1. Introduction
The ocean absorbs substantial carbon dioxide (CO2) from natural and anthropogenic emissions (Canadell et al. 2007). One report stated that it absorbs approximately 25% of anthropogenic CO2 emissions (Wieder et al. 2021). Water bodies cover more than 70.8% of Earth’s surface. It is a complex ecosystem containing organisms that affect the release and uptake of CO2 (Colossi Brustolin et al. 2019), making them a barrier to climate change (Wang et al. 2021). The ocean absorbs ∼30% of CO2 from fossil fuel emissions, 30% from the land, and the remainder stays in the atmosphere (Friedlingstein et al. 2022). Global remote sensing data and models estimated that the net primary production of the ocean is between 35 and –70 Pg of carbon (C) per annum (Carr et al. 2006). Thus, the water body is a significant carbon sink, influencing the global carbon cycle through the air–ocean CO2 exchange (Bates and Mathis 2009).
The tropical coastal ocean is a subset of the global ocean. It is rich in biodiversity and is linked to different biomes, including estuaries, rivers, tidal wetlands, and continental shelves (Ward et al. 2020; Yusup et al. 2018b). Numerous studies suggest that continental shelves and coastal areas capture CO2 and export the carbon to the open ocean (Chen 2004; Kahl et al. 2017). Upwelling processes, high sedimentation rates, and carbon-rich river water runoff make the coastal ocean a net carbon sink (Ye et al. 2021). Although the coastal ocean is 0.5% of the total ocean volume, it contributes > 30% of the entire air–ocean CO2 exchange. These exchange processes account for > 80% of the oceanic organic matter burial (Chen and Borges 2009).
The ability of the coastal ocean to act as a CO2 sink mainly depends on salinity, solubility, temperature, and the gas transfer velocity (Yang et al. 2021) through the wind-induced turbulent mixing in the atmospheric surface layer (ASL) (Mari 2008; Riebesell et al. 2007; Smetacek 1999). Recent reports stated that 30% of anthropogenic CO2 is absorbed by the sea (de Carvalho-Borges et al. 2018). The exchange occurs on weekly to monthly time scales (Friederich et al. 2008). The process also depends on phytoplankton biomass. Jyothibabu et al. (2015) found that the highest chlorophyll a concentration is in the shallow ocean rather than the deeper ocean. For example, the concentration of chlorophyll was higher on the coasts of Sri Lanka and were subject to the eddies of the Summer Monsoon Current. The study also showed that phytoplankton were prevalent in the southern Bay of Bengal (Jyothibabu et al. 2015). Through photosynthesis, phytoplankton fix the dissolved inorganic carbon into oxygen and organic nutrients in its chlorophyll a (Birch et al. 2021), which is transferred to the seafloor through a combination of gravitational settling and advective and diffusive transport. Therefore, the carbon cycle in the tropical coastal ocean is critical to the global carbon budget.
Some studies have reported the open-ocean CO2 flux of the Southeast Asia region (Table 1). However, the net flux at the tropical coast is underrepresented or unknown. In decadal studies, literature shows that the tropics should be a net-positive CO2 flux area because of the warmer water temperature. However, the flux might also depend on other subdecadal time scale factors, such as phytoplankton mass. The tropical coast is known to host large concentrations of chlorophyll a–containing organisms (Yusup et al. 2018a). This population suggests competing factors in the flux that could be dependent on the season and modulate the decadal pattern.
Literature values of the net CO2 flux in the South China Sea and Southeast Asia oceans. The Chukchi Sea is included for comparison. The primary method used to estimate the flux is the bulk transfer method using the sea–air pCO2 difference.
Carbon dioxide flux has been measured globally and multiple studies have detailed its variability and influencing factors (Fay et al. 2021). The researchers used different approaches, such as the eddy covariance method, water–air sampling, and remote sensing approaches (Harmon 2020). A study concluded that the eddy covariance method is reliable in quantifying air–sea exchanges of methane (CH4) at a coast (Gutiérrez-Loza et al. 2019). Deemer et al. (2016) described the application of micrometeorological methods to monitor CO2 fluxes, which can address many monitoring challenges by providing pseudocontinuous, long-term, spatially integrated flux measurements. Other studies reported that using the spatially integrated eddy covariance or acoustic monitoring resulted in higher flux measurements than did methods with less spatial and temporal coverages (e.g., floating chambers, thin boundary layer, inverted funnels). This is consistent with the low sampling coverage in stochastic systems, leading to flux underestimation (Wik et al. 2016). Furthermore, as compared with shipboard eddy covariance observations, measuring fluxes from a stationary tower does not require motion correction on the winds. This means that flux and measurements at a coastal location are potentially more accurate, especially at high wind speeds when the correction for a moving platform would become prominent.
Only a handful of studies report on the tropical coastal ocean CO2 flux; therefore, the exchange processes and their regulating factors are not well understood. The biological and physical factors determine whether the coast is a source or a sink. The monsoon seasons could also influence the elements that modulate the atmospheric and oceanic conditions. Hence, we aim to describe the variation of the CO2 flux measured over 5 years in the tropical coastal ocean using the eddy covariance and remote sensing methods and characterize CO2 flux variability on monthly and annual scales. An additional objective is to determine the underlying mechanisms regulating the CO2 flux. These findings could help us to understand the processes that affect the flux and the atmospheric and oceanic elements of the carbon cycle in the tropical coastal ocean. The new understanding could contribute to developing models that predict carbon fluxes in these areas.
2. Materials and methods
a. Site description
This study was carried out in the Teluk (bay) Aling, adjacent to the Centre for Marine and Coastal Studies (CEMACS), Universiti Sains Malaysia (USM), and on the island of Pulau Pinang. The coordinates of the site are 5°26′53″N and 100°11′36″E. It is approximately 500 km northwest of Kuala Lumpur. The site is directly connected to the Straits of Malacca, a critical sea trade route of high economic importance.
Besides logistical activities, the strait is also utilized for tourism and aquaculture activities; Malaysia’s leading economic sector is tourism. The sector contributes significantly to the country’s economic growth (Hanafiah et al. 2021; World Tourism Organization and Global Tourism Economy Research Centre 2019). In addition, Southeast Asia is the world’s second-most productive aquaculture region (Food and Agriculture Organization 2018); Indonesia, Vietnam, Thailand, and the Philippines are the other leading producers of marine products.
Malaysian aquaculture production, including in the state of Pulau Pinang, is dominated by inland and mangrove pond cultures, approximately 218 000 t, of which marine culture accounts for 55 506 t (Samah 2020). Malaysia is a maritime nation with a fishing industry that employs 134 000 people. Coastal fisheries contributed 65% of the total catch in 2012 as compared with 35% from deep-sea fishing (Yusoff 2015).
The study area sits on a continental shelf and represents a typical southern South China Sea tropical coastal environment. Figure 1 shows the study location. The coastal ocean surface is calm, and the waves are low, at about 0.25 m in height, due to calm winds that are, on average, 0.5 m s−1, originating from all four primary wind directions. The bathymetry is shallow (the mean water depth is 4 m), and the seabed slowly declines at a rate of 4 m km−1. The coast’s closest primary terrestrial carbon source stems from two large rivers: the Muda and Merbok Rivers. The average water discharge of these rivers is 100 m3 s−1 (Fatema et al. 2014). However, these rivers are 20 and 28 km northwest of the study site. The height of the water at this location is 1.5 m, and the underlying surface is sand.
Muka Head Station’s location is in the tropical coastal ocean of the Straits of Malacca (5°26′53″N, 100°11′36″E). The station is equipped with eddy covariance and Biomet systems.
Citation: Earth Interactions 27, 1; 10.1175/EI-D-22-0023.1
The climate is hot and humid throughout the year, with atmospheric temperatures ranging from 26° to 29°C and relative humidity ranging from 75% to 94%. The site receives a maximum global solar radiation of 1021 W m−2. The area is within the intertropical convergence zone (ITCZ); hence, it receives heavy precipitation: its monthly cumulative precipitation is between 6 mm in January and 448 mm in May. The Southeast Asia mesoscale cyclones regulate the monsoons that control the regional weather and consist of the northeast monsoon (NEM), the spring transitional monsoon (STM), the southwest monsoon (SWM), and the fall transitional monsoon (FTM) (Chenoli et al. 2018; Lim 2004; Sirinanda 1979). In the primary 4-month monsoons (December–March and June–September), the northeast monsoon and the southwest monsoon manifest stronger winds but low rainfall. However, the shorter, in-between, 2-month monsoons (April–May and October–November), the spring and fall transitional monsoons, are characterized by high precipitation and low winds (Yusup et al. 2018a, 2016, 2018b).
