| In the context of global warming,snow cover has received close attention as an element closely related to global climate change.The Tibetan Plateau is the plateau which has the highest elevation as well as the widest snow cover in the mid-latitudes of the northern hemisphere.The strong radiative forcing caused by snow has a huge impact on monsoons and climate systems in Asia and the world.MODIS snow products have high temporal and spatial resolution and are widely used in snow monitoring of the Tibetan Plateau.However,the product is blocked by clouds and there are a lot of data gaps,and the accuracy is low in mountainous areas with complex terrain such as the Tibetan Plateau.Based on a snow modeling algorithm under the framework of Hidden Markov Random Field(HMRF),this research proposes a snow spatio-temporal modeling algorithm based on "spectrum-spatio-temporal-terrain".The algorithm optimally combines the spectral background information,spatio-temporal background information of time-series MODIS snow products,and the terrain background information of the Tibetan Plateau to fill the data gaps in the original MODIS snow products.The algorithm adds solar radiation energy as terrain background information to the model,thereby adding corrections for the influence of complex terrain factors such as mountain shadows,elevation differences,and sunlight slope incident angles on snow distribution in the snow modeling.According to this algorithm,this study produced a long-term sequence of 17snow-covered years from 2002 to 2018,with no data vacancy snow products on the Tibetan Plateau.The product has a spatial resolution of 500 m and a temporal resolution of 1 day.The accuracy of the snow product was verified and evaluated using the snow depth data of the actual site and the Landsat-8 remote sensing image with a spatial resolution of 30 m.The accuracy evaluation and comparison of the snow product and the snow product produced by the HMRF snow modeling algorithm were carried out.Then,based on the snow products produced on the Tibetan Plateau,this study extracted the snow cover rate and snow phenology information,and temporal and spatial analysis was carried out.Correlation analysis between meteorological factors(including air temperature and precipitation rate)and the snow product were carried out too.The main conclusions are as follows:(1)Using the "spectrum-spatio-temporal-terrain" snow e modeling algorithm to fill in the data gaps in the MODIS snow products on the Tibetan Plateau,it can completely fill the data gaps in the original MODIS snow products.The overall accuracy during 2002-2018 hydrological year verified by the snow depth data of the measured site is up to 97.18%,and the overall accuracy during 2016-2018 hydrological year verified by the Landsat-8 snow recognition result is 81.26%.After analysis,the modeling algorithm can significantly improve the overall accuracy of the initial snow accumulation,the snow melting period and the nonsnow period,as well as the accuracy of low altitude areas,steep slope areas,sunny slope areas,and evergreen coniferous forest areas.(2)The snow cover on the Tibetan Plateau decreased slowly during the 2002-2018 snow cover years.Altitude impacted the snow cover most in the time preiod.The higher the altitude,the more snow cover days,the earlier the snow start time,the later the snow end time,and the longer the snow season lasts.In addition,the lake basin of the Qiangtang Plateau had little snow accumulation comparied with other area.Almost no snow accumulates in the Qaidam Basin throughout the year.In terms of time changes,the overall snow index of the Tibetan Plateau during this period of time changed slightly,and the snow cover showed a slow decreasing trend.In terms of spatial changes,snow phenology has significant changes in a few high-altitude areas,and almost no significant changes in low-altitude areas.(3)The response relationship between snow phenology and meteorological factors in the Tibetan Plateau is closely related to climate zoning.In the sub-frigid climate zone of the plateau,temperature is positively correlated with the number of days covered by snow and the duration of the snow season.In the temperate and arid regions of the plateau,there is a negative correlation.The response relationship between precipitation rate and snow phenology of the Tibetan Plateau is different in each climatic region.The higher the relative elevation of snow cover days and snow season duration in each climatic region,the more likely it is to show a negative correlation;The snow start date and snow end date are opposite. |
PDF Full Text Request