Abstract:Based on the daily temperature observation data of 5 stations in Shijiazhuang from 2008 to 2010，the average，extreme values and daily ranges of grass land temperature，as well as relationships between grass land temperature，ground temperature and air temperature under various meteorological conditions were analyzed．The results show that annual average of grass land temperature presented low in the north and high in the south，and highest in the south－west mountainous region．The maximum of annual mean grass land temperature was highest in the central region，the minimum of annual mean grass land temperature was lowest in the southwest mountainous region．The daily range of grass land temperature was affected by latitude，season and geographical conditions，and the diurnal amplitude was biggest in the central region．The diurnal amplitude of ground temperature was maximum，but minimum for air temperature．There was a positive correlation between grass land temperature and air temperature，ground temperature in a little cloudy days．The absolute difference of annual average value between ground temperature and grass land temperature was maximum when ground was covered with snow，but minimum in rainy days．

Based on the data ofdaily demand for electricity from June toAugustduring the period of2005- 2007 in Shijiazhuang, the meteorological electric quantity variedwithweather factorswas setapart, and the correlation coefficients betweenmeteorologicalelectric quantity on rainy days or cleardays and keymeteorological factors on the same periodwere calculated respectively, especially themeteorological electric quantity changing ruleswith temperature, humidity and precipitationwere studied. The results show that themeteorological electric quantity correlated positivelywith temperature, butnegativelywith relative humidity on rainy days. The correlation coef-
ficients between daily change of electric quantity and daily variation ofmeteorological elementswere calculated also, and itwas found that they had good correlation. Based on statistical analysis, themultiple regression predictionmodels forweather electric quantity and electric quantity daily change under the influence of compositiveweather elements on rainy days and cleardays respectivelywere established by usingOrigin 7. 5 software. The regression statistics and VAR analysis both indicate that all equations have passed the F test
withα=0. 0005. Themultiple correlation is higher than single correlation, and the consistent rate is highwhichmeans thismodel can be used as a reference for electric power department to attemper the electric power reasonably.