The diffusion fluxes of P in SWI
The diffusion fluxes of P in SWI are shown in Fig. 5. The values of the XW Reservoir and NZD Reservoir were significantly different. The only negative value appeared in XW-A, and the value of XW-C was much higher than others. XW-C was polluted by cage fish culture, which had an important impact on the eutrophication of XW Reservoir. Compared with the NZD Reservoir, the diffusion fluxes of P in the XW Reservoir were higher, and the value near the dam head was lower than that in the river region and transition region. The value of diffusion flux can describe whether the sediment is a sink or source of P (Sun et al. 2019). Only XW-A was negative, and all other sampling points were positive, indicating that the DGT-labile P in sediment in other areas was transformed from sediment to overlying water excluding XW-A.
As shown in Table 2, the diffusion flux values of NZD Reservoir were similar to the values of Dongting Lake in Hunan Province, China, but lower than that of other lakes and reservoirs in China and other countries. However, the values of XW Reservoir were larger than that of other lakes and reservoirs in China, which were similar to that of other countries, indicating that although many shallow lakes and reservoirs in China were seriously eutrophic, their internal pollution levels were relatively lower than that of XW Reservoir. In other words, attention should be paid to the control of P pollution in XW Reservoir.
Table 2
Diffusion fluxes of P in reservoirs and lakes
Location | P flux (ng·cm− 2·d− 1) | Method | Reference |
NZD Reservoir, LCR, China | 3.82–24.80 | DGT | This study |
XW Reservoir, LCR, China | -8.59-250.50 | DGT | This study |
Dongting Lake, Hunan, China | -2.70-19.70 | DGT | (Gao et al. 2016) |
Taihu Lake, Jiangsu, China | -21-65 | DGT | (Ding et al. 2015) |
Hongze Lake, Jiangsu, China | 17.20–79.30 | DGT | (Yao et al. 2016) |
Hongfeng Reservoir, Guizhou, China | 1–83 | DGT | (Wang et al. 2016) |
Lake Erie, Ohio, USA | 95–179 | DET | (Matisoff et al. 2016) |
Lake Peipsi, Estonia and Russia | 2.30-103.80 | discrete sediment pore water | (Tammeorg et al. 2015) |
Lake Okeechobee, Florida, USA | 83–240 | Aerobic and anaerobic condition | (Das et al. 2012) |
Lake Pontchartrain, Los Angeles, USA | 30–106 | Aerobic and anaerobic condition | (Roy et al. 2012) |
Effect Of Cascade Reservoirs On Release Of P
Due to the effect of cascade reservoir construction, most of the sediment was deposited in the upstream XW Reservoir, which led to regional sediment replenishment as the main sediment source of the downstream reservoir (Mu et al. 2020). The grain size from NZD Reservoir was coarser and contained high content of Al and Mn, and the grain size from XW Reservoir was mainly fine particles, and the content of Fe and Ca were relatively higher than that in NZD Reservoir. This indicated that the differences in physical and chemical properties in the sediment between XW and NZD reservoirs were related to the source of reservoir sediment. According to the above research, there was an obvious difference in the DGT-labile P between upstream and downstream reservoirs. The construction of cascade reservoirs not only led to different physical and chemical properties of sediment between the two reservoirs, but also led to significant differences in DGT-labile P between the two reservoirs.
The release of P in sediment involves the joint action of various geochemical processes, among which is closely related to geochemical cycles of Fe and Mn (Defforey and Paytan 2018, Smith et al. 2011). In this study, it could be observed that the DGT-labile Fe and Mn in XW and NZD reservoirs had the same variation trend with DGT-labile P excluding XW-C (Fig. 6). Similarly, DGT-labile Fe and Mn at other sampling points also had significant correlation with P (p < 0.05). The results indicated that FeOOH and MnOOH were reduced to Fe2+ and Mn2+, and led to the release of P (Chen et al. 2019). XW-C is located in the fish culture area, only a small part of the excessive bait was fed by fish, and most of the bait containing P was deposited at the bottom of the reservoir, which led to different vertical distribution and low correlation between DGT-labile P with Fe and Mn in XW-C.
The slopes of linear regression equation can reflect the degree of influence of DGT-labile Fe and Mn on P (Ding et al. 2016, Zeng et al. 2018). We could observe that the slopes between P with Fe in XW Reservoir were larger than that in NZD Reservoir, and the slopes between P with Mn in XW Reservoir were also larger than that in NZD Reservoir, which might be related to different sediment sources of the two reservoirs (Fig. 7). The sediment of NZD reservoir was supplemented by regional sediment, the release of DGT-labile P was more depended on the influence of sediment input, which led to different control factors of P. It is worth noting that in XW and NZD reservoirs, the slopes of P and Fe were larger than that of P and Mn. The results indicated that Fe oxides had stronger adsorption capacity for P in the sediment. In other words, the reductive dissolution of MnOOH dominated the release of P, which would take precedence over FeOOH (Postma and Appelo 2000). This is consistent with study by Li et al. (2021) on the coupling of Fe and Mn with P in the sediment.
Effect Of Cascade Reservoirs On Remobilization Of P
Bio-availability P (BAP) is the sum of active P forms in sediment, including P that can be directly utilized by organisms and P that can be converted into potential bioavailable P (Dan et al. 2020). DGT-labile P is an easily released or movable P component in sediment (Tao and Yali 2017). The BAP can be considered as the main source of the DGT-labile P. Correlation analysis also showed that there was a high correlation between DGT-labile P and BAP in each reservoir (p < 0.05). The DGT-labile P/BAP could be seen as the release of DGT-labile P by the unit BAP in the sediment, and it was not related to the absolute content of BAP in the sediment. The vertical variation of BAP and DGT-labile P/BAP in the two reservoirs is shown in Fig. 8. Two important conclusions can be drawn from the picture.
Firstly, with the increase in depth, BAP content was stable, indicating that the BAP input from outside was consistent and stable in the historical process of reservoir sediment deposition. However, the DGT-labile P/BAP in the two reservoirs had a clear tendency to increase with depth. The results show that with the increase of deposition time, more BAP was transformed into DGT-labile P, which was released into pore water. In other words, early diagenesis changed BAP remobilization into DGT-labile P. Secondly, the BAP in downstream NZD Reservoir was higher than that in upstream XW Reservoir, whereas the DGT-labile P concentrations in downstream NZD Reservoir were higher than that in upstream XW Reservoir. The DGT-labile P/BAP in upstream XW Reservoir was 7.8 times larger than that in downstream NZD Reservoir. This showed that, on the one hand, the cascade reservoir highlighted the importance of regional sediment. This could also be supported by previous studies. On the other hand, the cascade reservoirs weakened the remobilization of P in the sediment.