The widespread use of Aerodyne aerosol mass spectrometers (AMS) has greatly improved real-time organic aerosol (OA) monitoring, providing mass spectra that contain sufficient information for source apportionment. However, AMS field deployments remain expensive and demanding, limiting the acquisition of long-term datasets at many sampling sites. The offline application of aerosol mass spectrometry entailing the analysis of nebulized water extracted filter samples (offline-AMS) increases the spatial coverage accessible to AMS measurements, being filters routinely collected at many stations worldwide. PM1 (particulate matter with an aerodynamic diameter 10.5194/acp-17-117-2017lt; 1 μm) filter samples were collected during an entire year in Lithuania at three different locations representative of three typical environments of the southeast Baltic region: Vilnius (urban background), Rūgšteliškis (rural terrestrial), and Preila (rural coastal). Aqueous filter extracts were nebulized in Ar, yielding the first AMS measurements of water-soluble atmospheric organic aerosol (WSOA) without interference from air fragments. This enables direct measurement of the CO+ fragment contribution, whose intensity is typically assumed to be equal to that of CO+ 2. Offline-AMS spectra reveal that the water-soluble CO+ 2: CO+ ratio not only shows values systematically 10.5194/acp-17-117-2017gt; 1 but is also dependent on season, with lower values in winter than in summer. AMS WSOA spectra were analyzed using positive matrix factorization (PMF), which yielded four factors. These factors included biomass burning OA (BBOA), local OA (LOA) contributing significantly only in Vilnius, and two oxygenated OA (OOA) factors, summer OOA (S-OOA) and background OOA (B-OOA), distinguished by their seasonal variability. The contribution of traffic exhaust OA (TEOA) was not resolved by PMF due to both low concentrations and low water solubility. Therefore, the TEOA concentration was estimated using a chemical mass balance approach, based on the concentrations of hopanes, specific markers of traffic emissions. AMS-PMF source apportionment results were consistent with those obtained from PMF applied to marker concentrations (i.e., major inorganic ions, OC / EC, and organic markers including polycyclic aromatic hydrocarbons and their derivatives, hopanes, long-chain alkanes, monosaccharides, anhydrous sugars, and lignin fragmentation products). OA was the largest fraction of PM1 and was dominated by BBOA during winter with an average concentration of 2 μg m-3 (53 % of OM), while S-OOA, probably related to biogenic emissions, was the prevalent OA component during summer with an average concentration of 1.2 μg m-3 (45 % of OM). PMF ascribed a large part of the CO+ explained variability (97%) to the OOA and BBOA factors. Accordingly, we discuss a new CO+ parameterization as a function of CO+ 2 and C2H4O+ 2 fragments, which were selected to describe the variability of the OOA and BBOA factors.