[Effect of Fertilization on Soil Respiration and Its Temperature Sensitivity in a Sorghum Field]

The characteristics of soil respiration under the condition of fertilization have not been fully understood,especially for a long-term fertilization condition. In this study we measured both soil respiration using an LI-COR-6400-09 soil chamber attached to LI-COR-6400 portable photosynthesis system, and the vegetation spectrum using an ASD FieldSpec HandHeld2, in five different fertilization treatment fields. The soil respiration (R_s) and vegetation spectrum were simultaneously measured with two samples per month in the growing season in 2016 and 2017. The soil temperature at 10 cm depth (T₁₀) and moisture (W_s) for the surface of 10 cm were also measured simultaneously. The five different fertilization treatments included no fertilization (CK), inorganic fertilizer (INF), inorganic fertilizer+organic fertilizer (INF+M), inorganic fertilizer+organic fertilizer+straw turnover (INF+M+S) and organic fertilizer+straw turnover (M+S), and all treatments had been conducted since 2011. Based on those observation data, we made an analysis of R_s and its temperature sensitivity (Q₁₀) in the five different fertilization treatments. The results showed that no significant temporal change in R_s among the five treatments was found. No significant difference was found in R_s between the CK and INF treatments. Compared with the values of R_s in CK and INF, the R_s values in INF+M, M+S, and INF+M+S treatments increased by 28.2%-39.1%, 47.9%-76.0%, and 46.2%-50.8%, respectively. This indicated that use of organic fertilization and straw application increased R_s. Both the T_s and W_s showed 14%-96% and 6%-37% in R_s seasonal variations, respectively. Among the treatments, the correlation coefficient of the fitted equations between R_sand T_s was higher in the INF+M, INF+M+S and M+S treatment than in CK and INF, but was not between R_sand W_s. For the relationship between R_s and vegetation indexes we found that the correlation coefficients between R_s and the difference vegetation index (DVI), ratio vegetation index (RVI), and enhanced vegetation index (EVI), respectively, were higher than that of R_s and the normalized differential vegetation index (NDVI); and that the correlation coefficients between R_sand the red edge slope (D_red) and red edge area (S_red) were higher than between R_s and the red edge position (λ_red). This indicated that the treatments in INF+M+S increased the correlation coefficient between R_s and the spectrum characteristics index. The determination coefficient of the fitted equations including the feature spectral parameters, T₁₀, and W_svariables was higher than that of the equations only including both T₁₀ and W_s variables, or a single variable of T₁₀ or W_s. Compared with CK, the Q₁₀ value increased by 26%, 39%, 21%, and 37% for the INF, INF+M, INF+M+S, and M+S treatments, respectively. This indicated that temperature sensitivity Q₁₀ increased under the condition of fertilization treatments. The Shannon diversity index, bulk density, and soil organic matter were the main factors causing the difference in R_s, Q₁₀, and R₁₀, i.e., R_s at a temperature of 10℃, in the different treatments, which could explain the 97.6%, 78.2%, and 92.8% variations in R_s, Q₁₀, and R₁₀, respectively.