Ambient air pollution remains the major environmental cause of disease. Accurate assessment of population exposure and small-scale spatial exposure variations over long time periods is essential for epidemiological studies. We estimated annual exposure to fine and coarse particulate matter (PM_2.5, PM₁₀), and nitrogen oxides (NO_x, NO₂) with high spatial resolution to examine time trends 2000‒2018, compliance with the WHO Air Quality Guidelines, and assess the health impact. The modelling area covered six metropolitan areas in Sweden with a combined population of 5.5 million. Long-range transported air pollutants were modelled using a chemical transport model with bias correction, and locally emitted air pollutants using source-specific Gaussian-type dispersion models at resolutions up to 50 × 50 m. The modelled concentrations were validated using quality-controlled monitoring data. Lastly, we estimated the reduction in mortality associated with the decrease in population exposure. The validity of modelled air pollutant concentrations was good (R² for PM_2.5 0.84, PM₁₀ 0.61, and NO_x 0.87). Air pollution exposure decreased substantially, from a population weighted mean exposure to PM_2.5 of 12.2 µg m⁻³ in 2000 to 5.4 µg m⁻³ in 2018. We estimated that the decreased exposure was associated with a reduction of 2719 (95% CI 2046–3055) premature deaths annually. However, in 2018, 65%, 8%, and 42% of residents in the modelled areas were still exposed to PM_2.5, PM₁₀, or NO₂ levels, respectively, that exceeded the current WHO Air Quality Guidelines for annual average exposure. This emphasises the potential public health benefits of reductions in air pollution emissions.

环境空气污染仍然是疾病的主要环境原因。准确评估长时间内的人群暴露和小范围空间暴露变化对于流行病学研究至关重要。我们以高空间分辨率估算了细颗粒物和粗颗粒物（PM _2.5、PM ₁₀）和氮氧化物（NO _x、NO ₂）的年暴露量，以检查 2000-2018 年的时间趋势、符合世界卫生组织空气质量指南，并评估健康影响。建模区域覆盖瑞典六个大都市区，总人口达 550 万。使用具有偏差校正的化学传输模型对远距离传输的空气污染物进行建模，并使用分辨率高达 50 × 50 m 的特定源高斯型扩散模型对局部排放的空气污染物进行建模。使用质量控制的监测数据验证了建模的浓度。最后，我们估计了与人口暴露减少相关的死亡率下降。模型空气污染物浓度的有效性良好（PM _2.5 0.84、PM ₁₀ 0.61 和 NO _{x 0.87 的 R} ²）。空气污染暴露大幅下降，从 2000 年人口加权平均 PM _2.5暴露量 12.2 µg m ^-3降至 2018 年 5.4 µg m ^-3。我们估计暴露量的减少与 2719 的减少有关（95% CI 2046 –3055) 每年过早死亡。然而，2018年，模拟地区仍有65%、8%和42%的居民分别暴露在PM _2.5、PM ₁₀或NO ₂水平下，超出了现行世界卫生组织空气质量指南的年平均暴露水平。这强调了减少空气污染排放对公共健康的潜在好处。