科研之友
微信
新浪微博
Facebook
分享链接
摘要
Variations in snow cover and vegetation in the Province of Quebec, Canada, were characterized for a transect spanning from 50 degrees N to 60 degrees N during the International Polar Year field campaign of February 2008. The main objective of this study was to compare measured (AMSR-E) and modeled (MEMLS) brightness temperature (T-b) and to analyze differences in the in situ measurement of snow water equivalent (SWE) and vegetation. Sampling involved detailed snow measurements on the ground in four different ecological environments. Measured and modeled SWE were compared using a thermodynamic multilayered snow model (SNOWPACK) driven with North American Regional Reanalysis (NARR) data. The root mean square error (RMSE) of modeled data compared with measurements was 63 mm (30%). The simulated SWE was generally underestimated throughout the transect but stayed within the large standard deviation observed for measured SWE. In situ snow measurements were used as input to a microwave emission model (MEMLS) to simulate T-b. An innovative approach using calibrated near-infrared reflectance photographs was used to characterize the effective snow grain-size parameter needed for the radiative transfer model. Although some results provided T-b predictions similar to AMSR-E data for certain areas, large differences remained for the majority of sampling sites. The derived RMSE of 16 K and 32 K, respectively, for 18.7 and 36.5 GHz (vertical polarization) throughout the transect cannot be explained solely in terms of grain-size variations introduced into the simulations. Local variability in snow structure and thickness produced large variability (up to 60 K within one AMSR-E pixel) compared with AMSR-E T-b throughout the transect (15 K for 18.7 GHz and 35 K for 36.5 GHz).
- 出版日期2010
