The distribution of dissolved REEs varies from upstream to downstream especially during November and January. During this period, the total dissolved rare earth elements (ΣREEs) ranged from 368 to 9121 pmol∙L⁻¹, with a mean value of 2328 ± 1442 pmol∙L⁻¹, and LREE/HREE ratio ranging from 5.20 ± 0.97 to 22.41 ± 3.00 (Table 2). The LREE concentrations along the sampling stations ranged from 344 to 8724 pmol∙L⁻¹ and HREE from 23.5 to 593 pmol∙L⁻¹. According to McLennan (1989) [19] the abundance of total dissolved REEs strongly supports the huge concentration of Cerium (Ce). The concentrations of Ce also varied between 84 to 3237 pmol∙L⁻¹ with the highest mean concentration of 1680 ± 730 pmol∙L⁻¹ recorded in November 2014 (Table 2). These results were probably due to the high rainfall events [14]. Dissolution of Ce in January 2014 was observed to be statistically higher due to the higher monthly rainfall distribution (r = 0.328; p < 0.05) with 277 mm (Figure 2). The same results were obtained after the flooding disaster in December 2013 due to heavy rainfall (>1800 mm) as well as in November 2014 when the rainfall was about 288 mm. Hence these incidents have increased the REE contents in the water column.

There were no significant differences and distributions of REE contents among the stations from upstream to downstream (p > 0.05) except between sampling periods (p < 0.05) because the depth of the water column is very shallow e.g., less than 5 m during high tide which might have caused active resuspension. According to Shynu et al. (2011) [3] the resuspension activities of dissolved REEs in rivers might be from surface sediment to water column as well as re-suspended in subsurface waters from the river runoff. The resuspension and mixing process of REEs in the water column was also observed based on the significant statistical correlation between Al and Fe (R² = 0.65; p < 0.01) that was discussed previously in Bourg (1983) [15]. Regarding the concentration of Fe and Al that showed opposite trends to concentration of dissolved REEs, the species in Table 2 might be desorbed from particles or resuspended from surface sediment. This may have been caused by active river discharge transportation from the upper stream after incorporation of heavy rainfall and sources from groundwater discharge [12].

The sampling sites were divided into three zones which were river, estuarine and coastal (Table 1; Figures 3(b)-(d)) to examine the dissolved REE concentrations based on the salinity gradient whereby 27 - 33 psu, 13 - <27 psu and <13 psu represent coastal, estuary and river areas, respectively. The concentrations of ΣREEs slightly decreases towards the estuarine zone (mean range: 514 - 3713 pmol∙L⁻¹) and increases in the coastal zone (mean range: 387 to 6087 pmol∙L⁻¹).

The same scenarios were also observed by the other elements e.g., Ce, Nd, Eu and Ho in the coastal zone especially during monsoon events because of the abundant inorganic carbon in seawater during semidiurnal tides and the tidal pump phenomenon [20]. The river zones occupied by station 2 and station 3 have a water depth of less than 5 m along the sand bar. Hence, these stations experienced higher deposition of sediments and particles from the sand bar (Figure 1; Table 1). However, the fluctuating rainfall events affected the concentrations of suspended particulate matter in surface river discharges (Figure 3(a)).

According Agaki (2013) [21], it was indicated that the silicate or opal has the potential as an REE scavenger in the seawater column. Notwithstanding the presence of abundant biogenic SiO₂ also has a linear relationship with organic matter or organic carbon in the estuarine or marine environment. The distribution of silicate originated from terrestrial sources and detrital minerals from lithogenous and hydrogenous phases [22]. Sungai Balok is surrounded by a mangrove habitat and studies of estuarine fluxes by using REE proxies around the world found that the tropical river flux of trace metals in the dissolved phase have a high possibility of being from rainfall runoff and submarine groundwater discharge [23]. The concentration of biogenic SiO₂ from the wet season in January 2014 to dry season during April 2014 and July 2014 (Table 3) found that the high distribution of REEs in January 2014 was associated with the increase of biogenic SiO₂. While the decrease of REE concentrations in April 2014 and July 2014 were associated with decreasing biogenic SiO₂, which is known as an REE scavenger in the marine environment.

The high ratio value of HREEs in the dissolved phase especially at station ST 1 (Figure 4(h)) might be because of inorganic carbonate from seawater. The inorganic carbonate has been depicted as the greatest positive factor of Tm anomaly (>28) with similar patterns in Figures 4(d)-(f). A positive Gd anomaly in Figure 4(a) & Figure 4(b) with a range value of lower than 6 also indicated more natural freshwater from the hinterland to the study location [24], and the high positive anomaly value (>5) of Pr in Figure 4(d) was due to regenerated particle matter from surface sediments and nutrients [25]. Similarly, researchers have already discussed the enrichment pattern of MREEs and the enrichment pattern of MREE in this sampling locations also showed abundant positive anomaly of Eu (e.g., Figure 4(a); Figure 4(e) and Figure 4(i)) (i.e., [3] [14]).

