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

Chapter Six RESULTS AND DISCUSSION 6.1 Overview 88 6.2 The Sedimentary Record 88 6.3 Description of Sedimentary Profile 6.3.1 Core A 6.3.2 Core B 6.3.3 Core C 90 90 91 92 6.4 Organic Carbon Content 92 6.5 Evidence for Stream Acidification 6.5.1 Diatom Analysis 6.5.2 Geochemical Analysis 94 94 105 6.6 Limitations to Study 6.6.1 Representativeness of Diatoms 6.6.2 Preservation of Diatoms 117 118 119 87 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 88 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 Caesium137 analysis does not provide the age of a core at depth, but rather a timestratigraphic marker horizon, it would be a possible means of conducting this analysis. 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 89 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 90 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 91 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 different hydrological conditions. 92 %LOI (Core A) 50" 55" 60" %LOI (Core B2) %LOI (Core B1) 65" 0" 20" 40" 60" 80" 50" 60" %LOI (Core C) 70" 50" 55" 60" 65" 0" 2" 4" 6" Sample ID 8" 10" 12" 14" 16" 18" 20" 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. 93 %LOI (Core A) 50" 55" 60" %LOI (Core B1) 65" 0" 50" 100" %LOI (Core B2) 50" 60" %LOI (Core C) 70" 50" 55" 60" 65" 0" 2" 4" 6" Depth (cm) 8" 10" 12" 14" 16" 18" 20" 22" 24" 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 6.5.1 Diatom Analysis There were issues of diatom preservation in the Bukit Timah Jungle Falls sediments, as diatoms displayed signs of dissolution and a significant number of valves were broken (plate 6-4). When mounted, diatoms can either lie in a valve 94 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 95 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. 96 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. 97 a b c d e i f j g k l h 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). 98 a b c d e f g h i j k l m n o p 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. 99 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 100 Core A E. incisa E. af. paludosa E. flexuosa E. rhomboidea E. fallax E. curvata E. vanheurckii E. pectinalis E. parallela F. af. bicapitata Frustulia spp. Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) 0" 20" 0" 20" 0" 20" 0" 10" 0" 5" 10" 0" 5" 0" 5" 0" 5" 0" 5" 0" 5" 10" 0" Other diatoms Abundance (%) 5" 0" 5" 10" 2" 3" 4" 5" 6" Depth (cm) 7" 8" 13" 15" 17" 19" 20" 21" 22" Core B E. incisa E. af. paludosa E. flexuosa E. rhomboidea E. fallax E. curvata E. vanheurckii E. pectinalis E. parallela F. af. bicapitata Frustulia spp. Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) 0" 20" 0" 20" 0" 20" 0" 10" 0" 6" 0" 10" 0" 20" 0" 10" 0" 5" 0" 15" 0" Other diatoms Abundance (%) 10" 0" 15" 3" 4" 6" 8" 10" Depth (cm) 11" 12" 13" 15" 16" 17" 19" 20" 23" 24" Figure 6-3: Summary of the total percentage frequency of each diatom species with depth. 101 Core C E. incisa E. af. paludosa E. flexuosa E. rhomboidea E. fallax E. curvata E. vanheurckii E. pectinalis E. parallela F. af. bicapitata Frustulia spp. Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) 0" 20" 40" 0" 15" 0" 20" 0" 10" 0" 7" 0" 10" 0" 8" 0" 10" 0" 5" 0" 8" 0" Other diatoms Abundance (%) 5" 0" 1" 2" 3" 4" 5" 6" Depth (cm) 7" 8" 9" 10" 11" 12" 13" 14" 15" 17" 19" Figure 6-3 continued: Summary of the total percentage frequency of each diatom species with depth. 102 10" 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 103 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 2024cm. According to Newcastle University (2011b), Fragilaria bicapitata is found in 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. 104 Therefore, while the changes in the diatom assemblage at Jungle Falls valley are not drastic enough to enable the demarcation of assemblage zones, there appears to be potential evidence of acidification of the stream. 