Diatom and geochemical indicators of acidification in a tropical forest stream, singapore 6

An overall examination of the organic carbon content, geochemical and diatom data within the cores from Jungle Falls Valley indicate that the record is likely to be continuous and comple

Chapter Six

RESULTS AND DISCUSSION

6.1 Overview

Chapter six contains the results obtained in this study and discusses whether there is any evidence for the acidification of Jungle Falls stream within BTNR It is important to note that this study is not only about the potential

acidification of Jungle Falls stream per se, but also about investigating the

potential of paleolimnological indicators to track the acidification of freshwater ecosystems in Singapore and the surrounding region Firstly, there is a discussion on the quality of the sedimentary record obtained from Jungle Falls stream A description of the three cores is then provided, followed by the organic carbon content of the sediments within As the variations in organic carbon content are not extreme, the environment of the stream is unlikely to have changed significantly The variations in organic carbon content is thus used to correlate the cores

The results of the diatom analysis and trace metal analysis are shown Changes within the biological and geochemical profile of the cores demonstrate that there appears to be a record of atmospheric contamination and acidification within the sediments from Jungle Falls Valley It points to a rise in acidity and atmospheric contamination to the stream following the rapid industrialisation and urbanisation of Singapore in the 1960s There is also a possibility of recent acidification of the stream Finally, the chapter covers some limitations to the analysis which should be considered alongside the interpretation of the results

6.2 The Sedimentary Record

In paleolimnological studies, the quality of the sedimentary record is of utmost importance According to Smol (2008), it is the collection of the initial sediment cores that is the most critical step in the paleolimnological process This

is because it is near impossible to rectify any errors or problems encountered with the cores after collection and data analysis This core collection entails the

selection of an appropriate coring site which has accumulated sediments representative of overall limnological and environmental changes in the region

In Singapore, there is a lack of coring sites for the study of acidification (see section 3.2) However, the damming of Jungle Falls stream has provided an ideal location for the investigation of potential acidification of the area This is because it is located in a forested catchment that is not directly affected by anthropogenic activities, but may show effects of indirect anthropogenic atmospheric pollution and contamination The absence of suitable coring sites in the region, along with a lack of focus on acidification issues within Asia at the moment (see section 2.4), means that paleolimnological techniques have rarely been employed in regional acidification studies This study is then also about evaluating the potential of paleolimnological indicators to track acidification of tropical freshwater ecosystems, in this case, a stream, and is the first time such analysis is being carried out in Singapore

Another key assumption in this analysis is that the sedimentary record is continuous and complete This means that no erosion can have taken place and that there has been no hiatus in deposition One way of addressing this issue is

by dating the sediment core at regular high-resolution intervals While

Caesium-137 analysis does not provide the age of a core at depth, but rather a stratigraphic marker horizon, it would be a possible means of conducting this analysis

time-Caesium-137 is an artificially generated radioactive nuclide that has only been produced in significant quantities as a result of thermonuclear weapons testing which began in 1945 (Lowe and Walker, 1997) Atmospheric caesium-137 levels peak in 1963, after which, atmospheric levels declined significantly with successive nuclear ban treaties (Walker, 2005) This 1963 maximum is reflected

in sediment sequences and forms the distinctive marker horizon in the core While lead-210 is also common in paleolimnological acidification studies, these studies often look at longer sediment sequences of 100-150 years As the Jungle Falls dam is believed to be built during the late 1930s, caesium-137 would be an ideal choice for analysing this sediment core Unfortunately, this was beyond the scope of the study

Another indication of an incomplete record would be abrupt changes within the analysed data such as a sharp drop in %LOI or chemical concentrations An overall examination of the organic carbon content, geochemical and diatom data within the cores from Jungle Falls Valley indicate that the record is likely to be continuous and complete

6.3 Description of Sedimentary Profile

6.3.1 Core A

Being extracted from the side of the stream, sediment from Core A was homogenous and, as such, there were no visually distinguishable sedimentary layers The sediment was dark brown silt, containing decomposed detritus with the occasional roots, twigs and leaves (darkness 4, stratification 0, elasticity 3, dryness 2, humicity 3; plate 6-1), with a water content of approximately 80%

