In-Situ Variability of DOM in Relation with Biogeochemical and Physical Parameters in December 2017 in Laucala Bay (Fiji Islands) after a Strong Rain Event

: Heavy rain events alter the biogeochemical outﬂows, affects water quality and ecosystem health within the coastal waters of small Paciﬁc Islands. We characterized snapshots of the optical ﬁngerprints of dissolved organic matter (DOM) sources together with the select nutrients, biogeochemical and physical variables for 10 stations in December 2017 in Laucala Bay, Fiji Islands. DOM absorption coefﬁcients and ﬂuorescence components were determined via spectroﬂuorometry and Parallel Factor Analysis identifying four components: Type M, two terrestrial (humic, fulvic) components, and a protein component linked to marine biological activity. Associations of DOM together with climate variables and the other tested variables were determined via principal component, hierarchical cluster, and cross-correlation (Pearson) analysis. All component s (together with most tested variables) displayed higher values (plumes) at the southwest coast consistent with surface currents outﬂow during the wet season. Type M component associated with two allochthonous ﬂuorescent components signaling anthropogenic forcings via riverine outﬂows. Terrigenous inputs association with autochthonous chromophoric dissolved organic matter (CDOM) is indicative of tidal mixing, dilution, and bottom resuspension processes. Positive correlations of dissolved organic carbon (DOC) with nutrients (NO x , PO 4 ) elucidates DOM being utilized as energy sources. The positive correlation of DON with nutrients (NO x , PO 4 , Si(OH) 4 ) reﬂects the role of DON as a nutrient source consistent with chlorophyll plume formation.


Introduction
Like many developing countries, Fiji is a natural resource-dependent state experiencing intense competing pressures for agriculture, tourism, transport, mining, and many more land-use practices, which in total affects the health and quality of the surrounding water bodies [1][2][3][4]. Consequently, water clarity is affected by anthropogenic forces on terrestrial runoffs and riverine discharges, such as DOM (dissolved organic matter) sourced outflows from sewage, mining, sugarcane farming, land clearing, dredging, and natural processes, such as shoreline erosion, sediment resuspension from tidal action, coastal surges and wave action [4][5][6][7]. DOM is ubiquitous in aquatic systems and serves as the largest reservoir of dissolved organic carbon (DOC) in the oceans [8][9][10].
DOM contains a highly diverse mixture of molecules [10,11], which provide the largest source of energy and nutrients for the aquatic biomes. The large quantities of carbon (C),

In-Situ Sampling and Sea Measurements
Out of the 55 stations surveyed in Laucala Bay during the 5 BULA IRD cruises [41][42][43][44][45] surface sampling sites at the Suva Lagoon, a total of 10 stations were selected into 4 transects, to cover the Laucala Bay monitoring sites and organized along 4 transects T1-T4 (north to the south except for T4) ( Figure 1, Table 1). Stations were organized along 4 transects T1-T4 (north to the south except for T4). T1 (St 22-23) covered the Nasese passage from the coastal vicinity of the Vatuwaqa river mouth into the near reef vicinity while T2 (St 30-31-33) surveys the mouth of the Samabula River to the reef edge. T3 (  covers the coastal vicinity of the Nasinu and Vunidawa rivers to the reef edge while T4 (St [47][48][49] tries to cover the Nukulau and the Laucala passages. All transects cogitate river mouths and passages (including KWTP outfall) with stations positioned on a roughly regular grid from the coast to the barrier reefs into POW. Stations were grouped into three zones (coastal, CO; inner bay, IB; and outer bay, OB) as illustrated by black dashed lines in Figure 1 and defined by an imaginary line running at about a kilometer (km) from the coastline; IB running 2 km along the coastline while OB refers to an imaginary line running at about a km from the inner reef edges of the barrier reefs. The Rewa River could not be sampled at low tide.
The survey cruise was conducted onboard the MV Halimeda on 1 December 2017 from 09:00 h-17:00 h. The mean depth of the waters recorded at all stations was 15.13 m and the mean rainfall received during the past three months before the survey was 12.42 mm. Wind intensities fluctuated in 2017 recording monthly means of 3.90 ms −1 , 2.12 ms −1, and 3.91 ms −1 for October, November, and December, respectively. Strong rains of about 74.4 mm were observed on 1 December 2017 compared to the monthly mean of 13.65 mm for the same month. The mean sea level recorded for the past three months before the survey was 1.23 m ± 0.44 m while the sampling was conducted when the tide was decreasing from 1.6 m to 0.6 m, then rose to 0.8 m during the end of survey time.

