Both benthic and planktonic foraminifera were observed in all our samples. Under the microscope, foraminifera showed little evidence of mineralization and infilling.
For all the 18 samples, total foraminifera abundance varies from 29 ind./g to 1 986 ind./g. The abundance of planktonic foraminifera fluctuates from 15 ind./g to 1 880 ind./g and the abundance of benthic foraminifera ranges from 14 ind./g to 242 ind./g. Total foraminiferal fluxes fluctuate from 681 ind./(ka·cm2) to 28 519 ind./(ka·cm2). The planktonic and benthic foraminiferal fluxes vary from 359 ind./(ka·cm2) to 26 992 ind./(ka·cm2) and from 323 ind./(ka·cm2) to 3 278 ind./(ka·cm2), respectively. Total foraminiferal fluxes are higher during the glacial periods (MISs 2–4 and MIS 6) than during the interglacial periods (MIS 1 and MIS 5) (Fig. 3).
The benthic/total foraminifera ratio (A) and fragmentation index (B) from U1446 vary along marine isotope stages; benthic foraminiferal flux (C), planktonic foraminiferal flux (D), total foraminiferal flux and sedimentation rate (E) from U1446 vary along marine isotope stages. CaCO3 flux from NGHP-1-19B varies along ages (Phillips et al., 2014) (F); CaCO3 flux (G), total organic carbon (TOC) flux and sedimentation rate (H) from MD161-19 vary along ages (Da Silva et al., 2017). Horizontal lines in A, B, C, D and E indicate the average values of the benthic/total foraminifera ratio, FI, and fluxes for each marine isotope stage. Stars mark the four samples with photos taken and shown in Fig. 4. Graphic lithology of U1446 referred to Clemens et al. (2016).
The planktonic foraminifera account for 26%–96% of the total foraminifera, with an average value of (70±19)%. The planktonic foraminiferal at Site U1446 are typically tropical–subtropical assemblages (Ding et al., 2006). In total, 29 species of planktonic foraminifer were identified. The dominant species with average abundance larger than 5% include Globigerinita glutinata, Globigerinoides ruber, Globigerina bulloides, Neogloboquadrina dutertrei, Globigerinoides sacculifer, and Globorotalia menardii. The secondary level include nine species with abundance of 1%–5%: Pulleniatina obliquiloculata, Globigerinoides tenellus, Globigerinella calida, Globorotaloides hexagonus, Neogloboquadrina pachyderma (dex.), Globigerinella aequilateralis, Orbulina universa, Globigerina falconensis, and Globigerina rubescens. The remainder of the assemblages is composed of low abundant species with abundance less than 1% such as Beella digitata, Globigerinella adamsi, Globigerinoides conglobatus, Globoquadrina conglomerata, Globorotalia crassula, Globorotalia crassaformis, Globorotalia inflata, Globorotalia scitula, Globorotalia theyeri, Globorotalia tumida, Neogloboquadrina pachyderma (sin.), and Turborotalita quinqueloba.
Site U1446 was drilled at water depth of 1 430 m. It lies well above the modern lysocline, which was found between 2 000 and 2 800 m in the northern Bay of Bengal according to Cullen and Prell (1984) and Belyaeva and Burmistrova (1985). CaCO3 mass accumulation rates (MAR) was estimated at two sites adjacent to U1446, NGHP-1-19B at 18º58′N, 85º39′E, 1 422 m of water depth (Phillips et al., 2014) and MD161-19 at 18º59′N, 85º41′E, 1 480 m of water depth (Da Silva et al., 2017). The results show similar trends with our reconstructed foraminiferal fluxes at Site U1446 (Fig. 3), with relatively higher values during glacial periods than during interglacial periods.
