The pasture’s productivity simulated by the third-order polynomial growth curve mathematicalmodel in different years in Maqu grassland coincided with the actualvalue. The simulation results show that themaximum growth rate is70. 2 kg/hm2·d and the minimum is1. 6 kg/hm2·d. Themonthlymaximum yield is2 042. 5 kg/hm2and theminimum is57. 6 kg/hm2. FromMay to September, the 10 days accumulated temperaturewhich is above 0℃increased 1℃, the pasture’s drymatter increased 2. 54 kg/hm2, and the amount of the precipitation increased every 1 mm in every 10 days, the drymatter increased 7. 33 kg/hm2, aswell as the sunshine duration increased every one hour in every 10 days, the drymatter increased 3. 83 kg/hm2. Considering the difference of the climate
tendency for different factors, the ten days precipitation is themain factorwhich influences the pasture’s yield, and the accumulated temperature in every 10 days plays the second role. The climate resources utilization rate in high production year is higher than that in the low production year.

Based on temperature, p recip itation, sunshine hours, evaporation, ≥0 ℃ accumulated temperature and pasture’s yield data from Ap ril to Sep tember during 1985 - 2005 inMaqu grassland, the relation between meteorological factors and pasture’s yield was analyzed through the method of fuzzy cluster analysis. Results indicated that the configuration of light, heat and water has an obvious influence on pasture’s yield, and the pasture’s yield was ralated to ≥0 ℃ accumulated temperature. The gray - relation analysis showed that the p recip itation in Ap ril is crucial to pasture’s recovery and the temperature in June affects pasture’s tiller and heading most, and the p recip itation in July hasmarked influence on pasture’s heading and blooming. Both sunshine hours inMay, July and p recip itation in Sep tember has an influence on pasture’s growth. The negative or accumulated effects of climate change resulted in decrease of pasture’s yield inMaqu grassland. Due to the comp licated climates environment system, itwill have a p ractical significance to set up forecastmodel to forecast pasture’s yield inMaqu grassland.

The variations of the accumulated temperature and the sunlight hours inMaqu grassland in recent 40 yearswere analyzed by using a non - parameterMann - Kendall testmethod and wavelet analysis. The results show that annual variation of accumulated temperature tended to increase, and the tendency rate of ≥ 0 ℃ accumulated temperature is 59. 5 ℃/10 a, and 33. 7 ℃/10 a for ≥ 10 ℃ accumulated temperature. The period corresponding to high oscillation center of ≥10 ℃ accumulated temperature is two yearsmore than that of ≥0 ℃ accumulated temperature; the low and high frequency oscillation tended to increase at the same time. The annual trend of sunlight hours is also increasing, and the tendency rate is 40. 2 h /10 a. By Spearman’s order correlation analysis, we know that the positive matching characteristic of light, heat and water is remarkable; the synchronous changes p resented inMay to Sep tember forwater and heat, and June to July for light and water. The app rop riate p recip itation, higher air temperature and the adequate sunshine are helpful to pasture’s growth ofMaqu grassland.

By using the data of temperature，rainfall during 1967—2005 and soil water content during 1996—2005 from Maqu County，which is located in the upper stream of the Yelow River and is an important water replenishment region for the Yelow River，the cli-mate change and its impact on water resource there are analyzed ．Results show that the annual trend oftemperature is rising but rainfall is declining，and climate change results in the potential evaporation increase and the available water resources decrease．Th e polynomi·al fitting curve ofwater resource presented fluctuation wi出a two—vale an d one—peak．and the time scale ofannual chan ge in low water period prolonged doub led．