• Alonso, M. S., , J. L. Labajo, , and M. R. Fidalgo, 2003: Characteristics of the urban heat island in the city of Salamanca, Spain. Atmósfera, 16 , 137148.

    • Search Google Scholar
    • Export Citation
  • Bornstein, R., , and Q. L. Lin, 2000: Urban heat islands and summertime convective thunderstorms in Atlanta: Three case studies. Atmos. Environ., 34 , 507516.

    • Search Google Scholar
    • Export Citation
  • Bottyan, Z., , A. Kircsi, , S. Szegedi, , and J. Unger, 2005: The relationship between built-up areas and the spatial development of the mean maximum urban heat island in Debrecen, Hungary. Int. J. Climatol., 25 , 405418.

    • Search Google Scholar
    • Export Citation
  • Brown, D. G., , K. M. Johnson, , T. R. Loveland, , and D. M. Theobald, 2005: Rural land-use trends in the conterminous United States, 1950–2000. Ecol. Appl., 15 , 18511863.

    • Search Google Scholar
    • Export Citation
  • Dixon, P. G., , and T. L. Mote, 2003: Patterns and causes of Atlanta’s urban heat island–initiated precipitation. J. Appl. Meteor., 42 , 12731284.

    • Search Google Scholar
    • Export Citation
  • European Environment Agency, 2006: Urban sprawl in Europe—The ignored challenge. EEA Rep. 10/2006, 56 pp. [Available online at http://www.eea.europa.eu/publications/eea_report_2006_10/eea_report_10_2006.pdf].

    • Search Google Scholar
    • Export Citation
  • Grell, G. A., , J. Dudhia, , and D. R. Stauffer, 1995: A description of the fifth-generation Penn State/NCAR Mesoscale Model (MM5). NCAR Tech. Note NCAR/TN-398+STR, 117 pp.

    • Search Google Scholar
    • Export Citation
  • Grimmond, C. S. B., , and T. R. Oke, 1999a: Heat storage in urban areas: Local-scale observations and evaluation of a simple model. J. Appl. Meteor., 38 , 922940.

    • Search Google Scholar
    • Export Citation
  • Grimmond, C. S. B., , and T. R. Oke, 1999b: Evapotranspiration rates in urban areas. Impacts of Urban Growth on Surface Water and Groundwater Quality, J. B. Ellis, Ed., IAHS Publ. 249, 235–243.

    • Search Google Scholar
    • Export Citation
  • Hand, L. M., , and J. M. Shepherd, 2009: An investigation of warm-season spatial rainfall variability in Oklahoma City: Possible linkages to urbanization and prevailing wind. J. Appl. Meteor. Climatol., 48 , 251269.

    • Search Google Scholar
    • Export Citation
  • Huff, F. A., , and S. A. Changnon, 1973: Precipitation modification by major urban areas. Bull. Amer. Meteor. Soc., 54 , 12201232.

  • Jin, M., , and J. M. Shepherd, 2005: Inclusion of urban landscape in a climate model: How can satellite data help? Bull. Amer. Meteor. Soc., 86 , 681689.

    • Search Google Scholar
    • Export Citation
  • Jin, M., , J. M. Shepherd, , and C. Peters-Lidard, 2007: Development of a parameterization for simulating the urban temperature hazard using satellite observations in climate model. Nat. Hazards, 43 , 257271.

    • Search Google Scholar
    • Export Citation
  • Kaufmann, R. K., , K. C. Seto, , A. Schneider, , Z. Liu, , L. Zhou, , and W. Wang, 2007: Climate response to rapid urban growth: Evidence of a human-induced precipitation deficit. J. Climate, 20 , 22992306.

    • Search Google Scholar
    • Export Citation
  • Klysik, K., , and K. Fortuniak, 1999: Temporal and spatial characteristics of the urban heat island of Lodz, Poland. Atmos. Environ., 33 , 38853895.

    • Search Google Scholar
    • Export Citation
  • Lamptey, B. L., , E. J. Barron, , and D. Pollard, 2005: Impacts of agriculture and urbanization on the climate of the northeastern United States. Global Planet. Change, 49 , 203221.

    • Search Google Scholar
    • Export Citation
  • Masson, V., 2000: A physically based scheme for the urban energy budget in atmospheric models. Bound.-Layer Meteor., 94 , 357397.

  • Mayer, H., , A. Matzarakis, , and M. G. Iziomon, 2003: Spatio-temporal variability of moisture conditions within the urban canopy layer. Theor. Appl. Climatol., 76 , 165179.

    • Search Google Scholar
    • Export Citation
  • Mölders, N., , and M. A. Olson, 2004: Impact of urban effects on precipitation in high latitudes. J. Hydrometeor., 5 , 409429.

  • Montávez, J. P., , A. Rodríguez, , and J. I. Jiménez, 2000: A study of the urban heat island of Granada. Int. J. Climatol., 20 , 899911.

    • Search Google Scholar
    • Export Citation
  • Mote, T. L., , M. C. Lacke, , and J. M. Shepherd, 2007: Radar signatures of the urban effect on precipitation distribution: A case study for Atlanta, Georgia. Geophys. Res. Lett., 34 , L20710. doi:10.1029/2007GL031903.

    • Search Google Scholar
    • Export Citation
  • Oke, T. R., , R. A. Spronken-Smith, , E. Jauregui, , and C. S. B. Grimmond, 1999: The energy balance of central Mexico City during the dry season—Matching scales of observations and fluxes. Atmos. Environ., 33 , 39193930.

    • Search Google Scholar
    • Export Citation
  • Rosenfeld, D., 2000: Suppression of rain and snow by urban and industrial air pollution. Science, 287 , 17931796.

