The sampled river section is alongside the lower reach of the Yangtze River. The survey period is from 2016 to 2019. Sampling and study area were in four sections, namely, Hukou (HK), Anqing (AQ), Dangtu (DT), and Changshu (CS). Three sampling sites were set in each section for the larval fish, namely, the right bank (R), the left bank (L), and the middle site (M). Two survey sites were set in each section for the adult fish of silver carp survey (shown in magnified graph, Figure 1).

Environmental factors were tested in each sampling site including the following content. Water temperature (WT, °C), dissolved oxygen (DO, mg⋅L^–1), and pH were recorded using a multiparametric probe (WTW Multi 340i, MultiLine, Germany). Water transparency (WTP) was measured using a secchi disk. Current velocity (CV, m⋅s^–1) was measured using a flow meter Digital Flow Meter 23090 (KC Denmark A/S, Silkeborg, Denmark). Average WD (m) was measured using a depth sounder (S48-1X, Maikeyi, Beijing, China).

The fishing net includes stick-nets and multi-mesh gill nets. Among them, the length of a single stick-net is 75 m, the height of the net is 5 m, the mesh size of the block net is 3.5 cm, and the capsule mesh size of the bag net is 1 cm. The length and height of a single multi-mesh gill net are 100 and 2 m, and the mesh sizes are 2, 4, 6, 8, 10, and 14 cm, respectively. The investigation period was from January 2016 to December 2019; during the investigation, the multi-mesh gill nets were lowered to the designated investigation waters before 5:00 p.m., and the nets were closed before 8:00 a.m. on the next day. For each catch, the body length (BL) and weight were measured, and the biological characteristics were determined for the sampled sliver carp.

According to the annual average catch per unit effort (CPUE), the total number of fishing vessels and the annual fishing days of a single vessel, the proportion of weight and individual number in each section, and the average annual catch weight and numbers were estimated. Annual average natural resources in the sampled sections were estimated using FISAT II medium length stock analysis method. Among them, the limit BL is L∞ = 1,047.2 mm and k = 0.1603, and the relationship between BL and body weight is W = aL^b (a = 0.000040413, b = 2.7546; R² = 0.8652). The natural mortality estimation module was used in FISAT II and Pauley empirical formula was used to estimate the natural mortality of silver carp (Wang et al., 2019). The fishing mortality coefficient (FMC) of the maximum BL group is 0.5 as the initial value, do repeat calculation using iterative method, until the given FMC value and the weighted average value of the fishing death coefficient of each BL group converge, then the final FMC value can be determined.

Sampling was conducted daily during the fish spawning season. Sampling sites were set to ensure that they typified the four designated sections in terms of the larval fish development (Cao et al., 2007). The sampling process has been described as follows (Ren et al., 2016). Briefly, larval fish were collected using an ichthyologic trap net (0.5 mm mesh size and 4.9 m² area of the net mouth). A flow meter (Digital Flow Meter 23090, Denmark) was tied in the net mouth to measure the water flow volume. The sampling process was conducted for about 5 min on the spot using a traditional fishing boat against the current. Sampling time was set at every morning (08:00--11:00 a.m.) during the fish spawning season (May--August) during 2016--2019. The net was deployed about 20 m along the river bank side of the sampling sections. Larval fish sampling was conducted upstream using a traditional fish boat at a speed of 2 km/h. For each sample, after cleaning vegetation and foreign matters, ichthyoplankton were collected and then preserved in 75% ethanol for later classification and identification. At all sampling sections, WT, WF, and WTP were simultaneously measured, respectively. Daily water level and discharge were acquired from the Changjiang Hydrological Network.¹

The calculation method of larval fish density refers to the previous method (Cheng et al., 2013; Ren et al., 2016). According to the number of fish larvae collected, the sampling time, the difference between the beginning and end of the flow meter, and the active net area are calculated according to the following formula: Qi=(0.3×C⁢i×ai)/ti,Di=NiQi×ti, where Qi is the flow through the net mouth during the i time period (m³/s), C_i is the flow difference during the i time period in the flow meter, and a is the net port area of active gear (m²), 0.3 is the constant of the flow meter, and t is the time of collecting the fish larvae (s). Di is the density of fish larvae collected during the i time period (ind./m³) and Ni is the number of fish larvae collected in the i time period (ind.).

