Weak local upwelling may elevate the risks of harmful algal blooms and hypoxia in shallow waters during the warm season

Harmful algal blooms (HABs) and hypoxia, as common ecological disasters, are typically site-specific and recurrent, causing enduring environmental problems for coastal communities. Although these issues are often attributed to coastal eutrophication, in certain low-nutrient areas, such ecological disasters still frequently occur while the underlying cause is poorly understood. A prime example is the Qinhuangdao coastal waters in North China. This study intends to investigate the controlling factors of such incongruous ecological issues recurring in those low-nutrient areas with a case study of Qinhuangdao, utilizing numerical models and satellite observations. The result indicates that the weak tide-induced upwelling during summer creates favorable conditions (warm water with high transparency) for the occurrence of HABs and hypoxia in this region. It is due to that Qinhuangdao is precisely situated at the current amphidromic point of semi-diurnal tides, characterized by weak tide mixing. Likewise, the same story happens on the northern shelf of the Shandong Peninsula, where ecological problems are also prominent in China. The present study implies that shallow waters with weak local upwelling are susceptible to ecological issues during the warm season. This finding challenges the traditional view that strong-upwelling zones are more vulnerable to ecological disasters due to ample nutrient supply within the euphotic layer. It implies that tidal dynamics can greatly affect the vulnerability of coastal waters to ecological issues, which can be of significance to coastal management. Furthermore, the finding may have broader global applicability, given the ubiquity of tide-induced upwelling in various other coastal regions.


Introduction
Coastal waters are vital for global marine ecosystems, contributing to ∼25% of global ocean primary production (Smith and Hollibaugh 1993).Yet, they are facing increasing pressures from landsea interactions, leading to frequent ecological disasters, such as harmful algal blooms (HABs) and hypoxia (Smetacek andZingone 2013, Glibert et al 2018).These issues significantly impact on marine fisheries and have gained widespread attention from coastal communities (Azanza et al 2005, Richlen et al 2010).
Both HABs and hypoxia typically exhibit a recurrent pattern in specific locations, coming with severe environmental issues (Charlier et al 2006, Liu et al 2010).Such recurring ecological disasters are often attributed to coastal eutrophication, characterized by elevated nutrient levels or imbalanced nutrient ratios, especially in large estuaries and densely populated areas (Li et al 2014, 2015, Zheng and DiGiacomo 2020).Nevertheless, in certain regions characterized by recurrent ecological disasters, the eutrophication level may not necessarily be high.An exemplary instance can be found along the Qinhuangdao coast in China.
Qinhuangdao is a renowned tourist city in China, situated on the western coast of the Bohai Sea (BHS; figure 1(a)).Since 2009, Qinhuangdao coastal waters (QCWs) have suffered from the outbreak of severe brown tides during summer (Kong et al 2012, Zhang et al 2012, Xu et al 2017, Yu et al 2018).Statistical data covering over 50 yr demonstrated that QCW stands out as the most frequently affected area by HABs in the BHS, apart from the Bohai Bay (Song et al 2016, Li et al 2023).Satellite observations also confirmed the highly frequent algal blooms in this region (figure 1(c); He et al 2013, Zhai et al 2021b, 2023).Moreover, QCW is also known for extensive zones of hypoxia and acidification at the seabed during summer (Zhai et al 2012, Zhao et al 2017, Wei et al 2019, Song et al 2020, Chen et al 2022).However, it is odd that the average eutrophication level in QCW is not exceptionally high, particularly in comparison to the coastal bays in the BHS (Wang et al 2009, Yang et al 2022).This inconsistency implies that there might be other factors, rather than ecological factors, responsible for the high ecological risks in QCW.
Previous studies are primarily devoted to identifying the causative species of HABs (Kong et al 2012, Zhang et al 2012, Xu et al 2017) or the biogeochemical processes underlying hypoxia (Zhao et al 2017, Song et al 2020, Zhang et al 2022a).While there have been a few attempts to elucidate the high ecological risks in QCW (Cao et al 2018, Yao et al 2019, Wu et al 2022, Zhou et al 2023), the majority rely on qualitative speculations and can hardly provide a comprehensive explanation to the abnormally high ecological risks in this region.In other words, previous studies struggle to explain why QCW is so unique to be favored by the ecological disasters year after year.This situation greatly hinders the environmental protection and management in this region.
This study intends to investigate the underlying cause for the high risks of HABs and hypoxia recurring in these low-nutrient waters based on a case study of QCW.Numerical models and satellite observations will serve as primary research tools.Furthermore, we will approach this research from a perspective of hydrodynamics.An essential consideration is that our study area is characterized by low nutrient levels, implying that ecological factors may not be the predominant drivers.The outcomes of this study hold the potential to enhance our understanding of ecological disasters and contribute to coastal environmental management.This paper proceeds as follows.Section 2 introduces the study area.Section 3 is about the data and methods used in this paper, including satellite observations, hydrodynamic and ecosystem models.Section 4 presents the main findings.Section 5 discusses the significance of the findings.Finally, section 6 draws a brief conclusion of the paper.

Study area
The study area, QCW, is situated on the western side of BHS, which is the only inland sea in North China.The BHS is located on the continental shelf of the Northwest Pacific Ocean and is linked to the open sea via the Yellow Sea.Tidal waves originating from the open sea travel northward into the BHS, creating an intricate and robust tidal amphidromic system, primarily characterized by semidiurnal tides (Bian et al 2016).In the warm season, strong tidal motions lead to significant coastal tide-induced upwelling in the BHS and its surrounding areas, manifesting as conspicuous cold patches on temperature images at the surface (Lü et al 2010).The average depth of BHS is ∼18 m, and its shape resembles a triangle with three bays (Bohai Bay, Laizhou Bay, and Liaodong Bay) at its vertices.Furthermore, numerous rivers, including the Yellow River, Luanhe River, and Liaohe River, flow into this semi-enclosed shallow basin.Among these, the Yellow River is renowned for having the world's second-largest sediment discharge (Milliman and Meade 1983).

Satellite observations
The suspended particulate matter (SPM) were collected from the Global Ocean Colour, Bio-Geo-Chemical, Level 4 dataset provided by Copernicus Marine Environment Monitoring Service (CMEMS).This monthly dataset integrates ocean color observations from multiple sensors spanning from 1997 to the present with a spatial resolution of 0.05 • in both longitudes and latitudes.Daily products of sea surface temperature (SST) were acquired from the Operational Sea Surface Temperature and Ice Analysis available through CMEMS, covering the period from 2007 onwards, with a spatial resolution of 0.0417 • .

Hydrodynamic model
Modeling work is based on the Semi-implicit Cross-scale Hydroscience Integrated System Model (SCHISM; Zhang et al 2016).SCHISM is a 3D, hydrostatic, baroclinic model grounded on unstructured grids, which is a derivative product of the original semi-implicit Eulerian-Lagrangian finite-element model (Zhang and Baptista 2008).This model is designed for seamless simulation of 3D baroclinic circulation across creek-lake-river-estuary-shelf-ocean scales.This model has been extensively implemented in global oceans, demonstrating high reliability in hydrodynamic modeling (Wang et al 2020, Ye et al 2020, Li et al 2021).
SCHISM version 5.9 was implemented in the Bohai and North Yellow Sea (figure 1(a)).We designed a highly idealized experiment to explore the response of SST and SPM to tidal motions.The model grid consists of 30 841 nodes and 58 380 elements.The bathymetric data from the ETOPO1 dataset was employed in the model (Amante and Eakins 2009).We used atmospheric forcing data from the ERA-5 product by the European Centre for Medium-Range Weather Forecasts at a spatial resolution of 0.25 • .All hourly atmospheric variables were set to their average during the summer months (June, July, and August) from 1988 to 2022, with wind forcing deactivated.Four major semidiurnal tides (M 2 , S 2 , K 2 , N 2 ) from the FES2014 tidal product were implemented at the open boundary on the eastern side, excluding sub-tidal components.The initial temperature and salinity were set to 12 • C and 32 PSU, respectively, representing typical wintertime conditions (Zhou et al 2017).This idealized experiment lasts for 30 d, and the outputs at the 15th day are used for analysis.The objective of it is to reproduce the tide-induced upwelling triggered by semi-diurnal tides during summer in the BHS.
Furthermore, we introduced a sinking particle (hereafter pseudo SPM) to mimic the behaviors of SPM, using the generic tracer module of SCHISM.The sinking velocity is set to a typical value of 0.001 m s −1 (Nowald et al 2009).Initially, all particles are uniformly distributed throughout the model domain with a concentration of 10 units, without additional sources or sinks thereafter.

Ecosystem model
To investigate the ecological responses to different SPM conditions, we established a vertical onedimensional (1D) ecosystem model based on the carbon, silicon, nitrogen ecosystem module in SCHISM (Chai et al 2002, Liu et al 2018, Wang et al 2020).This adapted model considers 11 state variables, with two phytoplankton species (small phytoplankton, S1; diatom, S2), two zooplankton species (microplankton, Z1; mesoplankton, Z2), four inorganic nutrients (nitrate, NO 3 ; phosphate, PO 4 ; ammonium, NH 4 ; silicate, SiO 4 ), two kinds of detritus (detritus nitrogen, DN; detritus silicon, DSi), and dissolved oxygen (DOX).The model parameters are mostly determined based on previous studies (Zhang et al 2022a).Initial nutrient conditions are specified as typical values referring to previous observations (Feng et al 2007).The complete model description and configuration are given in the supporting information.

Environmental conditions in the QCW
The satellite observations reveal a warm water belt (∼23 • C) with high transparency (SPM < 2 mg l −1 ) in the QCW during summer (figures 2(a) and (b)).The warm water belt extends southwards from the Qinhuangdao to the Laizhou Bay, while the eastern side lies a notable cold water affected by the Liaodong upwelling (Zhang et al 2022b).Such environmental conditions undoubtedly favor the proliferation of phytoplankton in the QCW owing to high light availability and moderate temperature.In the coastal bays, water temperatures are much higher, especially in the bays on the southern side, where the water depth is shallower.In the Bohai and Laizhou Bay, the water is highly turbid (SPM > 10 mg l −1 ), partially owing to the sediment discharge of the Yellow River.Notably, SPM exhibits a reversed pattern in relation to water depth, with low/high SPM typically observed in deep/shallow waters.However, the QCW and the top of Laizhou Bay deviate from this pattern, as they exhibit high SPM concentrations despite the relatively shallow water depth.
The idealized model successfully reproduces the warm water with low concentrations of SPM in the QCW (figures 2(c) and (d)), despite the omission of spatio-temporal variability in heat fluxes.The cooling rate of water is evidently slower in the QCW and the North Shelf of the Shandong Peninsula (NSSP).The spatial pattern of pseudo SPM closely resembles the satellite-observed SPM, exhibiting higher concentrations in coastal bays and over the central shoal.These findings indicate the existence of a depth-dependent upwelling system throughout the entire Bohai basin during the summer, and the intensity of upwelling varies significantly within the basin.

Tidal modulation on the summertime SST and SPM
The average of daily maximum velocity in the first 30 d is used to represent the amplitude of tide currents (figure 3(a)).Two low-velocity areas appear at the QCW and NSSP.This pattern reflects the distribution of tidal current amphidromic points (CAPs) of semi-diurnal tides in the BHS (Bian et al 2016).The spatial patterns of CAPs and elevation amphidromic points (EAPs) are commonly coupled and regulated by multiple factors including basin geometry, bottom friction, and water depth (Taylor 1922, Hendershott and Speranza 1971, Davies and Jones 1995, Xia et al 1995).For gulfs in the Northern Hemisphere, the CAP and EAP typically appear at the left coast (facing the top of the gulf) due to the bottom friction (Fang and Wang 1966).This explains the emergence of CAP the Qinhuangdao coast.Evidently, CAPs correspond well with high-SST and low-SPM regions, where tideinduced upwelling is clearly weak due to weak tide currents (figure 3(d)).Figure S2 reveals that the transect temperature exhibits typical structures of tidal mixing fronts (Simpson and Sharples 2012).It should be noted that, the relatively mild slope of QCW may also contribute to the formation of weak local upwelling.
These findings demonstrate that semi-diurnal tide currents exert a robust modulation on the summertime SST and SPM in the BHS through a basinwide tide-induced upwelling system.In contrast, the amplitude of diurnal tide currents is considerably weak (Bao et al 2001), with minor contributions.Figure S3 shows the satellite-observed SST during the spring/neap tide and their differences, revealing that the summertime SST shows notable spring-neap tidal variability in the BHS.Besides, the regions with weak tidal currents exhibit smaller SST changes between the spring tide and neap tide periods.These results further suggest the crucial role of tide currents in the BHS.

Response of the marine ecosystem to different SPM conditions
This section examines the response of marine ecosystem to different SPM conditions (2 mg l −1 vs. 10.5 mg l −1 ) using a 1D ecosystem model.The value of 2 mg l −1 represents the SPM level in the QCW, while 10.5 mg l −1 is the average SPM level in the BHS during summer.Although the high summer SST in the QCW may potentially affect local ecosystem in many aspects, it is challenging to quantify the effect related to temperature considering that different phytoplankton species have varying optimal growth temperatures or biological responses (Huertas et al 2011, Edwards et al 2016).
As expected, low-SPM conditions promote the rapid growth of phytoplankton in the euphotic layer (figure 4(a)).This process is accompanied by intense nutrient depletion, resulting in the oligotrophic zone in the upper layer.Simultaneously, the decayed phytoplankton contributes a substantial amount of particulate detritus, some of which remineralize during settling while another portion undergoes decomposition in the sediment layer.This leads to the oxygen depletion and nutrient enrichment in the lower layer.The observations of Wang et al (2009) indicate that the regions with the highest concentrations of phosphate and silicate in the bottom layer of the BHS during summer are just found offshore of Qinhuangdao.
Notably, phytoplankton populations build up to the highest concentrations in the subsurface layer, especially for small phytoplankton (S1).This feature is common to oceanic HAB events in the world, and the major population is small phytoplankton such as dinoflagellates (Gentien et al 2005).Correspondingly, a oxygen-rich zone is evident in the subsurface layer.In-situ observations also corroborated that during summer, the subsurface layer off the Qinhuangdao coast exhibits higher concentrations of chlorophyll-a and DOX (Zhao et al 2020).In contrast, in the other regions of the BHS (figure 4(b)), the euphotic layer is too shallow to accommodate sufficient phytoplankton, and thus, the oxygen depletion at the bottom is significantly alleviated.
Here we implemented the 1D ecosystem model in the whole BHS, considering only variations in water depth and SPM concentration.As shown in figure

Discussions
Numerical experiment reveals a basin-wide tideinduced upwelling system in the BHS during summer, which can greatly affect the spatial pattern of SPM (figure 2).As we have noted, the SPM has a reversed pattern of the water depth, but also affected by the amplitude of tidal currents.Thus, it is feasible to establish a novel index to predict the spatial pattern of SPM following the idea of Simpson-Hunter index (Simpson and Sharples 2012), which is widely used to determine the locations of tide mixing fronts.This novel index is in the form of U m /H n , where U is the semi-diurnal tide velocity and H is the water depth, m and n are positive integers.To some extent, it reflects the competition of local bathymetry and tidal mixing in affecting the surface footprints of tidal-induced upwelling.Here we take an example with m = 1 and n = 2.It should be noted that while other values may achieve similar effects, the essence remains the same.
This novel index highly resembles the observed SPM in the China Seas in terms of the spatial pattern (figure 5).Many of the major features such as the low-SPM feature off the Qinhuangdao coast, are properly reproduced.It suggests that the summertime SPM pattern is largely determined by the water depth and the amplitude of semi-diurnal tide currents.As sediments from the Yellow River are not considered, the predicted result is lower at the Yellow River estuary.This indicates that the influence of the Yellow River on the spatial pattern of SPM should be localized.In the BHS where water depth is shallower, the modulation of semi-diurnal tides on the SPM pattern is more robust.The locations of CAPs correspond well to the low-SPM regions, including QCW, NSSP, and the top of Laizhou Bay.In contrast, the great depth is chiefly responsible for the large area of low-SPM waters south of the Shandong Peninsula.Interestingly, except for the top of Laizhou Bay, the other regions are renowned for recurrent ecological disasters.In September 2013, the hypoxia event occurred on NSSP triggered massive deaths of cultured sea cucumber (Liu et al 2014, Zhai et al 2021a, Wang et al 2022).Sun et al (2022) reported algal blooms and hypoxia on NSSP and attributed the bottom hypoxia to the decomposition of decayed phytoplankton.This process agrees well with our theoretical results.The southern waters of Shandong Peninsula are notorious for the outbreaks of the largest green tides in the world (Liu et al 2009(Liu et al , 2010)).Liu et al (2020) demonstrated that Ulva prolifera (the causative species of green tides) has remarkable light utilization efficiency, implying that the high light availability greatly promotes the green tide proliferation.Although satellite observations also reveal the high frequency of algal blooms at the top of Laizhou bay (figure 1(c); Song et al 2021), the case may be more complex here, as the sediment inputs from the Yellow River, and the shallow water depth can greatly complicate the variations of SPM in this region.
This study emphasizes that different areas may display substantial spatial disparities in their vulnerability to ecological disasters.These spatial differences are significantly influenced by tidal dynamic processes, especially in shallow waters.In essence, the coastal waters like QCW are more prone to ecological disasters, even under the same nutrient condition.Note that the occurrence of HABs or hypoxia in realistic scenarios should be the joint result of light regulation and nutrient inputs.For example, on a shorter time scale, a significant influx of nutrients from a single flood event may also elevate the risk of local ecological disasters.In terms of environmental management, our research finding suggests the need to establish region-specific thresholds for water quality standards to more effectively address ecological issues.Particularly, in high-vulnerability regions like QCW and NSSP, increased focus should be placed on water quality monitoring and supervision to mitigate potential ecological concerns.In this regard, this study may hold substantial implications for future coastal environmental management.
More importantly, traditional notion holds that the strong-upwelling zones tend to be hotspots of ecological disasters due to ample nutrient supply within the euphotic layer (McCabe et al 2016, Ryan et al 2017, Franco et al 2023), while the present study offers a novel insight that weak local upwelling may greatly elevate the risks of ecological disasters in shallow waters during the warm season.This finding may have broader global applicability, given that tideinduced upwelling is a common occurrence in coastal waters globally.Meanwhile, it should be noted that the present study is largely based on simplified models and does not fully account for factors such as atmospheric forcing, circulations, and river inputs.Despite the model results effectively replicating the primary observed features, it remains essential to approach the conclusions with caution.In the future, conducting more intricate modeling studies will still be necessary to comprehensively dissect the contributions of various physical processes.

Conclusions
The occurrence of coastal HABs and hypoxia is often attributed to eutrophication in nearshore areas.However, in some low-nutrient regions, such ecological disasters may also occur repeatedly.This study investigated such incongruous site-specific ecological disasters from a hydrodynamic perspective based on a case study along the Qinhuangdao coast, by employing numerical models and satellite observations.The research demonstrates that the unique hydrological feature of the QCW creates favorable environmental conditions (warm water with high transparency) for the HABs and hypoxia during the warm season.Specifically, QCW is precisely located at the CAP of semi-diurnal tides, characterized by weak tidal currents.Therefore, the tide-induced upwelling is exceptionally weak in this region at that time, which is largely responsible for the high-SST and low-SPM feature observed at the Qinhuangdao coast.
In general, our study reveals that weak local upwelling may elevate the risks of HABs and hypoxia in the shallow waters during the warm season.This finding challenges the traditional view that the strong-upwelling regions are prone to be hotspots of ecological disasters due to ample nutrient supply into the euphotic layer from the bottom layer.Given the widespread occurrence of tide-induced upwelling in coastal waters around the world, this discovery may have broader global relevance.Furthermore, it underscores the pivotal role of tidal dynamics in influencing ecological disasters in shallow waters.Tidal dynamics can significantly influence the susceptibility of coastal areas to ecological challenges.In the case of high-vulnerability regions, greater attention should be devoted to environmental management to mitigate potential ecological issues.

W
Figure 1.(a) Area map and bathymetry in the BHS, and the dashed black circle marks the coastal water of Qinhuangdao.(b) Triangular mesh used for hydrodynamic modeling.(c) Satellite-based frequency of algal blooms in the China Seas (He et al 2013).

Figure 2 .
Figure 2. (a) Multi-year averaged summer SST from the CMEMS product (2007-2022), (b) the same as (a) but for the SPM (1998-2022).Summer is defined as the months from June to August.(c) The SST on the 15th day simulated from the idealized experiment (averaged within the upper 5 m), (d) the same as (c) but for the pseudo SPM.

W
Wu et al 3. (a) Amplitude of semi-diurnal tide currents (m s −1 ) in the model, (b)-(e) transect diagrams of vertical velocity (m s −1 ) on the 15th day.The transect locations are marked on the subplot (a) with dashed blue lines.

Figure 5 .
Figure 5. (a) Amplitude of M2 tide velocity (U) from the FES2014 tidal product, (b) bathymetry (H) from the ETOPO1 dataset, (c) climatology SPM in summer from the CMEMS product (1998-2022), (d) ratio of M2 tide velocity to squared water depth (U/H 2 ).Green star denotes the location of Qinhuangdao, while the red circles represent the coastal areas with low-SPM concentration in North China.