Spatial patterns and trends in the concentration and quality of dissolved organic matter (DOM) are characterised across a tropical agricultural catchment using ultraviolet (UV)-visible absorbance, and fluorescence spectroscopy. Related determination of the environmental isotopes 18O and 2H clarify the dynamics of catchment water movement. Water samples were collected from the Kinabatangan River, Borneo, and selected tributaries in August and September 2008 in four regions with oil palm plantations (KB1, KB2, KB3 and KB4). The isotopic compositions of surface waters suggest that canals were characterised by a strong evaporative effect than tributaries and streams with more natural, forested vegetation. DOM was characterised by variations in UV absorbance and spectral slope. Individual fluorescence excitation–emission matrices were decomposed by Parallel Factor Analysis (PARAFAC) and three components extracted (C1, C2 and C3). Components C2 and C3 both appear to be derived from microbial sources and/or photo-degradation. The PARAFAC components indicate a clear trend of increasing DOM degradation as waters pass through the catchment. It is hypothesised that upstream DOM is rapidly photo and microbially degraded to less fluorescent DOM, while DOM concentration and character of DOM downstream is controlled by the hydrology, specifically by variations in the rate of water movement.

INTRODUCTION

Tropical wetlands are ecologically diverse (Junk 2002; Dudgeon 2003) and characterised by rapid nutrient recycling and processing (Hader et al. 1998). Typically, wetlands function as a carbon store and possible carbon sink (Richey et al. 2002; Alkhatib et al. 2007; Limpens et al. 2008) and where closely integrated with fluvial systems, they constitute an important source of carbon for marine environments (Stephens & Rose 2005). Tropical wetlands alone have been estimated to contribute ∼60% of the total water, sediment and organic carbon input to the ocean globally (Alkhatib et al. 2007).

Dissolved organic matter (DOM) is ubiquitous in aquatic systems (Baker & Spencer 2004; Evans et al. 2005; Oliveira et al. 2006) but many wetlands are characterised by the presence of particularly high quantities of DOM (Mladenov et al. 2007; Stern et al. 2007). DOM fluxes from wetlands represent an important carbon input to river systems, constituting its largest and most bioavailable pool (Wilson & Xenopoulos 2009). The presence of organic matter affects the transport of organic pollutants, particle surface and colloid chemistry, photochemistry of natural waters and nutrient availability in freshwater systems (Hope et al. 1994; Fellman et al. 2008b), while DOM also contributes to chemical processes in natural water bodies by altering surface-water acidity, and affecting metal speciation and ion-exchange between the water and sediment phase.

Significantly, however, to date DOM has been relatively poorly studied in tropical catchments. Here, DOM characteristics are likely to change substantially as a result of photochemical degradation. This reflects greater penetration of ultraviolet (UV)-B radiation, mixing within the water body, and input of terrestrial material (Findlay & Sinsabaugh 1999; Spencer et al. 2009). DOM may also degrade directly (by structural alteration), or indirectly (by reaction with free radicals created by the application of light). In catchment headwaters, DOM is generally dominated by inputs of terrestrially derived DOM and dissolved organic carbon (DOC) concentrations are high (Dalzell et al. 2009). In contrast, the controls on DOM downstream are often related to hydrological processes, particularly the rate of water movement in large rivers (Findlay & Sinsabaugh 1999). Moreover, light penetration at downstream sites is often less and mixing depths greater, decreasing the effects of photolysis. Photochemical reactions produce inorganic carbon, low molecular weight organic compounds, trace gases, phosphorus- and nitrogen-rich compounds (Cory et al. 2007; Winter et al. 2007; Kowalczuk et al. 2009).

DOM is an important primary food source for aquatic food webs (Pace et al. 2004) and ecosystem metabolism (Bradley et al. 2007), however, DOM quality and quantity is likely to be affected significantly by catchment management, including land use change, wetland drainage, river channelisation and flow regulation. This is particularly evident in South and Southeast Asia where significant areas of tropical forest have been recently cleared for agriculture (Mattsson et al. 2000; Sidle et al. 2006; Atapattu & Kodituwakku 2009). In Indonesia, for example, it has been estimated that ∼45% of the original peat swamp forest has been lost (Rixen et al. 2008) following development of rubber (Hevea brasiliensis) and oil palm plantations (Elaeis guineensis) (Hooijer et al. 2006). In Malaysia, changes in land use have been encouraged by government (Abdullah & Nakagoshi 2006), leading to the recent (and rapid) increase in oil palm plantations in Sabah (Malaysian Borneo). Land conversion has been particularly extensive in East Sabah; however, the environmental implications of this expanding agro-forestry industry specifically on DOM quality and quantity has yet to be quantified. This is important, both for this region and elsewhere in the Tropics: the extent of oil palm has increased rapidly in recent years in Southeast Asia, reflecting the market for oil palm both for vegetable oil and as a bio-fuel. In 2008, oil palm cultivation was estimated at >13.5 × 106 ha (Fitzherbert et al. 2008), and there are increasing pressures to cultivate the crop in other agriculturally suitable areas in Africa and South America. However, at present, the hydrological impacts of converting tropical forest to agriculture are relatively poorly understood (particularly with respect to DOM). On the Tekam River, Pahang (Figure 1), the removal of forest and development of oil palm plantations led to a significant increase in annual sediment loads from 28 t km−2 (mean of three years) to 414 t km−2 (Douglas 1999). In the same catchment, a large and sustained increase in annual runoff was observed after the fourth year of planting (Chuan 2003). For many reasons therefore, it is essential that the environmental impacts of oil palm are fully understood by studying impacted catchments in Southeast Asia so that any effects may be mitigated elsewhere.

Figure 1

The Lower Kinabatangan River catchment (above right and centre) and its regional location (top), with local site maps for: (a) Batu Putih (KB1); (b) Sukau (KB2); (c) Bilit (KB3); and (d) Abai (KB4).

Figure 1

The Lower Kinabatangan River catchment (above right and centre) and its regional location (top), with local site maps for: (a) Batu Putih (KB1); (b) Sukau (KB2); (c) Bilit (KB3); and (d) Abai (KB4).

This urgent need is addressed in this paper by characterising DOM in a tropical catchment which is currently experiencing rapid agricultural change from the logging of primary and secondary forest and the on-going development of oil palm plantations. The objectives of the paper are to characterise and interpret spatial patterns and trends in DOM (both concentration and quality) across a tropical catchment using UV-visible absorbance and fluorescence spectroscopy and infer differences in water movement through the catchment using environmental isotopes (18O and 2H).

MATERIALS AND METHODS

Study sites

DOM characteristics and stable isotopic variability were determined in selected downstream reaches of the Kinabatangan River and tributaries in Sabah, Malaysia. The Kinabatangan River (560 km in length) is the largest river in Sabah, with a catchment area of 16,800 km2, representing 23% of Sabah (Figure 1; Josephine et al. 2004). The area has a humid tropical climate with mean daily temperatures ranging from 22 to 32 °C and mean annual rainfall of between 2,500 and 3,000 mm (Boonratana 2000; Josephine et al. 2004). Heaviest rainfall occurs during the northeast monsoon between October and March at which time the floodplain and coastal plain are widely inundated and there is considerable inter-annual variability in rainfall.

The lower floodplain of the Kinabatangan extends over >2,800 km2, and is the largest wetland in Sabah (Figure 1). There are ∼20 ox-bow lakes in three main clusters: (i) above Batu Putih; (ii) between Batu Putih and Bilit; and (iii) between Bilit and Sukau, with a dense network of tributary streams and creeks that extend into an estuary downstream. The natural floodplain vegetation mainly comprises riverine and freshwater swamp forest, with some open reed swamp, and pristine lowland dipterocarp forest in areas that are not frequently inundated (Boonratana 2000). Although the catchment is relatively sparsely populated, it has been estimated that ∼26% of the catchment has been developed for oil palm plantations, mainly in the floodplain of the Lower Kinabatangan River (Josephine et al. 2004).