The monsoons regulate the oceanic circulation pattern of the area. The main circulation pattern in the northeast monsoon is the upwelling process, which is associated with positive wind stress curl. The opposite trend occurs in the southwest monsoon, which exhibits the downwelling process along the northwestward direction of the sea on the west coast of Malaysia (Tan 2005; Mandal et al. 2021; Gayathri et al. 2022; Tan et al. 2006).
The exchange of CO2 between the atmosphere and ocean and organic carbon transport into the deep sea is regulated by phytoplankton and zooplankton. Zooplankton serves as trophic bridges between primary producers and higher trophic levels and as recyclers of particulate carbon and nutrients into dissolved pools (Mitra et al. 2014; Steinberg and Landry 2017). Maznah et al. (2021) found that water quality parameters influence the number and variety of zooplankton in the coastal waters, and the community was impacted by water quality and sediment variables.
Phytoplankton are primary producers that add energy to the food chain. Phytoplankton composition, abundance, and biomass can be used to assess the ecosystem’s health (Zanuri et al. 2020). Reported phytoplankton, or blue-green algae, and dinoflagellates found in water include Skeletonema sp. 1, Skelotonema sp. 2, Gyrodinium sp. 1, Prorocentrum sp. 1, Protoperidinium sp. 1, Protoperidinium sp. 2, Karenia sp., and Thalassiosira spp. (e.g., T. nanolineata, T. densannula, T. gravida) (Javeed et al. 2018; Zanuri et al. 2020). They affect turbidity and the concentration of dissolved oxygen in the water.
Zooplankton are a secondary producer in the aquatic ecology. Zooplankton community traits can be employed as bioindicators of aquatic ecosystem health and play an essential part in the marine food chain (Steinberg and Landry 2017). The zooplankton in marine analogs are larger than the freshwater species, because the former comprises significant animal phyla (Davies et al. 2009; Maznah et al. 2021). Zooplankton are sensitive to nutrient levels, temperature, food availability, pollution, light intensities, predation, pH levels, and heavy metals. There are 49 taxa of zooplankton species in the area, including Arthropoda, Chordata, Coelenterata, Protochordata, Mollusca, Polychaeta, Hydrozoa, and Echinodermata (Wan Maznah et al. 2014).
b. Environmental measurements
Environmental parameters were measured using the “Biomet” system, a collection of meteorological sensors measuring weather parameters related to biological processes. Atmospheric temperature and relative humidity were measured using HMP155 (Vaisala, Inc.). The sensor was radiation shielded. The sensors’ accuracy is ±0.7°C and ±1.7%, respectively.
Global radiation and net radiometer were measured using a pyranometer (LI-200SL from Li-Cor, Inc.; error is <5 %) and a net radiometer (NR LITE 2 from Kipp and Zonen, Inc.; sensitivity is 13.6 μV W−1 m−2). Photosynthetic photon flux density (PPFD) was recorded using a quantum sensor (Li-Cor LI-190R); the unit of measurement is micromoles of photons per meter squared per second. The sensors were installed on an extended rod.
Water temperatures were measured at 0.5 m using a Li-Cor thermistor with ±1°C accuracy. A tipping-bucket rain gauge (TE525; Texas Instruments, Inc.) was used to collect 30-min cumulative precipitation in an open area. All sensors were factory calibrated before installation. The 1-min data were recorded using a datalogger (9210b Xlite; Sutron Corporation) and averaged every 30 min. A stable, alternating current electrical supply powered the station’s datalogger and sensors. The raw data were merged and transmitted to a data server. Other setup details are documented in previously published literature (Yusup et al. 2018a, 2019, 2020, 2018b).
c. Flux footprint
Approximately 50% of the station’s surface is marine, due to the station being situated next to the shore. Thus, fluxes from land surfaces need to be omitted (Morin et al. 2017). The two-dimensional footprint method was utilized to remove nonmarine fluxes (Kljun et al. 2015). The process determines the flux source for each observation. The technique is dependent on friction velocity and atmospheric stability.
The mean wind speed is 0.52 ± 0.56 m s−1. The wind speed range is 0.0015–4.478 m s−1. The boundary layer was mostly weakly unstable: z/L < −1: 4.58%; −1 < z/L < −0.1: 80.06%; −0.1 < z/L < 0.1: 8.80%; z/L > 0.1: 5.58%; and z/L > 1: 1.72%. We removed 14.1% of the CO2 flux because the ratio of friction velocity u* to wind speed U is >0.417, which is 150% of the average 0.278 ratio, following the procedure of van Dam et al. (2021) to discard potential land-based fluxes. The proportionality slope is 0.017. In addition, fluxes with a minimum value of 0.15 m s−1 were rejected to avoid insufficient turbulent mixing (Morin et al. 2017).
Ninety percent of fluxes originated from the marine surface 50 m southwest of the station (Fig. 2a). The diel cycle influenced the wind direction. Still, since the wind speeds are generally low, they were evenly dispersed around the station (Figs. 2b,c).
(a) The flux footprint around the station. The percentages denote the contribution to the flux; e.g., the 10% contour shows the location of the peak flux contributor at 90%. (b) The covariance spectra curve (purple) for z/L = −2.01, an unstable ASL, and spectra curve (green) for z/L = 0.08, a stable ASL. (c) The wind rose for u*.
Citation: Earth Interactions 27, 1; 10.1175/EI-D-22-0023.1
d. Vertical turbulent flux
The CO2 flux (μmol m−2 s−1) was measured using an eddy covariance system installed on a stable stainless-steel platform. Positive CO2 flux denotes emission, while negative flux means uptake. The unit was converted to millimoles per meter squared per day to facilitate flux comparison with the literature. The data were collected for 5 years, between November 2015 and November 2020. This platform sits at the end of a pier, so the system is directly over the coastal water. The eddy covariance system includes an open-path CO2/H2O gas analyzer (Li-Cor LI-7500A) and a 3D sonic anemometer (81000V; R. M. Young Co.). The gas analyzer measured CO2 and H2O mixing ratios, while the anemometer was used to measure three-dimensional wind velocities and wind direction. These instrument pairs were installed 4.1 m above the water surface. The gas analyzer and sonic anemometer were periodically calibrated: the instrument was sent for calibration every 2 years as recommended by the supplier. A datalogger (Li-Cor LI-7550 Analyzer Interface Unit) recorded the eddy covariance data at 20 Hz.
The software EddyPro (version 6; Li-Cor) was used to calculate the 30-min flux. Only 10% of gaps in the data were allowed per 30-min average. The mean vertical wind speed was set to zero using the double-rotation method (Wilczak et al. 2001). Before calculating the CO2 fluxes, the 20-Hz data were filtered by five quality control (QC) tests, including spike removal, amplitude resolution, dropouts, absolute limits, skewness, and kurtosis. The data were quality checked and flagged using QC flags following the procedures described in the literature (Foken et al. 2004; Vickers and Mahrt 1997). A QC flag of 0 denotes good-quality flux, whereas the flags of 1 and 2 mean intermediate and bad qualities, respectively. The data flagged as 2 were removed from the dataset. Intermediate-quality data (flag 1) were included in the analysis because the study’s goal is to determine concurrent trends among the flux and environmental factors, which the flag-1 data can address. Flag-0 data are commonly used for process-based experiments to develop parameterized equations. The procedure includes stationarity and turbulence development tests. The quality-control procedure implemented (Mauder et al. 2008) is commonly practiced in processing eddy covariance data (van Dam et al. 2021).
e. Measurement uncertainty of the CO2 flux
Since the eddy covariance method is a relatively new technique for quantifying oceanic CO2 flux, the associated errors were calculated to determine its measurement uncertainty. Three contributions to the measurement error were identified: uncertainties due to 1) linearity of the CO2 standard used for CO2 analyzer calibration, 2) the sensitivity of the sonic anemometer to wind direction, and 3) the sensitivity of the CO2 flux in different wind speed ranges. The CO2 calibration standard reports an accuracy of 0.2% for 100–3000 μmol mol−1. For the sonic anemometer, the standard deviation of the wind speed measured in multiple directions was 0.05 m s−1. The combination of the two uncertainties produced a standard uncertainty of ±0.26% at a coverage factor of 2, or 2 standard deviations. The uncertainty analysis was extended farther to include the sensitivity of the CO2 flux in different wind speed ranges, since the flux is susceptible to wind. The standard errors of fluxes were calculated for wind speed ranges of 0–1, 1–1.5, 1.5–2.0, 2.0–2.5, and >2.5 m s−1. The 0–1 m s−1 range has the most considerable uncertainty, at ±13%. Consolidation of all of the uncertainties gives a total standard uncertainty for the CO2 flux measurement of ±27% at a coverage factor of 2.