Phillips et al. (2014) suggested that the drastic increase of the CaCO3 MAR at Site NGHP-1-19B between 70 ka and 10 ka resulted from a higher productivity. Due to a weakened southwestern monsoon during the last glacial period, decreased freshwater input likely diminished stratification, allowing for increased mixing and nutrient availability, thus enhancing productivity (Phillips et al., 2014). Da Silva et al. (2017) also concluded that CaCO3 dissolution induced by anaerobic biogeochemical processes was unlikely and attributed the temporal variation of the CaCO3 MAR of MD161-19 to productivity rather than dissolution.
At Site U1446, however, the foraminiferal dissolution is striking. The microscope photos of two samples taken from interglacial intervals, U1446C-1H-1A 88–89 cm (0.89 m, in MIS 1) and U1446A-3H-3A 145–146 cm (19.01 m, in MIS 5), clearly show the poor state of foraminifera preservation (Figs 4a and c). Holes on chamber walls of many foraminifers and remaining keels of fully dissolved Globorotalia menardii specimens are clear evidence of a strong dissolution, which is readily reflected in the high values of the FI (22.7% and 29.5%). In addition, radiolarians and pyrite fragments are frequently observed in these samples. By contrast, the microscope photos of two samples from glacial intervals, U1446C-2H-5A 35–36 cm (14.51 m, in MISs 2–4) and U1446C-4H-3A 104–105 cm (32.06 m, in MIS 6), show much better foraminifera preservation (Figs 4b and d). The preservation of many more whole foraminifer specimens results in low FI values of 4.5% and 3.8%. Fragments and remains of gastropoda and bivalvia are also observed in these two samples (Figs 4b and d).
Microscope photos from samples of U1446C-1H-1A 88–89 cm (a), U1446C-2H-5A 35–36 cm (b), U1446A-3H-3A 145–146 cm (c) and U1446C-4H-3A 104–105 cm (d).
Overall, the FI record of Site U1446 ranges from 3.1% to 29.5%, with an average value of (12.8±7.4)%. The higher values (from 9.1% to 29.6%, average of around 17.1%) occur in MIS 1 and MIS 5, i.e., during the interglacial periods. This differs from the averaged values of around 5.5% and around 6.5% for MISs 2–4 and MIS 6 (Fig. 3). The benthic/total foraminifera ratio is higher during interglacial periods (from 0.29 to 0.73, average of 0.44) than glacial periods (from 0.04 to 0.2, average of 0.1) (Fig. 3). The benthic/total foraminifera ratio has been frequently interpreted as reflecting productivity factors such as food supply export to the seafloor (Berger and Diester-Haass, 1988), with a high ratio indicating high planktonic productivity. However, in samples from Site U1446, a clear inverse correlation between the benthic/total foraminifera ratios and total foraminiferal fluxes is observed (Fig. 3), suggesting that the benthic/total foraminifera ratio is likely related with carbonate dissolution and preservation at the seafloor rather than surface productivity and food supply to the deep sea. Total planktonic and benthic foraminiferal fluxes are much higher during glacial periods, varying from 6 264 ind./(ka·cm2) to 28 519 ind./(ka·cm2) with an average of around 16 076 ind./(ka·cm2), than those in interglacial periods, ranging from 681 ind./(ka·cm2) to 4 091 ind./(ka·cm2) with an average of around 2 526 ind./(ka·cm2) (Fig. 3).
The results of this study, therefore, show that dissolution is stronger during the interglacial periods (MIS 1 and MIS 5) than the glacial periods (MISs 2–4 and MIS 6), and the anti-correlation between dissolution indexes and foraminiferal fluxes suggests that dissolution has a significant impact on the foraminifera record at Site U1446 (Fig. 3).
Several mechanisms may be invoked to explain the dissolution above the lysocline. The most likely one is related to the decomposition of organic carbon inducing an increase of dissolved CO2 (decrease in pH) in the sediment interstitial water (Emerson and Bender, 1981; Peterson and Prell, 1985). Other mechanisms, more or less ultimately related to the release of CO2 during the oxidation of organic matter, have been invoked to explain above-lysocline dissolution: bacteria-induced dissolution (Freiwald, 1995), twilight zone dissolution (Schiebel et al., 2007) and CO2 maximum zone dissolution (Paulmier et al., 2011).