The Kinabatangan River is the main source of water for local communities; some river-water is also abstracted and piped to the city of Sandakan to the north (Boonratana 2000; Mansourian et al. 2003). Long-term monthly mean discharge for three gauging stations in Kinabatangan, Pagar (PGR), Balat (BLT) and Barek Manis (BM), are summarised in Figure 2. From 1979 to 2013, monthly mean river flow in the upper catchment (stations PGR and BLT) (Figure 1) is within the range 14.1–1,285.8 and 26.1–1,944.2 m3/s, respectively. Peak daily mean discharge was recorded in January 1986 at Balat (∼1,944 m3/s), and the lowest at Pagar in June 1998 (∼14 m3/s). A water quality study indicated that total suspended solids in the Lower Kinabatangan catchment, in Sukau, was Class IIB (max. conc. of 50 mg/L) according to the Malaysian Interim National Water Quality Standard (Harun 2006). This probably reflects widespread intensive commercial logging over much of the catchment since the 1980s and the increasing development of oil palm plantations (Boonratana 2000; Josephine et al. 2004).

Figure 2

Long-term mean monthly flow discharge for three gauging station at Kinabatangan: (a) PGR; (b) BLT; (c) BM. The catchment area for each station is 9,430, 10,800 and 12,300 km2, respectively. Arrows indicate flow discharges during fieldwork campaign.

Figure 2

Long-term mean monthly flow discharge for three gauging station at Kinabatangan: (a) PGR; (b) BLT; (c) BM. The catchment area for each station is 9,430, 10,800 and 12,300 km2, respectively. Arrows indicate flow discharges during fieldwork campaign.

Field-work design

Water samples were collected during summer baseflow conditions at selected locations in the Kinabatangan catchment in August and September 2008. Sampling design was significantly constrained by the difficulty of access, but the intention was to sample waters along a freshwater estuarine gradient and to capture the land use variations of DOM in the Lower Kinabatangan floodplain towards the Sulu Sea downstream. The nearest gauging station was at BM with an upstream catchment of 12,300 km2, which was located about ∼11 km from KB1-3 (Figure 1). The daily mean discharges observed during the fieldwork campaign are summarised in Table 1.

Table 1

Sampling date for each station and mean daily discharge at the nearest gauging station (BM) recorded by the Department of Irrigation and Drainage Malaysia

Sampling date (dd/mm/yy) Sampling station Average of daily mean discharge at BM (m3/s) 
23/08/08 Sg. Pin (KB1-1) 209.5 
23/08/08 Sg. Koyah (KB1-2) 209.5 
24/08/08 BS Mill (KB1-3) 204.5 
25/08/08 PS Plantation (KB1-4) 195.0 
25/08/08 Canals (KB1-6 to 8) 195.0 
27/08/08 DGFC (KB1-5) 260.2 
28/08/08 Sg. Resang (KB2-1) a 
28/08/08 Sg. Resik (KB2-2) a 
28/08/08 Kuala Sukau (KB2-3) a 
29/08/08 Balat Damit (KB4-1) a 
29/08/08 Sg. Merah (KB4-2) a 
02/09/08 Sg. Menanggol (KB2-5) 213.7 
02/09/08 Sg. Tenagang Besar (KB3-1) 213.7 
02/09/08 Sg. Tenagang Kecil (KB3-2) 213.7 
02/09/08 Malbumi Plantation (KB2-4) 213.7 
Sampling date (dd/mm/yy) Sampling station Average of daily mean discharge at BM (m3/s) 
23/08/08 Sg. Pin (KB1-1) 209.5 
23/08/08 Sg. Koyah (KB1-2) 209.5 
24/08/08 BS Mill (KB1-3) 204.5 
25/08/08 PS Plantation (KB1-4) 195.0 
25/08/08 Canals (KB1-6 to 8) 195.0 
27/08/08 DGFC (KB1-5) 260.2 
28/08/08 Sg. Resang (KB2-1) a 
28/08/08 Sg. Resik (KB2-2) a 
28/08/08 Kuala Sukau (KB2-3) a 
29/08/08 Balat Damit (KB4-1) a 
29/08/08 Sg. Merah (KB4-2) a 
02/09/08 Sg. Menanggol (KB2-5) 213.7 
02/09/08 Sg. Tenagang Besar (KB3-1) 213.7 
02/09/08 Sg. Tenagang Kecil (KB3-2) 213.7 
02/09/08 Malbumi Plantation (KB2-4) 213.7 

aData are unavailable.

Waters were sampled from sites that were closely associated with oil palm plantations and processing mills, as well as the main-stem of the Kinabatangan and tributary streams with varying proportions of oil palm plantation. Sample sites were initially selected according to their accessibility (within 3 hours by boat) and included: (i) the main stem of the Lower Kinabatangan River (26 samples); (ii) tributary streams (180 samples); and (iii) irrigation ditches associated with selected oil palm plantations (28 samples).

In total, 234 water samples were collected in 2 mL glass vials (in most cases from a boat) at three points across the river channel. Given uncertainties as to whether river waters were completely mixed, at each sampling point waters were sampled at points across the channel: adjacent to both riverbanks and in the channel mid-point. The small sample sizes also enabled rapid shipment of the water samples (unfrozen) to Birmingham, UK, for laboratory analysis. There are a number of logistical problems of sample storage and preservation when working in remote areas (e.g., Burke et al. 2002) although freezing has been applied successfully (Fellman et al. 2008a). In this case, freezing of samples was not feasible, however, nutrient concentration can be stabilised using inorganic acids (Harmel et al. 2006). Other physicochemical parameters, including pH and water temperature, were measured in situ using a Hanna multi-parameter water quality meter (Model HI9828). All samples for organic matter fluorescence analysis were filtered through Whatman GF/F syringe filters (nominal pore size 0.7 μm), and stored in the dark at 4 °C until analysis, which occurred within 1 month of sampling. A comparison of the effects of filtering in the field vs. filtering post-collection using triplicate grab samples demonstrated that the fluorescence of post-collection filtered samples was more variable than for field-filtered samples (<9% coefficient of variance for Parallel Factor Analysis (PARAFAC) C1 and <20% for C2 and C3 vs. <6% for all three components). This variability is taken into consideration in the subsequent interpretation.

Surface water samples were collected from different water body types as follows: (i) tributary streams (∼2–3 km above their confluence) (ST); (ii) effluent from oil palm plantation ditches (DC); (iii) main stem of the Kinabatangan (MS); and (iv) ox-bow lake. Ditches from oil palm plantations were found to have less vegetation cover, compared to other water body types. Sample sites have been grouped into four principal sampling regions (based on upstream–downstream gradient) (Figure 1):

  1. Batu Putih (KB1): samples were collected from tributary streams (KB1-1, KB1-2; 39 samples), effluent from an oil palm plantation mill and ditches (KB1-3, KB1-4, KB1-6 to KB1-8; 19 samples) and an ox-bow lake (KB1-5; 1 sample). The local population is ∼1,200–1,400, in a number of small settlements, along the river.

  2. Sukau (KB2): samples were collected from tributary streams (KB2-1, KB2-2, KB2-3, KB2-5; 63 samples), a drainage ditch associated with an oil palm plantation (KB2-4; 9 samples) and the main river (KB2-1, KB2-2, KB2-3, KB2-4, KB2-5; 15 samples). Sukau lies on the Kinabatangan River, 70 km upstream of Sandakan. The population of ∼2,000 primarily inhabit small settlements scattered along the riverbank. In places, untreated sewage discharges into the river, but some areas of riparian forest remain in the vicinity.