6.5.2 Geochemical Analysis The changes in trace metal concentrations with depth are shown in table 6-1 to 6-3 and figure 6-4 to 6-6. Figure 6-7 shows the trace metal concentration levels in Core A found by reprocessing the data. The graphs are interpreted from the bottom to the top, moving from the oldest sediment to younger sediments. Comparing the original data in figure 6-6 to the reprocessed data in figure 6-7, both concentration profiles display similar variation. Thus, zinc, sodium, potassium, iron and manganese values all start of high in both cores before decreasing. There are two peaks in both of the lead, zinc and sodium graphs, observed at depths of 4cm or 5cm and around 15cm; while the potassium graphs have just one peak at a depth of 15cm, with concentrations decreasing until 13cm before rising towards the surface. Manganese values in both datasets are practically identical. This suggests that the variation observed in the concentration profiles obtained for the three cores are accurate, and not due to any experimental errors. However, while the concentration profiles in the original and reprocessed data match well, the actual concentration values are significantly different. While sodium concentration levels range between 1ppm and 2ppm in the original data, they range between 30ppm to 50ppm in the reprocessed data, going as high as 80ppm. Potassium levels in the original data falls between 3ppm and 9ppm, but are between 10ppm and 15ppm in the reprocessed data. While the lead concentration profiles in both datasets have two peaks, and the first peak for both datasets have a similar concentration of 9ppm, the second peak in the profiles, at 105 Sample ID Depth (cm) Pb (ppm) Zn (ppm) Na (ppm) K (ppm) Fe (ppm) Mn (ppm) S (ppm) A1 2 5.87 0.87 1.66 6.26 595.83 0.86 142.53 A2 3 6.38 0.78 1.58 6.81 671.03 0.88 142.27 A3 4 7.33 0.53 1.69 5.34 536.00 0.78 150.01 A4 5 6.91 0.92 1.12 4.85 590.74 0.63 149.38 A5 6 7.09 0.91 1.39 5.34 596.08 0.63 143.04 A6 7 7.08 0.84 1.11 5.03 600.57 0.66 134.65 A7 8 6.18 0.86 1.26 4.65 590.16 0.71 135.12 A8 13 6.29 0.87 0.90 3.05 533.12 0.58 128.45 A9 15 9.76 1.36 2.30 6.01 699.51 0.87 146.37 A10 17 7.32 0.51 1.18 4.32 509.02 0.66 136.91 A11 19 6.56 0.92 1.54 4.75 578.43 0.79 138.91 A12 20 6.63 0.82 1.45 3.92 619.19 0.81 131.96 A13 21 5.89 1.29 1.51 6.26 746.80 1.33 135.60 A14 22 5.65 1.84 2.96 7.40 766.45 1.67 162.69 Table 6-1: Concentration (ppm) of trace metals with depth in Core A Sample ID Depth (cm) Pb (ppm) Zn (ppm) Na (ppm) K (ppm) Fe (ppm) Mn (ppm) S (ppm) B1 3 7.58 0.30 0.92 3.78 433.86 0.54 175.97 B2 4 7.42 0.54 1.16 4.88 460.16 0.53 174.35 B3 6 8.76 0.51 0.85 3.48 419.77 0.56 171.40 B4 8 8.82 0.83 0.95 3.36 412.39 0.71 171.60 B5 10 8.97 0.98 1.35 3.99 385.97 0.89 181.05 B6 11 8.88 1.10 1.38 4.12 415.34 1.40 149.02 B7 12 9.55 1.25 1.52 3.76 362.50 1.57 145.96 B8 13 8.98 1.30 1.75 4.00 348.05 1.59 158.25 B9 15 10.86 2.74 2.52 5.12 443.21 1.86 176.49 B10 16 10.04 2.08 1.26 3.19 337.59 1.49 177.77 B11 17 9.98 2.27 1.83 3.92 328.70 1.42 176.30 B12 19 8.28 1.96 1.21 3.15 296.37 1.30 180.21 B13 20 7.28 1.32 0.65 1.63 230.27 1.23 180.66 B14 23 6.48 0.62 0.36 0.70 204.75 0.83 83.77 B15 24 0.00 0.00 0.00 0.35 57.38 0.01 26.20 Table 6-2: Concentration (ppm) of trace metals with depth in Core B Sample ID Depth (cm) Pb (ppm) Zn (ppm) Na (ppm) K (ppm) Fe (ppm) Mn (ppm) S (ppm) C1 1 5.09 1.75 1.50 6.98 670.19 0.87 189.64 C2 2 5.25 0.41 1.22 5.87 625.62 0.91 184.30 C3 3 5.66 0.58 1.06 5.18 620.81 0.95 192.58 C4 4 6.14 1.32 1.06 6.32 636.08 0.88 181.62 C5 5 5.96 0.69 1.17 6.32 680.03 0.65 185.65 C6 6 5.64 0.63 1.22 5.71 646.30 0.58 177.46 C7 7 5.47 0.48 1.37 6.18 645.93 0.62 172.57 C8 8 5.33 0.70 0.90 5.60 703.88 0.50 159.69 C9 9 5.54 0.52 0.84 4.68 649.75 0.47 236.88 C10 10 5.98 1.03 2.37 6.28 558.85 0.44 154.41 C11 11 6.90 1.01 1.08 5.42 551.90 0.44 172.43 C12 12 7.20 1.19 1.28 4.72 568.77 0.42 173.66 C13 13 6.68 1.08 0.73 3.62 542.00 0.43 172.22 C14 14 6.66 0.93 0.86 4.61 560.50 0.45 166.54 C15 15 6.71 1.27 1.04 5.15 560.04 0.50 161.80 C17 17 7.39 1.27 1.12 3.40 436.59 0.82 141.76 C19 19 7.44 1.61 1.64 4.26 449.87 0.97 134.49 Table 6-3: Concentration (ppm) of trace metals with depth in Core C 106 Pb (Core B) Concentration (ppm) -1 4 9 Zn (Core B) Na (Core B) Concentration (ppm) Concentration (ppm) -0.5 0.5 1.5 2.5 -1 0 1 2 3 K (Core B) Concentration (ppm) 0 4 2 4 6 Fe (Core B) Mn (Core B) S (Core B) Concentration (ppm) Concentration (ppm) Concentration (ppm) 0 200 400 0 0.5 1 1.5 2 0 100 200 300 0 Depth (cm) 5 10 15 20 25 Figure 6-4: Concentration (ppm) of trace metals with depth in Core B Pb (Core C) Concentration (ppm) 4 5 6 7 8 0 Zn (Core C) Na (Core C) K (Core C) Fe (Core C) Mn (Core C) S (Core C) Concentration (ppm) Concentration (ppm) Concentration (ppm) Concentration (ppm) Concentration (ppm) Concentration (ppm) 0.5 1 1.5 2 0.5 1.5 2.5 3.5 3 5 7 9 400 500 600 700 0 0.5 1 1.5 2 100 150 200 250 300 