Sediment and nutrient removal by integrated tree-grass riparian buffers in Taihu Lake watershed, eastern China

Abstract

Loss of nitrogen (N) and phosphorus (P) from agricultural runoff is a leading cause of eutrophication in Taihu Lake, China. Riparian buffer strips can effectively remove sediment and nutrients from surface runoff. A field-scale study was conducted to assess the effectiveness of the integrated poplar (Populus)-grass buffer strips with different tree planting density in reducing sediment, N, and P from agricultural runoff during natural rainfall events in Taihu Lake watershed, eastern China. The experimental site consisted of four treatments, including a control plot with no buffer strip (T₀) and three poplar-grass buffer strips with different planting densities of 1,346 trees ha⁻¹ (T₁), 769 trees ha⁻¹ (T₂), and 577 trees ha⁻¹ (T₃) with planting spacing of 2 × 5 m (6.5 × 16.4 ft), 4 × 5 m (13.1 × 16.4 ft), and 6 × 5 m (19.7 × 16.4 ft), respectively. The integrated poplar-grass buffer strips significantly reduced runoff volume, sediment, total N, and total P masses by 49.3% to 61.6%, 78.7% to 90.2%, 57.2% to 71.5%, and 65.4% to 81.4%, respectively. With increasing poplar planting density in buffer strips, the canopy interception and maximum water storage capacity of litter layer increased, while the loss of sediments and nutrients from runoff decreased. There was significantly higher removal efficiency in buffer T₁ than that in buffers T₂ and T₃ for sediment, particle-bound N (PN), and particle-bound P (PP). There was no significant difference in removal efficiency for ammonium as N (NH₄-N), nitrate as N (NO₃-N), and phosphate P (PO₄-P) between buffers T₁ and T₂. According to different rainfall regimes, we found that the reduction in runoff volume, sediment, and nutrients load for buffers were higher under mild-long duration events (rainfall intensity class 2 to 4 mm h⁻¹ [0.078 to 0.15 in hr⁻¹]) and lower under intense-short duration events (rainfall intensity class 6 to 8 mm h⁻¹ [0.23 to 0.31 in hr⁻¹]). The removal efficiency of the buffers was negatively correlated with both precipitation and rainfall intensity, and the rainfall intensity had more effect on the removal efficiency than the precipitation. The results indicated that the impact of tree planting density on the effectiveness of buffer strips depended on the measured parameter. That is, the impact on sediment, PN, and PP was greater than that on the soluble N and P. The effect of tree planting density on the performance of buffer strips is mainly due to the interception of canopy and litter layer.