(*) single reading without replicate.

Sungai Balok drains towards the southern South China Sea (sSCS) and the effect of salt water intrusions is higher compared to the west coast of Peninsular Malaysia i.e., the Straits of Malacca [26]. The dissolved REEs have a strong affinity to particulate REEs and reactive particles in the marine environment [1]. Terrestrial runoff occurring during heavy rainfall may have brought an abundance of terrigenous detrital matter and particulate organic matter into the river and ocean environments. The negative statistical correlation was observed between particles and rainwater, which means particles such as dust, particulate matter and volatile gases were scavenged during wet deposition [27].

Ce species is dominant in certain conditions due to abundance among the REEs and in the environment it can be applied to determine the oxidation and reduction stage of cerium either Ce/Ce* < 1 and Ce/Ce* > 1 by Ce⁴⁺ and Ce³⁺, respectively. The adsorption of dissolved REEs will be scavenged to surface sediments, and the negative Ce anomaly (Ce/Ce*) along the stations ranging from 0.2 to 1.33 showed that oxidation process of stage IV occurred at Sungai Balok. Researchers Nozaki et al. (1997) [28] and Sholkovitz et al. (1999) [2] also explained the adsorption process could be detected by process of Ce oxidation in the marine environment. However, De Baar (1983) [29] noted that in the presence of oxygen, the seabed regeneration can reduce the Ce (IV) to Ce (III) especially in the river and terrestrial area. Sungai Balok also showed that the positive Ce/Ce* was abundant at the stations in the river and estuarine zones (i.e., ST3 to ST7) ranging from 1.0 to 1.33. The highest was recorded at ST7 in November 2013 and might be due to the weathering process as shown by lowest ratio value of Y/Ho (9.247 ± 4.026) and high contents of Mn (3.543 ± 2.74 μmol∙L⁻¹) compared to other sampling periods. According to Sholkovitz (1992) [4], the increased Mn oxide acts as the catalyst for supplying oxygen demand in surface water in order to maintain the dissolved REE in III charge. The moderate and strong significant correlations were determined among Ce, Eu, Nd and Ho, however, strong correlation (p < 0.001) was observed between LREE and total REE indicating chemical charge of REE (III) has strong binding affinity among dissolved REE members.

Europium (Eu) is a suitable terrestrial dust proxy available from the lithogenic sources such as igneous, granite rock, dust, hydrothermal and volcanic land (e.g., [4] [14]). The Eu anomaly has the ability to evaluate the lithogenic input through the ratio of Y/Ho [30], where the ratio Y/Ho < 55 in the study suggests a lot of freshwater input in the Sungai Balok river system (Figure 5) especially during the Northeast monsoon. The values of Eu anomalies also fluctuated in all the zones ranging 0.489 to 3.3 in the coastal zone, 0.836 to 2.767 in the estuarine zone, and 0.755 to 2.68 in the river zone (Table 2) which might be due to the fluctuation of monsoon events in the sSCS region.

The ratio of La/Yb_N is a common indicator to determine the enrichment of LREE over HREE [31] and the ratio was also used as a lithogenic proxy in Tokyo Bay by Takahashi and Noriki (2007) [6]. Meanwhile, the fluctuating rainfall events in the East coast of Peninsular of Malaysia have significant statistical correlation (p = 0.05) with La/Yb_N ratio and the ratios were high in January 2014 and July 2014 with values of 6.43 ± 2.47 and 7.098 ± 3.81, respectively. The overall ratios of La/Yb_N ranged between 0.69 to 11.57 throughout the sampling periods. The high amount of monthly rainfall distribution in January 2014 increased the soil humidity and dissolution of dissolved REE. Hence, the highest total concentration of 4825 ± 2460 pmol∙L⁻¹ was observed during that time. However, the high ratio of La/Yb_N in July 2014 (i.e., lower monthly rainfall) might be due to sudden rainfall during the sampling day along with concentration of 11 mm per day as reported by MMD (2015) [32] (Figure 2).

Sungai Balok falls into the southern South China Sea zone [14] and the amount of REEs discharged are still not well documented. Due to the abundance of Ce among REE members, it has been selected for the calculation of river fluxes. The flux of REEs was estimated using formula (1), while an inventory could be developed by formula (2) adopted from the published journal [28]. The inventory and flux of dissolved REEs were estimated using the Ce element in Table 4 and Table 5. By consideration of the residence time of Ce [33] where the value of Ce is 80 years while Nd is 270 years. Due to the unstable nature of Ce, it is not the main choice for the geochemistry scientists to determine the residence time in

*D; Average depth stations. Note: Inventory is obtained from Equation (2); The value of L is converted to cm³ and m is converted to cm, where 1L is equals to 1000 cm³ and 1m is equals to 100 cm.

