Retrieved of Snow Grain Size from MODIS Over Qiyi Glacier, Qilian Mountain

The 500 m resolution Moderate Resolution Imaging Spectroradiometer(MODIS) L1B scenes over Qiyi glacier,Qilian Mountain is used to estimate the snow grain size.The retrieval is based on ratios of near-infrared albedo to visible albedo which predicts snow reflectance as a function of snowpack grain size and ice absorption.Bow-tieeffects were removed by the ENVI software.Atmospheric effects are taken into account and corrected through the MODTRAN4+ model.By this approach snow optical grain size was determined by the volume-to-surface area ratio on account of non-oriented spheroids produce the same scattering results as spheres.The time series analysis of derived grain size shows a good sensitivity to snow melting and snow precipitation events.This inversion technique has been validated using a combination of ground-based grain size measurements derived from the handle lenses which measure the snow grain size over the top 1 cm layer.The retrievals correlate well with measurements when radii is in range of-0.1-1 mm,although retrieved optical snow grain size general is lower than physical measured grain size,the model may be need about a factor of 1.1.As part of validation analysis for the Qiyi glacier,the retrieved snow grain size from MODIS over selected sites in the period of September 7-15 in 2011 was compared to the temperature measurement of the automatic weather station near the selected sites.The relationship between the measured snow grain size,simulated snow grain size and average temperature patterns suggests that rapid destructive metamorphism of the fresh snow occurred when temperatures were near 0℃,once temperatures were higher than 0℃ and kept for a long time,the snow particle size change was evident,snow particle size increased significantly,and vice versa.But in this prossess,snow particle size also existed evolution process of deformation;snow particle size and temperature showed significantly positive correlation.The future work is to study the better cloud mask algorithm to remove the cloud effects and build the long and continuous time series of the snow grain size.