0 Depth (cm) 5 10 15 20 25 Figure 6-5: Concentration (ppm) of trace metals with depth in Core C 107 Pb (Core A) 5 7 9 Na (Core A) Zn (Core A) Concentration (ppm) Concentration (ppm) 11 0 1 K (Core A) Concentration (ppm) 2 0.5 1.5 2.5 Fe (Core A) Concentration (ppm) 3.5 3 5 7 Mn (Core A) Concentration (ppm) 9 400 600 800 S (Core A) Concentration (ppm) Concentration (ppm) 0 0.5 1 1.5 2 115 135 155 175 0 5 Depth (cm) 10 15 20 25 Figure 6-6: Concentration (ppm) of trace metals with depth in Core A R. Pb (Core A) Concentration (ppm) 6 7 8 9 R. Zn (Core A) R. Na (Core A) R. K (Core A) R. Fe (Core A) R. Mn (Core A) R. S (Core A) Concentration (ppm) Concentration (ppm) Concentration (ppm) Concentration (ppm) Concentration (ppm) Concentration (ppm) -0.5 0 0.5 1 30 50 70 90 0 5 10 15 20 550 650 750 850 0 0.5 1 1.5 2 90 110 130 150 0 5 Depth (cm) 10 15 20 25 Figure 6-7: Reproduced concentration (ppm) of trace metals with depth in Core A 108 a depth of 4cm, is markedly more drastic in the reprocessed data than the original data. This would indicate either that the sample digestion of 1ml H2O2 and 7ml HNO3 is inaccurate or that there was an issue in the reprocessing procedure. To assess the accuracy of the reprocessed results, an attempt was made to deduce what values might be expected from the Jungle Falls stream based on environmental conditions. In 2008, geochemical analysis was carried out on sediments from NSSF in order to potential environmental change in the area over the Late Quaternary (Oon, 2008). The sodium levels in the surface sediments of NSSF ranged between 1ppm to 3ppm and potassium levels were between 5ppm and 8ppm. As both these locations are similar, being waterlogged and on Bukit Timah Granite, we would expect similar concentration levels in both sites. As such, the reprocessed data for Na and K is likely to be inaccurate. Similar attempts to find a “typical” concentration range for lead, zinc, iron, manganese and sulphur have proved more difficult as such values vary greatly between environments. The majority of trace metal paleolimnological studies of acidification are based in Europe and North America and have trace metal concentration values that are significantly higher than Jungle Falls stream. This is because these sites have also received far higher atmospheric contamination that Jungle Falls stream. For instance, in a study of Loch Coire nan Arr in north-west Scotland, Rose and Rippey (2002) found that prior to atmospheric contamination, sediments had zinc and lead concentrations between 0µg/g and 25µg/g. Following the commencement of atmospheric pollution, concentrations of zinc and lead rose to between 40 and 55µg/g. Lead levels increasing by a mean of 18µg/g is actually considered low as most contaminated lakes studied in the UK 109 had a mean lead concentration increase of over 250µg/g (Rose and Rippey, 2002). Staying within Scotland, Loch Laidon had zinc concentrations that rose from 100µg/g to between 200µg/g to 400µg/g and lead concentrations that rose from less than 100µg/g to between 200µg/g to 450µg/g (Flower et al, 1988). In Siberia, at Lake Kholodnoye, lead concentrations were low at around 1µg/g prior to atmospheric contamination and rising to a maximum of 6µg/g after. However, in the same location, zinc levels were 150µg/g prior to contamination, reaching a maximum of 300µg/g after (Flower et al, 1997). Even within Singapore, “typical” trace metal concentrations are highly variable. In a study of heavy metal contamination in mangrove sediments, Dang et al (2005) found that lead concentrations ranged from 12.28µg/g to 30.98µg/g and zinc concentrations ranged from 51.24µg/g to 120.23µg/g. While no other comparable data is available for mangrove sediments in Singapore, Dang et al (2005) note that the concentrations of heavy metal contamination in mangrove sediments in Singapore are significantly lower than levels from marine sediments in and around Singapore. These marine sediments include data from a study by Goh and Chou (1997) who looked at heavy metal levels in marine sediments collected from twenty coastal locations around Singapore. They found that concentrations of zinc ranged from 94.9µg/g to 281.3µg/g and noted that zinc concentrations were significantly different among the sediments. Lead concentrations ranged from 1.4µg/g to 82.2µg/g (Goh and Chou, 1997). Sediments were also collected from Punggol Estuary to assess the