Except for February, salinity measured in other months at stations #2 and #4 never exceeded 0.24 (data not shown), suggesting the Changjiang freshwater end-member was well represented by samples from Xuliujing. In February 2010, a saltwater intrusion event was detected at Xuliujing and the salinity reached 1.95–2.05 at station #2 and 1.96–2.06 at station #4. Although the 12 field studies were all conducted during the spring tide periods, the ranges of variation in salinity for all samples from stations #2 and #4 within a month never exceeded 0.10 (data not shown), suggesting the influence of flood tides versus ebb tides was small.
 In most months the differences between the two average values of nitrate concentration at stations #2 and #4 (averaged by 39–42 measurements at the three depths) were not significantly different (P> 0.05, by paired-samples t test), except for December, February, March, and April (P < 0.05). In February, when saltwater intrusion was detected, the average nitrate value at station #2 was significantly higher than that at station #4 while in the other three months the average values at station #2 were significantly lower (Figure 3a). In all 12 months at station #4 and most months at station #2 (9 out of 12), no significant differences in nitrate concentration between surface and bottom were found (P> 0.05, by paired-samples t test); only in March, April, and May at station #2 were the nitrate concentrations at surface significantly lower than those at bottom (P < 0.05).
Figure 3. Variations of monthly average concentrations (±standard deviation) of nutrients measured at the two stations #2 and #4 during the sampling period, September 2009 to August 2010 and July 2007.
Download figure to PowerPoint
 For silicate, significant differences of average concentration between stations #2 and #4 were shown in the six months of September and January to May (P < 0.05): station #2 showed higher values in February, March, and May, whereas station #4 showed higher values in the other months (Figure 3b). In all 12 months at station #4 and in most months at station #2 (11 out of 12, including the three months of March, April, and May when significant surface-bottom differences were found for nitrate), there was no significant difference in silicate concentrations between surface and bottom (P > 0.05); and only in October at station #2 did silicate show significantly higher concentrations at bottom than at surface (P < 0.05).
 Unlike nitrate and silicate, in eight months (not October, March, June, or August) phosphate showed significant differences between average concentrations at stations #2 and #4 (P < 0.05). In six of these eight months (September, November, December, January, February, and May), the average phosphate concentrations at station #2 were significantly higher than those at station #4, whereas in the other two months the values for station #2 were significantly lower (Figure 3c). In all the 24 cases at stations #2 and #4, no significant differences in phosphate concentration between surface and bottom were found (P > 0.05).
 For ammonia, in eight months (not September, October, March, or May) significant differences were shown between average concentrations at stations #2 and #4 (P < 0.05). In three out of these eight months (June, July, and August), the average ammonia values at station #2 were significantly lower than those at station #4, whereas in the other months, significantly higher average values were displayed at station #2 (Figure 3d). For surface-bottom comparison, significant differences were found only in five cases out of 24 (P < 0.05), i.e., from March to May at station #2 and in October and March at station #4.
 Significant differences of average nitrite concentration between stations #2 and #4 were found in five months (P < 0.05), in which higher values were found at station #2 in November, December, and April and higher values were found at station #4 in March and August (Figure 3e). For surface-bottom comparison, significant differences were shown only at station #2 in October, March, and April, and at station #4 in June (P < 0.05).
 In summary, the Changjiang water at the river mouth was relatively well-mixed because during the year sampled insignificant or small differences of nutrient concentrations were generally found between stations #2 and #4 or between surface and bottom. For nitrate, the highest monthly value averaged by all measurements at stations #2 and #4 occurred in March (149.0 ± 5.5μM), whereas the lowest value was shown in September (93.0 ± 2.8 μM) (Figure 3a). Different from nitrate, silicate showed the highest (127.0 ± 4.3 μM) and lowest (90.7 ± 8.7 μM) monthly concentrations averaged at stations #2 and #4 in September and April, respectively (Figure 3b). Compared with nitrate and silicate, concentrations of phosphate showed a more complex seasonal variation pattern (Figure 3c). However, its overall seasonal variation was more similar to that of nitrate because it also showed the highest average value (1.94 ± 0.53 μM) in the dry month, November, and the lowest value (0.58 ± 0.18 μM) in the flood month, August. For ammonia and nitrite, their most striking seasonal variation pattern was that sharp concentration peaks appeared in winter (Figures 3d and 3e), and the monthly concentrations increased by more than one order of magnitude from flood months to dry months. For ammonia, the average values at stations #2 and #4 increased from the lowest value, 0.4 ± 0.2 μM in September, to highest, 30.9 ± 5.2 μM in February; for nitrite, the average values ranged from 0.05 ± 0.04 μM in September to 4.30 ± 0.64 μM in April.