Physical Parameters
A calibrated Sea-Bird SBE19 CTD (depth profiler device for detection of conductivity, temperature, and depth) and a YSI Model 30 handheld meter (for the surface waters) were used to measure temperature, salinity, dissolved oxygen, total dissolved solids in the liquid sample (TDS). The exact depth of the lagoon waters was also estimated as the last depth of the CTD profile.

Biogeochemical Parameters Sample Conditioning
Water samples were obtained at 1-m depth using a 2 L Niskin bottle to fill a 4.5 L gallon for Chlorophyll, Suspended matter concentration and absorption, and a 1 L clear bottle for DOC and CDOM/FDOM, 2 cryotubes (cytometry), and 2 × 40 mL bottles for nutrients [46]. The water samples (biogeochemical parameters) were filtered and conditioned in the Marine Studies Laboratory (USP) for further analyses in the chemistry laboratories MIO, Noumea, Toulon, and Marseille. Samples for nutrient analysis were prepared by addition of 20 µL of HgCl 2 to each 40 mL of unfiltered seawater samples that were collected and put in plastic bottles previously rinsed in 10% HCl and rinsed three times with the sample and then kept in the dark at 4 • C before analysis.
For suspended matter, a variable volume between 300 mL to 1 L of seawater was filtered with a NALGENE filtering unit onto a 47 mm filter diameter (0.45 µm of porosity) that was previously weighed when empty after 24 h in an oven at 60 • C (in triplicates). After filtration, 20 mL of ammonium formate was added and sipped in the filtration unit to dissolve additional sea salt on the filter. Filters were then put in Petri dishes in an oven at 60 • C for 24 h and kept dried until weighting. For Chla concentration, a variable volume from 300-1000 mL was filtered onto GFF Whatman 25 mm diameter (0.7 µm of porosity) and then immediately stored at −20 • C before analysis. For DOC analyses, 25 mL of seawater was filtered with the NALGENE precedingly rinsed twice with distilled water onto 47 mm diameter filters (0.2 µm of porosity) and the filtrate was conditioned in pre-combusted glass ampoules and capped with aluminum foil before use, flame-sealed after addition of 50 µL of 85% phosphoric acid to block bacterial activity. For CDOM/FDOM, the same filtration protocol was used except that f porosity) and the filtrate was conditioned into 25 mL pre-combusted (450 • C, 6 h) 25-mL Schott glass bottles and capped with aluminum foil before use. All filtered samples (DOC, CDOM/FDOM) were stored then kept in the dark at 4 • C before analysis.
For the absorption by particulates, a variable volume of seawater was filtered (from 0.3-1 L) onto GFF Whatman 25 mm Ø (0.7 µm of porosity) using a filtration unit (glass) and stored at −20 • C at the laboratory before analysis.

Biogeochemical Parameters Laboratory Methodology
The spectral absorption by CDOM(a CDOM , in m −1 ) was obtained from filtered seawater samples (on 0.2 µm filters) as in [16]. Briefly, spectrum measurements were conducted from 275 to 800 nm in 2 nm intervals using a liquid core waveguide system with an optical pathlength of 2 m (LLCW, World Precision Instruments, Inc., FL, USA) and a Hamamatsu spectrophotometer C10082CA, Hamamatsu, Japan. Between each processing of the sample, the capillary waveguide cell was flushed and filled with purified water used as blank according to the method described by [9]. After the acquisition of a dark (detector out) and the reference blank with Milli-Q water, each spectrum was measured in triplicate to ensure replicability. All measured spectra were corrected for residual absorbance, by assuming that the average of measured values over a 5 nm interval around 698 nm must be zero and shifting the spectra as proposed by [47]. CDOM absorption spectrum values extracted at different wavelengths include: a CDOM( 254), a CDOM( 350), a CDOM( 412) and a CDOM( 442) nm.

DOC (µM)
The DOC concentration was measured using a SHIMADZU TOC 5000 equipped with an autosampler. To eliminate dissolved organic carbonate, 50 mL of HCl was added followed by oxygen bubbling for ten minutes. The sample was analyzed by high-temperature catalytic combustion (650 • C). The obtained detection limit for this procedure was 0.15 µM (Mediterranean Institute of Oceanography, Campus de Toulon). Instrumental uncertainty on DOC values is 2-3 µM.