  • Rosenfeld, D., , W. L. Woodley, , D. Axisa, , E. Freud, , J. G. Hudson, , and A. Givati, 2008: Aircraft measurements of the impacts of pollution aerosols on clouds and precipitation over the Sierra Nevada. J. Geophys. Res., 113 , D15203. doi:10.1029/2007JD009544.

    • Search Google Scholar
    • Export Citation
  • Shem, W., , and M. Shepherd, 2009: On the impact of urbanization on summertime thunderstorms in Atlanta: Two numerical model case studies. Atmos. Res., 92 , 172189.

    • Search Google Scholar
    • Export Citation
  • Shepherd, J. M., 2005: A review of current investigations of urban-induced rainfall and recommendations for the future. Earth Interactions, 9 .[Available online at http://EarthInteractions.org].

    • Search Google Scholar
    • Export Citation
  • Stanhill, G., , and J. D. Kalma, 1995: Solar dimming and urban heating at Hong Kong. Int. J. Climatol., 15 , 933941.

  • Thielen, J., , W. Wobrock, , A. Gadian, , P. G. Mestayer, , and J-D. Creutin, 2000: The possible influence of urban surfaces on rainfall development: A sensitivity study in 2D in the meso-γ-scale. Atmos. Res., 54 , 1539.

    • Search Google Scholar
    • Export Citation
  • Trusilova, K., , M. Jung, , G. Churkina, , U. Karstens, , M. Heimann, , and M. Claussen, 2008: Urbanization impacts on the climate in Europe: Numerical experiments by the PSU–NCAR Mesoscale Model (MM5). J. Appl. Meteor. Climatol., 47 , 14421455.

    • Search Google Scholar
    • Export Citation
  • Unger, J., , Z. Sumeghy, , and J. Zoboki, 2001: Temperature cross-section features in an urban area. Atmos. Res., 58 , 117127.

  • World Resources Institute, 1996: World Resources 1996–97: The urban environment. World Resources Institute, 400 pp. [Available online at http://archive.wri.org/publication_detail.cfm?pubid=2872].

    • Search Google Scholar
    • Export Citation
  • View in gallery

    The current-state urban land and the increment of urban land (by 40%) at the spatial resolution of 10 km.

  • View in gallery

    The near-surface temperature and the energy balance of the urban surface in the reference run (NOU) and before (URB) and after (EURB) expansion. The data are 30-day averages of model output for (left) Milan in January 2005 and (right) Berlin in July 2005.

  • View in gallery

    Spatial distribution of statistically significant differences of diurnal temperature range (°C) between EURB-run and NOU-run for (a) January and (b) July simulations. Spatial distribution of statistically significant precipitation differences (%) between EURB-run and NOU-run for (c) January and (d) July simulations.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 32 32 9
PDF Downloads 6 6 0

On Climate Impacts of a Potential Expansion of Urban Land in Europe

View More View Less
  • 1 Max-Planck-Institute for Biogeochemistry, Jena, Germany
© Get Permissions
Full access

Abstract

Over the last two decades, a disproportional increase of urban land area in comparison with the population growth has been observed in many countries of Europe, and this trend is predicted to continue. The conversion of vegetated land into urban land leads to a higher proportion of impervious surface area, to decline and change of vegetation cover, to artificial heat sources, and therefore to changes in climate. This study focuses on the implications of the expansion of urban land for the European climate at the local and regional scales. Regional climate simulations with the fifth-generation Pennsylvania State University–NCAR Mesoscale Model (MM5) coupled to the Town Energy Budget model are used to isolate effects of urban land expansion on temperature and precipitation. The study suggests that the expansion of current urban land by 40% would lead to an enlargement of regions affected by thermal stress by a factor of 2, whereas the intensity of the thermal stress does not change significantly. Precipitation in urban areas would be reduced by 0.2 mm day−1 in summer as a result of disturbances of the water cycle caused by urban surfaces. The area in which precipitation was altered increased nearly linearly with the urban land increment.

Corresponding author address: Kristina Trusilova, MPI-BGC, Hans-Knoell Str. 10, Jena 07745, Germany. Email: ktrusil@bgc-jena.mpg.de

Abstract

Over the last two decades, a disproportional increase of urban land area in comparison with the population growth has been observed in many countries of Europe, and this trend is predicted to continue. The conversion of vegetated land into urban land leads to a higher proportion of impervious surface area, to decline and change of vegetation cover, to artificial heat sources, and therefore to changes in climate. This study focuses on the implications of the expansion of urban land for the European climate at the local and regional scales. Regional climate simulations with the fifth-generation Pennsylvania State University–NCAR Mesoscale Model (MM5) coupled to the Town Energy Budget model are used to isolate effects of urban land expansion on temperature and precipitation. The study suggests that the expansion of current urban land by 40% would lead to an enlargement of regions affected by thermal stress by a factor of 2, whereas the intensity of the thermal stress does not change significantly. Precipitation in urban areas would be reduced by 0.2 mm day−1 in summer as a result of disturbances of the water cycle caused by urban surfaces. The area in which precipitation was altered increased nearly linearly with the urban land increment.

