Spatial Biomass Production And Seasonal Nutrient Limitation Transitions In A Tributary of The Three Gorges Reservoir, China

The freshwater ecosystem characteristics in terms of nutrient inventory across seasons, spatial variations of chl-a biomass, and the phytoplankton community structure are prudent ecological assessment indices for a bloom management protocol. We evaluated the spatial and seasonal chl-a distribution under different nutrient conditions and phytoplankton community structure in a eutrophic Three Gorges reservoir tributary China. Result showed signicant variations in biomass production with the mainstream reaches severely affected. The nutrient addition bioassay demonstrated signicant stimulations on growth in both autumn and summer. The nutrient limitation pattern shifted from P in autumn and spring to N limitation during summer. Combined additions of trace metals with N, P, and Si in autumn and Fe alone enrichment in summer and spring showed maximum productivity. The phytoplankton community structure demonstrated strong sensitivities to seasonal variabilities with regime shift from Cyanophyta, dominated by the toxic and hypoxia generating, Microcystis spp in both autumn and summer, the Cryptophyta dominated by the Chroomonas acuta in spring to the Bacilliariophyta dominated by the genera, Cyclotella in winter. This reected the ability of the Bacilliariophyta to thrive under a low-temperature condition. Combined N&P led to signicant growth stimulation in summer while P alone controlled the bulk of the growth in autumn. The study points to the need for extending mitigation steps to the mainstream towards achieving lasting bloom management solution in the impacted tributary. The present study reports the spatial and seasonal distribution of chlorophyll-a (chl-a) induced by the remarkable uctuations in the different nutrient concentrations across the four seasons; autumn, winter, spring, and summer in XXB. Our study proves that the mainstream reservoir is severely eutrophic with signicant biomass production as chl-a during the intense summer bloom than the rest of the reaches. The phytoplankton community structure showed strong responsiveness to the seasonal variability with a regime shift from Cyanophyta, dominated by the Microcycstis spp in both autumn and summer. The study reveals relative stability in the concentrations of TN, while the TP, NO 3 , NH 4 , DTN, and DTP showed signicant seasonal variabilities. An NH 4 decit and NO 3 replete condition was reported during spring which probably initiated the total extinction of the Cyanophyta that show absolute preference to ammonium-nitrogen. It reckons that nutrient inputs regulations in the surrounding counties; Zigui and Xingshan in Xiakou Town though potent in the proximal tributaries would not be adequate to address the current state of bloom in the mainstream. It, therefore, points to the need for an extended mitigation design to the highly eutrophic mainstream which recharges the XXB through the backwater intrusion during the peak rise seasons. The study thus proposes broad monitoring programs while emphasizing that sectional ecological assessment would not effectively characterize the whole freshwater systems and recommend such extended assessment convention in the similar reservoirs and tributaries worldwide facing similar anthropogenic pressures.