  3. Bilit (KB3): samples were collected from streams (KB3-1, KB3-2; 27 samples) and the main river (KB3-1, KB3-2; 6 samples). This area is sparsely populated (∼300 inhabitants) and more riparian forest remains here. Characteristic vegetation includes three forest types: seasonal freshwater swamp, mixed dipterocarp forest (MDF) and MDF-limestone forest (Sabah Forestry Department 2001). The latter vegetation is associated with limestone outcrops that are also found in this area.

  4. Abai (KB4): samples were collected from streams (KB4-1, KB4-2; 51 samples) and the main river (KB4-1, KB4-2; 5 samples). Abai is relatively close to the estuary of the Sulu Sea, near a recently designated (2008) Ramsar wetland: Kinabatangan – Lower Segama. Although there are small settlements in this area, the population in 1996 was estimated at only ∼290 (Payne 1996). Freshwater swamp vegetation, including Nypa fruticans, is widespread along the estuary.

Laboratory analyses

Stable isotope analyses

Stable isotope analyses were undertaken in the Water Sciences laboratory at the University of Birmingham, UK, using a GV Instruments Isoprime isotope-ratio mass spectrometer (IRMS) connected to a Eurovector elemental analyser. Stable isotope values are expressed using the δ convention, where δ18O = (18O/16Osample)/(18O/16Ostandard)–1, and similarly for hydrogen isotopes (δ2H) expressed as ‰ (per mil) where the standard is Vienna standard mean ocean water (V-SMOW). For hydrogen isotope analysis, ∼0.3 μL of water was injected from sample vials on an autosampler into a column where reduction to hydrogen took place at 1,050 °C over a chromium metal catalyst. At least two successive analyses were made by repeat injections from the same vial. Internal (within-run) precision is 0.4 per mil for δD and overall (external) precision is ±1 per mil.

Oxygen isotope analyses were undertaken using an equilibration technique. 200 μL water samples were left to equilibrate (CO2) in a sealed container for 7 hours allowing the headspace CO2 to assume the δ18O composition of the water. The equilibrated CO2 was then analysed on the IRMS. The internal precision for δ18O is typically 0.08 per mil, external precision is better than 0.15 per mil.

Fluorescence spectroscopy and UV absorbance

DOM was characterised by fluorescence spectroscopy and UV-visible absorbance. Fluorescence intensity was measured using a Varian Cary Eclipse spectrophotometer equipped with a Peltier temperature controller. Emission scans were performed for wavelengths from 280 to 500 nm, with data collected at 2 nm intervals, and excitation wavelengths from 250 to 400 nm, at 5 nm intervals. Uncorrected spectra were combined to form an excitation–emission matrix (EEM). Water sample temperature was set to 20 °C and photomultiplier tube voltage was 725 V. Spectrophotometer output was monitored by regular determination of the Raman intensity of ultra-pure water in a sealed 10 × 10 mm cuvette at 348 nm excitation and 5 nm bandpass. A 1.5 mL (10 × 3.33 mm) cell was used to minimise both the sample volumes required. Due to the turbidity of the water samples, which remained in some samples even after filtration, manufacturer-supplied excitation filters (250–395 nm) were employed. An inner-filter effects correction has been applied to the data set following Ohno (2002): 
formula
1
where I is detected fluorescence intensity; I0 is fluorescence in the absence of self-absorption; b is the path length for both the excitation and emission beam; Aex is absorbance at excitation wavelength; and Aem is the absorbance at emission wavelength.

Absorption coefficients at 254 and 340 nm (a254 and a340) and spectral slope over the interval of 275–295 nm (S275−295) (Helms et al. 2008) were determined using a Lightwave spectrophotometer and 2 ml (10 × 5 mm; 5 mm path length) cuvettes. S275−295 was calculated by linear regression of the log-transformed a spectra. Distilled deionised water was used as a reference, and absorbance readings were corrected, where necessary, for long-term baseline drift. An inner-filter correction was also applied (mean UV absorption coefficient across all samples at a340 was 37.6 m−1).

Data analysis

Parallel factor analysis

PARAFAC was undertaken to separate EEMs statistically into their individual, underlying, fluorescent groups with specific excitation and emission spectra. This analysis provides a qualitative (fluorescent groups) and quantitative (fluorescent intensity of each group) model of the data and is ideally suited to detecting small, but potentially significant, differences in DOM composition.

PARAFAC modelling was undertaken following Stedmon & Bro (2008). Fluorescence EEMs were combined into a three-dimensional data array and decomposed to a set of trilinear terms and a residual array 
formula
2
where xijk is the fluorescence intensity for sample i at emission wavelength j and excitation wavelength k; aif, bif and ckf are the loading matrices. F is the number of model components, and eijk is residual noise (i.e., variability not explained by the model). Despite the use of a 250–395 nm excitation filter, initial PARAFAC analysis was confounded by scatter in the EEMs within the wavelength range of 250–280 nm (excitation) and 280–290 nm (emission). As a result, excitation wavelengths <290 nm were removed for PARAFAC analysis, and consequently it was not possible to investigate microbial fluorescence located at excitation wavelength 280 nm.
The model was defined and validated by split-half technique (Stedmon et al. 2003) in which the data array was divided into two halves and modelled separately. PARAFAC models, ranging from two to seven components, were derived for both data sets independently. Model validation was carried out by comparing the spectral shape of the components derived by the models. The model returns relative intensities of derived components (scores) because the specific absorption and quantum yield of fluorescence of individual components is unknown. The intensity of the nth component in a given sample, In, was calculated as the fluorescence intensity at the peak excitation and emission maximum of the nth component using 
formula
3
where Scoren is the relative intensity of the nth component, Exn (λmax) is the maximum excitation loading of the nth component, Emn (λmax) is the maximum emission loading of the nth component derived from the model. The total fluorescence intensity of a given sample was calculated as the sum of the components present in the samples 
formula
4

RESULTS

Stable isotope analyses

The stable isotope analyses (δ18O and δ2H) are summarised in Table 2 and Figure 3. The global isotopic composition of rainfall defines the global meteoric water line (GMWL, δ2H = 8 × δ18O + 10), while local, or regional, meteoric water lines have slightly different slopes and intercepts and can be determined from local precipitation. Araguás-Araguás et al. (1998) established a regional meteoric water line (RMWL) for neighbouring Sarawak (∼400 km from Kota Kinabatangan) (Figure 1), which provides a reference for this study (Table 2).