The CO2 flux was verified with the Surface Ocean CO2 Atlas (SOCAT) dataset (https://www.socat.info/index.php/data-access/). The nearest measurement pass by a scientific ship that collected the surface water partial pressure, or pCO2, data was on 27 April 2016. By back-calculating pCO2 using the eddy covariance CO2 flux data and the bulk transfer method, the eddy covariance surface water partial pressure of CO2 was estimated. The bulk transfer method with the cubic wind-parameterized transfer-velocity equation by McGillis et al. (2001) was used. Only two SOCAT data points were in synchronicity and proximity with the station: 1) 7.989°N, 97.534°E and 2) 7.946°N, 97.602°E. The SOCAT pCO2 values were 418.7 and 419.8 μatm, respectively. SOCAT reported the value in fugacity, which was converted to partial pressure before further analysis. The eddy covariance–derived pCO2 values are 421.8 and 442.6 μatm, 0.7% and 5% higher than the SOCAT values. This difference is within the measurement uncertainty (±27%) calculated for the eddy covariance system, even though eddy covariance tends to overestimate surface flux. Overestimation could be due to the instrument’s sensitivity and the potential presence of other nonseawater surface flux due to horizontal advection. However, the long distance between the sites and the lack of synchronous data require more data and analysis to confirm the accuracy of the eddy covariance measurement in comparison with the traditional measurement technique.
SOCAT CO2 flux measurement analyzes CO2 in the air, which involves the utilization of an infrared analyzer instrument installed on board research vessels. According to Pfeil et al. (2013), the quality control of SOCAT CO2 fugacity or partial pressure data was conducted by identifying irrational and outlier values based on location, time, temperature, pressure, and CO2 surface water concentration. For instance, the time and date must be identified to the correct location, and the atmospheric pressure must be in the range of 800–1100 hPa. Subsequently, CO2 fugacity recalculation was carried out as the final step to verify the value. After recalculation, the uncertainty of CO2 fugacity was less than 10 μatm or 2.5% (Yasunaka et al. 2018).
f. Decoupling the pCO2 derived and from the sea surface temperature
Applying the method of (Takahashi et al. 2002), the dependence of CO2 flux on sea surface temperature was removed so that the biological influence on the flux is emphasized. As stated in the previous subsection, CO2 flux is related to the partial pressure in the bulk transfer equation. According to those authors, the CO2 concentration difference at the water surface is 0.0423 °C−1, while the site’s surface temperature range was 4.2°C. This small range is because the site is in the tropics, where the monthly temperature is unvarying. Thus, the change in the temperature was minimal. However, applying the Takahashi decoupling method somewhat removed the contribution of temperature on surface CO2 concentration. Correlation analysis shows that the dependence was small: Pearson’s r between pCO2 and temperature before decoupling was 0.315 (p < 0.05), whereas after decoupling it was −0.178 (p > 0.05). Aside from the small temperature range, another reason the temperature’s effect is small is the strong biological drawdown signal obscuring the temperature-based drawdown trend. Thus, the influence of seasonal surface water temperature change on CO2 flux was minimal because the flux is dependent on other, more varying and correlated environmental factors, such as wind and chlorophyll a.
The flux with QC flags of 0 and 1 were used for time series analysis, whereas only the flux of QC flag 0 was used for process-based empirical relationship development. The data contain 0.031%, 73.53%, and 22.34% of flags 0, 1, and 2, respectively. The method by Moncrieff et al. (2004, 1997) was used to correct low- and high-frequency spectra. The Webb, Pearman, and Leuning (WPL) correction accounted for density fluctuations (Webb et al. 1980). The covariance spectra of the flux were routinely checked to ensure they obey the similarity theory constraints (Fig. 2b).
g. MODIS remote sensing data
Satellite-based remote sensing data provide upscaled environmental parameter values (Balsamo et al. 2018). The ocean’s optical characteristics and biochemical composition, including sea surface temperature (SST), chlorophyll a concentration, and photosynthetically active radiation (PAR), were obtained from the Moderate Resolution Imaging Spectroradiometer Aqua (MODIS-Aqua) dataset. The chlorophyll a concentration was estimated by measuring the visible leaving radiance from the chlorophyll a pigments. In contrast, the PAR was assessed using the top-of-the-atmosphere radiance in the visible spectrum. SST was calculated using an imaging radiometer that operated in the visible and infrared sections of the electromagnetic spectrum. The level-3 standard mapped image (SMI) data were downloaded from NASA’s OceanColor web. The MODIS-Aqua is the main instrument on NASA’s Earth Observing System (EOS) satellite. The sensor scans our site between 1400 and 1500 local time (UTC + 8). The resolution is 4 km, and one reading was taken per day. The equidistant cylindrical projection is used, and the coordinate system is the World Geodetic System 48 (WGS 48).
The OC3 algorithm (O’Reilly and Werdell 2019) was used to estimate the chlorophyll a concentration. Colored dissolved organic material (CDOM) and total suspended matter (TSM) can have negative impacts on the reflectance spectrum of wavelengths and cause MODIS chlorophyll a images to be less accurate. The algorithm can produce overestimated chlorophyll a concentration due to CDOM and TSM, reducing the ratio of blue to green bands (Siswanto et al. 2011). According to Gai et al. (2020), CDOM and TSM are higher around the coast, causing the suppression of spectral reflection along with the decreased performance of the OC3 algorithm. However, the research by Agele et al. (2013) showed a significant negative correlation between TSM and chlorophyll a on the east coast of Peninsular Malaysia. Although the area under study could be susceptible to CDOM and TSM contamination, the purpose of the present analysis is to ascertain the trend and relationship between chlorophyll a and CO2 flux, not the accurate measurement of the chlorophyll a concentration.
Because of MODIS’s scanning swath limitation, the closest remote sensing data point (6°N, 100°E) to the CO2 flux footprints is 64 km. We assessed the discrepancy between the remotely sensed sea surface and the in situ temperatures. We found a general consistency between the two measurements: the RMSE is 1.31°C, while r = 0.69 and is statistically significant. We also assessed the RMSE between PPFD (in situ) and PAR (MODIS) and found that it is 16.71 μmol m−2 day−1 and the r = 0.50. The results indicate the regularity between the in situ and remote sensing measurements.
h. Data processing and analysis
The computer software EddyPro (version 6) was used to process eddy covariance data. The data were consolidated and analyzed using R statistical software (R Development Core Team 3.0.1. 2013). The relationships between the in situ measurement and remote sensing data were calculated using the root-mean-square error (RMSE) and Pearson’s correlation coefficient r. The correlation indicates the variance proportion explained by the linear dependence between two independent variables (Haëntjens et al. 2017; Świrgoń and Stramska 2015).
3. Results and discussion
The CO2 air–sea exchange equilibrium occurs on time scales from weeks to months to years. The CO2 flux
Monthly averaged ASL parameters from 2016 to 2020: (a)
Citation: Earth Interactions 27, 1; 10.1175/EI-D-22-0023.1
Monthly averaged values from 2016 to 2020 for (a) PAR, (b) chlorophyll a, (c) Ta, (d) RH, (e) Ts, and (f) SST from MODIS. The solid blue line is the monthly averaged trend, and the red dashed line is the annually averaged trend.
Citation: Earth Interactions 27, 1; 10.1175/EI-D-22-0023.1
The 5-yr average of the CO2 flux is −0.089 mmol m−2 day−1, which indicates that this tropical coast is a carbon sink. However, based on the Table 1, the tropical ocean mostly acted as a weak CO2 source. The different signs of flux movement can be due to the localized nature of processes involving organic matter release at the shoreline, open-ocean water mixing, biological productivity, and upwelling. According to Chen et al. (2013), organic carbon flowing into the ocean through rivers causes the ocean to act as a CO2 source. However, higher biological productivity along with upwelling and undersaturation of CO2 in the water due to open-ocean water mixing result in lower CO2 partial pressures, making it a sink of CO2. Additionally, intense bloom of phytoplankton stimulated by nutrients from deep waters can also result in CO2 being absorbed by the ocean (Laruelle et al. 2014). Overall, the comparison highlights the weak exchange of CO2 in the tropical coastal ocean. It was also mentioned by Laruelle et al. (2014) that the CO2 flux near the equator was close to zero. The weak CO2 flux is due to the low wind speeds in the region (Chen et al. 2013). Furthermore, the disparities in flux estimation can also be caused by the application of disparate measurement methods and the application of different gas transfer velocity, wind-dependent parameterized equations to calculate the flux.
a. The monsoons and
The equatorial coastal region is hot and humid throughout the year. However, the influence of the four monsoons (northeast and southwest monsoons and the spring and fall transitional monsoons) on
The 5-yr average of the CO2 flux is −0.089 ± 0.024 mmol m−2 day−1; the error represents the measurement uncertainty at the 95% confidence interval. This value indicates that the tropical coastal ocean is a CO2 sink. Because of the monsoons, the tropical coast changes from a source to a sink of CO2. During the northeast monsoon, the coast is a strong CO2 sink at −1.647 ± 0.445 mmol m−2 day−1. Its sink capability decreased in the fall transitional monsoon at −0.016 ± 0.004 mmol m−2 day−1. However, in the southwest monsoon and spring transitional monsoon, the coast is a weak source of CO2 at 0.899 ± 0.243 and 0.827 ± 0.223 mmol m−2 day−1, respectively. The source value is similar to the one reported by Takahashi et al. (2002) for the Indian Ocean. This changing source-sink surface is possibly due to the variation of U (Chen et al. 2013; Laruelle et al. 2014).