Corresponding author address: Kristina Trusilova, MPI-BGC, Hans-Knoell Str. 10, Jena 07745, Germany. Email: ktrusil@bgc-jena.mpg.de

1. Introduction

Urban population is growing at a much faster rate than the earth’s population as a whole and by larger annual increments than ever before (World Resources Institute 1996). Given the future urbanization projections and estimations of impacts of individual cities on the environment, it becomes important to investigate effects of growing urban areas. The patterns of urban growth remain uncertain and vary among regions and countries. Whereas small cities may experience population densification, large metropolitan areas evolve by an expansion of less densely populated suburban land. In developed countries, a general trend toward less densely populated urban areas has been observed (European Environment Agency 2006; Brown et al. 2005); according to The Cities Alliance (http://www.citiesalliance.org), the average built-up area per person in European cities increased by more than 20% from 1990 to 2000. With the assumption that the population’s average income and demands for space are not changing, the growth of urban population would lead to an expansion of urban land into agricultural and forest areas. This expansion involves replacing vegetated land by heterogeneous surfaces partially covered by impervious materials.

Urban areas affect our environment in different ways: they disturb the water cycle by impervious surfaces, provide heat accumulation in construction materials, and serve as sources of air pollution. Previous studies showed that urbanization largely affects the energy budget of the surface (Grimmond and Oke 1999a; Oke et al. 1999) and air moisture (Mayer et al. 2003; Grimmond and Oke 1999b). The urban heat island (UHI), one of the most frequently studied climatic features of cities, was extensively observed (Bottyan et al. 2005; Alonso et al. 2003; Unger et al. 2001; Klysik and Fortuniak 1999) and modeled (Trusilova et al. 2008; Lamptey et al. 2005). Measurement studies on the distribution of summer precipitation conducted by Mote et al. (2007) and Hand and Shepherd (2009) identified significant impacts of urbanization on precipitation. However, estimates of urban impacts on precipitation vary greatly. It was found that increased surface roughness may lead to enhanced convergence (Thielen et al. 2000) and that the UHI may induce a convergence zone that initiates storms (Bornstein and Lin 2000). Shem and Shepherd (2009) used the Weather Research and Forecast Model (WRF)–National Centers for Environmental Prediction (NCEP)–Oregon State University–Air Force–Hydrologic Research Laboratory (Noah) Model coupled atmosphere–land model to investigate storms observed by Bornstein and Lin. By varying the size of Atlanta, Georgia, or removing it, they quantified the impact of surface fluxes and convergence on downwind precipitation. Rosenfeld et al. (2008) have conducted a measurement campaign and found that anthropogenic aerosols suppress precipitation in orographic clouds. Mölders and Olson (2004) performed model simulations showing that additional moisture from urban sources contributes to increased downwind precipitation.

Because the aerosol–precipitation feedbacks are still poorly understood and, thus, rarely included in regional climate models, we focus only on the effects of land use change on climate. Thus, the results of this study should be interpreted with care, keeping in mind that aerosol effects on surface energy balance and precipitation formation are not included in the models used here.

Recent studies (Jin et al. 2007; Jin and Shepherd 2005) have considered methods to include urban effects in climate models but have not gone beyond proposed methodologies or into quantitative analysis. Quantitative estimations of regional climate changes caused by urban land were made for the northeastern United States (Lamptey et al. 2005) and Europe (Trusilova et al. 2008). These studies have shown that the transformation of vegetated land into urban land reduces air moisture content, provides a stronger surface warming, and leads to shifts in precipitation patterns on the local and regional scales. However, responses of the climate in Europe to the future expansion of urban land have not yet been analyzed. The study presented here addresses this question and aims to provide an estimation of possible temperature and precipitation changes that result from an increased fraction of urban land in Europe. This study is a follow up to work previously published by Trusilova et al. (2008).

2. Materials and methods

a. Model

To isolate effects of the urban land cover on the climate, we use the limited-area numerical weather prediction fifth-generation Pennsylvania State University–National Center for Atmospheric Research (NCAR) Mesoscale Model (MM5; Grell et al. 1995) coupled to a single-layer urban canopy model, the Town Energy Budget model of Masson (2000). The validation of model results against temperature and snow-height measurements is described in the work of Trusilova et al. (2008). This coupling allowed representation of the impacts of urban land cover on the atmosphere and resolution of near-surface processes of heat and moisture transfer sufficiently. Performed model simulations represent responses of the atmospheric circulation to three different states of urbanization: 1) no urban land, 2) urban land at the extent as in 2000–05, and 3) expanded urban land. The first urbanization state with no urban land is defined as the reference state. The effects on climate from the other two are quantified as differences in climate variables such as temperature and precipitation from the reference state. To represent mentioned states of urbanization, we created corresponding land cover maps, which include respective fractions of urban land.

b. Mapping of urban areas

Existing land cover databases that include urban land cover categories define urban land differently. For this study, we define urban land as artificial surfaces defined in the Coordination of Information on the Environment (CORINE) Land Cover 2000 database (CLC2000; http://etc-lusi.eionet.europa.eu/CLC2000). There are 11 different classes of artificial surfaces at the spatial resolution of 250 m. According to the definition of artificial surfaces in CLC2000, urban land includes areas mainly occupied by dwellings and buildings, including their connected areas (associated lands, road network, and parking lots), rail networks, airport installations, river and sea port installations, industrial livestock-rearing facilities, construction sites, anthropogenic waste dump sites, urban parks, and sport and leisure facilities.

Three land-cover maps were created to represent the following states of urbanization: 1) a land cover map that includes no urban area (NOU map), 2) a land cover map with urban areas as in 2000–05 (URB map), and 3) a land cover map that represents expanded urban areas (EURB map).

The land cover map commonly used in MM5 (MM5 map) was derived from the Global Land Cover Characterization from the U.S. Geological Survey (GLCC-USGS). This land cover map includes 24 land cover categories with a single urban land cover class among them. We updated the mask of urban land class in the MM5 map for each model simulation. The NOU map was derived by replacing urban pixels in the MM5 map with the dominant land cover type of neighbor pixels of rural land.

