干旱区生态环境知识资源中心

Precipitable water vapor (PWV) was retrieved using direct solar irradiance at 938 nm measured by a multifilter rotating shadowband radiometer (MFRSR) at the Semi-Arid Climate and Environment Observatory of Lanzhou University (SACOL) located in the semi-arid area of northwest China from August 2007 to June 2010. Measurement also occurred at Zhangye, China, at the Atmosphere Radiation Measurements (ARM) Program’s Ancillary Facility during the dust period from April to June 2008. The line-by-line radiative transfer model (LBLRTM) code combined with the HITRAN 2004 spectral database is used to model the water vapor spectral transmittance throughout the 938-nm spectral response of MFRSR in the retrieval algorithm. Gaussian fitting is proposed to determine the daily calibration constant at the top of atmosphere for a long-term series under an obvious annual change in solar radiation.

PWV retrieved by MFRSR over SACOL shows that 90% of PWV values are smaller than 1.52 cm, and PWV distribution has a seasonal variation, with maximum in summer and minimum in winter. The comparisons between MFRSR and other measurements show a better agreement between MFRSR and sunphotometer (AERONET’s Cimel) PWV retrievals with relative bias of 2.9% and RMS difference of 9.1% than between MFRSR and microwave radiometer (MWR) with relative bias of 10% and RMS difference of 23% over SACOL, and an excellent agreement between MFRSR and sunphotometer with relative bias of 0.56% and RMS difference of 6.1% over Zhangye. To verify satellite PWV products over the semi-arid area of northwest China, the comparisons of PWV from MODIS and AIRS with MFRSR suggest that the agreement between satellite and MFRSR PWV retrievals is not as good as that between MFRSR and other ground-based instruments. MODIS appears to slightly underestimate PWV in a dry atmosphere but overestimate PWV in a moist atmosphere against MFRSR A method is proposed to correct MODIS PWV products. AIRS PWV products relative to MFRSR show a systematic underestimation. (C) 2015 Elsevier B.V. All rights reserved.