Table 2

Summary isotope data for the Lower Kinabatangan River catchment (standard deviations in parentheses) and comparison with data from Niah in Sarawak (from Stephens & Rose 2005)

  No. of samples       
Sampling station   From main river Sampling date (dd/mm/yy) pH Water temperature (°C) δ18O (‰) V-SMOW δ2H (‰) V-SMOW d-Excess (‰) 
Rainfall events              
Batu Putih (KB1):              
Rainfall (KB1-5) a 25/08/08 a a a a −4.0 a −25.4 a 3.9 a 
Sukau (KB2):              
Rainfall 1 (KB2-6) a 27/08/08 a a a a −2.5 (0.1) −7.9 (0.41) 10.7 (0.7) 
Rainfall 2 (KB2-6) a 28/08/08 a a a a −2.9 a −7.8 a 13.2 a 
Rainfall 3 (KB2-6) a 29/08/08 a a a a −6.5 a −36.5 a 11.3 a 
Rainfall 4 (KB2-6) a 30/08/08 a a a a −9.7 a −60.8 a 10.4 a 
Tributary streams              
Batu Putih (KB1):              
Sg. Pin (KB1-1) a 23/08/08 7.9 (0.3) 30.6 (0.7) −6.8 (0.1) −46.6 (0.44) 3.2 (0.5) 
Sg. Koyah (KB1-2) a 23/08/08 6.9 (0.3) 29.6 (2.9) −7.3 (0.1) −49.3 (0.25) 4.6 (0.6) 
DGFC (KB1-5) a 27/08/08 5.0 (0.04) 25.9 a −6.3 a −43.9 a 2.3 a 
Sukau (KB2):              
Sg. Resang (KB2-1) 28/08/08 6.5 (0.5) 29.1 (0.6) −7.9 (0.1) −50.0 (0.86) 8.4 (0.6) 
Sg. Resik (KB2-2) 28/08/08 7.4 (0.1) 30 (0.8) −7.9 (0.1) −51.0 (0.16) 7.2 (0.5) 
Kuala Sukau (KB2-3) 28/08/08 7.5 (0.1) 29.4 (0.6) −7.9 (0.1) −50.9 (1.0) 7.5 (1.3) 
Sg. Menanggol (KB2-5) 02/09/08 6.4 (0.04) 27.4 (1.1) −7.1 (0.2) −48.0 (1.23) 4.6 (0.6) 
Bilit (KB3):              
Sg. Tenagang Besar (KB3-1) 02/09/08 6.4 (0.2) 28.4 (0.4) −6.2 (0.5) −43.5 (4.59) 2.5 (1.7) 
Sg. Tenagang Kecil (KB3-2) 02/09/08 6.4 (0.03) 27.8 (0.4) −6.4 (0.8) −43.6 (4.58) 3.6 (1.0) 
Abai (KB4):              
Balat Damit (KB4-1) 11 29/08/08 6.2 (0.5) 28.5 (0.4) −7.7 (0.4) −49.7 (2.59) 6.9 (0.8) 
Sg. Merah (KB4-2) 29/08/08 6.5 (0.2) 29 (0.6) −7.8 (0.1) −50.7 (0.3) 6.5 (0.6) 
Ditches              
BS Mill (KB1-3) a 24/08/08 4.9 a 31.2 a −6.7 a −43.9 a 5.1 a 
PS Plantation (KB1-4) a 25/08/08 4.9 (0.2) 27.9 (0.9) −7.2 (0.6) -47.2 (3.3) 6.1 (1.3) 
Canal 1 (KB1-6) a 25/08/08 5.7 a 29.2 a −5.1 a −39.7 a -2.5 a 
Canal 2 (KB1-7) a 25/08/08 7.0 a 28.5 a −6.7 a −43.8 a 5.2 a 
Canal 3 (KB1-8) a 25/08/08 6.0 a 28.9 a −7.0 a −44.4 a 7.0 a 
Malbumi Plantation (KB2-4) 02/09/08 6.5 (0.1) 28.7 (0.5) −7.4 (0.1) −49.0 (0.66) 5.6 (0.6) 
Data from Niah (from Stephens & Rose 2005          
S. Niah 15 a 26/04/01 a a a a −7.9 (0.3) −49.0 (4.1) 13.5 (2.9) 
S. Niah a 27/04/01 a a a a −8.3 a −52.4 a 13.0 a 
South China Sea a 26/04/01 a a a a −4.6 a −29.3 a 7.0 a 
K. Niah a 26/04/01 a a a a −6.9 (3.1) −33.0 (19.1) 6.8 (5.0) 
G. Kira cave interior drip a 23/04/01 a a a a −6.7 (0) −38.9 (0.6) 14.2 (0.7) 
Rain outside West Mouth a 25/04/01 a a a a −10.8 (0.1) −71.0 (0.5) 14.2 (0.8) 
G. Kira cave mouth drip a 23/04/01 a a a a −8.2 (0) −53.3 (1.1) 11.8 (1.2) 
Rain outside Gan Kira a 23/04/01 a a a a −4.8 a −26.0 a 11.6 a 
Rainforest pool a 27/04/01 a a a a −7.3 a −44.6 a 13.3 a 
Rainforest stream (S. Subis) a 27/04/01 a a a a −8.4 a −54.7 a 12.1 a 
  No. of samples       
Sampling station   From main river Sampling date (dd/mm/yy) pH Water temperature (°C) δ18O (‰) V-SMOW δ2H (‰) V-SMOW d-Excess (‰) 
Rainfall events              
Batu Putih (KB1):              
Rainfall (KB1-5) a 25/08/08 a a a a −4.0 a −25.4 a 3.9 a 
Sukau (KB2):              
Rainfall 1 (KB2-6) a 27/08/08 a a a a −2.5 (0.1) −7.9 (0.41) 10.7 (0.7) 
Rainfall 2 (KB2-6) a 28/08/08 a a a a −2.9 a −7.8 a 13.2 a 
Rainfall 3 (KB2-6) a 29/08/08 a a a a −6.5 a −36.5 a 11.3 a 
Rainfall 4 (KB2-6) a 30/08/08 a a a a −9.7 a −60.8 a 10.4 a 
Tributary streams              
Batu Putih (KB1):              
Sg. Pin (KB1-1) a 23/08/08 7.9 (0.3) 30.6 (0.7) −6.8 (0.1) −46.6 (0.44) 3.2 (0.5) 
Sg. Koyah (KB1-2) a 23/08/08 6.9 (0.3) 29.6 (2.9) −7.3 (0.1) −49.3 (0.25) 4.6 (0.6) 
DGFC (KB1-5) a 27/08/08 5.0 (0.04) 25.9 a −6.3 a −43.9 a 2.3 a 
Sukau (KB2):              
Sg. Resang (KB2-1) 28/08/08 6.5 (0.5) 29.1 (0.6) −7.9 (0.1) −50.0 (0.86) 8.4 (0.6) 
Sg. Resik (KB2-2) 28/08/08 7.4 (0.1) 30 (0.8) −7.9 (0.1) −51.0 (0.16) 7.2 (0.5) 
Kuala Sukau (KB2-3) 28/08/08 7.5 (0.1) 29.4 (0.6) −7.9 (0.1) −50.9 (1.0) 7.5 (1.3) 
Sg. Menanggol (KB2-5) 02/09/08 6.4 (0.04) 27.4 (1.1) −7.1 (0.2) −48.0 (1.23) 4.6 (0.6) 
Bilit (KB3):              
Sg. Tenagang Besar (KB3-1) 02/09/08 6.4 (0.2) 28.4 (0.4) −6.2 (0.5) −43.5 (4.59) 2.5 (1.7) 
Sg. Tenagang Kecil (KB3-2) 02/09/08 6.4 (0.03) 27.8 (0.4) −6.4 (0.8) −43.6 (4.58) 3.6 (1.0) 
Abai (KB4):              
Balat Damit (KB4-1) 11 29/08/08 6.2 (0.5) 28.5 (0.4) −7.7 (0.4) −49.7 (2.59) 6.9 (0.8) 
Sg. Merah (KB4-2) 29/08/08 6.5 (0.2) 29 (0.6) −7.8 (0.1) −50.7 (0.3) 6.5 (0.6) 
Ditches              
BS Mill (KB1-3) a 24/08/08 4.9 a 31.2 a −6.7 a −43.9 a 5.1 a 
PS Plantation (KB1-4) a 25/08/08 4.9 (0.2) 27.9 (0.9) −7.2 (0.6) -47.2 (3.3) 6.1 (1.3) 
Canal 1 (KB1-6) a 25/08/08 5.7 a 29.2 a −5.1 a −39.7 a -2.5 a 
Canal 2 (KB1-7) a 25/08/08 7.0 a 28.5 a −6.7 a −43.8 a 5.2 a 
Canal 3 (KB1-8) a 25/08/08 6.0 a 28.9 a −7.0 a −44.4 a 7.0 a 
Malbumi Plantation (KB2-4) 02/09/08 6.5 (0.1) 28.7 (0.5) −7.4 (0.1) −49.0 (0.66) 5.6 (0.6) 
Data from Niah (from Stephens & Rose 2005          
S. Niah 15 a 26/04/01 a a a a −7.9 (0.3) −49.0 (4.1) 13.5 (2.9) 
S. Niah a 27/04/01 a a a a −8.3 a −52.4 a 13.0 a 
South China Sea a 26/04/01 a a a a −4.6 a −29.3 a 7.0 a 
K. Niah a 26/04/01 a a a a −6.9 (3.1) −33.0 (19.1) 6.8 (5.0) 
G. Kira cave interior drip a 23/04/01 a a a a −6.7 (0) −38.9 (0.6) 14.2 (0.7) 
Rain outside West Mouth a 25/04/01 a a a a −10.8 (0.1) −71.0 (0.5) 14.2 (0.8) 
G. Kira cave mouth drip a 23/04/01 a a a a −8.2 (0) −53.3 (1.1) 11.8 (1.2) 
Rain outside Gan Kira a 23/04/01 a a a a −4.8 a −26.0 a 11.6 a 
Rainforest pool a 27/04/01 a a a a −7.3 a −44.6 a 13.3 a 
Rainforest stream (S. Subis) a 27/04/01 a a a a −8.4 a −54.7 a 12.1 a 

