QUARTERLY JOURNAL OF THE ROYAL METEOROLOGICAL SOCIETY Q. J. R. Meteorol. Soc. 135: 1266–1276 (2009) Published online 10 July 2009 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/qj.446 Water vapour variability induced by urban/rural surface heterogeneities during convective conditions C. Champollion, a,b * P. Drobinski, a M. Haeffelin, a O. Bock, c J. Tarniewicz, c M. N. Bouin d,f and R. Vautard e a IPSL/Laboratoire de M´ et´ eorologie Dynamique, ´ Ecole Polytechnique/UPMC/ENS/CNRS, Palaiseau, France b G´ eosciences Montpellier, UM2/CNRS, Montpellier, France c IPSL/Service d’A´ eronomie, UPMC/UVSQ/CNRS, Paris, France d Laboratoire de Recherche en G´ eod´ esie, IGN/CNRS, Paris, France e IPSL/Laboratoire des Sciences du Climat et de l’Environnement, CEA/UVSQ/CNRS, Gif-sur-Yvette, France f Centre de M´ et´ eorologie Marine, CNRM, M´ et´ eo France, France ABSTRACT: Scientific interest in urban meteorology has increased because highly populated areas experience high vulnerability to pollution or heavy rain. However, compared to urban air quality or urban heat island (UHI) processes, the urban water vapour cycle is poorly understood because it has been investigated less due to the lack of upper-air measurements and the high sensitivity of surface measurements to local heterogeneities. In this paper, surface measurements of wind, temperature, pressure and humidity, as well as integrated water vapour (IWV) from GPS and MODIS and numerical simulations, have been used to investigate the urban cycle of water vapour in May and June 2004 during the VAPIC field experiment in the Paris area. The surface data show the typical characteristics of an urban area with the absence of water vapour sources and a UHI of about 6 ◦ C at night. The urban IWV distribution differs completely, with an urban IWV excess on average between 1600 and 0600 UTC (with a maximum of about 1.5 kg m −2 at 0600 and 1700 UTC). No IWV difference between the urban and rural areas is found in the middle of the day. The numerical simulations reproduce accurately the urban IWV anomaly. Shallow surface wind convergence associated with the UHI during nighttime provides moisture from the rural areas. Urban areas are therefore under wind convergence for most of the time. The rural water vapour sources and the depth of the convergence control the amplitude of the urban IWV excess. At about 1200 UTC, entrainment at the top of the urban boundary layer is the key process that inhibits the urban IWV excess observed at night. Copyright c  2009 Royal Meteorological Society KEY WORDS urban meteorology; water vapour budget; VAPIC experiment; GPS; mesoscale simulation Received 6 February 2007; Revised 23 February 2009; Accepted 29 April 2009 1. Introduction The worldwide increase of inhabitants in urban areas leads to a high level of vulnerability to natural hazards. Heatwaves, heavy rain and flooding produce more harm- ful effects in urban areas than in rural areas. Moreover, the peculiarity of urban areas may enhance convective precipitation (Thielen et al., 2000; Shepherd et al., 2002; Collier, 2006). The enhancement of convergence and con- vection above large urban area (Atlanta, New York, Saint Louis, Seoul) have indeed been used by several authors to explain the enhancement of precipitation (Baik et al., 2001; Rozoff et al., 2002; Dixon and Mote, 2003). How- ever, the urban area can also create a barrier effect which reduces frontal precipitation (Bornstein and Lin, 2000). There are also situations in which urban aerosol may suppress precipitation (Rosenfeld, 1999). Knowledge of the urban water cycle is thus a current research topic of ∗ Correspondence to: C. Champollion, G´ eosciences Montpellier, Uni- versit´ e Montpellier II, 34095 Montpellier C´ edex, France. E-mail: cedric.champollion@gm.univ-montp2.fr high interest to address how human activities can impact weather. In this context, the urban water cycle has not been extensively studied in spite of its importance. The urban effect on the temperature (the urban heat island, UHI) has been described for different climates and seasons by several authors (reviews by Oke, 1987; Arnfield, 2003). However the urban effect on water vapour has been doc- umented for a few cases but only based on surface mea- surements (Lee, 1991; J´ auregui and Tejeda, 1998; Unger, 1999; Deosthali, 2000; Unkaˇ sevic et al., 2001; Richards, 2005). Indeed, water vapour monitoring is complicated due to the specific constraints of urban areas: soundings are seldom released in city centres, the natural high vari- ability of water vapour is enhanced by the heterogeneities of the urban surface properties which affect the represen- tativeness of surface measurements of urban water vapour and impact on numerical modelling which requires spe- cific subgrid-scale parametrization (Masson, 2000, 2006; Dupont et al., 2004, 2006). Overall, studies found in the literature (all previous references in this paragraph) agree Copyright c  2009 Royal Meteorological Society