Plate 6-1: Sediment sample from Core A

6.3.2 Core B

The sediments from Core B, collected from the middle of the stream, were divided into three units The bottommost sediment, corresponding to a depth of 23-24cm, was light brown sand, containing decomposed detritus with roots, twigs and leaves (darkness 2, stratification 0, elasticity 0, dryness 3, humicity 0; plate 6-2) This layer had a 30% water content

Plate 6-2: Sediment sample from the base of core B, at the depth of 23-24cm

Above that, at a depth of 20-23cm, was a medium brown mixture of sand and silt, containing decomposed detritus with some roots, twigs and leaves (darkness 3, stratification 0, elasticity 1, dryness 3, humicity 1; plate 6-3) Water content of this layer was approximately 70%

Plate 6-3: Sediment sample from core B, collected at a depth of 20-23cm

Moving up past the bottom two sandy layers, the remaining sediment from Core B is similar to that from Core A – dark brown silt comprising decomposed detritus and the occasional root, twig and leaf (darkness 4, stratification 0, elasticity 3, dryness 2, humicity 3) Again, water content in this layer was around 80%

6.3.3 Core C

As with Core A, the sediment in Core C was again homogenous with no visible sedimentary layers The water content of sediment from Core C was also around 80% Thus, the sediment was dark brown silt with decomposed detritus and occasional roots, twigs and leaves (darkness 4, stratification 0, elasticity 3, dryness 1, humicity 3) There was a large piece of wood at the bottom of the core (at a depth 16-18cm) As such, this sample (C17) comprised of a 3cm section as opposed to the usual 1cm

6.4 Organic Carbon Content

Despite there being minimal visual changes within the cores, variations are present in the organic carbon content of the sediments %LOI was plotted against sample identification number for the three cores (figure 6-1) As the bottom three %LOI values of Core B are significantly lower (at 5.2%, 30.6% and 45.2%) than the other values, which range between 50-70%, Core B was plotted twice, once including the three points (B13, B14 and B15, Core B1) and once excluding them (Core B2) This will therefore enable the variations within Core B

to be seen more clearly

At first glance, there appears to be little correlation between the three cores However, adjustments have to be made before the cores can be compared This is because the cores were collected from three points behind the dam – the side, middle and just behind the dam (Plate 4-3) – each with slightly

Figure 6-1: %LOI graphs plotted against sample ID

There are similarities in the %LOI profiles of the three cores For instance, there

is a peak value at A5, B3 and C6, along with a trough at A7, B4 and C8 Another peak is present at A8, B8 and C13 These peaks and troughs were matched and the cores were corrected accordingly

The base of Core B was a sandy layer and it appears that this core had penetrated into the channel substrate, reflecting the ground conditions of the stream prior to impoundment Therefore, this was assumed to be the deepest layer The corrected data was then assigned depth values with the youngest sediments representing a depth of 1cm, as the cores extended to the surface See Appendix B for a graph that shows the corrected depths calculated from the original %LOI graph Figure 6-2 shows the %LOI with depth of the three cores following the adjustment process

Figure 6-2: %LOI with depth

Thus, it can be seen that %LOI is low at the bottom of the sedimentary profile, corresponding to the sandy layers in Core B, before rising rapidly and staying high, between 50%-70%, for the remainder of the profile There appears

to be three peaks in the sedimentary profile, at depths of 6cm, 13cm and 17cm However, as there is no perceptible upward or downward trend in the data and as the actual variation in organic carbon content values is low, this change in organic carbon content is unlikely to stem from a change in the environmental conditions within the basin Thus, the organic carbon content profiles are useful in correlating the sediment cores and adjusting sample depths according for comparison Because there is no significant change in %LOI, with values remaining high throughout the core, any variation seen in the biological and geochemical data from the cores is likely to be due to anthropogenic influences

