Severe Pollution in China Amplified by Atmospheric Moisture

In recent years, severe haze events often occurred in China, causing serious environmental problems. The mechanisms responsible for the haze formation, however, are still not well understood, hindering the forecast and mitigation of haze pollution. Our study of the 2012–13 winter haze events in Beijing shows that atmospheric water vapour plays a critical role in enhancing the heavy haze events. Under weak solar radiation and stagnant moist meteorological conditions in winter, air pollutants and water vapour accumulate in a shallow planetary boundary layer (PBL). A positive feedback cycle is triggered resulting in the formation of heavy haze: (1) the dispersal of water vapour is constrained by the shallow PBL, leading to an increase in relative humidity (RH); (2) the high RH induces an increase of aerosol particle size by enhanced hygroscopic growth and multiphase reactions to increase particle size and mass, which results in (3) further dimming and decrease of PBL height, and thus further depressing of aerosol and water vapour in a very shallow PBL. This positive feedback constitutes a self-amplification mechanism in which water vapour leads to a trapping and massive increase of particulate matter in the near-surface air to which people are exposed with severe health hazards.

The mass concentration of PM2.5 was measured by an Element Oscillating Microbalance (TEOM; Thermo Scientific Co., USA) instrument, with the model of R&P 1400a, which is tapered with a 2.5 μm cyclone inlet and an inlet humidity control system. The instrument was housed in an air-conditioned room and was operated with a hydrophobic filter material to reduce the humidity of incoming sampled air. The sample stream was preheated before entering the mass transducer, and semi-volatiles and water were not measured. The filter loading percentage and flow rates of TEOM were checked once a week, and the filter was replaced when the filter loading percentage was greater than 30% 1 .
A micro-pulse lidar (the model of MPL-4B, Sigmaspace Co., USA) was employed to study the evolution of PBL. The pulse repetition frequency of the MPL is 2500 Hz, with a wavelength of 532 nm of the laser beam. The peak value of the optical energy of laser beam is 8 μJ. The pulse duration was set to 100 ns, and the pulse interval was set to 200 ns, corresponding to a spatial resolution of 30 meters. The PBL height is detected by the sharp changes of the Lidar signals. The analysis of the PBL height by Quan et al 2 . suggests that the MPL instrument is only suitable for measuring the PBL height during the daytime. As a result, we will limit our analysis of PBL height from 8:00 to 18:00.
The gas chemical species were measured by several instruments. Nitrogen oxides (NO-NO2-NOx) were measured with a chemiluminescent trace level analyzer (TEI; Model 42iTL). The analyzer has a detection limit of 0.025 ppbv. Carbon monoxide (CO) was measured by an enhanced CO analyzer, with the Model 48iTL. The instrument uses a gas filter correlation technology, with a detection limit of 0.04 ppmv. Sulfur dioxide (SO2) was measured by a pulsed UV fluorescence analyzer (TEI; Model 43 i-TLE). The detection limit for the analyzer is 0.05 ppbv for 2-min integration with a precision of about 0.20 ppbv. Ozone (O3) was measured with a UV photometric analyzer (TEI; Model 49iTL), with a detectable limit of 0.05 ppbv.
Atmospheric visibility was observed by a PWD20 instrument (Vaisala Co., Finland).
The detecting range of visibility is from 10 to 20,000 meters. Other important meteorology variables (such as air temperature, relative humidity (RH), pressure, wind speed, and wind direction) were measured by WXT-510 (Vaisala Co., Finland).
The chemical composition of aerosol particles was measured by an Aerodyne   In this study, the default TUV model is modified for the calculation of particle effect on the solar radiation on the surface. The default model uses a vertical profile of aerosol optical depth (AOD), which represents a background clean air condition. In order to represent the aerosol condition in the polluted Beijing region, the AOD is calculated by the aerosol concentration measured in the P4 episode (as shown in Fig.   1A). The calculation is made based on the following steps and assumptions. Dd is the diameter of dry particles. The conversion between the mass to number concentrations are made by the assumption that the aerosol number distribution is following the Gaussian distribution 7 , with a mean radius and standard deviation 0.15 and 2.20 µm, respectively. An amplification factor is applied to the dry AOD accordingly. From the daily radio sounding data (at 8:00 and 20:00 local time), one can see that during the haze event, the vertical distribution of RH within the first 1 km (PBL or higher) was rather constant, as high as on the surface or even slightly increase in some cases (Fig. S1). In the current calculation, we thus treat the RH vertically homogeneous within the PBL (≤ 1 km). where H is the PBLH, P is the Pasquill stability level. Defined by six classes from A to F in terms of increasing order from very unstable (A), moderately unstable (B), slightly unstable (C), neutral (D), slightly stable (E) to moderately stable (F) 9 . T is the surface air temperature, Td is the surface air dew-point temperature, Uz is the mean wind speed (m s -1 ) at height of Z (Z=10 m), f is the Coriolis parameter (s -1 ), z0 is the surface roughness length (0.5 m in this study), Ω is the (7.29 × 10 -5 rad s -1 ). T, Td, and P are factors considering thermodynamic mechanism, while Uz and z0 are taking turbulence and dynamic mechanisms into account.
According to Wallace and Hobbs 10 , in the equation, T-Td can be replaced by (100-