在调查甘肃引黄灌区玉米产量、施肥量和土壤理化性状基础上，2015年确定60个地块作为研究对象，分层采集0～40 cm土层土壤样品，研究土壤三相比、紧实度、水稳性团聚体含量和有机质含量的变异规律及其与玉米产量、化肥偏生产力的关系。结果表明，0～40 cm土层土壤三相比δ值、土壤紧实度、土壤＞ 0.25 mm水稳性团聚体含量和有机质含量均呈中等变异特征。土壤紧实度随土层加深而显著增加，土壤三相比δ值、＞ 0.25 mm水稳性团聚体含量和有机质含量均有随着土层加深而降低的趋势。玉米产量与0～40 cm土层土壤三相比δ值和土壤紧实度呈极显著负相关，与土壤＞ 0.25 mm水稳性团聚体含量和土壤有机质含量呈极显著正相关；磷肥偏生产力与0～40 cm土层土壤三相比δ值呈极显著正相关，与土壤紧实度呈显著正相关，与土壤＞ 0.25 mm水稳性团聚体含量和土壤有机质含量呈极显著负相关；氮肥偏生产力与4个指标之间均达不到差异显著水平。玉米产量与土壤紧实度、土壤三相比δ值的关系可用一元二次方程来拟合，与土壤＞ 0.25 mm水稳性团聚体含量、土壤有机质含量的关系可用一元线性方程来拟合。多元逐步回归分析表明，玉米产量最终取决于土壤＞ 0.25 mm水稳性团聚体含量和土壤有机质含量两个关键因素。

Based on the investigations of corn yields，fertilizer application rates and soil physical and chemical properties in Gansu Yellow River irrigation district in 2015， the layered 0～40 cm soil samples were collected as the research object in 60 corn fields for the variability of soil three-phase，compaction，water-stable aggregate， organic matter content and their partial correlations with corn yield and partial factor productivity. The results showed that the averages of soil three-phase δ value，soil compaction， > 0.25 mm water-stable aggregate and organic matter content in 0-40 cm soil showed moderate variability. The soil compaction significantly increased with the increase of soil depth， but the soil three-phase δ value， > 0.25 mm water-stable aggregate and organic matter content decreased with the increase of soil depth. Corn yields were extremely negative correlated with 0～40 cm soil three-phase δ value and soil compaction， but extremely positive correlated with > 0.25 mm water-stable aggregate and organic matter content， significantly. Phosphate partial factor productivity were extremely positive correlated with three-phase δ value， positive correlated with soil compaction， but extremely negative correlated with > 0.25 mm water-stable aggregate and organic matter content 0～40 cm soil depth. There were no significant difference between nitrogen partial factor productivity and the four indexes. The relationships between corn yield and soil compaction， soil three-phase δ value could be fitted by quadratic unary equation， and that of > 0.25 mm water-stable aggregate and organic matter content could be fitted by unary linear equation. The results of multivariate stepwise regression analysis showed that > 0.25 mm water-stable aggregate and organic matter content were the key factors to decide corn yield.

S158；S513

农业农村部公益性行业(农业)科研专项(201503117)；甘肃省农业科学院重点研发计划(2020GAAS19)。