Throughout the sampling campaign, the ASL was strongly unstable (z/L = −1.5) with low u* (averaged at 0.041 m s−1). The higher water temperature than the atmospheric temperature generated an unstable atmosphere that typically exhibits low momentum transfer but high sensible heat transfer (Stull 1988). This condition could also promote gas exchange in the surface layer and a higher gas transfer velocity.
Carbon dioxide flux gradually increases from the beginning of the year and stabilizes in May, before decreasing in October. The increasing and declining phases feature extensive fluctuations in
The broadest
b. The monsoons, chlorophyll a, and environmental parameters
The chlorophyll a and PAR parameters represent the biological processes connected to
The correlation matrix of the CO2 flux and environmental parameters on the monthly and yearly time scales (three, two, and one asterisk indicate significance p < 0.0001, < 0.001, and < 0.01, respectively).
Chlorophyll a concentration (Figs. 4a,b) showed significant differences among the monsoons. The highest concentration of chlorophyll a is in the northeast monsoon, followed by the fall transitional monsoon, spring transitional monsoon, and southwest monsoon. The means are 1.355, 0.829, 0.556, and 0.490 mg m−3, respectively. It must be pointed out that the chlorophyll a concentration could be overestimated (Siswanto et al. 2011) or underestimated (Agele et al. 2013) due to remote sensing images being contaminated by dissolved organic material or total suspended matter; however, concentration value could not be verified because on-ground measurements were not performed. The maximum chlorophyll a season lagged PAR by one monsoon; the PAR was the highest in the spring transitional monsoon, followed by the northeast monsoon, southwest monsoon, and fall transitional monsoon. The averages are 50.59, 48.98, 43.54, and 43.44 μmol m−2 day−1, respectively. As mentioned, the variation in the PAR values could be due to the presence of clouds.
The CO2 flux is strongly, negatively correlated with chlorophyll a and weakly negatively correlated with PAR, although not statistically significantly so (Table 2). High chlorophyll a concentration suggests large phytoplankton density promoting the uptake of CO2. The large population also indicates an increase in nutrients in the ocean and a heightened PAR availability (Znachor and Nedoma 2010). Yet, the PAR was the highest in the spring transitional monsoon, the monsoon after the northeast monsoon, when chlorophyll a concentration and
Since higher water temperature might be related to higher CO2 partial pressure and higher CO2 flux, the relationship between temperature and chlorophyll a concentration was explored. Results showed that chlorophyll a concentrations decrease with increasing water temperatures (Table 2), while CO2 flux increases. The decreasing seawater temperature could be indicative of a general cooling of the shelf. This trend is similar to the one at the Patagonian shelf; the chlorophyll a concentration increased by 67.8% after the seawater temperature reduced to 0.78°C, due to the cooling of the shelf (Gregg et al. 2005). The results suggest that cooler temperatures promote the growth of the phytoplankton population, aside from increased upwelling processes caused by the dropping shelf temperature (Edwards et al. 2016). The optimum growth temperature range for two common species, Skeletonema sp. and Thalassiosira spp., is from 25° to 29°C (Tian et al. 2002; Cupp 1943; Tomas 1997). The average temperature was lower in NEM (29°C) than in SWM (32°C), suggesting that the growth rate was optimal in the NEM. Furthermore, CO2 will be less fixed in warm waters (Weiss 1974). Decoupling temperature’s effect on CO2 flux or partial pressure (section 2f) using the Takahashi method showed that the temperature impacted the CO2 flux or partial pressure. However, the temperature range was small, so the flux variation caused by water temperature was minimal. Thus, the temperature’s effect on CO2 flux might be masked by the impact of chlorophyll a.
During the northeast monsoon, the high precipitation could cause a high river influx from the estuaries to the ocean, but the events are sporadic (Yusup et al. 2018a). Precipitation increases the river discharge and surface runoff, contributing to more nitrogen and phosphate in the ocean (Prasanna Kumar et al. 2002). Some researchers found that the river discharge and precipitation contribute to high contents of organic material and nutrients in the sea (Bauer et al. 2013; Lorenz et al. 2018). This could increase chlorophyll a concentrations and primary production. There are two river systems near the site, the Muda and Merbok Rivers (Fatema et al. 2014), but the rivers are 24 km away. No data on the discharge rate and characteristics of the rivers are available for analysis.
Furthermore, more nutrients could be transported into the area because of upwelling caused by strong winds (Znachor and Nedoma 2010) and cooling shelf temperature (Gregg et al. 2005). Wind speed is strongly positively correlated with chlorophyll a concentration (Table 2). The wind of the northeast monsoon was intense and led to the upwelling process that brought nutrients from the seabed to the surface (Tan et al. 2006; Lachkar and Gruber 2013). The process sustains phytoplankton growth and intensifies chlorophyll a concentration. In the southwest monsoon, winds were weaker, and the upwelling process subsided, which coincided with lower chlorophyll a concentrations. Wind strength was moderately negatively correlated with CO2 flux, suggesting the stronger winds lead to increased atmospheric CO2 uptake.
The average in situ measurement for Ta and Ts, SST from MODIS, and the average relative humidity (RH) are presented in (Figs. 4c–f). The Ta and Ts are lowest at the beginning or middle of the northeast monsoon, and highest during the spring transitional monsoon or southwest monsoon (depending on the year). During the fall transitional monsoon and the end of the northeast monsoon, they are in a declining and ascending phase, respectively. The variation in Ta throughout the 5 years was 5°C as compared with Ts, which exhibits a lower variation of 4°C. SST’s seasonal trend followed Ts, albeit at lower variability. It has been reported that the water temperature of tropical seas and oceans typically ranges from 20° to 30°C and, in most cases, remains constant throughout the year (Sparks and Toumi 2020).
Relative humidity ranged from less than 70% to more than 95%, showing the highest variability among the other environmental factors. Unlike temperature, humidity is not associated with monsoons. However, relative humidity seems more variable in the northeast monsoon due to fewer clouds and high evaporation (Peterson et al. 2017). Thus, the humidity did not influence chlorophyll a concentration when compared with U or water temperature.
c.
The high, positive
Like on the monthly time scale, annual
On the yearly time scale, CO2 flux was significantly influenced by temperature (Table 2). The atmospheric temperature did not control the flux because the latter rose and fell from 2016 to 2020 without changing the CO2 flux trend. However, Ts and SST were steadily declining. The temperature effects on the CO2 flux could not be removed using the Takahashi method because it works only on monthly averaged data. Like chlorophyll a, water temperature did not spike with ENSO events, but it did rise at the end of the El Niño period of 2016 and La Niña of 2018. However, the trend could be due to seasonal effects and not ENSO. This result indicates water temperature regulates CO2 flux on the annual scale (Table 2) and suggests that physical processes overtake biological processes’ control on CO2 flux on annual time scales.
The years 2017 and 2016 recorded the most significant and second largest
Yearly averaged
d. Monthly and annually oscillations of
The continuous wavelet transform or wavelet analysis (Torrence and Compo 1998) was conducted to comprehensively identify CO2 uptake and outgassing cycles in the frequency domain. The mother wavelet Morlet was utilized for this analysis because it can detect peaks and valleys in the time series. Using this technique, we uncovered persistent periodicity in the dataset. The other analysis parameters set were four suboctaves per octave and lag 1 of 0.72, which is the default value for the autocorrelation of the red noise background. The 95% confidence spectrum exhibits that most of the peak power wavelet spectrum is statistically significant. In figures, the bright yellow shaded areas mean durations of intense fluctuation of CO2 flux, while the dark blue region implies muted CO2 flux variations.
