Multi-scale effects of landscape on nitrogen (N) and phosphorus (P) in a subtropical agricultural watershed: A case of Qi river basin (QRB), China

[Display omitted] •N and P parameters in the dry season were higher than those in the rainy season.•The correlation coefficient in the rainy season was greater than that in the dry season.•PLP, BLP, PD and SHDI were typically positively correlated with N and P.•NH4+-N, TN and TP were greatly affected by topographic and landscape metrics.•The main influencing factors of nutrient in the larger scale was more than others. Studying the temporal and spatial relationship between water body nutrients and landscape elements in the agricultural watershed will be helpful in revealing the process of non-point source pollution (NPSP). In order to quantitatively analyze the effects of landscape elements on nitrogen and phosphorus at multiple scales, the Qi River basin located at the tail of the Three Gorges Reservoir was selected as the study area, and various methods such as t-test, one-way analysis of variance, pearson correlation and redundancy analysis were used. The results showed that the nitrogen (N) and phosphorus (P) parameters in the dry season were higher than those in the rainy season, and the spatial differences of N and P in the watershed were significant (with the 95 % level). This may be due to the difference in surface runoff, which led to higher nitrogen and phosphorus concentrations in dry season than in rainy season. Built land percetage (BLP), the largest patch index and the paddy land percentage had the largest coefficients of variation (greater than 40 %). The 100 m buffer zone had the lowest correlation with the catchment, and the 300–1000 m buffer zone had the strongest correlation. The correlation coefficient in the rainy season was greater than that in the dry season. The interpretation rate of all selected indicators in the rainy season exceeded 51.87 % (significant with the 95 % confidence level), while the dry season on the same scale increased by 7.12 %-14.28 % (significant with the 95 % confidence level). Topography and land use/cover had obvious effects on N and P parameters in water bodies. N and P were typically positively correlated with paddy land percentage, dry land percentage and built land percentage, patch density and Shannon diversity index, but they had a negative correlation with the largest patch index. The optimal spatial scale of N and P management in the rainy season was within the buffer range of 100 m–500 m, while the optimal management scale in the dry season was the 500 m buffer and catchment. The amount of N and P e