Author/Authors :
Liangfu Chen a، نويسنده , , b، نويسنده , , *، نويسنده , , Yanhua Gao a، نويسنده , , c، نويسنده , , Lei Yang a، نويسنده , , d، نويسنده , , Qinhuo Liu، نويسنده , ,
Xingfa Gua، نويسنده , , b، نويسنده , , Guoliang Tian a، نويسنده , , b، نويسنده ,
Abstract :
In this paper, a MODIS-derived daily PAR (photosynthetically active radiation) simulation model from cloud-free image over land
surface has been developed based on Bird and Riordan’s model. In this model, the total downwelling spectral surface irradiance is
divided into two parts: one is beam irradiance, and another is diffuse irradiance. The attenuation of solar beam irradiance comprises
scattering by the gas mixture, absorption by ozone, the gas mixture and water vapor, and scattering and absorption by aerosols. The
diffuse irradiance is scattered out of the direct beam and towards the surface. The multiple ground–air interactions have been taken into
account in the diffuse irradiance model. The parameters needed in this model are atmospheric water vapor content, aerosol optical thickness
and spectral albedo ranging from 400 nm to 700 nm. They are all retrieved from MODIS data. Then, the instantaneous photosynthetically
available radiation (IPAR) is integrated by using a weighted sum at each of the visible MODIS wavebands. Finally, a daily
PAR is derived by integration of IPAR. In order to validate the MODIS-derived PAR model, we compared the field PAR measurements
in 2003 and 2004 against the simulated PAR. The measurements were made at the Qianyanzhou ecological experimental station, Chinese
Ecosystem Research Network. A total of 54 days of cloud-free MODIS L1B level images were used for the PAR simulation. Our results
show that the simulated PAR is consistent with field measurements, where the correlation coefficient of linear regression between calculated
PAR and measured PAR is 0.93396. However, there were some uncertainties in the comparison of 1 km pixel PAR with the tower
flux stand measurement.
2007 Elsevier Ltd. All rights reserved.