Nutrients (µM L −1 )
Nitrate + nitrite (reported as NO x ) concentrations were determined at nanomolar concentrations [48], phosphate (PO4 3 ) concentrations were determined according to [49] and silicates (Si(OH) 4 ) according to [50,51] with all analyses being conducted on a Bran-Luebbe III continuous flow autoanalyzer. Total dissolved nitrogen (TON) and phosphorus concentrations were determined as nitrate and phosphate (PO 4 ) concentrations after wet oxidation of water samples.

SPM-Dry Weight (mg L −1 )
Suspended particulate matter (SPM), a proxy for turbidity, was calculated as the difference between the weight of the dried filter and the empty dried filter and divided by the filtered volume (dry weight (mg L −1 ). Weighting was done on a microbalance with the precision of 0.02 mg L −1 [16].

FDOM (Fluorescent Dissolved Organic Matter)
This technique allows the obtaining of excitation-emission spectra as a function of the fluorescence intensity in a mixture of compounds. The Fluorescence Excitation-Emission Matrices of Fluorescence (EEMs: 3D) of the natural water samples were obtained using the Perkin Elmer LS 55 spectrofluorometer.
The acquisition method of the 3D fluorescence data was as follows: λEx 200-500 nm; λEm 280-550 nm with a 5 nm step for excitation and emission. A scanning rate of 1200 nm/min and a response time of 0.5 s [15,16]. The cell used was a quartz cell with a 1 cm optical path, previously washed with 10% HCl, rinsed three times with MQ water and calcined (500 • C, 4 h). Before starting the spectral data acquisition, the temperature of the sample had to be close to 20 • C. A thermostat bath set at 20 • C kept the tank and its container at this temperature throughout the acquisition. The acquisition of EEMs was carried out in triplicate. At the beginning of each series, a blank (MQ water) was also passed. The EEMs were normalized by measuring the intensity of the Raman peak from the 3D matrix to obtain Raman Unit (R.U.), using the excitation at 350 nm and dividing the EEMs by the area under the Raman peak. In this way, variations in the sensitivity of the instrument if any were corrected.

Statistical Analysis, Standardization, and Visualization
A total of 22 (N = 193) variables were compiled for the study. Maps of the spatial variability of select parameters from both the tested variables were visualized using Surfer 11 mapping software. Understanding the linkages of DOM sources with biogeochemical and meteorological parameters requires the standardization of data for further compilation methods. All variables from the dataset were standardized (mean of 0 and standard deviation of 1) using select statistical cluster classifier methods. The missing data were imputed by a chained equations algorithm [52] while the correlation and covariance matrices were compiled using cross-correlation, principal component analysis (PCA), and hierarchical cluster analysis (HCA). Results were presented as correlogram, dendrogram, boxplots, and PCA correlation circle. All numerical computations were analyzed using R program language software version 4.0. C1-type M peak (Ex = 300/Em = 420 nm) which is derived f from phytoplankton, two terrestrial-like humic compone = 420), a C4 humic-like (Ex = 280/Em = 510 nm) and a C3 ( like component.