effect that ongoing reclamation, dredging, construction and shipping activities had on sediments and how this contamination would affect phytoplankton and bacteria (Nayar et al, 2004). Lead concentrations here had a minimum of 1.18µg/g, a mean of 17.30µg/g and a maximum value of 156.50µg/g. 110 Sulphur, iron and manganese values are equally difficult to predict. In a study of acidification in Lilla Öresjön, Sweden, sulphur levels range from below 10µg/g to 30µg/g even as lead levels were between 50µg/g and 250µg/g and zinc levels between 100µg/g to 400µg/g (Renberg et al, 1990). Yet, in South Lake, New York, upon undergoing acidification, sulphur concentrations increase from around 100µg/g to a high of close to 300µg/g; while in Ledge Pond, Maine, sulphur concentrations go from between 100µg/g to 200µg/g up to between 300µg/g to 400µg/g (Mitchell et al, 1985). Iron concentrations are often high in sediments, and concentrations levels are also variable, with values of 1000µg/g (Friese et al, 1998) up to 50000µg/g (Pienitz et al, 2006). Manganese concentrations have also been recorded from lows of 2.3µg/g (White et al, 1989) to highs of 4000µg/g (Flower et al, 1997). Clearly, trace metal concentrations vary greatly between sites and without a close analogue to the Jungle Falls stream, assessing the reliability of the original and reprocessed data is complicated. Unfortunately, had the concentration values of the reprocessed dataset been similar to that of the original dataset, this would have given greater confidence in the digestion methodology employed to measure sulphur concentration. The significant difference between the original and reprocessed data could be due to the different digestion procedures employed. A problem with the digestion procedure would most likely lead to an incomplete sediment digestion. As the multi-element digestion procedure is based on EPA standardised methodology, it should be the sulphur digestion procedure that is incomplete. This would mean that the digestion for the reprocessed multi-element was incomplete and thus that the reprocessed multi-element data would have lower values. However, this is not the case as sodium and potassium values increase, not decrease. 111 The difference between the original and reprocessed data could also be linked to a calibration issue during either one of the procedures. Should a miscalibration be the cause, the concentrations of the trace metals in either dataset should all either go up or go down since they were all calibrated using the same multi-element solution. Instead, lead and zinc values drop while sodium and potassium values rise. Because the original data for lead, zinc, sodium, potassium, iron and manganese was processed from an EPA recommended methodology, this data is probably more accurate than the reprocessed data. However it is unclear which sulphur concentration values are more accurate – original or reprocessed. As such, both the sulphur datasets were considered in the interpretation of the geochemical analysis results. Therefore, figure 6-8 shows the reprocessed sulphur graphs for all three cores. R. Sulphur (Core A) Concentration (ppm) 90 110 130 150 10 R. Sulphur (Core B1) R. Sulphur (Core B2) Concentration (ppm) Concentration (ppm) 60 110 160 110 130 150 R. Sulphur (Core C) Concentration (ppm) 170 100 120 140 0 5 Depth (cm) 10 15 20 25 Figure 6-8: Reproduced concentration (ppm) of sulphur with depth All three cores are required in order to view the entire sedimentary sequence of Jungle Falls stream. Only Core B captures the deepest sediments at depths of 23cm and 24cm; while only Core C has the youngest sediments at a depth of 1cm. Core A provides an additional verification of the data. The 112 sediments from Jungle Falls stream have yet to be dated. However, assuming that the stream was impounded in the late 1930s, based on Lum and Sharp (1996, see section 4.5), that sedimentation has been continuous and that the basin is not at capacity, this would imply a 24cm sedimentary core accumulated over around 75 years, a estimated sedimentary accumulation rate of approximately 0.3cm/yr. While sedimentary accumulation rates would vary between catchments and environments, a comparison with other studies of sedimentary accumulation shows that this estimate of 0.3cm/yr is possible. For instance, Szczucinski et al (2009) record an accumulation rate of 0.20-0.46cm/yr off the coast of