The coefficient of comparison is between the average density of fertilized eggs in the sampling section and the density of fertilized eggs in fixed point. C=∑d¯di, where di is the fish fry density at fixed sampling points and d¯ is the average density of the fish fry at each sampling site in the cross-section.

Fertilized eggs runoff during collection period is M_i = D_i×Q_i×C_i, where Mi is the number of eggs passing through the section of the river during the i time period, Di is the density of eggs collected during the i time period, and Qi is the cross-section flow during the i time of collection.

Fertilized eggs runoff during non-collection period is M_{ii + 1} = (M_i/t_i+M_{i + 1}/t_{i + 1})t_{ii + 1}/2, where Mi, i+1 is the fertilized eggs runoff during collection interval of the i, i+1 time and ti, i+1 is the collection time interval of the i, i+1 time period. The total runoff of eggs collected in the river section is M = ∑M_i + ∑M_{i,i + 1}.

The density of fish larvae was calculated as the fish larvae number per 100 m³ of the filtered water. The detrended correspondence analysis (DCA) and distance-based redundancy analysis (db-RDA) (length of gradient < 3, suit for lineal model, otherwise using unimodal model) were used to unravel the variables explaining the differences in larval fish assemblages (Pan et al., 2015). Statistical significance was set at P = 0.05. GAM was used to analyze hydrological factors on the abundance of fish larvae (Daskalov, 1999). The smoothing splines (cubic spline) were used to represent the non-linear effect of predictors. The maximum degree of smoothing was set at 6 to avoid unrealistic patterns in the explanatory variables and to reduce over-fitting. According to the Akaike information criterion (AIC), the optimal number of nodes and the optimal model are determined (Akaike, 1973). As the original data are relatively discrete, the larval fish density and water environmental factors are processed using log (x+1) to meet the normal distribution of the analyzed data and using Monte-Carlo permutation test to sieve environmental factors (Pan et al., 2015). SPSS 19.0 and Excel 2016 were used for data analysis and drawing graphs, and Pearson correlation test was used for correlation analysis.

All sampling procedures were reviewed and approved according to the Regulations for the Administration of Affairs Concerning Experimental Animals, as approved and authorized by the State Council of the People’s Republic of China.

During the survey period from 2016 to 2019, according to the annual average CPUE in each river section, results showed that the annual average catch weight, the average annual catch number, and the BL range in the HK, AQ, DT, and CS sections were different (Table 1). Further analysis showed that the BL range of silver carp in the lower reach of the Yangtze River was 31.9–876 mm, and the average BL was 283.52 mm, of which 100–300 mm was the dominant BL group, accounting for 51.27%. The weight range of silver carp is 65.02–6,500 g, and the average weight is 700.99 g, of which 200–400 g group is the dominant weight group, accounting for 57.89%, and the average weight is 387.61 g (Figure 2).

The cumulative fishing mortality, annual total resources, and annual total tail of silver carp in HK, AQ, DT, and CS show significant river section differences (Table 2). The cumulative mortality of fishing in the HK section is significantly lower than that in other sections, and the difference between the other three river sections is not obvious. The average annual fishery resources and tail in each river section are CS > DT > HK > AQ. Interestingly, the total fishery resources of CS section are significantly higher than those of other sections.

A total of 1,139 silver carp larvae were collected during 226 days of field survey in HK section from 2016 to 2019. A total of 224 fertilized eggs were collected at different stages, including tail bud-emergence stage, blast cyst stage, and early gastrula stage. During the investigation period, the fish larvae and fertilized eggs of silver carp showed an increasing trend year by year. According to above mentioned methods, the estimated runoff of silver carp in the early life history passing through the HK section from 2016 to 2019 is 5.08 × 10⁸, 5.69 × 10⁸, 6.72 × 10⁸, and 7.32×10⁸, respectively. A total of 826 silver carp larvae and 126 fertilized eggs were collected, from 2016 to 2019, during 312 days of field survey in the AQ section. During the investigation period, the variation of fish larvae and fertilized eggs of silver carp showed a fluctuating trend annually. According to the number of larvae and fertilized eggs collected, the runoff of silver carp larvae passing through the AQ section from 2016 to 2019 is estimated to be 8.08×10⁸, 9.69 × 10⁸, 9.72 ×10⁸, and 10.32×10⁸, respectively. However, only 36 silver carp larvae and 2 fertilized eggs were collected during 236 days of field survey in DT and CS sections during the investigation period. By the same estimating method, the runoff passing through the DT and CS sections can be ignored.