aData are unavailable.

Figure 3

Plot of δ18O versus δ2H for sampling sites at the Lower Kinabatangan River catchment, and comparison with the regional meteoric water line (RMWL) and the meteoric water line for local precipitation.

Figure 3

Plot of δ18O versus δ2H for sampling sites at the Lower Kinabatangan River catchment, and comparison with the regional meteoric water line (RMWL) and the meteoric water line for local precipitation.

The δ18O of meteoric water in KB2 ranged from −9.7 to −2.5‰ while δ2H values ranged from −61 to −7.8‰, all lying parallel to, but slightly above, the RMWL. Precipitation on 29 August 2008 has a deuterium-excess of +11.3 compared to the RMWL published value of 9.2, indicating a slightly heavier vapour source for these short-term events compared to the longer-term mean. The large absolute range of values observed in Kinabatangan can be explained by the fact that the rainfall samples were not representative of complete events, but only a sub-sample.

Optical parameters

The results of the summer base-flow survey are presented here using descriptive statistics to identify trends in fluorescence and UV-visible absorption of waters sampled across the four study regions of the Lower Kinabatangan River catchment.

UV-visible absorption coefficients at 254 and 340 nm are summarised in Table 3, and varied considerably between sites. UV absorption has been found to correlate with DOC (Ahmad & Reynolds 1999; Tipping et al. 2009), and given the lack of direct DOC analyses (due to the small sample size) the results are interpreted here as indicative of relative variations in DOC. The greatest variability in absorption coefficients at 254 (340) nm were found at sites directly associated with oil palm plantations: KB1-6 to 8: 142.5 (47.9) m−1 and KB2-4: 123.7 (59.8) m−1.

Table 3

Summary mean of absorbance, spectral slope and PARAFAC data for the Lower Kinabatangan River catchment (standard deviation in parentheses)

  Absorption Absorption  PARAFAC components 
Sampling station No. of samples coefficient a254 (m−1coefficient a340 (m−1Spectral slope (nm−1Itot IC1 IC2 IC3 
Main stem 26 99.3 (21.4) 46.2 (12.2) 0.001 15.2 (1.2) 5.1 (0.6) 6.1 (0.7) 3.9 (0.7) 
Tributary streams           
Batu Putih (KB1):           
Sg. Pin (KB1-1) 21 97.6 (3.5) 33.6 (1.8) 0.013 62.5 (9.1) 20.4 (3.9) 28.1 (5.6) 14 (5) 
Sg. Koyah (KB1-2) 18 82.1 (3.4) 28.7 (1.4) 0.013 40.7 (4.8) 13.6 (1.7) 18.7 (2.3) 8.4 (1.8) 
DGFC (KB1-5) 62.2 a 20.3 a 0.019 27.0 (a5.9 (a9.1 (a12.0 (a
Sukau (KB2):           
Sg. Resang (KB2-1) 27 81.9 (3.8) 34.8 (1.8) 0.011 17.4 (5.4) 6.3 (2.5) 7.1 (2.2) 4.0 (1) 
Sg. Resik (KB2-2) 12 90.0 (1.7) 40.1 (1.1) 0.010 14.6 (0.8) 5.1 (0.3) 6.1 (0.5) 3.5 (0.3) 
Kuala Sukau (KB2-3) 18 86.8 (2.2) 40.6 (1.2) 0.009 15.1 (0.8) 5.3 (0.5) 6.4 (0.4) 3.4 (0.3) 
Sg. Menanggol (KB2-5) 18 106.8 (8.4) 47.6 (6.9) 0.010 23.9 (6.6) 9.6 (3.1) 10.1 (2.9) 4.2 (0.7) 
Abai (KB4):           
Balat Damit (KB4-1) 36 92.3 (10.5) 39.5 (3.7) 0.010 21.8 (8.8) 9.0 (4.9) 8.9 (3.7) 3.9 (0.7) 
Sg. Merah (KB4-2) 21 82.1 (6.3) 34.0 (2.7) 0.012 14.2 (1.8) 4.9 (0.2) 5.8 (0.7) 3.5 (1.7) 
Bilit (KB3):           
Sg. Tenagang Besar (KB3-1) 21 101.0 (5.1) 39.1 (3.9) 0.013 36.7 (11.8) 11.7 (3.7) 17.0 (5.7) 8.0 (3) 
Sg. Tenagang Kecil (KB3-2) 12 73.9 (11.0) 30.3 (6.5) 0.011 23.1 (5.3) 8.2 (2) 10.3 (2.7) 4.6 (0.8) 
Ditches           
BS Mill (KB1-3) 89.2 (2.2) 34.2 (2.0) 0.012 41.3 (0.5) 13.6 (0.5) 18.3 (0.3) 9.5 (1.0) 
PS Plantation (KB1-4) 13 81.6 (5.2) 30.9 (2.0) 0.013 26.5 (11) 8.7 (4.2) 10.8 (5.2) 7.1 (1.8) 
Canals (KB1-6 to 8) 142.5 (39.8) 47.9 (15.0) 0.013 57.4 (20.1) 24.4 (11.9) 24.4 (8.4) 8.6 (2.7) 
Malbumi Plantation (KB2-4) 12 123.7 (14.5) 59.8 (9.0) 0.008 17.3 (4.1) 6.3 (1.8) 7.1 (1.8) 4.0 (0.5) 
  Absorption Absorption  PARAFAC components 
Sampling station No. of samples coefficient a254 (m−1coefficient a340 (m−1Spectral slope (nm−1Itot IC1 IC2 IC3 
Main stem 26 99.3 (21.4) 46.2 (12.2) 0.001 15.2 (1.2) 5.1 (0.6) 6.1 (0.7) 3.9 (0.7) 
Tributary streams           
Batu Putih (KB1):           
Sg. Pin (KB1-1) 21 97.6 (3.5) 33.6 (1.8) 0.013 62.5 (9.1) 20.4 (3.9) 28.1 (5.6) 14 (5) 
Sg. Koyah (KB1-2) 18 82.1 (3.4) 28.7 (1.4) 0.013 40.7 (4.8) 13.6 (1.7) 18.7 (2.3) 8.4 (1.8) 
DGFC (KB1-5) 62.2 a 20.3 a 0.019 27.0 (a5.9 (a9.1 (a12.0 (a
Sukau (KB2):           
Sg. Resang (KB2-1) 27 81.9 (3.8) 34.8 (1.8) 0.011 17.4 (5.4) 6.3 (2.5) 7.1 (2.2) 4.0 (1) 
Sg. Resik (KB2-2) 12 90.0 (1.7) 40.1 (1.1) 0.010 14.6 (0.8) 5.1 (0.3) 6.1 (0.5) 3.5 (0.3) 
Kuala Sukau (KB2-3) 18 86.8 (2.2) 40.6 (1.2) 0.009 15.1 (0.8) 5.3 (0.5) 6.4 (0.4) 3.4 (0.3) 
Sg. Menanggol (KB2-5) 18 106.8 (8.4) 47.6 (6.9) 0.010 23.9 (6.6) 9.6 (3.1) 10.1 (2.9) 4.2 (0.7) 
Abai (KB4):           
Balat Damit (KB4-1) 36 92.3 (10.5) 39.5 (3.7) 0.010 21.8 (8.8) 9.0 (4.9) 8.9 (3.7) 3.9 (0.7) 
Sg. Merah (KB4-2) 21 82.1 (6.3) 34.0 (2.7) 0.012 14.2 (1.8) 4.9 (0.2) 5.8 (0.7) 3.5 (1.7) 
Bilit (KB3):           
Sg. Tenagang Besar (KB3-1) 21 101.0 (5.1) 39.1 (3.9) 0.013 36.7 (11.8) 11.7 (3.7) 17.0 (5.7) 8.0 (3) 
Sg. Tenagang Kecil (KB3-2) 12 73.9 (11.0) 30.3 (6.5) 0.011 23.1 (5.3) 8.2 (2) 10.3 (2.7) 4.6 (0.8) 
Ditches           
BS Mill (KB1-3) 89.2 (2.2) 34.2 (2.0) 0.012 41.3 (0.5) 13.6 (0.5) 18.3 (0.3) 9.5 (1.0) 
PS Plantation (KB1-4) 13 81.6 (5.2) 30.9 (2.0) 0.013 26.5 (11) 8.7 (4.2) 10.8 (5.2) 7.1 (1.8) 
Canals (KB1-6 to 8) 142.5 (39.8) 47.9 (15.0) 0.013 57.4 (20.1) 24.4 (11.9) 24.4 (8.4) 8.6 (2.7) 
Malbumi Plantation (KB2-4) 12 123.7 (14.5) 59.8 (9.0) 0.008 17.3 (4.1) 6.3 (1.8) 7.1 (1.8) 4.0 (0.5) 