Vietnam. A study of Tonle Sap, the ‘Great Lake’ in Cambodia, reveals sedimentary accumulation estimates as low as 0.01cm/yr to rates as high as 4cm/yr (Penny et al, 2005). As such, a base date of the late 1930s has been employed in the interpretation of these results. There are signs of atmospheric pollution and contamination in Jungle Falls valley based on the lead, zinc and sulphur concentration profiles. Concentration levels of these pollutants in the sediments start of low at the base of Core B, with both lead and zinc levels too low to be detected and sulphur concentration at 26.2ppm. Concentrations begin to rise dramatically before slowing down and peaking at a depth of 15cm. At this point, lead concentrations are at 10.9ppm, while zinc concentration is 2.74ppm and sulphur is at 176.49ppm. This rise at the base of the core could be due to changing catchment conditions compounded by the beginning of industrialisation in the country and consequently a rise in atmospheric pollution. Thus, sodium and potassium concentrations also begin low at around 0.5ppm and 1ppm respectively before rising and stabilising at around 1ppm for sodium and 4ppm for potassium. 113 The peak in values at a depth of 15cm could correspond to the mid to late 1960s. This was when industrialisation was rapid and pollution controls had yet to be implemented. Thus, atmospheric contamination and pollution was at its highest level. As sodium and potassium levels also peak at this depth, the question is whether this peak in lead, zinc, and sulphur is above that of natural variation. Unfortunately, levels of sodium, potassium, lead and zinc are all relatively low throughout the core, below 10ppm. Thus, any small variation can seem significant even though the change is as little as 1ppm or 2ppm. That being said, at this depth, sulphur levels increase by as much as 20-40ppm, which appears more significant. Furthermore, the diatom assemblage also changes at this depth. Diatom distribution variation could be caused by changes in temperature, turbulence, light availability, pH levels, nutrient availability and salinity (section 3.2). With the environment at BTNR unchanging over a short time-span of 70 years, any variation in diatom assemblage is likely due to anthropogenic influences. Thus, changing acidity is the most likely cause of the change in diatom assemblage, suggesting that atmospheric pollution was having an impact on the catchment and the acidity of the water was dropping. In the 1970s, with the enforcement of the Clean Air Act and effective management of industrialisation and urbanisation in Singapore, atmospheric pollution levels, and thus contamination into the catchment, would decrease. Therefore, while sodium and potassium levels appear to stabilise, or even increase (sodium and potassium levels are rising in Core A and potassium levels in Core C are also rising), lead and zinc levels are dropping. Sulphur levels at this point remain stable. 114 Approaching the top of the cores, lead levels remain low, returning to the concentration levels at the base of the core. This is because leaded petrol had been phased out in Singapore and atmospheric lead concentrations have shown a steady decline (figure 4-3). In contrast, zinc and sulphur levels appear to be increasing once again. This increase in zinc can be seen particularly in Core C (figure 6-5). In figure 6-5 and 6-8, sulphur levels also appear to be steadily rising. This rise in sulphur concentrations in Jungle Falls stream can also be seen in air quality monitoring data in Singapore. While average annual sulphur dioxide have been in the range of 10ppbv and 35ppbv since 1985, there has been a slight upward trend observed at most monitoring sites since 1991 (Bashkin, 2003). Again, it has to be acknowledged that sodium and potassium levels are increasing here, but the magnitude of the change is not as significant as that of zinc. Based on Table 6-3, from 2cm to 1cm in depth, concentration levels of sodium and potassium increase by 19% and 7% respectively. Zinc concentrations, on the other hand, increase by 324%. Figure 6-9 shows the reproduced sulphur concentration levels measured in Core A compared to available data on the changing total acidity levels in Singapore from Chin (2000). The reason why this core was chosen is because it best encapsulates the observations stated above, namely that sulphur levels peak at 15cm before dropping and stabilising, beginning to increase once again approaching the surface. As mentioned previously, while the sulphur concentration levels measured directly may be inaccurate due to incomplete digestion using 7ml HNO3 and 1ml H2O2, and the reprocessed sulphur concentration levels may be inaccurate due to the reprocessing procedure along with a digestion using 10ml of HNO3, however, the concentration profiles still show similar trends and it is this trend that is currently being examined rather than the absolute values. 