/1993 - Indiana University Grimmond and Oke USA, (Arcadia) lowrise 08/1993 (1995) Ma01 Marseille, France Compact R C 15.6 44 ● ● 06/2001 - UBC / Indiana Grimmond et al. (CAA) midrise 07/2001 University (2004) Mb03m Melbourne, Open R 12 40 ● ● 08/2003 - Monash University Coutts et al. (2007) Australia (Preston) lowrise 08/2004 Page 34

Appendix: Flux measurements in urban ecosystems

Sue Grimmond and Andreas Christen

(a) Code City, Country LCZ Land- zH zm E C A O Study period Operated by Primary Reference Site-Name use (b) (m) on flux measurements Mb03l Melbourne, Australia Open R 12 N/ ● ● Monash University Coutts et al. (2007) (Surrey Hills) lowrise A Me93 Mexico City, Mexico Compact R 18. N/ ● 11/1993 - University of British Oke et al. (1999) (School of Mines) midrise 4 A 12/1993 Columbia Me06 Mexico City, Mexico Compact R C 12 42 ● ● ● ● 03/2006; WSU/MCE2 (2006) Velasco et al. (2011) (Escandon) midrise 02/2009 - * MCE2/INE/SMADGF (2009*) Me03 Mexico City, Mexico Compact R S O 12 37 ● ● 04/2003 Washington State Velasco et al. (2005) (Iztapalapa) lowrise University Mi95 Miami, USA Open R 8 40 ● 05/1995 - UBC / Indiana Uni- Newton et al. (2007) (West Miami) lowrise 06/1995 versity Mn05 Minneapolis, USA Open R 6 40 ● ● ● 11/2005 - UC- Santa Barbara Peters and McFadden (Roseville, KUOM) lowrise 06/2009 (2012), this newsletter Mo07u Montréal, Canada Compact R 8 24 ● ● 10/2007 - McGill University Bergeron and Strachan (Rue des Ecores) lowrise 10/2009 (2009) Mo05 Montréal, Canada Compact R 9.5 21 ● ● 03/2005 - Environment Canada Lemonsu et al. (2008) (Rue Fabre) lowrise 04/2005 Mo07s Montréal, Canada Open R 7 24 ● ● 11/2007 - McGill University Bergeron and Strachan (Roxboro) lowrise 10/2009 (2009) Mu06 Münster, Germany Compact R C 15 65 ● ● 03/2011 - University of Münster Dahlkötter et al. (2010) (Manfred-von- midrise 06/2011 Richthofen-Kaserne) 08/2006 Na06 Nantes, France Sparsely R 6 27 ● ● ● 02/2006 - * IRSTV Ruban et al. (2011) (SAP / ONEVU) built Oa03 Ouagadougou, Burki- Not R N/ N/ ● 02/2003 Indiana University/ Offerle et al. (2005) na Faso (Residential) classified A A Göteborg University Ob10s Oberhausen, Germa- Sparsely R 6.2 30 ● 08/2010 - University of Duisburg Goldbach and Kuttler ny (suburban) built 07/2011 -Essen (2012) Ob10u Oberhausen, Germa- Com- R C 14 32 ● 08/2010 - University of Duisburg Goldbach and Kuttler ny (urban) pact 07/2011 -Essen (2012) midrise Ph11 Phoenix, USA Compact R 4.5 22 ● ● 12/2011 - * Arizona State Univer- (West Phoenix) lowrise sity Ro04 Rome, Italy Compact R C 20 56 ● 2004 -2006 Italian National Re- (Collegio Romano) midrise search Council Sl05 Salt Lake City, USA Open R 4.4 37 ● ● 2005, 2007 - * University of Utah Ramamurthy and (Murray) lowrise Pardyjak (2011) Page 35

Appendix: Flux measurements in urban ecosystems

Sue Grimmond and Andreas Christen

(a) Code City, Country LCZ Land- zH zm E C A O Study period Operated by Primary Reference Site-Name use (b) (m) on flux measurements Si06 Singapore, Compact R 9.1 21 ● 03/2006- National University of Roth et al. (2009) Singapore lowrise 03/2007; Singapore (Telok Kurau) 09/2007- 03/2008 Sw11 Swindon, UK Open R 5.5 13 ● ● 05/2011 - CEH Wallingford See this newsletter (BWY) lowrise Sy10u Syracuse, USA Not clas- C N/A N/A ● 06/2010 - * SUNY Buckley et al. (2011) (Center) sified

Sy10p Syracuse, USA Not clas- R N/A N/A ● 06/2010 - * SUNY Buckley et al. (2011) (Park) sified Ta06 Taichung, Compact R 28 56 ● ● 03/2006 - 04- National Chung Hsing Taiwan midrise 2006 University (University) Tk01 Tokyo, Japan Compact R 7.3 29 ● ● 2001 - 2007 Ehime University Moriwaki and Kanda (Kugahara) lowrise (2004) Tu90u Tucson, USA Light- R 5.2 25.6 ● 06/1990 Indiana University Grimmond and Oke (Sam Hughes) weight (1995) lowrise To04 Toulouse, Compact R C S 14.9 48 ● ● ● 02/2004 - CNRM / Météo- Masson et al. (2008) France midrise O 02/2005 France CNRS (Monoprix) Vl92 Vancouver, Large I 5.8 28 ● 07/1992 - Indiana U. / Univ. of Grimmond and Oke Canada lowrise 08/1992 British Columbia (1999) (Light Industri- al) Va08o Vancouver, Open R 5.8 29 ● ● 07/2008 - University of British Crawford et al. (2009) Canada lowrise 08/2008; Columbia (Oakridge) 07/2009 - 08/2009 Va08s Vancouver, Open R C 5.3 28 ● ● ● Intermittent University of British Christen et al. (2011) Canada lowrise since 1977; Columbia See also this newsletter (Sunset) 2001 - 2002; 05/2008 - *

(a) Local climate zone according to Stewart and Oke (2009) [updated version] (b) Land-use in the tower source area - R: residential, C: commercial, I: industrial, S: institutional, O: other (in most cases recreational).