Physical Conditions in Laucala Bay in December 2017
Salinity distributions (Table 2, Figure 3a (18.8) and near the vicinity of the mouth of the Vatuwaqa River. Generally, salinity is a measure of the total dissolved salts (TDS, see Table 2 for measured values) within the waters which means that for Laucala Bay, the TDS ( Figure 3d) is very much affected by the freshwater inputs from riverine discharges. This situation corresponds to an IRD cruise after rain in March 1999, when significant freshwater was exported from the Rewa River and produced a well-defined salinity gradient over the whole lagoon with salinity ranging from as low as 25.5 in the eastern part of Laucala Bay to a maximum of 35.5 in the western part of Suva Harbour [44]. orded in Laucala Passage (18.8) and near the vicinity of the mouth of the Vatuwaqa River. Generally, salinity is a measure of the total dissolved salts (TDS, see Table 2 for measured values) within the waters which means that for Laucala Bay, the TDS ( Figure 3d) is very much affected by the freshwater inputs from riverine discharges. This situation corresponds to an IRD cruise after rain in March 1999, when significant freshwater was exported from the Rewa River and produced a well-defined salinity gradient over the whole lagoon with salinity ranging from as low as 25.5 in the eastern part of Laucala Bay to a maximum of 35.5 in the western part of Suva Harbour [44].  Observations from both the surface and bottom temperature obtained from the CTD(depth profiler) and YSI (Yellow Springer Institute) Model 30 handheld water quality meter were similar (not shown). Surface temperature readings from the YSI meter recorded a minimum, maximum, mean of 26.9, 28, 27.5, and a standard deviation of 0.35 • C respectively (Table 2) which parallels previous [3] findings where the temperature was expected to be higher in the wet-warm season.
Dilution effects through the mixing of the surface and bottom layers could be attributed to causing these patterns as the surveyed stations were not so deep (5-30 m, Table 1). A mass of freshwater is thus observed via temperature distributions from the east to the west direction within the LBW (Figure 3b) which is consistent with the strong surface currents inflows (≈2 ms −1 and exacerbated by severe rain events [44]) from the RDW and Vunidawa river into the LBW through the Laucala passage. Dissolved oxygen, DO (Table 2, Figure 3c) which is essential to all marine life; exhibited a negative gradient offshore, that is, varies with salinity and TDS (Figure 3ad), and is consistent with previous findings [15,16].
The variability and distribution of the measured parameters (biogeochemical, absorption, and fluorescence) are summarized (Table 3) and compiled displayed as color maps (Figures 4 and 5). Slightly dissimilar patterns for Chla and SPM (Figure 4c,d) were observed. Chla had a maximum concentration in the CO (mean value of 3.3 ± 2 µg L −1 ) and IB (3.17 ± 2.33 µg L −1 ) and decreases in the OB (Table 3). SPM showed non significantly different concentrations in CO, IB, and OB (2 ± 0.4 mg L −1 , 1.70 ± 1.13 µg L −1 and 1.83 ± 1.13, Table 3). Chla and SPM had higher variability in the IB ( Figure 6D,F). SPM sources are more diverse in the CO (Figure 4a,b). This has been noted previously [2,3], as, during the rainy season, significant freshwater flow rates from the Rewa River plus the small rivers from the western coast of Laucala Bay strongly affected the distribution of SPM [44].            Table 3.  4 showed a patchier distribution. The minimum and maximum values of nutrient concentrations correspond to values observed during the rainy period at the BULA cruises [44,45].
Higher variability in concentrations was observed within the CO and IB stations and low variability in OB was observed (Figure 6A,C,G) for salinity, DO, chla and DON. SPM and DOC ( Figure 6F,H) displayed low variability in the CO contrasting higher variabilities for IB and OB stations. While temperature ( Figure 6B) displayed uniformity from the CO, IB, and OB; pheophytin ( Figure 6E) displayed higher variabilities in the CO and OB stations. It is striking that the highest percentage of pheophytin (Table 3) is observed in OB (30%, 14%, and 50% at CO, IB, and OB, respectively). This can be explained by the fact that waters flushed by fresh waters to the outside of the bay during decreasing tide are characterized by a high proportion of degraded phytoplankton.

CDOM Absorption and Fluorescence
CDOM absorption coefficients displayed large ranges in LBW for a CDOM (254) and a CDOM( 442) recording values from 1.57 to 1.73 and 0.48 to 0.12 m −1 respectively (Table 3). High variability in the coastal stations (CO) and low variability in the IB and OB ( Figure 6L Higher variability in concentrations was observed within the CO and IB stations and low variability in OB was observed (Figure 6a,c,g) for salinity, DO, chla and DON. SPM and DOC (Figure 6f,h) displayed low variability in the CO contrasting higher variabilities for IB and OB stations. While temperature (Figure 6b) displayed uniformity from the CO, IB, and OB; pheophytin (Figure 6e) displayed higher variabilities in the CO and OB stations. It is striking that the highest percentage of pheophytin (Table 3) is observed in OB (30%, 14%, and 50% at CO, IB, and OB, respectively). This can be explained by the fact that waters flushed by fresh waters to the outside of the bay during decreasing tide are characterized by a high proportion of degraded phytoplankton.   Distinctively, C1, C2, and C4 fluorescent components presented similar maps (Figure 8a,b,d), and recorded higher values at the west coast of the LBW near the vicinity of the mouths of Uluituni Creek and Samabula River. C3 presented a slightly different distribution in the South West of Laucala Bay with a lower value in front of Uluituni Creek. Lower values were recorded at the OB region within the reef edges and into the POW. Several identical distribution patterns were identified, one for the fluorescent components C2 and C4 (Figure 8b,d; Figure 6O,Q) displaying a higher concentration within the CO (mean values 0.15 ± 0.1 and 0.15 ± 0.09 01 R.U.) and IB (mean values 0.14 ± 0.06 and 0.14 ± 0.01 R.U.) region. Spatially distinguishable plumes at the western coast within the mouths of Uluituni Creek (in front of station 22) and Vatuwaqa River and a lower concentration at the OB (mean values 0.06 ± 0.01 and 0.06 ± 0.0201 R.U. respectively) were also observed.