aData are unavailable.

Several parameters can be derived from the UV-visible absorbance and the linearised gradient of absorbance, or spectral slope. Following Helms et al. (2008), the latter has been calculated between the wavelengths 275 and 295 nm. The spectral slope is plotted against absorption at 340 nm (a340) (indicator of DOM concentration) in Figure 4. The spectral slope is lowest in KB2 (KB2-1 to KB2-6), indicating DOM of a higher molecular weight, while samples from Batu Putih have the highest spectral slope (lowest molecular weight).

Figure 4

Correlation and comparison graphs between: (a) to (f) PARAFAC components C1, C2 and C3 against UV absorbance at 340 nm, respectively; (g) to (i) PARAFAC components C1, C2 and C3 against the spectral slope.

Figure 4

Correlation and comparison graphs between: (a) to (f) PARAFAC components C1, C2 and C3 against UV absorbance at 340 nm, respectively; (g) to (i) PARAFAC components C1, C2 and C3 against the spectral slope.

PARAFAC modelling

Three fluorescent components were identified by PARAFAC using EEMs of all samples collected from the study area. The ratio of the PARAFAC component intensity to the total fluorescence intensity for each sampling site is summarised in Table 3. The excitation and emission pairs of the main peak positions for each component are presented in Figure 5. The PARAFAC model identified three terrestrial peaks as characteristic fluorescent components in the Lower Kinabatangan River catchment. The mean total fluorescence intensity of samples collected is: IC2 > IC1 > IC3.

Figure 5

Fluorescence signatures of three identified PARAFAC model components. Contour plots present spectral shapes of excitation and emission of derived components. Line plots adjacent to each contour plot present split-half validation results for each identified component. Excitation (dotted line) and emission (solid line) loadings for each component, obtained from two independent PARAFAC models on random halves of the data array.

Figure 5

Fluorescence signatures of three identified PARAFAC model components. Contour plots present spectral shapes of excitation and emission of derived components. Line plots adjacent to each contour plot present split-half validation results for each identified component. Excitation (dotted line) and emission (solid line) loadings for each component, obtained from two independent PARAFAC models on random halves of the data array.

Component C1 contributed 35% of modelled fluorescence for the samples. It represents a combination of two non-separated peaks of different excitation wavelength: with a double excitation maxima at 345 and <290 nm (corresponding to the type A and C) and a single emission peak at 458 nm.

The excitation maximum for C2 occurred at 315 nm at 398 nm emission and the percentage contribution of modelled fluorescence is 43%, while component C3 occurred at a maximum excitation wavelength of <290 and 330 nm and emission wavelength of 360 nm, and contributes 22% of modelled fluorescence.

The fluorescent intensities of both PARAFAC components C1 and C2 were greatest at KB1, particularly in waters sampled from plantation ditches (highest intensity: 34.8 units) and in tributary stream KB1-1 (highest intensity: 23.2 units) (Table 3). Components C1, C2 and C3 are plotted against a340 in Figures 4(a)4(f), and theoretical lines of constant fluorescence per unit absorbance are shown in Figures 4(a)4(c). Water samples would be expected to lie on one of these lines if there were no change in DOM characteristics between sites, and if the intensity of both parameters reflected a dilution effect. Figures 4(a)4(f) indicate that only one subset of samples, those from the main stem of the Kinabatangan, lie along a dilution line. The results suggest that there is a loss of fluorescent DOM between the catchment tributaries and the main stem of the river. As a result, DOM in the main stem of the river is relatively less fluorescent per unit absorbance than expected. DOM hydrophilicity was found to decrease as DOM travels from Batu Putih to Abai, which may also reflect degradation of the DOM from PARAFAC component C1 (peak C) to PARAFAC component C2 (peak A). This trend is also evident in Figures 4(g)4(i) where PARAFAC components C1, C2 and C3 are plotted against spectral slope.

Figures 6(a) and 6(b) present the UV absorbance at 340 nm against spectral slope for each sampling station. High UV-visible absorption coefficient of a340 and low spectral slopes characteristics were found in waters sampled from the main stem of the Kinabatangan and from oil palm plantation ditches (DC) in KB2 and KB4, which could be indicative of reprocessed DOM. A comparison between type of sampling sites showed that UV absorbance at 340 nm for DC was more variable than for MS and ST samples <45% coefficient of variance for DC and <27% for MS and ST). S275−295 also exhibited similar variations where the coefficient of variance for DC was <23% and <20% for both MS and ST.

Figure 6

UV absorbance at 340 nm against spectral slope (S275−295) according to: (a) sampling area; (b) type of sampling site.

Figure 6

UV absorbance at 340 nm against spectral slope (S275−295) according to: (a) sampling area; (b) type of sampling site.

