UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Applications of U-decay series isotopes to studying the meridional overturning circulation and particle… Luo, Yiming 2013

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2013_spring_luo_yiming.pdf [ 5.16MB ]
Metadata
JSON: 24-1.0073792.json
JSON-LD: 24-1.0073792-ld.json
RDF/XML (Pretty): 24-1.0073792-rdf.xml
RDF/JSON: 24-1.0073792-rdf.json
Turtle: 24-1.0073792-turtle.txt
N-Triples: 24-1.0073792-rdf-ntriples.txt
Original Record: 24-1.0073792-source.json
Full Text
24-1.0073792-fulltext.txt
Citation
24-1.0073792.ris

Full Text

Applications of U-decay series isotopes to studying the meridional overturning circulation and particle dynamics in the ocean by  Yiming Luo  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  The Faculty of Graduate Studies  (Oceanography)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2013  © Yiming Luo 2013  Abstract Two important questions in the fields of paleoceanography and marine biogeochemistry (the reconstruction of past changes in the strength and geometry of the ocean’s overturning circulation and the quantification of particle flux to the seafloor) are addressed using three isotopes from the U-decay series (234Th, 230Th and 231Pa). Two-dimensional scavenging models of the Atlantic and Pacific Ocean were tuned to reproduce the 230Th and 231Pa seawater activity profiles measured in these oceans and used to establish the distribution of sediment  231  Pa/230Th generated by simple meridional  overturning circulation cells. The results indicate that circulation is the main factor controlling the distribution of sediment  231  Pa/230Th in the Atlantic and confirm the use of  this proxy as a paleocirculation tracer. In the Pacific, both circulation and boundary scavenging are important in determining the distribution of sediment 231Pa/230Th. Thorium-234 scavenging and moored sediment traps yield similar particle flux estimates in Saanich Inlet, on the coast of British Columbia. This study highlights the possibility of estimating the flux of organic carbon in coastal waters by simply measuring  234  Th and  POC on particles, which would provide a simple and rapid method for large scale monitoring. Measurements of 234Th and 230Th dissolved in seawater and adsorbed on three different size classes of particle were used to estimate particle flux in the epipelagic and mesopelagic zone of the ocean at station Papa. The results suggest that a significant fraction of the carbon flux can be associated with very large, rapidly-sinking particles with very low Th activities, and unaccounted for in Th-based flux estimates. ii  Preface Contributions of Authors  1) A version of chapter 2 was published as: Luo, Y., Francois, R., and Allen, S. E.: Sediment  231  Pa/230Th as a recorder of the rate of the Atlantic meridional overturning  circulation: insights  from  a  2-D  model,  Ocean  Sci.,  6,  381-400,  doi:  10.5194/os-6-381-2010, 2010.  The dissolved and particulate  231  Pa and  230  Th data were collected by Dr. Roger  Francois and Dr. Mike Bacon. I developed the model with the help from Dr. Susan Allen and I tuned the model with my supervisor Dr. Roger Francois. I wrote the manuscript with Dr. Roger Francois and Dr. Susan Allen.  Some of the results from chapter 2 were also used in  Gherardi, J.-M., Y. Luo, R. Francois, J. F. McManus, S.E. Allen, and L. Labeyrie 2010. Response to Comment by S. Peacock on “Glacial-interglacial circulation changes inferred from  231  Pa/230Th sedimentary record in the North Atlantic region”.  Paleoceanography, Vol. 25, PA2207, 5 PP., 2010 doi: 10.1029/2009PA001867.  and  iii  Lippold, J., J.-M. Gherardi, and Y. Luo (2011), Testing the 231Pa/230Th paleocirculation proxy: A data versus 2D model comparison, Geophys. Res. Lett., 38, L20603, doi: 10.1029/2011GL049282  2) A version of chapter 3 has been published in Nature Geosciences as: Lippold J., Luo Y., Francois R., Allen S.E., Gherardi J., Pichat S., Hickey B. and Schulz H.: Strength and geometry of the glacial Atlantic Meridional Overturning Circulation, Nature Geosci 5 (2012) 813-816.  J.L., Y.L. and R.F. developed the concept and designed the study. J.L., J.G., S. P. and B.H. performed  231  Pa/230Th measurements. J.L., Y.L. and R.F. performed opal  measurements. Y.L., R.F. and S.A. developed and applied the model. S.P., J.G and H.S. provided sample material and age models. J.L., Y.L. and R.F. wrote the manuscript.  3) A version of chapter 4 will be submitted as: Luo Y., Francois R. and Allen S.E. The influence of deep water circulation on the distribution of  231  Pa and  230  Th in the water  column and sediments of the Pacific Ocean.  The dissolved and particulate  231  Pa and  230  Th data were collected by Dr. Roger  Francois. I developed and tuned the model with Dr. Roger Francois and Dr. Susan Allen. The manuscript is being written in collaboration with Dr. Roger Francois and Dr. Susan Allen.  iv  Table of Contents Abstract…………………………………………………………………………….ii Preface……………………………………………………………………………..iii Table of Contents…………………………………………………………………v List of Tables…………………………………………………………………...……xiv List of Figures……………………………………………………………………….xvi Acknowledgements……………………………………………………………….xxvi Dedication………………………………………………………………………xxvii 1  Introduction……………………………………………………………….……..1  1.1 Ocean’s role in climate change……………………………………………………….1 1.2 U-series isotopes………………………………………………………………………4 1.3 Uranium budget in the ocean……………………………………………..…………7 1.4 Thorium and protactinium in the ocean…………..…………………………………9 1.4.1  Particle flux from the euphotic zone – (234Th:238U)…………………………..11  1.4.2  230  1.4.3  Sediment (231Paxs/230Thxs) to monitor past changes in the meridional overturning  Th normalization to reconstruct particle flux to the seafloor……………….12  circulation of the ocean…………………………………………………………...14 1.5 Thesis objectives………………………………………………..……………………17  v  2  Sediment  231  Pa/230Th as a recorder of the rate of the Atlantic meridional  overturning circulation: insights from a 2-D Model.........................................20 2.1 Introduction……………………………………………………………… ..……20 2.2 Model descriptions……………………………………………………………...21 2.2.1  Formulation……………………………………………..……………….…21  2.2.2  Overturning circulation…………………………………………..………..……..26  2.2.3  Parameterization……………………………………………………………. ..29  2.3 Dissolved  230  Th and  231  Pa water column profiles: Data-Model comparison...40  2.4 Fractionation factor: Data-Model comparison………………………………47 2.5 230Th and  231  Pa distribution in the control run……………………………..…50  2.5.1  Dissolved  230  2.5.2  Particulate  230  2.5.3  Dissolved  231  2.5.4  Particulate  231  2.5.5  Dissolved  231  2.5.6  Particulate  2.6 Sediment  231  Th…………………………………………………………..…51 Th……………………………………………………...………52  Pa……………………………………….……………………..55 Pa……….………………………………………………………56  Pa/ 230 Th……………………………………………………….58  231  Pa/ 230 Th……………………………………………………….61  Pa/230 Th: Data-Model comparison……………………………….63  2.7 Discussion………………………………………………………………………..66 2.7.1  The effect of AMOC on sediment  2.7.1.1 Vertical variations in sediment  231  Pa/230Th…………………………...……....66  231  2.7.1.2 Horizontal variations in sediment  Pa/230Th induced by the AMOC………….66 231  Pa/230Th induced by the AMOC……....68  vi  2.7.1.3 Changes in sediment  231  Pa/230Th resulting from changes in the rate of the  AMOC………………………………………………………………………..71 2.7.1.4 Changes in sediment  231  Pa/230Th resulting from changes in the geometry of the  AMOC………………………………………………………………….…….71 2.7.1.5 Possible sampling strategy to constrain past changes in AMOC from sediment 231  Pa/ 230 Th…………………………………………………………………...72 231  Pa/230Th………………………………73  2.7.2  The effect of AABW on sediment  2.7.3  The effect of particle composition on sediment  231  Pa/230Th……………….74  2.8 Conclusions………………………………………………………………… ..77 3  Strength and geometry of the Glacial Atlantic Meridional Overturning Circulation………………………………………………………………………80  3.1 Introduction……………………………………………………….……………..80 3.2 Materials and Methods…………………………………………...……………83 3.2.1  Core locations and chronology………………….…………..……………..83  3.2.2  Analytical methods ………………………………….………….……………84  3.2.2.1 231Pa/230Th data…………………………………………………….……….……85 3.2.2.2 Biogenic silica……………………………………………………..………..……85 3.2.3  Modeling……………………………………………………………………..85  3.3 Results and discussion………………………………………………………….90 3.3.1  The influence of biogenic silica on the distribution of sediment  231  Pa/230Th in  Atlantic sediments……………………………………………………………90  vii  3.3.2  The influence of AMOC on the distribution of sediment  231  Pa/230Th in Atlantic  sediments…………………………………………………………………….93 3.3.2.1 Holocene……………………………………………………………………..93 3.3.2.2 LGM………………………………………………………………………….96 3.3.3  Sensitivity test on the LGM 2-D scavenging model…………………………98  3.3.3.1 Sensitivity to sinking rate and fractionation factor…………………..……..98 3.3.3.2 Sensitivity to the strength of the overturning circulation cells (GNAIW; AABW)………………………………………………………………………100 3.3.3.3 Sensitivity to the geometry of the overturning circulation cells (GNAIW; AABW)……………………………………………………………………102 3.3.4  Can the Holocene circulation scheme explain the LGM observations by varying the scavenging parameters?…………………………………………………..104  3.4 Conclusions……………………………………………………………………107 4  The influence of deep water circulation on the distribution of  231  Pa and  230  Th in  the water column and sediments of the Pacific Ocean…………………….……109 4.1 Introduction……………………………………………………………………...109 4.2 Model descriptions…………………………………………………..…….……112 4.2.1  Overturning circulation……………………………………………………..112  4.2.2  Formulation of the two-dimensional scavenging model in the Pacific Ocean…………………………………………………………………… ..114  4.2.3  Parameterization of the scavenging model…………………………..……..116  viii  4.2.4  Estimating removal by “boundary scavenging” in the 2 -D scavenging model…………..……………………..…….…..…….…..………………....119  4.3 Results and discussion…..…….……..…….…..………………………………....119 4.3.1  230  Dissolved  Th and  231  Pa water column profiles…………………...………119  4.3.1.1 Data-model comparison………………...……………..…………...………119 4.3.1.2 The influence of boundary scavenging on the curvature of the  231  Pa seawater  profiles in the North Pacific……...…………..……...……………...………130 4.3.2  231  Pa and  230  4.3.2.1 Dissolved 4.3.2.2 Particulate 4.3.2.3 Dissolved 4.3.2.4 Particulate  Th sections generated by the Pacific 2-D scavenging model….131  230  Th...…………..……...……………………………………..…132  230  231  Th..……...……………………………………………………133  Pa……………………………………….…………….………135  231  Pa………………….…………….….……….……….………137  4.3.2.5 Dissolved Pa/Th….…….…………………….……………………………138 4.3.2.6 Particulate and sedimentary Pa/Th………………..………...………………139 4.3.3  Comparison between sediment Pa/Th measured in the Pacific and model predictions……………..………......……………..………...……………….143  4.3.4  Is the decreasing trend in Pa/Th with depth at mid-depths a result of the PMOC?.……………….……………………………..……………………….151  4.3.5  Sensitivity of sediment Pa/Th in low productivity regions to changes in the rate of PMOC…………….……………………………………………….………….152  ix  4.3.6  Relative sensitivity of sediment Pa/Th in low productivity regions to changes in PMOC and “boundary scavenging” – What is the most promising approach to constrain PMOC from sediment Pa/Th? …………………………...……157  4.4 Conclusions………………………………………………………………………159 5  A comparison of POC fluxes recorded by sediment trap and  234  Th: 238 U  disequilibrium in a coastal region (Saanich Inlet, British Columbia)……...161 5.1 Introduction…………………………………………………………………….…161 5.2 Measuring the sinking flux of carbon with sediment traps………………...163 5.3 Measuring the sinking flux of carbon using  234  Th/238U disequilibrium in surface  waters……………………………………………………………………...…..164 5.4 Materials and methods…. …. …. ………………………………………………..168 5.4.1  Study Site…………………………………………………………………..168  5.4.2  Sample collection, preparation and analyses………………………………170  5.4.2.1 Hydrography………………………………………………………..……..…170 5.4.2.2 Sediment traps………………………………………………………………171 5.4.2.3 234 Th method…………………………………………………………………172 234  Th: ……………………………………………………………..173  5.4.2.3.1  Total  5.4.2.3.2  Particulate  234  Th: …………………………………...…………………..174  5.5 The U-salinity relationship in Saanich Inlet waters…………………………175 5.6 Results and discussion……………………………………………….…………179 5.6.1  Salinity, temperature, density and O 2……………………………………….179  5.6.2  Sediment traps…………………………………………………….…………182 x  5.6.2.1 Sample mass and concentrations……………………………………………182 5.6.2.2 Fluxes from the sediment traps……………………………………………..185 5.6.3  234  Th deficits…………………………………………………………….……..192  5.6.3.1 Total  234  Th profiles………………………………………………………….193  5.6.3.2 Particulate 5.6.3.3 Dissolved  234  Th…………………………………………….……………….198  234  Th………………………………………………………….……..204  5.6.3.4 POC/ 234 Th ratio……………………………………………………………..205 5.6.3.5 Comparison between LVP samples and 25mm TQ samples.……………….209 5.6.3.6 Fluxes derived from  234  Th:238U disequilibria………………………………210  5.6.3.6.1  234  5.6.3.6.2  POC fluxes……………………………………………………………218  Th fluxes…………………………………………………………….210  234  5.6.4  Comparing POC flux measured with sediment traps and  5.6.5  Can POC fluxes in the coastal region be estimated by only measuring particulate 234  Th deficit…223  Th and POC? ………………………………………………………..226  5.7 Conclusions and future perspectives…………………………………………232 6  Particle fluxes and dynamics in the northeast Pacific Ocean from paired water column measurements of Th-230 and Th-234 activity……………….235  6.1 Introduction………………………………………………………………..………235 6.2 Materials and methods………………………………………….………..…...…238 6.2.1  Sample collection and preparation……………………………………...……….238  6.2.1.1 234 Th samples……………………………………………………..…………238 6.2.1.2 POC………………………………………..…………………………………242 xi  6.2.1.3 230 Th samples……………………………………………………..…………242 6.2.1.4 232Th, P, Al and Ca samples………………………..……………………..…244 6.2.2  Sample analysis……………………………………………………………..244  6.2.2.1 234 Th measurements…………………………………………………………244 6.2.2.2 POC measurements……………………………………………………….…245 6.2.2.3 230 Th measurements…………………………………………………………246 6.2.2.4 232Th, P, Al and Ca measurements………………………………..………246 6.3 Results……………………………………………………………………………249 6.3.1  234  6.3.2  POC……………………………………………………………………….….253  6.3.3  POC/234 Th.…………………………………………………………….……..255  6.3.4  230  6.3.5  POC/230 Th…….. ……………………………..……………………………..260  6.3.6  P and Ca……………………………………………………………..………261  6.3.7  Al………………………………………………………………………..……262  Th…………………………………………………………………………..249  Th…………………………………………………………………………..256  6.4 Discussions………………………………………………………………..…..…264 6.4.1  Th-234 fluxes ………………………………………………………………264  6.4.2  Fluxes of POC, Ca and Al estimated from the  234  Th:238U deficit………267  6.4.2.1 POC fluxes……………………………………………………….…….……267 6.4.2.2 Ca fluxes……………………………………………………………………..270 6.4.2.3 Al fluxes……………………………………………………………………..271 6.4.3  230  Th xs fluxes……………………………………………………………..…..276 xii  6.4.4  Fluxes of POC, Ca and Al estimated by normalization to  230  Thxs……..277  6.4.4.1 POC fluxes…………………………...………………………………….…..278 6.4.4.2 Ca fluxes…………………………………………………………………..….280 6.4.4.3 Al fluxes…………………………………………………………………..….282 6.4.5  Particle dynamics…………………………………………...…………….…283  6.4.5.1 Model A……………………………………………………………………...284 6.4.5.2 Model B………………………………………………………………….…..288 6.4.5.3 Model C………………………………………………………………………..294 6.5 Conclusions………………………………………………………….……..……303 7  Conclusions and perspectives…………………………………..…………...306  7.1 Summary of major findings and contributions…………………………...............306 7.2 Future research perspectives………………………. …………………...................311 Bibliography……………………………………………………………..………….314 Appendix A…………………………………………………………………………333 Appendix B…………………………………………………………………………346 Appendix C………………………………………………………………….….…..382 Appendix D………………………………………………………………….……...409 Appendix E…………………………………………………………………………432  xiii  List of Tables Table 1.1 Known inputs and outputs of U to the ocean……..……………………………..8 Table 1.2 Summary of major U, Th and Pa isotope data in world’s ocean………………10 Table 2.1 List of abbreviations and values for the model parameters…………………29 Table 2.2  230  Th and  231  Pa activities in sea water..…………………………………….….32  Table 2.3 “Equilibrium” Fractionation Factors……………..……………………….……39 Table 2.4 Holocene Pa/Th in 5 North Atlantic cores…………………………..…….…64 Table 3.1 List of abbreviations and values for the scavenging parameters……………..87 Table 3.2 Latitudinal variations of the equilibrium Fractionation Factors………….…89 Table 4.1 List of abbreviations and values for model parameters………………………117 Table 4.2 “Equilibrium” Fractionation Factors for the Pacific Model…………….……119 Table 4.3  230  Th and  231  Pa activities in sea water………………………….……………121  Table 5.1 Saanich Sampling schedule…………………………………………………171 Table 5.2 Dissolved U concentrations in Saanich Inlet……..…………………………177 Table 5.3 The comparison of fluxes…...……………..…………………………..……..190 Table 6.1 List of samples……………………………………………………………239 Table 6.2 List of Thorium isotope data……………..………………………………247 Table 6.3 POC and POC/234Th on particulate samples…………..…………………248 Table 6.4 Data for particulate P, Al and Ca………………..…………..……….…252 Table 6.5 POC and POC/230Thxs on particulate samples………….……..…………259 Table 6.6 Th-234 fluxes and POC fluxes based on  234  Th………..………………266 xiv  Table 6.7 Biogenic Ca/Th and Al/Th ratios for both  234  Th and 230Thxs……...…………271  Table 6.8 Biogenic Ca and Al fluxes calculated by 234Th and 230Thxs methods……...…274 Table 6.9 POC fluxes calculated by  230  Thxs normalization……………………………275  xv  List of Figures Figure 1.1 Uranium decay series………………..…………………………………………5 Figure 1.2 The principle of 234Th:238U method…..……………………………………….13 Figure 2.1 The scavenging model…………………………………………………….......25 Figure 2.2 Velocity vector plot………………………………………………………........27 Figure 2.3 Station locations for the water column profile..………………………..….....31 Figure 2.4 Dissolved 230Th and 231Pa at 60°- 70°N...…………..……………………......41 Figure 2.5 Dissolved  230  Figure 2.6 Dissolved  231  Th…………………………………………………………….....45 Pa…………………………………………………………….....46  Figure 2.7 Concentration profiles of dissolved  230  Th and 231Pa in the Southern Ocean…47  Figure 2.8 Distribution of fractionation factors……………………………………….....49 Figure 2.9 Dissolved  230  Figure 2.10 Particulate Figure 2.11 Total  Th section generated by the model………………………….....51  230  Th section generated by the model……………………….....52  230  Th concentration measured at Station KNR06-3……………….....53  Figure 2.12 Fraction of total  230  Th associated with particles generated by the model.....54  Figure 2.13 Dissolved 231Pa section generated by the model………………………….....55 Figure 2.14 Particulate  231  Pa section generated by the model…………..…………….....57  Figure 2.15 Fraction of total Figure 2.16 Total  231  Pa associated with particles generated by the model….57  230  Th concentration measured in the southwestern Atlantic……….....58  Figure 2.17 Dissolved  231  Pa/230Th section generated by the model………………….....59  Figure 2.18 Dissolved  231  Pa/230Th profiles in the North and Equatorial Atlantic…….60 xvi  231  Pa/230Th section generated by the model………………….....61  Figure 2.19 Particulate Figure 2.20 Sediment  231  Pa/230Th section generated by the model………………….....62  Figure 2.21 Sediment  231  Pa/230Th generated with an opal belt…………………..….....65  Figure 2.22 Lateral velocity profile…………………………………………….…….....69 Figure 2.23 Contrasting lateral velocity profiles between the control run…...…….....70 Figure 2.24 Sediment  231  Pa/230Th field generated without AABW………….……..74  Figure 2.25 The influence of Southern Ocean fractionation factor on the sediment 231  Pa/230Th in Atlantic sediments…………………………………………………….....76  Figure 3.1 Core locations…………………………………………………………….....84 Figure 3.2 Holocene overturning scheme………………………………………………86 Figure 3.3 LGM overturning scheme……………………………………………………87 Figure 3.4 Overturning scheme used to test the very shallow………………………….90 Figure 3.5 Correlation between sediment  231  Figure 3.6 Correlation between sediment  231  Figure 3.7  Pa/230Th and  230  Th-normalized opal flux...91  Pa/230Th and opal concentration…………93  231  Pa/230Th versus water depth……………….………………………….....94  Figure 3.8 Correlation between sediment  231  Pa/230Th and model outputs……………95  Figure 3.9 Variations in the linear correlation between the sediment  231  Pa/230Th database  and model output…………………………………………………………………….....99 Figure 3.10 Variations in the mean square weighed deviation (mswd) between the sediment  231  Pa/230Th database and model output………………………………….…99  Figure 3.11 Fit between observations and model outputs generated with the optimal LGM model geometry with varying GNAIW and AABW strengths……………………….....100 xvii  Figure 3.12 Fit between observations and model outputs generated with a very narrow GNAIW as a function of GNAIW and AABW strengths……………………..….....102 Figure 3.13 Sediment  231  Pa/230Th superimposed to the sediment  231  Pa/230Th section…105  Figure 3.14 Linear correlation coefficient and mean square weighed deviations obtained when correlating the sediment database with sediment  231  Pa/230Th generated by the  Holocene circulation scheme with varying particle sinking rates and fractionation factors……………………………………………………………………….........106 Figure 4.1 Velocity vector plot………………………………………………………...113 Figure 4.2 Scavenging and transport model in each model grid box……………............116 Figure 4.3 Station locations………………………………………….………………......121 Pa and  230  Figure 4.5 Dissolved  230  Figure 4.4  231  Figure 4.6 Dissolved  Th profiles measured in the Southern Ocean……………..128  Th and  230  Th and  231  231  Pa at equatorial Pacific……….…….…………...129  Pa at station Aloha………………………………...130  Figure 4.7 Changes in the curvature of the dissolved  231  Pa profile generated at 21°N with  26Sv PMOC and varying effective rate constant for removal to the margins……….....131 Figure 4.8 Dissolved  230  Th section generated by the model…………………………...132  Figure 4.9 Difference between the dissolved  230  Th concentration generated by the 2-D  model and the concentration predicted in the absence of circulation………………...133 Figure 4.10 Particulate  230  Th section generated by the model……………..…………...134  Figure 4.11 Fraction of total  230  Th associated with particles generated by the model…134  Figure 4.12 Dissolved 231Pa section generated by the model…………………………...136  xviii  Figure 4.13 Difference between the dissolved  231  Pa concentration generated by the 2-D  model and the concentration predicted in the absence of circulation……………....136 Figure 4.14 Particulate  231  Pa section generated by the model………………………….137  Figure 4.15 Fraction of total Figure 4.16 Dissolved  231  Figure 4.17 Dissolved  231  Figure 4.18 Particulate  231  231  Pa associated with particles generated by the model…138  Pa/230Th section generated by the model…………………...138 Pa/230Th below 1000m……………………………………....140 Pa/230Th section generated by the model…………………...141  Figure 4.19 Sedimentary 231Pa/230Th section generated by the model………………....142 Figure 4.20 Sediment  231  Pa/230Th generated by the 2D model at 21°S and 21°N……142  Figure 4.21 Distribution of core-top Pa/Th data……………………………..………...144 Figure 4.22 Sediment Pa/Th measured in the Pacific vs. sediment Pa/Th generated by the 2D scavenging model…………………………………………………………….….....145 Figure 4.23 Pa/Th in core tops versus depth……………………………………….....147 Figure 4.24 (a)  230  Th-normalized opal flux against sediment Pa/Th in the equatorial  Pacific; (b) Difference between Pa/Th measured and estimated from  230  Th-normalized  fluxes as a function of depth and the linear regression…………………………...….....148 Figure 4.25 Vertical sediment Pa/Th profiles generated between 35°N and 45°S by the 2D scavenging model in the absence of PMOC and with varying boundary scavenging strength…………………………………………………………………………….....151 Figure 4.26 Variations in sediment Pa/Th as a function of latitude in low productivity regions predicted by the 2D scavenging model with varying PMOC rates…………154  xix  Figure 4.27 Changes in the sediment Pa/Th vertical gradient between deep and intermediated depth as a function of latitude and PMOC strength………………155 Figure 4.28 Difference in sediment Pa/Th generated by the 2D scavenging model with PMOC = 26Sv and 13Sv…………………………………………….…………….....156 Figure 4.29 Variations in sediment Pa/Th as a function of latitude in low productivity regions predicted by the 2D scavenging model with varying boundary scavenging….158 Figure 4.30 Changes in the difference in sediment Pa/Th between 3000 and 4750m as a function of latitude for a fixed PMOC (26Sv) and varying effective removal rate constant to the margins……………………………………………………………………….....159 Figure 5.1 Map of Saanich inlet and coastal southwestern British Columbia………….169 Figure 5.2 Time-series hydrographic data from surface to 200m during the 2 year period………………………………………………………………………………...181 Figure 5.3 Seasonal changes in Opal% and OM% in the material collected by sediment traps at 3 depths in Saanich Inlet…………………….………………………………...184 Figure 5.4 (a) Seasonal variations in the OC/N ratio of sediment trap material collected at the three depths. (b) OC versus N for all the sediment -trap samples……186 Figure 5.5 Mass fluxes recorded by sediment traps deployed at three different depths over the 2-year time series…………………………………………………………...187 Figure 5.6 Opal fluxes recorded by sediment traps deployed at three different depths over the 2-year time series……………………………………………………………...188 Figure 5.7 POC fluxes recorded by sediment traps deployed at three different depths over the 2-year time series…………………………………………………………………...189 xx  Figure 5.8 Total  234  Th activities measured……………………………………………...194  Figure 5.9 Average total 234Th measured at a given depth……………………………...195 Figure 5.10 Total  234  Th during (a) late winter, (b) early spring, (c) late summer or early  fall…………………………………………………………………………………...197 Figure 5.11 Monthly composite time-series of total Figure 5.12 Average fraction of particulate  234  Th……………………………197  234  234  Th (% of total  Th) as a function of water  depth………………………………………………………………………………....198 Figure 5.13 Particulate  234  Th activities over an annual cycle…………………………199  Figure 5.14 Average particulate Figure 5.15 Particulate  234  Th measured at a given depth……………………199  234  Th during (a) late winter, (b) early spring, (c) late summer or  early fall………………………………………………………………………………...200 Figure 5.16 Monthly composite time-series of particulate 234Th………………………..201 Figure 5.17 Average fraction of dissolved 234Th as a function of water depth………….201 Figure 5.18 Dissolved  234  Th activities over an annual cycle…………………………...202  Figure 5.19 Average dissolved 234Th profile…………………………………………….202 Figure 5.20 Dissolved  234  Th during (a) late winter, (b) early spring, (c) late summer or  early fall………………………………………………………………………...............203 Figure 5.21 Monthly composite time-series of dissolved  234  Th………………………204  Figure 5.22 POC/234Th on particulate samples over an annual cycle………………..206 Figure 5.23 Average of all the POC/234Th measured in particles collected at a given depth………………………………………………………………………………..206  xxi  Figure 5.24 POC/234Th during (a) late winter, (b) early spring, (c) late summer or early fall………………………………………………………………………................207 Figure 5.25 Monthly composite time-series of POC/234Th……………………………208 Figure 5.26 Comparison between POC, particulate  234  Th and POC/234Th obtained by  collecting particles with Large Volume in-situ Pumps (LVP) and small volume filtration………………………………………………………………………................209 Figure 5.27  234  Th fluxes calculated from estimates of  238  U seawater concentration  obtained from equation 5.6 (blue line) and 5.10 (red line)……………………………212 Figure 5.28 Average calculated  234  Th flux profile…………………………………213  Figure 5.29 (a) 234Th fluxes in late winter and during the spring bloom; (b) 234Th fluxes in late summer and early fall………………………………………………………………214 Figure 5.30 Th-234 fluxes calculated with the steady state model vs the non-steady state model………………………………………………………………………...........216 Figure 5.31 Ratio of fluxes estimated with the non-steady state model to those estimated with the steady state model from  238  U seawater concentration estimates obtained from  equation 5.5………………………………………………………………………..........217 Figure 5.32 Average POC fluxes derived at a given depth with the steady state model from 238  U seawater concentration estimates obtained from equation 5.5………………219  Figure 5.33 (a) POC fluxes in late winter and during the spring bloom, (b) POC fluxes in late summer or early fall………………………………………………………………..220  xxii  Figure 5.34 Contour plot of the seasonal and depth variation in POC fluxes derived with the steady state model from 238U seawater concentration estimates obtained from equation 5.5………………………………………………………………………..................221 234  Figure 5.35 Average POC fluxes derived from 238  Th deficit assuming steady state from  U seawater concentration estimates obtained from equation 5.6 and average POC fluxes  measured by sediment traps…………………………………………………………..222 Figure 5.36 Two year time series records of POC fluxes measured with sediment traps and 234  corresponding fluxes derived from  Th deficits…………………………………223  Figure 5.37 (a) Ratios of organic carbon fluxes obtained with equation 5.12 and 5.5; (b) ratios of organic carbon fluxes obtained with equation 5.13 and 5.5……………228 Figure 5.38 Comparison of organic carbon fluxe time series obtained with sediment traps, equation 5.5 (blue dots) and equation 5.13 (black dots) from  238  U seawater concentration  estimates obtained from equation 5.6 for all cases……………………………………230 Figure 5.39 (a) POC fluxes estimated from equation 5.5 and  238  U seawater concentration  estimated from equation 5.6 versus POC fluxes from the sediment traps; (b) POC fluxes estimated from equation 5.13 and  238  U seawater concentration estimated from equation  5.6 versus POC fluxes from the sediment traps………………………………………231 Figure 6.1 Model (A) used to describe thorium cycling in seawater…………………237 Figure 6.2 Total  234  Th, particulate  234  Figure 6.3 Profiles of fine particulate  Th and  238  U profiles at OSP……………………250  234  Th obtained by small volume filtration on 25mm  TQ filtration, by LVPs on Supor filters, and large (>53 μm) particulate  234  Th collected  with LVPs on Nylon mesh…………………………………………………………….251 xxiii  Figure 6.4 Profiles of  234  Th in large particles and “extra-large” particles……………251  Figure 6.5 POC profiles and POC/234Th for the particles collected on 25mm TQ filters, the fine particles collected by LVP on Supor filters, and Nitex mesh………………...254 Figure 6.6 Redfield ratio derived from the 25mm TQ and LVP Supor data by assuming the POC/234Th are the same on those two different types of fine particles……………255 Figure 6.7 Profiles of dissolved Figure 6.8 The fraction of total Figure 6.9 Profiles of  230  Th,  230  Th on fine particles and large particles…257  230  Th associated with fine particles………………257  230  Th on fine, large and XL particles…………………………258  Figure 6.10 POC/230Th for particles collected by LVP on Supor filters and Nitex mesh.......................................................................... .............................258 Figure 6.11 P, Ca, and Al concentrations associated with fine, large and XL particles………………………………………………………………………................263 Figure 6.12 Th-234 fluxes based on 234Th:238U deficit………………………………….265 Figure 6.13 POC fluxes derived from the 234Th flux and POC/234Th ratio on the 25mm TQ filters.……………………………………………………………………….............268 Figure 6.14 POC fluxes derived from the POC/234Th ratio on all three samples………270 Figure 6.15 Biogenic Ca fluxes derived from the biogenic Ca/234Th ratio on LVP-Supor and mesh samples………………………………………………………………………271 Figure 6.16 Al fluxes derived from the Al/234Th ratio on LVP samples………………272 Figure 6.17 POC fluxes derived by  230  Th normalization on fine particles compared with  the POC fluxes estimated by other means……………………………………………278  xxiv  Figure 6.18 Biogenic Ca fluxes derived by derived by  Thxs normalization compared to those  234  Th:238U and sediment traps………………………………………………282  Figure 6.19 Al fluxes derived by 234  230  230  Thxs normalization compared to those derived by  Th:238 U……………………………………………………………………… …..283  Figure 6.20 Output from model A with K1 = 0.5/y and K-1 = 1.6/y; B1 = 3/y and B-1 = 150/y; and S = 150m/day………………………………………………………...286 Figure 6.21 Model B…………………………………………………………………..288 Figure 6.22 The optimal run of model B………………………………………………290 Figure 6.23 Model B for POC dynamics………………………………………………291 Figure 6.24 POC concentrations generated by Model B………………………………293 Figure 6.25 Model C used to describe the complex thorium cycling in seawater……294 Figure 6.26 Model C used to describe the complex POC dynamics in seawater………295 Figure 6.27 Th concentrations produced by model C…………………………………296 Figure 6.28 POC concentrations produced by model C………………………………297 Figure 6.29 POC fluxes calculated by [POC]*S from Model C………………………298 Figure 6.30 POC/230Th ratios on fine particles produced by model C compared to that measured on LVP-Supor samples………………………………………………………301 Figure 6.31 POC fluxes based on  230  Th normalization from model C output compared to  the POC fluxes results by other means…………………………………………………301 Figure 6.32 POC fluxes based on  230  Th normalization compared to the ‘real’ POC fluxes  calculated by [POC]*S from model output and the sediment trap data………………302  xxv  Acknowledgements I would like to sincerely thank my advisor, Dr. Roger Francois, for his support and guidance on my research. His contagious energy and passion on science have shown me how to be a great scientist. He always enlightens me with his fresh ideas when I got stuck with my research. I feel so lucky to have Roger as my supervisor not only because of his support on my research but also because of his encouragement and help on my life when I felt down. Special thanks are given to my committee members: Dr. Susan Allen, Dr. Lisa Miller, Dr. Kristin Orians and Dr. Evgeny Pakhomov for their valuable and constructive comments, suggestions and proofreading of my thesis. In particular, Susan’s help with the set-up of my 2D scavenging model and Lisa’s help with the beta counting of my  234  Th  samples are very crucial to my Ph.D. thesis. I also greatly appreciate the help from all the past and present lab members of Roger Francois’s research group in the department of earth and ocean sciences. I want to thank Maureen Soon, Samuel Jaccard, Kristina Brown, Bart De Baere, Drew Snauffer, Genna Patton and Aram Goodwin for their generous help during my study here. I would like to thank all my collaborators Joerg Lippold, Jeanne Gherardi and Sylvain Pichat for their priceless input to all our published and unpublished research fruits. I also would like to thank the Journal of Ocean Science and Nature Geosciences for giving me the permission to reproduce the figures in this thesis. Lastly, but most importantly, I would like to express my deep and sincere gratitude to all my family members for their support for the success of the Ph.D. research work. xxvi  Dedication  To my parents & my wife Yu Ling  xxvii  Chapter 1 Introduction  The ocean impacts global climate in several important ways: its biogeochemistry and overturning circulation strongly influence the atmospheric CO 2 level and greenhouse warming (Sigman et al., 2010), while its surface and overturning circulation contribute significantly to the redistribution of solar heat on the surface of the planet (Ganachaud and Wunsch, 2000). In this thesis, the importance of these fundamental processes for climate evolution is further explored by applying U-decay series isotope systematics, which provides some of the most important geochemical tools to estimate rates of processes occurring in the modern and past oceans (Cochran, 1992; Henderson and Anderson, 2003). This introduction describes the link between ocean processes and climate and some of U-series isotopes applications to study these processes, focusing on thorium (Th) and protactinium (Pa) isotopes.  1.1 Ocean’s role in climate change The ocean strongly affects the evolution of Earth’s climate on decadal to multi-millennial timescales. This stems from the fact that the ocean is a very large reservoir of heat 1  (Rahmstorf, 2002) and carbon (Sigman et al., 2010), capable of responding to forcing on these timescales. In particular, it is believed that the ocean plaid a key role in the climatic variability associated with the waxing and waning of the ice ages during the Quaternary (Rahmstorf, 2002; Sigman et al., 2010). While it is well established that the timing of Quaternary climate variability is controlled by periodic changes in the eccentricity of Earth’s orbit, the tilt of its axis of rotation, and the precession of the equinoxes, the amplitude of glacial-interglacial climate changes appears to be controlled to a large extent by physical, chemical and biological processes taking place in the ocean. Abrupt climate changes that punctuated the ice ages and deglaciations have also been attributed to the response of the ocean to addition of freshwater from collapsing ice sheets (Schmittner et al., 2002; Clark et al., 2002). Decadal climate oscillations, such as the El Nino Southern Oscillation and Pacific Decadal Oscillation, are also strongly influenced by ocean processes (McPhaden, 1999) and the ocean also plays a key role in mitigating global warming from anthropogenic CO2 emissions by absorbing a significant fraction of the emitted CO2 (Sabine and Tanhua, 2010) and by storing some of the heat induced by greenhouse warming (Meehl et al., 2011). Among the ocean processes that play a critical role in climate regulation, the Ocean’s meridional overturning circulation (MOC), also known as the thermohaline circulation, is particularly important for redistributing solar heat on the surface of the planet, and particularly, in delivering heat from low to high latitudes in the North Atlantic (Rahmstorf, 2002). This heat transport has a profound impact on the regional climate of land masses in the northern hemisphere and the response of the MOC to global warming is one of the 2  main uncertainties in our prediction of the future evolution of climate. Heat and moisture transport by the MOC is also believed to be fundamentally important in understanding the waxing and waning of the northern continental ice sheets during the Quaternary and the recurrence of abrupt, millennial-scale climatic events in the past. Moreover, changes in the overturning circulation are also believed to have significant impacts on the concentration of CO2 in the atmosphere because it determines the rate at which deep waters with high metabolic CO2 content return to the surface.  