CDOM Absorption and Fluorescence
The other two fluorescent components C1 and C3 ( Figure 6N,P) recorded less similar distributions displaying higher plumes at the west coast of LBW but slightly lower contributions (mean values 0.08 ± 0.03 and 0.07 ± 0.0101 R.U.) at the OB region. A notable feature observed here was the similarity (mean values of 0.11 R.U. for all contributions, Table 3) in distribution patterns for all fluorescent components ( Figure 6N  The other two fluorescent components C1 and C3 (Figure 6n,p) recorded less similar distributions displaying higher plumes at the west coast of LBW but slightly lower contributions (mean values 0.08 ± 0.03 and 0.07 ± 0.0101 R.U.) at the OB region. A notable feature observed here was the similarity (mean values of 0.11 R.U. for all contributions, Table 3 Figure 6R,S) with higher concentrations were observed at the CO and dispersed evenly into the IB which contrasts the low concentrations at the OB region for both variables.

Association and Dependence of Environment Variables
Cross-correlation coefficients for all the test variables (N = 22) are presented in the correlogram (Figure 9). Positively correlated variables are displayed in green pies (highly correlated are darker with smaller pie cuts) and negative in red pies. The correlation coefficient(r) is labeled inside the pie and the sizes of the circles are equivalent to the value of the coefficient. Distinctive correlation patterns displayed by the physical parameters, temperature recording negative correlations with all the other variables (biogeochemical, bio-optics, fluorescence, and nutrients) with notable significant correlation with DON, a CDOM (412)/a CDOM (442), and a CDOM (350). Salinity on the other hand also displayed a negative correlation with all variables. Significant correlations were observed with DO and Si(OH) 4 . For the biogeochemical variables, DOC was moderately correlated with DON while Chla recorded high correlations with a CDOM (254), all fluorescent components; and nutrients but moderate for NOx. Pheo was highly correlated with SPM, a CDOM (350), a CDOM (412), and aCDOM(442) while SPM was highly correlated with a CDOM (412), Si(OH) 4, and Pheo.   PCA was performed for the standardized dataset (N = 193) and imputed together with the climate variables: Sea level (SL), rainfall, and wind datasets (standardized) for the past three months before the survey and presented (Figure 10a) in a correlation circle.  Table 3. For the nutrients results, NOX was highly correlated with all CDOM spectral absorption components and three fluorescent components, namely; C1, C2, and C3. Si(OH) 4 and PO 4 had a very high correlation with DON. Si(OH) 4

Association with the Climate Variables-PCA Results
PCA was performed for the standardized dataset (N = 193) and imputed together with the climate variables: Sea level (SL), rainfall, and wind datasets (standardized) for the past three months before the survey and presented (Figure 10a) in a correlation circle. Quality representation (cos2) of the tested variables via the principal components, Dim 1 and Dim2 has seen low representations for DO, wind, and salinity (<0.25, arrows midway in the circle) contrasting higher values (>0.65, arrows reaching near the circle) for PO4, DON, Si(OH) 4

Association with DOM Sources-HCA Results
HCA was performed (k = 5, method = complete/furthest distance) with the most correlated (pies in Figure 9a) variables to trace linkages of tested variables with the sources of DOM obtained from the laboratory analytical results and presented (Figure 10b) as a hierarchical dendrogram. Two main clusters were identified, one enclosed in the light blue rectangle, the other is divided into 3 clusters: red, green, and dark blue. The light blue (single fork/branch) cluster comprises pheophytin while the red (double fork) cluster consists of TOC and DOC. The red, green, and dark blue are branched as a single big cluster.
The green cluster consists of SPM as a single fork/cluster and branches into another cluster which consists of aCDOM(254) as a single cluster which in turn branches into Chla and C3-protein-like as a separate (2 fork) cluster. The green and the dark blue cluster are branched as a single big cluster. On the other hand, the dark blue cluster consists of two separates clusters both cluster again branching out into two clusters.
One links fluorescent components together in a single cluster: C1-type M (maybe linked to pollution) as a single cluster which then branches (2 forks) into C2-fulvic terrestrial(origin) and C4-humic terrestrial(origin) as a separate cluster. The component cluster  4 . Sea level and temperature, however, displayed negative correlations with all the biogeochemical and nutrient parameters tested for PCA.

