4.3.1 Nondiversion water period
Fig. 3 shows that there were significant differences in the chemical types of water
in the Old Canal and in the Nanyang sublake and that the SO42- concentrations in the Old Canal were lower than those in the Nanyang sublake. Therefore, the Old Canal was not the main source for the Nanyang sublake, and the impact of the Old Canal on sulfates in the Nanyang sublake could be ignored during the nondiversion water period.
In addition, the converging area of the Sihe River and the Nanyang sublake was the boundary of the distribution of the gypsum salts in the WSD, and there was more gypsum salts distributed in the area of southwest of the Sihe River (Qi, 2010 ). Therefore, SO42- from evaporite dissolution was not considered as the main source of sulfate in the Nanyang sublake, except in the converging area of the Sihe River and the Nanyang sublake.
According to the spatial distribution of the adjacent main SO42- source, the Nanyang sublake was divided into three parts: NYSL-1, NYSL-2 and NYSL-3 (Fig. 7(a) ). NYSL-1 was the mixing area of the Guangfu River, the Zhushui River and the Nanyang sublake. NYSL-2 was the convergence area of the Sihe River and the Nanyang sublake. NYSL-3 was the mixing area of the Zhuzhaoxin River, the Old Wanfu River and the Nanyang sublake. We hypothesized that the δ34SSO4 values in each region reached the value observed in the Nanyang sublake during the nondiversion water period. The δ34SSO4 values were measured to the northeast and southwest of the Nanyang sublake due to the complex SO42- concentration distributions in the Nanyang sublake. In the NYSL-1 and NYSL-2 regions, the δ34SSO4 value was 12.50‰, and the δ34SSO4 value was 13.35‰ in the NYSL-3 (Table 1 ).
The SO42- contribution rates of each main source can be calculated by the following formulas (Li et al., 2011 ):
δ34Slake34Si*A+δ34Sj*B (1)
A + B= 100% (2)
δ34Slake = δ34Sgypsum * C + δ34Sk *D (3)
C + D = 100% (4)
where δ34Slake is the value of δ34SSO4 in the lake; δ34Si, δ34Sj and δ34Sk are the values of δ34SSO4 from different inflowing river inputs; and δ34Sgypsum is the value of δ34SSO4 in gypsum (15.70‰ here). A, B, C and D are the contribution rates. Eq. (1) and Eq.(2) were used for NYSL-1 and NYSL-3, and Eq. (3) and Eq. (4) were used for NYSL-2.
All the calculated contribution rates are summarized in Table 2 , and the maximum contribution of SO42- in the Nanyang sublake came from the Old Wanfu River input and was as high as 32.56%, followed by the contribution of evaporite dissolution (25.60%). Therefore, the inflowing river inputs were the main sulfate source in the Nanyang sublake in the nondiversion water period.