To represent the state of urbanization in 2000–05, an updated mask of urban land was needed because the MM5 map strongly underestimates urban land in Europe. We used an urban mask that was derived by merging urban masks of different land cover databases (Trusilova et al. 2008): GLCC-USGS, Global LandCover 2000 (from the Joint Research Centre of the European Commission Directorate General; GLC2000), the Moderate Resolution Imaging Spectroradiometer (MODIS) land cover types map from the National Aeronautic and Space Administration, the LandScan population dataset (LANDSCAN) from the Oak Ridge National Laboratory, and nighttime light emissions data (NIGHTLIGHT) from the Defense Meteorological Satellite Program. This new urban mask was superimposed on the NOU map to create the URB map.

For the map of expanded urban land (the EURB map), we used the MM5 map for vegetated land combined with a new mask of expanded urban land. The mask of the expanded urban land was created using a proxy indicator of the urban area extent, the urban score. The urban score map was calculated using urban masks from the GLCC-USGS, GLC2000, MODIS, LANDSCAN, and NIGHTLIGHT datasets with a spatial resolution of 1 km. In contrast to the other databases, NIGHTLIGHT and LANDSCAN have continuous data fields. Thus, appropriate thresholds have to be set for extracting relevant urban masks from these datasets.

The CLC2000 database provides a harmonized, reliable, and comparable snapshot of land cover for 2000 for Europe (29 countries) based on high-resolution satellite data with high geometric quality. The CLC2000 land use classification was made using not only the satellite images but also local knowledge about particular land use at the national level. This last fact distinguishes the CLC2000 data from other global land use classifications mentioned above, which are based solely on remote sensing image data. Therefore, the CLC2000 database includes multiple urban land use classes that are often omitted in other databases. We chose the CLC2000 to validate urban masks of other land cover databases.

Each of the individual urban maps (imap) was compared with CLC2000, and the accuracy of mapping of the urban class was calculated as the number of pixels for which CLC2000 and imap agree (both indicate the urban land cover category for the pixel). The degree of the match between the urban mask of imap and the urban mask of CLC2000 (Table 1) was calculated as
i1558-8432-48-9-1971-e1
where imap = GLCC-USGS|GLC2000|MODIS|NIGHTLIGHT|LANDSCAN, Nurban(imap&CLC2000) = number of pixels that are defined as “urban” in both imap and CLC2000, and Ntotal(imap) = total number of urban pixels in imap only.
From the definition, if the urban masks in imap and CLC2000 match for all pixels, then P(imap) = 1; otherwise, P(imap) < 1. Here P(imap) can be understood as the probability of an urban pixel in imap being urban in reality. Each urban pixel in imap was set to P(imap). Given that LANDSCAN and NIGHTLIGHT have continuous values, P(imap) was calculated for each possible value. The urban score map (USM) was calculated as the mean of values P(imap) of all individual imaps for each pixel in row i and column j:
i1558-8432-48-9-1971-e2
From the definition, the maximum possible value of USM is 100% when all individual maps agree on the same urban mask and the minimum value is 0 for pixels that are not identified as urban by any imap.
The mask of expanded urban land for the EURB map was derived from the USM by setting a threshold that masks urban areas that are 2 times as large as in the URB map. The optimal threshold (thEURB) was calculated by minimizing the difference between the total urban area in USM(thEURB) and 2 times the total area of urban mask in the URB map:
i1558-8432-48-9-1971-e3
where EURB = USM(thEURB), thEURB = 25%, and UrbanArea(x) or UrbanArea[x] = total area (km2) of urban land in map x = EURB|URB.

The total area of urban land contained in the EURB map is larger than in the URB map and can be interpreted as the expanded urban land. The total urban area in the EURB map accounts for ∼180% of CLC2000 total urban area. This is explained by the fact that the CLC2000 database has a spatial resolution of 250 m and includes a larger number of small urban areas, which are omitted on the 1-km resolution in the URB and EURB maps.

The derived urban masks in the URB and EURB maps were upscaled to the resolution of 10 km of the model domain (Fig. 1). Because of the upscaling by the principle of dominant land cover type, multiple small urban areas were not included into the urban masks with the coarser resolution. The URB and EURB maps respectively contain 2.8% and 3.9% of land that is classified as urban (Table 2). The ratio of the total urban area in the URB map to the total urban area in the EURB map was 1:1.4, which means that the urban land was enlarged by 40% of its original size. The increase of the urban-to-rural land border (perimeter) was ∼30%.

c. Modeling protocol

Three sets of model simulations, which make use of the three different land cover maps, were performed: the baseline simulation NOU run used the NOU map with no urban land, the URB run used the URB map, and the EURB run used the EURB map of expanded urban land.

The model domain of 361 × 283 grid cells included the most urbanized areas of Europe, with the grid size of 10 km and 23 vertical σ levels. It was nested in an intermediate domain with a spatial resolution of 30 km. At lateral boundaries, the model was constrained every 12 h (at 0000 and 1200 UTC of each simulated day) by the NCEP final analysis data (FNL ds083.2; http://dss.ucar.edu/datasets/ds083.2/). The model setup of physical parameterization schemes is described in detail in the previous study by Trusilova et al. (2008). This setup was used for the current study without further modifications.

