Predicting Long-Term Hydrological Change Caused By Climate Shifting In The 21st Century In The Headwater Area of The Yellow River Basin

Jingyi Hu (  2269722132@qq.com ) Xi'an Jiaotong University https://orcid.org/0000-0001-5212-0440 Yiping Wu Xi'an Jiaotong University https://orcid.org/0000-0002-5163-0884 Pengcheng Sun Xi'an Jiaotong University Fubo Zhao Xi'an Jiaotong University Ke Sun Xi'an Jiaotong University Tiejian Li Tsinghua University Bellie Sivakumar Indian Institute of Technology Bombay Linjing Qiu Xi'an Jiaotong University Yuzhu Sun Xi'an Jiaotong University Zhangdong Jin Chinese Academy of Sciences Institute of Earth Environment

Our modeling result will provide a proper perspective for investigating the main 151 influencing climate factors of the hydrological components, which is not only useful 152 for people to formulate suitable strategies and policies in semi-arid area, but also key 153 to the sustainable development of the eco-environment in the YRB. With this in mind, 154 the goal of the present study was to assess the hydrological responses to the future 155 projected climate in the HYRB during the near-future period (2020-2059) and far-156 future period (2060-2099). The assessment was made for three RCP scenarios (RCP 8 / 57 2.6, 4.5, and 8.5) using an ensemble of eight downscaled GCMs and SWAT modeling.

158
The specific objectives were: (1) to validate the suitability and performance of the 159 SWAT model in simulating the hydrological processes in the HYRB; (2) to predict the where SW is the soil water content, is the time t (days) for the simulation period, R    Table 2). The Sequential Uncertainty

221
Fitting version 2 (SUFI-2) algorithm was adopted for the parameter optimization in this 222 study (Yang et al., 2008 used bilinear interpolation to obtain high-resolution data that could be used in  were calculated as follows: where m is the month m, and are the corrected GCMs precipitation and 254 temperature, 0 and 0 are initial GCMs precipitation and temperature, ℎ 255 and ℎ are GCMs precipitation and temperature data in historical period, ℎ 0 and 256 0 are CMA precipitation and temperature data.

257
The GCMs data were evaluated by comparing with the CMA data during the where represents the value of independent variable in t period, represents the  We also used Mann-Kendall nonparametric rank test to analyze the trend of 278 hydrological and meteorological elements (Kendall and MauriceG, 1979). The rank 279 correlation test for two sets of observations X = 1 , 2 , … and Y = 1 , 2 , … is formulated as follows. The statistic S is calculated as follows: and is similarly defined for the observations in Y. Under the null hypothesis that X 283 and Y are independent and randomly ordered, the statistic S tends to normality for large 284 n, with ( ) = 0 and variance given by: The significance of trends is tested by comparing the standardized test statistic Z with 286 the standard normal variate at the desired significance level. Z is calculated as: | | ≥ 1.64 means that the confidence level in the current test is more than 95% 288 (p<0.05). water in this area was higher. Figure 4 h showed that the area with decreased soil water 330 was greater than the increased one, and we could also find the same result from Table   331 4.

332
Water yield refers to the capacity of a catchment to supply water (Arnold et al., 1998).

333
The average water yield in study area was 205 mm with a range of 147 to 305 mm   temperature, which meant that although the rainfall increased in this region, the increase 404 of ET due to raise of temperature played a greater role. Figure S5 showed that the decrease of soil water was predicted to be mainly in the west, middle and export areas 406 of the basin, while it would increase slightly in the southeastern region. Compared with 407 the near future period, the increment of soil water in southeast may decrease during the 408 far future period, and even turn to a decrease.

409
Under the combined effects of increased temperature and variations in precipitation, 410 the water yield showed a decrement of 16.5-20.1% during the near future period 411 (Figure 7). At the end of this century, due to the increase of precipitation, the water 412 yield would be no longer continuously reduced, and the decline rate was similar to that 413 the near future period (15-19.5%). Table 6 indicated that water yield had a larger range 414 of variation and correlation than AET and soil water. because of the small value of absolute water yield in winter. Besides, water yield was 421 projected to decrease from May to August in each scenario. Figure S6 was the change 422 of water yield during two future periods. We found that water yield in most HRUs 423 would decrease under three RCPs, which was related to the obvious increase of AET.

424
The water yield was predicted to increase only in a few areas, mainly distributed in the  Furthermore, Figure 6 showed that projected increment of maximum temperature were 448 slower than that of minimum temperature, which is consistent with most areas around the world and might lead to a decline in diurnal temperature range (DTR) and (2) (3) where . and , are measured and simulated streamflow at each time step ; ( 1 , 2 ), 21 = ( 2 , 1 ), 22 = ( 2 , 2 ) ( Figure S2).

619
First, linear interpolation is performed in the -direction: