Winter bloom initiation with water column stabilization and improvement of light environment in a turbid shallow coastal water

Abstract

We conducted continuous mooring observation from autumn to winter of fiscal year 2020 to elucidate the mechanism of red tide development in the inner Ariake Sea, a very turbid shallow coastal water in Japan. The red tide dominated by Skeletonema spp. (mainly Skeletonema dohnii) developed at first neap tide after the annual minimum water temperature. Red tides at similar times of the year have been frequently observed here. Formation of two physical environments favorable for phytoplankton proliferation played a trigger role. One was stabilization of water column due to net heat flux transition through the sea surface from cooling to heating in mid-winter. Another was deepening of euphotic layer up to or exceeding water depth at the neap tide. Since the inner Ariake Sea has the small heat capacity due to its shallowness, the air and water temperature fluctuated almost in tandem, and reached their respective lowest values with a short time lag. The sea-surface heat flux, a main factor governing water temperature fluctuations, was dominated by latent heat and showed the highest correlation with the difference between atmospheric and sea-surface specific humidity. After mid-January, the atmosphere stabilized as the air temperature exceeded the water temperature, and the sea-surface cooling due to the latent heat weakened. With the heat flux change from negative to positive, the water column was stabilized. Then, winter bloom occurred during the neap tide when the compensation depth became deep with the decrease in suspended sediment concentration.

Appendices

Appendix 1: Calculation method for sea-surface heat flux

Calculation method for sea-surface heat flux is based on the coupled ocean–atmosphere response experiment version 3.0 (Fairall et al. 2003). Sea-surface heat flux is defined as follows:

$$Q=\mathrm{SW}+\mathrm{LW}+\mathrm{SE}+\mathrm{LA},$$

(4)

where Q is net heat flux, SW is short-wave radiation, LW is long-wave radiation, SE is sensible heat and LA is latent heat. The equation for LA is the same as Eq. 3 in the main text. The SW, LW, and SE are defined as follows:

$$\mathrm{SW}=0.94\cdot \mathrm{sorad},$$

(5)

$$\mathrm{LW}=s\sigma {\left(\mathrm{SST}+273.15\right)}^{4}\cdot \left(0.39-0.05\cdot \sqrt{\mathrm{vapp}}\right)\cdot \left(1.0-\mathrm{delta}\cdot \mathrm{cld}\right)+4.0s\sigma {\left(\mathrm{SST}+273.15\right)}^{3}\cdot \left(\mathrm{SST}-\mathrm{Tair}\right),$$

(6)

$$\mathrm{SE}={\rho }_{\mathrm{a}}{C}_{\mathrm{P}}{C}_{\mathrm{H}}V\left(\mathrm{SST}-\mathrm{Tair}\right),$$

(7)

where sorad is solar radiation (W m⁻²), s ratio of the radiation of the sea surface to a black body (= 0.97), $\sigma$ Stefan–Boltzmann coefficient (= $5.6705\times {10}^{-8}$ W m⁻² K⁻⁴), SST sea surface water temperature (°C), vapp vapor pressure (hPa), delta value calculated by $\mathrm{delta}=0.00427\cdot \mathrm{xlat}+0.5036,$ where xlat is latitude of location where the heat flux is calculated, cld amount of cloud, $\mathrm{Tair}$ air temperature (°C), ${\rho }_{\mathrm{a}}$ air density (kg m⁻³), ${C}_{\mathrm{P}}$ specific heat (J kg⁻¹ K⁻¹), ${C}_{\mathrm{H}}$ bulk coefficient for LW, V wind strength (m s⁻¹).