According to previous investigations, it was found that urban land is most likely to modify the atmospheric circulation in winter (Montávez et al. 2000) and in summer (Bottyan et al. 2005; Unger et al. 2001). According to this finding and in an attempt to reduce large model computational costs, we performed an ensemble of six model realizations for January and six realizations for July over 2000–05 for each model scenario.

d. Analysis of results

Effects of the urban land on the climate were detected with significance tests of the differences in the near-surface temperature and precipitation between the URB and EURB runs and the NOU run. The differences of climate variables between scenarios are denoted with indices URB-NOU and EURB-NOU for pairs of URB or EURB runs and the baseline NOU run, respectively. Different significance tests were used for the temperature and precipitation effects: the Mann–Whitney U test for analysis of the near-surface temperature and the sign test for analysis of precipitation differences.

We chose to demonstrate changes of precipitation as the relative quantity (%) calculated as
i1558-8432-48-9-1971-e4
where PRNOU = daily precipitation in the baseline NOU run.

This was done because absolute precipitation amounts vary strongly over Europe and the same precipitation change may be insignificant in wet climates whereas it may be crucial in dry regions. Thus, plotting relative precipitation change shows the places most affected by precipitation changes due to urban growth. However, we still provided quantitative estimates of precipitation changes in millimeters per day.

To characterize the change in the land area where climate is affected by urban areas, we used the regional effect index (REI) suggested by Trusilova et al. (2008). The REI is calculated for the pair of model runs urb and NOU (urb = URB|EURB) as the ratio of the total area over which significant differences in the climate variable x between the two runs were found to the total area of urban land:
i1558-8432-48-9-1971-e5
where Aaff_rur(x, urb) = total rural area beyond cities over which significant differences of x between the urb and NOU runs were found and Aurb = total area of urban land in the urb run.

From the definition of REI(x, urb), it is always greater than or equal to 1.0, assuming that urban land is always affected. If no rural land is affected by changes of x and Aaff_rur(x, urb) → 0, then there is no significant regional effect on the variable x and REI(x, urb) → 1. If REI(x, urb) is significantly greater than 1.0, x is altered on the regional scale. The significance threshold for REI(x, urb) was set to 0.05, so that REI(x, urb) > 1.05 means urban land in the urb run has a significant effect at the regional scale.

In addition we calculate the ratio r(x) of total areas over which the variable x was altered by urban land in the URB and EURB simulations:
i1558-8432-48-9-1971-e6
where Aaff_tot(x, urb) = total area over which significant differences between the urb and NOU runs were found for variable x [Aaff_tot(x, urb) = Aaff_rur(x, urb) + Aurb(x, urb)]. If the ratio r(x) is close to the ratio of the urban land expansion (i.e., 1:1.4), then the area over which x is affected is linearly proportional to the total urban area. Otherwise, the relationship is not linear.

Quantitative estimates of average impacts of urbanization on climate variables were calculated as the average over all cities in the model domain. For a detailed demonstration of urban effects we chose four cities: two in a temperate zone with cold and humid winters [Berlin, Germany, (an inland city) and London, United Kingdom (a city located close to the coast)] and two in a drier temperate climate [Milan, Italy, (a city located close to the mountains and the coastline) and Madrid, Spain (an inland city)].

3. Results and discussion

a. Effects of urban growth on near-surface temperature

The analysis of the NOU-run and URB-run model simulations showed that urban land modifies the atmospheric circulation and leads to changes in the near-surface temperature and precipitation (Trusilova et al. 2008). As the presence of urban land contributed ∼1°C to the increase of the minimum diurnal near-surface temperature Tmin, the subsequent expansion led to a weakening of this effect. For example, in Milan, the winter Tmin increment was greater than 2.2°C and less than 1.4°C before and after the expansion, respectively (Table 3). The reason for the Tmin increase can be attributed to the changes in the surface energy balance between the NOU, URB, and EURB runs: in the EURB run the ground heat flux was higher during the daytime than in the URB run. The increase of the latent heat flux in the EURB run conditioned additional surface cooling.

Standard deviations for estimates of Tmin increased from the URB simulation to the EURB simulation. This indicates that the larger urban surfaces do not directly imply stronger UHI and may in fact initiate more heterogeneous temperature gradients.

Although the difference between TminURB-NOU and TminEURB-NOU was not statistically significant, the changes of Tmin in the EURB run were found over larger suburban areas than in the URB run (Table 4). The value REI(Tmin, EURB) = 2.56 for July shows that significant changes of Tmin were found over the rural area that were as large as 2.56 − 1.00 = 1.56 times the total urban area; the expansion of urban land from 159 100 to 222 800 km2 caused the total area affected by Tmin anomalies to grow from 168 500 to 570 100 km2. In this case, the total area over which Tmin was altered increased by 238% in response to the 40% increase in urban land. For January, the increase accounted for 156%.

In winter, the enlargement of urban surface in the EURB run provided an increase in the maximum diurnal near-surface temperature Tmax of the same magnitude as the URB run (Table 3); in both runs the Tmax increment was less than 0.4°C on average. The variability of Tmax was high over the model domain because of heterogeneous climate conditions across Europe, and it resulted in large uncertainties of urban effect estimates on Tmax (Table 3).

In response to the heat storage in artificial materials, the diurnal temperature peak is more likely to be offset so that the temperature reaches its highest value Tmax in a city several hours later than in a rural area. This is most likely to happen when there is a strong contrast between urban and rural surface properties (i.e., albedo, roughness, and heat capacity; Trusilova et al. 2008). For example, in Madrid Tmax was reduced by 0.03° and 0.54°C before and after urban expansion, respectively (Table 3). In London and Berlin—cities with cold and humid winters—Tmax changed neither with the presence of urban land nor with its growth, whereas the Tmax increment in Milan exceeded 0.2°C (Table 3; Fig. 2a).