 During the semidiurnal tides for sampling in each month, the nine SPM filters from station #4 at high, intermediate, and low tides (three samples at surface, middle, and bottom on each occasion) were selected to measure SPM abundance, POC and PN contents, and their stable isotope ratios (δ13C and δ15N). The highest monthly SPM abundance averaged by the nine measurements appeared in the flood month, September; whereas the lowest value was in the dry month, February (Figure 4a). Corresponding to the highest SPM abundance, POC and PN contents displayed the lowest values in September (Figures 4b and 4c). On the other hand, the highest average POC and PN values were found in February, when the lowest SPM abundance appeared. The highest monthly POC and PN values in February were both about twice their lowest values in September. The average monthly POC/PN molecular ratios ranged between 8.7 ± 0.9 in August and 12.5 ± 1.7 in April; however, the ranges within each month were always larger than those between months (Figure 4d). The highest average δ13C value appeared in September, whereas the lowest value was in December (Figure 4e). Statistically significant negative relationships were found between the monthly average values of δ13C and POC (n = 12, R2 = 0.78, P < 0.001) and between δ13C and PN (n = 12, R2 = 0.61, P = 0.003). Compared with the other indices, the variation of δ15N seemed to be most complex and displayed a twin-peak pattern during the year sampled (Figure 4f). In the four months with intermediate Changjiang discharges (October, November, April, and May), the highest δ15N average values, >5.2‰, were found; whereas the lowest values, <3.5‰, were found during both the flood months of August and September and the dry months of December to March.
Figure 4. Variations of the monthly average values (±standard deviation) of SPM abundance, POC and PN contents, POC/PN molecular ratio, and stable isotope values (δ13C and δ15N) of POM measured at station #4 during the sampling period, September 2009 to August 2010.
Download figure to PowerPoint
 The monthly Changjiang fluxes of nutrients, POC, and PN shown in Table 1 were calculated by multiplying the average monthly concentrations over all the measurements at stations #2 and #4 (measured at Xuliujing) by the average monthly Changjiang discharges (measured at Datong from CWRC). The Changjiang monthly discharge fluxes varied in a wide range (the maximum Changjiang monthly discharge in July 2010 was more than five times the minimum in January 2010, as is shown in Table 1), hence most chemicals, including nitrate, silicate, POC, and PN, showed highest monthly fluxes in the flood month of July and lowest fluxes in the dry months of December to February. Ammonia and nitrite, which displayed sharp, high peaks in dry months, showed the highest and lowest fluxes in March to April and in October to November, respectively. Finally, phosphate showed the highest monthly flux in May and the lowest flux in January.
Table 1. Monthly Fluxes of Freshwater, Nutrients, DIN, POC, and PN Discharged by the Changjiang From September 2009 to August 2010
|Monthly Flux||Freshwater (×1010 m3 month−1)||NO3−||SiO32−||PO43−||NH4+||NO2−||DIN||POC||PN|
|(×109 mol month−1)|
 A wide range of methods have been proposed for calculating annual river fluxes using concentration and flow data, including interpolation and extrapolation methods [Littlewood et al., 1998; Webb et al., 2000]. In this study, the monthly averaged concentrations of nutrients, POC, and PN have been obtained by our measurements. The daily freshwater fluxes at Datong are measured by CWRC in weekdays (242 out of the 365 days for the year sampled from September 2009 to August 2010, http://yu-zhu.vicp.net/). In each month, if the daily discharge fluxes were not provided by CWRC for the sampling days of our field studies, the discharge values in the next days are used instead as Qi (instantaneous discharge at the sampling time). Since a good rating curve between the logarithmically transformed values of concentration and flow cannot be found for most chemicals (P > 0.05, by least squares regression) except for phosphate (n = 12, R2 = 0.46, P = 0.015), in Table 2 the annual fluxes are calculated only by the five interpolation methods, Method 1 to Method 5, proposed by Webb et al. . The comparisons between results in Table 2 show that the five interpolation methods give similar annual flux results, with relative standard deviation values of <4% for nitrate, silicate, DIN, POC, and PN; 7% for phosphate; 12% for nitrite; and 28% for ammonia (n = 5). These annual flux values are believed to be the most accurate estimates, up to now, for the Changjiang because they are obtained from more intensive sampling than previous studies on both spatial (at the three depths of surface, middle, and bottom, as well as at the two stations near the mouth of the Changjiang) and temporal scales (at high, intermediate, and low tides and in each month of the year sampled). No matter which method is used, the annual DIN fluxes are always higher than the corresponding PN fluxes by more than one order of magnitude (Table 2).
Table 2. Annual Fluxes of Nutrients, DIN, POC, and PN Discharged by the Changjiang From September 2009 to August 2010, Calculated According to the Interpolation Methods (Method 1–Method 5) Proposed by Webb et al. 
|(×109 mol yr−1)|
|Relative Standard Deviation||3%||2%||7%||28%||12%||4%||3%||3%|