115 Year' Sulphur'Concentra;on'(ppm)' 1973$ 1975$ 1977$ 1979$ 1981$ 1983$ 1985$ 1987$ 1989$ 1991$ 1993$ 1995$ 90$ 240$ 70$ 220$ 50$ 200$ 30$ 180$ 10$ 160$ !10$ 140$ !30$ 120$ !50$ 100$ !70$ 80$ Acidity'(µg/m3)' 1971$ 260$ !90$ 18$ 16$ 14$ 12$ 10$ 8$ Depth'(cm)' Sulphur$Concentra:on$with$Depth$ 6$ 4$ 2$ 0$ Annual$Urban$Adidity$Level$ Figure 6-9: Reproduced sulphur concentration levels from Core A compared to total acidity levels measured in Singapore (data for total acidity measurements from Chin, 2000). The sulphur concentration profile of the Jungle Falls catchment sediments in figure 6-9 appears to follow the total acidity levels measured in Singapore closely. Of particular note is the increase in total acidity from 1991 onwards in urban areas following the increase in sulphur concentration levels at a depth of 4cm. The reproduced sulphur concentration levels stop at a depth of 2cm as Core A has not captured the top surface sediments, unlike Core C. Note that this graph, would imply that sedimentation within the impoundment ceased by the mid-1990s. This could be because the dam reached full capacity during this time period. Aside from atmospheric contamination and changes in the erosional intensity, addressed by looking at the sodium and potassium levels in the sediments, another reason for the alteration of sedimentary trace metal concentrations would be diagenetic surface effects, namely redox-recycling processes. The redox-driven cycles of iron and manganese changes the amount 116 of sorption occurring and could lead to a mobilisation of lead and zinc from the sediment to the porewaters (Flower et al, 1997). This would then lead to a migration of trace metals with manganese and iron (Cornwell, 1986). Thus, a similarity between iron, manganese and other trace metal profiles would suggest that iron and manganese oxide enrichments have influenced trace metal distribution (Cornwall, 1986). Even though iron concentration profiles in Core A does appear to mirror lead and zinc profiles, it does not appear to do so in Core B and Core C, while manganese profiles are dissimilar. As such, the lead and zinc profiles in the Jungle Falls sediment are not likely to be affected by diagenetic impact. While diatom analysis is the preferred method for investigation paleolimnological acidification, it is interesting to note that geochemical evidence could be more sensitive to basin changes that biological changes. For instance, in Lake Kholodnoye, Siberia, Flower et al (1997) note that lead and zinc concentrations have increased since the 1920s due to atmospheric pollution, with the level of contamination accelerating after 1970. However, this atmospheric contamination of the lake was insufficient to impact the diatom community within and, as diatoms are also highly sensitive to environmental change, Flower et al (1997) believe that the lake ecosystem is currently not degraded by the atmospheric pollution into the catchment. This could explain why, though sulphur and zinc levels appear to be increasing at the top of the Jungle Falls sediment core, a concurrent signal is not seen in the diatom assemblage. 6.6 Limitations to Study The issues associated with the geochemical analysis of the core is mainly focussed on the accuracy of the results, and this has been discussed extensively above. Some caution is also required in the interpretation of the diatom assemblage as a lack of diatom studies in the region makes it harder to draw 117 conclusions on the change in diatom assemblage and preservation issues may have led to a biased assemblage. 