Introduction
Eutrophication, which adversely affects both the chemical and ecological status of aquatic ecosystems worldwide has remained unabated despite several mitigation approaches (Nwankwegu et al., 2019).
Aside from the physicochemical and chlorophyll-a (chl-a) concentration, other water quality indices for bloom assessment are till date controversial (Sòria-perpinyà et al., 2019). While the former focuses on the quanti cation of the nutrient inventory, nitrogen (N), phosphorus (P), as well as some physical parameters including; pH, dissolved oxygen, and (DO), the later provides information on the relative phytoplankton community biomass and assesses the eutrophic status of aquatic systems. The biomass production as chl-a represents a more direct measure of the extent of primary productivity and bloom severity in the eutrophic freshwaters. The magnitude of biomass yield during an intense bloom can also provide a presumptive ecological assessment on the speci c genera/species dominating a bloom (Nwankwegu et al., 2019). The physiochemical index merely assesses the implications of the increasing anthropogenic activities including N& P inputs loading, as well as some pH and DO-altering chemicals.
This index, unlike the chl-a concentration certainly does not address concerns relating to the phytoplankton community structure. The chl-a estimation is thus most common index for the trophic state categorization, impact indication, and water quality management (Zou et al., 2020; Sòria-perpinyà et al., 2019; Carneiro et al., 2014). Further, based on the European Water Framework Directive, chlorophyll-a and cyanobacteria have been adopted in the ecological classi cation of freshwaters e.g., lakes (Søndergaard et al., 2011) perplexed by the multiple climatic, anthropogenic, and environmental stressors.
In the eutrophic Xiangxi Bay (XXB), a Yangtze River tributary in Hubei Province, China, bearer of the world's largest impoundment project, the Three Gorges Dam , extended monitoring campaigns in terms of the spatial biomass production and the seasonal nutrient limitation transitions have not been considered in the past 15 years. We are also not aware of any recent study that has considered the full report of the XXB seasonal pattern of nutrient limitation shift, microelements stimulation on growth, and spatial biomass yields, particularly during the productive period. Several previous studies in the tributary have evaluated the spatial analysis of biomass approximately 15 years ago e.g. spring bloom and nutrient limitation (Fang et (Nwankwegu et al., 2020b). Spatial biomass accumulation during productive periods, the seasonal nutrient limitation, as well as growth stimulation by microelements are particularly imperative to infer the contemporary eutrophication impacts in the vulnerable ecosystem.
In the present study, we, therefore, studied the spatial variations in the distribution of chl-a biomass along the XXB major representative reaches to the mainstream during the intense summer bloom. We also evaluated the seasonal dominance in the phytoplankton taxonomic groups across seasons (autumn, winter, spring and summer), and ascertained the seasonal variations in the critical nutrient pool controlling phytoplankton growth. Further, we constituted nutrient addition bioassays involving the principal nutrients (N&P), as well as the micro-nutrients; silicon (Si), iron (Fe), zinc (Zn), manganese (Mn), and copper (Cu) to assess their responses on growth promotions while estimating the synergy among the different nutrient combinations. We hypothesized a signi cant difference in biomass production between the mainstream and the other points within Xiangxi Bay. The present study would provide useful insights into effectively understanding the aquatic ecosystem functioning while complementing the current . The bay is heavily impacted by multiple anthropogenic stressors including; the direct discharge of untreated domestic sewage, agricultural runoffs due to the commercial row-crop agriculture (Nwankwegu et al.,202b) dominated by leguminous crops and potentially increasing the ecosystem N budget. The in-situ laundry services by the local people of Xingshan and Zigui counties also accounts for the enormous phosphorus (P) input into the system. The annual phosphorus load has been estimated to 29.78 tons (Nwankwegu et al., 2020a) and 61-40% of TN and TP inputs have been attributed to the nonpoint sources from the upper XXB tributary (Yang et al., 2015). The mainstream nutrient structure has been described far beyond the internationally recognized eutrophication threshold (Zhou et al., 2011;Nwankwegu et al., 2020a). There are also several industrial localization within the XXB bank, typical examples are the Shu Kongping chemical and mining industrial in Gufu and Xiakou towns, the end of backwater zone, which are the potential sources of anthropogenic nutrients, especially phosphorus leaching from waste rocks into the XXB watershed (Jiang et al., 2016). The nutrient characteristics without a proportionate tidal inundations make the bay a typical nutrient-saturated system (Dai et al., 2010;Zhijing et al., 2012) and the harmful cyanobacterial blooms have remained unabated in recent years despite several mitigation attempts.

Nutrient addition bioassay experiment
Sampling was conducted over one year from August 2018 to August 2019 aimed at covering the four seasons; autumn, winter, spring and summer. The sampling points for the spatial chl-a distribution during the intense summer bloom were; the downstream reach (XX01), the middle reach of XXB (XX06) otherwise known as the XXB con uence, which is 18.8 km from the TGR, and the mainstream TGR (XX09), 32 km from XXB. The collected samples were returned to the platform at XX06 for incubation (see Fig.S1). The location XX06 was used in other seasons while the bloom intensity was low. The navigation of the sparsely distributed sampling points was made possible with the use of a steamer vessel belonging to China's Three Gorges University, Yichang, Hubei province. Nutrient addition bioassay experiments were deployed in XXB during autumn (20th August -15th September 2018), winter (20th December 2018-9th January 2019), spring (5th April-April 2019), and summer (19 July-1st August 2019). Subsurface (approximately 0.2 m) water samples were collected into a cleaned plastic container (280 L) which was 0.01 N HCl-sterilized and XXB water-rinsed. The chlorophyll-a (chl-a) biomass surface scum was scooped off during water samples collection to obtain a moderate initial biomass density of approximately 10 µgL − 1  during summer and autumn when primary production was most severe with the characteristic large surface cyanobacterial "mats". Water samples pre-screening for zooplankton removal was not considered. This is because the modi cation signi cantly alters the baseline phytoplankton assemblages while increasing the bias associated with the extrapolation of the bioassay outcomes to the natural community structure of phytoplankton Domingues et al., 2016). In natural eutrophic freshwater, the complex interactions often do not stop the harmful algal bloom expansion. So, removal of the non-plankton diversity by ltration can produce results which cannot accurately represent the natural ecosystem characteristics. Further, particulate nutrients and some planktonic algae may be held within the bloom bio lm and are removed during ltration . This potentially creates arti cial phytoplankton community structure and nutrient limitation pattern that could differ signi cantly with the ecosystem actuality. The pooled water sample was continuously stirred for even distribution of nutrients, particles, and phytoplankton species while dispensing 4 L aliquot into acid (0.01 HCl) and then XXB water-