Carbon uptake by the ocean occurs through a range of mechanisms that are often referred to as 'carbon pumps'. We distinguish the “Solubility Pump”, which is controlled by temperature, the “Biological Pump”, which is controlled by the balance between the export flux of organic carbon from surface to deep water and the return to the surface of CO2 produced by the decomposition of organic matter in the deep sea by the meridional overturning circulation, and the “Carbonate Pump”, controlled by the formation and dissolution of calcium carbonate, which dictates the alkalinity of seawater. Reference is also sometime made to the “Continental Shelf Pump” to emphasize the particular role of highly productive shelf waters and marginal seas (Lee et al., 2011). Because the dissolved inorganic carbon content of the ocean is sixty times larger than the CO 2 content of the atmosphere, ocean circulation, biological productivity, and the chemical and physical properties of seawater largely dictate the level of atmospheric CO 2 and its variability through time on time scales greater than the mixing time of water in the ocean (~ 1000 years). Likewise, increasing sea surface temperature and pH as a result of anthropogenic CO2 emissions will likely produce changes in marine ecosystems that could significantly 3  affect the ocean’s biological, carbonate and solubility pumps in a way which would hamper CO2 uptake by the ocean and exacerbate the climatic impact of future emissions.  While it is clear that oceanic processes largely control the level of atmospheric CO 2 on a range of timescales, the details of the physical, biological and chemical mechanisms involved and their complex interactions are still poorly understood. Such an understanding is, however, crucial to make more reliable predictions regarding the evolution of climate in the coming centuries and guide society in taking appropriate actions. Two distinct aspects of this overarching question are addressed in this thesis. In Chapter 2, 3 and 4, I contribute to the development of a new tool to constrain past changes in the meridional overturning circulation on glacial-interglacial timescales, while in chapter 5 and 6, I further develop an approach to better assess the export of marine particles from surface to deep water. To tackle these questions, I make use of isotopes from the uranium decay series, which provide a unique tool box to quantify the rates of various oceanic processes, and in particular the rate of the overturning circulation and the sinking flux of particles, as described in the following sections of this chapter.  1.2 U-series isotopes Uranium has two long-lived isotopes initiating a decay chain (Fig. 1.1): 4.47 109 y) and  235  238  U (half-life =  U (half-life = 7.04 108 y). They were produced billions of years ago in  supernovae by nucleosynthesis, but because of their very long half-lives, they still persist to this day. When left undisturbed, all the daughters in the series (Fig. 1.1) have the same 4  decay rate (or the same “activity”). The decay series is then said to be in “secular equilibrium”. In this situation, each radioactive isotope of the series decays at a rate which is equal to the rate at which it is produced (which is the rate at which its parent is decaying). Ultimately, the decay rates of all the daughters are dictated by the decay rate of the initial parent (238U or 235U), which decreases very slowly with time.  Figure 1.1: Uranium decay series 5  The usefulness of these isotopes in environmental science stems from the fact that processes on Earth’s surface often separate daughters from parents, thereby perturbing secular equilibrium (Bourdon et al., 2003). This arises because the successive decay series daughters consist of different elements with very different chemical properties (Fig. 1.1). For instance, while uranium and radium are relatively soluble in water, thorium, protactinium, polonium and lead have very low solubility and are readily removed from aqueous solutions by adsorption on surfaces of particles or sediments. On the other hand, radon is a gas which diffuses into the atmosphere from the rocks or water where it is produced from the decay of radium. When an environmental process (e.g., particle scavenging, air-sea exchange, etc.) separates a daughter from its parent, the activity of the daughter isotope (i.e., the rate at which it decays) decreases and becomes lower than the activity of its parent isotope. Assuming steady-state, we can write a simple mass balance equation stating that the rate at which the daughter is produced (i.e., the rate at which the parent decays or the activity of the parent; parent Nparent) must be equal to the rate at which the daughter decays (i.e., the activity of the daughter; daughter Ndaughter) plus the rate at which the daughter is removed by the environmental process (R daughter). Thus: Rdaughter (atoms.min-1) = parent Nparent - daughter Ndaughter where  is the decay constant, N is the number of atoms and R is the rate of removal or addition of the daughter isotope by the environmental process of interest  6  Measuring the difference in activity between parents and daughters in environmental samples thus provides a means of quantifying the rate of removal (or addition) of the daughter from (to) the sample. In turn, determining the rate of removal of the daughter by a given process under different environmental conditions provides quantitative information on the relative intensity of this process under these environmental conditions and adds insight into the mechanisms involved.  1.3 Uranium budget in the ocean Uranium is supplied to the ocean by rivers at an estimated rate of 1.1*1010g/year (±35%) (Cochran, 1992; Palmer and Edmond, 1993). Additional sources from wind-blown dust and ground water discharge are less important and difficult to estimate (Dunk et al., 2002; Henderson and Anderson, 2003). In most of the ocean where oxidizing conditions prevail, uranium is soluble by forming soluble uranyl-carbonate species in sea water. Uranium is removed from the ocean mainly by burial in anoxic or suboxic sediments after reduction to its insoluble tetravalent state in pore waters, or by the alteration of ocean basalts.  Estimates of the major terms of the U budget are given in Table 1.1 (Dunk et al., 2002). Within relatively large uncertainties (e.g., the total removal flux is only known within a factor of 2, Henderson and Anderson, 2003), the oceanic U budget appears to be balanced, and the oceanic residence time of U has been estimated at 320-560 kyr (Dunk et al., 2002). With such a long residence time, U is a conservative element with a nearly constant 7  concentration in seawater (3.3 ppb). Estimating the U sedimentary sink during glacial periods suggests that seawater U concentration did not change significantly during the 100 kyr glacial-interglacial cycle (Rosenthal et al., 1995).  Table 1.1: Known inputs and outputs of U to the ocean. Some other fluxes such as groundwater input and input or removal in estuaries are poorly known and are not included in this table (Henderson and Anderson, 2003).  8  1.4 Thorium and protactinium in the ocean  Thorium is found at the +4 oxidation state and forms a highly insoluble neutral hydroxide (Th(OH)40) in seawater (Langmuir and Herman, 1980). As a result, it has a strong tendency to adhere to particle surfaces (Cochran, 1992). Thorium isotopes produced by uranium decay are thus removed rapidly from the water column by particle scavenging. Among the Th isotopes commonly used in oceanographic studies,  232  Th is supplied from  continental sources by rivers and aeolian dust and initiates its own decay series, while 230  Th and 234Th are added to seawater by decay of U isotopes (234U and 238U, respectively).  Since uranium is conservative, these thorium isotopes are thus produced uniformly over the entire water column, which is a key advantage for using these isotopes to quantify particle flux and ocean circulation. Because of its long half-life (75.69 ±0.23 ka; Cheng et al., 2000),  230  Th is mostly removed from seawater by scavenging. Its high affinity for  particle surfaces results in short residence times in seawater, ranging from a few months in surface waters to a few decades in deep waters (Table 1.2). In contrast, removal of the short lived  234  Th (half-life: 24.1 days) from seawater is mostly by radioactive decay over  most of the water column, but in surface waters, both scavenging and decay become important. Thorium-234 is thus in secular equilibrium with its parent isotope  234  U over  most of the water column, except at the surface where rapid scavenging results in a residence time similar to  234  Th’s half-life. Thorium-234 is best suited for studying faster  processes occurring in the upper ocean (e.g., export flux; Buesseler et al., 2006 and references therein) and surface sediments (bioturbation; Aller and DeMaster, 1984), while the longer half-life of 230Th provides constraints on the rate of slower processes occurring 9  deeper in the water column (e.g., scavenging, Bacon and Anderson, 1982; circulation, Marchal et al, 2000) and in sediments (e.g., sedimentation; Bacon, 1984). Table 1.2: Summary of major U, Th and Pa isotope data in world’s ocean. Nuclide  Concentration in the ocean  Residence time  Half life  (g/g)  (dpm/T)  (years)  (years)  238  3.3*10-9  2.5*103  4*105  4.46*109  Burial & Basalt  235  2.4*10-11  3.5*103  4*105  2.34*107  Burial & Basalt  234  2*10-13  2.9*103  1.9*105  2.45*105  Burial, Basalt, Decay  234  (1.9-4.8)*10-20  (~1 – 2.5)*103  ~0.04 – 0.1  6.6*10-2  Decay & Scavenging  230  (0.1-3.3)*10-17  (0.5-15)*10-1  0.1 - 55  7.57*104  Scavenging  231  (0.1-5.7)*10-18  (0.1-6)*10-1  1 - 240  3.25*104  Scavenging  U U U Th Th Pa  Mode of removal  Protactinium-231 is also produced uniformly in seawater by the decay of 235U, and it has a relatively long half-life (32.71 ± 0.11 ka; Robert et al., 1969). As for removed from the water column by scavenging, but 230  230  Th, it is mostly  231  Pa is not as particle reactive as  Th. As with Th4+, Pa forms insoluble Pa(V) hydroxide complexes in aqueous condition.  Its generally lower particle reactivity may stem from the strong complexing tendency of fluoride ions for Pa(V), which produces a range of negatively charged ions (Pal’shin et al., 1970). When the surface available for adsorption is silica, however, this difference disappears. In the ocean, silica is an important biomineral produced mostly by diatoms in the more productive regions of the ocean, such as upwelling regions and the southern Ocean. Although this observation has been confirmed in field (Walter et al., 1997) and 10  laboratory (Gueguen and Guo, 2002) studies, its chemical underpinning is not yet fully understood, but may be due to colloidal conditions at the silica-water interface, which reduces the hydrolysis of the Pa(V) (Roberts, 2008). Nonetheless, outside the oceanic regions dominated by diatoms (mainly the southern ocean and the North Pacific), the contrast between the solubility of Pa and Th forms the basis for several oceanographic and paleoceanographic applications. Below, I review briefly some of the main applications of Th and Pa isotopes in marine studies.  1.4.1  Sediment (231Paxs/230Thxs) to monitor past changes in the meridional overturning circulation of the ocean.  Uranium-234 decay produces 2.67 × 10-2 dpm/m3/y of  230  Th while  235  U decay produces  2.46 × 10-3 dpm/m3/y of 231Pa resulting in a constant production rate ratio of (231Pa/230Th) = 0.092 (dpm/dpm). However, the  231  Pa/230Th ratio measured in marine sediments mostly  deviates from this ratio. This is because the two isotopes have different residence times in the water column (Table 1.2). Protactinium-231 is usually less particle-reactive, and because of the resulting longer residence time in the water column, it is more effectively transported laterally by advection or diffusion than 230Th. As a result, sediment 231Pa/230Th is higher than the production rate ratio in regions that receive laterally transported while sediments underlying regions exporting  231  231  Pa,  Pa have ratios < 0.092 (e.g., Francois,  2007). 11  The lateral transport of  231  Pa can be driven by turbulent mixing between areas of  contrasting scavenging intensity, the latter depending on particle flux and composition. Scavenging intensity increases with particle flux and the concentration of opal in the settling particles. The latter reflects the greater affinity of  231  substrates (Chase et al., 2002). As a result, there is a net  Pa for silica than for other  231  Pa transport from the low  productivity central gyre regions of the ocean towards the margins, a process called “boundary scavenging” (Bacon, 1988), which has been used to explain the sediment 231  Pa/230Th distribution pattern in the Pacific (Yang et al., 1986; Anderson et al., 1990) and  to estimate variations in opal flux in the past (Anderson et al., 2009).  Alternatively, lateral transport of  231  Pa can be driven by deep water circulation and  particularly the meridional overturning circulation of the ocean. This type of lateral transport forms the basis for using sediment  231  Pa/230Th to reconstruct past changes in the  Atlantic Meridional Overturning Circulation [AMOC] (Yu et al., 1996; Marchal et al., 2000; McManus et al., 2004; Gherardi et al., 2005, 2009; Hall et al., 2006; Negre et al., 2010) and this is the application which is further developed in this thesis.  1.4.2  Particle flux from the euphotic zone – (234Th:238U)  Particle export from the euphotic zone is a key aspect of the biological pump which sequesters carbon to the deep sea, and it has become a focus in the research of many oceanographers interested in the fate of anthropogenic CO2 added to the atmosphere. 12  Measuring the disequilibrium between the  234  Th and  238  U in surface waters has become a  tool of choice to estimate this process.  Depth  Activity  Figure 1.2: The principle of 234Th:238U method (Weinstein et al., 2005). M is the element of interest in the particles that scavenged the Th in the water column. C M is the concentration of this element in the particles and ApTh is the activity of  234  Th scavenged by the same particles. P Th is the flux of  234  Th  derived from the 234Th:238U deficit in upper water column. P M is the flux of M.  The  234  Th activity in seawater results from the balance between production from  238  U  decay, removal by radioactive decay and scavenging on sinking particles, and transport by advection and diffusion. The temporal change in total  234  Th activity in the mixed layer is  expressed by: әTh/әt = λAU- λATh-P+V (1.1) 13  where AU and ATh are dissolved 238U and total 234Th activities integrated over the depth of the mixed layer (dpm/m2), respectively, λ is the decay constant of 234Th (=0.02876/day), P is the net removal flux of  234  Th by scavenging (dpm/m2/d), and V is the sum of the  advective and diffusive fluxes. Advection and diffusion can often be neglected, particularly when the water column is stratified (Savoye et al, 2006). Therefore, under these conditions, by assuming steady state conditions, equation 1.1 can be simplified as: P =λAU- λATh (1.2)  Since  238  U is conservative in the ocean, its activity in the water column can be derived  from salinity. The net  234  Th flux at the base of a layer (P) equals the  integrated over that layer multiplied by the decay constant of total  234  Th deficit  234  Th. Measuring profiles of  234  Th activity in the upper water column and the ratio of organic carbon  concentration to  234  Th activity in sinking particles provides an estimate of the export flux  of organic carbon (Fig. 1.2). The application of  234  Th:238U method expanded during the  JGOFS program (Buesseler et al., 1992) and has been increasingly used since (Waples et al., 2006).  1.4.3  230  Th normalization to reconstruct particle flux to the seafloor  Th-230 is produced in the water column by decay of  234  U. Its removal by scavenging is  best described by reversible adsorption on the surface of settling particles (Bacon and 14  Anderson, 1982; Nozaki et al., 1987), which predicts a gradual increase in dissolved and particulate  230  Th concentration or activity with depth. This prediction has now been  confirmed by multiple measurements in the world ocean (e.g., Francois, 2007).  Because of its very high particle reactivity, the residence time of  230  Th in the water  column is short (at most 40 years in deep waters), and considering its very long half-life, almost all of the removal is due to scavenging. As a result of its short residence time, 230  Th is effectively removed by scavenging into the underlying sediment before it can be  extensively transported laterally by advection or diffusion. This means that the scavenged flux of  230  Th to the seafloor is always nearly equal to its production rate in the overlying  water, which is a simple linear function of water depth (Bacon, 1984): PTh =Z (234U) λ230 (1.3) where PTh is the production rate of 230Th in the oceanic water column (dpm/m2/y), Z is the water depth (m), (234U) is the activity of 234U in seawater (= 2750 dpm/m3) and λ230 is the decay constant of 230Th (y-1).  If removal by scavenging were instantaneous, the vertical flux of  230  Th would be exactly  equal to the production rate: FTh=PTh. Of course, this is only an approximation, since the removal is not instantaneous. Nonetheless, it is fast enough that F Th rarely deviates from PTh by more than 30% (Henderson et al., 1999; Yu et al., 2001). This forms the basis for an important application in paleoceanography in which  230  Th is used as a constant flux  15  tracer to normalize the flux of other sedimentary constituents accumulating on the seafloor (Bacon, 1984; Francois et al., 2004).  Accurately estimating past variations in the flux of biogenic and lithogenic particles reaching the seafloor is central to better understanding the response of the ocean to past environmental conditions and assess its role in the evolution of these environmental conditions. Sedimentation rate on the seafloor is however controlled not only by the vertical rain rate of particles from the overlying surface, which is generally the quantity of interest when assessing past environmental conditions, but also on complex interactions between bottom topography and ocean currents that redistribute sediments on the seafloor after their initial deposition. Because post- or syn-depositional transport by bottom currents redistribute equally  230  Th and all other sediment constituents (except in cases of  significant size fractionation), normalizing the sediment constituents to that  230  230  Th by assuming  Th flux to the seafloor is equal to its rate of production in the overlying water can  therefore correct for lateral redistribution by bottom currents. This method was first proposed by Bacon (1984) and further developed in the following two decades (Francois et al., 2004). It has now become a standard tool in paleoceanography to assess past changes in the flux of particles reaching the seafloor and preserved in sediments.  16  1.5 Thesis objectives  Applications of U-decay series isotopes in oceanographic and paleoceanographic studies have multiplied in recent decades, but key questions remain unresolved. For instance, the relative importance of particle composition and ocean circulation in controlling sediment 231  Pa/230Th is still being debated (e.g., Keigwin and Boyle, 2008; Gherardi et al., 2009;  Peacock et al., 2009), as well as the validity of  230  Th-normalization to reconstruct particle  flux from the sedimentary record (Francois et al., 2008; Lyle et al., 2008). The exact interpretation of sediment  231  Pa/230Th in terms of changes in deep water circulation has  also been questioned (Thomas et al., 2006; Negre et al., 2010). Likewise, while calculating the  234  Th export flux from the  234  Th/238U disequilibrium in surface waters is  straightforward, estimating the associated flux of carbon is much more problematic (Rutgers van der Leoff et al., 2006; Buesseler et al., 2006), stemming from uncertainties in the dynamics and composition of the particles removing Although  234  Th from surface waters.  234  Th provides a useful tool to estimate export flux from surface water, it does  not provide information on the fluxes of particles in deeper water, where most of the carbon remineralization occurs. Yet, the depth of organic matter remineralization is the key to establishing the efficiency of export production in sequestering carbon to the deep sea. These are the questions and problems that this thesis attempts to address. It is subdivided into 5 chapters following this introduction and concludes with a brief “Conclusions and Perspectives” section.  The five core chapters are: 17  Chapter 2: Sediment  231  Pa/230Th as a recorder of the rate of the Atlantic Meridional  Overturning Circulation: insights from a 2-D model  In this chapter, I develop a 2D scavenging model to establish the meridional and vertical distribution patterns of sediment  231  Pa/230Th generated by simple Atlantic meridional  overturning cells.  Chapter 3: Strength and geometry of the Glacial Atlantic Meridional Overturning Circulation  In this Chapter, I report results from a collaboration with colleagues who measured 231  Pa/230Th in sediment samples from the Holocene and last glacial sections of multiple  cores from the Atlantic to constrain the rate and geometry of the Atlantic Meridional Overturning Circulation during the last ice age. My contribution was in using my model to better interpret the data generated by my co-authors. Chapter 4: The influence of deep water circulation on the distribution of 231Pa and 230Th in the water column and sediments of the Pacific Ocean  In this chapter, I expand the 2D scavenging model to include a representation of the Pacific Ocean to explore the potential of sediment 231Pa/230Th in constraining past changes in the overturning circulation in this ocean.  Chapter 5: Comparison of POC fluxes measured with sediment trap and  234  Th:238U  disequilibrium in a coastal setting (Saanich Inlet, British Columbia) 18  In this chapter, I report a time-series study (Feb/2009 to Feb/2011) comparing fluxes measured by 234Th scavenging and sediment traps in Saanich Inlet, British Columbia.  Chapter 6: Particle fluxes and dynamics in the northeast Pacific Ocean studied by paired measurements of thorium isotope activities  In this chapter, I explore the possibility of combining measurements of  234  Th and  230  Th  dissolved in seawater and adsorbed on three different size classes of particle to estimate particle flux in the mesopelagic zone of the ocean to assess the efficacy of the biological pump.  19  Chapter 2 Sediment 231Pa/230Th as a recorder of the rate of the Atlantic Meridional Overturning Circulation: insights from a 2-D model  2.1 Introduction  Ocean circulation plays an important role in climate control by transferring solar heat from low to high latitudes (Ganachaud and Wunsch, 2000). In particular, rapid changes in the strength and geometry of the Atlantic Meridional Overturning Circulation (AMOC) have been invoked to explain the abrupt variations in climate that have punctuated the last ice age and deglaciation (Schmittner et al., 2002; Clark et al., 2002). However, documenting the link between changes in climate and ocean circulation still remains a major challenge in paleoclimatology (Lynch-Stieglitz et al., 2007). Past changes in circulation were first inferred from the sedimentary records of nutrient proxies (Boyle and Keigwin, 1987). While these tracers provide important information on changes in the geometry of the overturning circulation, they do not constrain changes in the rate of overturning (Legrand and Wunsch, 1995). To address this problem, several kinematic tracers of ocean circulation 20  231  Pa/230Th ratio of Atlantic  are being investigated (Lynch-Stieglitz et al., 2007). The  sediments is one of these tracers. This proxy has recently been used to investigate past changes in the rate of the AMOC from the last glacial maximum to present (McManus et al., 2004; Hall et al., 2006; Gherardi et al., 2005; 2009). Because both  231  Pa and  have uniform production rates (from the decay of dissolved uranium) and longer residence time than  230  Th in the water column, the AMOC exports  230  Th  231  Pa has a  231  Pa more  effectively from the Atlantic into the Southern Ocean (Yu et al., 1996; Francois, 2007). The modern rate of overturning results in the mean residence time of deep water in the 231  Atlantic roughly equivalent to the mean residence time of 200 years), so that nearly half of the  231  Pa in the water column (~  Pa produced in Atlantic water is exported to the  southern ocean with the water in which it formed. On the other hand, with its much shorter residence time (~ 30 years), nearly all of the  230  Th produced in this water is  removed into the sediments of the Atlantic and little is exported to the Southern Ocean. As a result, the  231  Pa/230Th ratio in Atlantic sediments is, on average, about half the  production rate ratio (0.092 dpm/dpm) of these two isotopes in the water column. Faster rates of overturning export a larger fraction of  231  Pa and further decrease  231  Pa/230Th,  while slower rates of overturning increase this ratio (Marchal et al., 2000; Siddal et al., 2007).  Application of this simple principle is, however, complicated by two factors. First, sedimentary  231  Pa/230Th is not only controlled by the rate of overturning, but also by the  removal rate of the two isotopes from the water column by particle scavenging. Scavenging rates, which are controlled by the flux and composition of settling particles 21  (Bacon, 1988; Walter et al., 1997; Chase et al., 2002; 2003), dictate the residence time of 231  Pa in seawater and the extent to which it can be exported from the Atlantic by the  AMOC (Yu et al., 1996). On the other hand,  230  Th has a residence time sufficiently short  to severely limit its redistribution by circulation and mixing, even when the rate of overturning is fast (Francois et al., 2004). It is possible to assess the impact of changes in particle scavenging by analyzing the composition of the sediment, which informs us on changes in particle flux and composition at the site of study and their possible overprint on sediment  231  Pa/230Th at this location (Gherardi et al., 2009), at least to the extent that  we can take into account the effect of diagenesis. However, the extent to which scavenging can also affect sediment  231  Pa/230Th further “downstream” in the overturning  circulation cell still needs to be investigated. The second point of contention is the extent to which sediment 231Pa/230Th integrates circulation rates over the overlying water column. In a recent study using a 1-D scavenging model, Thomas et al. (2006) have argued that sedimentary  231  Pa/230Th may only record overturning occurring in about 1000m of water  overlying the analyzed sediment and shallower overturning cannot be recorded in deep sediments.  Several studies have investigated the distribution of  231  Pa and  230  Th in the ocean using  three dimensional circulation models based on simplified dynamics (Henderson et al., 1999; Siddall et al., 2005; 2007) or the primitive equations (Dutay et al., 2009). In this study, we take a very different approach and develop a simple 2-D scavenging model to establish the patterns of  231  Pa/230Th distribution that can be generated by an ascribed  overturning circulation. The results provide possible explanations for some of the existing 22  field observations in the water column and sediments and a baseline for further evaluating the influence of the other factors that affect the distribution of 231Pa/230Th in the real ocean. They also suggest sampling strategies to maximize the information on paleocirculation that could be obtained from a very limited sediment database.  2.2 Model descriptions  Water column profiles of dissolved and particulate  230  Th and  231  Pa concentration indicate  that these two isotopes are removed from seawater by reversible scavenging (Bacon and Anderson, 1982; Nozaki et al., 1987). We use the same formalism to describe scavenging imbedded in a 2-D circulation scheme to investigate how the concentration of 230  231  Pa and  Th in the water column and sediments can potentially be affected by changes in  circulation and scavenging rate.  2.2.1 Formulation  The scavenging model used for both 230  231  Pa and  230  Th is shown in figure 2.1. Dissolved  Th and 231Pa concentrations ([X]d; where X represents 230Th or 231Pa) are controlled by  the production rates of the respective nuclides (P X; dpm.m-3.y-1), their adsorption (K1X) and desorption (K-1X) rate constants (y-1), and the transport rates imposed by the circulation scheme (V; m.y-1), while  230  Th and  231  Pa particulate concentrations ([X]p) are  controlled by the adsorption/desorption rate constants, transport rates and the sinking rates 23  (S; m.y-1) of the particles that scavenge the two nuclides from the water column. At steady-state, we can write: Px - K1x[X]d + K-1x[X]p + VΔ[X]d = 0  (2.1)  K1x[X]d - K-1x[X]p + VΔ[X]p + dFlux/dZ = 0  (2.2)  dFlux/dZ = S ([X]p(i+1) - [X]p(i))  (2.3)  Where X represents 230Th or 231Pa, Z is water depth (m), i is the vertical index, and Δ is an “upwind” difference divided by the grid spacing (Press et al., 1992). The model uses a uniform grid with a horizontal grid spacing of 2.5 degrees latitude and a vertical grid spacing of 250 m.  24  Figure 2.1: The scavenging model consists of a meridional section (from 70°N to 70°S) evenly distributed into 56*20 grids (20 layers evenly distributed over 5000m depth and 56 columns evenly distributed over the Meridional section; 2.5 latitude per column). In each box, X represents  230  Th or  231  Pa. Xd’s = dissolved concentration (dpm/m3). Xp’s = particulate concentration (dpm/m3). Px =  production rate from U decay (dpm/m3/y). S = sinking rates of particles (m/y). V = transport rates (m/y). The MATLAB code of the model is reported in Appendix A.  These equations are used to calculate the concentration of dissolved and particulate and  231  230  Th  Pa as a function of depth and latitude. Using the upwind scheme with a horizontal  velocity u = 5.3 10-3 m/s and a horizontal grid spacing x = 278 103 m, the inherent 25  mixing in our model (Kdiff) is ~ 800 m2/s (Kdiff = u x/2; based on equivalence of the upwind  scheme  applied  to  an  advective-reactive  equation  and  an  analytic  diffusive-advective-reactive equation; e.g. Press et al., 1992). This is in the upper range of the along-isopycnal tracer diffusivities reported for the southern ocean (100-800 m2/s; Zika et al., 2009). Initial tests indicate that using a smaller grid size to decrease the model’s diffusivity does not result in significant differences in the model results.  2.2.2 Overturning Circulation  The 2D meridional overturning circulation scheme (control run) used in this study is ascribed within a meridional section in the Atlantic Ocean (constant depth of 5000m from 70°N to 70°S) and based on the meridional overturning transports for the North Atlantic reported by Talley (2003). It consists of two meridional overturning cells flowing in opposite directions (Fig. 2.2). The Atlantic Meridional Overturning Circulation (AMOC) is initiated by the formation of 20.5 Sv of North Atlantic Deep Water (NADW) (Friedrichs and Hall, 1993; Macdonald, 1998; Talley et al., 2003) resulting from water flowing north in the upper 1500m of the water column and sinking between 60°N and 70°N. This latitudinal range coincides roughly to the latitudes where deep water forms in the Labrador and Nordic seas. The site of deep water formation (60°N ~ 70°N) is represented by one homogenized region between 250 to 4250 m depth to represent rapid deep water convection.  26  Flux at Equator  Atlantic Flux and Velocity Vector  10  20  1500  20  10 5  -5  15  15  -5  10  5  3750 0  -5  5  2000 2500 3000  5  10  3250  4250  15  10  2250  Depth(m)  10 0  Depth  1750  2750  1000  15  20  5  -5  500  15  20  20  20  5  10  15  20 15  1250  5  10  15  10 5-5  750  5  0  250  0  5  0  3500 4000  0  -5  4500  4750 5000 -4  60S 50S 40S 30S 20S 10S 0 10N 20N 30N 40N 50N 60N Latitude  -2  0 2 Flux(Sv)  4  6  Figure 2.2: (a) Velocity vector plot. Size of the arrows is proportional to the transport rates used in the model. (b) Overturning fluxes in the model at the equator.  Water from this homogeneous region is then transported horizontally to the south at different rates (Fig. 2.2b). The depth distribution of lateral transport was chosen so that the model generates dissolved  230  Th and  231  Pa profiles consistent with observations (see  below). At 10°N, the NADW flow increases to 22.5 Sv with the addition of 2 Sv from the Antarctic Bottom Water (AABW) between 10°N and 35°N. Two Sverdrups of AABW are added further south, resulting in a total flow of 24.5 Sv of NADW, which is close to the NADW strength (23±3 Sv) estimated from the World Ocean Circulation Experiment (WOCE) data (Ganachaud and Wunsch, 2000). NADW starts to gradually upwell at 42.5°S towards a mixing zone (i.e. one homogeneous region) located above 1000m between 67.5°S and 57.5°S. Water from this mixing zone feeds surface and intermediate water forming the shallow return limb of the AMOC. 27  The second overturning cell is initiated by 8 Sv of AABW, originating from the same mixing cell, flowing into the southernmost region (67.5°- 70°S) and sinking directly to 3500 m (Fig. 2.2). Four Sverdrups are transported northward between 3500m and 4500m depth and entrained in the upwelling NADW south of 40°S. The remaining 4 Sv are transported northward below 4500m. This northward flow is gradually attenuated by entrainment to the NADW and disappears at 37.5°N, which is roughly consistent with hydrographic observations (Sloyan and Rintoul, 2001). In this study, we do not specifically represent the Antarctic Intermediate Water (AAIW). Although the rate of AAIW formation may affect the  231  Pa/230Th of sediment deposited at intermediate depths  in the South Atlantic, preliminary model runs indicate that this water mass has little or no effect on the 231Pa/230Th of deep sea sediments.  2.2.3 Parameterization  Among all the parameters needed to constrain the model shown in Figure 2.1, the production rates for 230Th and 231Pa are best known since they are essentially constant and only depend on the well-established concentration of 234U and 235U in seawater (Delanghe et al., 2002; Robinson et al., 2004). The other parameters (Table 2.1) are associated with greater variability and uncertainties.  Sinking rate (S): Most estimates of the average sinking rate of fine particles (S) obtained from water column profiles of particulate  230  Th (e.g. Krishnaswami et al., 1981; Rutgers  van der Loeff and Berger, 1993; Scholten et al., 1995; Moran et al., 2001) range between 28  400-800 m.y-1. Since there are no clear indications of systematic variability in this parameter, we chose a uniform and intermediate value of 500 m.y-1 (Table 2.1) K1Th and K-1Th: The adsorption (K1) and desorption (K-1) rate constants for  230  Th have  been estimated using a reversible scavenging model (Bacon and Anderson, 1982; Nozaki et al., 1987; Clegg and Whitfield, 1991; Clegg et al., 1991) and mostly range from 0.2 to 0.8 y-1 for K1Th and 1 to 3 y-1 for K-1Th. We chose values within this range (Table 2.1) which generate dissolved  230  Th profiles broadly consistent with water column profiles  measured at several locations in the Atlantic and in the Southern Ocean (Fig. 2.3; Table 2.2). K1Th is lower in the Southern Ocean than in the Atlantic, consistent with the data of Chase et al. (2002). We also used higher K1Th in the upper 500m to reflect the increase in K1Th with particle concentrations (Bacon and Anderson, 1982).  Table 2.1: List of abbreviations and values for the model parameters.  Variables  Symbol  Control run  Units  231  PPa  0.00246  dpm/(m3*yr)  230  PTh  0.0267  dpm/(m3*yr)  Particle sinking rate  S  500  m/yr  Pa production rate Th production rate  230  Th adsorption rate (70°N-50°S) 0-250m 250-500m > 500m  K1Th  1.0  1/yr  K1  Th  0.75  1/yr  K1  Th  0.5  1/yr  K1Th  230  Th adsorption rate (50°S-70°S) 0-250m 250-500m > 500m  0.6  1/yr  K1  Th  0.45  1/yr  K1  Th  0.3  1/yr  230  Th desorption rate (70°N-70°S) 29  Variables  Symbol All depths  Control run  Units  1.6  1/yr  K1Pa  0.08  1/yr  K1  Pa  0.06  1/yr  K1  Pa  0.04  1/yr  K-1  Th  231  Pa adsorption rate (70°N-42.5°S) 0-250m 250-500m > 500m  231  Pa adsorption rate (42.5°-45°S) 0-250m 250-500m > 500m  K1Pa  0.2  1/yr  K1  Pa  0.15  1/yr  K1  Pa  0.1  1/yr  K1Pa  231  Pa adsorption rate (45°-47.5°S) 0-250m 250-500m > 500m  0.3  1/yr  K1  Pa  0.225  1/yr  K1  Pa  0.15  1/yr  K1Pa  0.44  1/yr  K1  Pa  0.33  1/yr  K1  Pa  0.22  1/yr  1  1/yr  231  Pa adsorption rate (47.5°S-70°S) 0-250m 250-500m > 500m  231  Pa desorption rate (70°N-70°S) All depths  K-1Pa  K1Pa and K-1Pa: The adsorption and desorption rate constants for  231  constrained and we selected their values so as to obtain dissolved  231  Pa are even less Pa concentration  profiles (Table 2.2) and fractionation factors (Table 2.3) that are also broadly consistent with observations in the field.  30  Figure 2.3: Station locations for the water column profiles used to constrain the parameters in the model. Nordic Seas: N-A, N-B, N-C (Moran et al., 1995; 1997; 2002); Western Atlantic: W-A to W-F  31  (Table 2.2); Eastern Atlantic: E-A to E-F (Table 2.2); Southern Ocean: S (Rutgers van der Loeff and Berger, 1993).  