Background Climate change (CC) and land use and land cover change (LUCC) are two driving factors affecting streamflow within a watershed. The Fu River Watershed (FRW) was selected as the study area due to its critical location to the upper Yangtze River. This study aims to quantify streamflow reduction under the sole and combined effects of CC and LUCC at multiple spatial scales.

Methods Two combined used techniques, Mann-Kendall test and Sen′s slope estimator, were selected for trend analyses. Double Mass Curve (DMC) and Pettitt test were used for the abrupt point detection. Transition matrix was adopted to analyze the variation of land use type before and after the detected abrupt year. Two time periods were divided by the abrupt year and four different scenarios were set up to simulate the sole and combined effects of CC and LUCC. Streamflow discharges for different scenarios were driven by SWAT simulation.

Results 1) During the entire period, annual streamflow showed a significant reduction trend with a confidence level of P < 0.01 and annual precipitation exhibited a downward trend, respectively. The year 2000 was detected as the abrupt point, and the whole time series presented characteristics of "decreasing at the first stage and increasing at the second stage". 2) The primary land use type of the FRW was farmland (58.80%), which decreased by 427.98 km² during the study period. The areas of unutilized and urban land were relatively small, yet with largest growth rates (218.80% and 195.58%, respectively). 3) The R², E_ns, and PBIAS were selected as model evaluation indices, the values for which were 0.89, 0.88, and -0.2% for calibration period, respectively, and 0.82, 0.82, and 6.8% for validation period, respectively. This demonstrated that the SWAT could be satisfactorily used in FRW. 4) Four scenarios were constructed during the basic period (1980-2000) and the varied period (2001-2015), and results showed that: at the whole basin scale, water yield(WYLD) under scenario 1(with isolated variation of precipitation), scenario 2(with isolated variation of land use) reduced by 10.47% and 1%, respectively. However, WYLD under scenario 3(with combined variations of precipitation and land use) reduced by 16.01%. At the sub-basin scale, the increase of unutilized in the upstream increased the risk of soil erosion. On the other hand, the expansion of urban land at downstream increased the WYLD, while the effects of Grain-to-Green project on both bank regions of the river channel significantly reduced streamflow discharge.

Conclusions Both CC and LUCC will affect WYLD of FRW at basin scale and sub-basin scale. It is remarkable that the combined effects of CC and LUCC are not equal to sum of corresponding sole effect, which is potentially due to the interactions between the climate and land use types. This research may provide reference for water resource management, soil and water conservation and reservoir streamflow forecast of the Three Gorges Dam.