| In order to explore the effects of different nitrogen application rates on farmland soil respiration (RS) and quickly and accurately estimate RS, this study took winter wheat in Guanzhong area as the research object, observed the changes of RS under five nitrogen application rates, and examined the effects of environmental factors (soil temperature and soil moisture) and crop factors (leaf area index, above\|ground biomass, SPAD) on RS. An estimation model of farmland soil respiration based on soil temperature and vegetation index was established in Guanzhong area. The experiment was set as five nitrogen application rates under straw returning. They were conventional nitrogen application SN200 (200 kg·hm-2), optimal nitrogen application SN150 (150 kg·hm-2), 60% optimal nitrogen application SN120 (120 kg·hm-2), 50% optimal nitrogen application SN100 (100 kg·hm-2) and no nitrogen application SN0 (0 kg·hm-2). The results showed that RS increased first and then decreased with the growth period under different nitrogen application rates, and the addition of nitrogen promoted RS emission. The mean value of RS in each treatment observation period was: SN200 (3.68 μmol·m-2·s-1) > SN150 (3.40 μmol·m-2·s-1) > SN120 (3.06 μmol·m-2·s-1) > SN100 (2.70 μmol·m-2·s-1) > SN0 (2.21 μmol·m-2·s-1). The near\|infrared reflectance of winter wheat canopy at jointing stage and heading stage was significantly different under different nitrogen application rates, and the reflectance from high to low was SN200>SN150>SN120>SN100>SN0, while the difference was not significant at filling stage and maturity stage. Soil temperature significantly affected RS (P<0.01), but soil moisture had no significant correlation with RS (P0.05). Leaf area index, aboveground biomass, SPAD and vegetation index were significantly correlated with RS (P<0.05). Through multiple model evaluation, the optimal farmland soil respiration estimation model based on vegetation index and soil temperature was established, which was significantly higher than the single factor model based on soil temperature, and the accuracy of the model could reach more than 0.6 (n=120). |