Association with DOM Sources-HCA Results
HCA was performed (k = 5, method = complete/furthest distance) with the most correlated (pies in Figure 9a) variables to trace linkages of tested variables with the sources of DOM obtained from the laboratory analytical results and presented (Figure 10b) as a hierarchical dendrogram. Two main clusters were identified, one enclosed in the light blue rectangle, the other is divided into 3 clusters: red, green, and dark blue. The light blue (single fork/branch) cluster comprises pheophytin while the red (double fork) cluster consists of TOC and DOC. The red, green, and dark blue are branched as a single big cluster.
The green cluster consists of SPM as a single fork/cluster and branches into another cluster which consists of a CDOM (254) as a single cluster which in turn branches into Chla and C3-protein-like as a separate (2 fork) cluster. The green and the dark blue cluster are branched as a single big cluster. On the other hand, the dark blue cluster consists of two separates clusters both cluster again branching out into two clusters.
One links fluorescent components together in a single cluster: C1-type M (maybe linked to pollution) as a single cluster which then branches (2 forks) into C2-fulvic terres-trial(origin) and C4-humic terrestrial(origin) as a separate cluster. The component cluster link with the nutrients (2 forks): Si (OH) 4 and PO 4 as the other cluster. Both the nutrients and component clusters make a big single cluster which branches with another big cluster branching in two clusters. One branch contains CDOM parameters as a single cluster: a CDOM (350) (single cluster), a CDOM (412), and a CDOM (442) (double fork cluster) linked to DON and NOx as another single cluster.
The measurement and distribution of various environment variables are needed to assess environmental constraints and to identify naturally or potentially disturbed major sources of terrigenous and anthropogenic inputs. Sources of CDOM could therefore be inferred from the absorption coefficients and their relation with the environment. This cruise could not sample the Rewa River at low tide. Therefore, spatial distributions obtained in December 2017 during this particular cruise do not include the inputs of the Rewa River after a strong rain [42] which allows estimating other sources in Laucala Bay.

CDOM Inputs Enhancement and Conservative Dilution by Heavy Rain
Rainfall impacts on riverine terrigenous outflows in tropical coastal lagoons in the Pacific have been documented in various studies [15,16]. Observation of high positive correlations of rainfall with DON (r = 0.74, p < 0.05), nutrients (PO 4 , Si(OH) 4 , NOx; 0.74 ≤ r ≤ 0.62, p < 0.05) and moderate positive correlations with C2-humic terrestrial, C4-fulvic terrestrial fluorescent components were recorded. Terrigenous CDOM enhanced by heavy rain is consistent with the results of previous studies [15,16]. The dilution effects of mixing of the oceanic waters masses from SHW and POW together with freshwater discharges from RDW, Uluituni Cr., Vatuwaqa, Samabula, Nasinu, Vunivadra channel, and Vunidawa (hereby termed riverine outflows) was evident from the surface salinity and temperature signals (positive gradients coast-offshore, Figure 3a,b) across the bay.
Periods of high rainfall events exacerbate the mixing effect [15], with a lower mean surface salinity (23.2 ± 0.33) recorded for LBW comparable to previous records of 24.8 during wet conditions [3]. The current study recorded even lower salinity (19.7 ± 2.82) values for CO and IB stations again highlighting the influence of freshwater riverine discharges into the LBW. Observations of significant negative correlations (Figure 3a) of salinity with most tested variables (e.g., C2-fulvic terrestrial and C4-fulvic terrestrial) indicate the terrestrial origins of CDOM [15]. Similarly, the type M component which may be linked to pollution also recorded a significant negative correlation with salinity which may also correspond to anthropogenic (terrestrial) sources. Thus, the negative correlations between salinity and the absorption coefficients of CDOM indicate that the conservative dilution of LBW through the riverine discharges plays a vital role in controlling the distribution of CDOM absorption and fluorescence [15,16].