The fractionation factor is defined as (Anderson et al., 1983): F = ([231Pa]d/[230Th]d)/([231Pa]p/[230Th]p)  (2.4)  F has been directly measured in the Atlantic and southern ocean (Walter et al., 1997; Moran et al., 2001; Chase et al., 2002). Particle composition affects the fractionation factor (F) due to the stronger affinity of opal for  231  Pa. In carbonate dominated regions, F  is much higher than in opal dominated regions, where F is close to 1. We have adjusted the adsorption and desorption rate constants with latitude (Table 2.1) to produce systematic variations in the “equilibrium” fractionation factor which broadly reflect the field observations (Moran et al., 2002; Walter et al., 1997; Table 2.3). Table 2.2: 230Th and 231Pa activities in sea water (dpm/1000kg).  W-A: Station EN407-3 (39°28’N; 68°22’W) Diss. 230Th  depth m  Diss. 231Pa  dpm/1000kg (±95% CI)  250  -  500  0.140  751  -  0.062  ± 0.004  ± 0.005  0.110  ± 0.005  0.187  ± 0.007  0.147  ± 0.006  1001  0.204  ± 0.008  0.134  ± 0.006  1250  0.161  ± 0.006  0.092  ± 0.005  1501  0.265  ± 0.009  0.148  ± 0.009  1800  0.299  ± 0.008  0.166  ± 0.007  2200  0.338  ± 0.011  0.176  ± 0.007 32  depth  Diss. 230Th  m 2500  Diss. 231Pa  dpm/1000kg (±95% CI) 0.353  2750  ± 0.011 depth 0.329 ± 0.019  2980  0.295  0.181 m  ± 0.009  ± 0.007 depth 0.165 ± 0.011 0.140  ± 0.005  dpm/1000kg (±95% CI) W-B: Station EN407-4 (38°36’N; 68°53’W) m depth m  Diss.  230  Th  Diss. 231Pa  50  dpm/1000kg (±95% CI) dpm/1000kg (±95% CI) 0.044 ± 0.002 0.030 ± 0.004  200  0.090  ± 0.004  0.064  ± 0.005  400  0.116  ± 0.005 m 0.113  ± 0.007  600  0.238  ± 0.008  0.168  ± 0.009  800  0.235  ± 0.009  0.177  ± 0.006  1000  0.242  ± 0.009  0.152  ± 0.008  1200  0.206  ± 0.008  0.140  ± 0.006  1400  0.247  ± 0.008  0.147  ± 0.006  1600  0.280  ± 0.012  0.170  ± 0.005  1800  0.299  ± 0.010  0.155  ± 0.006  2000  0.303  ± 0.011  0.156  ± 0.007  2200  0.340  ± 0.011  0.193  ± 0.008  2400  0.335  ± 0.014  0.187  ± 0.009  2600  0.337  ± 0.014  0.182  ± 0.007  2800  0.321  ± 0.016  0.165  ± 0.007  3000  0.344  ± 0.015  0.156  ± 0.006  3200  0.276  ± 0.011  0.144  ± 0.007  3400  0.227  ± 0.013  0.125  ± 0.005  3470  0.196  ± 0.010  0.129  ± 0.006  33  W-C: Station KNR07-4 (01°34’N; 23°38’W) depth  Diss. 230Th  m  Diss. 231Pa  dpm/1000kg (±95% CI)  50  0.054  ± 0.003  0.038  ± 0.007  400  0.130  ± 0.003  0.122  ± 0.009  800  0.218  ± 0.004  0.253  ± 0.017  1100  0.284  ± 0.005  0.310  ± 0.021  1500  0.372  ± 0.008  0.314  ± 0.013  1800  0.430  ± 0.009  0.290  ± 0.011  2100  0.461  ± 0.007  0.326  ± 0.031  2400  0.489  ± 0.007  0.323  ± 0.019  2700  0.490  ± 0.007  0.348  ± 0.020  3000  0.477  ± 0.007  0.314  ± 0.015  3400  0.502  ± 0.006  0.304  ± 0.017  3800  0.571  ± 0.008  0.264  ± 0.015  W-D: Station KNR07-3 (01°12’S; 25°29’W) depth  Diss. 230Th  m  Diss. 231Pa  dpm/1000kg (±95% CI)  50  0.063  ± 0.003  0.051  ± 0.004  400  0.140  ± 0.006  0.128  ± 0.009  800  0.223  ± 0.005  0.234  ± 0.014  1000  0.274  ± 0.004  0.292  ± 0.012  1300  0.325  ± 0.007  0.291  ± 0.016  1500  0.377  ± 0.008  0.275  ± 0.013  2000  0.446  ± 0.009  0.268  ± 0.017  2500  0.449  ± 0.006  0.300  ± 0.016  3000  0.425  ± 0.006  0.283  ± 0.012  3500  0.470  ± 0.006  0.236  ± 0.012  4000  0.572  ± 0.014  0.214  ± 0.012  4500  0.800  ± 0.009  0.263  ± 0.013  34  W-E: Station KNR07-2 (03°44’S; 27°58’W) depth  Diss. 230Th  m  Diss. 231Pa  dpm/1000kg (±95% CI)  50  0.051  ± 0.003  0.052  ± 0.011  300  0.174  ± 0.006  0.104  ± 0.010  900  0.271  ± 0.007  0.271  ± 0.012  1100  0.295  ± 0.008  0.308  ± 0.017  1600  0.431  ± 0.010  0.292  ± 0.017  2100  0.490  ± 0.012  0.278  ± 0.014  2600  0.532  ± 0.011  0.321  ± 0.015  3100  0.598  ± 0.019  0.291  ± 0.016  3600  0.594  ± 0.011  0.227  ± 0.014  4000  0.690  ± 0.011  0.229  ± 0.019  4400  0.848  ± 0.015  0.266  ± 0.014  5000  0.845  ± 0.018  0.262  ± 0.010  W-F: Station KNR07-1 (07°10’S; 31°15’W) depth  Diss. 230Th  m  Diss. 231Pa  dpm/1000kg (±95% CI)  50  0.062  ± 0.003  0.036  ± 0.006  450  0.172  ± 0.008  0.200  ± 0.016  900  0.304  ± 0.010  0.330  ± 0.018  1350  0.389  ± 0.007  0.282  ± 0.015  1756  0.485  ± 0.009  0.306  ± 0.020  2250  0.492  ± 0.014  0.279  ± 0.011  3150  0.567  ± 0.017  0.278  ± 0.013  3556  0.568  ± 0.010  0.238  ± 0.010  4000  0.734  ± 0.011  0.262  ± 0.012  4456  0.908  ± 0.013  0.267  ± 0.016  5000  0.813  ± 0.013  0.254  ± 0.012  35  E-A: Station EN328-9 (45°32’N; 21°24’W) depth  Diss. 230Th  m  Diss. 231Pa  dpm/1000kg (±95% CI)  50  0.030  ± 0.001  0.066  ± 0.004  400  0.139  ± 0.003  0.097  ± 0.007  800  0.194  ± 0.003  0.146  ± 0.008  1000  0.230  ± 0.004  0.168  ± 0.008  1500  0.250  ± 0.004  0.163  ± 0.009  2000  0.312  ± 0.005  0.169  ± 0.010  2500  0.282  ± 0.004  0.176  ± 0.011  3000  0.244  ± 0.009  0.198  ± 0.012  3500  0.272  ± 0.005  0.269  ± 0.014  3827  0.332  ± 0.005  0.300  ± 0.014  E-B: Station EN328-7 (31°00’N; 31°02’W) depth  Diss. 230Th  m  Diss. 231Pa  dpm/1000kg (±95% CI)  50  0.058  ± 0.002  0.069  ± 0.007  400  0.151  ± 0.003  0.078  ± 0.005  800  0.193  ± 0.004  0.147  ± 0.006  1000  0.279  ± 0.006  0.205  ± 0.007  1500  0.361  ± 0.007  0.237  ± 0.011  2000  0.429  ± 0.007  0.274  ± 0.014  2500  0.504  ± 0.008  0.283  ± 0.012  3000  0.651  ± 0.010  0.323  ± 0.011  3500  0.795  ± 0.014  0.364  ± 0.014  4000  0.775  ± 0.012  0.365  ± 0.013  4375  0.767  ± 0.017  0.343  ± 0.010  36  E-C: Station EN328-4 (22°00’N; 36°31’W) depth  Diss. 230Th  m  Diss. 231Pa  dpm/1000kg (±95% CI)  50  0.062  ± 0.002  0.058  ± 0.005  400  0.174  ± 0.003  0.096  ± 0.006  800  0.231  ± 0.004  0.190  ± 0.008  1000  0.275  ± 0.005  0.240  ± 0.011  1300  0.401  ± 0.007  0.289  ± 0.011  1500  0.453  ± 0.008  0.323  ± 0.013  2000  0.689  ± 0.013  0.414  ± 0.012  2500  0.814  ± 0.011  0.449  ± 0.014  3000  0.902  ± 0.012  0.423  ± 0.013  4000  0.983  ± 0.012  0.368  ± 0.013  4997  0.900  ± 0.009  0.292  ± 0.014  5506  0.843  ± 0.010  0.280  ± 0.010  E-D: Station KNR07-9 (12°56’N; 23°21’W) depth  Diss. 230Th  m  Diss. 231Pa  dpm/1000kg (±95% CI)  50  0.061  ± 0.003  0.067  ± 0.011  450  0.142  ± 0.004  0.126  ± 0.017  900  0.233  ± 0.005  -  1300  0.334  ± 0.006  0.349  ± 0.031  1700  0.461  ± 0.008  0.429  ± 0.030  2100  0.561  ± 0.009  0.408  ± 0.028  2500  0.545  ± 0.008  0.415  ± 0.027  3000  0.679  ± 0.011  0.437  ± 0.026  3500  0.770  ± 0.009  0.446  ± 0.038  4000  0.699  ± 0.008  0.332  ± 0.013  4500  0.657  ± 0.012  0.260  ± 0.013  4700  0.590  ± 0.008  0.239  ± 0.010  -  37  E-E: Station KNR07-6 (10°04’N; 23°14’W) Diss. 230Th  depth m  Diss. 231Pa  dpm/1000kg (±95% CI)  50  0.088  ± 0.002  0.077  ± 0.007  450  0.164  ± 0.003  0.132  ± 0.014  900  0.228  ± 0.005  0.227  ± 0.013  1300  0.350  ± 0.008  0.324  ± 0.026  1700  0.466  ± 0.007  0.356  ± 0.021  2100  0.541  ± 0.008  0.352  ± 0.031  2500  0.620  ± 0.010  0.431  ± 0.028  3000  -  0.384  ± 0.027  3500  0.731  ± 0.010  0.366  ± 0.025  4000  0.660  ± 0.009  0.302  ± 0.016  4500  0.707  ± 0.011  0.266  ± 0.012  5000  -  0.240  ± 0.015  -  -  E-F: Station KNR07-5 (07°50’N; 24°37’W) depth  Diss. 230Th  m  Diss. 231Pa  dpm/1000kg (±95% CI)  50  0.056  ± 0.002  0.057  ± 0.006  450  0.150  ± 0.004  0.086  ± 0.007  900  0.254  ± 0.006  0.274  ± 0.013  1300  0.369  ± 0.008  0.310  ± 0.020  1700  0.459  ± 0.008  0.298  ± 0.018  2100  0.554  ± 0.010  0.390  ± 0.020  2500  0.586  ± 0.013  0.350  ± 0.019  3000  0.671  ± 0.013  0.358  ± 0.019  3500  0.665  ± 0.011  0.359  ± 0.023  4000  0.640  ± 0.010  0.300  ± 0.023  4500  0.690  ± 0.010  0.273  ± 0.019  4700  0.702  ± 0.015  0.248  ± 0.017  38  Station KNR06-3 (29°32’S; 43°20’W)  Total 230Th  depth m  Total 231Pa  dpm/1000kg (±95% CI)  12  0.096  ± 0.003  0.045  ± 0.006  401  0.197  ± 0.004  0.070  ± 0.008  797  0.316  ± 0.005  0.140  ± 0.008  1202  0.494  ± 0.006  0.260  ± 0.012  1600  0.572  ± 0.008  0.353  ± 0.016  1998  -  0.307  ± 0.015  2200  0.690  ± 0.009  0.311  ± 0.014  2400  0.692  ± 0.009  0.344  ± 0.020  2800  0.746  ± 0.010  0.376  ± 0.017  3197  -  0.308  ± 0.012  3598  0.961  ± 0.011  0.331  ± 0.014  3944  1.412  ± 0.014  0.326  ± 0.015  -  -  Table 2.3: “Equilibrium” Fractionation Factors.  Latitude  “Equilibrium Fractionation Factor”  70°N – 42.5°S  7.8  42.5°S-45°S  3.1  45°-47.5°S  2.1  47.5°S-50°S  1.4  50°S-70°S  0.9  The “equilibrium” fractionation factor is the fractionation factor that would be measured if particles were in equilibrium with surrounding seawater. In this case [X]p/[X]d = K1X/K-1X and F = (K-1Pa K1Th)/(K1Pa K-1Th). As we will discuss below, however, F measured in the field is also affected by particle sinking rates and circulation. The “equilibrium” 39  fractionation factors used in our control run are set at 7.8 in all waters situated north of 42.5°S. Further south, they decrease gradually to reach a minimum of 0.9 south of 50°S. In order to calculate the transport rates (V; m.y-1) needed to obtain the desired water transport fluxes (Sv), we fixed the width of the Atlantic basin in our model at 3000 km.  2.3 Dissolved 230Th and 231Pa water column profiles: Data-Model comparison  We used water column data (dissolved  230  Th and  231  Pa profiles; fractionation factors) to  constrain the circulation and scavenging parameters in our model. Dissolved 231  230  Th and  Pa profiles from the North and Equatorial Atlantic (Table 2.2; Fig. 2.3) were measured  following the ICP-MS isotope dilution method described by Choi et al. (2000). Samples were collected in 1998 (KNORR 159-7), 1999 (ENDEAVOR 328), and 2005 (ENDEAVOR 407). In this section, we present the fit between field data and those generated by our control run and discuss the processes that generate them.  Simple scavenging models using constant S, K1 and K-1 and neglecting circulation predict a linear increase in dissolved and particulate  230  Th and  231  Pa concentrations versus depth  (Bacon and Anderson, 1982; Bacon et al., 1985; Nozaki et al., 1987):  [X]p = [PX / S] Z  (2.5)  [X]d = [PX / K1] + [(K-1 PX)/(K1 S)] Z  (2.6)  Where Z is depth. 40  However, most 230Th and 231Pa seawater profiles measured in the ocean display significant deviations from linearity because the effect of circulation can rarely be neglected. The dissolved  230  Th and  231  Pa concentration profiles obtained with our model using the  parameters listed in Table 2.1 also deviate often from linearity and are broadly consistent with observations.  Diss. 231Pa: dpm/1000kg  Diss. 230Th: dpm/1000kg 0.0 0.0  0.2  0.4  0.6  0.8  0.1  0.2  0.3  0.4  0  1.0  0  1000  2000  Control run: 60°70°N  Depth (m)  Depth (m)  1000  2000  No circulation  IOC93-2: 54°30'N; 48°28'W 3000  Labrador: 58°11'N; 50°52'W  3000  Labrador: 58°11'N; 50°52'W  IOC93-13: 64°48'N; 6°12'W  Control run: 60°-70°N No circulation 4000  4000  Figure 2.4: Dissolved 230Th and 231Pa obtained with the control run at 60°- 70°N and measured in the Norwegian Sea (IOC93-13: Moran et al., 1995) and the Labrador Sea in 1993 (IOC93-2; Moran et al., 1997) and 1999 (Labrador: Moran et al., 2002).  The model reproduces reasonably well the water column profiles measured in the Labrador and Norwegian Sea (Fig. 2.4). Shallow waters entering the Nordic Seas to 41  produce deep water have low  230  Th and  231  Pa concentrations and deep winter convective  mixing results in low and nearly constant concentration profiles. Concentrations are higher at shallow depths and lower in deep waters than predicted by the scavenging model in the absence of vertical mixing. The fit of the modeled  230  Th is best with the profiles  measured in the Labrador Sea in 1993 (Moran et al., 1997) and in the Norwegian Sea (Moran et al., 1995). The  230  Th concentrations measured in the Labrador Sea in 1999 are  significantly higher and have been attributed to a temporary cessation of deep water convection in the Labrador Sea during that period (Moran et al., 2002). The build-up of 231  Pa resulting from the same effect is expected to be much smaller (the response time  depends the residence time and is longer for  231  Pa; see below), and we find a reasonable  fit between the model and the 1999 Labrador Sea measurements of dissolved  231  Pa  (although the model generates somewhat higher concentrations than observed).  The concentration deficit in deep waters generated in the Nordic and Labrador Seas spreads southward with the North Atlantic Deep Water. During transit to the Southern Ocean, the newly formed deep water is continuously subjected to the particle rain that originates from surface waters and which scavenges the  230  Th and  231  Pa continuously  produced in the water column. When particles reach the depth of the newly formed deep water where the  230  Th and  231  Pa seawater concentrations are below steady-state  concentrations dictated by scavenging, desorption from particles is enhanced and a fraction of the  230  Th and  231  Pa scavenged at shallower depths is released to the deep  waters instead of being removed into the underlying sediments. Thus, the  230  Th and  231  Pa  concentrations in newly formed deep waters gradually increase during transit to the 42  southern ocean until the concentration at steady state with respect to scavenging is regained, at which point the water column profiles have relaxed back to linearity (Francois, 2007). The gradual relaxation of the profiles to linearity has been described for each isopycnal by adding a lateral transport term to the 1-D scavenging model, as first proposed by Rutgers van der Loeff and Berger (1993): X]t/t = PX – S (K[X]t)/Z + (i[X]t – [X]t)/w = 0  (2.7)  Where i[X]t and [X]t are total 230Th or 231Pa concentration measured at two locations on the same isopycnal with i[X]t the concentration in the upstream source region and w is the “transit time” of water between these two sites. In deep waters, K (= [X] p/[X]t) is nearly constant. Integrating the above equation thus gives: [X]t  (PX w + i[X]t) (1 – e-Z/wSK)  (2.8)  In the absence of circulation or mixing, and assuming a constant K, the reversible scavenging model predicts that ss[X]t  PX Z / SK, where ss[X]t is the total concentration of 230  Th or  231  Pa in seawater at steady state with respect to scavenging. Z/SK is thus the  residence time with respect to addition by uranium decay and removal by scavenging when the profile has regained linearity (i.e. when it has regained steady state with respect scavenging), defined as ss = ss[X]t / PX. Therefore, ss = Z / SK and: [X]t  (PX w + i[X]t) (1 – e-ss/w)  (2.9)  43  Equation (2.9) predicts that the radioisotope profiles relax back to linearity more slowly with increasing ss and therefore water depth. Profile linearity is thus regained closer to the source at shallower depths, and  230  Th regains linearity faster than  231  Pa because of its  shorter ss. The shapes of the 230Th profiles measured in the Atlantic are in agreement with this simple conceptual model and are also reproduced in the control run (Figs. 2.5 and 2.6). The seawater data show clearly the gradual southward relaxation of the profiles towards linearity. Linearity is regained faster for 230Th and at shallower depths. We also note that the profiles from the western Atlantic display a greater deficit farther south, reflecting the stronger ventilation of the western Atlantic basins. The profiles obtained from the model (Figs. 2.5c and 2.6c) show similar trends with dissolved  230  Th and  231  Pa concentrations  close to those observed in the ocean.  230Th 0.0  0.2  230Th  - dpm/1000kg West 0.4  0.6  0.8  1.0  0.0  1.2  EN407-3: 39°28'N; 68°22'W EN407-4: 38°36'N; 68°53'W KNR07-3: 1°12'S; 25°29'W KNR07-4: 1°34'N; 23°38'W KNR07-2: 3°44'S; 27°58'W KNR07-1: 7°10'S; 31°15'W No circulation  1000  Depth (m)  Depth (m)  2000  - dpm/1000kg East 0.4  0.6  0.8  1.0  1.2  0  0  1000  0.2  2000  3000  3000  4000  4000  5000  5000  EN328-9: 45°32'N; 21°24'W EN328-7: 31°00'N; 31°02'W EN328-4: 22°00'N; 36°31'W KNR07-9:12°56'N; 23°21'W KNR07-6:10°04'N; 23°14'W KNR07-5:7°50'N; 24°37'W No circulation  44  230Th 0.0  0.2  - dpm/1000kg Model 0.4  0.6  0.8  1.0  1.2  0  56N 51N 44N  1000  34N 6S No circulation  Depth (m)  2000  3000  4000  5000  Figure 2.5: Dissolved  230  Th measured (a) in the western Atlantic, (b) in the eastern Atlantic, and (c)  produced with the control run  0.0 0  Depth (m)  1000  2000  0.2  231Pa  - dpm/1000kg West 0.4  0.6  EN407-3: 39°28'N; 68°22'W EN407-4: 38°36'N; 68°53'W KNR07-3: 1°12'S; 25°29'W KNR07-4: 1°34'N; 23°38'W KNR07-2: 3°44'S; 27°58'W KNR07-1: 7°10'S; 31°15'W No circulation  0.0  0.8 0  1000  Depth (m)  231Pa  2000  3000  3000  4000  4000  5000  5000  0.2  - dpm/1000kg East 0.4  0.6  0.8  1.0  EN328-9: 45°32'N; 21°24'W EN328-7: 31°00'N; 31°02'W EN328-4: 22°00'N; 36°31'W KNR07-9:12°56'N; 23°21'W KNR07-6:10°04'N; 23°14'W KNR07-5:7°50'N; 24°37'W No circulation  45  231Pa 0.0  0.2  - dpm/1000kg Model 0.4  0.6  0.8  0  56N 51N 44N  1000  34N  Depth (m)  6S No circulation  2000  3000  4000  5000  Figure 2.6: Dissolved  231  Pa measured (a) in the western Atlantic, (b) in the eastern Atlantic and (c)  produced with the control run.  Further south, where the deep waters start to upwell, their relatively high  230  Th  concentrations exceed the concentrations predicted by the scavenging model in absence of circulation, resulting in convex dissolved profiles (Francois, 2007). This is again clearly seen in measured seawater profiles (Rutgers van der Loeff and Berger, 1993) and model results (Fig. 2.7a). This trend is less apparent for response time, preventing the  231  Pa (Fig. 2.7b) because of its slower  231  Pa profiles from regaining linearity before reaching the  Southern Ocean.  46  231  230  Pa - dpm/1000kg  Th - dpm/1000kg 0.0  0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8  0.2  0.4  0.6  0.8  0  0  ANT VIII/3 1785 (55° 06'S; 27°51'W) 64S-Model  1000  1000  No circulation  2000  m  m  2000  3000  3000 ANT VIII/3 1785 (55° 06'S; 27°51'W)  4000  4000  64S-Model  No circulation  5000  5000  Figure 2.7: Concentration profiles of dissolved  230  Th and  231  Pa measured (Rutgers van der Loeff and  Berger, 1993) and modeled (control run) in the southern ocean  2.4 Fractionation factors: Data-Model comparison  Fractionation factors (F) are most often obtained by measuring dissolved and particulate 231  Pa and  230  Th concentrations in the same seawater sample and applying equation 4 (e.g.  Anderson et al., 1983; Walter et al., 1997; Moran et al., 2002). These measured values are generally viewed as being mostly controlled by particle composition, with opal having a much lower F than the other major constituents of marine particles (Chase et al., 2002; Guo et al., 2002; Geibert and Usbeck, 2004). Our model reflects the generally accepted view that F is much lower in the opal dominated Southern Ocean than in the 47  carbonate-dominated Atlantic Ocean and we chose adsorption and desorption rate constants to generate “equilibrium” fractionation factors broadly consistent with field observations (Table 2.3). The fractionation factor generated by the model using equation (4) indicates, however, that F is also significantly affected by the sinking rate of particles and ocean circulation (Fig. 2.8). This is because the chemical equilibrium between particles and seawater cannot be reached when particles sink through vertical dissolved 230  Th and  231  Pa concentration gradients. In the absence of circulation, we can rearrange  equations 2.4 – 2.6 to show that: K1Th (S + K-1Pa Z) F = ----------------------K1Pa (S + K-1Th Z)  (2.10)  If the two nuclides have identical desorption rate constants, F would be independent of sinking rates in the absence of circulation. However, if K-1Pa < K-1Th and K1Pa < K1Th (Table 2.1) then F calculated with sinking particles rises well above equilibrium values and gradually decreases towards the equilibrium value with depth (Fig. 2.8a). When particles sink through the Atlantic Meridional Overturning cell, the fractionation factors estimated from equation 4 drop below the equilibrium fractionation factor within the core of the NADW (Fig. 2.8b). The fractionation factors measured in the field are therefore not directly comparable to those derived from equilibrium absorption experiments conducted in the laboratory (e.g. Geibert and Usbeck, 2004; Guo et al., 2002).  48  Figure 2.8: Distribution of fractionation factors: (a) obtained in our model with sinking particles but without circulation; (b) obtained with sinking particles in the control run with AMOC (NADW: 21.5Sv; 49  AABW: 8Sv). Note that sinking rates and circulation significantly affect the fractionation factor defined as (231Pad/230Thd)/(231Pap/230Thp) (see text for explanation).  In contrast to the F generated by our model, the fractionation factors measured by Moran et al. (2002) and Scholten et al. (2008) in the equatorial and southern Atlantic increase with depth down to ~ 1500 m and stay roughly constant or decrease further down. Also, the natural variability in F is much larger than the range observed in our model. The reason for this discrepancy could be depth variation in particle composition, a factor that is not taken into account in our model. Scholten et al. (2008) invoke a drop in the opal content of particles to explain the trend. However, the few available data on the composition of suspended particles (in the Sargasso Sea, Sherrell and Boyle, 1992; and in the North Pacific, Sherrell et al., 1998) do not show a clear trend with depth (except for one profile taken in spring 1991 in the North Pacific). Clearly, more data are needed before adding this variable in any model and this discrepancy must be left unresolved for now.  2.5 230Th and 231Pa distribution in the control run:  Since our control run is broadly consistent with the limited water column data that are available, we can discuss the general distribution of  230  Th and  231  Pa generated by the  model with some level of confidence.  50  2.5.1 Dissolved 230Th:  The model clearly generates the downward penetration of low dissolved  230  Th by deep  convection in the high northern Atlantic (Fig. 2.9). The horizontal isolines between 20°N and 30°S indicate however that the vertical dissolved  230  Th profiles quickly regain  linearity, as is observed in the field data. South of 30°S, dissolved  230  Th concentrations  start to increase at all depths as a result of deep water upwelling (Fig. 2.2). The increase in dissolved  230  Th concentration is enhanced south of 50°S by the lower adsorption rate  constants imposed in this region to reflect the dominance of biogenic silica (Table 2.1), while the formation of AABW results in dissolved  230  Th maxima at intermediate depths,  similar to observations (Fig. 2.7a).  Figure 2.9: Dissolved 230Th section generated by the model 51  230  In surface water, dissolved  Th concentration is significantly higher in the Southern  Ocean, as has been noted in field data (Rutgers van der Loeff and Berger, 1993; Walter et al., 2001; Chase et al., 2003). 2.5.2 Particulate 230Th:  The pattern of distribution of particulate of dissolved  230  Th concentration (Fig. 2.10) is similar to that  230  Th. There is a conspicuous maximum in particulate  230  Th just north of the  230  southern opal belt, which is a result of the increase in dissolved resulting from deep water upwelling. The sharp drop in particulate  Th concentration  230  Th further south is a  direct consequence of the lower K1Th in the southern ocean.  Figure 2.10: Particulate 230Th section generated by the model 52  Total  230  Th concentration profiles measured in the western (Table 2.2; Fig. 2.11) and  eastern (Scholten et al., 2008) South Atlantic display a near bottom maximum similar to that generated by the model. This result may help explain the presence of a near-bottom maximum in total  230  Th when there is no clear evidence for the presence of a nepheloid  layer (Scholten et al., 2008). However, our model generates these near-bottom maxima further south than observed, suggesting that a better representation of the AMOC in our model may require that the shoaling of the deep limb of the overturning cell starts further to the north.  Figure 2.11: Total  230  Th concentration measured in the southwestern Atlantic at Station KNR06-3  (Table 2.2) showing a near-bottom maximum similar to that generated in the South Atlantic by our model.  53  The fraction of total  230  Th in particulate form generated by the model ([ 230Th]p/ [230Th]t)  ranges from 0.18 to 0.22 at low latitudes (Fig. 2.12), which is also conforming to field observations (Bacon and Anderson, 1982; Moran et al., 2002). In the Southern Ocean below 1000m, the model produces somewhat lower fractions in particulate form (0.16-0.18), reflecting the lower affinity of biogenic silica for Th (Table 2.1). Somewhat higher fractions are generated in the upper water column of the Southern Ocean (0.22) and in the Nordic Sea (0.28) reflecting deep convection and longer residence time of particles in these waters.  Figure 2.12: Fraction of total 230Th associated with particles generated by the model  54  2.5.3 Dissolved 231Pa: As for dissolved 230Th, the model produces a clear downward penetration of low dissolved 231  Pa by deep convection at high northern latitudes (Fig. 2.13). However, following  expectations and observations, the minimum associated with the core of the NADW propagates much further south, reaching the southern ocean. South of 30°S, dissolved 231  Pa concentrations start to increase as a result of deep water upwelling but the effect is  not as pronounced as for  230  Th because of the higher adsorption rate constants imposed in  the southern ocean to reflect the dominance of biogenic silica (Table 2.1).  Figure 2.13: Dissolved 231Pa section generated by the model  55  Surface water dissolved 231Pa concentrations are significantly higher in the southern ocean, but the effect is less pronounced than for 231  230  Th because of the higher scavenging rate of  Pa.  2.5.4 Particulate 231Pa:  The most prominent feature in the distribution of particulate  231  Pa concentration is the  concentration maximum in the southern ocean (Fig. 2.14), resulting from the higher K1Pa used in this region. The fraction of particulate  231  Pa generated by the model north of 45°S  remains uniform between 0.04 and 0.05 (Fig. 2.15), in general agreement with observations (Moran et al., 2002), while the higher values generated in the southern ocean (0.16 – 0.20) are consistent with some of the extreme values reported by Rutgers van der Loeff and Berger (1993). Profiles of total  231  Pa generated in the south Atlantic in the  model are intermediate between measurements made in the western and eastern side of the basin (Fig. 2.16). The lower concentrations measured in the western Atlantic than in the eastern Atlantic suggest that the Deep Western Boundary Current rather than boundary scavenging plays a major role in controlling the distribution of  231  Pa in the water column  of this region.  56  Figure 2.14: Particulate 231Pa section generated by the model  Figure 2.15: Fraction of total 231Pa associated with particles generated by the model  57  Figure 2.16: Total  230  Th concentration measured in the southwestern Atlantic at Station KNR06-3  (Table 2.2) and in the eastern South Atlantic (Scholten et al., 2008) compared to the model results at 40 ºS.  2.5.5 Dissolved 231Pa/230Th:  Modeled dissolved  231  Pa/230Th ratios systematically decrease with water depth in the  North and Equatorial Atlantic, while this trend is less pronounced in the South Atlantic (Fig. 2.17). Data from nine of the North and equatorial Atlantic stations presented in Table 2.2 reflect this trend with a clear decrease in dissolved (Fig. 2.18). In shallower water, dissolved  231  Pa/230Th with depth below 500m  231  Pa/230Th is more variable. This may be a 58  result of the short residence times of  230  Th and  limited lateral transport. Shallow dissolved  231  Pa at these shallow depths and their  231  Pa/230Th are likely to be more affected by  local changes in particle composition. The lack of a clear trend with depth below 1000 m generated by the model in the South Atlantic is consistent with the observations of Scholten et al. (2008).  Figure 2.17: Dissolved 231Pa/230Th section generated by the model  The model also predicts that the highest ratios would be found in the surface water of the South Atlantic (Fig. 2.17). Walter et al. (2001) report an increasing trend in surface water dissolved 231Pa/230Th from 0.5 to 2.0 between 65°S and 40°S (their figure 2.4c). However, water column profiles from the South Atlantic available to date (Moran et al., 2002) fail to 59  document the predicted large ratios in surface water. High ratios are generated in our model because surface waters from the southern ocean with relatively high dissolved 230  Th and  231  Pa concentrations are advected north. Since  by scavenging, dissolved  230  Th is more quickly removed  231  Pa/230Th initially increases to eventually decrease farther  north as the scavenging of 231Pa “catches up” with that of 230Th. Evidently, the complexity of surface water movement in the South Atlantic cannot be fully captured in our simple 2D model and these very high surface values may be artifacts of our simplified circulation. This question needs to be further explored with three dimensional models.  Figure 2.18: Dissolved 231Pa/230Th profiles measured at 9 stations in the North and Equatorial Atlantic (Table 2.2). 60  2.5.6 Particulate 231Pa/230Th: The distribution of particulate 231Pa/230Th generated by the model is shown in Figure 2.19. We can take these values as representing the  231  Pa/230Th that sediments would have if they  were deposited at a given depth and latitude.  Figure 2.19: Particulate 231Pa/230Th section generated by the model  However, as mentioned when discussing the fractionation factors, settling particles in our model are not in chemical equilibrium with surrounding waters. When particles reach the seafloor, they could possibly come into equilibrium with bottom waters. Whether they do or not depends on how long they are in contact with bottom waters before burial as a 61  result of sedimentation and bioturbation. With the rate constants used in our model, it would take 1-3 years (depending on initial conditions) for surface sediments to be within 95% of their equilibrium value with bottom waters. We can calculate sediment  231  at equilibrium with bottom waters using [X] p/[X]d = K1X/K-1X for  231  230  Th and  Pa/230Th Pa (Fig.  2.20).  Figure 2.20: Sediment  231  Pa/230Th section generated by the model assuming that sediment reach  equilibrium with bottom water (see text for explanations).  Partial equilibration would result in sediment  231  Pa/230Th intermediate between values  reported in Figure 2.19 and 2.20. The difference is relatively small in deep water but significantly larger in shallower waters. This is consistent with the observation of 62  Scholten et al. (2008) who remarked that, at shallow depths, particles are significantly lower than  231  Pa/230Th in suspended  231  Pa/230Th in surface sediments (their figure 5) and  suggest that surface sediments do reach equilibrium with bottom water.  2.6 Sediment 231Pa/230Th: Data-Model comparison In this section, we compare the distribution of sediment  231  Pa/230Th generated by the  model with 231Pa/230Th measured in Atlantic sediments as a test for further validation. The distribution of particulate (Fig. 2.19) and sediment (Fig. 2.20) 231Pa/230Th generated by the model is clearly controlled both by circulation and particle composition. We find the lowest values near the base of the two overturning cells just downstream of the sites of deep water formation and the highest values in the Southern Ocean. The low values are clearly generated by the overturning circulation cells, while the high values in the southern ocean are a direct consequence of particle composition.  Sediment  231  Pa/230Th generally decreases with depth, a pattern dictated by trends in  dissolved  231  Pa/230Th which is generated by the overturning circulation. A similar  decreasing trend from ~ 0.13 at ~ 1000m to ~ 0.04 at ~ 5000m has been reported by Scholten et al. (2008) for surface sediments in the South Atlantic. Holocene  231  Pa/230Th  from the five North Atlantic cores discussed by Gherardi et al. (2009) also show a similar trend, with values approaching the production rate ratios for the two shallower cores and lower values for the three deeper cores (Table 2.4).  63  Table 2.4: Holocene Pa/Th in 5 North Atlantic cores (Gherardi et al., 2009) Core  Position  Water Depth (m)  Holocene 231Pa/230Th  DAPC2  58°58’N, 09°36’W  1709  0.093 ±0.001  MD95-2037  37°05’N, 32°01’W  2150  0.093 ±0.004  SU81-18  37°46’N, 10°11’W  3135  0.064 ±0.005  SU90-44  50°01’N, 17°06’W  4279  0.052 ±0.004  OCE326-GGC5  33°42’N, 57°35’W  4550  0.054 ±0.004  Values reported for core tops from the Nordic Seas range from 0.07 to 0.09 (Yu et al., 1996). Our model generates these values with an equilibrum fractionation factor of 7.8, somewhat higher than the fractionation factors measured in the Labrador Sea (3-7; Moran et al., 2002). Significantly higher sediment  231  Pa/230Th have been reported, however, just  south of Iceland and the Denmark Strait (0.10-0.15; Yu et al., 1996; Anderson, pers. comm.) but they are generally found in sediments deposited between 1500m and 2000m water depth and seems confined to a relatively small area where Leinen et al. (1986) report opal concentration (carbonate-free wt %) of up to 20%. With the fractionation factors reported in Table 2.3 and Figure 2.8, our model generates sediment  231  Pa/230Th  below the production rate ratio at this depth range just south of the site of deep water formation (Fig. 2.20). The model generates the high values reported in this region only if we lower the equilibrium fractionation factor to 3.9 (Fig. 2.21). 64  Figure 2.21: (a) Sediment  231  Pa/230Th generated with an opal belt just south of the site of deep water  formation. (b) Differences in the sediment 231Pa/230Th field generated in with and without the northern opal belt. 65  2.7 Discussion 2.7.1 The effect of AMOC on sediment 231Pa/230Th  In the absence of any circulation, the model generates a field of constant sediment 231  Pa/230Th equal to the production rate ratio (0.092). In this case, changes in the  fractionation factor (Table 2.3) produce changes in the dissolved 230Th and 231Pa fields but not in the particulate fields. The distribution of particulate and sediment  231  Pa/230Th  reported in Figure 2.19 and 2.20 should thus provide information on the ocean overturning circulation. The model results clearly indicate, however, that the relationship between sediment  231  Pa/230Th at any given site and the overturning circulation is very complex, as  was also noted by Siddall et al. (2007). Sediment  231  Pa/230Th depends not only on the rate  of the overturning and particle scavenging, but also on the detailed geometry of the overturning cell and the distance between the coring site and the site of deep water formation. Sediment  231  Pa/230Th reaches a minimum at a depth dictated by the geometry  of the overturning cell and at a latitude dictated by the position of the site of deep water formation and the strength of the overturning circulation (Fig. 2.22). Clearly, it is impossible to constrain the history of changes in the AMOC from the evolution of 231  Pa/230Th at one site, as was attempted by McManus et al. (2004).  2.7.1.1 Vertical variations in sediment 231Pa/230Th induced by the AMOC  The use of sediment  231  Pa/230Th to reconstruct past changes in the AMOC relies on the  longer residence time of  231  Pa in the water column. While the short residence time of 66  230  Th severely limits the extent to which it can be laterally transported after its production  by uranium decay, the longer residence time of  231  Pa results in its extensive redistribution  by ocean circulation.  According to equation (2.9), the  231  Pa profiles relax back to linearity at a rate that  decreases as ss (the residence time with respect to addition by uranium decay and removal by scavenging in the absence of circulation or mixing) increases.  ss is  proportional to water depth. Therefore, if the rate of lateral volume transport were the same at all depths, the fraction of the  231  Pa production that is laterally transported with the  water would increase with depth. This effect contributes to the general decrease with depth in dissolved (Fig. 2.17) and particulate (Fig. 2.19)  231  Pa/230Th generated by the  model and measured in sediments (Table 2.4). Very little 231Pa can be laterally exported by circulation at shallow depths but an increasing fraction can be exported with increasing depth. Sediment 231Pa/230Th integrates the lateral export of  231  Pa over the entire overlying  water column. The integration in terms of lateral volume transport, however, is not linear but weighted by ss. At similar rates, shallow overturning cells lower sediment  231  Pa/230Th  at the base of the cells less than deeper overturning cells. The relationship between changes in sediment  231  Pa/230Th with depth and changes in lateral volume transport with  depth is therefore complex and difficult to intuit. In our control run, sediment  231  Pa/230Th  reaches its lowest value at the depth where we find the highest rate of lateral volume transport (Figs. 2.22a, b), but, this is not always the case. For instance, if we use the zonally integrated overturning rates recently derived from the ECCO consortium dataset 67  (Wunsch and Heimbach, 2006), the lowest sediment  231  Pa/230Th is reached 1000m below  the depth of maximum lateral volume transport (Figs. 2.23a,b). 2.7.1.2 Horizontal variations in sediment 231Pa/230Th induced by the AMOC Sediment 231Pa/230Th also changes systematically with latitude or distance from the site of deep water formation. Latitudinal changes in sediment  231  Pa/230Th at the depth where the  minimum ratio is reached documents an initial decrease with distance from the site of deep water formation, followed by an increase (Fig. 2.22c). Dissolved  230  Th and  231  Pa  concentrations are low throughout the water column at the site of deep water formation (Fig. 2.4). Because of its shorter ss,  230  Th concentration increases faster to reach its  steady-state concentration with respect to scavenging (equation 2.9), thereby gradually decreasing dissolved, particulate and sediment  231  Pa/230Th. Once dissolved  reached its maximum value, the slower increase in dissolved  230  Th has  231  Pa results in a slow  increase in 231Pa/230Th further downstream.  