DISCUSSION

The environmental isotopes provide useful information on water movement through the catchment. Isotope data from tropical catchment are scarce (Stephens & Rose 2005), and hence the results of this study are useful in helping understand the modern stable isotope composition of Borneo. The isotopic composition of precipitation samples varies significantly during the period of sampling, and becomes isotopically lighter over time. Deuterium isotope values within single rainfall events have been found to vary by as much as 30‰ in 15 min within a single temperate rainfall event (Dansgaard 1964). Isotopic variability within single rainfall events in tropical regions will also be strongly controlled by the isotopic composition of precipitation during the monsoonal periods reflecting ‘amount effect’ (Dansgaard 1964; Darling et al. 2005). The isotopic composition of waters sampled from the Kinabatangan River and tributaries are isotopically lighter than the annual mean composition of precipitation, intersecting the RMWL at ∼50‰ (δ2H) and −8‰ (δ18O). These waters evolve away from the MWL at a lower slope due to evaporation. Some tributaries respond more strongly (i.e., with a reduced slope) indicating greater evaporation across their sub-catchments which may be due to vegetation cover and/or higher evaporation. Significantly, waters sampled from drainage ditches associated with the oil palm plantations were among those waters that were isotopically evolved. This possibly reflects a reduced cover of understorey vegetation and hence a greater evaporation effect.

Spectral slope and UV absorbance at 340 nm yield useful information about DOM characteristics and have been found to correlate strongly with the molecular weight of fulvic acid isolates (Baker et al. 2008; Helms et al. 2008). UV absorbance at 340 nm has been found to be a useful surrogate for DOC concentration (Spencer et al. 2009). An indicative four times difference in molecular weight is seen between the results and the spectral slope of Suwannee River natural organic matter (Helms et al. 2008). Other studies have reported S275−295 to be ∼13–17 × 10−3nm−1 in freshwater samples from Chesapeake Bay, USA (Helms et al. 2008), 12.2–19.9 × 10−3nm−1 in the Yukon River, Alaska (Spencer et al. 2009) and ∼18–19 × 10−3nm−1 from eutrophic Lake Taihu, China (Zhang et al. 2009b). Spectral slope has been found to increase with irradiation (e.g., Helms et al. 2008; Zhang et al. 2009a, b), but the ditches (DC) and main stem river-water samples (MS) were characterised by a lower spectral slope (7–19 × 10−3nm−1), especially in the main river compared to the tributaries. This suggests that DOM in the main stem and oil palm plantation ditches of the Lower Kinabatangan River catchment in particular samples from KB2 is of relatively high DOC and molecular weight. A quantitative aquatic carbon budget for the Langat River watershed in Malaysia (Figure 1) indicated that although C3 plant-derived matter was the primary source of carbon in wetland areas, sewage treatment and landfill sites in the lower reaches of the river provided a significant additional source of organic carbon (Lee et al. 2013). The relatively non-fluorescent and high-molecular-weight DOM at the Lower Kinabatangan River catchment may be indicative of relatively stable organic complexes such as humic substances (HS) given the high concentration of fine sediment within the colloidal (<0.7 μm) size range, rather than as a result of photo-degradation processes. This is consistent with a wetland study in Mukah, Sarawak (Figure 1) which found DOC concentrations varied between 18.9 and 75.3 mgC/L with a significant correlation with soil-derived humic substances (Watanabe et al. 2012). This study also suggested that humic substances mainly drive DOC photoreactivity in wetlands and HS tend to increase with increasing precipitation and/or temperature.

The PARAFAC analysis provides further information on the origin and biogeochemistry of DOM at each site, as well as any systematic differences in DOM characteristics throughout the catchment (particularly from sites upstream to downstream; KB1 to KB4). The spectral characteristics of C1 resemble those of Kowalczuk et al. (2009), Luciani et al. (2008), Stedmon & Markager (2005) and Yamashita et al. (2008) and can be associated with terrestrially derived OM, which occurs in a range of aquatic environments. Both components C2 and C3 have been described by Coble (1996) and Parlanti et al. (2000) as marine in origin. Stedmon et al. (2003) suggested that this component was observed in samples which were terrestrially origin OM while more recently, Fellman et al. (2010) identified this peak as ultraviolet A; a low molecular weight component which they attributed to microbial processing. This component is common in marine environments associated with biological activity but has also been found in wastewater, wetland and agricultural environments. Thus, it is likely that components C2 and C3 found in samples from the Kinabatangan catchment is reprocessed DOM derived from microbial and/or photo-degradation processes.

The PARAFAC components C1, C2 and C3 identified in this study, indicate that the majority of DOM in the Kinabatangan catchment comprises terrestrially derived material, and that the relative loss of PARAFAC component C1, relative to absorbance, is due to a change in DOM characteristics. Upstream sites on the Kinabatangan (i.e., KB1) which include several sites closely associated with oil palm plantations (KB1-3, KB1-4, KB1-6 to 8), have a high fluorescence PARAFAC component C1 per unit absorbance and high spectral slope (Figures 6(a) and 6(b); Table 3), which is indicative of organic matter with relatively low molecular weight. In contrast, sites on the Kinabatangan downstream (i.e., at KB2) have a high UV-visible absorption coefficient and low spectral slope, which is indicative of a higher molecular weight DOM.

Although the effects of sunlight on DOM degradation by bacteria are various and contrasting (Cory et al. 2013), both components C2 and C3 could be indicative of microbial and/or photo-degradation processes, particularly given the high solar radiation that is typical of tropical areas. Significantly though, the variation in fluorescence relative to UV-visible absorbance between tributary sites and the main stem of the Kinabatangan, reveal a loss of highly fluorescent DOM within the catchment. This loss, and the general spatio-temporal variability of DOM characteristics, will reflect the interaction between at least three sources of variability of DOM: (i) spatial variation in DOM source; (ii) the effects of transport (e.g., temporal degradation from photo-degradation); and (iii) differences among DOM in the propensity for biological and photochemical removal.

The results suggest that in the Kinabatangan catchment, rapid oxidation by photochemical and more especially microbial processes to produce carbon dioxide (Cory et al. 2007) preferentially breaks down the aromatic carbon-containing molecules which account for the fluorescent properties of DOM. DOM in waters sampled from low-order tributaries would be expected to be dominated by inputs of terrestrially derived DOM, which would include DOM that is derived from activities associated with oil palm plantations. This DOM is rapidly photo- and bio-degraded to less fluorescent, stable DOM, which is probably present as fine colloidal complexes. The concentration and character of DOM downstream is then likely to reflect hydrological controls, particularly the rate of water movement (Findlay & Sinsabaugh 1999).

CONCLUSIONS

The results and analyses presented in this paper provide important baseline data on variations in DOM quantity and quality in a degraded tropical catchment in the island of Borneo. The findings from this study provide further information on DOM characteristics in tropical agricultural catchments. These environments have been hitherto neglected in studies of DOM characterisation and hence the results presented here are particularly relevant to other degraded aquatic ecosystems, particularly those potentially impacted by the development of oil palm plantations. The results suggest that by characterising DOM using absorbance and fluorescence, it is possible to differentiate between the DOM characteristics of individual sub-catchments of the Kinabatangan River. The analysis is helped significantly by the additional information provided from environmental isotopes, and there is evidence of progressive photo- and microbial-degradation of DOM downstream. While this is important in itself, the focus in this study as been on baseflow conditions, and more work is urgently required to investigate any seasonal trends in DOM characteristics in catchments such as the Kinabatangan.