The statistical analysis revealed no significant differences between the magnitudes of Tmax induced by urban land before and after expansion. Thus, the expansion of urban land does not significantly change urban Tmax. However, this urban growth leads to an expansion of the area of urbanization-modified Tmax by 80% in winter (Table 4).

In summer, changes of Tmax were similar across two geographical regions: 1) eastern and central Europe (the ECEU region) and 2) southern Europe (the SEU region). For both the URB and EURB runs, within the ECEU region Tmax increased by more than 0.7°C, whereas within the SEU region Tmax was reduced by more than 1.1°C (Table 3). For example, in Berlin (ECEU) the Tmax increment was greater than 0.2°C before the expansion and in Madrid (SEU) the change of Tmax from the baseline was negative (Table 3). Similar to the winter simulations, the diurnal temperature peak was delayed by several hours responding to the urban heat storage. The heat storage as well as the sensible heat flux was larger in the expanded cities than in the cities of the URB run (Fig. 2). A case of the urban cooling effect in Madrid in July 2005 was also demonstrated by Trusilova et al. (2008).

In the URB and EURB runs changes of Tmax were detected in urban areas and in the close suburban surroundings, thus indicating no strong regional-scale effects (Table 4). From the analysis of the magnitude and extent of TmaxURB-NOU and TmaxEURB-NOU, we found that the expansion of urban land does not significantly change urban Tmax. The urban expansion by 40% provided 166% increase of the area of urbanization-modified Tmax in summer.

The changes in Tmin and Tmax caused the diurnal temperature range (DTR) to decrease within urban land and its surroundings (Figs. 3a,b). DTRURB-NOU and DTREURB-NOU decreased by more than 0.7°C in January and 1.2°C in July (Table 3). For four chosen cities—Berlin, London, Milan, and Madrid—the DTR decreased; however, for Madrid and Milan this reduction was stronger then for London and Berlin in both seasons (Table 3). Because the urbanization-induced effects on Tmin and Tmax were not significantly different between the URB and EURB runs, DTR also did not change significantly with the expansion of the urban land. The area of urbanization-modified DTR was enlarged by 170% and 172% in January and July, respectively (Table 4).

The assumed 40% expansion of urban land did not significantly contribute to the magnitude of urbanization-induced thermal stress but did result in a considerable increase of area over which temperature regimes were affected by the urban land use. The analyzed cases showed that the area of urbanization-affected climate increased by more than 100% in response to the 40% urban land increment. This nonlinear proportion suggests that growing cities alter thermal regimes over distances larger than may be expected.

b. Effects of urban growth on precipitation

Both model simulations that include urban land produced less precipitation (PRURB-NOU < 0 and PREURB-NOU < 0) than the baseline simulation (Figs. 3c,d), with a stronger average reduction of precipitation over the expanded urban land (Table 3). The high spatial variability of precipitation rates across the model domain produced large uncertainties in estimates of precipitation response to urban land.

Although the simulated effects PRURB-NOU and PREURB-NOU were of the same order of magnitude in winter, the expansion of urban land caused a greater change in summer precipitation (Table 3). Whereas urban land provided some deficit of urban precipitation all over the model domain on average, some large cities such as London and Berlin showed enhanced precipitation in winter (Table 3). This local increase was mostly found in the temperate zone in which winters are cold and humid and the thermal perturbation of the boundary layer initiated by the UHI may lead to enhanced convection (Bornstein and Lin 2000).

In dry climates within large urban areas such as Madrid and Milan, winter precipitation was reduced because of changes in the hydrological cycle. Because the low-level moisture is one of the most important factors for UHI-induced precipitation (Dixon and Mote 2003), the large surface runoff in urban areas leads to reduced surface evaporation and a deficit of moisture availability for convective precipitation formation. As evidence, the latent heat flux in the URB and EURB runs was very small (Fig. 2) as compared with the baseline simulation. Because of the larger size of urban areas in the EURB-run, the total water surface runoff was also larger than in the URB run and resulted in a stronger deficit of surface evaporation.

The strongest reduction of precipitation in July was found in northern Italy and accounted for 0.41 ± 0.35 and 0.67 ± 0.59 mm day−1 before and after the expansion, respectively. The majority of cities all over Europe showed some precipitation reduction in the dry season that was amplified by the cities’ expansion. For example, in London and in Madrid the summer precipitation deficit was amplified with the growth of urban areas (Table 3). In a similar way, Kaufmann et al. (2007) found a causal relationship from temporal and spatial patterns of urbanization to temporal and spatial patterns of precipitation during the dry season in the Pearl River delta of China. The authors suggested that urbanization-related changes in surface hydrology led to the precipitation deficit.

REI values indicate that the land area over which cities influence precipitation has increased with urban growth (Table 4). The increment of the area of precipitation change and the magnitude of precipitation suggest that the expansion of urban land has almost a linear effect on precipitation. However, the intensified urban drought in summer may be evidence of the high sensitivity of convective precipitation to the growth of urban land.

4. Summary and outlook

This study suggests that the size of urban areas has a significant influence on near-surface temperatures and summer precipitation. Our numerical simulations showed that the expansion of existing urban areas leads to a disproportional enlargement of the area over which near-surface temperature is affected. The proportion of land over which the diurnal temperature range was significantly affected grew by more than a factor of 2 in response to the 40% urban land increase. Thus, the rural areas adjacent to cities are very sensitive to the urban expansion because rural temperatures are strongly influenced by thermal regimes of cities.

A strong reduction of summer precipitation was attributed to the lack of surface evaporation, which affected the formation of convective precipitation. The area over which precipitation was affected increased in almost linear proportion to the urban land increment.