6.5 Evidence for Stream Acidification

view (front) or a girdle view (side), often depending on which has the larger and/or flatter surface Diatoms mounted on their girdle side were not identified or counted This is because the girdle band of a diatom has less intricate patterns than valves and are also illustrated less often in the taxonomic literature, making

them harder to identify (Battarbee, 2001; Blanco et al, 2008) As such, diatom

girdles are often ignored in counting (Battarbee, 2001; Crosta and Koç, 2007; Jordan and Stickley, 2010)

Plate 6-4: Diatoms showing signs of dissolution and breakage All diatoms at 400x magnification.

Unfortunately a significant proportion (30%-60%) of diatoms in this study appeared in girdle view Yet, this was not entirely unexpected as “frustules of

genera with wide girdle bands (especially Eunotia) usually settle from suspension

in girdle view” (McBride, 1988) Microscope slides from a previous study in Nee Soon Swamp Forest (NSSF), Singapore, with an assemblage also dominated by

Eunotia species (72.4%), had a similar proportion of diatoms in girdle view (Oon,

2010) In this situation, a potential method to increase the number of diatoms that

can be counted is to separate the diatom valves from the girdles using an ultrasonic bath (McBride, 1988) Unfortunately, a side-effect of the sonication of a

diatom suspension is the fracturing of valves (Battarbee et al, 2001; Serieyssol et

al, 2011) As valve breakage is already an issue within the Jungle Falls diatom

assemblage, it was determined that sonication of the diatom suspensions was not recommended

Broken valves were only counted when more than half of the valve was present and identifiable There were between 150-200 diatoms per slide from Core A and 200-500 diatoms per slide from Core B and Core C The sandy layers

in Core B yielded the most number of diatoms, with 641 in the lowest slide, followed by 570 and 503 in the slides above These concentrations are low, further implying that diatom preservation at Jungle Falls stream is an issue

As mentioned in section 5.3.3, diatoms are usually counted until a predetermined target is reached, typically between 300-600 diatoms per slide This number ensures that enough diatoms are counted for the entire assemblage

to be represented (Battarbee et al, 2001) In this study, the entire slide had to be

counted and even so, none of the slides from Core A have diatom numbers approaching the recommended target values When counting diatoms under a

microscope, Battabee et al (2001) also recommends three or four diatoms per

field of view Such a concentration was not possible in this study Even though a higher diatom concentration and count could be achieved by dropping more than 400µl of each diatom suspension onto the coverslip, this was not possible as the other components in the suspension, such as the mineral debris, would also have increased in concentration and obscured the diatoms present

There is one other study in Singapore that attempted to look at a sedimentary diatom record to track environmental change An analysis of diatoms from NSSF found that “none of the samples analysed contained sufficient quantities of undamaged diatoms to warrant further, detailed analysis of diatoms”

(Taylor et al, 2001: 274) Another examination of the NSSF sediments, conducted

by Oon (2010), found that only the topmost sediment sample collected contained sufficient diatoms to enable some analysis to be conducted, and even these sediments displayed preservation problems This points to potential issues with diatom preservation in Singapore in general

Nevertheless, there were sufficient diatoms in each slide from this study

to permit counting and analysis In total, 40 diatom species were present in the

sediment samples The assemblage was dominated by Eunotia species, which

constituted around 80% of the diatoms present in slides 12 diatom species were

selected for the analysis of changing diatom assemblage with depth – Eunotia

incisa, Eunotia paludosa, Eunotia flexuosa, Eunotia rhomboidea, Eunotia fallax, Eunotia curvata, Eunotia vanheurckii, Eunotia pectinalis, Eunotia parallela, Fragilaria af bicapitata; Frustulia rhomboides and Frustulia af rhomboids var crassinervia (plate 6-5) These species accounted for around 90% of the diatoms

in the slides, with the other 29 diatom species making up the last 10% As mentioned in section 5.2.3, there was difficulty identifying diatoms in this study

