Estimation of the terms acting on local surface one-hour temperature variations in Paris region: the specific contribution of clouds

Abstract

Abstract. Local temperature variations at the surface are mainly dominated by small-scale processes coupled through the surface energy budget terms, which depend mostly on radiation availability and thus cloud processes. A method to determine each of these terms based almost exclusively on observations is presented in this paper, with the main objective to estimate their importance in hourly surface temperature variations at the SIRTA observatory, near Paris. Almost all terms are estimated from the multi-year dataset SIRTA-ReOBS, following a few parametrizations. The four main terms acting on temperature variations are radiative forcing (separated into clear-sky and cloud radiation), atmospheric heat exchange, ground heat exchange, and advection. Compared to direct measurements of hourly temperature variations, it is shown that the sum of the four terms gives a good estimate of the hourly temperature variations, allowing a better assessment of the contribution of each term to the variation, with an accurate diurnal and annual cycles representation, especially for the radiative terms. A random forest analysis shows that whatever the season, clouds are the main modulator of the clear sky radiation for 1-hour temperature variations during the day, and mainly drive these 1-hour temperature variations during the night. Then, the specific role of clouds is analyzed exclusively in cloudy conditions considering the behavior of some classical meteorological variables along with lidar profiles. Cloud radiative effect in shortwave and longwave and lidar profiles show a consistent seasonality during the daytime, with a dominance of mid- and high-level clouds detected at the SIRTA observatory, which also affects surface temperatures and upward sensible heat flux. During the nighttime, despite cloudy conditions and having a strong cloud longwave radiative effect, temperatures are the lowest and are therefore mostly controlled by larger-scale processes at this time.