Abstract. Estimating ambient PM_2.5 (fine particulate matter) concentrations in India over many years is challenging because spatial coverage of ground-based monitoring, while better recently, is still inadequate and satellite-based assessment lacks temporal continuity. Here we analyze MERRA-2 reanalysis aerosol products to estimate PM_2.5 at hourly scale to fill the space-time sampling gap. MERRA-2 PM_2.5 are calibrated and validated (r = 0.94, slope of the regression = 0.99) against coincident in-situ measurements. We present the first space-time variability of ambient PM_2.5 diurnal pattern in India for an 18-year (2000–2017) period. Diurnal amplitude is found to be quite large (> 30 μg m⁻³) in the Indo-Gangetic Basin (IGB) and western arid regions of India. PM_2.5 is found to decrease over the western dust source region and increase over the Himalayan foothills and parts of IGB and central India primarily in the morning and evening hours. This increasing trend at an annual scale is primarily governed by a large increase in concentration during Oct–Feb that can be attributed to a combination of the rise in emission and declining boundary layer height. Our results suggest that the satellite-based concentration estimates (typically representative of late morning to early afternoon hours) are lower (magnitude depends on the place and season) than the 24-hour average concentration in most parts of India. In the future, the integration of reanalysis data in concentration modeling may assist in reducing the uncertainty in estimates of air pollution concentration patterns in India and elsewhere.