Spatial coherence of nucleation and ultrafine particle concentrations over eastern North America (Invited)

Particle size distributions from sites in eastern North America imply new particle formation (NPF) is frequently observed, however the absence of an observing network precludes direct estimation of spatial scales of NPF. Inferences regarding the spatial scales of NPF and resulting ultrafine particle (UFP, i.e. sub-100 nm diameter) concentrations are presented based on in situ and remote sensing observations along with simulations conducted with the GLOMAP global chemistry model and the WRF-CHEM regional chemistry model. Analyses of in situ measurements indicate; (i) NPF and UFP concentrations in two long-term data sets (one collected at a site in southern Indiana (MMSF) and the other 1500 km to the northeast, close to Toronto) indicate a high probability of simultaneous NPF and high coherence of UFP concentrations. (ii) The spatial extent of NPF estimated from the temporal duration of events at MMSF and back-trajectories show minimum spatial scales of 340 to 1000 km. (iii) Co-occurrence of NPF along an 80 km transect in southern Indiana but site-to-site variability in UFP concentrations with data from the Indianapolis metropolitan area differing substantially from two non-urban locations. The implied regional scale NPF and relatively high spatial correlation of UFP concentrations are used in part to justify development of a physically-consistent, but statistically-based, proxy for UFP concentrations where the predictors are spatially averaged remote sensing observations. The proxy is built on data from MMSF and includes predictors known to control NPF; Aerosol Optical Depth and Ångstrom exponent from MODIS (as proxies for the condensational sink), and ultraviolet radiation flux and SO2 concentrations from the Ozone Monitoring Instrument, and NH3 concentrations from Tropospheric Emission Spectrometer as indicators of ternary nucleation. The model is applied on a grid of 100x100 km across the whole of eastern North America and the resulting UFP concentrations are cross-compared with output from GLOMAP. The results indicate: (i) Consistent with in situ observations, there is evidence that UFP concentrations, and by association NPF events, are relatively coherent across large spatial scales over eastern North America. (ii) Consistent with in situ observations, NPF intensity peaks during spring and summer. (iii) Satellite-derived estimates of UFP concentrations exhibit a high degree of coherence with GLOMAP in terms of seasonality and spatial patterns. But there is a systematic positive bias in the proxy estimates that may be due to differences in the nucleation mechanism applied, background concentrations associated with primary emissions and/or errors in the condensation and coagulation sinks. To evaluate the spatial variability (i) in UFP concentrations at scales below those represented in the satellite proxy and/or GLOMAP and (ii) of UFP that can be attributed to causes not linked to NPF, we are conducting simulations at 12 km with WRF-CHEM. Initial results suggest a substantial fraction of the non-NPF-related variability in near-surface total UFP concentrations on scales below 100x100 km can be attributed to variations in Aitken mode primary emissions and/or boundary layer dynamics determined by land-use patterns.