Being in the Indian summer monsoon core region, Site U1446 is heavily influenced by monsoon changes (Clemens et al., 2016). Previous studies suggested that the Indian summer monsoon had been stronger during interglacial periods than during glacial periods (Duplessy, 1982; Caley et al., 2011; Phillips et al., 2014). The enhanced Indian summer monsoon precipitation could affect the inputs of nutrients from increased river discharge and result in higher surface productivity (Sarma et al., 2013). In turn, such a higher productivity may have caused more organic carbon to reach the seafloor. The decomposition of labile organic carbon, by consuming O2 and releasing CO2, may have driven a stronger dissolution of foraminifera in the sediments and anoxia of the bottom water environment. The stratification induced by enhanced monsoon precipitation might also have blocked the vertical exchange of O2 and CO2 (O’boyle and Nolan, 2010; Sarma et al., 2013), enhancing the increase in bottom water CO2 associated to the decay of organic matter in the upper sediments. The observation of abundant pyrite fragments in interglacial samples is also a clear argument that suggests the occurrence of more anoxic bottom water environment during interglacial periods. When the monsoon became weaker, the lower supply of organic material and freshwater resulted in the better preservation of foraminiferal tests during glacial periods, which is observed in this study.
Because dissolution of carbonate material was stronger during interglacial periods than glacial periods, the higher sedimentary rates in MIS 1 and MIS 5 compared to MISs 2–4 and MIS 6 of Site U1446 are explained as reflecting the increase in terrestrial material input associated to the periods of strong summer monsoon. The TOC fluxes from adjacent Site MD161-19 (Fig. 3; Da Silva et al., 2017) do show higher values during interglacials than glacials. However, TOC fluxes estimated for Core MD161-19 combine the contribution of marine-produced organic carbon, with that from terrestrial organic carbon (Da Silva et al., 2017), which is more refractory (Canfield, 1994; Balakrishna and Probst, 2005). In addition, Site U1446 being located in a strongly incised continental slope (Clemens et al., 2016) and being characterized by strongly variable sedimentation rates, its carbon-carbonate preservation history may not be faithfully deduced from observations made on the nearby Site MD161-19. Thus, additional analyses, such as TOC, is called for performing on U1446 samples directly.
The planktonic and benthic foraminiferal fluxes, and the relative abundance of high-productivity foraminifer species are usually considered as being robust paleo-productivity indexes (Herguera and Berger, 1991; Ding et al., 2006). Nevertheless, in some occasions, dissolution can strongly bias these proxies, as clearly shown in our observations of Site U1446 samples. Furthermore, the developments of benthic foraminifera populations and the type of species on the seafloor are also affected by the decrease in O2 content associated to the development of anoxic conditions (Naidu and Malmgren, 1995; Chen et al., 2011). Therefore, to better understand the mechanisms that drive above-lysocline dissolution at Site U1446, some additional work is mandatory: (1) precisely reconstruct the evolution of organic matter content and its oxygen, hydrogen and nitrogen compositions (i.e., to decipher the relative contribution of labile versus refractory organic matter, and marine versus continental organic matter); and (2) reconstruct paleoproductivity based on proxies that are not susceptible to strong dissolution bias (e.g., siliceous productivity indicators).
The strong difference of glacial period and interglacial period in foraminiferal dissolution and preservation that we have observed in Site U1446 can severely bias foraminifer-based micropaleontological and geochemical analyses, and noticeably impact paleoceanographic and paleoclimatic reconstructions in this region. In order to assess this potential impact and try to correct for those effects of preservation and dissolution, further studies will be necessary to tackle the impact of dissolution on key geochemical proxies (e.g., the ratio of Mg/Ca based thermometer).