As such, diatoms marked with “af.” indicates that these diatom species have not yet been recorded in Singapore, and thus only have an affinity with the identified species

Plate 6-5: Predominant diatoms present in the sediment cores (a) Eunotia parallela; (b) Eunotia

pectinalis; (c) Eunotia curvata; (d) Eunotia flexuosa; (e) Fragilaria af bicapitata; (f) Frustulia rhomboides; (g) Frustulia af rhomboids var crassinervia; (h) Eunotia fallax; (i) Eunotia rhomboidea;

(j) Eunotia incisa; (k) Eunotia af paludosa; (l) Eunotia vanheurckii All diatoms at 400x

magnification

Plate 6-6 contains a selection of the other diatoms present in this study See appendix C for a full list of diatoms found in the Bukit Timah Jungle Falls sediment samples Some diatoms could not be identified due to the lack of a regional identification key, along with poor image quality (plate 6-7) Unfortunately, a better microscope would be needed in order to view the details required for identification with regard to some diatoms These would include being able to magnify the diatoms 1000x and viewing the striations within each diatoms As seen in plate 6-7, the outlines of these diatoms are not unique or peculiar enough for identification and an image of the striations within would greatly aid identification A better microscope, and view of the striations in the diatoms, would also help confirm the identification of some diatoms where only

the outlines are perceivable such as Surirella af angusta, Navicula af subtilissima and Achnanthes af helvetica (plate 6-6)

h

Plate 6-6: A selection of other diatoms present in the sediment cores (a) Pinnularia abaujensis; (b)

Navicula cryptocephala; (c) Eunotia af papilio; (d) Eunotia hexaglyphis; (e) Eunotia serra; (f) Pinnularia braunii; (g) Hantzschia amphioxys; (h) Eunotia camelus; (i) Eunotia sp.; (j) Surirella af angusta; (k) Navicula af subtilissima; (l) Eunotia sp.; (m) Pinnularia microstauron; (n) Amphora angusta; (o) Fragilaria af lapponica; (p) Achnanthes af helvetica All diatoms at 400x magnification

Plate 6-7: A selection of unidentified diatoms in the sediment cores

However, the fact that there are as yet unidentified diatoms should not have an impact on this study This is because the unidentified diatoms each comprise such a small proportion of the diatoms in the assemblages (less than 5 per slide) that they are unlikely to have much significance

h

p

The diatom assemblage with depth is shown in figure 6-3 As Frustulia

rhomboides and Frustulia af rhomboides var crassinervia both comprise a small

percentage of the assemblage, they were combined to form Frustulia spp

(species) While there does not appear to be any distinctive assemblage zones, the diatom data from Jungle Falls Valley still contains significant results The diatom flora and assemblage are representative of the environment they are

found in Eunotia species, in general, are strong indicators of an acidic,

freshwater, ogliotrophic environment which is oxygen-rich and poor in organic nitrogen compounds; though some species can thrive in other environments as

well (Van Dam et al, 1994)

From figure 6-3, it can be seen that the assemblages are all dominated by

Eunotia incisa, Eunotia af paludosa, and Eunotia flexuosa Eunotia incisa is an

acidophilous freshwater species which has been observed in pH levels as low as 4.7 (Ortiz-Lerín and Cambra, 2007) In studying the diatom assemblages found in

selected Welsh lakes, Round (1990) found Eunotia incisa in acid lakes in that

have a pH as low as 4.9 and, unlike the other acidophilous species found, did not

extend into less acid lakes that had pH values of 5.4 or 6.1 Battarbee et al (2011) recorded this species at an optimum of 5.2 and Dixit et al (2002) at a pH optimum of 5.7 This makes Eunotia incisa potentially indicative of acidification in

