The cloud—water resources of the Qilian Mountain were analyzed by using the observed data of eight meteorological stations from 1960 to 2002 and the NCEP／NCAR data(2．5。×2．5。)for the period of 1970—1997．Results show that the water vapor amount gradually increased from spring to summer，and then decreased；the precipitation there was concentrated in May to Septembe r which was in the proportion of 86．8％ to annual precipitation．Comparing the preceding period (1970—1986)with the succeeding period (1988—1997)of climatic change in Northwest China，the annual mean precipitation there increased more than 8．1％ in the succee—ding period (mainly in spring and summer)，which was beneficial to the ecological improvement there．The precipitation in the middle part of the Qilian Mountain was much more than that in eastern and western part．The annual water vapor input was 885．4×10 m ．which mostly cflme from the levels under 6O0 hPa，an d it divergenced over 600 hPa at most time of a year．The water vapor mainly canle from longitudinal input．

The cloud—water resources of the Qilian Mountain were analyzed by using the observed data of eight meteorological stations from 1960 to 2002 and the NCEP／NCAR data(2．5。×2．5。)for the period of 1970—1997．Results show that the water vapor amount gradually increased from spring to summer，and then decreased；the precipitation there was concentrated in May to Septembe r which was in the proportion of 86．8％ to annual precipitation．Comparing the preceding period (1970—1986)with the succeeding period (1988—1997)of climatic change in Northwest China，the annual mean precipitation there increased more than 8．1％ in the succeeding period (mainly in spring and summer)，which was beneficial to the ecological improvement there．The precipitation in the middle part of the Qilian Mountain was much more than that in eastern and western part．The annual water vapor input was 885．4×10 m ．which mostly cflme from the levels under 6O0 hPa，an d it divergenced over 600 hPa at most time of a year．Th e water vapor mainly canle from longitudinal input．

According to "the criterion of surface meteorological observation"，visibility and wind speed are important targets to classify dust一stormy weather.  Using the data of visibility and wind speed observed in meteorological stations，we try to construct duststormy weather index to determine the intensity of dust一stormy weather phenomenon，moreover we can measure the strong or weak state of dust一stormy weather happened anywhere.  We have obtained thefollowing conclusions ahe dust一sand index can corre}ond with the intensity of dust一stormy weather ;our analysis to state of dust一stormy weather in different area is in corre}ondence with real weather facts.  The correlative analysis between sand一dust index and urban airpollution index indicated their affinity. It ispossible to build urban air pollution statistical forecasting model with the sand一dust index.