The analysis of the monthly
Wavelet analysis results of 2016–20 monthly (a)
Citation: Earth Interactions 27, 1; 10.1175/EI-D-22-0023.1
It also displayed a critical 3-month (0.25 period) CO2 flux variability, consistent with the monsoons. In this period band, an increased variability period is followed by a decreased variability phase, supporting the increasing, stable, and declining phases discussed previously. However, the wavelet analysis indicates that the ENSO events could have influenced CO2 flux, but its effect was indirect. New trends were observed in the 1.5- and 4.5-month bands in 2018 (0.125 and 0.375 period) that replaced the 3-month band. The change was possibly due to the La Niña phase of the 2018 ENSO event. This condition suggests a disruption of the monsoonal cycle. At the end of 2018 and the start of 2019, the variability was intense within the 3- and 6-month bands (0.25–0.5 period). The intensity implies enhanced fluctuations of CO2 in this period. This trend was potentially due to the El Niño phase of the 2018 ENSO event. This shows that future ENSOs could result in the augmented release and carbon uptake in this tropical coastal region, which suggests rapid carbon changes in the water and disturbances of the seasonal cycle.
Overall, a cross-wavelet analysis of chlorophyll a also showed statistically significant similar trends of identical behavior with CO2: increased variability in the 1- and 0.25-period bands. Some deviations from seasonal trends are also observed for chlorophyll a, similar to CO2 flux in 2017 and 2018: in 2017, the 4.5-month (0.375 period) band is prominent, highlighting that the chlorophyll a fluctuations were large; however, in 2018, the changes shift to the 1.5-month band (0.125 period). Therefore, the 4.5-month band in 2017 is associated with the 3-month band of CO2 flux. Similarly, the 1.5-month chlorophyll a band in 2018 related to the 1.5-month flux. Note that, to a lesser extent, the 3-month CO2 flux bands can also be associated with the 1.5-month chlorophyll a band. The pattern suggests the small but quick chlorophyll a concentration perturbations that occurred in the same 2018 period that changed CO2 flux variations. This period coincided with the La Niña phase of the 2018 ENSO event. It is noteworthy that after 2019, the high chlorophyll a concentration did not trigger enhanced CO2 flux uptake, but the bands of both chlorophyll a and CO2 flux extended from the 3- to 6-month bands. This analysis supports that the monthly variability of chlorophyll a is associated with fluctuations in CO2 flux.
For PAR, the 12- and 6-month bands are prominent (1 and 0.5 periods, respectively). However, the 6-month variability subsided to 1.5 month in 2019 and onward. The 1.5-month band also appeared in 2018, indicating significant PAR changes between months. In mid-2017, the La Niña and El Niño events of the 2018 ENSO could have caused the 6-month band to give way to the 1.5-month band. This could be due to more clouds that increased PAR’s variability and reduced availability. Therefore, PAR’s fluctuations could be tied to CO2 flux through chlorophyll a.
The noteworthy band for seawater temperature is the 12-month (1 period) band, which refers to the annual cooling, and its associated perturbations, of the seawater from 2016 to 2020. The cooling was variable in 2017 but tapered off starting from 2018 to 2020; the temperature variability in the 3-month (0.25 period) band triggered the interyear cooling but became absent in the proceeding years.
The analysis highlights the strong annual and seasonal features of CO2 flux. The rise and fall of CO2 depend on periodic changes in chlorophyll a. Global-scale events, such as ENSO, increased the variability of chlorophyll a and PAR on the 1.5-month scale. The La Niña phase of the 2018 ENSO triggered CO2 flux changes to shift to the 1.5-month band and induced a CO2 outgassing scenario in 2018 (Table 3). This period is followed by El Niño of the 2018 ENSO, which further expanded the CO2 fluctuation from the 3-month to the 6-month band. This condition resulted in CO2 uptake in the early part of 2019. These ENSO cycles are embedded within the typical monsoonal 3- and 12-month cycles.
e. Environmental drivers of
Monthly and annual scales pivot the environmental controls on
On the monthly scale, chlorophyll a concentration shows substantial negative correlations with
Correlation plots of the CO2 flux and selected environmental parameters on the monthly time scale.
Citation: Earth Interactions 27, 1; 10.1175/EI-D-22-0023.1
Carbon dioxide flux switched from strongly, negatively correlated with U (r = −0.52; p < 0.001) on the monthly time scale (Fig. 6) to positively correlated with U (r = 0.43; p > 0.01) on the yearly time scale. Wind exerts itself on
Water temperature (SST and TS) regulated
Wind speed is also related to chlorophyll a, suggesting the presence of wind-driven upwelling contributions to
Chlorophyll a concentration was also inversely proportional to SST (r = −0.52; p < 0.0001) on the monthly time scale; it was weakly, inversely proportional to Ts. The relationship strengthens on the yearly time scale (r = −0.79; p > 0.01). Because of the unimodal nature of the growth curve of phytoplankton, the high SST will reduce the phytoplankton density and decrease chlorophyll a concentration and vice versa (Edwards et al. 2016). Also, the decreasing seawater temperature could enhance upwelling (and nutrient availability) due to the cooling of the shelf, increasing chlorophyll a concentration (Gregg et al. 2005). Furthermore, high SST will induce the release of CO2 into the atmosphere because high temperature reduces water CO2 solubility and causes the ocean to outgas CO2 (Karlin et al. 2015).
4. Conclusions
We described the variation of 5 years’ CO2 flux measurements using the eddy covariance and remote sensing methods. The low-resolution spatial scale of the remote sensing dataset can produce signals and trends that correlate with in situ measurements but on monthly and yearly time scales.
The 5-yr average of the CO2 flux is −0.089 ± 0.024 mmol m−2 day−1, which indicates that the tropical coastal ocean is a moderate carbon sink. The seasonal cycle controls three main phases of the flux, which are the increasing, stable, and decreasing stages. The rising and declining phase is characterized by the erratic nature of the average flux, which increases and decreases rapidly. The stable phase did not exhibit similar erraticism. The rising phase is in the spring transitional monsoon; the mean is 0.822 ± 0.222 mmol m−2 day−1. The stable phase is in the southwest monsoon (mean = 0.899 ± 0.242 mmol m−2 day−1), while the decreasing phase is in the fall transitional monsoon (mean = −0.016 ± 0.004 mmol m−2 day−1). The northeast monsoon (mean = −1.647 ± 0.445 mmol m−2 day−1) showed decreasing and increasing trends. The yearly flux trend is not as apparent as the monthly trend, but it shows that the CO2 flux decreases yearly.
The results indicate the cyclical nature of the flux was affected by strong biological and physical controls depending on the time scale, specifically chlorophyll a, U, SST, and Ts. Biological controls on CO2 flux are more prominent than physical controls on the monthly time scale and vice versa for the yearly time scale. Wavelet analysis shows that the CO2 flux variation is mainly regulated by chlorophyll a and indirectly by PAR, possibly due to changes associated with El Niño–Southern Oscillation. On the monthly time scale, sea surface temperature did not affect the fluxes as much as chlorophyll a, but temperature’s control on the flux became apparent on the yearly time scale.
These findings could help us to understand the processes that affect CO2 flux and the atmospheric component of the carbon cycle at the tropical coast and could contribute to the development of models to predict carbon fluxes in the area.
Acknowledgments.
We recognize the Ministry of Education Malaysia, which awarded us the Malaysian Research University Network Long-Term Research Grant Scheme (MRUN-LRGS; Grant 203.PTEKIND.6777006) that made this research possible. We thank our industry partner Elite Scientific Instruments Sdn. Bhd. for contributing the sensors to make our measurements.
Data availability statement.
The data used in this research can be retrieved online (https://atmosfera.usm.my).
REFERENCES
Agele, D., T. Lihan, A. R. Sahibin, and Z. Rahman, 2013: Application of remote sensing techniques for studying the spatial & temporal variability of the Pahang River plume, Malaysia. J. Basic Appl. Sci. Res., 3, 44–52.
Balsamo, G., and Coauthors, 2018: Satellite and in situ observations for advancing global Earth surface modelling: A review. Remote Sens., 10, 2038, https://doi.org/10.3390/rs10122038.
Bates, N. R., 2006: Air-sea CO2 fluxes and the continental shelf pump of carbon in the Chukchi Sea adjacent to the Arctic Ocean. J. Geophys. Res., 111, C10013, https://doi.org/10.1029/2005JC003083.
Bates, N. R., and J. T. Mathis, 2009: The Arctic Ocean marine carbon cycle: Evaluation of air-sea CO2 exchanges, ocean acidification impacts and potential feedbacks. Biogeosciences, 6, 2433–2459, https://doi.org/10.5194/bg-6-2433-2009.
Bauer, J. E., W.-J. Cai, P. A. Raymond, T. S. Bianchi, C. S. Hopkinson, and P. A. G. Regnier, 2013: The changing carbon cycle of the coastal ocean. Nature, 504, 61–70, https://doi.org/10.1038/nature12857.
