The cloud variations under subtropical high(STH) conditions during summers over a ten-year period are studied using combined data from the International Satellite Cloud Climatology Project and the National Centers for...The cloud variations under subtropical high(STH) conditions during summers over a ten-year period are studied using combined data from the International Satellite Cloud Climatology Project and the National Centers for Environmental Prediction.The results reveal that clouds mainly experience an isolated evolution in the STHs,which is designated in this study by the 1540 gpm geopotential lines at 850 hPa.In the STH domain throughout the Northern Hemisphere,the average amount of total clouds exceeds 30%.Low clouds dominate in the STH domain,contributing over 60%of total cloud amount within the Pacific subtropical high and over 40%within the Atlantic subtropical high.The prevalence of low clouds in above regions is determined by the circulation pattern around 150°-180°E and 850 hPa,which suppresses both the upward development of the cloud tops and the water vapor divergences near the surface.Furthermore,clouds present great geographical incoherence within the STH domain.In the eastern STHs,the amount of middle and low clouds increases to peak in the early morning and decreases to a trough in the afternoon,while the amount of high clouds remains stable throughout the day.Conversely,in the western STHs,the diurnal amplitude of low and middle clouds is less than three,while high clouds dramatically reach the maximum in the afternoon and drop to the minimum in the evening.Among the nine cloud categories,stratocumulus clouds with greater optical thickness account for the most under STH conditions,no matter their occurrence or amount,causing more shortwave cloud radiative forcing to cool the local atmosphere and surface as a consequence.展开更多
This paper analyzes the correlation between variations of total and low cloud amounts and the varying features of aerosols related to urban development of Beijing by using the cubic spline fitting method based on the ...This paper analyzes the correlation between variations of total and low cloud amounts and the varying features of aerosols related to urban development of Beijing by using the cubic spline fitting method based on the monthly meteorological data of temperature,humidity,precipitation,clouds,and aerosol optical depth （AOD） during 1950-2005.The statistics on the development of the city of Beijing in this period,including the total industrial output,population,residential housing development,highway construction,charcoal production,etc.,is revealed.Accompanying the rapid urban development of Beijing over the past 55 years or so,the urban aerosol concentration and composition have changed.The results indicate that：1） there is a general trend of climate warming and drying in Beijing;2） the total cloud amount in all seasons declines drastically,but lower cloud amount climbs up slightly;3） the high correlations between cloud amount and the indices of Beijing urban development such as the housing area,charcoal production,and road construction show that the variation of cloud amount is closely related to the urban development;4） the changing trend of AOD coincides more closely with the variation of low cloud amount.The evident drop of total cloud amount is in agreement with the trend of climate warming and drying,while the slight growth of low cloud amount is likely caused by more haze and fog occurrences in the lower troposphere in association with the pollution responsible for the＂darkening＂of Beijing and the surrounding areas.展开更多
Topic modeling is a mainstream and effective technology to deal with text data, with wide applications in text analysis, natural language, personalized recommendation, computer vision, etc. Among all the known topic m...Topic modeling is a mainstream and effective technology to deal with text data, with wide applications in text analysis, natural language, personalized recommendation, computer vision, etc. Among all the known topic models, supervised Latent Dirichlet Allocation （sLDA） is acknowledged as a popular and competitive supervised topic model. How- ever, the gradual increase of the scale of datasets makes sLDA more and more inefficient and time-consuming, and limits its applications in a very narrow range. To solve it, a parallel online sLDA, named PO-sLDA （Parallel and Online sLDA）, is proposed in this study. It uses the stochastic variational inference as the learning method to make the training procedure more rapid and efficient, and a parallel computing mechanism implemented via the MapReduce framework is proposed to promote the capacity of cloud computing and big data processing. The online training capacity supported by PO-sLDA expands the application scope of this approach, making it instrumental for real-life applications with high real-time demand. The validation using two datasets with different sizes shows that the proposed approach has the comparative accuracy as the sLDA and can efficiently accelerate the training procedure. Moreover, its good convergence and online training capacity make it lucrative for the large-scale text data analyzing and processing.展开更多
基金supported by the Special Funds for Public Welfare of China(Grant No.GYHY-QX-2007)the National Natural Science Foundation of China(Grant Nos.40730950,40675027,and 40805007).