During the investigation period, different BL groups of silver carp showed different distribution patterns in the sampled sections (Figure 2). Silver carps of different BL groups in the HK section showed a downward trend year by year. The juvenile silver carp in the below 200 mm BL group in the AQ section were significantly higher than those in other BL groups and showed a downward trend year by year. The silver carp in the 200–400 mm BL group showed an upward trend yearly, and the interannual distribution in the 400–600 mm and above 600 mm BL groups has no significant difference. The ratio of juvenile silver carp in the below 200 mm BL group in the DT section showed an obvious upward trend year by year, while the 200–400 and 400–600 mm BL groups showed a downward trend. The ratio of juvenile silver carp in the below 200 mm BL group and adult silver carp in the above 600 mm BL group in the CS river section is small and showed an obvious upward trend year by year. The 400–600 mm BL group has an absolute advantage in the surveyed section, but the resources showed a downward trend year by year.

Silver carp of different BL groups also have obvious seasonal differences between the investigated sections. Seasonal variations in the adult fish in different sections are shown in Figure 3. During spring, autumn, and winter, the above 600 mm BL group of silver carp in the CS section was significantly higher than that in other sections. The 400–600 mm BL group in each section in summer was obviously high. In autumn, the silver carp in the below 200 mm BL group was dominant in the HK, AQ, and DT sections, but the ratio in the CS section was lowest. In winter, the ratio of juvenile silver carp in the HK and AQ sections in the below 200 mm BL group was high, and the proportion of adult silver carp (400–600 mm and >600 mm BL group) in the DT and CS sections was high. The results showed that in the river-lake connected sections such as HK and AQ, the proportion of juvenile silver carp in the below 200 mm BL group was significantly higher than that in other investigated sections.

During the survey time, as for the larval fish results, the highest abundance of larval fish and fertilized eggs appeared in the HK and AQ sections. Comparing the average abundance of larval fish in each sampling section exhibited the trend as HK > AQ >> DT > CS. Interestingly, the highest abundance of the larval fish appeared in the HK section, and the lowest abundance appeared in the CS section, which was almost never sampled. Furthermore, the larval fish and fertilized eggs were continuously sampled from early June to early August, and peaked in early July. The GAM results showed that the hydrological factors significantly influence the abundance of fish larvae (Figure 4; P < 0.05). In the HK and AQ sections, the contribution orders were the daily increasing rate of water level, WTP, water temperature, and the daily increasing rate of discharge (Table 3, P < 0.05). Among them, the total deviation interpretation rate by the optimized GAM was 67.2 and 83.7%, respectively. Results revealed the concrete influence of hydrological conditions on the abundance of the larval carps (Figure 4).

The silver carp is a type of semi-migratory fish species between rivers and lakes, and it grows up, matures in floodplain lakes, and spawns in rivers (often triggered by increasing temperature and rising water levels). Due to overfishing, water pollution, habitat loss, and hydroelectric engineering such as the TGD, the natural resources of the four Chinese carps have declined considerably during the past decades. Studies showed that the closer to the dam, the more obvious the effect is. Most studies on the complementary mechanism of the silver carp focused in the up and middle reaches, and the lower reach of the Yangtze River has often been ignored. Some studies have indicated that the recruit population contribution has shifted from the middle reaches to the lower reaches. In this case, studies on the complement mechanism of the silver carp in the lower reach of the Yangtze River became even more important. Furthermore, the hatching of fertilized eggs and the development of fish larvae were transferred by the downstream drift and subsequently juvenile fish were moved into nursery areas and grown up, and such biological process probably happens in the lower reaches. For this reason, clarifying the growth and death characteristics of silver carp in the early life stage and its supplementary contribution to the adult population will provide a scientific and basic data support for the protection of silver carp resources in the Yangtze River.