Water sample analysis
In-situ measurement of pH, temperature, dissolved oxygen (DO), and electrical conductivity were performed using multi-sensor Yellow Spring Instrument ((YSI Quatro, 18G100594, USA)) automatic calibration aided. Total nitrogen (TN), total phosphorus (TP) and dissolved nutrients TDN, and TDP were analyzed by combined persulfate digestion and further followed by spectrophotometric analysis for nitrate-nitrogen (NO 3 − -N). Ammonium nitrogen (NH 4 + -N) was analyzed using the indophenol blue method

Statistical analysis
Each set of data in the experiments was collected in three replicates and the analytical result was the mean of three measurements. The standard deviations (error bars) and statistical signi cance (5 % level of signi cance) were analyzed with origin software. Differences in the growth responses (chl-a) between various treatments were analyzed by one-way ANOVA. The Post Hoc Multiple comparisons of treatment means across seasons were performed by Tukey's least signi cant difference procedure. The classical growth rate (µ) equation mathematically expressed as; µ = ln(X 2 -X 1 )/(T 2 -T 1 ) was adopted in estimating the growth rates based on chl-a in uenced by the nutrient enrichments across the treatments and sampling points. Where X 1 is the concentration of chl-a at (T 1 ), and X 2 is the concentration of chl-a at the time (

Seasonal nutrient dynamics
The nutrients concentrations (Fig. 3) (Fig. 3a). However, the reliability of TN/TP ratios (stoichiometry) as an index for nutrient limitation has been previously questioned (Lv et al., 2011). The maximum concentration of NO 3 (2.98 ± 0.28 µgL − 1 ) and the least concentration in NH 4 (0.28 ± 0.09 µgL − 1 ) was reported in spring indicating a nitrate replete and ammonium de cit condition during spring bloom while the minimum concentration of NO 3 (0.81 ± 0.02 µgL − 1 ) was observed during summer. This further explains why in Fig. 2 above, Fe addition with NO 3 led to signi cant biomass production. The maximum concentration in NH 4 (1.08 ± 0.07 µgL − 1 ) was reported in winter probably re ecting the total absence of ammonium-dependent phytoplankton groups e.g., the Cyanophyta in winter (Zhou et al., 2012). The dissolved N (DTN) and P (DTP) also showed strong uctuations across seasons. The maximum DTN concentration (1.58 ± 0.09 µgL − 1 ) was obtained in autumn and least in spring (Fig. 3b) . In this case, the protracted N utilization without replenishment and an increasing P supply from sediment and the geogenic resources e.g., phosphorus rocks can exacerbate a critical Nlimitation in freshwaters. On the whole, the spatial nutrient response pattern during summer strongly reveals N-limitation indicating that N enrichment would greatly control growth while prolonging primary productivity in the entire system.