文摘The cloud variations under subtropical high(STH) conditions during summers over a ten-year period are studied using combined data from the International Satellite Cloud Climatology Project and the National Centers for Environmental Prediction.The results reveal that clouds mainly experience an isolated evolution in the STHs,which is designated in this study by the 1540 gpm geopotential lines at 850 hPa.In the STH domain throughout the Northern Hemisphere,the average amount of total clouds exceeds 30%.Low clouds dominate in the STH domain,contributing over 60%of total cloud amount within the Pacific subtropical high and over 40%within the Atlantic subtropical high.The prevalence of low clouds in above regions is determined by the circulation pattern around 150°-180°E and 850 hPa,which suppresses both the upward development of the cloud tops and the water vapor divergences near the surface.Furthermore,clouds present great geographical incoherence within the STH domain.In the eastern STHs,the amount of middle and low clouds increases to peak in the early morning and decreases to a trough in the afternoon,while the amount of high clouds remains stable throughout the day.Conversely,in the western STHs,the diurnal amplitude of low and middle clouds is less than three,while high clouds dramatically reach the maximum in the afternoon and drop to the minimum in the evening.Among the nine cloud categories,stratocumulus clouds with greater optical thickness account for the most under STH conditions,no matter their occurrence or amount,causing more shortwave cloud radiative forcing to cool the local atmosphere and surface as a consequence.
基金Supported by the Special Grant in Atmospheric Sciences of the China Meteorological Administration(GYHY200706036)the National"973"Program of China(2011CB403404)the International Cooperation Project on Monsoon Monitoring (200 9DFA21430)
文摘This paper analyzes the correlation between variations of total and low cloud amounts and the varying features of aerosols related to urban development of Beijing by using the cubic spline fitting method based on the monthly meteorological data of temperature,humidity,precipitation,clouds,and aerosol optical depth （AOD） during 1950-2005.The statistics on the development of the city of Beijing in this period,including the total industrial output,population,residential housing development,highway construction,charcoal production,etc.,is revealed.Accompanying the rapid urban development of Beijing over the past 55 years or so,the urban aerosol concentration and composition have changed.The results indicate that：1） there is a general trend of climate warming and drying in Beijing;2） the total cloud amount in all seasons declines drastically,but lower cloud amount climbs up slightly;3） the high correlations between cloud amount and the indices of Beijing urban development such as the housing area,charcoal production,and road construction show that the variation of cloud amount is closely related to the urban development;4） the changing trend of AOD coincides more closely with the variation of low cloud amount.The evident drop of total cloud amount is in agreement with the trend of climate warming and drying,while the slight growth of low cloud amount is likely caused by more haze and fog occurrences in the lower troposphere in association with the pollution responsible for the＂darkening＂of Beijing and the surrounding areas.
基金This work was supported in part by the National Natural Science Foundation of China under Grant Nos. 61572226 and 61876069, and the Key Scientific and Technological Research and Development Project of Jilin Province of China under Grant Nos. 20180201067GX and 20180201044GX.
文摘Topic modeling is a mainstream and effective technology to deal with text data, with wide applications in text analysis, natural language, personalized recommendation, computer vision, etc. Among all the known topic models, supervised Latent Dirichlet Allocation （sLDA） is acknowledged as a popular and competitive supervised topic model. How- ever, the gradual increase of the scale of datasets makes sLDA more and more inefficient and time-consuming, and limits its applications in a very narrow range. To solve it, a parallel online sLDA, named PO-sLDA （Parallel and Online sLDA）, is proposed in this study. It uses the stochastic variational inference as the learning method to make the training procedure more rapid and efficient, and a parallel computing mechanism implemented via the MapReduce framework is proposed to promote the capacity of cloud computing and big data processing. The online training capacity supported by PO-sLDA expands the application scope of this approach, making it instrumental for real-life applications with high real-time demand. The validation using two datasets with different sizes shows that the proposed approach has the comparative accuracy as the sLDA and can efficiently accelerate the training procedure. Moreover, its good convergence and online training capacity make it lucrative for the large-scale text data analyzing and processing.