Birch, H., D. N. Schmidt, H. K. Coxall, D. Kroon, and A. Ridgwell, 2021: Ecosystem function after the K/Pg extinction: Decoupling of marine carbon pump and diversity. Proc. Roy. Soc., 288B, 20210863, https://doi.org/10.1098/rspb.2021.0863.
Burrows, E. H., J. L. Bubier, A. Mosedale, G. W. Cobb, and P. M. Crill, 2005: Net ecosystem exchange of carbon dioxide in a temperate poor fen: A comparison of automated and manual chamber techniques. Biogeochemistry, 76, 21–45, https://doi.org/10.1007/s10533-004-6334-6.
Canadell, J. G., D. E. Pataki, R. Gifford, R. A. Houghton, Y. Luo, M. R. Raupach, P. Smith, and W. Steffen, 2007: Saturation of the terrestrial carbon sink. Terrestrial Ecosystems in a Changing World, J. G. Canadell, D. E. Pataki, and L. F. Pitelka, Eds., Global Change—The IGBP Series, Springer, 59–78.
Carr, M.-E., and Coauthors, 2006: A comparison of global estimates of marine primary production from ocean color. Deep-Sea Res. II, 53, 741–770, https://doi.org/10.1016/j.dsr2.2006.01.028.
Chen, C.-T. A., 2004: Exchanges of carbon in the coastal seas. The Global Carbon Cycle: Integrating Humans, Climate, and the Natural World, C. B. Field and M. R. Raupauch, Island Press, 341–351.
Chen, C.-T. A., and A. V. Borges, 2009: Reconciling opposing views on carbon cycling in the coastal ocean: Continental shelves as sinks and near-shore ecosystems as sources of atmospheric CO2. Deep-Sea Res. II, 56, 578–590, https://doi.org/10.1016/j.dsr2.2009.01.001.
Chen, C. T. A., T.-H. Huang, Y.-C. Chen, Y. Bai, X. He, and Y. Kang, 2013: Air–sea exchanges of CO2 in the world’s coastal seas. Biogeosciences, 10, 6509–6544, https://doi.org/10.5194/bg-10-6509-2013.
Chenoli, S. N., P. R. Jayakrishnan, A. A. Samah, O. S. Hai, M. Y. A. Mazuki, and C. H. Lim, 2018: Southwest monsoon onset dates over Malaysia and associated climatological characteristics. J. Atmos. Sol.-Terr. Phys., 179, 81–93, https://doi.org/10.1016/j.jastp.2018.06.017.
Chou, W.-C., G.-C. Gong, D. D. Sheu, C.-C. Hung, and T.-F. Tseng, 2009: Surface distributions of carbon chemistry parameters in the East China Sea in summer 2007. J. Geophys. Res., 114, C07026, https://doi.org/10.1029/2008JC005128.
Churilova, T. Y., V. V. Suslin, N. A. Moiseeva, and T. V. Efimova, 2020: Phytoplankton bloom and photosynthetically active radiation in coastal waters. J. Appl. Spectrosc., 86, 1084–1091, https://doi.org/10.1007/s10812-020-00944-0.
Colossi Brustolin, M., I. Nagelkerken, C. Moitinho Ferreira, S. Urs Goldenberg, H. Ullah, and G. Fonseca, 2019: Future ocean climate homogenizes communities across habitats through diversity loss and rise of generalist species. Global Change Biol., 25, 3539–3548, https://doi.org/10.1111/gcb.14745.
Cupp, E. E., 1943: Marine plankton diatoms of the west coast of North America. Bulletin of the Scripps Institution of Oceanography, Vol. 5, No. 1, 237 pp., https://escholarship.org/uc/item/922945w8.
Davies, O. A., J. F. N. Abowei, and B. B. Otene, 2009: Seasonal abundance and distribution of plankton of Minichinda stream, Niger Delta, Nigeria. Amer. J. Sci. Res., 2, 20–30.
de Carvalho-Borges, M., I. B. M. Orselli, M. L. de Carvalho Ferreira, and R. Kerr, 2018: Seawater acidification and anthropogenic carbon distribution on the continental shelf and slope of the western South Atlantic Ocean. J. Mar. Syst., 187, 62–81, https://doi.org/10.1016/j.jmarsys.2018.06.008.
Deemer, B. R., and Coauthors, 2016: Greenhouse gas emissions from reservoir water surfaces: A new global synthesis. BioScience, 66, 949–964, https://doi.org/10.1093/biosci/biw117.
Edwards, K. F., M. K. Thomas, C. A. Klausmeier, and E. Litchman, 2016: Phytoplankton growth and the interaction of light and temperature: A synthesis at the species and community level. Limnol. Oceanogr., 61, 1232–1244, https://doi.org/10.1002/lno.10282.
Ekayanti, N. W., and A. Rahman as-Syakur, 2011 : CO2 flux in Indonesian water determined by satellite data. Int. J. Remote Sens. Earth Sci., 8, 1–12, https://doi.org/10.30536/j.ijreses.2011.v8.a1736.
Fatema, K., W. W. Maznah, and M. M. Isa, 2014: Spatial and temporal variation of physico-chemical parameters in the Merbok Estuary, Kedah, Malaysia. Trop. Life Sci. Res., 25, 1–19.
Fay, A. R., and Coauthors, 2021: SeaFlux: Harmonization of air–sea CO2 fluxes from surface pCO2 data products using a standardized approach. Earth Syst. Sci. Data, 13, 4693–4710, https://doi.org/10.5194/essd-13-4693-2021.
Foken, T., M. Göockede, M. Mauder, L. Mahrt, B. Amiro, and W. Munger, 2004: Post-field data quality control. Handbook of Micrometeorology, X. Lee, W. Massman, and B. Law, Eds., Atmospheric and Oceanographic Sciences Library, Vol. 29, Springer, 181–208.
Food and Agriculture Organization, 2018: The State of World Fisheries and Aquaculture. Food and Agriculture Organization, 227 pp.
Friederich, G. E., J. Ledesma, O. Ulloa, and F. P. Chavez, 2008: Air–sea carbon dioxide fluxes in the coastal southeastern tropical Pacific. Prog. Oceanogr., 79, 156–166, https://doi.org/10.1016/j.pocean.2008.10.001.
Friedlingstein, P., and Coauthors, 2022: Global carbon budget 2021. Earth Syst. Sci. Data, 14, 1917–2005, https://doi.org/10.5194/essd-14-1917-2022.
Gai, Y., D. Yu, Y. Zhou, L. Yang, C. Chen, and J. Chen, 2020: An improved model for chlorophyll-a concentration retrieval in coastal waters based on UAV-Borne hyperspectral imagery: A case study in Qingdao, China. Water, 12, 2769, https://doi.org/10.3390/w12102769.
Gayathri, N. M., A. V. Sijinkumar, B. N. Nath, K. Sandeep, and K. Y. Wei, 2022: A ∼30 kyr sub-centennial to millennial Indian summer monsoon variability record from the southern Andaman Sea, northeastern Indian Ocean. Palaeogeogr. Palaeoclimatol. Palaeoecol., 590, 110865, https://doi.org/10.1016/j.palaeo.2022.110865.
Gregg, W. W., N. W. Casey, and C. R. McClain, 2005: Recent trends in global ocean chlorophyll. Geophys. Res. Lett., 32, L03606, https://doi.org/10.1029/2004GL021808.
Gutiérrez-Loza, L., M. B. Wallin, E. Sahlée, E. Nilsson, H. W. Bange, A. Kock, and A. Rutgersson, 2019: Measurement of air-sea methane fluxes in the Baltic Sea using the eddy covariance method. Front. Earth Sci., 7, 93, https://doi.org/10.3389/feart.2019.00093.
Haëntjens, N., E. Boss, and L. D. Talley, 2017: Revisiting ocean color algorithms for chlorophyll a and particulate organic carbon in the Southern Ocean using biogeochemical floats. J. Geophys. Res. Oceans, 122, 6583–6593, https://doi.org/10.1002/2017JC012844.
Hanafiah, M. H., M. R. Jamaluddin, and V. Kunjuraman, 2021: Qualitative assessment of stakeholders and visitors perceptions towards coastal tourism development at Teluk Kemang, Port Dickson, Malaysia. J. Outdoor Recreation Tourism, 35, 100389, https://doi.org/10.1016/j.jort.2021.100389.
Harmon, T. C., 2020: Carbon gas flux to and from inland waters: Support for a global observation network. Limnology, 21, 429–442, https://doi.org/10.1007/s10201-020-00623-1.