Jungle Falls stream

Eunotia af paludosa has been found in Korea (Liu et al, 2011) and

Northern Thailand, Borneo and Indonesia (Patrick, 1936) However, a study of diatoms from Singapore and Peninsular Malaysia does not include this species (Wah, 1988) It is a freshwater species that is “often associated to mosses in acid waters of low mineral content, also in bogs and small streams” (Patrick and Reimer, 1966 cited in Ortiz-Lerín and Cambra, 2007: 426) It is an acidobiontic

species that occurs at a pH less than 5.5 (Van Dam et al, 1994), though it can be found in pH levels as high as 7.1 (Liu et al, 2011)

Eunotia flexuosa is an acidophilous freshwater species (Van Dam et al,

1994) It has a possible pH optimum between 4.3 to 6.5 (Liu et al, 2011) and Dixit

et al (2002) found that in Killarney Lake, Canada, it had a pH optimum of 6.0 In a

study of Tuckean Swamp, Australia, Taffs et al (2008) note a shift in diatom assemblage after 1970, whereupon Eunotia flexuosa became dominant They

infer this zone to be affected by land use change and the most acidic, with pH values between 3.5 to 4.5

A change can be seen in Core B as the sediments move from the sandy

layer below to the slit layer above In the sandy layers, Eunotia vanheurckii is abundant, comprising 25% of diatoms in the lowest layer This is also an acidophilous species However, unlike the above species, Eunotia vanheurckii does not seem to thrive at acidity levels below a pH of 5 Battarbee et al (2011) records it at an optimum of 5.9, Rosén et al (2008) at a pH of 5.8, Uutala et al (1994) at 5.6 and Dixit et al (2002) at 5.2 With Eunotia flexuosa notably only

comprising 2% of the diatom assemblage, the data seems to imply a stream pH

of about 5 or higher Moving up the core, Eunotia vanheurckii levels drop rapidly

to 16.5% at 23cm and 8% at 20cm It remains at between 7% and 9% at a depth

of 13cm to 19cm Above that, abundance drops to between 3% to 5% This is a similar proportion as the abundance recorded in Core A and Core C

Percentage of Eunotia curvata and Eunotia parallela are also lower at the base of the core Eunotia curvata is an acidophilous to pH indifferent species widely distributed in waters of low mineral content (Czarnecki et al, 1978) Dixit et

al (2002) finds it has an optimum of pH 5.6, while Uutala et al (1994) records it at

a pH optimum of 5.4 to 5.7 Less information is available on Eunotia parallela

besides it being an acidophilous species (Van Dam et al, 1994); though Liu et al

(2011) states that it has a rather large pH range of 4.3 to 9.1 Diatom evidence seems to point to a lowering pH within the stream over time

This view is further strengthened by the increase in Eunotia flexuosa

abundance moving up all cores, from the low of 2% to as high as 22%, though it

averages around 15% In contrast, proportion of Fragilaria af bicapitata

decreases moving up in Core A and Core B, though this difference is not apparent in Core C This is probably because Core C only reaches a depth of

19cm and the highest levels of Fragilaria bicapitata are recorded between 24cm According to Newcastle University (2011b), Fragilaria bicapitata is found in

20-ogliotrophic environments, and are a circumneutral freshwater species This, once again, implies a drop in pH as the species abundance decreases up-core

Eunotia incisa levels are low at the same depth of 15cm to 24cm In Core

A, it comprises approximately 15% to 18% of the assemblage In Core B, it comprises around 20% of the assemblage Again, this is not reflected in Core C

for the same reason as Fragilaria bicapitata Past a depth of 15cm, Eunotia incisa

abundance rises as high as 30% in Core A, and remains at around 28% and as

high as 32% in core B, remaining at around 30% Eunotia incisa abundance in

Core C is also similar, with an average abundance of approximately 30%, and going as high at 40% In a study of lakes in the Cairngorm and Lochnagar areas

of Scotland, Jones et al (1993) found a rise in the abundance of Eunotia incisa,

among other species, which corresponded to a decline in pH by 0.5 units While

Jones et al (1993) had a different diatom assemblage from that found at Bukit

Timah, this shows that a decline of 0.5 pH units, from 5.5 to 5.0, can cause an

increase in the dominance of Eunotia incisa