Seasonal dynamics in biomass production as chl-a
The phytoplankton biomass (Fig. 4) considerably varied across seasons. Each season showed peculiar responses on growth induced by the different nutrient enrichments. In autumn, nutrient additions involving N, P, and in combinations revealed biomasses as chl-a in N (48.47 ± 6.05 µgL − 1 ), P (59.76 ± 6.73 µgL − 1 ), and N + P (68.40 ± 5.73 µgL − 1 ). This indicated that all the nutrient enrichment stimulated growth responses which differed signi cantly with the ambient chl-a concentrations both in the initial and control conditions. Although nutrient promotions on growth by N and P separate additions during autumn showed no signi cant difference (p > 0.05), the nutrient response adequately showed P-limited than N limited growth condition. However, the XXB nutrient limitation pattern slightly differed from the previous study by Paerl et al., (2011) in a shallow eutrophic freshwater, Lake Taihu where P limitation was reported in winter-spring but it is important to recall that these two freshwater systems have different ecosystem behaviours both in depth, hydrodynamics, nutrient uxes, and phytoplankton species structure. For example, a consistent P-limitation of phytoplankton growth in eight (8)  Internal nutrient loading is a potential process regulating phosphorus dynamics contributing up to 86% P reduction from legacy stores, phytoplankton, chlorophyll-a, and cyanobacterial blooms (Radbourne et al., 2019). Qin et al., (2020) recently showed that while P limitation predominates deep lakes and reservoirs, N limitation predominates shallows lakes and reservoirs. This indicates that a deviant nutrient limitation pattern is possible and largely dependent on the water depth. In the same study, the authors argued therein, that the biogeochemical mechanisms associated with water depth essentially control the nutrient dynamics in the freshwater systems. It further demonstrated that in shallow systems usually characterized by mixing depth > maximum depth, sediment exchanges and water column in uences are dynamic thus often exacerbate potential N loss (denitri cation) and enormous P release from the legacy store through precipitation, leading to low N:P ratio and consequently elicits N limitation. Conversely, in the deep systems, which are characterized by mixing depth < mean depth, the hypolimnion boundary receive minimal turbulent/perturbation seasonally while the maximum hydrodynamic actions are concentrated on the epilimnion. The retarded N loss with the increased P loss often through sedimentation and immobilization directly triggers the elevated N:P ratio causing P limitation to prevail. The release of the high level of remobilized P from sediment which signi cantly encourages maximal primary productivity . This invariably suggests that light availability which is critical to chlorophyll-a dynamics decreases with depth in the eutrophic freshwaters. The N&P simultaneous addition in autumn led to maximum growth with P potentially stimulating optimal growth than N in the combinations. It thus reveals that in the light of the nutrient structure, P-alone rather than the traditional association of N&P could drive to a logaritmic biomass production in XXB during autumn.
In winter, all the nutrient additions including N, P and their contemporaneous additions resulted in growth responses which did not differ (p > 0.05) with the control. It was evident to deduce that nutrient limitation did not characterize winter bloom in XXB indicating that the limitation could be attributed to other factors including the dramatically reduced photosynthetically active radiation (PAR) and water temperature. In spring, a similar nutrient limitation characteristic as autumn was observed but the magnitude of biomass production in both seasons signi cantly differed (p < 0.05). In summer, nutrient limitation shifted to N as N-alone addition caused a growth response which did not differ signi cantly with N&P combined addition but differed signi cantly with both P-alone addition and the control. The N limitation reported in summer corroborated previous studies in the different freshwater systems (Paerl et al., 2011;. In Waquoit Bay, USA, the accumulation of phytoplankton biomass in brackish and saline water was limited by the supply of nitrate during warm months (Tomasky et al., 1999).
The effect of different nutrient combinations with the essential micro-nutrients including Fe (both alone and in combinations), Si, Mn, Zn, and copper showed strong variability to seasons (Fig. 5).

Seasonal taxonomic dynamics
Signi cant seasonal variabilities in the population of each taxonomic group were observed (Fig. 6). Five principal phytoplankton taxa notably; Cyanophyta, Baccilliariophyta, Chlorophyta, Cryptophyta, and Pyrrophyta were identi ed in XXB although, in autumn, the community structure included a few populations of the Euglenophyta and Xanthophyta. The pattern of taxonomic dominance showed strong sensitivity to variations in season. Each season, therefore, selected the dominance of one or more taxa but not all at the same time except in autumn where all the taxonomic groups were represented in the largest amount relative to other seasons. This indicates that autumn condition supports a wide range of phytoplankton taxa. During autumn, the community structure revealed signi cant cell densities in both the Cyanophyta (3.8×10 7 cellsL − 1 ) and Chlorophyta (3.3×10 7 cellsL − 1 ). A relatively lower cell densities were observed in Bacilliariophyta (1.00×10 7 cellsL − 1 ), Cryptophyta (3.10×10 6 cellsL − 1 ), and Pyrrophyta

Species transitions across seasons
Species/genera dominance varied strongly across seasons ( Table 2)

Conclusion
The present study reports the spatial and seasonal distribution of chlorophyll-a (chl-a) induced by the remarkable uctuations in the different nutrient concentrations across the four seasons; autumn, winter, spring, and summer in XXB. Our study proves that the mainstream reservoir is severely eutrophic with signi cant biomass production as chl-a during the intense summer bloom than the rest of the reaches.        Supplementary.docx