We found that the diurnal temperature range was reduced by 0.08 C° in January and by 0.23 C° in July as a result of the expansion of urban areas. This relatively small effect becomes important when we take into account the growing proportion of land over which it occurs. The maximum reduction of summer precipitation provided by cities’ growth was greater than 0.2 mm day−1 (6 mm month−1) in the area of northern Italy. This reduction was local in character but could potentially affect many city’s inhabitants and suburban agricultural lands. However, the estimation of precipitation changes from this study should be taken with care because the precipitation microphysics was not resolved explicitly because of the coarse resolution of the model; extreme highly localized precipitation events are likely to be omitted in this estimation.

These findings suggest that local modifications of land cover such as a switch from vegetated to urban land may significantly alter temperature and precipitation of areas much larger than the cities themselves. Therefore, urbanization may cause significant effects on climate at the regional scale and thus should not be neglected in regional climate forecasts by climate scientists or policy makers when cities in the region of interest are expected to grow.

In this study we have focused on one of the most obvious urbanization-driven modifications of our environment—changes in land cover—and its implications for temperature and precipitation. As was pointed out in the review by Shepherd (2005), other shortcomings of urbanization such as air pollution may also affect urban and rural environments. Air pollution causes significant impacts on cloud and precipitation formation (Dixon and Mote 2003; Rosenfeld 2000; Huff and Changnon 1973) through disturbances in cloud droplet nucleation and on the surface–atmosphere energy budget through shielding shortwave radiation (Stanhill and Kalma 1995). These effects should be taken into account in future research when their mechanisms can be better understood and parameterized within weather/climate models.

Acknowledgments

We thank the Max-Planck-Gesellschaft for providing the scholarship for Kristina Trusilova, DKRZ (Deutsches Klimarechenzentrum GmbH) for providing computer facilities, and the MM5 development team at NCAR for the model support. We especially thank Dr. J. Schumacher for consulting on the statistics and Prof. S. Grimmond for fruitful discussions and useful suggestions on the modeling work.

REFERENCES

  • Alonso, M. S., , J. L. Labajo, , and M. R. Fidalgo, 2003: Characteristics of the urban heat island in the city of Salamanca, Spain. Atmósfera, 16 , 137148.

    • Search Google Scholar
    • Export Citation
  • Bornstein, R., , and Q. L. Lin, 2000: Urban heat islands and summertime convective thunderstorms in Atlanta: Three case studies. Atmos. Environ., 34 , 507516.

    • Search Google Scholar
    • Export Citation
  • Bottyan, Z., , A. Kircsi, , S. Szegedi, , and J. Unger, 2005: The relationship between built-up areas and the spatial development of the mean maximum urban heat island in Debrecen, Hungary. Int. J. Climatol., 25 , 405418.

    • Search Google Scholar
    • Export Citation
  • Brown, D. G., , K. M. Johnson, , T. R. Loveland, , and D. M. Theobald, 2005: Rural land-use trends in the conterminous United States, 1950–2000. Ecol. Appl., 15 , 18511863.

    • Search Google Scholar
    • Export Citation
  • Dixon, P. G., , and T. L. Mote, 2003: Patterns and causes of Atlanta’s urban heat island–initiated precipitation. J. Appl. Meteor., 42 , 12731284.

    • Search Google Scholar
    • Export Citation
  • European Environment Agency, 2006: Urban sprawl in Europe—The ignored challenge. EEA Rep. 10/2006, 56 pp. [Available online at http://www.eea.europa.eu/publications/eea_report_2006_10/eea_report_10_2006.pdf].

    • Search Google Scholar
    • Export Citation
  • Grell, G. A., , J. Dudhia, , and D. R. Stauffer, 1995: A description of the fifth-generation Penn State/NCAR Mesoscale Model (MM5). NCAR Tech. Note NCAR/TN-398+STR, 117 pp.

    • Search Google Scholar
    • Export Citation
  • Grimmond, C. S. B., , and T. R. Oke, 1999a: Heat storage in urban areas: Local-scale observations and evaluation of a simple model. J. Appl. Meteor., 38 , 922940.

    • Search Google Scholar
    • Export Citation
  • Grimmond, C. S. B., , and T. R. Oke, 1999b: Evapotranspiration rates in urban areas. Impacts of Urban Growth on Surface Water and Groundwater Quality, J. B. Ellis, Ed., IAHS Publ. 249, 235–243.

    • Search Google Scholar
    • Export Citation
  • Hand, L. M., , and J. M. Shepherd, 2009: An investigation of warm-season spatial rainfall variability in Oklahoma City: Possible linkages to urbanization and prevailing wind. J. Appl. Meteor. Climatol., 48 , 251269.

    • Search Google Scholar
    • Export Citation
  • Huff, F. A., , and S. A. Changnon, 1973: Precipitation modification by major urban areas. Bull. Amer. Meteor. Soc., 54 , 12201232.

  • Jin, M., , and J. M. Shepherd, 2005: Inclusion of urban landscape in a climate model: How can satellite data help? Bull. Amer. Meteor. Soc., 86 , 681689.

    • Search Google Scholar
    • Export Citation
  • Jin, M., , J. M. Shepherd, , and C. Peters-Lidard, 2007: Development of a parameterization for simulating the urban temperature hazard using satellite observations in climate model. Nat. Hazards, 43 , 257271.

    • Search Google Scholar
    • Export Citation
  • Kaufmann, R. K., , K. C. Seto, , A. Schneider, , Z. Liu, , L. Zhou, , and W. Wang, 2007: Climate response to rapid urban growth: Evidence of a human-induced precipitation deficit. J. Climate, 20 , 22992306.