68  Figure 2.22: (a) Lateral velocity profile in the control run between 60ºN and 35ºN. (b) Vertical sediment 231Pa/230Th bathymetric profiles generated by the model at different rates of overturning and at the latitudes where lowest sediment 231Pa/230Th is found. (c) Latitudinal sediment 231Pa/230Th profiles for different rates of overturning at the depth where the lowest sediment  231  Pa/230Th is found (3625m)  (red symbols represent the lowest depth of maximum lateral velocity (a), minimum 231Pa/230Th (i.e. the depth for the latitudinal profiles) (b) and latitude of minimum  231  Pa/230Th (i.e. the latitudes for the  vertical profiles) (c))  69  Figure 2.23: (a) Contrasting lateral velocity profiles between the control run, the overturning profile from the ECCO consortium (14Sv, Wunsch and Heimbach, 2006) and an arbitrary shallower overturning cell (20.5 Sv). (b) Vertical sediment the latitude where the lowest  231  Pa/230Th generated by three overturning profiles at  231  Pa/230Th is found. (c) Latitudinal sediment  generated by three overturning cells at the depth where the lowest sediment  231  231  Pa/230Th profiles  Pa/230Th is found (red  symbols represent the lowest depth of maximum lateral velocity or the base of the shallow overturning cell (a), minimum  231  Pa/230Th (i.e. the depth for the latitudinal profiles) (b) and latitude of minimum  231  Pa/230Th (i.e. the latitudes for the vertical profiles) (c)).  70  2.7.1.3 Changes in sediment  231  Pa/230Th resulting from changes in the rate of the  AMOC Increasing the rate of overturning in the control run without changing the geometry of the overturning cell has several effects on the distribution of Atlantic sediment 231Pa/230Th: (1) It pushes the zone of minimum  231  Pa/230Th farther away from the site of deep water  formation (Fig. 2.22c); (2) The latitudinal minimum in sediment decrease, but instead increases (Figs. 2.22b, c) (3) Sediment  231  Pa/230Th does not  231  Pa/230Th also increases at  the site of deep water formation and directly south of it (Fig. 2.22c); (4) The vertical gradient of sediment 231  231  Pa/230Th at the latitude corresponding to the minimum sediment  Pa/230Th increases (Fig. 2.22b); (5) the largest decrease in sediment  231  Pa/230Th  downstream of the deep water formation zone is found in the Southern and equatorial region (Fig. 2.22c). Even without changing the geometry of the overturning cell and particle scavenging, the same value of sediment  231  Pa/230Th can be generated at one site  by different rates of overturning. For instance, the same value of 0.052 is produced at latitude 36.25°N at 3635m with overturning rates of 10.25 Sv and 30.75 Sv (Fig. 2.22c). This observation reinforces the fact that sediment  231  Pa/230Th at one site cannot uniquely  constrain the rate of the AMOC.  2.7.1.4 Changes in sediment  231  Pa/230Th resulting from changes in the geometry of  the AMOC  We find systematic changes in the distribution of sediment  231  Pa/230Th when we impose a  shallower overturning cell without changing the rate of overturning: (1) The depth of 71  minimum sediment  231  Pa/230Th tends to shoal (Figs. 2.23a, b), although that might not be  always the case; (2) The latitudinal gradient at the depth of minimum sediment decreases (higher sediment 231  231  Pa/230Th  231  Pa/230Th in the North Atlantic and lower sediment  Pa/230Th in the South Atlantic, Fig. 2.23c) because  Pa has a shorter ss in shallower  231  water and is less effectively exported horizontally; (3) Sediment  231  Pa/230Th increases  rapidly with depth below the base of the overturning cell (Fig. 2.23b), largely corroborating the finding of Thomas et al. (2006) that the sediment  231  Pa/230Th signal  generated by a shallow overturning circulation is, if not totally absent, at least strongly attenuated in sediments deposited more than 1000m below the base of the overturning cell.  2.7.1.5 Possible sampling strategy to constrain past changes in AMOC from sediment 231  Pa/230Th  These results suggest a possible sampling strategy to constrain past changes in the rate and geometry of the AMOC. A series of bathymetric profiles down the eastern and western slope of the North Atlantic, the Mid Ocean Ridge, or the flanks of seamounts, with due attention to possible changes in sediment composition, could document the vertical and horizontal sediment 231Pa/230Th gradients and the depth of minimum sediment 231  Pa/230Th for different time slices. The shape of the vertical profiles would inform us on  the geometry of the meridional overturning cells, while the gradients (horizontal and vertical) would provide constraints on the rate of the overturning. Figure 2.22 and Figure 2.23 also suggest that sediment  231  Pa/230Th at the site of deep water formation may be 72  sensitive to the rate and depth of the AMOC. Whether these simple systematic trends can be reproduced in more complex circulation models, however, still needs to be verified. 2.7.2 The effect of AABW on sediment 231Pa/230Th  Figure 2.20 clearly indicates that the overturning cell initiated in the Southern Ocean by the formation of AABW significantly contributes to lowering sediment  231  Pa/230Th in the  South Atlantic. If we eliminate the formation of AABW, sediment 231Pa/230Th in the South Atlantic significantly increases (Fig. 2.24). The process whereby AABW is producing these low sediment  231  Pa/230Th is the same as for the northern overturning cell but the  effect is found at greater depth and is less pronounced because of the smaller flow of water involved and the higher initial dissolved  230  Th and  231  Pa in the water that generates  AABW. The low sediment 231Pa/230Th (< 0.05) in the deep Southeast Atlantic (Scholten et al., 2008) are consistent with the importance of AABW in generating low  231  Pa/230Th in the South  Atlantic and suggest that sedimentary records in this region, if unaffected by changes in opal flux, could generate important constraints on variations in the rate of formation of this important water mass (Negre et al., 2010).  73  Figure 2.24: Sediment  231  Pa/230Th field generated in the control run without the formation of the  AABW.  2.7.3 The effect of particle composition on sediment 231Pa/230Th  As already indicated above, in the presence of circulation and/or mixing, localized changes in particle composition and fractionation factors produce dramatic but localized changes in sediment  231  Pa/230Th (Fig. 2.21). Such changes can be taken into account by  analyzing the opal content of the sediment from which the  231  Pa/230Th record is obtained  (Gherardi et al., 2009) with, however, one important caveat. Opal is undersaturated 74  throughout the ocean and much of it dissolves before burial. Below a certain threshold in opal flux and sediment mass accumulation rates, opal is not preserved in sediments but the 231  Pa/230Th generated by the presence of opal in sinking particles could persist. We could  further address this question by using a diagenetic model (e.g. Khalil et al., 2007) to estimate the opal concentration in sinking particles reaching the seafloor from sediment mass accumulation rates and use this information to estimate the range of possible fractionation factors to be applied at this site using the sediment trap data compilation of Chase et al. (2003). However, distinguishing between the importance of changes in circulation and opal flux will eventually be best addressed by generating a database large enough to obtain a near-synoptic view of the spatial distribution of sediment  231  Pa/230Th  for each time slice of interest, since the distribution generated by the overturning circulation is clearly distinct from the distribution generated by the distribution of opal productivity in the ocean.  While Figure 2.21 clearly demonstrates the potential impact of localized variations in fractionation factors, it also shows, and maybe more importantly, that such changes in the North Atlantic have little impact on the  231  Pa/230Th deposited downstream (231Pa/230Th <  0.002; Fig. 2.21b).  75  Figure 2.25: The influence of Southern Ocean fractionation factor on the sediment  231  Pa/230Th in  Atlantic sediments. (a) Lateral velocity field used to conduct the experiments. (b) Vertical sediment 231  Pa/230Th gradients generated by the control run (1*FF) and when the Southern Ocean equilibrium  fractionation factor (FF=0.9) is doubled (2*FF) or halved (0.5*FF). (c) Sediment  231  Pa/230Th produced  in the North Atlantic and the Southern Ocean under these three scenarios.  This is however not the case when we change the fractionation factor in the Southern Ocean (Fig. 2.25). Doubling the equilibrium fractionation factor in the Southern ocean from 0.9 to 1.8 not only decreases sediment  231  Pa/230Th in the Southern Ocean from ~ 0.3 76  to ~ 0.2 but also uniformly increases sediment  231  Pa/230Th along the latitudinal transect of  the Atlantic by nearly ~ 0.01. Reducing the fractionation factor increases the Southern Ocean  231  Pa sink and decreases  231  Pa/230Th in the Atlantic. However, the slope of the  latitudinal gradient of sediment 231Pa/230Th in the Atlantic is not significantly affected and could still be used to constrain the rate of the overturning. Nonetheless, accurately assessing the extent of the southern ocean  231  Pa sink will be important to evaluate the rate  of AMOC.  2.8 Conclusions  We have developed a simple 2D scavenging model to address some of the questions that have been raised concerning the use of sediment  231  Pa/230Th as a paleocirculation tracer  (Keigwin and Boyle, 2008; Scholten et al., 2008; Lippold et al., 2009). Although our circulation model is clearly too simple to capture all the complexity of ocean circulation, it reproduces many of the features observed in the distribution of dissolved 231  Pa and sediment  230  Th and  231  Pa/230Th and provides a tool to start assessing the relative  importance of circulation and particle scavenging in controlling the distribution pattern of sediment 231Pa/230Th in the Atlantic.  The circulation scheme imposed in our model broadly reflects the flow of the main deep Atlantic water masses (NADW, AABW). The detailed geometry of the two overturning cells and the parameters of the imbedded scavenging model have been tuned to reproduce 77  the broad features of the distribution of dissolved 230Th and 231Pa and fractionation factors measured in the water column to date. The model produces a general decrease in dissolved, particulate and sediment  231  Pa/230Th with depth, which is consistent with field  observations (Fig. 2.18; Scholten et al., 2008; Gherardi et al., 2009). It also produces patterns in the distribution of sediment  231  Pa/230Th which could be used to distinguish the  circulation signal from the effect of particle scavenging. The model output also suggests sampling strategies to optimize the information in past circulation that could be derived from sediment 231Pa/230Th. The most robust circulation signals generated by the model are the vertical and horizontal sediment  231  Pa/230Th gradients, which changes systematically  with the rate and geometry of the AMOC (Figs. 2.22, 2.23). However, we still need to establish whether these diagnostic trends can also be produced with more complex 3D circulation models.  We have used our 2D model to test the extent to which changes in fractionation factor can obliterate the patterns of sediment  231  Pa/230Th generated by the overturning circulation.  While it is clear that changes in particle composition in the North Atlantic can change sediment 231Pa/230Th locally, our model indicates that the  231  Pa/230Th pattern generated by  circulation further downstream is not significantly affected. This may be different for the Southern Ocean, which is the main sink for 231Pa in our model. Changing the fractionation factor in the Southern Ocean offsets  231  Pa/230Th but has little impact on the gradients  below 1500m, and the information on the rate and geometry of the overturning circulation is still preserved.  78  Our 2-D model largely corroborates the results from the 1-D model of Thomas et al. (2006) and indicates that the sediment  231  Pa/230Th signal is rapidly attenuated in sediment  deposited below the base of the overturning cell. Finally, low sediment  231  Pa/230Th in the  South Atlantic (Scholten et al., 2008) appears to be due to the formation of AABW, which suggest that the  231  Pa/230Th sedimentary record in this region, just north of the zone  influenced by biogenic silica, could be used to constrain past changes in the rate of formation of this water mass (Negre et al., 2010).  79  Chapter 3 Reconstruction of the strength and geometry of the Glacial Atlantic Meridional Overturning Circulation using sediment 231Pa/230Th  3.1 Introduction  The Atlantic Meridional Overturning Circulation (AMOC) has a major influence on Earth’s climate due to its role in the large scale redistribution of heat, CO 2 and nutrients (Rahmstorf, 2002; Sigman et al., 2010). The Cd/Ca and δ13C records in benthic foraminifera indicate that the nutrient depleted North Atlantic Deep Water (NADW) was displaced upwards during the Last Glacial Maximum (LGM) by nutrient rich waters from the Southern Ocean. This shallower glacial NADW is called the Glacial North Atlantic Intermediate Water (GNAIW). These results are now well established (Boyle and Keigwin, 1987; Curry and Oppo, 2005; Marchitto and Broecker, 2006), but while they document changes in the geometry of the overturning circulation, they do not constrain its rate, and there is no consensus on the strength of the AMOC during the LGM (Kitoh et al., 2001; Lynch-Stieglitz et al., 2007; Burke et al., 2011; Hesse et al., 2011). 80  The ratio  231  Paex,0/230Thex,0 (activity ratio of unsupported  231  Pa and  230  Th in sediments  decay-corrected to the time of deposition; 231Pa/230Th hereafter), could potentially provide estimates of past rates of AMOC (Yu et al, 1996; Marchal et al., 2000). Pa-321 and Th-230 are produced in seawater from the radioactive decay of dissolved uranium and rapidly removed to the underlying sediment by adsorption on settling particles. Because  231  Pa is  removed at a slower rate than 230Th, a larger fraction of the 231Pa produced in the Atlantic is exported to the Southern ocean by the AMOC. The average 231Pa/230Th of modern Atlantic sediments is thus lower than the production rate ratio of the two isotopes (0.092) and varies with the NADW formation rate (chapter 2). Applying this approach to several Atlantic cores revealed variations in sediment  231  Pa/230Th during the last deglaciation, which was  interpreted as reflecting large changes in the rate and geometry of the AMOC (McManus et al., 2004; Gherardi et al., 2005; 2009). Sediment 231Pa/230Th, however, is also affected by the preferential scavenging of  231  Pa by  biogenic silica, which could lead to erroneous interpretation of changes in circulation (Keigwin and Boyle, 2008; Lippold et al., 2009). The role of biogenic opal in controlling 231  Pa/230Th is evident in the Southern Ocean where settling particles have high opal  concentrations, and periods of higher  231  Pa/230Th at several locations in the Atlantic have  been attributed to increased abundance of diatoms or opal (Hall et al., 2006; Keigwin et al., 2007; Bradtmiller et al., 2007; Lippold et al., 2009; Anderson et al., 2009). Recent interpretation of 231Pa/230Th in terms of paleocirculation considered this possible overprint by ascertaining that changes in  231  Pa/230Th did not coincide with changes in the sediment  content of biogenic silica (Gherardi et al., 2005; 2009; Gihou et al., 2010). However, this 81  argument is weakened by the fact that opal preserved in sediment is only a fraction of the settling particle flux that originally scavenged  231  Pa from the water column. Dissolution  during early diagenesis could remove opal while leaving the scavenged  231  Pa in the  sediment.  Clearly, both the effect of circulation and opal flux must be taken into account to interpret sediment  231  Pa/230Th, and we must identify the conditions under which one becomes  dominant to interpret changes in this ratio. I developed a two-dimensional scavenging model to derive the sediment  231  Pa/230Th distribution generated by simplified meridional  overturning cells (chapter 2). The most striking trend generated by the model is a vertical gradient in sediment 231Pa/230Th, a trend which has now been documented in the sediments deposited in the western equatorial Atlantic (Lippold et al., 2010). This vertical trend is generated by the southward flow of NADW and the longer residence time of respect to reversible scavenging compared to  231  Pa with  230  Th (Francois et al., 2007; Gherardi et al.,  2010; chapter 2). Building on these results, we have assembled a large 231Pa/230Th and opal sediment database (Appendix B) to establish the extent to which sediment 231Pa/230Th in the Atlantic reflects the vertical trend predicted by the 2D scavenging model, to identify areas where enhanced scavenging obscures the circulation signal, and to reconstruct the strength and geometry of the AMOC during the LGM.  82  3.2 Materials and methods  3.2.1 Core locations and chronology  Sediment  231  Pa/230Th from 51 Holocene and 47 last glacial maximum core sections are  reported in Appendix B, along with core locations, water depths, sampling intervals, age 231  Pa/230Th data are averaged over 0-7 ka and LGM data over  information (Holocene 19-24 ka.), % opal,  230  Th-normalized opal fluxes, analytical errors and references. The  cores are broadly distributed within the Atlantic (Fig. 3.1) to establish whether the basin-scale distribution patterns in sediment  231  Pa/230Th expected from the impact of  variations in the AMOC (chapter 2) can be broadly recognized in Atlantic sediments.  83  Figure 3.1: Core locations for Holocene and LGM 231Pa/230Th compilation identifiable by core number (first column) in Appendix B. Colour code indicates data available for the two time periods (black), the Holocene only (red), or the LGM only (blue). Open squares indicate cores influenced by opal (preserved opal flux > 0.2 g/cm2 ka) and open triangles indicate cores affected by boundary scavenging.  3.2.2 Analytical Methods  In addition to compiling literature data, additional cores were analyzed (Appendix B) using the following methods. 84  3.2.2.1 231Pa/230Th data:  My colleague Joerg Lippold and Jeanne Gherardi measured the  231  Pa and  230  Th on all the  new samples reported in Tables a, b, Appendix B. Their method is also included in Appendix B.  3.2.2.2 Biogenic silica:  Biogenic silica was measured following the method described by Mortlock and Froelich, (1989) and Müller and Schneider, (1993). Samples were weighed (in the 20 mg range) in centrifuge tubes. Hydrogen peroxide (H2O2) and HCl were added to samples in order to remove organic and inorganic carbon, respectively. After removing the dissolved phase by centrifuging the samples, opal was extracted from the samples by adding Na2CO3 and heating to 85ºC in a water bath. Dissolved silica was measured colorimetrically at 812 nm on a LKB spectrophotometer. This method has a precision of about 10%.  3.2.3 Modeling  To establish the extent to which the distribution of sediment explained by lateral  231  Pa/230Th could be  231  Pa transport with the AMOC, the data (Appendix B) are compared  with the output of two-dimensional scavenging models (chapter 2). The Holocene data were compared to the model output with a circulation scheme based on modern circulation as described in chapter 2. The circulation in the Holocene Atlantic presented here (Fig. 3.2) has been slightly modified from chapter 2 to add a representation of the 85  Antarctic Intermediate Water (AAIW). The scavenging parameters (Table 3.1) were established by fitting model output to dissolved  230  Th and  231  Pa profiles measured in the  Atlantic (see chapter 2)  Atlantic Flux and Velocity Vector  -5  15  10  10  20  5 15  5  5 10 15 20  -5  750  10  5  0  250  20  20  1250  10 5  5  15  0  10  -5  10  -5  10  5  3250  5  3750 4250  5  15  15  -5  2750  20  15 10  0  2250  20  20  1750 Depth  15  -5  0  5  -10  0  -15  0  4750  60S 50S 40S 30S 20S 10S 0 10N 20N 30N 40N 50N 60N Latitude  -20  Figure 3.2: Holocene overturning scheme used for the model-data comparison shown in Fig. 3.8,  NADW: 20.5 Sv, AABW: 8 Sv, AAIW: 10 Sv  86  Atlantic Flux and Velocity Vector  -1 5 -5  0 -1 0  5  25  25  15 25  10  20  15  15  -5  5  10  5  5  2750  0  0  Depth  20 10  15  1750 2250  25  20 10  5  25 10  -10  1250  -5  -15-5  750  5 15 20  10  20 15  5 15 20  10  10  5  -15 -10 5 20 15  0  250  -5  3250 0  3750  -10  0  0  0  4250 4750  -15  0 60S 50S 40S 30S 20S 10S  -20  0 10N 20N 30N 40N 50N 60N Latitude  Figure 3.3: LGM overturning scheme used for the model-data comparison shown in Fig. 3.8,  GNAIW: 25 Sv, AABW: 4 Sv  Table 3.1: List of abbreviations and values for the Holocene and LGM scavenging parameters (chapter 2). Higher K1Pa south of 47.5°S and between 55°and 60°N (Holocene only) represent the higher opal concentrations of particles settling in the Southern Ocean and in the Northern North Atlantic. The change in the position of the Southern opal belt during the LGM is represented by a northward shift of the southern region with higher K1Pa. Lower K1Th at 50°S-55°S for the LGM accounts for the high percentage of opal in the settling material of this region (Asmus et al., 1999)  Variables  Symbol  Holocene  231  PPa  0.00246  same  dpm/(m3∙yr)  230  PTh  0.0267  same  dpm/(m3∙yr)  Particle sinking rate  S  500  same  m/yr  Pa production rate Th production rate  230  Th adsorption rate (70°N-50°S)  0-250 m  K1  Th  1.0  LGM  same  Units  1/yr 87  Variables 230  Th adsorption rate (50°S-55°S)  230  Th adsorption rate (55°S-70°S)  230  Th adsorption rate 250-500 m 0-250 m 230 Th adsorption rate > 500 m 0-250 m 230 Th desorption rate (70°N-70°S) 231  Pa adsorption rate (70°N-60°N)  231  Pa adsorption rate (60°N-55°N) all depths 0-250 m 231 Pa adsorption rate (55°N-42.5°S) 231  Pa adsorption rate (42.5°S-45°S) 0-250 m 231 Pa adsorption rate (45°S-47.5°S) 231  Pa adsorption rate (47.5°S-50°S) 0-250 m 231 Pa adsorption rate (50°S-70°S) 0-250 m 231 Pa adsorption rate 250-500 m 0-250 m 231 Pa adsorption rate > 500 m 231  Pa desorption rate (70°N-70°S)  Symbol  Holocene  LGM  Units  K1  Th  1.0  0.6  1/yr  K1  Th  0.6  same  1/yr  K1  Th  75% of 0-250  same  1/yr  K1  Th  50% of 0-250 m value 1.6 m value 0.08  same  1/yr  same  1/yr  same  1/yr  K-1  Th  K1  Pa  K1  Pa  0.16  0.08  1/yr  K1  Pa  0.08  same  1/yr  K1  Pa  0.08  0.2  1/yr  K1  Pa  0.08  0.44  1/yr  K1  Pa  0.2  0.44  1/yr  K1  Pa  0.44  same  1/yr  K1  Pa  75% of 0-250  same  1/yr  K1  Pa  50% of 0-250 m value 1 m value  same  1/yr  same  1/yr  K-1  Pa  all depths  The overturning circulation scheme for the LGM (Fig. 3.3) was obtained by systematically changing its strength and geometry to optimize the fit with the LGM sediment data (Appendix B). The optimal LGM geometry consists of a shallower northern overturning cell reaching down to 3500 m depth in the North Atlantic and gradually shoaling to 2000 m in the South Atlantic. The LGM 2-D scavenging model also accounts for shifts in the position of the southern opal belt by 5 degrees to the north (Gersonde et al., 2003) and northern opal region was removed on account of sea ice cover (Table 3.1, Table 3.2). Fractionation factors were kept constant over the rest of the Atlantic Ocean (52.5°N – 40°S; Table 3.2). The position of the zone of deep water formation in the North Atlantic was also shifted to the south by 10 degrees (Labeyrie et al., 1992; Sarnthein et al., 2003) 88  Table 3.2: Latitudinal variations of the equilibrium Fractionation Factors (FF) used in the model. FF is the fractionation factor that would be measured if particles were in equilibrium with surrounding seawater and is calculated from the  231  Pa and  230  Th adsorption and desorption rate constant: FF =  (231Pa/230Th)diss / (231Pa/230Th)part. Differences between Holocene and LGM account for shifts in the position of high biogenic opal flux regions. For the LGM, the southern opal belt was shifted ~5°to the north (Asmus et al., 1999; Sarnthein et al., 2003) and the northern opal region was removed to account for the lower preserved opal fluxes due to sea ice cover. Fractionation factors were kept constant over the rest of the Atlantic Ocean (a reasonable assumption considering that the mean concentrations of opal in the Holocene and LGM sediment reported in Appendix B are similar: Holocene mean: 2.6 %, n=37, 1 SD=1.6 %; LGM mean=3.7 %, n=32, 1 SD=1.9 %). The position of the zone of deep water formation was also shifted to the south by 10 degrees during the LGM (Labeyrie et al., 1992).  Latitude 70°N – 60°N 60°N – 52.5°N 52.5°N – 40°S 40°S – 42.5°S 42.5°S – 45°S 45°S – 47.5°S 47.5°S – 50°S 50°S – 55°S 50°S – 70°S  FF Holocene 7.8 3.9 7.8 7.8 7.8 7.8 3.1 1.4 0.9  FF LGM same 7.8 same 5.2 3.1 1.4 1.4 0.9 same  One of the alternate geometries tested against the sediment database is shown in Fig. 3.4, with a northern overturning cell restricted to the upper 2000 m of the water column. This geometry was recently suggested based on a model – data comparison for benthic 13C (Hesse et al., 2011). 89  Atlantic Flux and Velocity Vector  15  5 15  2020  15  5 15  2020  15  1250  10  20  15  10  15  10  10  5  5  -15 0 10  750  10  10  5  5 10 -5-15 -10 -1 0 -5  250  5  5  Depth  1750 2250  -15 -1 0  -5  0  0  0 0  -5  2750 -5  0  3250 3750  -10 4250  0  0  4750  0  -15  60S 50S 40S 30S 20S 10S 0 10N 20N 30N 40N 50N 60N Latitude  Figure 3.4: Overturning scheme used to test the very shallow (within the upper 2000 m of the water column) Glacial North Atlantic Intermediate Water (GNAIW) deduced by comparing data vs model results for benthic foraminifera 13C (Hesse et al., 2011).  3.3 Results and Discussion  3.3.1 The influence of biogenic silica on the distribution of sediment  231  Pa/230Th in  Atlantic sediments When plotting 231Pa/230Th versus biogenic silica flux for Holocene and LGM core sections, a correlation is found only when including sites from the southern ocean and North Atlantic with biogenic opal flux exceeding 0.2 g/cm 2·ka (Fig. 3.5). Sedimentary 231Pa/230Th is also higher compared to other cores from similar water depths at the African margins (Scholten et al., 2008; Christl et al., 2009; Lippold et al., 2012) even though opal fluxes are relatively 90  small, reflecting the effect of boundary scavenging (Anderson et al., 1983). However, the lack of correlation between opal flux and 231Pa/230Th for most of the Atlantic sites indicates that variations in scavenging intensity in most of the Atlantic Ocean, where opal fluxes are < 0.2 g/cm2·ka, has a negligible impact on the distribution of sediment 231Pa/230Th.  Figure 3.5: Correlation between sediment  231  Pa/230Th and  230  Th-normalized opal flux is found during  the Holocene (red) and LGM (blue) when including cores with high opal flux (black regression lines). The correlation disappears when only considering the main Atlantic basin (red, blue regression lines), where opal fluxes are lower than 0.2 g/cm2 ka. Error bars indicate 2 SE for both opal flux and 231Pa/230Th. Significance of correlation is estimated by the linear correlation coefficient r and the p-value. Breaks in axes are inserted to display the Southern Ocean core. 91  A similar result is obtained when plotting sediment  231  Pa/230Th vs % opal (Fig. 3.6),  indicating that opal starts to significantly influence sediment  231  Pa/230Th only when its  concentration exceeds 8%. Since most Atlantic sediments have biogenic silica concentration below this threshold, the large scale distribution of sediment  231  Pa/230Th in  most of the Atlantic must be controlled by another factor.  92  Figure 3.6: Correlation between sediment  231  Pa/230Th and opal concentration is found during the  Holocene (a) and LGM (b) when including cores with opal concentration > 8% (black regression lines). The correlation disappears when only considering the main Atlantic basin (red, blue), where opal concentrations are lower. Error bars indicate 2 SE for both 231Pa/230Th and opal flux.  3.3.2 The influence of AMOC on the distribution of sediment  231  Pa/230Th in Atlantic  sediments  3.3.2.1 Holocene: The  231  Pa/230Th ratios measured in most of the Holocene sediment  samples clearly decrease with water depth down to 4000 m (Fig. 3.7a), consistent with the prediction of the 2D scavenging model (chapter 2). Below ~ 4000 m, the trend disappears or reverses, reflecting the influence of Antarctic Bottom Water (AABW). 93  Figure 3.7: 231Pa/230Th versus water depth. (a) Holocene sediment 231Pa/230Th generally decreases with water depth (red circles). (b) LGM 231Pa/230Th in the main Atlantic basin (blue squares) decreases with depth down to 2500 m and then increase in deeper waters. Values deviating from these general trends are from regions with high opal flux (open symbols) or from the African margin (black triangles). Note the break in the horizontal axis. Individual measurements have been averaged for both time periods (Appendix B). Error bars in 231Pa/230Th indicate 2 SE.  94  There are, however, several samples that deviate from this general trend and display higher ratios: samples immediately south of Greenland/Iceland, from the African margin, and the only sample from the southern ocean opal belt included in this study. At these sites, higher scavenging intensity by opal or particle flux increases 231Pa/230Th above the value expected from the overturning circulation alone.  Figure 3.8: Correlation between Holocene (a) and LGM (b) sediment  231  Pa/230Th and model outputs  from grid cells closest to core locations.  To further test the importance of the AMOC in controlling sediment  231  Pa/230Th in the  Atlantic and eventually to assess the strength of the glacial AMOC, we compare the Holocene  231  Pa/230Th database to the output of the Holocene 2-D scavenging model (Fig.  3.8). In this model, we use 20.5 Sv NADW and 8 Sv AABW (Talley et al., 2003), and adsorption/desorption rate constants vary with latitude to reflect the presence of opal in the North Atlantic and the Southern Ocean (Table 3.1; 3.2). The model generates a decreasing 95  trend in sediment Pa/Th with depth (chapter 2), as observed in the data (Fig. 3.7a), and we find a good correlation between sediment data and model output generated at the closest locations with respect to latitude and depth (Fig. 3.8a). This result supports the interpretation that the distribution of 231Pa/230Th in Atlantic sediments deposited in regions with low opal flux can be largely attributed to the AMOC. The offset between data and model output is attributed to boundary scavenging, which is not included in the model. Boundary scavenging removes to the margins some of the 231Pa produced in the water of the NADW as it transits through the Atlantic, accounting for the systematically lower 231  Pa/230Th in the data compared to model output.  3.3.2.2 LGM: In contrast to the Holocene, LGM sediment 231Pa/230Th decreases with depth to 2500 m only and starts increasing below (Fig. 3.7b). As for the Holocene, sites south of Iceland and Greenland diverge for the LGM from the general trend but less so than during the Holocene, reflecting lower opal fluxes. Plotting  231  Pa/230Th versus biogenic silica flux  in glacial sediments (Fig. 3.5) reveals a weak but significant correlation with the inclusion of the Southern Ocean site. Three high  231  Pa/230Th cores south of Iceland have high opal  fluxes, suggesting an opal influence. However, two cores from the NW Atlantic with similar opal fluxes have low 231Pa/230Th, casting doubt that opal is the prevailing cause for these higher values. High 231Pa/230Th in the cores south of Iceland may also indicate a zone of deep water formation, where sediment  231  Pa/230Th ratios are expected to be higher  (chapter 2). Two cores taken in the upwelling region off Namibia also clearly deviate from the general trend due to boundary scavenging. As for the Holocene, there is no significant  96  correlation between preserved opal flux and sediment 231Pa/230Th where the flux of opal is below 0.2 g/cm2·ka or where % opal is less than 8% (Fig. 3.6b). The change in the depth gradients in sediment 231Pa/230Th during the LGM (Fig. 3.7b vs Fig. 3.7a) is generally consistent with a shallower glacial AMOC driven by Glacial North Atlantic Intermediate Water (GNAIW). We tested several glacial AMOC geometries in the 2D model and found that the best fit with the data is obtained when the core of the glacial northern sourced AMOC is at 2000 m. However, to generate the comparatively low 231  Pa/230Th at North Atlantic sites between 3000 and 3600 m depth (Fig. 3.7b) and produce  a good fit between data and model output (Fig 3.8b), the base of the overturning cell must have reached down to this depth before gradually shoaling to the south (Fig. 3.3). Paleonutrient proxies suggest that GNAIW extended down to ~2500-3000 m depth only (Curry and Oppo, 2005; Marchitto and Broecker, 2006), and recently Hesse et al. (2011) inferred an even shallower glacial overturning by comparing the output of a 3-D circulation model and a δ13C database (Fig. 3.4). If our interpretation is correct, the low benthic δ13C of these cores must reflect high preformed nutrients in the deepest and densest waters involved in the glacial northern AMOC.  The fit between sediment  231  Pa/230Th generated by the optimal 2D glacial overturning  geometry (Fig. 3.3) and the sediment database is shown in Fig. 3.8b. As for the Holocene, the offset from the 1:1 line is attributed to boundary scavenging, which is not represented in the model. This offset is less pronounced for the LGM, suggesting a weaker boundary scavenging effect at the African margin (Lippold et al., 2012). 97  3.3.3 Sensitivity test on the LGM 2-D scavenging model  The sensitivity of the LGM model output to the scavenging parameters (sinking rate, fractionation factor), and to the strength and geometry of the overturning cells (GNAIW and AABW) was tested by comparing data vs model output, as shown in Fig. 3.8, under different configurations, and reporting the fit for each run either as the linear correlation coefficient (r) or the mean square weighed deviation (mswd, Powell et al., 2002) of the linear regression between data and model output.  3.3.3.1 Sensitivity to sinking rate and fractionation factor: Using the optimal geometry of the glacial AMOC (Fig.3.3), we systematically varied the sinking rate of particles and the fractionation factor in the main Atlantic basin, and compared sediment data with model outputs as done for Figure 3.8. Plotting the linear correlation coefficient (r; Fig. 3.9) and the mean square weighed deviation (mswd; Fig. 3.10) obtained when regressing the output for each model run against the sediment database indicates that the best fit is obtained with a sinking rate of 500m/y and a fractionation factor of 7.8. These parameters were thus used to assess the effect of overturning strength and geometry.  98  sensitivity test of model scavenging parameters 1.00  0.80  r  0.60 FF=3.4  0.40 FF=5.1 FF=7.8  0.20 FF=10.5 FF=15.6  0.00 0  100  200  300  400 500 600 particle sinking rate [m/year]  700  800  900  Figure 3.9: Variations in the linear correlation (r) obtained between the sediment  1000  231  Pa/230Th database  (Appendix B) and model output using the overturning geometry and strength shown in Fig. 3.3 and systematically varying the sinking rate of particles and fractionation factor (FF).  sensitivity test of model scavenging parameters particle sinking rate [m/year] 0  100  200  300  400  500  600  700  800  900  1000  0  50  mswd  100 FF=3.4  150 FF=5.1 FF=7.8  200 FF=10.5 FF=15.6  250  Figure 3.10: Variations in the mean square weighed deviation (mswd) between the sediment 231  Pa/230Th database (Appendix B) and model output using the overturning geometry and strength  shown in Fig. 3.3 and systematically varying the sinking rate of particles and fractionation factor (FF). 99  3.3.3.2 Sensitivity to the strength of the overturning circulation cells (GNAIW; AABW)  Using the optimal geometry of the glacial AMOC (Fig. 3.3) and the scavenging parameters determined above, we systematically varied the strength of the northern overturning cell, initiated by the formation of GNAIW, and the southern overturning cell, initiated by the formation of AABW, and quantified the fit between sediment data and model outputs using the mean square weighed deviation (mswd) of the regression.  Figure 3.11: Fit between observations and model outputs generated with the optimal LGM model geometry with varying GNAIW and AABW strengths. Agreement between measurements and model is 100  quantified using mean square weighted deviation (mswd). Smaller deviations are obtained with GNAIW > 20 Sv. The fit is less sensitive to variations in AABW. The model output with the Holocene circulation scheme yields a weaker agreement to the LGM observations (black square) further indicating that a stronger and shallower AMOC is necessary to explain the LGM measurements.  The mswd between data and model increases sharply as soon as the rate of GNAIW drops below 22 Sv and is optimal for GNAIW rates ranging between 22 and 30 Sv (Fig. 3.11). The model-data fit provides a sharp lower limit for the rate of GNAIW (20-22 Sv), but the upper limit is less precisely constrained. This is because sediment  231  Pa/230Th is more  sensitive to changes in the rate of AMOC when it is weaker (Yu et al., 1996). When GNAIW drops below 22 Sv, the model is unable to generate the low sediment  231  Pa/230Th  observed in the glacial sediments deposited in the equatorial region.  