Javeed, A., S. Salleh, A. Darif, and M. Mohammad, 2018: Preliminary observation of tropical diatom Thalassiosira spp. from Teluk Bahang, Penang. Scr. Biol., 5, 7–11, https://doi.org/10.20884/1.sb.2018.5.1.668.
Jyothibabu, R., P. N. Vinayachandran, N. V. Madhu, R. S. Robin, C. Karnan, L. Jagadeesan, and A. Anjusha, 2015: Phytoplankton size structure in the southern Bay of Bengal modified by the Summer Monsoon Current and associated eddies: Implications on the vertical biogenic flux. J. Mar. Syst., 143, 98–119, https://doi.org/10.1016/j.jmarsys.2014.10.018.
Kahl, L. C., A. A. Bianchi, A. P. Osiroff, D. R. Pino, and A. R. Piola, 2017: Distribution of sea-air CO2 fluxes in the Patagonian Sea: Seasonal, biological and thermal effects. Cont. Shelf Res., 143, 18–28, https://doi.org/10.1016/j.csr.2017.05.011.
Karlin, L. N., V. N. Malinin, and S. M. Gordeeva, 2015: On the influence of the sea surface temperature on the carbon dioxide exchange with the atmosphere. Oceanology, 55, 11–19, https://doi.org/10.1134/S0001437015010099.
Kljun, N., P. Calanca, M. W. Rotach, and H. P. Schmid, 2015: A simple two-dimensional parameterisation for Flux Footprint Prediction (FFP). Geosci. Model Dev., 8, 3695–3713, https://doi.org/10.5194/gmd-8-3695-2015.
Lachkar, Z., and N. Gruber, 2013: Response of biological production and air–sea CO2 fluxes to upwelling intensification in the California and Canary Current Systems. J. Mar. Syst., 109–110, 149–160, https://doi.org/10.1016/j.jmarsys.2012.04.003.
Laruelle, G. G., R. Lauerwald, B. Pfeil, and P. Regnier, 2014: Regionalized global budget of the CO2 exchange at the air‐water interface in continental shelf seas. Global Biogeochem. Cycles, 28, 1199–1214, https://doi.org/10.1002/2014GB004832.
Li, Q., X. Guo, W. Zhai, Y. Xu, and M. Dai, 2020: Partial pressure of CO2 and air-sea CO2 fluxes in the South China Sea: Synthesis of an 18-year dataset. Prog. Oceanogr., 182, 102272, https://doi.org/10.1016/j.pocean.2020.102272.
Lim, J. T., 2004: Weather and Climate of Malaysia. University of Malaya Press, 170 pp.
Lorenz, M., J. Bliefernicht, B. Haese, and H. Kunstmann, 2018: Copula‐based downscaling of daily precipitation fields. Hydrol. Processes, 32, 3479–3494, https://doi.org/10.1002/hyp.13271.
Mandal, S., N. Behera, A. Gangopadhyay, R. D. Susanto, and P. C. Pandey, 2021: Evidence of a chlorophyll “tongue” in the Malacca Strait from satellite observations. J. Mar. Syst., 223, 103610, https://doi.org/10.1016/j.jmarsys.2021.103610.
Mari, X., 2008: Does ocean acidification induce an upward flux of marine aggregates? Biogeosciences, 5, 1023–1031, https://doi.org/10.5194/bg-5-1023-2008.
Mauder, M., T. Foken, R. Clement, J. A. Elbers, W. Eugster, T. Grünwald, B. Heusinkveld, and O. Kolle, 2008: Quality control of CarboEurope flux data–Part 2: Inter-comparison of eddy-covariance software. Biogeosciences, 5, 451–462, https://doi.org/10.5194/bg-5-451-2008.
Mauri, E., P.-M. Poulain, and Ž. Južnič‐Zonta, 2007: MODIS chlorophyll variability in the northern Adriatic Sea and relationship with forcing parameters. J. Geophys. Res., 112, C03S11, https://doi.org/10.1029/2006JC003545.
Maznah, W. O. W., N. A. Kassim, and Z. Din, 2021: Zooplankton community structure in relation to the water quality and seston fatty acid content in the coastal waters of Penang, Malaysia. Sains Malays., 50, 1577–1588, https://doi.org/10.17576/jsm-2021-5006-06.
McGillis, W. R., J. B. Edson, J. E. Hare, and C. W. Fairall, 2001: Direct covariance air–sea CO2 fluxes. J. Geophys. Res., 106, 16 729–16 745, https://doi.org/10.1029/2000JC000506.
Mitra, A., and Coauthors, 2014: The role of mixotrophic protists in the biological carbon pump. Biogeosciences, 11, 995–1005, https://doi.org/10.5194/bg-11-995-2014.
Moncrieff, J. B., and Coauthors, 1997: A system to measure surface fluxes of momentum, sensible heat, water vapour and carbon dioxide. J. Hydrol., 188–189, 589–611, https://doi.org/10.1016/S0022-1694(96)03194-0.
Moncrieff, J. B., R. Clement, J. Finnigan, and T. Meyers, 2004: Averaging, detrending, and filtering of eddy covariance time series. Handbook of Micrometeorology, X. Lee, W. Massman, and B. Law, Eds., Atmospheric and Oceanographic Sciences Library, Vol. 29, Springer, 7–31.
Morin, T. H., G. Bohrer, K. C. Stefanik, A. C. Rey-Sanchez, A. M. Matheny, and W. J. Mitsch, 2017: Combining eddy-covariance and chamber measurements to determine the methane budget from a small, heterogeneous urban floodplain wetland park. Agric. For. Meteor., 237–238, 160–170, https://doi.org/10.1016/j.agrformet.2017.01.022.
O’Reilly, J. E., and P. J. Werdell, 2019: Chlorophyll algorithms for ocean color sensors—OC4, OC5 & OC6. Remote Sens. Environ., 229, 32–47, https://doi.org/10.1016/j.rse.2019.04.021.
Peterson, I., B. Greenan, D. Gilbert, and D. Hebert, 2017: Variability and wind forcing of ocean temperature and thermal fronts in the slope water region of the northwest Atlantic. J. Geophys. Res. Oceans, 122, 7325–7343, https://doi.org/10.1002/2017JC012788.
Pfeil, B., and Coauthors, 2013: A uniform, quality controlled Surface Ocean CO2 Atlas (SOCAT). Earth Syst. Sci. Data, 5, 125–143, https://doi.org/10.5194/essd-5-125-2013.
Prasanna Kumar, S., P. M. Muraleedharan, T. G. Prasad, M. Gauns, N. Ramaiah, S. N. de Souza, S. Sardesai, and M. Madhupratap, 2002: Why is the Bay of Bengal less productive during summer monsoon compared to the Arabian Sea? Geophys. Res. Lett., 29, 88–1–88–4, https://doi.org/10.1029/2002GL016013.
Qu, B., J. Song, X. Li, H. Yuan, N. Li, and Q. Ma, 2013: pCO2 distribution and CO2 flux on the inner continental shelf of the East China Sea during summer 2011. Chin. J. Oceanol. Limnol., 31, 1088–1097, https://doi.org/10.1007/s00343-013-2225-9.
R Development Core Team 3.0.1., 2013: A language and environment for statistical computing. R Foundation for Statistical Computing, accessed 10 February 2015,https://www.gbif.org/tool/81287/r-a-language-and-environment-for-statistical-computing.
Riebesell, U., and Coauthors, 2007: Enhanced biological carbon consumption in a high CO2 ocean. Nature, 450, 545–548, https://doi.org/10.1038/nature06267.
Sailley, S. F., and Coauthors, 2013: Comparing food web structures and dynamics across a suite of global marine ecosystem models. Ecol. Modell., 261–262, 43–57, https://doi.org/10.1016/j.ecolmodel.2013.04.006.
Samah, R., 2020: Increasing the fish production with good aquaculture practices. Int. J. Bus. Technol. Manage., 2, 27–33.
Sirinanda, K. U., 1979: The role of applied climatological research in alleviating the hazardous impact of the monsoons in South and Southeast Asia. GeoJournal, 3, 137–146, https://doi.org/10.1007/BF00257702.
Siswanto, E., and Coauthors, 2011: Empirical ocean-color algorithms to retrieve chlorophyll-a, total suspended matter, and colored dissolved organic matter absorption coefficient in the Yellow and East China Seas. J. Oceanogr., 67, 627–650, https://doi.org/10.1007/s10872-011-0062-z.
Smetacek, V., 1999: Diatoms and the ocean carbon cycle. Protist, 150, 25–32, https://doi.org/10.1016/S1434-4610(99)70006-4.
Sparks, N., and R. Toumi, 2020: Pacific subsurface ocean temperature as a long-range predictor of South China tropical cyclone landfall. Commun. Earth Environ., 1, 30, https://doi.org/10.1038/s43247-020-00033-2.