6.6.1 Representativeness of Diatoms As can be seen from the description of the ecological preferences of the various diatom species in section 6.5.1, a single species can have different pH optimums depending on the location they are found in. This is because diatoms behave differently in different environments. Thus, it should be noted that comparing a local assemblage to species in another area could lead to a misinterpretation of data (Crosta and Koç, 2007). Jackson and Overpeck (2000) explain why species behaviour varies by expanding on G.E. Hutchinson’s concept of fundamental and realised spaces of taxa. According to them, every species has a specific environmental tolerance which gives them a range of possible environments to live in – their fundamental niche space (figure 6-10). Every environment will also have a specific condition of variables, the realised environmental space. Where the realised environmental space intersects the fundamental niche space, a potential niche space occurs where the species will appear and thrive. In that potential niche space, species will only populate a certain area – the realised niche. Thus, this resultant niche would be different in different realised environmental spaces. In other words, while there is a rich bank of information to tap into pertaining to diatom species and their ecological tolerances and preferences, it is preferable to use modern diatom analogues from the same geographic region as the study site (Jones, 2007). Thus, in order to fully understand the cause for the floristic changes seen in BTNR, there is a need for a regional diatom database, rather than a global one, and surface samples need to be collected and compiled. This entails choosing appropriate reference sites that encompass the range of environmental 118 Environmental(Variable(2( Environmental(Variable(1( Realised( Environmental( Space( Fundamental( Niche(Space( Potential(Niche( Space( Realised(Niche( Figure 6-10: Fundamental versus realised spaces (Jackson and Overpeck, 2000) conditions expected and comprise the taxa encountered in the BTNR Jungle Falls cores (Smol, 2008). 6.6.2 Preservation of Diatoms As mentioned in section 6.5.1, there are some concerns with the preservation of diatoms in Jungle Falls stream as a significant proportion of valves were broken, some valves showed signs of dissolution and diatom concentrations were low. While diatom valves can be broken during the treatment of sediments for diatom analysis, such as if the samples were centrifuged rather than allowed to settle overnight prior to decantation (Battarbee et al, 2001), however, Blanco et al (2008), experimenting with diatom slide quality when different treatments were used, found that treatment of samples had no 119 significant effect on the proportion of broken frustules. Steps were also taken to minimise diatom breakage (see section 5.3.3). Diatom preservation has been linked to six main factors – pH/alkalinity levels, salinity, temperature, silica content of water and water movements along with the shape of the fossils themselves (Ryves et al, 2006; Flower and Ryves, 2009). In general, preservation is often best in “cold, soft water lakes typical of boreal latitudes and poorest in warm alkaline or saline lakes in low latitudes” (Battarbee et al, 2001: 169). Silica dissolution increases exponentially once pH rises above 9 (Ryves et al, 2006; Barker, 1992). However, with Jungle Falls stream being acidic, pH is unlikely to be a factor in the poor preservation of diatoms. Dissolution also increases at higher salinities, even when salinity changes are small (Flower and Ryves, 2009). As a freshwater stream, salinity would not be significant in this study. Increasing temperature will increase dissolution rates as higher temperatures accelerate chemical reactions (Lewin, 1961). Furthermore, when studying diatom preservation, it was found that higher temperatures encouraged greater bacterial action on the organic matrix surrounding each diatom frustule. This increased particulate organic carbon hydrolysis by bacteria led to faster exposure of the siliceous fraction of diatoms, causing more rapid frustule dissolution (Bidle et al, 2002). According to Bidle et al (2002), with every 15oC rise in temperature, opal dissolution rates increase approximately 10-fold. While their study was based on marine diatoms, the same principle would apply to freshwater diatoms. For instance, in a study of sediments from North African lakes, Flower et al (2001) report that preservation was poor in some cores, likely due to high temperatures, along with water movement, pH, bioturbation and low diatom productivity. Logan et al (2010) reported that diatoms preservation from Moreton Bay, Australia, was