Considering that the cross section of the GNAIW overturning cell was smaller than today, these results imply a stronger mean flow of northern-sourced deep water during the LGM. In contrast, limits on the strength of the AABW could not be clearly established. Although a weak AABW overturning cell produces a slightly better agreement with observations (Fig. 3.11), the differences are very small. This may be because the available database is skewed to the northern and equatorial region of the Atlantic (Fig. 3.1) where the distribution of sediment  231  Pa/230Th is dominated by the strength (and geometry) of GNAIW. Increasing  the sediment database in the South Atlantic may improve our estimates of the strength of the AABW during the LGM.  101  3.3.3.3 Sensitivity to the geometry of the overturning circulation cells (GNAIW; AABW)  The distribution of  231  Pa/230Th in glacial Atlantic sediments confirms that the glacial  AMOC was shallower but suggests a deeper overturning in the North Atlantic than inferred from nutrient proxies (Hesse et al., 2011).  Figure 3.12: Fit between observations and model outputs generated with a very narrow GNAIW (Fig. 3.4) as a function of GNAIW and AABW strengths. Agreement between measurements and model is quantified using mean square weighted deviation (mswd), which is invariably large compared to the optimal LGM circulation scheme. 102  To further test whether the glacial sediment  231  Pa/230Th database is compatible with the  consistently shallow overturning cell inferred from 13C, we calculated the msdw of the correlation between sediment data and model output using the geometry shown in Fig. 3.4, in which the penetration of GNAIW is restricted to the upper 2000m over the entire Atlantic and the core of the overturning cell is at 750 m. The results (Fig. 3.12) indicate a very poor fit, primarily because the low  231  Pa/230Th values observed below 2000m in the glacial  sediment database cannot be reproduced using this geometry. This confirms that a relatively deep overturning in the North Atlantic is needed to explain the observations and the discrepancy with 13C is best explained by invoking changes in preformed nutrients in the densest water of the GNAIW, maybe because of extensive ice cover in the zone where these denser waters formed. If this interpretation is correct, it could have important implications for our understanding of the marine carbon cycle and the estimation of CO 2 uptake during the LGM (Kwon et al., 2012).  The fit between observational data and model output is also shown on the sections reported in Fig. 3.13 showing sediment 231Pa/230Th sections generated by the three model geometries (Fig. 3.2 - Holocene circulation; Fig. 3.3 - optimal LGM circulation; and Fig. 3.4 - very shallow GNAIW) and superimposing on them the sediment database (plotted at a position defined by their depth and latitude) using the same colour scale.  Fig. 3.13a clearly shows that the very shallow overturning cell derived from the distribution of 13C alone is unable to generate the low 231Pa/230Th in the sediment of the North Atlantic observed below 2000m, even if the GNAIW is increased to 30 Sv. On the other hand, Fig. 103  3.13b shows that the optimal LGM circulation scheme (Fig. 3.3) is able to produce these low values, as already indicated by the fit in Fig.3.8b. This figure also shows that if GNAIW is reduced, the model cannot generate the low 231Pa/230Th observed in the equatorial region. Finally, Fig. 3.13c shows that while the Holocene model (Fig. 3.2) reproduces well the Holocene database, it is not able to generate the distribution observed in the glacial core sections. In particular the Holocene model cannot reproduce the relatively high sediment 231  Pa/230Th found in the deep North Atlantic during the LGM. This misfit is also shown on  Fig. 3.11, where the mswd of the correlation between the LGM sediment database and the Holocene model input show a relatively poor fit compared to the optimal LGM model output.  3.3.4 Can the Holocene circulation scheme explain the LGM observations by varying the scavenging parameters?  Thus far, we have shown that when using the scavenging parameters that best reproduce the LGM data with the optimal LGM circulation, the Holocene circulation scheme generates a distribution of sediment  231  Pa/230Th that fits the LGM data base poorly (Fig. 3.11). The  question remains, however, whether the Holocene circulation scheme could generate a 231  Pa/230Th distribution similar to LGM observations with a different set of scavenging  parameters.  104  Figure 3.13: Sediment  231  Pa/230Th superimposed to the sediment  231  Pa/230Th section generated by the  2-D scavenging model. 105  a FF=3.4  0.80  FF=5.1  linear correlation coefficient r  FF=7.8 FF=10.5  0.60  FF=15.6 LGM FF=7.8  0.40  0.20  0.00 0  100  200  300  400  500  600  700  800  900  1000  900  1000  particle sinking rate [m/year]  -0.20  b particle sinking rate [m/year] 0  100  200  300  400  500  600  700  800  5 25 45 65  mswd  85 105 125  FF=3.4  145  FF=5.1  165  FF=7.8 FF=10.5  185  FF=15.6 LGM FF=7.8  Figure 3.14: Linear correlation coefficient (a) and mean square weighed deviations (b) obtained when correlating the sediment database with sediment  231  Pa/230Th generated by the Holocene circulation  scheme with varying particle sinking rates and fractionation factors. 106  To test this possibility, we determined the linear correlation coefficient (r) and the mswd of the correlation obtained between the LGM dataset and the model output generated by the Holocene circulation scheme with varying sinking rates and fractionation factors. The results (Fig. 3.14) indicate a much poorer fit over a wide range of sinking rates and FF than can be obtained with the optimal LGM circulation scheme, reinforcing the conclusion that the LGM database reflects a change a circulation as illustrated in Fig. 3.3.  3.4 Conclusion  The results from this study indicate that while the concentration of biogenic silica in particles can affect sediment  231  Pa/230Th, its influence becomes significant only when the  preserved opal flux exceeds 0.2g/cm2ka or when the opal concentration in sediment exceeds 8%. Below these threshold values, the distribution of sediment 231Pa/230Th in the Atlantic is controlled by the differential transport of the two isotopes by the AMOC.  Since the opal content of most Atlantic sediments is below the threshold at which it controls 231  Pa/230Th, this study confirms that past changes in Atlantic sediment  231  Pa/230Th can be  interpreted as changes in the strength and geometry of the AMOC, if the cores analysed do not include those from the African coastal upwelling regions and contain less than 8% opal.  The main conclusions derived by comparing the extended database reported in Appendix B with output from our 2-D scavenging models are as follows. 107  (1) The GNAIW formation rate during the LGM was > 20Sv compared to 20Sv for the modern NADW. Because of its smaller cross section, the flow of water must have been faster and the glacial AMOC more vigorous during the LGM. However, while sediment 231  Pa/230Th provides a sharp lower limit for the rate of GNAIW (20-22 Sv), the upper limit  is less precisely constrained (~25-30 Sv).  (2) The glacial AMOC was shallower than during the Holocene, but the GNAIW sunk to greater depth in the North Atlantic than suggested by 13C, which implies higher preformed nutrients in densest (deepest) waters of the GNAIW.  The existing database is, however, unable to constrain changes in the rate of AABW. This is in part because of the relatively poor data coverage from the deep South Atlantic, which could be remedied by analysing additional cores from this region.  108  Chapter 4 The influence of deep water circulation on the distribution of 231Pa and 230Th in the water column and sediments of the Pacific Ocean  4.1 Introduction The ocean’s deep water circulation plays a prominent role in climate regulation. The overturning circulation of the Atlantic is a major contributor to heat transport to northern latitudes, while the relative isolation of deep water in the Pacific results in greater sequestration of carbon, contributing to lowering atmospheric CO 2. It is therefore important to document past changes in ocean circulation in both oceans to understand climate evolution, especially on glacial-interglacial time scales and during abrupt climate changes. Past changes in ocean circulation have been inferred from nutrient proxies (e.g., Boyle and Keigwin, 1987), but poor carbonate preservation severely limits their application in the Pacific Ocean. In addition, although these tracers provide crucial information on changes in water mass distribution, they cannot constrain changes in the rate of deep water circulation (Legrand and Wunsch, 1995). In response to the latter 109  problem, attempts are being made to develop kinematic tracers of past ocean circulation (Lynch-Stieglitz et al., 2007).  In particular, sedimentary  231  Pa/230Th (activity ratios of  231  Pa and  230  Th not supported by  U decay in the sediment mineral lattice and decay-corrected to the time of deposition, Pa/Th hereafter) has been used to evaluate past changes in the rate of the Atlantic Meridional Overturning circulation (AMOC) (Yu et al., 1996; Marchal et al., 2000, Mcmanus et al., 2004; Hall et al., 2006; Gherardi et al., 2005, 2009; Negre et al., 2010; chapter 3). This approach stems from the observation that Atlantic sediments have, on average, Pa/Th ratios lower than the fixed production rate ratio of 0.092 (Yu et al., 1996). Pa-231 and Th-230 are produced at constant rates in seawater from the decay of dissolved 235  U and  234  U, respectively. Pa-231 is less particle-reactive than Th-230 and has a longer  residence time in the water column before removal to the sediment by scavenging. As a result, the AMOC exports  231  Pa into the southern ocean more effectively, which produces  a 231Pa deficit in Atlantic sediments controlled by the rate of overturning (Yu et al., 1996; Marchal et al., 2000; Siddall et al., 2007). This simple interpretation, however, can be obscured by changes in particle flux and opal content, which affect the residence time of 231  Pa in the water column and overprint sediment Pa/Th independently of circulation  (Walter et al., 1997; Chase et al., 2002, 2003; Gil et al., 2009; Guo et al., 2002; Lippold et al., 2009).  To further address this question, we recently developed a 2D scavenging model to investigate the influence of the overturning circulation on the distribution of Pa/Th in 110  Atlantic sediments (chapter 2). The model was tuned using vertical profiles of dissolved 231  Pa and  230  Th measured in the Atlantic and it predicts variations in Pa/Th that are  recognized in the sediment, in particular, a decrease with depth which cannot be readily explained by invoking particle or opal flux as the main controlling factors (Gherardi et al., 2005; 2009; Lippold et al., 2011; chapter 3). Here, we use a similar approach to argue that deep water circulation must also be considered when interpreting the sedimentary Pa/Th record in the Pacific Ocean.  In the Pacific, sediment Pa/Th tends to be low in the central basins and higher near the continental margins (Yang et al., 1985; Walter et al., 1999). This distribution of sediment Pa/Th is commonly attributed to ”boundary scavenging”, a process that enhances the scavenging of the less particle-reactive  231  Pa in zones of higher particle flux (Yang et al.,  1985; Bacon, 1988; Taguchi et al., 1989; Lao et al., 1992; Yu et al., 2001; Roy-Barman, 2009), higher opal flux (Pichat et al., 2004; Bradtmiller et al., 2006), or higher opal and MnO2 concentration in settling particles (Anderson et al., 1983; Walter et al., 1999; Chase et al., 2002). These areas often coincide with continental margins. In contrast to the Atlantic, the residence time of deep water in the Pacific (~ 600 years) is much longer than the lateral diffusive mixing time in ocean basins (~ 100 years; Anderson et al., 1990), allowing a full expression of boundary scavenging (Yu et al., 2001). The Pacific Meridional Overturning Circulation (PMOC) has, thus far, not been considered as a significant factor affecting the large scale distribution of Pa/Th in Pacific sediments, in sharp contrast to the on-going debate concerning the influence of the AMOC in Atlantic sediments (e.g., Peacock et al., 2010; Gherardi et al., 2010). Building on chapter 2, we 111  have developed a Pacific 2D scavenging model with a schematic overturning circulation based on observations and tuned to produce water column profiles of dissolved 231  230  Th and  Pa similar to those measured in the Pacific Ocean. The results show that the PMOC, in  addition to boundary scavenging, has a significant impact on the distribution of Pa/Th in Pacific sediments.  4.2 Model descriptions  4.2.1 Overturning circulation.  The 2-D Pacific Meridional Overturning Circulation (PMOC) scheme (Figure 4.1) connects the meridional section in the Atlantic Ocean (constant 5000m depth from 70°N to 70°S, evenly divided into 20 layers and 56 columns; chapter 2) to a meridional section of the Pacific Ocean (constant 5000m depth and 4000km width from 55°N to 70°S evenly divided into 20 layers and 50 columns).  112  Flux and Velocity Vector 5  0  250  -10 -10  -20  -10  -15 1250  0  0  -5  -10  0  -2 0  -20  -5  -20 -15  -20  -5  -1 5 -25 -20 -10 -5 -15 5 0  750  -5  -5  -15  -15  5  -5 -20  0  -1 0  -25  -1 0  1750  -20  -15  -15  -2 5  -10  3250  -5 -15  -5  -20  5  0  2750  -10 -10 -5  -1 5  -15  -2 5  -10  -10 -5  Depth  2250  5  -5 -20 0  3750  -20  0  5  -15 -1 0  4750 N.A. 60N  -10  -5  0  0  4250  -5 50N  40N  30N  20N  10N  0  10S  20S  30S  40S  50S  60S S.O. 60S Latitude  50S  -25 40S  30S  20S  10S  0  ATLANTIC  10N  20N  30N  40N N.P.  PACIFIC  Figure 4.1: Velocity vector plot for the Atlantic (Antarctic Intermediate Water (AAIW): 10 Sv; North Atlantic Deep Water (NADW): 20.5 Sv; Antarctic Bottom Water (AABW): 8 Sv) and the Pacific (Antarctic Intermediate Water (AAIW): 3 Sv; North Pacific Intermediate Water (NPIW): 4 Sv; Lower Circumpolar Deep Water (LCDW)/Antarctic Bottom Water (AABW): 26 Sv).  The circulation in the Atlantic section is reported in chapter 3. The Atlantic and the Pacific sections are connected by a single mixing box from 0 to 1000m at 57.5°S-70°S in the Atlantic and 62.5°S-70°S in the Pacific. The geometry and transport rates of the different Atlantic water masses are based on published field observations (Friedrichs and Hall, 1993; Macdonald, 1998; Talley, 2003; Ganachaud and Wunsch, 2000) and discussed in chapter 2. Pacific Meridional Overturning Circulation (PMOC) is initiated by 26Sv of surface and intermediate water sinking in the southernmost column (67.5°S – 70°S) and 113  flowing northward below 3000m to represent the Lower Circumpolar Deep Water (LCDW) and Antarctic Bottom Water (AABW) (Sloyan and Rintoul, 2001). This Southern Component Water mass (SCW) gradually upwells in the north Pacific to support the southward flow of North Pacific Deep Water (NPDW) between 1000m and 3000m. Most of the SCW (25Sv) upwells south of 45°N (Macdonald, 1998) while the remaining 1Sv is set to upwell further north. North Pacific Deep Water (NPDW) begins to upwell at 45°S toward the surface and subsurface of the southern ocean. At shallower depths, the North Pacific Intermediate Water is represented by a 4Sv meridional overturning cell above 1000m north of the equator (Kawabe and Fujio, 2010). In the south Pacific, the upper 1000m are occupied by a 3Sv overturning cell initiated by the formation of AAIW at 40°S (Talley, 2003).  4.2.2 Formulation of the two-dimensional scavenging model in the Pacific Ocean The scavenging model used in our study (Fig. 4.2) is based on the principle of reversible scavenging (Bacon and Anderson, 1982; Nozaki et al., 1987) and builds on the Atlantic model in chapter 2. In contrast to the Atlantic, however, where boundary scavenging can be neglected to a first approximation, boundary scavenging is much more pronounced in the Pacific Ocean and must be taken into account. Since this removal process cannot be explicitly represented in the 2-D meridional overturning scheme, we have added a removal term in each Pacific grid between 40°S and 35°N to reflect the lateral removal of 230  Th and  231  Pa from the central gyres to the margins. Dissolved  230  Th and  231  Pa  concentrations in the Pacific 2-D model ([Xd] in dpm.m-3) are thus dictated by their 114  production rates from uranium decay (PX; dpm.m−3.y−1), the adsorption (KX1) and desorption (KX−1) rate constants (y−1), the transport rates imposed by the circulation scheme (V; m.y−1), and the removal rate to the margins by boundary scavenging, which we assume, as a first approximation, to be proportional to the concentration of dissolved nuclide (R[Xd]). Particulate 230Th and  231  Pa concentrations ([Xp]) are controlled by the  adsorption/desorption rate constants, transport rates and the sinking rate (S; m.y−1) of the particles that scavenge the two nuclides from the water column. At steady-state, we can write: Px − K1x [Xd] + K−1x [Xp] + VΔ[X]d - R [Xd] = 0  (4.1)  K1x [Xd] − K−1x [Xp] + VΔ[Xp] + dFlux/dZ = 0  (4.2)  dFlux = S( [Xp](i+1) − [Xp](i) )  (4.3)  where X represents  230  Th or  231  Pa, Z is water depth (m), i is the vertical index, Δ is an  “upwind” difference divided by the grid spacing (Press et al., 1992), and R (y -1) is the effective removal rate constant to the margins, taken to be invariant between 40°S and 35°N. Using the upwind scheme with a horizontal velocity u=4×10 −3 m/s and a horizontal grid spacing Δx = 278×103 m, the inherent mixing in our model (Kdiff) is ~600 m2s-1 (Kdiff = u x/2); based on equivalence of the upwind scheme applied to an advective-reactive equation and an analytic diffusive-advective-reactive equation (e.g., Press et al., 1992)]. This is in the range of the along-isopycnal tracer diffusivities reported for the southern ocean (100~800m2s−1; Zika et al., 2009). Unlike lateral transport to the margins, the meridional lateral transport by turbulent mixing is therefore implicit in the model. 115  Figure 4.2: Scavenging and transport model in each model grid box. Xd = the concentration of dissolved  230  Th or 231Pa (dpm/m3). Xp = the concentration of particulate  230  Th or  231  Pa (dpm/m3). Px =  production rate of 230Th or 231Pa (dpm/m3/y). K1/K-1 = adsorption/desorption rate of 230Th or 231Pa (1/y). S = sinking rate of scavenging particles (m/y). V = transport rate by circulation (m/y). R is the effective removal rate constant to the margins.  4.2.3 Parameterization of the scavenging model.  The parameters for the scavenging model in the Pacific have been selected following chapter 2, but the adsorption rate constants for  231  Pa and 230Th had to be lowered in order  to generate dissolved 231Pa and 230Th profiles similar to water column profiles measured in the North Pacific (Table 4.1). This is consistent with the general view that particle scavenging is, on average, more pronounced in the Atlantic compared to the Pacific because of its smaller volume and stronger aeolian input. The Southern Ocean’s opal belt is represented by higher K1Pa south of 45°S (Table 1b).  116  Table 4.1a: List of abbreviations and values for the Atlantic model parameters. Variables  Symbo  Control run  Units  231  P l Pa  0.00246  dpm/(m3*y  230  PTh  0.0267  3 dpm/(m *y r)  Particle sinking rate  S  500  m/yr r)  230  K1Th  1  1/yr  230  K1Th  0.6  1/yr  230  K1Th  75% of 0-250 m 1/yr  230  K1Th  50% value of 0-250 m 1/yr  230  K-1Th  1.6 value  231  K1Pa  0.08  231  Pa adsorption rate (60°N-55°N & 45°S -47.5°S) 0-250m  K1Pa  0.2  1/yr  231  Pa adsorption rate (47.5°S-50°S) 0-250m 0-250m  K1Pa  0.32  1/yr 1/yr  231  K1Pa  0.44  1/yr  231  K1Pa  0.36  1/yr  231  K1Pa  0.28  1/yr  231  K1Pa  75% of 0-250 m 1/yr  231  K1Pa  50% value of 0-250 m 1/yr  231  K-1Pa  1 value  Pa production rate Th production rate  Th adsorption rate (70°N-55°S) 0-250m Th adsorption rate (55°S-70°S) 0-250m Th adsorption rate 250-500m Th adsorption rate > 500m Th desorption rate (70°N-70°S) All depths Pa adsorption rate (70°N-60°N & 55°N -45°S)  Pa adsorption rate (50°S-57.5°S) 0-250m Pa adsorption rate (57.5°S-60°S) 0-250m Pa adsorption rate (60°S-70°S) 0-250m Pa adsorption rate 250-500m Pa adsorption rate > 500m Pa desorption rate (55°N-70°S) All depths  1/yr  1/yr  Table 4.1b: List of abbreviations and values for the Pacific model parameters. Variables  Symbo  Control run  Units  231  P l Pa  0.00246  dpm/(m3*yr  230  PTh  0.0267  3 dpm/(m *yr )  Particle sinking rate (55°N-35°N)  S  750  m/yr )  Particle sinking rate (35°N-70°S)  S  500  m/yr  Pa production rate Th production rate  117  Variables  Symbo  Control run  Units  R l  0.0025 (AVG)  1/yr  230  ThS-55° adsorption rate (55°N-70°S) 0-250m (40° N)  K1Th  0.6  1/yr  230  K1Th  75% of 0-250 m 1/yr  230  K1Th  50% value of 0-250 m 1/yr  230  K-1Th  1.6 value  1/yr  231  K1Pa  0.2  1/yr  231  K1Pa  0.16  1/yr  231  K1Pa  0.048  1/yr  231  K1Pa  0.2  1/yr  231  K1Pa  0.32  1/yr  231  K1Pa  0.44  1/yr  231  K1Pa  0.36  1/yr  231  K1Pa  0.28  1/yr  231  K1Pa  75% of 0-250 m 1/yr  231  K1Pa  50% value of 0-250 m 1/yr  231  K-1Pa  1 value  Boundary  scavenging  removal  rate  Th adsorption rate (55°N-70°S) 250-500m Th adsorption rate (55°N-70°S) > 500m Th desorption rate (55°N-70°S) All depths Pa adsorption rate (55°N-50°N) 0-250m Pa adsorption rate (50°N-35°N) 0-250m Pa adsorption rate (35°N-45°S) 0-250m Pa adsorption rate (45°S-47.5°S) 0-250m Pa adsorption rate (47.5°S-50°S) 0-250m Pa adsorption rate (50°S-57.5°S) 0-250m Pa adsorption rate (57.5°S-60°S) 0-250m Pa adsorption rate (60°S-70°S) 0-250m Pa adsorption rate 250-500m Pa adsorption rate > 500m Pa desorption rate (55°N-70°S) All depths  constant  1/yr  Similarly, K1Pa is higher between 35°N and 55°N to reflect higher opal fluxes in the North Pacific. The equilibrium fractionation factors F = (K-1Pa K1Th)/(K1Pa K-1Th) generated by these rate constants range from 7.8 between 45°S and 35°N to 2 north of 50°N to a minimum of 0.9 between 50°S and 57.5°S (Table 4.2). The fractionation factor in the Southern Ocean opal belt reaches a minimum between 50°S and 57.5°S in both Atlantic and Pacific sector, to reflect the latitude of the region of highest opal flux. There is also an opal region in the North Atlantic (60°N-52.5°N) (Chapter 3). 118  Table 4.2: “Equilibrium” Fractionation Factors for the Pacific Model. Latitude  “Equilibrium Fractionation Factor”  55°N – 50°N  1.88  50°N – 35°N  2.34  35°N – 45°S  7.8  45°S – 47.5°S  2  47.5°S – 50°S  1.24  50°S – 57.5°S  0.9  57.5°S – 60°S  1.1  60°S – 70°S  1.65  4.2.4 Estimating removal by “boundary scavenging” in the 2-D scavenging model This section is the contribution from my PhD committee member Susan Allen and it is therefore reported in Appendix C.  4.3 Results and discussion 4.3.1 Dissolved 230Th and 231Pa water column profiles  4.3.1.1. Data-model comparison  We use published (Nozaki et al., 1987; Chase et al., 2003) and new (Table 4.3) 230  231  Pa and  Th seawater profiles (Fig. 4.3) to constrain the circulation and scavenging parameters in 119  our 2-D model. The dissolved  231  Pa and 230Th profiles reported in Table 3 were measured  by isotope dilution ICP-MS as described by Choi et al. (2000).  In the absence of circulation, the reversible scavenging model predicts linear increases in dissolved and particulate  231  Pa and  230  Th concentrations with depth if the scavenging  parameters (K1, K-1, S) remain constant (Bacon and Anderson, 1982; Nozaki et al., 1987): [X]p = [PX/S]*Z  (4.4)  [X]d = [PX/K1] + [(K−1PX)/(K1S)]*Z  (4.5)  We use these linear profiles as reference (“no-circulation” profiles) in Figure 4.4-4.5 to highlight the influence of circulation on the  230  Th and 231Pa profiles. Profiles deviate from  linearity when influenced by circulation and relax gradually back to linearity with an e-folding time equivalent to the residence time of the nuclide with respect to scavenging (e.g., Francois, 2007). When nuclide concentrations in seawater exceed the value predicted by the “no-circulation” profile, they tend to gradually decrease as water masses age (because the rate of adsorption on particles exceeds the rate of desorption). On the other hand, when nuclide concentrations are lower than predicted by the “no-circulation” profiles, they tend to gradually increase as water masses age.  120  Figure 4.3: Station locations for the water column profiles used to constrain the parameters in the model: North Pacific: CE-8, CE-13 (Nozaki et al. 1987); ALOHA (Table 4.3); South Pacific: SAZ (Table 4.3); and AESOPS stations MS-1 to MS-5 (Chase et al., 2002)  Table 4.3: 230Th and 231Pa activities in sea water (dpm/1000kg). Station PaPa (50°N, 145°W) depth  Total 230Th  Total 231Pa  m  dpm/1000kg (±95% CI)  200  0.092  ± 0.005  -0.002  ± -0.005  400  0.134  ± 0.008  0.005  ± 0.004  700  0.253  ± 0.014  0.026  ± 0.006 121  depth  Total 230Th  m  dpm/1000kg (±95% CI)  1000  Total 231Pa 0.302 ± 0.019  1300 1600 1900  Total 231Pa  Total 231Pa 0.056 ± 0.006  m 0.336 ± 0.021 0.088 depth depth 0.448 ± 0.025 0.098 dpm/1000kg (±95% CI) 0.543 ± 0.03 0.128  ± 0.011 ± 0.010 ± 0.012  2200  0.543 m  ± 0.03  0.132  ± 0.013  2500  0.550  ± 0.033  0.167  ± 0.020  2800  0.638  ± 0.035  0.153  ± 0.014  3100  0.720  ± 0.038  0.184  ± 0.015  3400  0.719  ± 0.039  0.192  ± 0.015  3700  0.799  ± 0.038  0.209  ± 0.018  4000  0.907  ± 0.042  0.209  ± 0.020  4315  1.165  ± 0.064  0.217  ± 0.016  Station Aloha (22°45’N, 158°W) depth  Total 230Th  Total 231Pa  m  dpm/1000kg (±95% CI)  25  0.04949  ± 0.00222 0.011  ± 0.015  100  0.06284  ± 0.00252 0.045  ± 0.014  150  0.05663  ± 0.00766 0.025  ± 0.015  200  0.08648  ± 0.00438 0.038  ± 0.015  250  0.10694  ± 0.00436 0.028  ± 0.014  300  0.11576  ± 0.00362 0.041  ± 0.014  400  0.15316  ± 0.00349 0.065  ± 0.015  500  0.17743  ± 0.00426 0.096  ± 0.015  122  depth  Total 230Th  m  dpm/1000kg (±95% CI)  650  Total 231Pa Total 231Pa 0.26129 ± 0.00521 0.133 ± 0.017  800 1150 1500  Total 231Pa  m 0.32618 ± 0.00804 0.236 ± 0.019 depth depth 0.47534 ± 0.00839 0.401 ± 0.023 dpm/1000kg (±95% CI) 0.59073 ± 0.00858 0.509 ± 0.022  1800  0.77509 m  ± 0.01  0.549  ± 0.019  2100  0.86417  ± 0.01218 0.581  ± 0.022  0.581  ± 0.023  2450 2800  1.09756  ± 0.01817 0.568  ± 0.023  3200  1.37107  ± 0.01983 0.568  ± 0.023  3600  1.58197  ± 0.02386 0.508  ± 0.022  4000  1.64832  ± 0.01838 0.466  ± 0.022  4410  1.66059  ± 0.02484 0.427  ± 0.019  4610  1.69769  ± 0.02938  4710  1.69085  ± 0.02694 0.430  ± 0.023  4760  1.63658  ± 0.02988 0.395  ± 0.021  4790  1.7175  ± 0.032  ± 0.020  0.392  Station SAZ2002-54S (54°S, 142°E) depth Total 230Th  Total 231Pa  m  dpm/1000kg (±95% CI)  30  0.096  ± 0.003  0.090  ± 0.012  100  0.094  ± 0.004  0.087  ± 0.011  200  0.230  ± 0.007  0.127  ± 0.009  400  0.328  ± 0.011  0.205  ± 0.017  123  depth Total 230Th  Total 231Pa  m  dpm/1000kg (±95% CI)  599  Total 231Pa 0.350 ± 0.010  800 1102 1400  Total 231Pa 0.213 ± 0.011  m 0.435 ± 0.009 0.246 ± 0.008 depth depth 0.515 ± 0.013 0.298 ± 0.022 dpm/1000kg (±95% CI) 0.594 ± 0.010 0.308 ± 0.009  1702  0.641 m  ± 0.010  0.330  ± 0.015  2101  0.733  ± 0.025  2601  0.792  ± 0.013  0.363  ± 0.014  3183  0.861  ± 0.022  0.384  ± 0.010  Station SAZ2002-61S (61°S, 142°E) depth Total 230Th  Total 231Pa  dpm/1000kg (±95% CI) m 21  0.271  ± 0.027  152  0.149  ± 0.012  401  0.405  ± 0.012  0.220  ± 0.012  801  0.566  ± 0.010  0.273  ± 0.011  1200  0.712  ± 0.022  0.315  ± 0.013  1600  0.774  ± 0.020  0.330  ± 0.012  2002  0.826  ± 0.021  0.330  ± 0.011  2399  0.917  ± 0.019  0.363  ± 0.012  2802  1.007  ± 0.029  0.365  ± 0.013  3200  1.091  ± 0.016  0.372  ± 0.010  3599  1.174  ± 0.021  0.397  ± 0.018  4003  1.194  ± 0.032  0.447  ± 0.015  124  Station SAZ2002-64S (64°S, 142°E) depth Total 230Th  Total 231Pa  dpm/1000kg (±95% CI) m 30  0.183  ± 0.006  0.089  ± 0.006  152  0.143  ± 0.007  0.186  ± 0.007  401  0.599  ± 0.013  0.245  ± 0.013  599  0.686  ± 0.012  0.276  ± 0.012  801  0.737  ± 0.012  0.299  ± 0.012  1200  0.756  ± 0.012  0.323  ± 0.012  1601  0.912  ± 0.013  0.362  ± 0.013  2002  0.995  ± 0.017  0.350  ± 0.017  2401  1.079  ± 0.017  0.370  ± 0.017  2803  1.110  ± 0.012  0.378  ± 0.012  3201  1.107  ± 0.016  0.386  ± 0.016  3549  1.023  ± 0.012  0.367  ± 0.012  Station SAZ2002-66S (66°S, 142°E) depth Total 230Th m 50  Total 231Pa  dpm/1000kg (±95% CI) 0.459  ± 0.01  0.180  ± 0.011  200  0.407  ± 0.010  0.158  ± 0.011  400  0.422  ± 0.010  0.157  ± 0.020  600  0.434  ± 0.011  0.167  ± 0.015  749  0.486  ± 0.012  0.168  ± 0.011  125  The  230  Th and  231  Pa profiles produced by the Pacific model between 54°S and 66°S  (Figure 4.4c, f) exhibit convex shapes similar to those measured in Pacific sector of the southern ocean (Fig. 4.4a, b, d, e). The measured and modeled concentrations in this region are higher than expected for the “no-circulation” profiles, particularly for  231  Pa.  These high concentrations can be explained by upwelling of Upper Circumpolar Deep Water (UCDW), which brings deep waters with high nuclide concentrations to shallower depths (Rutgers van der Loeff and Berger, 1993; Chase et al., 2003; Francois, 2007), and by the lower affinity of marine particles for 231Pa north of the opal belt (Walter et al., 1997; Chase et al., 2002), allowing  231  Pa to build-up in the NPDW before it returns to the  Southern Ocean. The formation of AABW from shallow waters also influences the  230  Th  and 231Pa profiles by depressing concentrations in bottom waters.  Profiles measured in the western tropical Pacific (Fig. 4.5) and at station ALOHA (Fig. 4.6) also show curvatures that are reproduced in the model and can be explained by deep water upwelling in the North Pacific. Southern Component Water (SCW) spreads northwards and shoals to intermediate depths in the north Pacific to form NPDW which flows to the south between 1000 and 3000 m depth. The maximum dissolved  231  Pa  concentration at intermediate depth is due to upwelling, which brings deep seawater with relatively high  231  Pa to intermediate depth and decreases the sinking rate of small  scavenging particles.  126  230Th  0.2  0.4  0.6  0.8  1  1.2  0  1.4 0  500  500  1000  1000  1500  1500  2000  2000  Depth (m)  0  2500 SAZ2002 66S,142E 3000  0.4  (dpm/T)  0.6  0.8  1  1.2  1.4  MS-1 53S,17.5W MS-2 57S,170W MS-3 60S,170W  3500  SAZ2002 61S,142E  MS-4 63S,170W  SAZ2002 54S,142E  4000  0.2  230Th  2500 3000  SAZ2002 64S,142E  3500  4000  MS-5 66S,169W  Standard  Standard  4500  4500  230Th  0  0.2  0.4  (dpm/T)  0.6  0.8  1  1.2  1.4  0 500 1000 1500  Depth (m)  Depth (m)  0  (dpm/T)  2000 2500  Model 54S  Model 59S  Model 64S  Model 66S  3000 3500 4000  Standard  4500  127  0  0.1  0.2  231Pa  (dpm/T)  0.3  0.4  231Pa  0.5  0.6  0.7  0  0  0.1  0.2  (dpm/T)  0.3  0.4  0.5  0.6  0.7  0 Standard K1=0.024  Standard K1=0.024  Standard K1=0.14  Standard K1=0.14  500  500  Standard K1=0.22  Standard K1=0.22  MS-1 53S,17.5W SAZ2002 54S,142E  1000  1000  MS-2 57S,170W  SAZ2002 61S,142E  MS-3 60S,170W 1500  SAZ2002 64S,142E SAZ2002 66S,142E  2000  Depth (m)  Depth (m)  1500  MS-5 66S,169W 2000  2500  2500  3000  3000  3500  3500  4000  4000  4500  4500  231Pa  0 0 500 1000  0.1  0.2  MS-4 63S,170W  (dpm/T)  0.3  0.4  0.5  0.6  0.7  Standard K1=0.024 Standard K1=0.14 Standard K1=0.22 Model 49S Model 54S  1500  Model 59S  Depth (m)  Model 66S 2000 2500  3000 3500 4000  4500  Figure 4.4: Total 231Pa and 230Th profiles measured in the Southern Ocean (a) & (d) at 142°E longitude from 54°S to 66°S during the 2002 Sub-Antarctic Zone program (SAZ) (Table 4.2), (b) & (e) during the AESOPS program (Chase et al., 2003), and (c) & (f) predicted by our 2-D scavenging model between 54°S and 66°S. The linear profiles are concentrations predicted by the reversible scavenging 128  model in absence of circulation with K1 = 0.3 y-1 for 230Th, and with 0.14 < K1 < 0.22 (southern ocean) and K1 = 0.024 y-1 (47.5°N – 45°S) for 231Pa.  Figure 4.5b and 4.6b indicate that 231Pa concentrations are lower than the “no circulation” values below ~2500 m. We thus expect  231  Pa concentration to increase as deep water is  upwelled to intermediate depths. The 230Th profile at Station ALOHA also shows a convex shape because of the upwelling, but the curvature is not as pronounced as for 231Pa due to the greater particle reactivity and shorter residence time of Th in the water column, allowing it to relax towards linearity more quickly.  230Th 0  0.4  0.8  231Pa  (dpm/T) 1.2  1.6  0  2  0.2  0.4  (dpm/T)  0.6  0.8  1  1.2  0  0  CE-8 12.5N,173E 1000  1000  CE-13 12N,152E Model 14N  2000  2000  Depth (m)  Depth (m)  Standard  3000  CE-8 12.45N,173E 4000  3000  4000  CE-13 12N,152E  5000  Model 14N  5000  Standard 6000  Figure 4.5: Dissolved  6000  230  Th (a) and  231  Pa (b) at station CE-8 (12.45°N, 173°E) and CE-13 (13.12°N,  152°E; Nozaki et al., 1987) compared to model results at 14°N.  129  230Th 0.4  0.8  1.2  1.6  0  2  0  0  500  500  1000  1000  1500  1500  2000  2000  2500  3000  3500  Depth (m)  Depth (m)  0  231Pa  (dpm/T)  Aloha Feb/2002  2500  3000  Model 21N 3500  Model 31N 4000  4000  Standard 4500  4500  Standard *  0.2  0.4  (dpm/T) 0.6  0.8  1  Aloha Feb/2002 Model 21N Model 31N Standard Standard *  5000  5000  Figure 4.6: Dissolved  230  Th (a) and  231  Pa (b) at station Aloha compared to model results at 21°N and  37.5°N.  4.3.1.2. The influence of boundary scavenging on the curvature of the  231  Pa seawater  profiles in the North Pacific  The dissolved  231  Pa seawater profiles generated by the 2D scavenging model suggest that  PMOC is the primary factor producing the curvature often reported in the dissolved  231  Pa  seawater profiles in the North Pacific. To test whether boundary scavenging could also influence the shape of these profiles, the scavenging model was run with PMOC at a fixed rate of 26Sv while varying the effective removal rate constant to the margins (R) by a constant multiplier (Fig. 4.7). The results indicate that the curvature generated by the  130  PMOC is more pronounced when R is smaller indicating that boundary scavenging tends to reduce the curvature generated by circulation.  Dissolved 231Pa (dpm/T) 0  0.2  0.4  0.6  0.8  1  1.2  1.4  1.6  0 Aloha Feb/2002 Standard Standard*  1000  Model 21N (No R) Model 21N (1/2*R) Model 21N (CTR run)  Depth (m)  2000  Model 21N (2*R)  3000  4000  5000  Figure 4.7: Changes in the curvature of the dissolved  231  Pa profile generated at 21°N with 26Sv  PMOC and varying effective rate constant for removal to the margins.  4.3.2 231Pa and 230Th sections generated by the Pacific 2-D scavenging model  The sections generated by the Atlantic section of the 2-D scavenging model have been discussed in details in previous chapters. Here, we are discussing the influence of the overturning circulation in the Pacific section.  131  4.3.2.1 Dissolved 230Th The dissolved 230Th generated by our 2D scavenging model (Fig. 4.8) gradually increases in the SCW as this water mass spreads northward, reaching a maximum in the zone of deep water upwelling in the North Pacific.  Pacific dissolved 230Th 0  1.8  0.4  0.2  0.2  500  0.2  1.6  0.4 0.4  1000  0.6  0.8  0.6  1.4  1500  0.4  0.8  0.6 0.6  1.2  2000  1.2  0.8  1  0.8  1  Depth  1  1  0.8 2500  1.4  1  1. 2  0.8  1.4  1 3000  0. 6  6 1.  0.8  0.6  1.2  1.2  3500  1.2  0.4  4000  55s  45s  35s  25s  15s  5s  5N  0.2  1.6  1.2  0.8  1.4  65s  1  1.4 1.6  1  4500  15N  25N  35N  45N  Latitude  Figure 4.8: Dissolved 230Th section generated by the model.  Comparing the dissolved 230  230  Th concentrations generated by the model to the dissolved  Th concentrations expected in absence of circulation (Fig. 4.9) indicates that, while  SCW starts at concentrations lower than the “no-circulation” values in the southern ocean, its dissolved  230  Th concentrations quickly reach the “no-circulation” value and then  increase slowly to the north. The reason for this unexpected result (i.e., that dissolved 230  Th concentration continues to increase as SCW moves northward even though it  already exceeds the concentration at steady state with respect to scavenging) is the deep water upwelling in the North Pacific, which generates NPDW with a large dissolved  230  Th  132  excess, which is transmitted to deeper waters by scavenging. The NPDW entrains high 230  230  Th water southward between 1000m and 3000m, but the high  Th dissolved  concentration is quickly attenuated because of the thorium particle reactivity . The AAIW and NPIW above 1000m in the model have limited influence on the dissolved  230  Th  distribution due to the short residence time of 230Th in shallow waters.  Dissolved Th, circu-no circu 0  0.5  0.1  0 0.4  0.3  0  0  0  500  0.2  0.1  0.4  0.2  0.1  0  0.1  -0.3  0.5  0.3  -0 . 1 -0 .2  4500  0.4  0.4  3 -0 . -0 .4  0.2  0  0 65s  55s  45s  35s  25s  15s  5s  5N  15N  25N  35N  -0.1  -0.2  0.2  0.1  3500  4000  0  1 0.  0.20.3  3000  0.1  0.2  -0.1-0.2  0.5  0  0.5  0.3  2500  0.4  0.3  0  Depth  2000  0.2  0.1  0.1  0  0.3 0.2  0.2 0.3  0.1 1500  0.3  0  1000  45N  -0.4  -0.5  Latitude  Figure 4.9: Difference between the dissolved  230  Th concentration generated by the 2-D model and the  concentration predicted in the absence of circulation. Positive values indicate that dissolved  230  Th  concentration in the presence of circulation exceeds the steady state value with respect to scavenging.  4.3.2.2 Particulate 230Th  The section of particulate 230  230  Th concentration (Fig. 4.10) is similar to that of dissolved  Th. The most conspicuous feature is the maximum in the deep North Pacific.  133  Pacific particulate 230Th 0  0.3  0.05 500  0.0 5  0.05  0.25  1000  0.1  0.05 1500  0.1  0.1  0.2  0.15 5 0.1  0.3  0.2 3500  0.1  5 0.1  0.25  0.2  4000  0.05  5 0.2  65s  55s  45s  35s  25s  15s  5s  5N  15N  25N  0.25  0.3  0.15  4500  0.15  0.2  0.25  0. 1  0.2  3000  0.15  0.2  2500  0.15  Depth  5 0.1  2000  35N  45N  Latitude  Figure 4.10: Particulate 230Th section generated by the model.  Particulate 230Th/Total 230Th 0  0.12  0.13  0.12  0.21  0.11  0.13  0.2  0.12  0.1 7  3 0.1  0.12  14 0.  1000  19 08. 0.1  4 0.113 0.  0.14 0.13  0.12  0.12  0.22 0.2 0.2 1 0.15 0.16  0.12  500  0.14 0.13  0.19  0.13  1500  0.14  0.18  4 0.1  0.15  Depth  0.13  0.14 2000  0.17  0.16  0.14  2500  0.15  0.1 5  3000  0.15  0.17 0.16  0.14  0.14 3500  0.1 5  0.13  4000  16 0.  0.12  5 0.1  4500  0.11 65s  55s  45s  35s  25s  15s Latitude  5s  5N  15N  25N  35N  45N  Figure 4.11: Fraction of total 230Th associated with particles generated by the model.  