湖北省径流系数时空演变特征与参数区域化研究
Study on the Spatio-Temporal Evolution Characteristics and Parameter Regionalization of Runoff Coefficient in Hubei Province

径流系数的取值可直接关系到无资料区域水文计算的可靠性和科学性，如何确定这一参数仍是当前水文工作者面临的棘手问题。本文以湖北省县级行政区与水资源四级区套合形成的210个基本单元为研究对象，收集了2001~2023年的水文气象和下垫面等数据，采用Mann-Kendall趋势检验法分析湖北省径流系数时空变化特点，结合随机森林识别径流系数的主控影响因子，通过多元线性逐步回归方法构建径流系数区域化预测模型，实现了全省各研究单元径流系数的定量估算。最终结果表明：其一，时间变化上，湖北省多数区域2001~2023年径流系数呈现不显著变化趋势，只有荆门东宝区、随州广水等局部区域呈显著减小趋势，黄石大冶、黄冈罗田等东南部区域呈显著增大趋势，2010年之后波动明显加剧。其二，在空间分布上，全省径流系数总体上呈南高北低的分布格局，鄂西南和鄂东南多年平均径流系数超过0.54，而鄂北岗地仅有0.24~0.31，这个分布基本上符合湿润区的产流规律。其三，降水量、潜在蒸散发、归一化植被指数是湖北省径流系数的主控因子，其中降水量为所有区域的核心影响因子，鄂西山区还同时受蒸散发制约，城镇化水平较高的武汉城市圈受归一化植被指数影响明显，而坡度、土壤质地等因子对年尺度径流系数的解释力较低。其四，通过构建的区域化多元线性回归模型对全省各研究单元径流系数的模拟平均相对误差为12.7%，其中鄂西生态区精度最好(平均误差8.3%)，鄂北农业区精度相对差一点(平均误差19.7%)，不过整体来说能满足水文工程的应用需求。这项研究工作可以为湖北省无资料流域水文设计、水资源管理提供科学支撑。
The value of the runoff coefficient is directly related to the reliability and scientific validity of hydrological calculations in ungauged regions, and determining this parameter remains a challenging issue for hydrologists. This study took 210 basic units formed by the spatial overlay of county-level administrative divisions and water resources level-4 sub-zones in Hubei Province as the research objects, and collected hydrological, meteorological, and underlying surface data of Hubei Province from 2001 to 2023. The Mann-Kendall trend test was employed to analyze the spatio-temporal variation characteristics of the runoff coefficient in Hubei Province, and the random forest algorithm was used to identify its dominant influencing factors, and a regionalized prediction model of the runoff coefficient was constructed via multiple linear stepwise regression, enabling quantitative estimation of the runoff coefficient for all research units across the province. The results indicate that: 1) In terms of temporal variation, the runoff coefficient in most areas exhibited no significant trend during 2001 to 2023, with only part areas such as Dongbao District in Jingmen and Guangshui in Suizhou showing a significant decreasing trend, while southeastern regions such as Daye in Huangshi and Luotian in Huanggang showed a significant increasing trend, with fluctuations becoming more pronounced after 2010. 2) In terms of spatial distribution, the runoff coefficient generally presented a pattern of being higher in the south and lower in the north, with the multi-year average runoff coefficient exceeding 0.54 in southwestern and southeastern Hubei, but only ranging from 0.24 to 0.31 in the northern Hubei upland, which is