ACKNOWLEDGEMENTS

We thank the Universiti Malaysia Sabah, Sabah Forestry Department and Sabah Wildlife Department for permitting this research to be undertaken in the Kinabatangan River Catchment. Thanks are also extended to Mr Kevin Burkhill and Ms Arnie Abdul Hamid for drawing Figure 1; Mr Zainal Abidin Jaafar, Ms Asnih Etin, Mr Mansur Ismail and Mr Ismail Abdul Hamid for their great help with the field sampling.

REFERENCES

REFERENCES
Araguás-Araguás
L.
Froehlich
K.
Rozanski
K.
1998
Stable isotope composition of precipitation over southeast Asia
.
J. Geophys. Res.
103
,
721
728
,
742
.
Boonratana
R.
2000
A study of the vegetation of the forests in The Lower Kinabatangan Region, Sabah, Malaysia
.
Mal. Nat. J.
54
,
271
288
.
Burke
P. M.
Hill
S.
Iricanin
N.
Douglas
C.
Essex
P.
Tharin
D.
2002
Evaluation of preservation methods for nutrient species collected by automatic samplers
.
Environ. Monit. Assess.
80
(
2
),
149
173
.
Dansgaard
W.
1964
Stable isotopes in precipitation
.
Tellus
16
,
436
468
.
Darling
W. G.
Bath
A. H.
Gibson
J. J.
Rozanski
K.
2005
Isotopes in water
. In:
Isotopes in Palaeoenvironmental Research
(
Leng
M. J.
, ed.).
Springer
,
Dordrecht, The Netherlands
.
Fitzherbert
E. B.
Struebig
M. J.
Morel
A.
Danielson
F.
Brühl
C. A.
Donald
P. F.
Phalan
B.
2008
How will oil palm expansion affect biodiversity?
Trends Ecol. Evol.
23
(
10
),
538
545
.
Hader
D. P.
Kumar
H. D.
Smith
R. C.
Worrest
R. C.
1998
Effects on aquatic ecosystems
.
J. Photochem. Photobiol. B Bio.
46
,
53
68
.
Harmel
R. D.
Cooper
R. J.
Slade
R. M.
Haney
R. L.
Arnold
J. G.
2006
Cumulative uncertainty in measured streamflow and water quality data for small watersheds
.
Trans. ASABE
49
,
689
701
.
Harun
S.
2006
Aquatic Insects and Water Quality of the Lower Kinabatangan River. Unpublished MSc Thesis
,
Universiti Malaysia Sabah
,
Malaysia
.
Hooijer
A.
Silvius
M.
Wösten
H.
Page
S. E.
2006
Peat-CO2, assessment of CO2 emissions from drained peat lands in SE Asia
.
Delft Hydraulics Report Q3943
,
41
.
Hope
D.
Billett
M. D.
Cresser
M. S.
1994
A review of the export of carbon in river water: fluxes and processes
.
Environ. Pollut.
84
,
301
324
.
Josephine
R.
Alfred
R. J.
Indran
R.
2004
Integrated river basin management planning for the Kinabatangan Catchment, Sabah: approach and strategy
.
Paper presented on World Water Day 2004
.
World Wide Fund Malaysia (WWFM)
.
Lee
K. Y.
Syakir
M. I.
Clark
I. D.
Veizer
J.
2013
Isotope constraints on the aquatic carbon budget: Langat Watershed, Malaysia
.
Aquat. Geochem.
19
(
5–6
),
443
475
.
Limpens
J.
Berendse
F.
Blodau
C.
Canadell
J. G.
Freeman
C.
Holden
J.
Roulet
N.
Rydin
H.
Schaepman-Strub
G.
2008
Peatlands and the carbon cycle: from local processes to global implications – a synthesis
.
Biogeoscience
5
,
1475
1491
.
Luciani
X.
Mounier
S.
Paraquetti
H.
Redon
R.
Lucas
Y.
Bois
A.
Lacerda
L.
Raynaud
M.
Ripert
M.
2008
Tracing of dissolved organic matter from the SEPETIBA Bay (Brazil) by PARAFAC analysis of total luminescence matrices
.
Mar. Environ. Res.
65
,
148
157
.
Mansourian
S.
Davison
G.
Sayer
J.
2003
Bringing back the forests: by whom and for whom? In
:
Proceedings of an International Conference on Bringing Back the Forests: Policies and Practices for Degraded Lands and Forests
,
Kuala Lumpur
,
7–10 October 2002
, pp.
27
140
.
Mladenov
N.
McKnight
D. M.
Macko
S. A.
Norris
M.
Cory
R. M.
Ramberg
L.
2007
Chemical characterization of DOM in channels of a seasonal wetland
.
Aquat. Sci.
69
,
456
471
.
Oliveira
J. L.
Boroski
M.
Azevedo
J. C. R.
Nozaki
J.
2006
Spectroscopic investigation of humic substances in a tropical lake during a complete hydrological cycle
.
Acta. Hydrochim. Hydrobiol.
34
,
608
617
.
Pace
M. L.
Cole
J. J.
Carpenter
S. R.
Kitchell
J. F.
Hodgson
J. R.
van de Bogert
M. C.
Bade
D. L.
Kritzberg
E. S.
Bastviken
D.
2004
Whole-lake carbon-13 additions reveal terrestrial support of aquatic food webs
.
Nature
427
,
240
243
.
Payne
J.
1996
Sabah Biodiversity Conservation Project, Malaysia: Kinabatangan Multi Disciplinary Study
.
Ministry of Tourism and Environmental Development, Sabah & Danish Co-operation for Environment and Development (DANCED)
.
Richey
J. E.
Melack
J. M.
Aufdenkampe
A. K.
Ballester
V. M.
Hess
L. L.
2002
Outgassing from Amazonian rivers and wetlands as a large tropical source of atmospheric CO2
.
Nature
416
,
617
620
.
Rixen
T.
Baum
A.
Pohlmann
T.
Balzer
W.
Samiaji
J.
Jose
C.
2008
The Siak, a tropical black water river in central Sumatra on the verge of anoxia
.
Biogeochemistry
90
,
129
140
.
Sabah Forestry Department
.
2001
Conservation Areas, Information and Monitoring System (CAIMS)
.
Available at
: .
Sidle
C. S.
Ziegler
A. D.
Negishi
J. N.
Nik
A. R.
Siew
R.
Turkelboom
F.
2006
Erosion processes in steep terrain – Truths, myths, and uncertainties related to forest management in Southeast Asia
.
For. Ecol. Manage.
224
,
199
225
.
Spencer
R. G. M.
Aiken
G. R.
Butler
K. D.
Dornblaser
M. M.
Striegl
R. G.
Hernes
P. J.
2009
Utilizing chromophoric dissolved organic matter measurements to derive export and reactivity of dissolved organic carbon exported to the Arctic Ocean: A case study of the Yukon River, Alaska
.
Geophys. Res. Lett.
36
,
1
6
.
Tipping
E.
Corbishley
H. T.
Koprivnjak
J.-F.
Lapworth
D. J.
Miller
M. P.
Vincent
C. D.
Hamilton-Taylor
J.
2009
Quantification of natural DOM from UV absorption at two wavelengths
.
Environ. Chem.
6
(
6
),
472
476
.
Watanabe
A.
Moroi
K.
Sato
H.
Tsutsuki
K.
Maie
N.
Melling
L.
Jaffé
R.
2012
Contributions of humic substances to the dissolved organic carbon pool in wetlands from different climates
.
Chemosphere
88
(
10
),
1265
1268
.
Winter
A. R.
Fish
T. A. E.
Playle
R. C.
Smith
D. S.
Curtis
P. J.
2007
Photodegradation of natural organic matter from diverse freshwater sources
.
Aquat. Toxic.
84
,
215
222
.