| In this paper, strong alkaline soil respiration rate has been measured in spring, summer, autumn under sunny weather condition by using LI-8100 soil carbon flux meter, temperature (air temperature and soil temperature), humidity (relative humidity and soil moisture) were also recorded simultaneously, and the correlation among above factors has been analyzed to identify the key factors influencing soil respiration in different season, and establish a variety of regression models. On the basis of the accuracy test and simple principle, the optimal model to predict the seasonal soil respiration has been selected. The results show that: (1) Although the temperature and humidity are the main factors affecting strong alkaline soil respiration in different seasons, but the temperature has relatively larger effect, where the air temperature is the biggest direct impact factor on soil respiration in spring and autumn, while in summer, the soil temperature is the greatest direct impact factor, and soil moisture is the largest indirect factor in all season. (2) The best prediction model of soil respiration in spring and autumn is the two-factor equation composed of air temperatures at 10 cm layer and soil moisture, which has a smaller RMSE (0.159 and 0.259, respectively), and the RPD>2(2.9,2.094, respectively), has a better predictive accuracy. The best soil respiration predictive model for summer season is composed of air temperature at 10 cm, surface temperature, relative humidity and soil moisture, RMSE is 0.248, RPD>2(of 2.406). (3) The trend of soil respiration variation and their impact factors was similar in spring, but different in summer and autumn. In spring the soil respiration model has the highest prediction accuracy (92.67%), followed by summer (84.99%) and poor in fall (77.23%). |