The field survey results showed that the closer the section is to the estuary, the fewer the larvae silver carp is, and the bigger BL group gradually increased. Adult silver carps can swim freely and have strong control ability in water. What is more, the adult silver carps tend to choose habitats with slow water flow, rich biological bait, and less human interference. In the lower reach of the Yangtze River from HK to the estuary, there are many sandbars, developed water network, distributed thousands of bays, and rich bait biological sources along the river. This kind of habitat is an important feeding ground for the silver carp (Shuai et al., 2018). Results in this study found that adult silver carps were widely distributed in the investigated sections. In the section with sandbar habitat, such as HK, AQ, DT, and CS, the distribution density of the adult fish was high, and the distribution of BL group was rich. Interestingly, the adult silver carp resources are highest in the CS section. The CS section is wide, with obvious river diversion, and has the characteristics of rising and falling tides. Some studies have shown that the tide can significantly promote the enrichment of plankton, which are the main feed resource for the silver carps. Therefore, comparing with other sections, the CS section has more abundant bait organisms and is more suitable for the fish growth. Another important reason for more big silver carps in this section may be the numerous and effective stock enhancement in the lower reaches of the Yangtze River in recent years.

The spatial heterogeneity of the fish larvae is mainly affected by the hydrological situation and the topography of the water area. The sandbars obviously existed in the HK and AQ sections and divided water flow into two or more streams. This habitat usually can form the “vortex water,” which is necessary for the silver carp spawning behavior. Based on the study result, it can be speculated that the spawning ground are distributed in Jiangzhou, Longping, and Qizhou waters, which contain many sandbars, with eddies or bends in the river center (unpublished data). There is no spawning ground found below AQ section. The main reason is that after silver carp spawning, the life history stages of drifting and first feed must be completed in the early development stage; the drifting distance from AQ to the estuary cannot meet the developmental needs of the silver carp from the fertilized eggs to the initial feed stage (Shuai et al., 2018). The HK section is connected to the Poyang Lake, which is the river-lake connected section in the lower reach of the Yangtze River, and the AQ section is a relatively stable bifurcated river with many complex beaches. This complex ecosystem with the characteristics of connecting river and lake (RL) can provide good nursing and feeding places for silver carps.

Silver carp has the migratory habit between river and its connected water body, and its distribution pattern has obviously seasonal differences (Zhang et al., 2021). Because the adult fish may choose the complex section for winter migration, in spring, it was found that the ratio of the adult fish with a lower BL was higher in the HK and AQ sections. Summer is the spawning season, the WT and the WL begin to rise in the river, and under the accelerating current, the adult fish begin spawning. For this reason, the number of middle and high BL groups increased generally in the HK and AQ sections. After the spawning period, the hatched fish larvae successively entered Poyang Lake or Wanhe River for nursing and feeding. However, the CS section has a little salinity and is not fit for the hatched fish larvae. Therefore, the ratio of the below 200 mm BL group was lowest in the CS section. It can be seen that the seasonal distribution characteristics of adult silver carp are closely related to their reproductive biological characteristics and river-specific habitat.

Drifting development extensively exists in the early stage of river fish species, and its growth and development is also affected by abiotic and biological factors (Nagrodski et al., 2012). Abiotic factors include WL, riverbed shape, WT, and hydrological conditions. Biological factors include fish development stage, habitat preference, and interspecific relationship (Jager and DeAngelis, 2018). The temporal dynamics of the silver carp in the early life history stage can directly reflect the varying characteristics of the adult fish spawning time. The WT for the silver carp spawning is generally required to be above 18°C; in this study, the fish eggs in the HK section were investigated from late May to mid-July, and the WT fluctuated in the range of 20.4–28.7°C. In the fish spawning season, the changes in river hydrology, such as the increase of water flow, the rise of water level, and the acceleration of river flow velocity, are important and essential factors causing the spawning peak of the fish (Zhang et al., 2021). In 2019, the runoff and WL began to rise in late May and remained high until the middle of July, which induced three fish egg peaks in the HK and AQ sections. Related studies have shown that WT, WF, and WTP were essential environmental factors affecting the community structure of fish larvae (Machado et al., 2017). Furthermore, the GAM results showed that the hydrological factors have the different contribution rate to the abundance of fish larvae, which was in accordance with the reproductive biology of the silver carp (Kun et al., 2019).