In-situ CDOM from Biogenic Sources
Very weak negative correlations ( Figure 9) were observed between salinity and Chla (r = −0.33, p < 0.05); also with C3-(protein-like) component indicating possible links with the in-situ production within the bay or from oceanic origins [16] and through bottom re-suspensions [15]. Conspicuously, very high concentrations were observed for Chla concentrations in the CO (5.24 µg L −1 ) and IB (5.69 µg L −1 ) stations( Figure 6D) contrasting lower values observed in other tropical lagoons (e.g., the eastern lagoon of New Caledonia, 1.44 µg L −1 [16]). Owing to the high Chla biomass observations and its high correlation ( Figure 9, r = 0.76, p < 0.05) with C-3 protein-like component, an interpretation of in-situ CDOM production from phytoplankton biodegradation [15,16] was diagnosed. Confirmations of CDOM inputs from biogenic sources were visualized in the HCA results (Figure 10b green clusters) where Chla and C3protein-like-component were clustered as a single cluster. Chla also recorded high correlations with absorption coefficient a CDOM (254) (r = 0.72, p < 0.05) and SPM (r = 0.6, p < 0.05) which is indicative of possible linkages through re-suspension within the LBW. The high Chla biomass observed correlated very well with nutrients observations (Si(OH) 4 :PO 4 :NO x ) indicating phytoplankton plumes induced from nutrient runoffs into the LBW [43]. Nutrients concentrations reported earlier [4] were comparable to the findings of the current study (e.g., Mean NO x values recording 1.31 and 1.75 µM respectively) which is indicative of high terrestrial runoffs. Nevertheless concentrations of Si(OH) 4 were relatively high (mean = 78.98 µM). The nutrient levels recorded were 10 folds higher than those recorded in unpolluted waters in Fiji [3,4] which may also indicate (apart from riverine discharges) possible links to effluents discharge from KWTP [4,43].
3.6. South-West Plume Formation during Heavy Rain via the LBW Distributions (Figures 4-7) for most tested variables (includes all biogeochemical, CDOM absorptions, fluorescent components, and nutrients) displayed a peculiar pattern during heavy rain events. Plume formation observed at the west to northwest of the LBW along the Nasese passage is evident and is consistent with the surface current outflows through the Nasese passage to SHW during the wet season [44]. Strong surface currents (≥2 ms −1 ) inflows from the RDW into LBW from the north-west direction into the bay then west to south-west (SW) outflows via the Nasese passage into SHW. The moderate SW outflows (≤0.9-1.2 ms −1 ) through (combined with the geomorphological aspect of the bay and riverine outflows from the western coast) the narrow Nasese passage may partially explain the surface accumulation of most particulates (plume formation) at the west coast along T1. Particulate distributive patterns that display the plume formation outflows in the west coast of LBW include TOC, DON, Chla, all nutrients (PO 4 , NO x , Si(OH) 4 ), all CDOM absorptions coefficients, and all fluorescent components (C1-type M, C2-humic terrestrial, C3protein-like, C4-fulvic terrestrial). Sea level observations displayed significant negative correlations with almost all the biogeochemical and bio-optical variables (Figure 9a) indicating high oceanic influx during high water levels and high freshwater influx during low water levels.

Association of Terrigenous and Oceanic Inputs
PCA assessments (Figure 10a) reveal the various associations of the physical, biogeochemical, DOM absorption/fluorescence, and nutrient parameters to select climatic parameters (seal level winds and rainfall). Higher correlated variables were then demarcated through HCA (Figure 10b) to further detect their linkages. Oceanic inputs were easily identifiable via the association of Chla and C3protein-like components which were clustered together as once cluster (green rectangle in Figure 10b) together with SPM and absorption coefficient a CDOM (254). Association with sea level derived from PCA (Figure 10a) ropes in the mixing of oceanic inputs and terrestrial inputs via bottom re-suspension and tidal influx within the bay thus displayed by the significant correlation with SPM (a proxy for water clarity). SPM associations in the oceanic cluster might be an indication that origins of CDOM may differ (due to OM and microbial decomposition as in [15,16]) within the water column which exacerbates by rain events. Significant positive correlations of DO with rainfall and also with autochthonous/allochthonous CDOM (Figure 9a) settles the argument well [48]. Associations with optical properties of allochthonous CDOM were reproduced and visualized via HCA (Figure 10b) classifier method indicated by the blue cluster.
Fluorescent components of C2-fulvic like and C4-humic-like are clustered together with C1-type M indicative of their terrestrial (Figure 10b, blue rectangle) origins. Other clusters that may have linkages to the same sources include Si(OH) 4 and PO 4 . Absorption coefficients a CDOM (412), a CDOM (442), and a CDOM (350) branch out into another single cluster which also derives their terrestrial origins. NO x and DON branch out from the absorption cluster (Figure 10b) indicative of the similarities of their sources. Finally, the sources of the 4 different clusters (Figure 10b) within the terrestrial (blue) cluster may differ or the same warranting further studies for verification. Notably, so, the marine (green rectangle, Figure 10b) and terrestrial clusters (blue rectangle, Figure 10a) are again clustered together as one big cluster which may be indicative of mixing via tidal influx and bottom re-suspension combined with rain and wind derived riverine discharges [48]. Another may depict the parachotonous origins of CDOM within the LBW. Nutrients and metal additives, for example, in coastal waters through river outflows have also been reported to alter the in-situ ambient DOM pool [6]. Finally, the heterogeneity of CDOM origins in LBW arises from the interconnectivity through the transport of terrigenous inputs, decomposition of OM, and primary production activities within the interconnected biomes.

