LAUSR

Abstract In the framework of the INDAAF (International Network to study Deposition and Atmospheric chemistry in AFrica) programme, part of the ACTRIS European Research Infrastructure for the long-term observation of Aerosols, Cloud, and Trace gases, this paper aims to study trends and seasonal variability of surface atmospheric NO2 and HNO3 concentrations, and OMI (Ozone Monitoring Instrument) NO2 over 6 sites in tropical Africa. Sites are located in west and central Africa to represent the major African biomes: dry savanna (Banizoumbou, Niger and Katibougou, Mali), wet savanna (Djougou, Benin and Lamto, Cote d’Ivoire) and forest (Bomassa, Republic of Congo and Zoetele, Cameroon). Ground-based NO2 and HNO3 concentrations were obtained over the period 1998–2015 using INDAAF passive samplers at a monthly basis, and NO2 Vertical Column Densities (VCDs) from OMI for a 1° grid cell around each sites were obtained from 2005 to 2015. Mean annual NO2 concentrations ranged from 2.3 ± 1.2 to 0.9 ± 0.4 ppb from dry savannas to forests, representing a north south gradient. In dry savannas, we observe two concentration peaks of NO2 appearing at the beginning and the end of the wet season both for ground-based and satellite measurements, whereas at wet savannas and forest sites, NO2 concentrations are highest in the dry season. The seasonality of surface NO2 observations provide further evidence for a large role of microbial soil NOx emissions in dry savannas and of biomass burning NOx emissions in wet savanna and forest sites. Mean annual HNO3 concentrations ranged from 0.4 to 0.5 ppb in dry and wet savannas to 0.2–0.3 ppb in forest. In dry ecosystems, higher HNO3 concentrations are measured in the early wet season, consistent with NO2 results. The analysis of a long-term dataset of surface O3 concentrations indicates that HNO3 production can mainly be explained by the photooxidation of NOx. Mann-Kendall and Seasonal Kendall statistical tests showed that NO2 surface concentrations have a significant decreasing seasonal and annual trends at multiple sites (p-value