When analyzing the spatial distribution of fish larvae abundance in the HK and AQ sections, it is found that the fertilized eggs were collected in the three sites except the Waterway of Poyang Lake in the HK section. It can be inferred that the supplement population of silver carp only came from the upstream section of the Yangtze River. Due to the different natural conditions such as riverbed shape, river flow velocity, and watershed dynamics between various sites in the section, the drift density of fish larvae varied greatly between different sites on the same river section. Interestingly, the egg density in the river center is significantly higher than that on both banks, which is consistent with the biological characteristics of silver carp egg development and drifting (Wang et al., 2019).

Understanding the complementary mechanisms is critical to the effective conservation of wild fish populations (Goldenberg et al., 2019). There were numerous evidences that qualified relationships existed between hydrology and density of fish larvae (Doyle et al., 2002; Jenkins et al., 2015). When a natural flood with water velocity ≥0.7 m/s occurred in the Mississippi River, it can stimulate the sliver carp spawning behavior (Shuai et al., 2018). The project of biological regulation from TGD also can lead hydrological changes and stimulate the silver carp to spawn in the spawning season (Zhang et al., 2021). Unfortunately, it has been proven by field surveys that the construction of TGD had a dramatic influence on the abundance of silver carp larvae (Zhang et al., 2020). Dams in the Yangtze River have modified water hydrology and flow and then affected the silver carp migration and reproduction (Fu et al., 2003). Because the lower reaches of the Yangtze River are far away from the TGD, the hydrological characteristics are less affected (Doyle et al., 2002). Furthermore, the connectivity between Poyang Lake and the Yangtze River is conducive to the fish reproduction. The RL ecosystem forms the reason for the adult amount of resources of silver carp in the surveyed sections. Furthermore, the most suitable discharge of spawning habitats of silver carp in the Yangtze River were 15,000–21,300 m³/s (Yu et al., 2018). Due to the inflow of tributaries along the river, this is a kind of essential topographic habitat for forming the bubble water (Shuai et al., 2018). In this water area, the adult sliver carp spawned and the fish in early life history stages developed. However, the complementary quantity and quality derived from the adult population absolutely depends on the parent fish. How many parent fish lived in the Yangtze River and whether they can spawn successfully should be considered systematically. Therefore, continuous field surveys on larvae and adult fish resources in the whole Yangtze River Basin and exploring the complementary relationship will be considered persistently.

Although only 4 years of systematic data were included in this study, we proposed that the adult silver carp has the distinctive preference in the habitat selection. Furthermore, being the overfishing in the Yangtze River, the adult fish that can spawn in the spawning period all come from the up reach of the river, and downstream fish have less complement contribution to the larval fish population. In contrast, it is also found that the river flows, such as discharge, artificial flood, or flood pulses, are factors that effectively determined larval fish abundance in the Yangtze River. Based on the above results, we can propose an exchange regulating method that the parent fish can intercourse each other between the upstream and downstream reaches in the Yangtze River during their breeding season. By doing these, their population, perhaps, expands quickly and their breeding population can recover in the future, and it would be a preferable method to control the propagation of silver carps (Ban et al., 2019; Wang et al., 2019). Our study provides preliminary clues to understand how the adult silver carps recruit and migrate between sections along the Yangtze River. Besides, understanding the breeding characteristics of the silver carps in their native range and their protecting biography are also of great significance. Future research will focus on the detail complementary mechanism silver carp both in the natural habitat and invasion waters.

This study has carried out research on the temporal and spatial dynamics of adult and early resources of silver carps and analyzed the supplementary contribution of early resources in different sections of the lower reaches of the Yangtze River to adult populations. The research results provided theoretical support for the formulation of fishery resources conservation and management measures in the Yangtze River. The accumulation of these basic data can also provide a scientific basis for resource evaluation and species protection strategies of silver carps in the Yangtze River. Poyang Lake has typical habitat characteristics of connecting rivers and lakes; it is the key ecosystem to maintain the successful supplement of silver carp early resources to adult fish resources. However, how to ensure the connectivity of the RL effectively, so as to enhance the silver carp resources in the Yangtze River, should be focused on especially. Therefore, wading projects such as sand mining, shipping, and aquatic operations in silver carp migration channels, spawning grounds, and feeding grounds must be prohibited during the spawning period. At the same time, the correlation analysis between adult and fish larvae resources of silver carp is conducive to clarify the complementary mechanism. In short, establishing a reliable biological data and explaining the supplementary contribution of early resources to adult resources will be the key content of later research.