The fraction of total  230  Th associated with particles ([230Th]p/[230Th]t) ranges from 0.12 to  0.15 over most of the Pacific (Fig. 4.11), which is lower than generated by the model in the Atlantic (0.18-0.22; chapter 2) but still conforms to field observations (Bacon and Anderson, 1982; Moran et al., 2002). Higher fractions are generated in the upper (<1000m) 134  water column of the Southern Ocean (>0.22), reflecting the convection and the resulting longer residence time of particles in the water column. 4.3.2.3 Dissolved 231Pa  Compared to dissolved  230  Th, the distribution of dissolved  231  Pa concentration is more  strongly influenced by PMOC, because of its longer residence time in the water column. Pa-231 concentrations gradually increase along the path of the SCW, reaching a maximum at mid-depth in the North Pacific where deep waters upwell (Fig. 4.12). This is consistent with the observation that dissolved  231  Pa concentrations in the deep Pacific remain well  below the concentrations predicted in the absence of circulation below 2500m depth (Figure 4.5b, 4.6b, 4.13). As the SCW spreads northward and leaves the opal dominated region of the southern ocean, the “no-circulation” concentrations for 231Pa increase sharply, as K1Pa decreases north of the opal belt (Table 4.1b). Because the “no-circulation” concentrations exceed the 231Pa concentrations in deep water, the latter increase northward. Above 3000m, the initially high dissolved south as dissolved contrast to  231  Pa concentration of NPDW decreases to the  231  Pa is slowly scavenged from the southward flowing water mass. In  230  Th, however, mid-depth waters with relatively high  231  Pa concentration  propagate far into the South Pacific.  135  Pacific dissolved 231Pa 0  0. 1  0.1  0.1  0.55  0.2  500  0.2  0.5  0.2  0.3  0.3  0.2  1000  0.45  0.4  0.1  0.4  0.3  0.4 1500  0.3  0.4  0.5  0.3  0.35  0.5  0.3  0. 6  0.2  2500  0.5  Depth  2000  0.25  0.5  3000  0.2  0.4  0.4  0.3  0.3  0.6  3500  0.15  0.5  0.4  0.5  0.3  4000  0.1  4500 65s  55s  45s  35s  25s  0.05  0.2  0.4  15s  5s  5N  15N  25N  35N  45N  Latitude  Figure 4.12: Dissolved 231Pa section generated by the model.  Dissolved Pa, circu-no circu 0.5 0  0  0 0.1  0  0  0.4  0  -0.1  500  0  2 0.  0.1  0.1 -0.1 0  0.2  0.1 0.3  0.2  2500 0.1  -0.1  -0 .1  0.1  0  55s  45s  0.3  -0 .2  0.1 0 -0.1  65s  -0 .3  -0.4 -0.5 -0.6 -0.7  4500  -0.3  0.2  0.1  -0.1 0 -0.2 -0.5  0 -0 .1 -0 .2  -0.4  -0 .3  35s  25s  15s  5s  5N  15N  25N  35N  -0.2  0.2  0.1 -0.4  -0.2  2 0.  3500  0.2  .3 -0  3000  4000  0 0.1  0.1  0  0.1  0.2  2000 Depth  0.2  0.1 0  0.2  0.2  1500  0.3  0.1  1000  45N  -0.5  Latitude  Figure 4.13: Difference between the dissolved 231Pa concentration generated by the 2-D model and the concentration predicted in the absence of circulation. Negative values indicate that dissolved concentrations are lower than the steady state (non-circulating)  231  231  Pa  Pa concentration with respect to  scavenging, allowing 231Pa to increase along the path of the water masses.  136  4.3.2.4 Particulate 231Pa  Particulate  231  Pa concentrations (Fig. 4.14) are mostly controlled by the presence of opal,  which has a high affinity for Pa (Walter et al., 1997; Chase et al., 2002). Accordingly, concentrations are highest in the southern ocean and in the North Pacific, resulting from the higher K1Pa in these regions.  Pacific particulate 231Pa 0  0.07  0.01  500  2 0.0  0.06  1000  0.03  0.01 0.01  0.04  0.02  0.05  0.02  0.06  0.01  01 0.  0.05 0.07  0.02  3500  0.06  0.03 0.02  0.02  0.01  0.03 0.04  0.04  0.03  4000  0.03  0.03  3000  0.05 0.06  2500  0.0 4  Depth  0.03 0.04  0.03  2000  0.05  0.04  1500  0.02  0.01  4500 65s  55s  45s  35s  25s  15s  5s  5N  15N  25N  35N  45N  Latitude  Figure 4.14: Particulate 231Pa section generated by the model.  The fraction of particulate  231  Pa generated by the model between 55°S and 45°N ranges  between 0.02 and 0.04 (Fig. 4.15) with higher values in the North Pacific (0.06-0.08) and the Southern Ocean (0.10–0.18). The latter are consistent with the values from the Atlantic sector of the southern ocean reported by Rutgers van der Loeff and Berger (1993).  137  Particulate 231Pa/Total 231Pa  0.06 0.04  0.04 0.06  0.12  2500  0.02  0.02  0.1  0.04  0.14 0.16  0.04 0.06 0.08 0.1 0.12  6 0.0  0.08  0.1 8  0.12  0.06  0.04 0.06 0.08 0.1 0.12  4000  65s  55s  0.04  45s  35s  25s  15s  5s  5N  15N  25N  0.08  0.18 0.16 0.14  4500  0.06  0.12  0.14 0.16  3000  0.04 0.06  Depth  0.14  0.02  0.12  0.12  0.14  0.1  2000  3500  0.16  0.02  0.06  14 0.  1000  1500  0.12  0.08  500  0.02  0.02  12 0.  0.1  0.18  0.08  0  35N  0.02  45N  Latitude  Figure 4.15: Fraction of total 231Pa associated with particles generated by the model.  4.3.2.5 Dissolved Pa/Th  Between the northern and southern opal belts, modeled dissolved Pa/Th ratios decrease with water depth below 1000m (Fig. 4.16).  Pacific dissolved 231Pa/230Th 0  1.4  1.2  1  10.8 0.6  0.2  500  0.6  0.4  0.6  1  4 0.  0.8 0.6  0.8  1.2  0.4  0.6  1000  1500  1  0.4 0.4  2500  0.8  0.6  0.4  Depth  0.2  0.6  0.6  2000  3000  0.6  3500  0.4  0.2  0.4  0.4  4000  4500  0.2 65s  55s  45s  35s  25s  15s Latitude  5s  5N  15N  25N  35N  45N  Figure 4.16: Dissolved 231Pa/230Th section generated by the model. 138  While this decrease within the intermediate waters is found over the entire Pacific, it is more pronounced south of the equator (Fig. 4.17a). In the opal belts of the southern ocean and the North Pacific, the Pa/Th ratios are lower (Fig. 4.16) because of the higher Pa adsorption rates and the ratios are nearly uniform with depth (Fig. 4.17c). Available field data show similar vertical trends (Fig. 4.17b, d). In the southern Pacific, maximum dissolved Pa/Th ratios occur at the surface just north of the opal belt and a sharply decrease with depth, while within the opal belt, dissolved Pa/Th is low and nearly constant with depth (Fig. 4.17d). In the subtropical Pacific, dissolved Pa/Th ratios decrease with depth from a maximum of ~0.8 at 1000 m in both the model (Fig. 4.17a) and the field data (Fig. 4.17b). Model and field data diverge in shallower water and may reflect the larger errors associated with field measurements of very low concentrations or misrepresentation of the shallow overturning cells in the model.  4.3.2.6 Particulate and sedimentary Pa/Th The distribution of particulate Pa/Th produced by the model is shown in Figure 4.18. Between the southern and northern opal belts, the model predicts a decrease with depth below 1500m south of the equator and a northward decrease below 1500m in the North Pacific.  139  Dissolved 231Pa/230Th 0.4  0.6  0.8  1  0.2  1000  1000  1500  1500  2000  2000  2500  2500  Depth (m)  Depth (m)  0.2  Dissolved 231Pa/230Th  3000  3500  0.6  0.8  1  3000  3500  Model 24S  CE-13 12N,152E  Model 9S  4000  0.4  4000  CE-8 12.5N,173E Model 14N 4500  4500  Aloha Feb/2002  Model 21N 5000  5000  Dissolved 231Pa/230Th  Dissolved 231Pa/230Th 0.2  0.4  0.6  0.8  0  1 1000  1500  1500  2000  2000  2500  2500  Depth (m)  Depth (m)  0 1000  3000  3500  Model 66S  3500  4000  Model 56S 4500  Model 53S  4500  Model 44S 5000  5000  Figure 4.17: Dissolved  0.4  0.6  0.8  1  3000  Model 61S 4000  0.2  MS-1 53S,17.5W MS-2 57S,170W MS-3 60S,170W MS-4 63S,170W MS-5 66S,169W SAZ2002 54S,142E SAZ2002 61S,142E SAZ2002 64S,142E SAZ2002 66S,142E  231  Pa/230Th (a) generated by the model between the two opal belts below  1000m; (b) measured in the subtropical North Pacific below 1000m; (c) generated by the model in the 140  Southern opal belt (red) and just to the north (blue), (d) measured in the Pacific section of the southern ocean (MS stations are from Chase et al., 2002; SAZ stations are reported in Table 3) to the north and south of the Polar Front.  To generate sections of sediment Pa/Th in the 2D model, we take into account the fact that sinking particles are not in equilibrium with seawater (chapter 2). Upon reaching the seafloor, particles are not usually buried immediately and reach equilibrium with bottom waters (chapter 2). We can calculate the Pa/Th of sediment deposited at a given latitude and depth and at equilibrium with bottom waters using [X]p/[X]d=KX1/KX−1 for 231  230  Th and  Pa (Fig. 4.19).  Pacific particulate 231Pa/230Th 0.14  0.07  0.08  0.07  0.12  0.1  0.09  0.11  0.06  0. 08  0.05  0.1  0.14  0.13 0.12 0.09  0.13 0.12 0.10.11 0.080.09  0.11  0.12 0.13  0.08  0. 07  2500  0.11 0.1  0.14  2000  0.06  0.07  1500  0.13  0.1 1  0.1 0.08  1000  Depth  1 0.1  0.12  09 0.  0.1  0.12  0.1 4  0.11  0.07 0.08 0.14  0.1  0.13 0.09  500  09 0.  0.12  14 0.  0.13 0.12 0 0.1 0.08 0.0 .1 1 9  0  65s  55s  45s  0.1  5 0.0  0.05  0.14  0.11  0.05 0.04  0.06  0.09  4500  0.07  0.13 0.12  0.05 0.06 0.070.08 0.09 0.110.1 0.12 0.13  4000  06 0.  0.04  0.14  0.06  0.12 0.13  3500  0.11 0.1 0.08 0.070.05 0.06  0.07  3000  0.04 35s  25s  15s Latitude  5s  5N  15N  25N  35N  45N  Figure 4.18: Particulate 231Pa/230Th section generated by the model.  Partial equilibration would result in sediment Pa/Th intermediate between values reported in Figs. 4.18 and 4.19. The distribution pattern of sediment Pa/Th is very similar to that of particulate Pa/Th but with more pronounced vertical gradients. This is because settling 141  particles are farther from chemical equilibrium with ambient seawater at shallower depths (chapter 2), reflecting the trend generated by the model for dissolved Pa/Th, the vertical gradients of sediment Pa/Th are stronger south of the equator (Fig. 4.20).  Pacific sediment 231Pa/230Th  0. 11  0.12  0.06  0.11  0.14  0.06 0.05  45s  35s  25s  15s Latitude  5s  5N  15N  25N  0.06  0.05  0.12  0.14  0.13 0.12 0.1 0.09 0.11 0.07 0.06 0.05 0.04  0.04  0.07  0.13  0.13 0.12 0.1 0.09 0.11  Depth  0.08 55s  0.13  65s  0.05  0.11 0.1 0.08 0.06 0.07 0.04  0.14  4500  0.0 5 0.04  0.08  0.11  0.12 0.05 0.04  0.0 6 0.06  4000  0.12  0.07  3500  0.1  0.09  0.09  3000  0.11  0. 13  0.13 0.11 0.1 0.08 0.06 0.07  0.07  0.12  4 0.1  0.08  0.14  2500  0.12  8 0.0  1500  0.13  0.09  0.09  0.09  2000  0.13  0.05 7 0.0  0.14  0.08  0.07 0.08  0. 1 0.08 1  0.07  0.08  0.09 0.1  1000  0.12  0.09  0.06  0.13 0.12  0.09  0.14  0.1  14 0. 0.1  500  3 0.1 1 0.1  0.12  3 0.1  0.14  11 0. 0 .1  0. 07  0  35N  0.04  45N  Figure 4.19: Sedimentary 231Pa/230Th section generated by the model.  Sediment 231Pa/230Th 0.02 1500  0.03  0.04  0.05  0.06  0.07  0.08  0.09  0.1  2000  Depth (m)  2500 3000  3500 4000  Model 21S Model 1S  4500  Model 1N Model 21N  5000  Figure 4.20: Sediment 231Pa/230Th generated by the 2D model at 21°S and 21°N. 142  4.3.3 Comparison between sediment Pa/Th measured in the Pacific and model predictions  We use a compilation of surface sediment Pa/Th (Fig. 4.21) reported by Lao et al. (1992) and Pichat et al. (2004) (and references therein) and recent measurements published by Bratmiller et al. (2006)  (see Appendix C) to investigate whether the depth dependency  predicted by our 2D model can also be recognized in the Pacific sediment Pa/Th database. In the eastern equatorial Pacific, Pichat et al. (2004) found a linear correlation (r2 = 0.69) between the surface sediment Pa/Th and export production estimated from satellite data. Likewise, Bradtmiller et al. (2006) found a linear correlation (r2 = 0.66) between Holocene sediment Pa/Th and  230  Th-normalized opal flux in cores spanning the entire  equatorial Pacific region. These results indicate that particle flux is an important factor controlling sediment Pa/Th, at least in the equatorial upwelling region. However, Pichat et al. (2004) also noted that variations in sediment Pa/Th could not be readily explained by variations in productivity estimated from ocean color in the Western Equatorial Pacific. They tentatively attributed this observation to underestimations of the export flux from satellite data in this region, but other factors could also be important.  143  Figure 4.21: Distribution of core-top Pa/Th data. Data taken from Lao et al. (1992), Pichat et al. (2004) and Bradtmiller et al. (2006), and references therein. Blue diamonds = cores from low productivity regions at depth < 5000m; white diamonds = cores from low productivity regions at depth > 5000m (i.e., deeper than the 2D scavenging model); Orange squares = cores from the eastern equatorial upwelling region with high Pa/Th; grey squares = cores from the eastern equatorial upwelling region with low Pa/Th; brown squares = cores from the western equatorial Pacific region with high Pa/Th; green triangles = cores from the northeast Pacific margin; red triangle = core from the southeast Pacific margin; yellow triangle = core from the western equatorial Pacific margin; magenta squares = cores affected by biogenic silica;  light blue squares: cores at the boundary between subtropical and  subpolar regions; black circles = cores from low productivity sites where Pa/Th is anomalously high; green diamonds = cores from low productivity sites where Pa/Th is anomalously low (see also Fig. 24; cores affected by hydrothermal activity (Pichat et al., 2004) are not included). 144  The correlation between the entire database and our model output (i.e., sediment Pa/Th generated by the model at the same depth and latitude as the cores) is not very high (Fig. 4.22; r2 = 0.12) and the sediment data are often higher than the model output. This is not surprising since many of the cores analyzed come from areas close enough to the coast to be directly affected by boundary scavenging, which cannot be represented in the 2D scavenging model.  0.2 Low Prod. (>5000m)  Low Prod. Opal 0.15 Opal Boundary  model Pa/Th  N. Pac Margin S. Pac Margin  0.1 y = 1.01x R²= 0.77  Asian Margin EEP EEP - Lo  0.05  WEP Hi Pa/Th  Too Hi - Lo P 0 0  0.05  0.1  0.15  0.2  Too Lo - Lo P  Measured Pa/Th  Figure 4.22: Sediment Pa/Th measured in the Pacific vs. sediment Pa/Th generated by the 2D scavenging model at the same latitude and depth. Symbols are the same as in Fig. 4.21. [Blue diamonds = cores from low productivity regions at depth < 5000m; white diamonds = cores from low productivity regions at depth > 5000m (i.e. deeper than the 2D scavenging model); Orange squares = 145  cores from the eastern equatorial upwelling region; grey squares = cores from the eastern equatorial upwelling region with lower Pa/Th; brown squares = cores from the western equatorial Pacific region with high Pa/Th; green triangles = cores from the northeast Pacific margin; red triangle = core from the southeast Pacific margin; yellow triangle = core from the western equatorial Pacific margin; magenta squares = cores affected by biogenic silica; green squares: cores at the boundary between subtropical and subpolar regions; black circles = cores from low productivity sites where Pa/Th is anomalously high; green diamonds = cores from low productivity sites where Pa/Th is anomalously low].  We note, however, that a group of data points plot close to the 1:1 line (blue and white diamonds in Figs. 4.21 and 4.22; Modeled Pa/Th = 1.01 (Measured Pa/Th); r2 = 0.77), indicating that they are well reproduced by the model. These cores were taken in low productivity regions of the Pacific (Fig. 4.21), and their Pa/Th clearly decreases with depth between 2000m and 5000m (Fig. 4.23), as predicted by the 2D scavenging model. Some of these cores were taken below the maximum depth represented in our 2D model (5000m). Because observed sediment Pa/Th does not generally decrease between 5000 and 6000m, we compared Pa/Th measured in sediment below 5000m to model results obtained at 5000m (white diamonds; Fig. 4.22) but excluded them in the linear regressions (Fig. 4.23; r2 = 0.88). Comparing this regression line with model output on Fig. 4.20 confirms that our 2D model reproduces reasonably well the vertical sediment Pa/Th gradient observed in the data, at least in low-productivity areas with low sedimentary Pa/Th ratios. The shallower low productivity cores (blue diamonds) in Fig. 4.23 come from the western Pacific Ocean. Those data are thus best compared to model predictions near the equator (Fig. 4.20; Model 1S and 1N). 146  Measured sediment Pa/Th 0.00 0  0.05  0.10  0.15  0.20 Low prod. (>5000m) Low prod.  1000 Opal Opal Boundary  2000  Depth (m)  R²= 0.8809  N. Pac Margin  3000 S. Pac. Margin  R²= 0.8015  Asian Margin  4000 EEP  5000  EEP-Lo WEP Hi Pa/Th  6000  Too Hi - Lo P Too Lo - Lo P  7000  Figure 4.23: Pa/Th in core tops versus depth (symbols are identical to Fig. 4.21).  When plotting Pa/Th from the low productivity sites against satellite-derived export production, no significant correlation can be found (r2=0.005), possibly indicating that proximal scavenging has little effect on sediment Pa/Th in low productivity regions In the other regions of the Pacific, sediment Pa/Th is significantly higher than predicted by the low productivity depth trend (Fig. 4.23). Sediment Pa/Th from core tops in the eastern equatorial upwelling region also decreases with depth (r 2 = 0.80), but Pa/Th at a given  147  depth is much higher than in low productivity regions, reflecting the direct effect of boundary scavenging.  y = 1.1673x - 0.0454 R2 = 0.6646  0.2  0.1  230  Th-normalized opal flux  0.3  0 0.05  0.1  0.15  0.2  231  Pa/230Th  Figure 4.24a: 230Th-normalized opal flux against sediment Pa/Th in the equatorial Pacific (Bradtmiller et al., 2006)  0.04  y = -4E-05x + 0.1095  0.03  R2 = 0.8751  Th Pa/  -0.01 -0.02  231  230  0.01    0.02 0  -0.03 -0.04 -0.05 1500  2000  2500  3000  3500  4000  4500  Depth (m)  Figure 4.24b: Difference between Pa/Th measured and estimated from  230  Th-normalized fluxes as a  function of depth and the linear regression (data from Bradtmiller et al., 2006; two yellow data points show the data from western Pacific and the others are data from eastern Pacific). 148  In addition to correlating with water depth, Pa/Th at these sites also correlates, but less tightly, with export production estimated from satellite data (r 2 = 0.59), as already noted by Pichat et al. (2004). Although export production and water depth also weakly correlate (r2 = 0.38) at these core locations, both water depth and particle flux appear to play a role in determining sediment Pa/Th in the Pacific equatorial upwelling region.  This is further confirmed when revisiting the data reported by Bradtmiller et al. (2006), who found a correlation between  230  Th-normalized opal fluxes and sediment Pa/Th using  11 cores spanning the entire equatorial Pacific (Fig. 4.24a). Plotting the differences between measured Pa/Th and the regression line reported in Figure 4. 24a against water depth (Fig. 4.24b) reveals a clear correlation (r2 = 0.88), confirming that water depth is also important in dictating sediment Pa/Th in the equatorial Pacific.  The three relatively shallow cores from the eastern equatorial Pacific with lower Pa/Th than neighboring cores (grey squares; Fig. 4.21-4.23) come from the Panama Basin just north of the equator, a region with relatively low productivity between the equatorial upwelling and the Costa Rica upwelling dome, and may thus reflect lower scavenging intensity at this particular location. Alternatively, lower Pa/Th could also be explained if the upper section of these cores is missing and the sediment analyzed is not modern. This is probably the explanation for two of the cores with very low Pa/Th taken from the low productivity region (green diamonds; Fig. 4.21-4.23), where core tops can be lost more readily due to lower sedimentation rates.  149  More intriguing, five core tops from the western equatorial Pacific (brown squares; Fig. 4.21 & 4.23) have significantly higher Pa/Th ratios than expected from their depth, even though they were collected in regions where satellite derived export production estimates are low. This observation was already discussed by Pichat et al. (2004) and is now confirmed by similarly high values from the Bradtmiller et al. (2006) data set. Pichat et al. (2004) suggest that the deep chlorophyll maximum observed in this region may account for a higher export production than is deduced from ocean color. Clearly, this feature needs further investigation by conducting water column measurements in the region.  The 2D model also generates sediment Pa/Th lower than measured at three low productivity sites (black circles; Fig. 4.21-4.23). Two of the cores were taken in the vicinity of Hawaii and may reflect a local effect on particle flux scavenging or underwater volcanoes, but there is no explanation for the high value observed the central equatorial Pacific core (4810 m).  The Pa/Th of sediment deposited at ocean margins and opal dominated regions (Fig. 4.21) are invariably higher than in the low productivity regions (Fig. 4.22, 4.23) reflecting the effect of enhanced scavenging. Some cores taken at the boundary between the subtropical and subpolar region (light blue squares) show the same depth dependency as the low productivity cores, suggesting that opal flux at these sites is too small to significantly affect Pa/Th in the underlying sediments.  150  4.3.4 Is the decreasing trend in sediment Pa/Th with depth below 1500m a result of the PMOC?  The vertical gradient in sediment Pa/Th is one of the key features generated by the meridional overturning circulation in the Atlantic Ocean (chapter 2, 3; Lippold et al., 2011). To assess whether the vertical gradient in sediment Pa/Th observed in the Pacific and generated by the 2D scavenging model can also be attributed to the overturning circulation, we ran the scavenging model without PMOC while maintaining boundary scavenging.  Sediment Pa/Th generated by different Boundary scavenging without PMOC 0.00 1500  2000  0.01  0.02  0.03  0.04  0.05  0.06  0.07  0.08  0.09  0.10  0.5*R without PMOC  Depth (m)  2500  1*R without PMOC 3000  3500  2*R without PMOC  4000  4500  Figure 4.25: Vertical sediment Pa/Th profiles generated between 35°N and 45°S by the 2D scavenging model in the absence of PMOC and with varying boundary scavenging strength. 151  Under these conditions (Fig. 4.25), the model does not produce the decrease in sediment Pa/Th with depth below 1500m observed in the data or in the model when PMOC is present. As expected, increasing boundary scavenging decreases sediment Pa/Th in the low productivity regions represented in the 2D model, but the decrease happens at all depths. The vertical gradient in sediment Pa/Th in low productivity regions thus seems to be a direct consequence of the overturning circulation.  4.3.5 Sensitivity of sediment Pa/Th in low productivity regions to changes in the rate of PMOC  The ultimate goal of this study is to assess whether sediment Pa/Th could potentially provide constraints on past changes in the rate of PMOC. To assess whether sediment Pa/Th in the Pacific could be used as a paleocirculation proxy, we must establish whether the vertical and latitudinal distributions of sediment Pa/Th in low productivity regions are sufficiently sensitive to PMOC strength and the extent to which they are also influenced by factors others than the PMOC.  When varying PMOC from 13 Sv to 39 Sv in the model, and keeping the same water mass distribution and boundary scavenging intensity, the latitudinal gradients of sediment Pa/Th at different depths change significantly (Fig. 4.26). With 39 Sv PMOC (i.e., a 50% increase compared to today), sediment Pa/Th at 4750m is low and almost constant (~ 0.03) with latitude. As PMOC intensity decreases, sediment Pa/Th at 4750m increases gradually to a maximum at mid-latitudes in the northern hemisphere. 152  The model predicts that a 50% reduction in PMOC rate from 26Sv to 13Sv would generate an increase in sediment Pa/Th of ~ 0.01 at 20-30°N in deep water (4.26a). In contrast, at the base of the NPIW (~3000m), high rates of PMOC result in a sharp northward decrease in sediment Pa/Th from ~0.075 to ~0.035 between 40°S and 30°N (4.26b). Slowing down PMOC results in nearly constant and high (~0.075) sediment Pa/Th south of the equator and a sharp decrease further north, where SCW upwells. These model outputs suggest that the rate of PMOC would be more easily estimated from changes in sediment Pa/Th in the northern subtropical Pacific than in the South Pacific. Slower rates of PMOC would generate higher sediment Pa/Th at all depths in the North Pacific. Confirmation that the signal is generated by changes in PMOC rate could be obtained by determining the latitudinal gradients of sediment Pa/Th at a given depth. In addition to generating higher sediment Pa/Th in the northern subtropical Pacific, slower rates of PMOC should also produce an increase in the northward gradient of sediment Pa/Th in deep water. On the other hand, faster rates of PMOC would decrease sediment Pa/Th in the northern subtropical Pacific and reduce the latitudinal gradient in deep water but generate a sharp northward decrease in sediment Pa/Th at mid-depth.  As already noted in Fig. 4.19 and Fig. 4.20, the vertical gradient of sediment Pa/Th between 5000m and 3000m is larger in the South Pacific and decreases gradually to the North (Fig. 4.27).  153  Pa/Th  (a) Sedi Pa/Th @ 4500-4750m 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02  13Sv 26Sv 39Sv Data  -50  -40  -30  -20  -10 0 Longitude  10  20  30  20  30  20  30  Pa/Th  (b) Sedi Pa/Th @ 2750-3000m 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 -50  13Sv  26Sv  39Sv  Data  -40  -30  -20  -10 0 Longitude  10  Pa/Th  (c) Sedi Pa/Th @ 1500-1750m 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02  13Sv 26Sv 39Sv -50  -40  -30  -20  -10 0 Longitude  10  Figure 4.26: Variations in sediment Pa/Th as a function of latitude in low productivity regions predicted by the 2D scavenging model with varying PMOC rates. Squares are measured sediment Pa/Th from the low productivity zone.  154  Sedi Pa/Th @ 3000m - Sedi Pa/Th @ 4750m 0.06  39 Sv  0.05  26 Sv  13 Sv ΔPa/Th  0.04 0.03 0.02 0.01 0.00 -50  -40  -30  -20  -10 0 Longitude  10  20  30  Figure 4.27: Changes in the sediment Pa/Th vertical gradient between deep and intermediated depth as a function of latitude and PMOC strength.  However, the relationship between the vertical gradients in sediment Pa/Th and the rate of PMOC is complex. An increase in the rate of PMOC from 26 to 39Sv would decrease the vertical gradients by ~ 0.01 between 40°S and 20°N, but a similar decrease between 30°S and the equator could be generated by a decrease in PMOC from 26Sv to 13Sv. These two situations could be distinguished from the vertical gradient at 10 – 20°N. High rates of PMOC would result in lower vertical gradients in this region than low rates of PMOC (Fig. 4.27). However, the predicted changes in gradients are rather small and it may be difficult to precisely measure them to constrain the rate of PMOC.  While the available data capture the decrease in sediment Pa/Th with depth (Fig. 4.23), they are not precise enough to show the expected latitudinal trends (Fig. 4.26). This database has been obtained over the course of several decades using different techniques 155  (radiometric and mass spectrometric) and without inter-calibration. The latter is particularly crucial to obtain a precise data base (Anderson et al., 2012). Establishing whether the latitudinal trends can be recognized in the sediments of the Pacific Ocean will require careful collection of surface sediment with appropriate coring techniques, assessing the impact of bioturbation, particularly in low sedimentation rate cores, and a systematic coring program based on our understanding of the main flow path of AABW and LCDW.  Figure 4.28: Difference in sediment Pa/Th generated by the 2D scavenging model with PMOC = 26Sv and 13Sv. Faster PMOC results in a substantial increase in sediment Pa/Th in the southern ocean.  The 2D model also suggests that lower rates of PMOC would result in a substantially smaller  231  Pa export from the Pacific into the Southern Ocean, where it is effectively 156  scavenged by opal (Fig. 4.28). Sediment Pa/Th in the Pacific sector of the Southern ocean appears to have been significantly lower during the last glacial maximum than during the Holocene (Chase et al., 2003), which would be consistent with a significant decrease in the rate of PMOC (Jaccard et al., 2010). Such a trend could be further confirmed by a simultaneous increase in Pa/Th at mid-depth in mid-northern latitudes at 30 – 40 °N (Fig. 4.28).  4.3.6 Relative sensitivity of sediment Pa/Th in low productivity regions to changes in PMOC and “boundary scavenging” – What is the most promising approach to constrain PMOC from sediment Pa/Th?  In the Pacific, the sediment Pa/Th in low productivity regions could also be significantly affected by changes in boundary scavenging. It is therefore important to establish how the signal generated by PMOC could be modified by possible changes in Pa scavenging at the margins.  If we keep PMOC constant but multiply the effective removal rate constant to the margins (R) by a constant, sediment Pa/Th decreases with increasing R, as expected (Fig. 4.29). The changes in sediment Pa/Th are more pronounced at intermediate depths, resulting in large changes in the vertical gradients (Fig. 4.30), but the latitudinal gradients change little (Fig. 4.29). Latitudinal gradients in sediment Pa/Th in deep waters are thus sensitive mainly to changes in the rate of PMOC and largely unaffected by changes in boundary scavenging and could potentially be used to reconstruct past changes in the rate of the 157  overturning circulation in the Pacific Ocean. In contrast, changes in sediment Pa/Th at intermediate depth and changes in the vertical gradient in sediment Pa/Th are more sensitive to changes in boundary scavenging than they are to changes in PMOC.  Pa/Th  Sedi Pa/Th @ 4500-4750m 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 -50  -40  0.5R  1R  2R  Data  -30  -20  -10 0 Longitude  10  20  30  20  30  20  30  Pa/Th  Sedi Pa/Th @ 2750-3000m 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02  Pa/Th  -50  -40  0.5R  1R  2R  Data  -30  -20  -10 0 Longitude  10  Sedi Pa/Th @ 1500-1750m  0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02  0.5R 2R -50  -40  -30  1R  -20  -10 0 Longitude  10  Figure 4.29: Variations in sediment Pa/Th as a function of latitude in low productivity regions predicted by the 2D scavenging model with varying boundary scavenging (R = effective removal rate constant to the margins) 158  Sedi Pa/Th @ 3000m - Sedi Pa/Th @ 4750m 0.07 3000m-4750m 0.5R  0.06  3000m-4750m 1R  0.05 ΔPa/Th  3000m-4750m 2R 0.04 0.03 0.02 0.01  0 -50  -40  -30  -20  -10 0 Longitude  10  20  30  40  Figure 4.30: changes in the difference in sediment Pa/Th between 3000 and 4750m as a function of latitude for a fixed PMOC (26Sv) and varying effective removal rate constant to the margins.  4.4. Conclusions  The 2D scavenging model indicates that the Pacific meridional overturning circulation can explain the curvature in the seawater dissolved  231  Pa profiles often observed in the North  Pacific. The model also indicates that changes in the rate of PMOC can influence the distribution of sediment Pa/Th, particularly in low productivity regions. The latter raises the prospect for developing a paleocirculation proxy for the Pacific Ocean.  The results obtained in this study suggest that past changes in the rate of the PMOC would be more easily estimated from changes in the longitudinal gradient of sediment Pa/Th at depth, following the main flow path of deep water. On the other hand, changes in 159  sediment Pa/Th at intermediate depths and vertical gradients in sediment Pa/Th appear to be more sensitive to the intensity of boundary scavenging. Although PMOC seems essential to generate a vertical sediment Pa/Th gradient in low productivity regions, this gradient is strongly modulated by boundary scavenging.  Sediment Pa/Th in the equatorial upwelling regions is affected both by particle flux and water depth, indicating that changes in circulation could also affect sediment Pa/Th in this region, and interpretation of changes in this ratio in terms of particle scavenging alone maybe unwarranted.  The existing sediment Pa/Th database, although good enough to document the relatively large vertical gradient in low productivity areas, is not sufficiently precise and accurate to discern the expected latitudinal gradients. Documenting these relatively subtle changes to document past changes in PMOC will require a concerted effort involving careful planning of a targeted coring program and inter-calibration of measurements.  160  Chapter 5 A comparison of POC fluxes recorded by sediment traps and 234Th:238U disequilibrium in a coastal region (Saanich Inlet, British Columbia)  5.1 Introduction  The role played by the coastal zone in the marine carbon cycle is disproportionate to its surface area. While accounting for only 8% of the surface area of the whole ocean, it contributes up to 30% of oceanic primary production and 80% of the organic carbon burial in marine sediments (Liu et al., 2000, Middelburg et al., 1997, Holt et al., 2009). However, notwithstanding the clear quantitative importance of the coastal zone for the carbon budget of the ocean and its potential role for sequestering anthropogenic CO2, there is not, as yet, a consensus on whether this oceanic region is a net source or a net sink for atmospheric CO2, and what fraction of the organic carbon produced in the coastal zone or added from nearby landmasses is exported to the deep sea, buried in margin sediments, or remineralized in coastal waters.  The reason for the relatively poor characterization of the carbon cycle at ocean margins is 161  that coastal zones are complex and the processes regulating the fate of carbon are highly variable both in time and space. This complicates the interpretation and the integration of instantaneous or localized measurements of quantities such as primary production, export flux and burial rates. Although satellites provide a powerful tool to start addressing both spatial and temporal variability, they do not provide direct estimates of flux and are prone to biases resulting from non-biotic influences on the optical properties of surface waters (Carr et al., 2006). Clearly, there is a need to develop means of estimating carbon fluxes (uptake, export, burial) in coastal waters using methods that provide some level of integration to generate flux estimates that could be more meaningfully parameterized in global carbon models.  According to Tsunogai et al. (1999), the biological pump operating in the coastal zone (i.e., the “continental shelf pump”) is distinct from the more classical biological pump of the ocean interior in that it requires lateral transport to offshore waters to remove carbon from the atmosphere on centennial to millennial timescales. Therefore, even though coastal production is high, its efficacy at sequestering atmospheric CO 2 depends on physical factors which control the transport of carbon to the deep sea. The “continental shelf pump” is primarily driven by isopycnal transport of dissolved inorganic carbon (DIC) from shelf bottom waters to greater depths offshore. Dissolved inorganic carbon (DIC) concentration in bottom shelf waters is largely controlled by the remineralization of particulate organic carbon (POC) settling from surface coastal waters. It is also possible that a fraction of the settling POC is laterally exported to the continental slopes. Measuring the sinking flux of carbon in the coastal zone is thus an essential part of better characterizing the continental 162  shelf pump and its variability under different environmental conditions (high vs low latitude, wide vs narrow shelves; proximity to major rivers and canyons, circulation, etc.).  In this chapter, two methods of estimating the sinking flux of organic carbon are compared in a coastal setting: bottom tethered sediment traps and  234  Th/238U  disequilibrium. Both approaches provide means of integrating the sinking flux of carbon over several weeks and documenting seasonal variations. However, both methods are also associated with significant uncertainties, particularly when applied in coastal waters. By comparing directly the two approaches in this setting, a reasonable agreement, while not proving their accuracy, would nonetheless bring more credence to the validity of the fluxes measured by either method.  5.2 Measuring the sinking flux of carbon with sediment traps  Sediment traps of various designs have been widely used to measure the flux of sinking particles in the ocean. They range from free-drifting floating traps, which are deployed from surface buoys for a few days at a time (a more recent design consists of neutrally buoyant sediment traps; Buesseler et al., 2000), to bottom-tethered time-series traps that are typically deployed in the deep ocean for extended periods of time (Honjo et al., 2008). In coastal waters, bottom-tethered traps of simple designs, as used in this study, are often deployed for extended periods of time and turned-over on a regular basis to collect time-series samples (e.g., Timothy et al., 2003). 163  While conceptually straightforward, the measurement of particle flux with sediment traps is subject to significant uncertainties. When deployed in dynamic regions with strong currents, the flux measured by sediment traps often seems to exceed or underestimate the true vertical flux (e.g., Buesseler, 1991). This is mostly because of hydrodynamic biases resulting from the formation of eddies near the mouth of the trap (Butman et al., 1986). Another difficulty with sediment traps is the presence of “swimmers”, living organisms that swim into the trap and die if the trap has been poisoned (thereby increasing the measured carbon flux), or consume the organic matter accumulating in the trap in the absence of poison (thereby decreasing the measured carbon flux). These biases are more severe in shallow waters (Yu et al., 2001) and therefore particularly problematic in coastal waters, which are relatively shallow and dynamic environments.  5.3 Measuring the sinking flux of carbon using  234  Th/238U disequilibrium in surface  waters  In addition to sediment traps, the  234  Th:238U disequilibrium method has also been applied  extensively to estimate the sinking flux of carbon in the ocean (Buesseler et al., 2006 and references therein). While sediment traps can provide flux estimates at any depths, the 234  Th/238U method is mostly restricted to the upper water column, where particle flux is  high enough to produce a measurable deficit in 234Th.  