In order to explore the influencing factors of organic carbon transfer (Msoc) in rill erosion in alpine soil and to construct an organic carbon migration model that conforms to regional erosion characteristics, this study selected four factors that affect soil erosion: gravel cover degree, slope length, slope, and flow rate (Rc,L, S, Q). A full factorial design method was used to conduct runoff erosion experiments. The main results were as follow: (1) Msoc varies parabolically with Rc, and there was a positive correlation between Msoc and Rc when Rc below the critical value, and a negative correlation between Rc and Msoc above the critical value, with a critical value of around 45%; Msoc increased with S, and the rate of increase gradually decreased. L and Q had a small impact on Msoc. (2) The sensitivity and contribution rate of the factors to Msoc were consistent, with Rc (68.69%) > S (18.97%) > Q (6.00%) > S × Q (2.09%) > Rc × S(2.08%) > Rc × Q (1.18%) > L (0.91%). The interactions of Rc × L, S × L, and L × Q had no significant effect on Msoc (P > 0.05). (3) Rc, S, and Q had a cumulative contribution rate of 93.66% to Msoc, which can explain most of the Msoc variation. A redictive model for Msoc was established based on these three factors: Msoc= 0.6 231Rc + 1.998 S + 6.885 Q - 0.006 849 Rc2 - 0.06 073 S2 - 1.108 Q2. The model was tested with measured values, and the R2 value was 0.8 352, the NSE value was 0.722 545, and the P-value of the Mann-Whitney test was 0.724, which indicated that the model was basically applicable and achieve high prediction accuracy. Compared with the various test indicators of the full factor model, some factor models had fewer independent variables, simpler equations, and basically no decrease in accuracy. They had a high explanatory power for the dependent variables, making them more suitable for predicting Msoc in the southeast of Tibet, where information acquisition was difficult and the environment was complex.
CacuciD G, Ionescu-BujorM. A comparative review of sensitivity and uncertainty analysis of large⁃scalesystems - II: Statistical methods [J]. Nuclear Scienceand Engineering, 2017, 147(3) : 204-217.
[24]
赵紫岩.基于DOE和代理模型的结构优化设计方法及应用研究[D].西安:西安建筑科技大学,2017.
[25]
KinnellP I A. Runoff as factor influencing experimentally determined interrill erodiblities[J].Australian Journal of Soil Research, 1993, 31(3): 333-342.
[26]
HuangChihua. Empirical-analysis of slope and runoff for sediment delivery from interrill areas[J]. Soil Science Society of America Journal, 1995, 59(4): 982-990.
[27]
ZhangXunchang, NearingM A, MillerW P, et al. Modelling interrill sediment delivery[J].Soil Science Society of America Journal, 1998, 62(2): 438-444.
[28]
ZhangGuanghui, WangLili, TangKeli, et al. Effects of sediment size on transport capacity of overland flow on steep slopes[J]. Hydrological Sciences Journal, 2011, 56(7): 1289-1299.
BrakensiekD.L. and Rawls W J. Soil containing rock fragments: effects on infiltration. In: J. Poesen and H. Lavee (editors), Rock fragments in soils: surface dynamics[ J]. Catena, 1994,23: 99-110.
[34]
ValentinC. Surface sealing as affected by various rock fragment covers in West Africa[J]. Catena, 1994,23:87-97.
[35]
Poesen, J, et al. The hydrological response of soil surfaces to rainfall as affected by cover and position of rock fragments in the top layer[J]. Earth Surf. Process. Landforms, 1990,15: 653-671.
[36]
Van OostK, QuineT A, GoversG,et al.The impact of agricultural soil erosion on the global carbon cycle[J]. Science,2007, 318( 5850) : 626-629.