Appendix 2: Estimation of PAR extinction coefficient (${K}_{d}$)

Field observation was conducted at eight stations near Sta. H (open circles in Fig. 1c) on April 26, 29, and May 6 and 10, 2021. Vertical profiles of turbidity (Turb_RINKO) and PAR intensity were measured with RINKO Profiler (JEF Advantech Co.) and DEFI2-L (JEF Advantech Co.), respectively. Turb_RINKO was converted to SS concentration. By fitting exponential approximation curb (Eq. 1) to the PAR vertical profiles, ${K}_{d}$ was obtained as a function of surface SS concentration (mean value from the sea surface to 1.0 m depth) as follows (Fig. 9):

$${K}_{d}=0.0484\cdot \mathrm{SS}+0.6501.$$

(8)

Fig. 9

Relationship between the SS concentration at the sea surface and the PAR extinction

Appendix 3: Evaluation of sidelobe effect of ADCP

Measurement errors in current by upward-looking ADCP due to acoustic doppler reflections near the sea surface (Sidelobe phenomenon) are well known. Its influence is generally remarkable within 10% of the water depth from the sea surface. To examine the sidelobe effects, the velocity data 1.0 m below the sea surface at Sta. H measured by ADCP was compared with the velocity data at the same depth and location by an electromagnetic current meter (Compact-EM by JFE Advantec Co.). Figure 10 showed the comparison result of the current speed in direction of principle axis of M2 tidal current ellipse (counterclockwise rotation by 117.75° from the east–west direction) at 1.0 m below the sea surface from May 5 to 10, 2021. The mainstream flow measured by ADCP was almost the same in magnitude and phase as that by Compact-EM. It can be said that the influence of side lobe was very small at this depth.

Appendix 4: Contribution of horizontal advection to water temperature variation

Here, we evaluated effect of heat flux due to horizontal advection on water temperature change. Since there was no data on horizontal gradient in water temperature needed for its evaluation, we evaluated the contribution of horizontal advection by investigating how well the sea-surface net heat flux (NHF) could reproduce the water temperature change. In particular, we focused on short-term water temperature fluctuations based on the fact that phytoplankton proliferation occurred on a short time scale less than a week. Temperature change ($\Delta T$) due to NHF was evaluated as follows.

$$\Delta T=\frac{\mathrm{NHF}}{{\rho }_{0}{c}_{p}H},$$

(9)

where ${\rho }_{0}$ is reference density (= 1020 kg m⁻³), ${c}_{p}$ heat capacity (= 4000 J kg⁻¹ deg⁻¹) and $H$ mixing layer depth (= water depth in this case), respectively. We compared the sea-surface water temperature observed at Sta. H (SST_Obs) with that estimated by only NHF using the observed SST (SST_Obs) as the initial value (SST_Est). Figure 11 showed the comparison of the SST_Est after 25 and 50 hours (SST_25hour and SST_50hour, respectively) with SST_Obs from December 10, 2020 to February 9, 2021. SST_25hour and SST_50hour agreed well with SST_Obs (MAE was 0.32° and 0.46°, respectively). The same result was obtained even if the estimation period was extended to 150 h (MAE = 0.73°), indicating that the heat flux through the sea surface was  a dominant factor in the water temperature change during the study period. In other words, this result indicates that the effect of heat flux due to the horizontal advection was relatively small.

Fig. 10

Time series of the observed current velocity by ADCP and Compact-EM in principle axis of M2 tidal current ellipse (counterclockwise rotation by 117.75° from the east–west direction) at 1 m below the sea surface at Sta. H from May 5 to 10 2021

Fig. 11

Comparison of the estimated SST by only NSF after 25 and 50 hours (SST_25hour and SST_50hour, respectively) with the observed SST (SST_Obs) from December 10, 2020 to February 9, 2021. The observed SST was used as an initial SST value for SST_25hour and SST_50hour

Fig. 12

Time series of a the water temperature at Sta. H, b air temperature at Sta. Shimabara (solid lines are raw data and thick dashed lines are curves approximated by sine function), and c an enlarged view of the approximated curves of the air and water temperature around their respective minimum value

Appendix 5: Temperature synchronization of water and air

Here, we evaluated the synchronism of sea water temperature and air temperature. Figure 12 showed the temporal variation of (a) water temperature 1.0 m below the sea surface at Sta. H and (b) air temperature at Sta. Shimabara (Fig. 1b). The thick dashed lines in the figures represent curves approximated by a sine function. To improve approximation accuracy of the sine curve, the  water and air temperature obtained in fiscal year 2021 were added for the analysis. Figure 12c showed an enlarged one of the approximated curves of the air and water temperature around their minimum values. It can be seen that the water temperature reached its lowest value almost at the same time as the air temperature.