Long time series spatialized data for urban climatological studies: A case study of Paris, France

Abstract

Historically, the urban climate and its interaction with regional environment and larger‐scale meteorological phenomena have been studied by comparing weather stations in urban and surrounding rural areas. Nevertheless, spatialized meteorological data are now available with different spatial resolutions, temporal frequencies, and time depths. Here with the Paris area as a case study, three products were selected for their comparable spatial resolution of 1 km, their daily frequency, and a minimum 15‐year time period with substantial common overlap: the MODIS products for land surface temperature (2000–2016), daily maps of minimum and maximum 2‐m air temperatures (2000–2017) based on a statistical method of spatialization, and the COMEPHORE analysis for precipitation (1997–2012). Several spatialized indicators were used to monitor the urban climate and its impact on local climate. The intensity and spatial extent of both the surface urban heat island (SUHI) and the urban heat island of urban canopy layer (UHI), as well as the effect of the city on precipitation, were characterized. Based on these indicators, a climatological study was made for the Paris urban area and its surroundings. The comparison of (S)UHI climatologies showed that both phenomena (qualified by their intensity and spatial extent) have comparable seasonalities and orders of magnitude at night. The maximum intensity is reached in July, with more than 60% of the city affected by high temperatures. However, the phenomena differ during the day: the SUHI is mostly driven by the differences in surface properties between urban and rural environments, whereas the UHI is more influenced by general wind circulation and local atmospheric turbulence. Finally, the COMEPHORE analysis highlighted the influence of the Paris urban area on rainfall. It is observed throughout the year with an excess in daily precipitation of 29% downwind of the city, with some seasonal variability showing a more pronounced effect in summer.

Intensité et spatialisation de l’îlot de chaleur urbain parisien à partir de données participatives

“…Les résultats sont en accord avec l'étude réalisée récemment par B. Le Roy et al (2019) sur la période 2000-2016, qui ont comparé les données Météo France interpolées aux données MODIS. Enfin, même si les données participatives utilisées ici nécessitent une analyse critique conséquente (qui pourra ultérieurement être affinée) et donc même s'il faut être plutôt prudents, nous aimerions insister sur les potentialités et les opportunités qu'elles offrent, par leur résolution spatio-temporelle, et il serait intéressant d'approfondir la dynamique d'apparition et disparition de l'îlot de chaleur en couplant ces données avec des simulations numériques.…”

“…D'autres approches s'appuient sur des mesures mobiles, principalement réalisées lors de conditions radiatives (e.g. les travaux de thèse de Foissard, 2015et Le Bras, 2015, sur la modélisation (Wouters et al, 2013 ;Kastendeuch et Najjar, 2015 ;Le Bras, 2015) ou encore sur l'utilisation de données satellitaires (Dousset et al, 2011 ;Madelin et al, 2017 ;Le Roy et al, 2019).…”

Co-Visualization of Air Temperature and Urban Data for Visual Exploration

“…Climate data visualization mostly exists to validate and communicate about simulation results, and especially about air temperature profiles, dynamic wind representation [13] or possible air corridors [4], with the expectations to validate input/output data and models characteristics. Many representation types are used, most of all as vertical and horizontal cross-sections, long time series spatialized data [12] or global trends into long time series [8]. Sets of multi-dimensional spatial 2D views from very small scales (at European level for instance), to wider scales (at city scales for instance) are very often proposed to visualize the spatial distribution of air temperature coming from simulation models integrating morphological indicators, in addition to visualizations of their temporal profile distributions through graphs and histograms [12,25].…”

“…Many representation types are used, most of all as vertical and horizontal cross-sections, long time series spatialized data [12] or global trends into long time series [8]. Sets of multi-dimensional spatial 2D views from very small scales (at European level for instance), to wider scales (at city scales for instance) are very often proposed to visualize the spatial distribution of air temperature coming from simulation models integrating morphological indicators, in addition to visualizations of their temporal profile distributions through graphs and histograms [12,25]. Spatio-temporal animations are very useful to improve the global perception of a spatio-temporal phenomenon and its general patterns, but it remains difficult to detect and analyze changes, at any scales.…”

“…Co-visualizing simulated climate data with morphological urban data would allow supporting the validation of climate simulation models, by comparing simulation results with morphological indicators used as simulation input data, such as a suitable visual analysis of relationships between UHIs and the urban environment. Regarding specifically temperature data, existing visualizations work as final media, to present global trends, breaks and changes, based on long-time series of temperature values as impactful stripes [8] or as spatialized data [12,25], or to inform and explain the temperature profiles as vertical and horizontal cross-sections at city scale. The visualization of plane representations of air temperature into the city have been proposed, projecting temperature into 3D urban space [20], or representing local effects of urban structures by representing temperature on static/animated vertical cross-sections [10].…”