Note: Period was divided to two seasons which were during monsoon from November to January while after monsoon from April until July of sampling periods. Average Ce is obtained from average concentration within four months. The conversion factors 1/N of sampling periods are obtained by calculated the days of sampling period time to value 86400 of second per days. Fluxes is obtained from Equation (1) after entering the value of conversion factor, average daily river discharge and dissolved Ce concentration.

marine, coastal and ocean areas and mostly use the Nd as a proxy. However, from this finding the dissolved REE river flux was identified by measuring the accumulation duration over the sampling period conducted in Table 5, but not the remaining dissolved REE or resident at the sampling point.

Inventory = C REE × D (2)

where, the C is the concentration of dissolved REE (pmol∙L⁻¹) in the water column, D is the depth of water column (m). Therefore, an inventory could also be produced by multiplication of the C which is concentration of dissolved REE (pmol∙L⁻¹) in the water column with D which is the depth of water column (m).

The accumulation of dissolved REEs could be represented by the dissolved Nd and Eu fluxes with the overall domination during the Northeast monsoon which was also supported by the previous findings in November 1996 [34] and in August 1997 [13]. Although Nd fluxes were widely used to measure the river and seawater fluxes for dissolved REE deposition, the Eu flux could potentially be used to measure the lithogenic input into the river system. According to Nozaki et al. (1999) [28] the South China Sea is the origin of the fluvial and coastal input that has been feeding the Northern South China Sea and also the Sulu Sea at a rate of 1.7 Nd pmol∙cm²∙yr⁻¹, while other Nd flux contributions were also 3.2 pmol∙cm⁻²∙yr⁻¹ [5] and 1.2 pmol∙cm⁻²∙yr⁻¹ [35]. Similarly, newer studies on the dissolved Nd flux budget in the Bay of Bengal also reported 1.9 to 6.0 × 10⁻⁷ mol∙m⁻²∙yr⁻¹ or 1.9 to 6 mmol∙cm⁻²∙yr⁻¹ caused by high lithogenic input from the Ganges-Brahmaputra river [8].

The calculated inventories of dissolved Ce in January and November 2014 were 586.5 pmol∙cm⁻² and 643.4 pmol∙cm⁻², respectively (Table 4). The results from Sungai Balok might be due to the dilution of river water because of frequent heavy rain during the Monsoon and freshwater discharge during weathering processes the in southern South China Sea. Hence, the river flux deposition increased during both periods of November 2013 to January 2014 and November 2014 to January 2015 along the value of 39.14 nmol∙cm⁻²∙yr⁻¹ and 59.78 nmol∙cm⁻²∙yr⁻¹ (Table 5). Moreover, the inventory of dissolved Ce also showed a strong correlation with monthly rainfall distribution (R² = 0.52). This statistical correlation showed the possibility of higher rainfall distributions in November until January that might have caused the increased dissolved Ce flux compared to other periods. Overall the dissolved Ce fluxes over seasons as shown in Table 5 were greater compare to previous reports from Nozaki et al. (1999) [28]. The Nd was used to budget the deposition fluxes in offshore areas and oceans by Yu et al. (2017) [8] because of the longer residence time in deep locations i.e., it was estimated about 1.5 - 2.6 years. Therefore, it was not practical to state that where the depth of the river is shallower, the residence time is shortest. Furthermore, according to Greaves et al. (1994) [35] the reason to use Nd for estimating flux is because Nd and Sm are good indicators for river evolution. However, most of the tropical rivers are still in pristine condition [36], probably due to wash out by a lot of freshwater and abundant of rainfall. This consequently delivered abundant lithogenic sources especially Al and Ce from crustal and most of the Asean continental parts. Despite the unstable Ce element caused by oxidation and reduction, the study has considered to choose it based on the findings of similar river fluxes assessed by Sholkovit et al. (1999) [2] where Eu followed the same behavior of Ce.

According to Sholkovits and Szymczak (2000) [37], the Indonesian archipelagos, Papua New Guinea (PNG), Peninsular Malaysia and Borneo were believed to supply more than 20% of global sediment input into the sea. The island weathering activities explained that the terrestrial sources draining into rivers and the huge river discharges of more than 2 × 10¹⁵ m³∙s⁻¹ fold was due to storms and inundation [2]. The PNG and Indonesia archipelagoes were also fed by sources of volcanic and hydrothermal eruptions. Thus, river input has made the fluvial flux amount increase to 1.4 × 10¹⁶ pmol∙yr⁻¹ of Eu or estimated to be about 1.4 × 10¹⁰ pmol∙cm⁻²∙yr⁻¹. These values are greater compared to this present study and might be due to lithogenic and terrigenous volcanic sources around the Indonesian archipelagoes. Furthermore, according to Yu et al. (2017) [8], the amount of fluvial fluxes were also affected by biological uptake from organisms such as plankton, process of iron hydroxide co-precipitation and colloidal coagulation in the marine environment and physical factors such as canyons and tidal changing phenomena. However, the abundance of basaltic and granitic rocks [14] along the East Coast of Peninsular Malaysia should be considered for the increased dissolved Ce Fluxes as well as rainfall events that make Sungai Balok unique.