    • Search Google Scholar
    • Export Citation
  • Klysik, K., , and K. Fortuniak, 1999: Temporal and spatial characteristics of the urban heat island of Lodz, Poland. Atmos. Environ., 33 , 38853895.

    • Search Google Scholar
    • Export Citation
  • Lamptey, B. L., , E. J. Barron, , and D. Pollard, 2005: Impacts of agriculture and urbanization on the climate of the northeastern United States. Global Planet. Change, 49 , 203221.

    • Search Google Scholar
    • Export Citation
  • Masson, V., 2000: A physically based scheme for the urban energy budget in atmospheric models. Bound.-Layer Meteor., 94 , 357397.

  • Mayer, H., , A. Matzarakis, , and M. G. Iziomon, 2003: Spatio-temporal variability of moisture conditions within the urban canopy layer. Theor. Appl. Climatol., 76 , 165179.

    • Search Google Scholar
    • Export Citation
  • Mölders, N., , and M. A. Olson, 2004: Impact of urban effects on precipitation in high latitudes. J. Hydrometeor., 5 , 409429.

  • Montávez, J. P., , A. Rodríguez, , and J. I. Jiménez, 2000: A study of the urban heat island of Granada. Int. J. Climatol., 20 , 899911.

    • Search Google Scholar
    • Export Citation
  • Mote, T. L., , M. C. Lacke, , and J. M. Shepherd, 2007: Radar signatures of the urban effect on precipitation distribution: A case study for Atlanta, Georgia. Geophys. Res. Lett., 34 , L20710. doi:10.1029/2007GL031903.

    • Search Google Scholar
    • Export Citation
  • Oke, T. R., , R. A. Spronken-Smith, , E. Jauregui, , and C. S. B. Grimmond, 1999: The energy balance of central Mexico City during the dry season—Matching scales of observations and fluxes. Atmos. Environ., 33 , 39193930.

    • Search Google Scholar
    • Export Citation
  • Rosenfeld, D., 2000: Suppression of rain and snow by urban and industrial air pollution. Science, 287 , 17931796.

  • Rosenfeld, D., , W. L. Woodley, , D. Axisa, , E. Freud, , J. G. Hudson, , and A. Givati, 2008: Aircraft measurements of the impacts of pollution aerosols on clouds and precipitation over the Sierra Nevada. J. Geophys. Res., 113 , D15203. doi:10.1029/2007JD009544.

    • Search Google Scholar
    • Export Citation
  • Shem, W., , and M. Shepherd, 2009: On the impact of urbanization on summertime thunderstorms in Atlanta: Two numerical model case studies. Atmos. Res., 92 , 172189.

    • Search Google Scholar
    • Export Citation
  • Shepherd, J. M., 2005: A review of current investigations of urban-induced rainfall and recommendations for the future. Earth Interactions, 9 .[Available online at http://EarthInteractions.org].

    • Search Google Scholar
    • Export Citation
  • Stanhill, G., , and J. D. Kalma, 1995: Solar dimming and urban heating at Hong Kong. Int. J. Climatol., 15 , 933941.

  • Thielen, J., , W. Wobrock, , A. Gadian, , P. G. Mestayer, , and J-D. Creutin, 2000: The possible influence of urban surfaces on rainfall development: A sensitivity study in 2D in the meso-γ-scale. Atmos. Res., 54 , 1539.

    • Search Google Scholar
    • Export Citation
  • Trusilova, K., , M. Jung, , G. Churkina, , U. Karstens, , M. Heimann, , and M. Claussen, 2008: Urbanization impacts on the climate in Europe: Numerical experiments by the PSU–NCAR Mesoscale Model (MM5). J. Appl. Meteor. Climatol., 47 , 14421455.

    • Search Google Scholar
    • Export Citation
  • Unger, J., , Z. Sumeghy, , and J. Zoboki, 2001: Temperature cross-section features in an urban area. Atmos. Res., 58 , 117127.

  • World Resources Institute, 1996: World Resources 1996–97: The urban environment. World Resources Institute, 400 pp. [Available online at http://archive.wri.org/publication_detail.cfm?pubid=2872].

    • Search Google Scholar
    • Export Citation
Fig. 1.
Fig. 1.

The current-state urban land and the increment of urban land (by 40%) at the spatial resolution of 10 km.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2108.1

Fig. 2.
Fig. 2.

The near-surface temperature and the energy balance of the urban surface in the reference run (NOU) and before (URB) and after (EURB) expansion. The data are 30-day averages of model output for (left) Milan in January 2005 and (right) Berlin in July 2005.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2108.1

Fig. 3.
Fig. 3.

Spatial distribution of statistically significant differences of diurnal temperature range (°C) between EURB-run and NOU-run for (a) January and (b) July simulations. Spatial distribution of statistically significant precipitation differences (%) between EURB-run and NOU-run for (c) January and (d) July simulations.

Citation: Journal of Applied Meteorology and Climatology 48, 9; 10.1175/2009JAMC2108.1

Table 1.

Degree of the match of urban masks P between each individual urban mask (imap) and the reference urban mask EU-CORINE.

Table 1.
Table 2.

Comparison of urban land fraction in the URB and EURB maps of urban land.

Table 2.
Table 3.

Effects of urban land and its expansion on the near-surface temperature and precipitation.

Table 3.
Table 4.

Differences in the spatial extent of the effects on the near-surface temperature and precipitation from actual (URB map) and expanded (EURB map) urban land. The ratio of the total urban land in the URB map to the total urban area in the EURB map is 1:1.4. Here, AreaURB(x) and AreaEURB(x) are the total area over which changes of x are found for the URB and EURB runs, respectively.

Table 4.
Save