Contribution of cloudy/clear-sky partitions and aerosol variability to the solar irradiance measured over 2010-2020 in Northern France

Abstract

The absorption and scattering by aerosols and clouds greatly affect surface solar irradiance and the Earth's energy budget. Large uncertainties still remain on the current estimates of these radiative effects because of an incomplete knowledge of their spatial and temporal variabilities. We analyze coincident daytime ground-based measurements of aerosol optical properties from AERONET photometer and of direct and diffuse horizontal surface irradiances from a set of pyrheliometer and pyranometer routinely performed at the ATOLL (Atmospheric Observations in LiLle) platform in north of France over the period 2010-2020. The site is located in a highly populated area greatly influenced by clouds and aerosol pollution. In order to isolate the radiative effect of aerosols from that of cloud occurences, a separation between cloudy-sun, clear-sun with surrounding clouds and clear-sky moments is performed using two cloud detection algorithms at 1-min resolution. The measurements are further analyzed using a radiative transfer code to simulate the spectrally integrated solar global horizontal irradiance and its components in clear-sky conditions with and without aerosols (pristine like conditions). Our analysis shows that on average in ATOLL over the period 2010-2020, the sky is cloudy 89% of the time with around 67% of cloudy-sun situations and 22% of clear-sun with clouds moments. The proportion of clear-sky conditions is relatively low (11%) with a minimum in winter (6%) and a maximum in spring (15%). In summer, we observe over the period a robust increasing trend in measured total irradiances in all-sky conditions of +5.2 ± 1.8 W/m²/year. This evolution is mainly explained by a positive trend in the occurrence of clear-sky situations (+0.7 ± 0.3% per year) to the detriment of cloudy-sun moments. In spring, we highlight a high variability of cloud occurences and mean solar irradiances. Indeed, the mean proportion of clear-sky moments varies more than fourfold between 2013 (8%) and 2020 (35%), leading to corresponding all-sky irradiance extrema of 285 and 389 W/m² respectively. Moreover, in clear-sky conditions, an important variability is observed between the maximum of seasonal global irradiance of spring 2018 (522 W/m²) and the minimum of spring 2014 (435 W/m²). The latter variability is emphasised by a significative positive trend in direct irradiances observed for springtime clear-sky conditions of +5.3 ± 2.3 W/m²/year. A sensitivity analysis based on our radiative simulations shows that it is partly explained by a significant decrease in measured AOD440 nm (-0.006 ± 0.002 per year) and a change in the proportion of high aerosol loads over 2010-2020. This is also consistent with a negative trend of the diffuse component (-1.2 ± 0.4 W/m²/year) observed for clear-sky conditions in spring. Finally, besides showing the highest proportion of clear-sky moments, spring is also the most polluted season in aerosol, with more than 80% of AOD440 nm higher than 0.1. This translates to an average seasonal maximum of aerosol direct radiative effect in spring of -22.6 W/m² (-6.1%), with a loss of -69.3 W/m² (-19.2%) of direct irradiance partially compensated by an increase of the diffuse radiation of +46.6 W/m² (101.0%).