气候变化和人类活动影响使得部分地区水文系列资料的一致性受到影响，从而使得这些地区逐渐变成缺资料地区或无资料地区，无资料地区水文预测（PUB）问题一直是国际水文的研究热点，而山丘区的无资料小流域PUB更是其中的难点之一。分别选取了资料条件相对丰富的江苏省北部的黑林、南部的横山水库、中田舍等3个代表性的山丘区小流域，通过考虑前期30 d降雨量、增加经流曲线数（curve number,CN)分级和对SCS模型的常数项进行修正，得到适用于江苏省山丘区小流域的SCS产流模型；并将其与基于地形地貌参数的Nash汇流模型结合，构建了山丘区小流域的SCS-Nash水文模型，并选取江苏中部的中田舍小流域进行验证。结果表明:构建的SCS-Nash模型具有较好的适用性，可为研究区域小流域的PUB问题提供参考。

The impact of climate change human activities has affected the consistency of historical hydrologic data,which has gradually turned these areas into datadeficient areas or areas without data.The problem of Prediction in Ungauged Basins (PUB) has been a research hotspot in the world.The PUB problem in small watersheds in hilly areas is one of the difficulties.In recent years,some hydrological models based on the physical characteristics of watersheds have been applied to hydrological simulations in the areas lacking data or without data.This reduced the dependence of hydrological simulations on historical data,and provided a good opportunity to study the mechanism of runoff generation and convergence in areas without data. An SCS model is used based on constant term revision and a Nash unit line based on topography and geomorphology characteristics.Three typical small watersheds in hilly areas of Jiangsu Province were selected for application research,aiming to explore and solve the problem of runoff yield and confluence and flood forecasting of small watersheds in hilly areas in the shortage of data.The improvement of the SCS model mainly includes the following three points.The revision of the SCS model mainly uses the previous impact rainfall Pa instead of the previous 5 d rainfall total in the original model to determine the degree of soil moisture in the previous period and increase the classification of the number of runoff CN.Combining the reestimated previous soil moisture degree,linear interpolation is used to increase the classification of the number of runoff curves,reduce the level of CN value,and then revise the CN value table;establish an empirical relationship with the CN value reflecting the topographical features.According to the small watershed data with measured rainfall and runoff data,the S value in the SCN model is calculated,and then the constant term in the SCN relationship is revised.using the rich observed rainfall data in small watershed,figure out the S value in the SCS model,and then revised the constant term of the SCS model. Taking 3 small watersheds in Jiangsu Province as the research area,the SCS runoff generation model was improved in terms of determining the degree of soil moisture in the early stage,increasing the CN value classification and revising the constant term.Finally,the constant term of the SCS runoff generation model was determined as 160.The revised SCS runoff model was combined with the Nash confluence model based on topographic and geomorphic parameters to construct the SCSNash hydrological model of the hilly small watersheds.The test results showed that the SCSNash model had good applicability in flood simulation and forecasting,and the calculation accuracy was slightly improved. The results showed that the SCSNash hydrological model had certain precision.Especially for the simulation and forecast of the peak discharge,the relative errors of 8 out of 9 floods are less than 20 %.Considering that the predicting accuracy of the peak discharge is more important,the SCSNash hydrological model have some applicability in the flood forecast of small watershed in the hilly areas of Jiangsu Province.Due to the lack of hydrological data in small watersheds in hilly areas,the study did not collect sufficient hydrological data.Although the research results have been used for reference in the flood forecast of small watersheds in Jiangsu Province,the conclusion is preliminary,and a more comprehensive conclusion needs further study.

国家重点研发计划（2016YFC0402709）；国家自然科学基金（41730750）；江苏省水利科技项目（2017037）