exceedingly poor due to high temperatures and 120 saline waters. Being a tropical freshwater dam, temperature would be an important factor in the poor preservation of diatoms at BTNR. Furthermore, when a body of water is undersaturated with respect to silica, any silica surface that is exposed will undergo dissolution (Bidle et al, 2002). It is possible that the water in Jungle Falls stream has a low silica content, enhancing silica dissolution. This problem could be worsened with water movements. Should the sediments in the stream be re-suspended frequently, this would “inhibit the buildup of dissolved silica in upper sedimentary pore waters, which might slow or halt dissolution of a sedimented valve” (Ryves et al, 2006: 1361). Thus, the damming of the Jungle Falls stream, while enabling the collection of sediments, vital for the investigation of potential acidification in the stream, could also lead to a poor preservation of diatoms. This is because the drainage pipe that allows water to flow out of the impediment is located at the bottom of the brick wall. Consequently, there is a constant flow of water through the sedimentary matrix in order to exit the dam, preventing the build-up of silica in sedimentary pore waters and increasing diatom dissolution and breakage. As the bottom sandy layers in Core B are probably below this drainage pipe, this could explain why B13, B14 and B15 had the highest number of diatoms, reaching a maximum value of 641 diatoms in B15. According to Flower (1993), even in areas with a low pH level, if ground water movements occur, dissolution levels of diatoms will be extensive. Water depth and wind speed are also important factors in diatom preservation. In high-energy environments, such as nearshore, shallow and wave-mixed ones within lakes, valves are more likely to break apart, increasing the speed of dissolution (Ryves et al, 2006). Ryves et al (2006: 1361) found that “poor preservation in some relatively low-alkalinity and low-salinity shallow lakes ([...]... 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, 19 36) 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... small streams” (Patrick and Reimer, 1 966 cited in Ortiz-Lerín and Cambra, 2007: 4 26) It is an acidobiontic 100 Core A E incisa E af paludosa E flexuosa E rhomboidea E fallax E curvata E vanheurckii E pectinalis E parallela F af bicapitata Frustulia spp Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%)... 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... 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... anthropogenic influences Thus, changing acidity is the most likely cause of the change in diatom assemblage, suggesting that atmospheric pollution was having an impact on the catchment and the acidity of the water was dropping In the 1970s, with the enforcement of the Clean Air Act and effective management of industrialisation and urbanisation in Singapore, atmospheric pollution levels, and thus contamination into... (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... 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... because leaded petrol had been phased out in Singapore and atmospheric lead concentrations have shown a steady decline (figure 4-3) In contrast, zinc and sulphur levels appear to be increasing once again This increase in zinc can be seen particularly in Core C (figure 6- 5) In figure 6- 5 and 6- 8, sulphur levels also appear to be steadily rising This rise in sulphur concentrations in Jungle Falls stream... mangrove sediments in Singapore, Dang et al (2005) note that the concentrations of heavy metal contamination in mangrove sediments in Singapore are significantly lower than levels from marine sediments in and around Singapore These marine sediments include data from a study by Goh and Chou (1997) who looked at heavy metal levels in marine sediments collected from twenty coastal locations around Singapore. .. 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) 98 a b c d e f g h i j k l m n o p Plate 6- 6: A selection of other diatoms present in the sediment cores (a) Pinnularia abaujensis; (b) Navicula cryptocephala; (c) Eunotia af papilio; ... 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. .. 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. .. curvata E vanheurckii E pectinalis E parallela F af bicapitata Frustulia spp Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance (%) Abundance