DOM Responses as Energy and Nutrient Sources
Due to the inherent molecular complexity of DOM which contains functional groups as well as essential nutrients such as nitrogen, N, and phosphorous, P [49][50][51][52]; it is not answering to quantify the portion of the DOM pool which contains organic nutrients N and P [53]. Instead, the concentration of organic nutrients must be calculated by difference as illustrated by [6,51]  ; where TON-total organic nitrogen; DIN-dissolved inorganic nitrogen). Nevertheless, since the biogeochemistry of DOM [16] is linked to the nutrients found in the coastal environment (e.g., NO x ) we can assess the responses of DOM through CDOM and DOC with the tested nutrients variables [54]. Similarly, examining how DOC and CDOM respond to the other environment variables (e.g., physical and climatic parameters) will portray a better insight into how DOM responds to the fluctuating environment and climate conditions [55]. The positive correlations of DOC with nutrients (NO x : r = 41, p < 0.05; PO 4 : r = 0.31, p < 0.05) elucidates DOM being utilized as energy sources while the negative correlation with Si(OH) 4 (r = −0.08, p < 0.05) signals DOM being the sources of nutrients in the LBW [6,54]. Conversely, the highly positive correlation of DON with the nutrient variables NO x (r = 0.93, p < 0.05); PO 4 (r = 0.74, p < 0.05) and moderate positive correlation with Si(OH) 4 (r = 0.46, p < 0.05) reflects the role of DON as source of nutrients for the LBW [6]. DON, therefore, serves as a nutrient (N) source for microalgae which may partially explain the high recordings of Chla (max = 5.69 µg L −1 ) in the Laucala Bay waters.

Conclusions
The instantaneous overview of the snapshots of optical fingerprints of CDOM variability during heavy rain events animating the superposition of daily disturbances within the LBW was investigated and presented for the first time. Four fluorescent intensities were identified, type M, linked to the degradation of organic matter; fulvic-like, protein-like, and humic-like. Type M, fulvic-like and humic-like were sourced from terrigenous origins while the protein-like component ties up with marine biogenic sources. The overall dynamics and nature of the biogeochemistry and bio-optical characteristics of the LBW are heavily dictated by the frequency and intensities of the climate variables tested (rainfall, wind, and SL). Notable findings include the formation of the south-west plume formation (heavy rain) for CDOM fluorescence and absorption parameters (and other biogeochemical parameters) characterized by strong inflows of riverine discharges into LBW and moderate outflows through the narrow Nasesese passage into SHW. The Suva Barrier Reefs system limits the dispersal of particulates into POW which complements their accumulation in the west of the bay.
The recorded high Chla values coincided with the accumulation of nutrients in the bay which may be sourced by terrigenous (riverine discharges) and anthropogenic inputs. A significant correlation of Chla and DO reveals primary production and decomposition within the marine biomes. Negative gradients of the shore for salinity and temperature signifies the conservative dilution of freshwater from riverine discharges which in turn is proportional to rainfall intensities. The tidal influx and bottom re-suspension are indicative for mixing as denoted by negative correlations with the tested variables and wind have a moderate influence on the spatial distribution patterns of the optical fingerprints of CDOM. Thus the findings provided baseline information for the biogeochemical dynamics and updates the physical and nutrients findings compiled for the Fiji waters database. DOM was being utilized as both energy and nutrients sources in the LBW as observed from the responses of DOC/CDOM to the nutrients variables NOx, PO 4, and Si (OH) 4 . Also, DON was seen as a major source of organic nutrients in the LBW as observed with its responses to the tested nutrient variables. The high nutrient loads (i.e., in the form of DON, etc.) transported by the rivers may partially explain the high phytoplankton (Chla) biomass within the Laucala bay coastal waters.  Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.
Data Availability Statement: All data needed to evaluate the conclusions in the paper are presented in the paper.