Thorium-234 is a relatively short-lived (half-life: 24.1 days) isotope of thorium produced 164  by decay of dissolved  238  U in seawater. Uranium is soluble in seawater and has a long  residence time in the ocean (200,000-400,000 years; Ku et al., 1977; Dunk et al., 2002). As a result, its concentration is virtually proportional to salinity (Owens et al., 2011; Chen et al., 1986), and the  234  Th formation rate is thus fairly uniform (i.e., proportional to  salinity) throughout the ocean.  The production rate of  234  Th (atoms/m3) is the rate at which its parent  238  U decays, which  is also known as the activity of 238U: A238 = 238 x N238  (5.1)  where 238, N238 and A238 are the decay constant (min-1or d-1), concentration (atoms/m3) and activity (dpm/m3) of 238U. When sinking particles remove 234Th from seawater in a closed system at steady-state, the rate of production of  234  Th must be equal to the sum of its rate of decay and removal by  scavenging: 238 x N238 = 234 x N234 + r234  (5.2)  or  A238 = A234 + r234  (5.3)  where 234, N234, A234 and r234 are the decay constant (d-1), total (dissolved + particulate) 165  concentration (atoms/m3), activity (dpm/m3) and removal rate by scavenging (atoms/m 3.d) of  Th. Multiplying each term by 234 and rearranging equation 5.3 expresses the  234  removal flux of 234Th in dpm/m3.d (R234): R234 = 234 x (A238 - A234)  When scavenging removes  (5.4)  234  Th at a rate which is similar to or greater than its rate of  decay, as is the case in most oceanic surface waters, the activity of 234Th in seawater drops below that of  238  U, and this measurable deficit can be used to estimate R 234. Integrating  over the depth of the euphotic zone (or the depth where a measureable found) gives the removal rate of  234  Th deficit is  234  Th from the upper water column, expressed in  dpm/m2.d 𝑧  FluxThz= 234∫0 (A238 − A234) dz  (5.5)  where z is the depth of integration.  While A234 is measured at discrete depths, A238 is often calculated assuming conservative behavior of U in seawater (Chen et al., 1986): A238 (dpm l-1) = 0.0704dpm l-1 x salinity  (5.6)  The 234Th flux obtained from equation 5.5 is then multiplied by the POC/234Th ratio of the particles that sink from the surface and scavenge  234  Th to calculate the export flux of  carbon: 166  FluxPOCz= FluxThz *(CPOCz/CThz)  (5.7)  where FluxPOCz and FluxThz are the sinking fluxes of POC and  234  Th at depth z, CPOCz and  CThz are the POC and 234Th concentrations of the sinking particles at depth z. 234  Complications arise, however, when the system is open, and  Th is added or removed  from the area of study by advection or mixing (Buesseler et al., 1995; Bacon et al., 1996; Cochran et al., 1995), or when the system is not at steady state (e.g., at the onset or at the end of a bloom). In these situations, equation 5.5 must be modified (e.g., Savoye et al., 2006): 𝑧  FluxThz= 234∫0 (A238 − A234]) dz + V – A234/t (5.8)  where V is the sum of the advective and diffusive fluxes of  234  Th (dpm/m2/d) to or from  the study site and A234/t is the change in 234Th activity with time at the study site. Another complication arises from uncertainties regarding the (CPOCz/CThz) ratio that should be used to convert  234  Th fluxes into carbon fluxes (Buesseler et al., 2006). This  ratio often increases with particle size but not always. Generally, the ratio obtained from large particles (>53um) sampled with large volume pumps is used, based on the assumption that these are the particles that actually remove  234  Th from surface waters, but  very different estimates of carbon flux can be obtained depending on the method used to collect particles for measuring CPOCz/CThz.  167  Application of the  234  Th:238U method in coastal regions may be further complicated by  other factors such as possible deviations from the  238  U-salinity relationship caused by  fresh water input with variable U content (Owens et al., 2011; Rutgers van der Loeff et al., 2006; Palmer and Edmond, 1993). In situations where anoxia develops in the water column, as is the case for Saanich Inlet, possible removal of U from the water column is also a concern. There are also potential calibration problems caused by high and variable particle load on the filters used for particulate of dissolved  234  Th measurements and variable recovery  234  Th by co-precipitation due to complexation by dissolved or colloidal  organic matter. It is therefore recommended to directly measure the uranium concentration of anoxic or low salinity waters, to carefully control the particle load on filters and to use a yield monitor during the co-precipitation of  234  Th from seawater samples (Rutgers van  der Leoff et al., 2006).  5.4 Materials and methods  5.4.1 Study Site  Saanich inlet is located in southern Vancouver Island, British Columbia, Canada (Fig. 5.1). It has a maximum depth of 230 m and is connected to Georgia Strait through Satellite Channel and a relatively broad and shallow (80 m) sill which restricts deep water renewal. River runoff at the head of the inlet is very small (Stucchi and Whitney, 1997) and the largest sources of fresh water are the Cowichan and Fraser Rivers (Fig. 5.1) adding 168  freshwater through Satellite Channel (Herlinveaux, 1962). The tidal-induced overflow of the water from Georgia Strait results in the continuous and vigorous renewal of the waters above the sill depth, but deep water renewal only takes place in late summer or fall. Wind-driven upwelling along the coast of the northeast Pacific Ocean and enhanced estuarine circulation from the freshet of the Fraser River draw dense and oxygenated waters through Juan de Fuca Strait and into the Strait of Georgia (Masson & Cummins, 1999). These waters are then able to flush the anoxic water that develops through the course of the year in the deep Saanich basin (Anderson and Devol, 1973).  Figure 5.1: Map of Saanich inlet and coastal southwestern British Columbia. The star indicates the position of the sediment trap mooring for this study. SN-9 and SN0.8 are the sediment trap mooring locations in Timothy et al. (2003).  169  Previous investigations have shown that primary production in Saanich Inlet is highly seasonal and generally higher than in neighboring Georgia Strait (Timothy and Soon, 2001). Diatoms contribute most of the yearly primary production, with occasionally significant growth of dinoflagellates and nanoflagellates (Timothy et al., 2003).  5.4.2. Sample collection, preparation and analyses 5.4.2.1 Hydrography Monthly temperature, salinity and oxygen profiles were obtained during most of the cruises (Table. 5.1) using a CTD deployed over the entire water column at the study site. Salinity and oxygen were calibrated on a regular basis with measurements from bottle samples collected at several depths during each cruise. Table 5.1: Sampling schedule for sediment traps, 234Th deficit measurements and hydrography.  2009 M  2010 A  M  J  x  x  J  A  S  O  N  D  J  F  2011 M  A  M  x  x  -  J  J  A  S  O  N  D  J  F  x  x  x  x  Trap 234  Th  -  x  x  CTD  Grey indicates that data were obtained. White indicates no measurement for a given month. For 234Th; (x) indicates that the data are used in the following discussion. In two instances (-) [March-09 and May-10], the data are not considered. The first set of total 234Th samples was not spiked with  230  Th so 170  that the recovery during co-precipitation could not be estimated. The May2010 samples were damaged during shipping between UBC and IOS and the accuracy of their counts is suspect. Taken together, the 234  Th measurements yield a nearly complete monthly coverage of a full year if we take: Mar (2010),  Apr (2010), May (2009), Jun (2009), Jul (gap), Aug (2010), Sep (gap), Oct (2009), Nov (2010), Dec (2010), Jan (2011), Feb (2011). This is the sequence of samples that was used to produce the time-series reported in Figs. 5.11, 5.16, 5.21 and 5.25.  5.4.2.2 Sediment traps  The deployment of sediment traps began on February 10, 2009 and continued uninterrupted until February 9, 2011 (Table D1, Appendix D). The sediment traps were made from PVC tubing (Timothy and Soon, 2001) with an inside diameter of 14 cm and a height of 50 cm (aspect ratio = 3.6). In order to decrease mixing (Gardner, 1980), a baffle grid (1.5 cm squares) was placed at the opening of each trap and another approximately 10 cm above the base. The sediment traps were deployed in pairs at three depths (50 m, 115 m and 180 m) on a mooring located at 48.59°N, 123.50°W (water depth: 230 m; Fig. 5.1), and serviced on a monthly basis (Table 5.1). Concentrated sodium chloride (300 g/l) was added at the bottom of all traps before deployment and sodium azide was added to one trap of each pair to retard bacterial degradation (Knauer & Asper, 1989).  Upon recovery, the samples were sieved using a Nylon screen with 200μm mesh size to remove the swimmers, centrifuged, freeze-dried and ground with a pestle and mortar for subsequent analysis. Total carbon and nitrogen were measured by gas chromatography on a model 1106 Carlo Erba CHN analyser with a precision of ±1.3% for carbon and ±2% for 171  nitrogen (Verardo et al., 1990). Inorganic carbon was measured with a Coulometrics Inc. CO2 coulometer (precision ±2%, derived from 2SD% of standards determination) and organic carbon calculated by difference. Biogenic silica was measured following the method described by Mortlock and Froelich (1989) or Müller and Schneider (1993). Samples were weighed (~ 20 mg) in centrifuge tubes. Hydrogen peroxide and HCl were added to remove organic and inorganic carbon, respectively. After removing the dissolved phase by centrifugation, opal was extracted with Na 2CO3 at 85ºC in a water bath. Dissolved silica was measured colorimetrically at 812 nm on a LKB spectrophotometer. This method has a precision of about 10%. 5.4.2.3 234Th method  Th-234 disequilibrium was measured on multiple occasions during the sediment trap deployment period (Table 5.1). Sea water samples were collected with GO-FLO bottles (General Oceanics) attached to a cable and closed with a messenger. Total  234  Th was  obtained by MnO2 co-precipitation on 2L samples (Buesseler et al., 2001) and filtration onto a 25mm diameter Tissuequartz filter (25mm TQ samples hereafter), while particulate 234  Th samples were obtained by filtering 4-8L of seawater onto 25mm diameter  Tissuequartz filters. The samples were then counted for beta emission using a Risø low level multi-counter (DTU Nutech).  172  5.4.2.3.1 Total 234Th:  Upon return to the laboratory at the end of the day of sampling, a graduated cylinder was used to measure 2L of sea water (precision: ±10 ml) and transferred into a Nalgene bottle. The samples were acidified to pH2 with 2N HCl and spiked with 1g of a 10 dpm.g-1 solution of  230  Th in 8N HNO3. After 12 hours to allow isotopic equilibration, the pH was  raised back to 8 with concentrated NH4OH before adding 100μl KMnO4 (3g/l) and 100μl MnCl2 (8g/l) to generate a MnO2 precipitate. The Nalgene bottles were then shaken vigorously and left to equilibrate for 1hr. They were subsequently heated in a water bath at 80ºC for an additional hour to reduce filtration time, following Cai et al. (2006). After cooling in an ice bath, the MnO2 precipitates were filtered onto tissue-quartz filters (25 mm diameter, 1μm nominal pore size), oven dried at ~50 º C, and covered with Polyethylene wrap (1.11mg/cm2) for beta counting. Due to the relatively low counts generated by our 2 L samples, we did not cover our samples with Al foil to shield possible beta emissions from contaminants, and thus, some of our  234  Th counts may have been  overestimated by 0-0.4 dpm (L. Miller, pers. comm.). The samples were counted three times in the course of 234  40  234  Th decay to verify that the beta emissions were primarily due to  Th. Background emissions from possible longer lived beta emitting contaminants (e.g.,  K) were measured after ten  234  Th half-lives. Counting efficiencies were established for  each counter by processing 2 L samples of acidified deep water (> 400 m, details in Chapter 6) collected in the NE Pacific at station P26 and stored in the laboratory for an extended period of time so that  234  Th was in equilibrium with  238  U. After the last count, 173  the samples were spiked with 200 mg of a solution containing 690 dpm/g of  229  Th in 8 N  HNO3 and digested on a hot plate with 20ml concentrated HNO3, 5 ml concentrated HClO4 and 2 ml concentrated HF. The Th was then purified by anion-exchange to measure  230  Th (and therefore  234  Th) recovery during the Mn oxide co-precipitation,  following Pike et al. (2005). This small volume method provides a fast and convenient way to measure total  234  Th and minimizes possible procedural errors caused by  self-absorption when larger volumes, necessitating larger amount of MnO2, are used (Cai et al., 2006). 5.4.2.3.2 Particulate 234Th:  Two different methods were used to collect marine particles for beta counting:  Approximately 6-8L of seawater were measured with a graduated cylinder (±10 ml) and filtered onto a pre-combusted (350ºC, 4hrs) TQ filter (25 mm diameter, 1 μm pore size). The samples were then oven-dried at ~50ºC.  On several occasions, samples were also obtained using large volume in-situ pumps deployed at the same depths where the water samples were taken. Particles were collected by filtering relatively large volumes (generally > 100 L) of seawater onto 142 mm diameter Tissue Quartz filters with 1μm pore size. The filters were dried at ~50ºC in an oven and a 25mm diameter punch was taken for 234Th counting.  174  Both the 25mm filters and 25mm punches were covered with Polyethylene wrap and counted three times. Counting efficiencies for the particulate samples were determined using a standard  238  U/FeO filter made according to Rutgers van der Loeff and Moore  (1999). Our particulate  234  Th measurements were consistent with those of other  laboratories during the GEOTRACES inter-calibration exercise (Maiti et al., 2012). Because of the relatively large particle load on the filters, contribution from dissolved 234  Th adsorbed on the filter during filtration was considered negligible. Several “dip”  blanks were collected during the deployment of the large volume pumps and were found to have similar activities as the blank TQ filters (< 0.1 cpm). Following completion of counting, at least 8 months after the initial sample collection, POC and PON were measured by gas chromatography on a model 1106 Carlo Erba CHN analyser.  5.5 The U-salinity relationship in Saanich Inlet waters  As indicated in section 5.3, when applying the  234  Th/238U disequilibrium method to  coastal waters, in addition to the usual precautions regarding blanks, background corrections, yield during co-precipitation, and calibration, special attention must be given to the U/salinity correlation used to estimate seawater U activity from salinity measurements. Deviations from the correlation observed in open ocean waters could arise as a result of local freshwater input with relatively high U content or, in the case of Saanich Inlet, removal of U from anoxic seawater. 175  Although the deeper waters of Saanich Inlet are anoxic, their frequent renewal suggests that their residence time in the inlet is too short to produce a substantial decrease in uranium concentration. This is corroborated by estimates of the removal rate of authigenic uranium in the anoxic sediments of Saanich Inlet (~5,000 dpm.m -2.y-1; Anderson et al., 1989). If the authigenic U removed in Saanich sediment is uniformly removed from a 100m layer of anoxic water which gets renewed every year, the resulting maximum decrease in U concentration at the end of the year would be 50 dpm/m 3, i.e., 0.05 dpm/l. If deep water renewal happens only every two years, then the expected maximum drop in concentration at the end of two years would be 0.1 dpm/l. This is relatively small compared to total U activity in Saanich water. Moreover, there is no evidence from the analysis of material collected with sediment traps that U is scavenged uniformly from the anoxic zone (Anderson et al., 1989). Instead, it seems that a large fraction of U reaching the sediment is included in particles formed in surface waters or added by an unknown mechanism at the sediment water interface.  In coastal waters with relatively low salinity, as found in Saanich Inlet, deviations from the conventional U-salinity relationship could be potentially important if there is a local source of U-enriched freshwater. Dissolved uranium concentrations measured in the water column of Saanich Inlet in April 2009 (Amini and Holmden, pers comm) are significantly higher than expected from salinity and equation 5.6 (Table 5.2), suggesting that Saanich Inlet waters are influenced by a freshwater end-member with a high U concentration. Assuming conservative mixing with a seawater end-member with a salinity of 35 and a U  176  concentration of 2.46 dpm/l (35 x 0.0704), we can calculate the U concentration of the freshwater end-member needed to explain the observed concentrations (Table 5.2): A[U]rw = (35 x (A[U]meas – (0.0704dpm l-1 x S))) / (35 – S)  (5.9)  where S is the salinity of the sample. Table 5.2: Dissolved U concentrations in Saanich Inlet (123°30.2’N; 48°34.6’W) measured on 10 April, 2009 (M. Amini and C. Holmden, pers. comm), and estimated based on two different U – salinity relationships (see text for explanations). Depth  Salinity  [O2]  [U]meas  A[U]meas  A[U]s  A[U]rw  A[U]  m  psu  mM  ppb*  dpm.L-1*  dpm.L-1**  nmol.kg-1#  dpm.L-1##  10  29.959  382.5  3.01  2.20  2.11  3.5  2.34  115  30.906  10.3  3.25  2.37  2.18  9.3  2.37  200  31.329  1.2  3.13  2.28  2.21  -  2.38  (*) Measured (**) Calculated from equation 5.6 (#) [U] in the freshwater end-member calculated by assuming conservative mixing between a seawater end-member with a salinity of 35‰ and a uranium concentration of 2.46 dpm.l-1 (equation 5.9) (##) Calculated from equation 5.10  The calculated freshwater end-member value is lower for the surface (10m) sample, which could reflect a local source (e.g., the Cowichan River) mostly affecting the salinity of surface water, while a more regional source is affecting the sample taken at 115 m. For the latter, the Fraser River is the most obvious potential source. However, the calculated value is much higher than obtained in preliminary measurements of Fraser River water (0.5-1 nmol.kg-1; Peuker-Ehrenbrink and Voss, pers comm), suggesting another U source, 177  possibly associated with groundwater. Interestingly, the U concentration at 200 m is ~ 0.1 dpm/l lower than at 115m, possibly reflecting removal of U from bottom waters by diffusion into anoxic sediments.  Considering the very limited number of uranium concentration measurements from Saanich Inlet available to date and our ignorance regarding the source of the U-enriched freshwater end-member, the 234Th deficit was calculated using two different approaches to provide an estimate of the error that could result from uncertainties in the exact U-salinity relationship. In the first approach, we calculate the  234  Th deficit using equation 5.5. As  shown in Table 5.2, this calculation yields U concentrations that are 0.1 – 0.2 lower than observations. Alternatively, we derived a regional U-salinity relationship based on the U concentration and salinity measured at 115m, which yields a freshwater end member with a U concentration of 2.20 ppb or 1.66 dpm/l (or 9.3 nmoles.kg-1; Table 5.2): A[U] (dpm L-1) = (1.66dpm l-1 x ((35-S)/35)) + (2.46dpm l-1 x S/ 35) (5.10) where S is the salinity of the sample. This calculation slightly overestimates the U content of bottom waters, possibly depleted in U by diffusion in sediments, and in surface water, possibly diluted by a more local source of freshwater with lower U concentration.  178  5.6 Results and discussion  5.6.1 Salinity, temperature, density and O2  The lowest surface salinities in Saanich Inlet occur in winter when local precipitation is highest (Fig. 5.2a). The temperature profiles reflect seasonal heating (Fig. 5.2b) with maximum surface temperatures between June and August and minimum between December and February. The entire water column at the study site is stratified year round with very shallow mixing layers (Fig. 5.2c; Timothy and Soon, 2001). The temperature of the deep water in the basin varies little from month to month. Small increases in salinity and density in late summer or early fall are the results of deep water renewal events (Anderson and Devol, 1973).  179  30  2429 23 2825 2627  23 27  23 29 28 27  30  26  24  20  30 2928 26 27 25 24  24 26 28 25  Salinity SI 31  30  40  29  29  30  30  23  60  26  28 25 24  2429 23 2825 26 30 27  24 25 27 26 28  24 26 28 3025 23 27 31  31  31  27  2330 29 28 27  100  30 31 29  Depth (m)  80  31  120  26 23  29  140  26  Jun/2009  30  Oct/2009  2429 23 2825 26 30 27  24 26 28 3025 23 27 31  Apr/2009  Dec/2009  24  2330 29 28 27  200  31  31 2928 26 25 27 24  Aug/2009  180  25  25 24 31  160  23  Feb/2010 Apr/2010 Cruise Month  Jun/2010  Aug/2010  Oct/2010  Dec/2011  Feb/2011  Temperature SI 6  10  15  8  8  12  16  8  12  12 68  20  14  12  10  16  10  14  10  8  10  40  10  13  8  12  6  11  8  100 8  10  8  68  9  120  8 6  8  140  7  6  6  160 6  Apr/2009  Jun/2009  8  200  8  180 86  Depth (m)  6  80  6  10  60  Aug/2009  Oct/2009  5  Dec/2009  Feb/2010 Apr/2010 Cruise Month  Jun/2010  Aug/2010  Oct/2010  Dec/2011  Feb/2011  180  Density SI 24  20 21  23.5  21  19 20 22 21  23  23  21 20  23 22  22 20 21 19  20  22  19  22  23  23  40 22.5  19 20 22 21 23  24 19  20.5  20  23 20 2419  23 22  140  Apr/2009  Jun/2009  Oct/2009  19 24 20 22 21 23  2419  22 20 21 19  Aug/2009  180  Dec/2009  19.5  19 2223  24  20 21  160  200  21 24  24  24  120  222321  21 20  21.5 24 22 20 21 19  Depth (m)  80  100  22  20 19  23 22  23  23  60  Feb/2010 Apr/2010 Cruise Month  18.5 Jun/2010  Aug/2010  Oct/2010  Dec/2011  Feb/2011  Oxygen SI 3  9  4  10  11  6 5  2 4  6  7  3  8  4  6  5  5  3  10  5  1  20  6  5  4 3  6  6  7  8  5  65 7 89 10 9  3  3  1 2  8  4  4  60  3  4  4 2  2  Depth (m)  1  3  3  2  100  7  4  3  3  80  9  4  2  1  4  4  1  2  40  2 1  1  1  6  5 120 4 140 3 160 2 180  200  1  Apr/2009  Jun/2009  Aug/2009  Oct/2009  Dec/2009  Feb/2010 Apr/2010 Cruise Month  Jun/2010  Aug/2010  Oct/2010  Dec/2011  Feb/2011  Figure 5.2: Time-series hydrographic data from surface to 200m during the 2 year period: (a) Salinity (psu), (b) temperature (°C), (c) density () and (d) oxygen (ml/L).  Anoxia develops in the water column below the sill depth (Fig. 5.2d) and has been attributed mainly to high primary production, which results in high export flux of carbon, 181  and weak estuarine circulation, which limits lateral export of phytoplankton biomass from the inlet (Timothy and Soon, 2001). High primary productivity in Saanich Inlet appears to be mostly sustained by lateral transport of nutrients from Georgia Strait (Timothy and Soon, 2001). The boundary of the anoxic zone ([O2] < 10 μmol/kg), defined by Kamykowski and Zentara, 1990) is generally between 120m and 130m in the winter seasons but rises to shallower depths in late summer and early fall, as anoxic water is pushed upward by denser water overflowing from the sill. The oxygen concentration of deep water does not increase substantially during these events because of the presence of reducing chemical species (H2S, CH4, Mn2+, etc.) that react with the incoming oxygen.  5.6.2 Sediment traps  5.6.2.1: Sample mass and concentrations  The weight of material collected for each deployment and the concentration of organic carbon (POC), total nitrogen (PON), inorganic carbon (PIC) and biogenic silica (BSi) are reported in Table D1, Appendix D. Only POC and BSi are discussed in this chapter. Calcium carbonate (CaCO3) contributed less than 2% of the mass flux on average as foraminifera and coccolithophorids are rare. The percentage of PIC was only used to estimate organic carbon from total carbon measured with the CN Analyzer. Nitrogen (PON) was used to calculate the C/N ratio of the settling material.  As indicated above, sediment traps were deployed in pairs at each depth. A concentrated NaCl solution was added to the bottom of the two traps and NaN3 was added to one of 182  them to retard bacterial degradation (Knauer & Asper, 1989). However, there is no significant difference in the % POC measured by the two sets of sediment traps (%POCNaN3-%POCNaCl = 0 ± 1.5 %), corroborating the findings of Timothy et al. (2003). From January 2010, the samples from the non-NaN3 treated traps were used to measure oxygen isotopes in biogenic silica (De Baere et al., in prep). The data plotted in the figures of this chapter are therefore just from the NaN3 treated traps.  Seasonal variations in % opal and % organic matter are shown in Fig. 5.3. Percent organic matter was estimated from %POC x 1.85, the conversion factor used in Saanich Inlet by Timothy et al. (2003). Biogenic silica concentrations show a clear seasonal pattern with high concentrations in spring, a gradual decrease through summer and minimum values in winter. The settling material intercepted by the trap deployed at 50 m tends to have higher opal concentration, reflecting either dissolution or dilution by laterally transported material in the deeper traps. On the other hand, organic matter concentrations reach a maximum in late summer, which is observed at all depths in 2009 but only in the shallow trap in 2010. The relatively low concentration of organic matter collected during the spring bloom is likely due to dilution by biogenic silica. The higher concentrations in late summer suggest a shift in plankton community toward species other than diatoms.  183  Seasonal compositional characteristics of settling particles in Saanich 60.00% Shallow Trap @ 50m 50.00% Middle Trap @ 115m 40.00% Opal%  Deep Trap @ 180m  30.00% 20.00% 10.00% 0.00% Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  40.00% Shallow Trap @ 50m Middle Trap @ 115m  30.00% OM%  Deep Trap @ 180m 20.00%  10.00%  0.00% Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  Figure 5.3: Seasonal changes in (a) Opal% and (b), OM% in the material collected by sediment traps at 3 depths in Saanich Inlet.  Previous studies conducted in Saanich Inlet have revealed relatively high OC/N ratio in the organic matter collected with sediment traps (Timothy and Soon, 2001; Timothy et al., 2003). Our results (Fig. 5.4) confirm this observation, yielding a molar OC/N ratio of 8.7 which is consistent with value of 8.5 reported at the head of Saanich Inlet by Timothy et al., 2003. This relatively high ratio must reflect addition of terrigenous organic matter (with a high C/N ratio), as would be expected for coastal waters. Terrigenous organic matter contributions are expected to be higher in winter when productivity is low and 184  continental run-off is highest and are expected to be lowest during the spring blooms. Lower C/N ratios are found in spring (Fig. 5.4a). Samples collected in summer 2009 had higher C/N than those collected in summer 2010, suggesting higher input of terrigenous organic matter in 2009.  5.6.2.2 Fluxes  Fluxes of total mass, biogenic silica, and organic carbon recorded at the three depths between February 2009 and February 2011 are shown in Figs. 5.5, 5.6 and 5.7, respectively, and reported in Table D2; Appendix D. Opal and organic carbon fluxes are clearly highest in spring at all depths, as a result of diatom blooms. In comparison, seasonal changes in total mass fluxes are muted, particularly at 50 m (Fig. 5.5a). Total mass fluxes are also relatively high between November 2009 and January 2010. Since the increase in opal and organic carbon fluxes was less pronounced during that period, the winter increase in total mass fluxes can be attributed to an increase in the flux of lithogenic material, possibly associated with higher run-off or resuspension of sediments from the nearby sill. The main component of biogenic Si in Saanich is diatom frustules (Sancetta & Calvert, 1988). Particulate organic carbon fluxes at 50 m showed a seasonal pattern similar to that of Biogenic Si (Fig. 5.6a, 5.7a), despite the fact that terrigenous and other marine OM (e.g., flagellates and zooplankton) were also caught by the sediment traps. Both BSi and POC fluxes roughly reflected the yearly cycle of primary production (Timothy et al., 2003), increasing in March–April and decreasing in September–October. Although the flux of BSi exhibits an rapid decrease in June-July, the POC flux still 185  remained at high levels until after September, supporting the possibility that flagellates play an increasingly important role after the spring diatom bloom (Hobson and Mcquoid, 2001; Timothy and Soon, 2001).  13 12 11  Molar OC/N  10 9 8 7 Shallow Trap @ 50m  6 5  Middle Trap @ 115m  4  Deep Trap @ 180m  3 Dec-08  Mar-09  Jul-09  Oct-09  Jan-10  May-10  Aug-10  Nov-10  Feb-11  15 OC = 8.7124 N + 0.2462 R²= 0.9378  OC%  10  5  0 0  0.5  1  1.5  N% * (12/14)  Figure 5.4: (a) Seasonal variations in the OC/N ratio of sediment trap material collected at the three depths. (b) OC versus N and OC/N values for all the sediment-trap samples. Weight percent N is multiplied by 12/14, so the slope of the regression line provides the estimated OC/N molar ratio. 186  Mass fluxes recorded by sediment traps deployed at 3 different depths  Mass Flux (g/m2/day)  9 8  Shallow Trap @ 50m  7  Middle Trap @ 115m  6  Deep Trap @ 180m  5 4 3 2 1 0 Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  Mass Flux (g/m2/day)  9 8  Shallow Trap @ 50m  7  Middle Trap @ 115m  6  Deep Trap @ 180m  5 4 3 2 1 0 Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  Mass Flux (g/m2/day)  9 8  Shallow Trap @ 50m  7  Middle Trap @ 115m  6  Deep Trap @ 180m  5 4 3 2 1 0 Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  Figure 5.5: Mass fluxes recorded by sediment traps deployed at three different depths over the 2-year time series. The data from shallow, middle and deep traps are highlighted in 5.5(a), 5.5(b) and 5.5(c), respectively.  187  Opal fluxes recorded by sediment traps deployed at 3 different depths 3.5 Shallow Trap @ 50m  Opal Flux (g/m2/day)  3  Middle Trap @ 115m  2.5  Deep Trap @ 180m  2 1.5 1 0.5 0 Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  3.5 Shallow Trap @ 50m  Opal Flux (g/m2/day)  3  Middle Trap @ 115m  2.5  Deep Trap @ 180m  2 1.5 1 0.5 0 Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  3.5 Shallow Trap @ 50m Opal Flux (g/m2/day)  3  Middle Trap @ 115m  2.5  Deep Trap @ 180m  2  1.5 1 0.5 0 Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  Figure 5.6: Opal fluxes recorded by sediment traps deployed at three different depths over the 2-year time series. The data from shallow, middle and deep traps are highlighted in fig. 5.6(a), 5.6(b) and 5.6(c), respectively.  188  POC fluxes recorded by sediment traps deployed at 3 different depths 0.6 Shallow Trap @ 50m POC Flux (g/m2/day)  0.5  Middle Trap @ 115m Deep Trap @ 180m  0.4 0.3  0.2 0.1 0 Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  0.6  POC Flux (g/m2/day)  Shallow Trap @ 50m 0.5  Middle Trap @ 115m  0.4  Deep Trap @ 180m  0.3 0.2 0.1 0 Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  0.6  POC Flux (g/m2/day)  Shallow Trap @ 50m  0.5  Middle Trap @ 115m  0.4  Deep Trap @ 180m  0.3 0.2 0.1 0 Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  Figure 5.7: POC fluxes recorded by sediment traps deployed at three different depths over the 2-year time series. The data from shallow, middle and deep traps are highlighted in fig. 5.7(a), 5.7(b) and 5.7(c), respectively.  189  Fluxes are generally highest at 115m depth and the contrast in flux with water depth is larger for total mass than for biogenic silica and organic carbon. A similar trend was observed by Timothy et al. (2003) and attributed to the re-suspension of sediment from the sill and lateral transport at mid-depth into Saanich Inlet.  Table 5.3: The comparison of fluxes for mass (a), biogenic opal (b) and organic carbon (c) between this study and those reported by Timothy et al. (2003), in which the depth deployments of the traps were 50m (shallow), 135m (mid) and 180m (deep) at station SN-0.8 (48.55°N, 123.55°W) and 45m (shallow), 110m (mid) and 150m (deep) at station SN-9 (48.67°N, 123.51°W). 5.3a: Annual mass fluxes (g/m2/d) Stations  This study (48.59°N; 123.50°W)  SN-0.8  SN-9  Sampling period  03/2009-02/2010  03/2010-02/2011  01/1984-12/1989 08/1983-12/1989  Shallow Trap  3.09  2.63  1.63  5.67  Mid Trap  4.91  4.31  2.33  11.1  Deep Trap  3.87  3.83  2.40  12.7  SN-9  5.3b: Annual biogenic opal fluxes (g/m2/d) Stations  This study (48.59°N; 123.50°W)  SN-0.8  Sampling period  03/2009-02/2010  03/2010-02/2011  01/1984-12/1989 08/1983-12/1989  Shallow Trap  1.02  0.76  0.669  1.84  Mid Trap  1.41  1.09  0.750  2.70  Deep Trap  1.06  0.95  0.680  2.80  190  5.3c: Annual OC fluxes (g/m2/d)  Stations  This study (48.59°N; 123.50°W)  SN-0.8  SN-9  Sampling period  03/2009-02/2010  03/2010-02/2011  01/1984-12/1989 08/1983-12/1989  Shallow Trap  0.229  0.21  0.168  0.351  Mid Trap  0.294  0.232  0.178  0.543  Deep Trap  0.214  0.206  0.177  0.498  The mass fluxes between March/2010 and February/2011 are slightly lower than fluxes measured between March/2009 and February/2010 (Table 5.3). Since the fluxes of biogenic silica and organic carbon are also slightly lower during the second year, this difference is at least partly due to slightly lower export productivity. This difference is however small compared to the difference in annual fluxes between this study and the earlier sediment trap study conducted in Saanich Inlet between January 1984 and December 1989 (Timothy et al., 2003). During this study, two sediment trap moorings were deployed in the deep basin of Saanich Inlet (Fig. 5.1), one near the sill (SN-9) and one near the head of the inlet (SN 0.8), while our sediment trap mooring was located between these two stations at the center of the inlet (Fig. 5.1). Fluxes for total mass, biogenic opal and organic carbon at our station fall between the fluxes measured at SN-0.8 and at SN-9. This trend coincides with the location of the stations which suggests that fluxes increase towards the sill at the mouth of the inlet. This increase has been  191  attributed both to higher primary production (Timothy and Soon, 2001) and closer proximity to the source of sediment from the sill (Timothy et al., 2003). 5.6.3 234Th deficits  Th-234 deficits in the water column were measured several times between March 2009 and February 2011 (Table 5.1; Table D3, Appendix D). Samples from May/2010 were damaged during shipment while the Mar/2009 samples were not spiked with  230  Th. These  two sets of samples are not considered in the following discussion. By combining the measurements conducted on Mar/2010, Apr/2010, May/2009, Jun/2009, Aug/2010, Oct/2009, Nov/2010, Dec/2010, Jan/2011 and Feb/2011 we can obtain a nearly complete composite picture of the annual cycle of  234  Th deficit and partitioning between dissolved  and particulate phases.  The samples were counted long enough to obtain counting errors within 3-8% for all the samples. The errors associated with counting efficiency and decay curve regression resulted in a final propagated error of about 8% for total 234  Th. Because dissolved  234  Th and 15% for particulate  234  Th was calculated by difference and the fraction of  234  Th  dissolved is generally smaller than the fraction adsorbed on particles, the relative error on dissolved 234Th is large and variable. The average dissolved 234Th calculated by difference in this study is 0.26 ± 0.21 dpm/l (1 standard deviation, n = 97) and the absolute error is 0.13 ±0.05 dpm/l. The average relative error on dissolved 234Th is thus 56%.  192  Particulate samples were also collected by Large Volume Pump (LVP) in Mar/2009, Dec/2010 and Jan/2011 on 150mm Tissuequartz filter (Table D4, Appendix D). Subsamples (LVP samples hereafter) were punched from the 150 mm filters for  234  Th  analysis, which provided a comparison to the results from the particulate samples obtained by small volume filtration onto 25 mm diameter TQ filters (25mm TQ samples). 5.6.3.1 Total 234Th profiles Total 234Th activities generally show large deficits in the water column (Fig. 5.8; Table D3, Appendix D) compared to the  238  U activities derived from equations 5.6 and 5.10,  indicating that the particle flux in Saanich Inlet is nearly always high enough to scavenge a large fraction of the 234Th before it decays.  193  Total 234Th activity (dpm/L) 0  0.5  1  1.5  2  2.5  3  0 Dec 20 Nov 40  Oct  60  Aug  Jun  Depth (m)  80  May 100 Apr 120  Mar  140  Feb Jan  160  U-238 (a) 180 U-238 (b) 200  Figure 5.8: Total  234  Th activities measured in Mar/2010, Apr/2010, May/2009, Jun/2009, Aug/2010,  Oct/2009, Nov/2010, Dec/2010, Jan/2011 and Feb/2011. The  238  U activities are estimated using  equation 5.6 [U-238 (a)] and 5.10 [U-238 (b)].  The averages of all data measured at a given depth +/– one standard deviation are shown in Figure 5.9. In general, total  234  Th activities are lower in the upper 100 m of the water  column (~ 0.5 – 1 dpm/l) and higher but more variable towards the bottom.  194  Total 234Th (dpm/L) 0  0.5  1  1.5  2  2.5  3  0 20 40 60  Depth (m)  U-238 (a) 80 U-238 (b) 100 120 140 160 180  200  Figure 5.9: Average total 234Th measured at a given depth (dpm/l). The two black lines show the one standard deviation envelope. The 238U activities are estimated using equation 5.6 [U-238 (a)] and 5.10 [U-238 (b)].  Total  234  Th activities in the upper 100m tend to be somewhat higher in late winter (Fig.  5.10a). The deficit is however still pronounced, indicating continued scavenging by particles. In early spring, total  234  Th in the upper 100 m decreases, indicating enhanced  scavenging during the spring bloom. This trend is clearly illustrated by the May-09 and April-10 profiles (Fig. 5.10b). The latter, however, show significantly higher total deeper water. In late summer or early fall, total  234  Th in  234  Th profiles tend to show intermediate  values in the upper 100 m but high values in deeper water (Fig. 5.10c). The intermediate total 234Th activity in the upper 100 m is consistent with the intermediate fluxes generally 195  recorded at that time by the sediment traps (Fig. 5.5), while the high  234  Th measured at  depth could be attributed to resuspension of recently deposited sediment as a result of deep water renewal in the deep basin of the inlet.  