Steinberg, D. K., and M. R. Landry, 2017: Zooplankton and the ocean carbon cycle. Annu. Rev. Mar. Sci., 9, 413–444, https://doi.org/10.1146/annurev-marine-010814-015924.
Stull, R. B., 1988: Mean boundary layer characteristics. An Introduction to Boundary Layer Meteorology, R. B. Stull, Ed., Atmospheric Sciences Library, Vol. 13, Springer, 1–27.
Świrgoń, M., and M. Stramska, 2015: Comparison of in situ and satellite ocean color determinations of particulate organic carbon concentration in the global ocean. Oceanologia, 57, 25–31, https://doi.org/10.1016/j.oceano.2014.09.002.
Takahashi, T., and Coauthors, 2002: Global sea–air CO2 flux based on climatological surface ocean pCO2 , and seasonal biological and temperature effects. Deep-Sea Res. II, 49, 1601–1622, https://doi.org/10.1016/S0967-0645(02)00003-6.
Tan, C. K., J. Ishizaka, S. Matsumura, F. M. Yusoff, and M. I. H. Mohamed, 2006: Seasonal variability of SeaWiFS chlorophyll a in the Malacca Straits in relation to Asian monsoon. Cont. Shelf Res., 26, 168–178, https://doi.org/10.1016/j.csr.2005.09.008.
Tan, T. K., 2005: Spatial and temporal variation of satellite ocean color in the eastern Indian Ocean. Ph.D. dissertation, Nagasaki University, 94 pp.
Tian, Y., Z. Mingjiang, and Q. Peiyuan, 2002: Combined effects of temperature, irradiance and salinity on growth of diatom Skeletonema costatum. Chin. J. Oceanol. Limnol., 20, 237–243, https://doi.org/10.1007/BF02848852.
Tomas, C. R., 1997: Identifying Marine Phytoplankton. 1st ed. Academic Press, 874 pp.
Torrence, C., and G. P. Compo, 1998: A practical guide to wavelet analysis. Bull. Amer. Meteor. Soc., 79, 61–78, https://doi.org/10.1175/1520-0477(1998)079<0061:APGTWA>2.0.CO;2.
van Dam, B., and Coauthors, 2021: Global trends in air‐water CO2 exchange over seagrass meadows revealed by atmospheric eddy covariance. Global Biogeochem Cycles, 35, e2020GB006848, https://doi.org/10.1029/2020GB006848.
Vickers, D., and L. Mahrt, 1997: Quality control and flux sampling problems for tower and aircraft data. J. Atmos. Oceanic Technol., 14, 512–526, https://doi.org/10.1175/1520-0426(1997)014<0512:QCAFSP>2.0.CO;2.
Wan Maznah, W. O., K. Yahya, and A. H. Ismail, 2014: Zooplankton community structure in the coastal waters of Penang National Park, Penang, Malaysia, and their relation with water quality conditions. Annual Int. Conf. on Ecology, Biodiversity and Environment (CEBE 2014), Singapore, Global Science and Technology Forum, 42–50.
Wang, F., and Coauthors, 2021: Technologies and perspectives for achieving carbon neutrality. Innovation, 2, 100180, https://doi.org/10.1016/j.xinn.2021.100180.
Ward, N. D., and Coauthors, 2020: Representing the function and sensitivity of coastal interfaces in Earth system models. Nat. Commun., 11, 2458, https://doi.org/10.1038/s41467-020-16236-2.
Webb, E. K., G. I. Pearman, and R. Leuning, 1980: Correction of flux measurements for density effects due to heat and water vapour transfer. Quart. J. Roy. Meteor. Soc., 106, 85–100, https://doi.org/10.1002/qj.49710644707.
Weiss, R. F., 1974: Carbon dioxide in water and seawater: The solubility of a non-ideal gas. Mar. Chem., 2, 203–215, https://doi.org/10.1016/0304-4203(74)90015-2.
Wieder, W. R., Z. Butterfield, K. Lindsay, D. L. Lombardozzi, and G. Keppel‐Aleks, 2021: Interannual and seasonal drivers of carbon cycle variability represented by the Community Earth System Model (CESM2). Global Biogeochem. Cycles, 35, e2021GB007034, https://doi.org/10.1029/2021GB007034.
Wik, M., B. F. Thornton, D. Bastviken, J. Uhlbäck, and P. M. Crill, 2016: Biased sampling of methane release from northern lakes: A problem for extrapolation. Geophys. Res. Lett., 43, 1256–1262, https://doi.org/10.1002/2015GL066501.
Wilczak, J. M., S. P. Oncley, and S. A. Stage, 2001: Sonic anemometer tilt correction algorithms. Bound.-Layer Meteor., 99, 127–150, https://doi.org/10.1023/A:1018966204465.
World Tourism Organization and Global Tourism Economy Research Centre, 2019: UNWTO/GTERC Asian Tourism Trends—2018 edition. UNWTO/GTERC Rep., 128 pp., https://doi.org/10.18111/9789284420032.
Yang, M., T. J. Smyth, V. Kitidis, I. J. Brown, C. Wohl, M. J. Yelland, and T. G. Bell, 2021: Natural variability in air–sea gas transfer efficiency of CO2. Sci. Rep., 11, 13584–, https://doi.org/10.1038/s41598-021-92947-w.
Yasunaka, S., and Coauthors, 2018: Arctic Ocean CO2 uptake: An improved multiyear estimate of the air–sea CO2 flux incorporating chlorophyll a concentrations. Biogeosciences, 15, 1643–1661, https://doi.org/10.5194/bg-15-1643-2018.
Ye, Y., B. Liang, S. Li, and J. Hu, 2021: A study on the response of carbon cycle system in the Pearl River Estuary to riverine input variations. J. Mar. Syst., 215, 103498, https://doi.org/10.1016/j.jmarsys.2020.103498.
Yusoff, A., 2015: Status of resource management and aquaculture in Malaysia. Southeast Asian Fisheries Development Center Aquaculture Department Institutional Repository Doc., 13 pp., https://core.ac.uk/download/pdf/77980929.pdf.
Yusup, Y., W. A. A. Alqaraghuli, and A. F. M. Alkarkhi, 2016: Factor analysis and back trajectory of PM and its metal constituents. Environ. Forensics, 17, 319–337, https://doi.org/10.1080/15275922.2016.1177757.
Yusup, Y., A. F. M. Alkarkhi, J. S. Kayode, and W. A. A. Alqaraghuli, 2018a: Statistical modeling the effects of microclimate variables on carbon dioxide flux at the tropical coastal ocean in the southern South China Sea. Dyn. Atmos. Oceans, 84, 10–21, https://doi.org/10.1016/j.dynatmoce.2018.08.002.
Yusup, Y., J. S. Kayode, and A. F. M. Alkarkhi, 2018b: A methodological approach to the air-sea energy fluxes data collection and analysis at the tropical coastal ocean. MethodsX, 5, 448–453, https://doi.org/10.1016/j.mex.2018.05.003.
Yusup, Y., J. S. Kayode, and A. F. M. Alkarkhi, 2019: Monsoonal and annual energy fluxes variations in the atmospheric surface layer over a coastal region of the tropical ocean. Reg. Stud. Mar. Sci., 29, 100707, https://doi.org/10.1016/j.rsma.2019.100707.
Yusup, Y., N. K. Ramli, J. S. Kayode, C. S. Yin, S. Hisham, H. Mohamad Isa, and M. I. Ahmad, 2020: Atmospheric carbon dioxide and electricity production due to lockdown. Sustainability, 12, 9397, https://doi.org/10.3390/su12229397.
Zanuri, N. B. M., M. B. Abdullah, N. A. M. Darif, N. Nilamani, and A. T. S. Hwai, 2020: Case study of marine pollution in Teluk Bahang, Penang, Malaysia. IOP Conf. Ser., 414, 012032, https://doi.org/10.1088/1755-1315/414/1/012032.
Zhai, W.-D., and Coauthors, 2013: Seasonal variations of air-sea CO2 fluxes in the largest tropical marginal sea (South China Sea) based on multiple-year underway measurements. Biogeosciences, 10, 7775–7791, https://doi.org/10.5194/bg-10-7775-2013.
Znachor, P., and J. Nedoma, 2010: Importance of dissolved organic carbon for phytoplankton nutrition in a eutrophic reservoir. J. Plankton Res., 32, 367–376, https://doi.org/10.1093/plankt/fbp129.
Zou, K. H., K. Tuncali, and S. G. Silverman, 2003: Correlation and simple linear regression. Radiology, 227, 617–628, https://doi.org/10.1148/radiol.2273011499.