Total 234Th (dpm/L) 0.00 0  0.50  1.00  1.50  2.00  2.50  20 40  Depth (m)  60 80  Jan-11  100 Feb-11  120 140 160 180 200  Total 234Th (dpm/L) 0.00 0  0.50  1.00  1.50  2.00  2.50  20 40  Depth (m)  60 80 100  May-09 Apr-10  120 140 160 180 200  196  Total 234Th (dpm/L) 0.00 0  0.50  1.00  1.50  2.00  2.50  20 40  Depth (m)  60 80  Aug-10  100  Oct-09  120 140 160 180 200  Figure 5.10: Total 234Th during (a) late winter, (b) early spring, (c) late summer or early fall. The thin black lines are the extrema of total 234Th from Figure 5.9.  Total Th234 SI (dpm/L) 20  2  0.6  0. 8  1  1  0.4  0.6  1.8  40  0.8 0.8  1.6 0.8  70  0.8  0.8  1.2  0.8  100  0.8  0.6  Depth (m)  0.6  1.4  1  1  1 1  0.6  0.8  8 0.  0.6  1. 2  1  115  1  1.2 1.4  3  1  4  5  6  7  0.6 1.2  1.6 1.8  160  1.6 1.8 2  1.4  0.8  2 1.  130  4 1.  2  8  9  10  0.4 11  12  1  2  Cruise Month  Figure 5.11: Monthly composite time-series of total 234Th. Numbers on the horizontal axis correspond to the months of the year. The figure was drawn by combining the measurements conducted on Mar/2010, Apr/2010, May/2009, Jun/2009, Aug/2010, Oct/2009, Nov/2010, Dec/2010, Jan/2011 and Feb/2011.  197  These general trends can also be seen in the composite picture of the monthly changes in total 234Th over an entire year (Fig. 5.11). We clearly see the lower total  234  Th activities in  spring and higher values in deep water over most of the year, particularly during the spring bloom and the late summer deep water renewal. 5.6.3.2 Particulate 234Th profiles  Total  234  Th activity is partitioned between dissolved (< 1 μm) and particulate (> 1 μm)  phases. In Saanich Inlet, the particulate  234  Th activity is generally higher than the  dissolved 234Th activity. Averaging all available data at a given depth (Fig. 5.12) indicates that the fraction of particulate  234  Th increases from 48% at the surface to 85% near the  oxic/anoxic interface and decreases back to 60% towards the bottom.  % particulate 234Th 0  20  40  60  80  100  0 20 40  Depth (m)  60 80 100 120 140  160 180 200  Figure 5.12: Average fraction of particulate 234Th (% of total 234Th) as a function of water depth.  Particulate  234  Th activities vary from ~ 0.2–1.7 dpm/l (Fig. 5.13) and generally increase  with depth (Fig. 5.14). Temporal changes in particulate  234  Th generally follow those  observed for total 234Th (Fig. 5.15; Fig. 5.16). The increasing trend with depth is observed 198  234  through the year and particulate  Th dominates the deep total  234  Th maxima observed in  spring and late summer. Particulate 0  0.5  234Th  1  activity (dpm/L)  1.5  2  2.5  3  0  Dec Nov  20  Oct 40 Aug  Depth (m)  60  Jun  80  May  100  Apr  120  Mar  140  Feb Jan  160  U-238 (a)  180  U-238 (b) 200  Figure 5.13: Particulate 234Th activities over an annual cycle. The 238U activities are estimated using equation 5.6 [U-238 (a)] and 5.10 [U-238 (b)].  Particulate 234Th (dpm/L) 0.0  0.5  1.0  1.5  2.0  2.5  0 20 40  Depth (m)  60 80 100 120 140 160 180 200  Figure 5.14: Average particulate  234  Th measured at a given depth (dpm/l). The two black lines show  the averages ±one standard deviation. 199  Particulate 234Th (dpm/L) 0.0  0.5  1.0  1.5  2.0  2.5  0 20  Depth (m)  40 60 80  Jan-11  100  Feb-11  120 140 160 180 200  Particulate 234Th (dpm/L) 0.0  0.5  1.0  1.5  2.0  2.5  0 20  Depth (m)  40 60 80  May-09  100  Apr-10  120 140 160 180 200  Particulate 234Th (dpm/L) 0.0  0.5  1.0  1.5  2.0  2.5  0 20 40  Depth (m)  60  80  Aug-10  100  Oct-09  120 140 160 180 200  Figure 5.15: Particulate 234Th during (a) late winter, (b) early spring, (c) late summer or early fall. The thin black lines are the extrema of particulate 234Th from Fig. 5.14. 200  Particulate Th234 SI (dpm/L) 1.8 0. 4  20  0. 4  1.6 4 0.  0.4  40  6 0.  1.4  0.6 0.4  70  1.2  Depth (m)  0.6  1 100 6 0.  8 0.  0.8 0.8  115  0.6  1.4  3  4  5  6  1  7  8  9  0.2  1  1.2  1.6  1.4  0. 8  1.2  0.8  1.2  160  0.4  0.8  130 1  0.6  0.8  0. 6  0.8  1  10  11  12  1  2  Cruise Month  Figure 5.16: Monthly composite time-series of particulate  234  Th. Numbers on the horizontal axis  correspond to the months of the year. The figure was drawn by combining the measurements conducted on Mar/2010, Apr/2010, May/2009, Jun/2009, Aug/2010, Oct/2009, Nov/2010, Dec/2010, Jan/2011 and Feb/2011.  % dissolved 234Th 0  10  20  30  40  50  60  0 20 40  Depth (m)  60 80 100 120  140 160 180 200  Figure 5.17: Average fraction of dissolved 234Th (% of total 234Th) as a function of water depth. 201  Dissolved 234Th (dpm/L) 0  0.5  1  1.5  2  2.5  3  0  Dec  20  Nov  40  Oct Aug  60 Depth (m)  Jun 80 May 100  Apr  120  Mar  140  Feb  160  Jan  180  U-238 (a) U-238 (b)  200  Figure 5.18: Dissolved 234Th activities over an annual cycle. The 238U activities are estimated using equation 5.6 [U-238 (a)] and 5.10 [U-238 (b)].  Dissolved 234Th (dpm/L) 0.00 0  0.50  1.00  1.50  2.00  2.50  20 40  Depth(m)  60 80 100 120 140 160 180 200  Figure 5.19: Average dissolved  234  Th profile (dpm/l). The two black lines show the average ± one  standard deviation.  202  Dissolved 234Th (dpm/L)  Depth (m)  0.0  0.5  1.0  1.5  0 20 40 60 80 100 120 140 160 180 200  2.0  2.5  Jan-11 Feb-11  Dissolved 234Th (dpm/L)  Depth (m)  0.0  0.5  1.0  1.5  0 20 40 60 80 100 120 140 160 180 200  2.0  2.5  Apr-10 May-09  Dissolved 234Th (dpm/L)  Depth (m)  0.0  0.5  1.0  1.5  0 20 40 60 80 100 120 140 160 180 200  2.0  2.5  Aug-10 Oct-10  Figure 5.20: Dissolved 234Th during (a) late winter, (b) early spring, (c) late summer or early fall. The thin black lines are the extrema of dissolved 234Th from Fig. 5.19. 203  5.6.3.3 Dissolved 234Th profiles 234  The average fraction of total  Th in the dissolved pool decreases from 53% in surface  water to 15% near the oxic/anoxic interface and then increases to 38% towards the bottom (Fig. 5.17). Dissolved 234Th activities vary from below detection limits to ~ 0.8 dpm/l (Fig. 5.18) and are generally lowest near the oxic/anoxic boundary (Fig. 5.19). Seasonal variations in dissolved 234Th are not as clear as for total or particulate  234  Th, as a  result of lower activities and larger errors on the measurement (Fig. 5.20). The clearest seasonal signal is found in the upper 50 m of the water column where dissolved  234  Th is  lowest in spring and highest in winter, following primary production (Fig.5.21). The minimum dissolved 234Th at mid-depth appears to be a year-round feature (Fig. 5.20), but attenuated during deep water renewal (Fig. 5.21).  Dissolved Th-234 SI (dpm/L) 0.5  0.2  0.2  0.1  0.3  0  0.3  20  0.6 0.4  40 0. 4  0.3  0.4  0.2  0.2  0  0.1 0  0.2  0  -0.1  0.4  0.1  0.1  0.2  0.2  0  115  0.3  0.3  -0 .2  0.2 0.1  0 0.4  1 0.  0.5  4  0. 3  0.3 0. 4  0.3  5  0.5 0.6  0.2  6  7  8  9  0.3  130  0.2  -0 .1  3  0.4  0.2  0.1  100  0.1  0.2  0.3  0.3  Depth (m)  0.3  70  160  0.5  0. 3  0.3  0.1  0.2  0.4  10  11  12  1  2  0  Cruise Month  Figure 5.21: Monthly composite time-series of dissolved  234  Th. Number on the horizontal axis  correspond to the months of the year. The figure was obtained by combining the measurements 204  conducted on Mar/2010, Apr/2010, May/2009, Jun/2009, Aug/2010, Oct/2009, Nov/2010, Dec/2010, Jan/2011 and Feb/2011.  5.6.3.4 POC/234Th ratio The POC/234Th ratio (μgC/dpm) of particles collected in Saanich Inlet generally varies between ~ 30 and 120 gC/dpm (Fig. 5.22), except for four abnormally high values (20m in Mar/2010, 80m in Apr/2010, 10m in Nov/2010 and 100m in Dec/2010), possibly due to the accidental inclusion of zooplankton in the samples. Neglecting these four samples, the average POC/234Th ratio generally decreases with depth with a possible secondary maximum near the oxic/anoxic boundary (Fig. 5.23). The POC/234Th ratios observed in Saanich Inlet are within the range reported elsewhere (Buesseler et al., 2006) and are not particularly high. While we might expect that particles from productive coastal waters would have relatively high POC/234Th because of their high POC concentration, the ratios measured in Saanich Inlet are only slightly higher than those measured in spring in the surface waters of station Papa (Chapter 5). A possible explanation is the addition of lithogenic particles by rivers, which could lower the POC concentration of particles in Saanich Inlet, while keeping the adsorption of dissolved 234Th (Buesseler et al., 2006). Seasonal changes in the POC/234Th ratio are well defined above the oxic/anoxic interface with higher ratios during the spring bloom, lower ratios in the winter and intermediate ratios in late summer or early fall (Fig. 5.24; 5.25). Below 120 m, the seasonal variability 205  is less pronounced. Ratios of POC/234Th are lower than in shallower water, but there is a hint of slightly higher ratios in late summer (Fig. 5.25). In particular, the samples from the deep basin collected in October-2009 (Fig. 5.24c) have higher ratios, which could possibly reflect sediment resuspension of high POC particles during deep water renewal. POC/234Th on particulate samples (μg/dpm)  Depth (m)  0  100  200  300  400  500  600  0  Dec  20  Nov  40  Oct  60  Aug  80  Jun  100  May  120  Apr  140  Mar  160  Feb  180  Jan  200  Figure 5.22: POC/234Th (gC/dpm) on particulate samples over an annual cycle.  POC/234Th on particulate samples (μg/dpm) 0  20  40  60  80  100  120  0 20 40  Depth (m)  60 80 100 120 140 160 180 200  Figure 5.23: Average of all the POC/234Th measured in particles collected at a given depth (μg/dpm). The two black lines show the average ±one standard deviation. 206  POC/234Th on particulate samples (μg/dpm) 0.0  20.0  40.0  60.0  80.0  100.0  120.0  0 20  Depth (m)  40 60 80  Jan-11  100 Feb-11  120 140 160 180 200  POC/234Th on particulate samples (μg/dpm) 0.0  20.0  40.0  60.0  80.0  100.0  120.0  140.0  0 20  Depth (m)  40 60 80  May-09  100  Apr-10  120  140 160 180 200  POC/234Th on particulate samples (μg/dpm)  Depth (m)  0.0  20.0  40.0  60.0  80.0  100.0  0 20 40 60 80 100 120 140 160 180 200  120.0  Aug-10 Oct-09  Figure 5.24: POC/234Th during (a) late winter, (b) early spring, (c) late summer or early fall. The thin black lines are the standard deviations of POC/234Th from Fig. 5.23. 207  High POC/234Th ratios appear sporadically (e.g., Feb-11 at 115m; Fig. 5.24.a) and result from relatively high POC concentrations (Table D3; Appendix D). Such variability could be due to the inclusion of zooplankton or large aggregates in the samples and could be eliminated by pre-filtering the samples with a coarser mesh filter (e.g., Nitex: 53 m). The very high POC/234Th ratio measured at 115 m in November 2010 (Fig. 5.25) could in part reflect such an occurrence. On the other hand, relatively high POC/ 234Th between 70 and 120 m could also reflect slow upwelling of fine particles resuspended from the sediments during deep water renewal.  POC/Th-234 SI (ug/dpm) 20  80  70  80  10 0  50  110  60  70  90  120  60  70  80  60  70  0 11  12 0  40  80  90  100  50 70  60  90  80  50  40  90  90  30  70  70  70  60 50 60  50 70  50 60  80  40 30  40  50 50  40  30  4  5  6  7  30  50  50  30  3  10 0  60  130  160  70  70  80  60  70  70  115  90  50  60  100  70 60  70  80  80  60  80 80  Depth (m)  60  50  80  40  70  80  90  80  60  60  8  9  10  11  12  1  2  Cruise Month  Figure 5.25: Monthly composite time-series of POC/234Th. Numbers on the horizontal axis correspond to the months of the year. The figure was obtained by combining the measurements conducted on Mar/2010, Apr/2010, May/2009, Jun/2009, Aug/2010, Oct/2009, Nov/2010, Dec/2010, Jan/2011 and Feb/2011. Rarely sampled depths (10m and 180m) and samples suspected to be affected by capture of zooplankton during filtration are excluded from this figure).  208  5.6.3.5 Comparison between LVP samples and 25mm TQ samples.  POC and particulate  234  Th were measured on three occasions (March 2009, December  2010, and January 2011) using samples collected with large volume in-situ pumps (Table D4, Appendix D). The  234  Th and POC profiles are generally similar to those obtained by  small volume filtration, but with some significant differences (Fig. 5.26).  Figure 5.26: Comparison between POC, particulate  234  Th and POC/234Th obtained by collecting  particles with Large Volume in-situ Pumps (LVP) and small volume filtration, as described in the method section. 209  With the exception of one sample collected at 100m during the March-2009 sampling, when an abnormally high POC value was obtained with the LVP, the agreement between the two methods for POC concentration is reasonable. On the other hand, particulate  234  Th  activity is consistently higher on the LVP filters (by 22 ± 28% (1 SD)), resulting in somewhat lower POC/234Th ratios. The reason for this difference may be due to adsorption of dissolved  234  Th onto the filter. The ‘dip’ blanks used to correct for  234  Th  adsorption on the LVP filters were obtained by counting the filters loaded on LVPs which were deployed to the same depth but had 0 liter water passed through. Therefore they were very low and may underestimate dissolved  234  Th adsorption during filtration.  Because the samples obtained with the in-situ pumps are few, the carbon fluxes discussed below are all based on the samples obtained by small volume filtration. 5.6.3.6 Fluxes derived from 234Th:238U disequilibria 5.6.3.6.1 234Th fluxes 234  Th fluxes were first calculated using equation 5.5, which assumes a closed system at  steady state. The 234Th:238U deficits were calculated by integrating the difference between measured total  234  Th activities and  238  U activities using the trapezoid approximation of  Equation 5.5. U-238 activities were calculated from salinity using equation 5.6 and 5.10. The difference provides an estimate of the uncertainty stemming from not having measured 238U activity in all samples.  210  The  234  Th fluxes increase nearly linearly with depth and are high year-round compared to  the open ocean sites (e.g., station Papa in Chapter 6). Th-234 fluxes calculated with equation 5.6 are 15 % lower than those calculated with equation 5.10 (Fig. 5.27). Improving the accuracy of  Th fluxes to better that 15% would thus require measuring  U in seawater more extensively than has been done in this study.  234Th  fluxes in Jun-09 (dpm/cm2/day)  234Th  fluxes in May-09 (dpm/cm2/day) 0.2  0.4  0.6  0.8  1.0  0.0  0.4  0.6  0.8  1.0  0.0 0  20  20  20  40  40  40  60  60  60  80 100 120  Depth (m)  0  80 100 120  140  160  160  180  180  180  200  200  200  0.2  0.4  0.6  234Th  0.0  0.2  0.4  0.6  fluxes in May10(dpm/cm2/day)  0.8  0.0  0  0  20  20  20  40  40  40  60  60  60  100  Depth (m)  0  80  80 100  0.8  234Th  fluxes in Apr-10 (dpm/cm2/day)  0.8  0.6  120  140  0.0  0.4  100  160  fluxes in Mar-10 (dpm/cm2/day)  0.2  80  140  234Th  Depth (m)  0.2  fluxes in Oct-09 (dpm/cm2/day)  0  Depth (m)  Depth (m)  0.0  234Th  Depth (m)  238  234  0.2  0.4  0.6  0.8  80 100  120  120  120  140  140  140  160  160  160  180  180  180  211  234Th  234Th  fluxes in Aug-10 (dpm/cm2/day) 0.2  0.4  0.6  0.8  0.0  0.4  0.6  0.8  0.0 0  20  20  20  40  40  40  60  60  60  80 100  Depth (m)  0  80 100  120  140  140  140  160  160  160  180  180  180  0.4  0.6  0.8  0.0  20  20  40  40  60  60  Depth (m)  0  100  0.2  0.4  0.6  0.8  80 100  120  120  140  140  160  160  180  180  Figure 5.27:  0.8  fluxes in Feb-11 (dpm/cm2/day)  0  80  0.6  234Th  fluxes in Jan-11 (dpm/cm2/day) 0.2  0.4  100  120  0.0  0.2  80  120  234Th  Depth (m)  0.2  fluxes in Dec-10 (dpm/cm2/day)  0  Depth (m)  Depth (m)  0.0  234Th  fluxes in Nov-10 (dpm/cm2/day)  234  Th fluxes calculated from estimates of  238  U seawater concentration obtained from  equation 5.6 (blue line) and 5.10 (red line).  Seasonal changes in the fluxes of shows the average  234  Th are relatively small but well defined. Fig. 5.28  234  Th flux profile (± 1 standard deviation). The relative standard  deviation at different depths varies between 6 and 13%. Within this relatively small range, however, there is a clear seasonal signal. The  234  Th fluxes measured in late winter are at  the lower range of measured fluxes (Fig. 5.29a), while those measured during the spring 212  bloom are at the upper range (Fig. 5.29a). Fluxes measured in late summer are intermediate (Fig. 5.29b). Linear fits of the data shown in Fig. 5.29a yield: Jan-11: 234Th flux = 34.1 10-4 z  r2 = 0.997  Feb-11: 234Th flux = 33.8 10-4 z  r2 = 0.999  May-09: 234Th flux = 44.7 10-4 z  r2 = 0.997  Apr-10: 234Th flux = 42.3 10-4 z  r2 = 0.988  234  Th fluxes during spring blooms are thus on average 27% higher ([(44.7 + 42.3) / (34.1  + 33.8)] = 1.27) than in winter.  234Th  0.0  0.2  fluxes (average ± 1 SD) dpm/cm2/d 0.4  0.6  0.8  0 20 40  Depth (m)  60 80 100 120 140 160 180 200  Figure 5.28: Average calculated  234  Th flux profile (red line) ± 1 standard deviation (black lines),  assuming steady state. Fluxes were calculated from  238  U based on equation 5.5. The blue line  represents the maximum 234Th flux (234 A238) by assuming 234Th is completely depleted. 213  234Th  0.0  0.2  fluxes (dpm/cm2/d) 0.4  0.6  0.8  0 20 40  Depth (m)  60 80  Jan-11  100  Feb-11  120  May-09 Apr-10  140 160 180 200  234Th  0.0  0.2  fluxes (dpm/cm2/d) 0.4  0.6  0.8  0 20 40  Depth (m)  60 80 100 Aug-10 120  Oct-09  140 160 180 200  Figure 5.29: (a)  234  Th fluxes in late winter and during the spring bloom, (b)  234  Th fluxes in late  summer and early fall. The thin black lines are the standard deviations of the average 234Th fluxes from Fig. 5.28. 214  Since  234  Th profiles were measured during consecutive months on several occasions  (Table 5.1), the non-steady-state model (equation 5.8, assuming V = 0) can be used to refine our calculation of  234  Th flux in June-09, April-10, December-10, January-11 and  February-11. We used the equation of Buesseler et al. (1992), also reported in Gustafsson et al. (2004): FluxThz= 234 (zi –zi+1) ((A238 (1- e-t) + 1A234 e-t – 2A234) / (1 – e-t)  (5.11)  where zi and zi+1 are the depths of integration, 1A234 and 2A234 are the  234  Th activity  measured at the start and end of the sampling period under consideration, and t is the length of that period (days). For each of these months, the 234Th fluxes derived from equation 5.11 are very close to the fluxes calculated with the steady state model (Fig. 5.30). The fluxes calculated with the non-steady state model are within 15% of those calculated with the steady state model (Fig. 5.31). The main reason for the reasonable accuracy obtained with the steady-state model is the generally low total  234  Th activities in Saanich Inlet water, so that the relative  changes in activity between consecutive months are small compared to the large  234  Th  deficit.  215  Figure 5.30: Th-234 fluxes calculated with the steady state model vs the non-steady state model (equation 5.11). Blue diamonds and red squares are  234  Th fluxes obtained from  238  U seawater 216  concentration estimates obtained from equation 5.6 and 5.10, respectively. The black line is the 1:1 relationship.  234Th  0.4  0.6  fluxes - non steady state/steady state 0.8  1  1.2  1.4  0  Depth (m)  20 40  Jun-09  60  Apr-10  80  Dec-10  100  Jan-11  120  Feb-11  140 160 180 200  Figure 5.31: Ratio of fluxes estimated with the non-steady state model to those estimated with the steady state model from 238U seawater concentration estimates obtained from equation 5.5.  In our calculation of 234Th fluxes with equation 5.11, we assumed that exchanges of  234  Th  between Saanich Inlet and surrounding areas are negligible. Timothy et al. (2003) indicated that primary production is higher within the inlet than outside. We would 234  therefore expect some level of addition of total  Th from Georgia Strait into Saanich  Inlet. We have not measured 234Th outside Saanich Inlet and cannot ascertain whether this is the case. However, if there is addition of  234  Th into Saanich Inlet from neighboring  regions, it should be small considering the low total 234Th concentration typically found in these coastal waters.  217  5.6.3.6.2 POC fluxes  Particulate organic carbon (POC) fluxes were calculated by multiplying the  234  Th fluxes  obtained with the steady state model at a given depth and the POC/234Th ratio of the particles collected at the same depth (equation 5.7). The average POC fluxes increase with depth from the surface to the oxic/anoxic interface (Fig. 5.32). Below the oxic/anoxic interface, the POC fluxes decrease slightly but they increase again towards the bottom.  In open ocean waters, the flux of organic carbon generally increases with depth within the euphotic zone, where the rate of photosynthesis exceeds the rate of respiration, and then decreases with depth as a result of organic matter remineralization (e.g., Bacon et al., 1996; Chapter 6). Since the depth of the euphotic zone in Saanich Inlet is always shallower than 20 m (Timothy and Soon, 2001), the increase in organic carbon flux with depth down to 120 m can only be explained by lateral transport of organic matter that has been previously deposited at shallower depths for a period of time long enough to allow decay of 234Th that had been scavenged during sinking to the initial site of deposition.  218  POC fluxes (μmol/cm2/d) 0.0  0.5  1.0  1.5  2.0  2.5  3.0  3.5  4.0  0 20 40  Depth (m)  60 80 100 120 140 160 180  200  Figure 5.32: Average POC fluxes derived at a given depth (μmol/cm2/d) with the steady state model from 238U seawater concentration estimates obtained from equation 5.5. The two black lines show the average ±one standard deviation.  POC fluxes (μmol/cm2/d) 0.0  0.5  1.0  1.5  2.0  2.5  3.0  3.5  4.0  0 20 40  Depth (m)  60 80  Jan-11 Feb-11  100  May-09  120  Apr-10  140 160 180 200  219  POC fluxes (μmol/cm2/d) 0.0  0.5  1.0  1.5  2.0  2.5  3.0  3.5  4.0  0 20  Depth (m)  40 60  Aug-10  80  Oct-09  100 120 140 160 180 200  Figure 5.33: (a) POC fluxes in late winter and during the spring bloom, (b) POC fluxes in late summer or early fall. The thin black lines are the standard deviations of POC fluxes from Fig. 5.32.  There are clear seasonal variations in POC flux from the surface to the oxic/anoxic interface (Fig. 5.33, 5.34). Late winter and spring bloom POC fluxes are at the lower and higher ends of the POC flux range, respectively (Fig. 5.33a), while POC fluxes during late summer are intermediate (Fig. 5.33b). Seasonal variations in POC fluxes are less clear in deeper waters. The fluxes in deeper waters tend to be lower during winter and spring and higher in late summer/early fall. The latter may reflect resuspension of sediments by deep water renewal.  220  POC Flux SI (umol/cm2/day) 20  3.5 0.5 1  40  1 1  5 0.  3  1.5 1.5  60  1 2  2.5 70 1  2  1.5  Depth (m)  2  1.5  80  1  1.5  2  2. 5  1.5  2.5  3.5  2  3  100  115  5 1.  2  1  2  3  5 2.  3 2.5  130  4  2  5  6  1.5  3  2.5  160  1 2  2.5  7  8  9  10  0.5 1.5  11  12  1  2  Cruise Month  Figure 5.34: Contour plot of the seasonal and depth variation in POC fluxes derived with the steady state model from 238U seawater concentration estimates obtained from equation 5.5.  Timothy et al. (2003) attempted to estimate the export ratio (e-ratio) of organic carbon (the ratio of the settling flux of POC to net production in the euphotic zone) in Saanich Inlet by dividing annually averaged primary production measurements obtained by  14  C  incubations (Timothy and Soon, 2001) by the annually averaged flux of POC measured with a trap deployed at 50 m depth at stations SN-9 and SN-0.8 (Fig. 5.1). In doing so, they found surprisingly low e-ratios (0.12 at SN-9; 0.09 at SN0.8). Although they attempted to correct for the presence of terrigenous organic matter in the trap samples using 13C, they made the assumption that the flux of organic carbon intercepted by the 50 m trap originated entirely from the overlying surface water. However, the increase in 234  Th-based POC flux between 20 m and 50 m (i.e. below the euphotic zone) indicates  that some organic matter intercepted by the 50 m traps may have been laterally 221  transported instead of being exported from overlying surface water. This would yield even lower e-ratios. Primary production was not measured at our study site, but if we take the average between Timothy and Soon’s sites of study (SN-9: 13 mol/cm2/d; SN-0.8: 9 mol/cm2/d) and compare it to our  Th-based POC flux at 20m (0.5 mol/cm2/d) we  234  find an e-ratio of 0.5/11 = 0.045. This must be taken as a maximum since we did not account for contribution of terrigenous organic matter in the export flux. This suggests that there is extensive nutrient recycling in the surface waters of Saanich inlet (or that sediment traps and the  234  Th deficit method are missing an important part of the sinking  flux (see Chap 6)).  POC fluxes (μmol/cm2/d) 0.0  0.5  1.0  1.5  2.0  2.5  3.0  3.5  4.0  0 20 40  Depth (m)  60 80 100  Th-234 Trap  120 140 160 180 200  Figure 5.35: Average POC fluxes derived from 234Th deficit assuming steady state from 238U seawater concentration estimates obtained from equation 5.6 and average POC fluxes measured by sediment traps. The two black lines show the average ±one standard deviation for the fluxes derived from 234Th. The error bars on the average trap-recorded POC fluxes represent ±one standard deviation.  222  4  μmol/cm2/d  Trap 50m 3 Trap 115m 2 Trap 180m 1 Th-234 50m* 0 Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  Apr-11  4  μmol/cm2/d  Trap 50m 3 Trap 115m 2 Trap 180m 1 Th-234 115m 0 Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  Apr-11  4  μmol/cm2/d  Trap 50m 3  Trap 115m 2 Trap 180m 1 0 Feb-09  Th-234 180m* May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  Apr-11  Figure 5.36: Two year time series records of POC fluxes measured with sediment traps and corresponding fluxes derived from  234  Th deficits, using the steady state model and  238  U seawater  concentration estimates obtained from equation 5.5. Sediment trap fluxes at 50 m are compared to average 234Th-based fluxes measured at 40m and 60/70m. Sediment trap fluxes at 180 m are compared to 234Th-based fluxes measured at 180 m or 160m when no samples were obtained at 180m.  5.6.4 Comparing POC flux measured with sediment traps and 234Th deficits. The POC fluxes derived from  234  Th assuming steady state are in reasonable agreement  with the POC fluxes measured with sediment traps (Fig.5.35). The average POC flux 223  measured at 50 m with 234Th is lower than the flux intercepted by the shallow trap, but the two methods still agree within a factor of 2 and their ranges overlap. The difference between the two methods at 50 m is further illustrated by plotting the two data sets versus time (Fig. 5.36). The agreement between the two methods is better in deeper water, particularly at mid-depth, but on occasion  234  Th-based carbon fluxes are more than twice  the corresponding sediment traps carbon fluxes (Fig. 5.36). These occurrences (115m in March-2010, November-2010, Feburary-2011 and 180m in October-2009) are the results of high POC/234Th and could be attributed to the capture of zooplankton on the filters. If correct, a simple way to avoid this problem would be to screen the seawater samples before filtration (although this may also point to a more serious problem with the  234  Th  method that will be discussed in Chapter 6). Fig 5.36 also confirms that at 50 m, the 234Th method produces flux estimates that are significantly lower than those obtained with sediment traps. At this point, the reason for this difference is unclear. The difference could be explained if the 50 m trap intercepted more than the vertically settling flux (i.e. over trap) or if the  234  Th method underestimates the vertical POC flux. Lateral advection or  mixing of 234Th from Saanich Inlet into Satellite Channel or Georgia Strait could result in such an underestimation, but confirmation requires measurements of  234  regions. This explanation seems unlikely, considering the low dissolved  234  Th in these  Th activities  that are expected in these waters. Alternatively, the difference between the two methods could be explained by focusing of the POC flux between surface and 50 m depth. In fjords, the sinking flux of particles is increasingly focused with depth due to the gradual decrease in the width of the fjord (Wassmann, 1991; Timothy et al., 2003). Particle focusing thus 224  increases the flux of refractory particles with depth (by a factor inversely proportional to the decrease in the inlet’s width with depth) and mitigate (or reverse) the decrease in flux with depth of particles subjected to remineralization or dissolution. On the other hand, particle focusing is not affecting the POC/234Th ratios of sinking particles but increases 234  the activity of particulate  Th with depth. This would result in a decrease in POC flux  estimated from 234Th with depth, as long as sinking and focusing occurs rapidly compared to the half-life of 234Th. These contrasting responses to focusing by the two methods could help explain the difference in fluxes recorded at 50 m.  In addition to focusing, the increase in particle flux with depth in the upper 120 m of the water column could also be due to lateral transport of particles at depth. This process is different from sediment “focusing” discussed above. Focusing is syndepositional (i.e. it occurs as particles settle) but lateral transport is postdepositional (i.e. it occurs after initial deposition at another site). If particles were first deposited at shallower depth (e.g., on the sill) for a period of time long enough to allow decay of the excess  234  Th they scavenged  from the water column, their subsequent resuspension, lateral transport and sinking would result in a gradual increase in particle flux with depth which would be recorded by both methods. The sediment traps would intercept both the vertical and lateral flux, while lateral transport of “aged” particles would not affect particulate  234  Th concentration in  seawater but would increase the POC/234Th ratio of the particles. The better agreement between the POC fluxes estimated from  234  Th and measured by sediment traps deployed  at mid depth (Fig. 5.35) suggests that the increase in flux with depth between 50 m and 120 m recorded by the  234  Th deficit method is mainly due to this mechanism. This stands 225  to reason since the depth of the sill, probably the main source of material supplied laterally to Saanich Inlet, is at 70 m depth.  5.6.5 Can POC fluxes in coastal regions be estimated by only measuring particulate 234  Th and POC?  The activity of dissolved  234  Th in Saanich Inlet is consistently low (Fig. 5.18, 5.19) and is  almost completely depleted at mid depth (Fig. 5.21). Seasonal variability in dissolved 234  Th is thus also very small and difficult to precisely measure by difference (Fig. 5.20).  This suggests that one might obtain as good an estimate of carbon sinking fluxes in coastal waters by only measuring  234  Th and POC on particles. If such is the case, the  approach could provide a relatively simple and rapid method to survey and monitor the regional sinking flux of carbon or any other particle constituent in the coastal zone.  To assess the potential of this approach, we calculated carbon fluxes from our data, assuming that dissolved  234  Th is zero or always equal to the average activity measured  during this study (0.26 ±0.21 dpm/l).  In this case, equation 5.5 becomes: 𝑧  FluxThz= 234∫0 (𝐴238 − 𝑃𝐴234) dz  (5.12)  or 𝑧  FluxThz= 234∫0 (𝐴238 − 𝑃𝐴234 − 0.26) dz  (5.13) 226  where pA234 is the activity of particulate 234Th. In Saanich Inlet, average seawater activities are (Table D3; Appendix D): A238 = 2.26 ±0.11 dpm/l A234 = 0.92 ±0.34 dpm/l p  A234 = 0.66 ±0.32 dpm/l  Therefore, if we assume that dissolved  234  Th = 0 (equation 5.12), we would overestimate  the 234Th flux (and the POC flux) by: [2.26 (±0.11) – 0.66 (±0.32)] ----------------------------------- = 1.19 ±0.41 [2.26 (±0.11) – 0.92 (±0.34)] On the other hand, equation 5.13 would evidently give the right flux but with an uncertainty of 41%. [2.26 (±0.11) – 0.66 (±0.32) – 0.26] ------------------------------------------- = 1.00 ±0.41 [2.26 (±0.11) – 0.92 (±0.34)] To further assess the potential of this approximation, equations 5.12 and 5.13 were applied to each data point reported in Table D3 in Appendix D to calculate organic carbon fluxes which were then be compared to the fluxes obtained with equation 5.5. The results confirm that equation 5.12 overestimates the flux of organic carbon reaching 180m by 20% (Fig. 5.37a). While equation 5.13 provides an accurate mean flux at 180m, it overestimates fluxes in the upper 50m by ~ 10% (Fig. 5.37b) because the dissolved  234  Th  concentrations at these depths are higher than 0.26 dpm/l (Fig. 5.19). If we were to 227  substitute 0.26 in equation 5.13 by the mean dissolved  234  Th concentration in the upper 50  m, the equation would evidently provide a more accurate 234Th flux estimate.  POC fluxes - Simplified model (Equation 5.12) / Steady state model (Equation 5.5) 0  0.2  0.4  0.6  0.8  1  1.2  1.4  1.6  0 20  AVG  40  Dec Nov  Depth (m)  60  Oct 80  Aug Jun  100  May  120  Apr 140  Mar  Feb  160  Jan  180  POC fluxes - Simplified Model (Equation 5.13)/steady state model (Equation 5.5) 0  0.2  0.4  0.6  0.8  1  1.2  1.4  1.6  0 20  AVG  40  Dec Nov  Depth (m)  60  Oct 80  Aug  Jun  100  May  120  Apr  140  Mar Feb  160  Jan  180  Figure 5.37: (a) Ratios of organic carbon fluxes obtained with equation 5.12 and 5.5; (b) ratios of organic carbon fluxes obtained with equation 5.13 and 5.5. The  234  Th fluxes are derived from  238  U  seawater concentration estimates obtained from equation 5.6 for both cases. 228  The purpose of this exercise is to show that in productive coastal waters, measuring total 234  Th by the time consuming MnO2 co-precipitation method in every sample collected  will not increase significantly the precision and accuracy of the calculated fluxes. Collecting particles by filtering 4-8 L of seawater to measure particulate concentration, complemented by a few measurements of total  234  234  Th and POC  Th in the seasons with  highest and lowest primary productivities to constrain the mean value of total 234Th would suffice. The time saving that this approach affords would provide a means to map and routinely monitor the POC fluxes over large swaths of coastal waters. To further illustrate the validity of this simplified approach, organic carbon fluxes measured with sediment traps, equation 5.5 and equation 5.13 are compared on figure 5.38 and 5.39. They show that the agreement between sediment traps and fluxes derived from equation 5.5 is similar to the agreement between sediment traps and fluxes derived from equation 5.13, confirming that simply measuring systematically  234  Th and POC on  suspended particles only, complemented by a few measurements of total 234Th to estimate its mean concentration, suffices to assess the seasonal amplitude of annual POC flux in the coastal zone.  229  4  μmol/cm2/d  Trap 50m 3 Trap 115m 2  Trap 180m  1  Th-234 50m  0 Feb-09  Th-234 50m * May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  Apr-11  4  μmol/cm2/d  Trap 50m 3  Trap 115m  2  Trap 180m Th-234 115m  1  Th-234 115m * 0 Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  Apr-11  4  μmol/cm2/d  Trap 50m 3  Trap 115m  2  Trap 180m Th-234 160/180m  1  Th-234 160/180m * 0 Feb-09  May-09  Aug-09  Dec-09  Mar-10  Jun-10  Sep-10  Jan-11  Apr-11  Figure 5.38: Comparison of organic carbon fluxe time series obtained with sediment traps, equation 5.5 (colored dots) and equation 5.13 (black dots) from 238U seawater concentration estimates obtained from equation 5.6 for all cases.  230  POC fluxes - derived from equation 5.5 vs from sediment traps (μmol/cm2/d)  2.5  POC fluxes from  3.5  234Th  4.5 4  3 y = 1.0121x R²= 0.6166  2 1.5 1  0.5 0 0  0.5  1  1.5 2 2.5 3 3.5 POC fluxes from sediment traps (μmol/cm2/d)  4  4.5  POC fluxes - derived from equation 5.13 vs from sediment traps (μmol/cm2/d)  2.5  POC fluxes from  3.5  234Th  4.5 4  3 y = 1.0004x R²= 0.4536  2 1.5 1 0.5 0 0  0.5  1  1.5  2  2.5  POC fluxes from sediment traps  3  3.5  4  4.5  (μmol/cm2/d)  Figure 5.39: (a) POC fluxes estimated from equation 5.5 and  238  U seawater concentration estimated  from equation 5.6 versus POC fluxes from the sediment traps; (b) POC fluxes estimated from equation 5.13 and  238  U seawater concentration estimated from equation 5.6 versus POC fluxes from the  sediment traps. The data points influenced by high POC/234Th (115m in March-2010, November-2010, Feburary-2011 and 180m in October-2009) possibly reflect capture of zooplankton or other large particles on the filters are marked green. The surface data points (shown in Fig. 5.38), which show lower POC fluxes estimated from  234  Th than from sediment traps (excluding October-2009,  March-2010, April-2010 and December-2010) are marked blue. Those two sets of data are not included in the linear regression. 231  5.7 Conclusions and future perspectives. A 2-year time series of monthly sediment trap sampling provides a continuous record of POC flux which can be compared to concurrent estimates obtained from  234  Th deficits in  the water column. The organic carbon fluxes obtained by both methods are similar. In agreement with previous sediment trap experiments conducted in Saanich Inlet (Francois, 1988; Timothy et al., 2003), the present data indicate that POC fluxes vary in concert with surface water primary productivity, and particle fluxes are highest at mid-depth as a result of lateral transport of sediment winnowed from the sill by tidal currents. The present study also confirms a gradual increase in particle and POC flux from the head of the inlet towards the sill. The similarity of the fluxes measured by sediment traps and 234Th deficits does not prove but at least supports the validity of both approaches for measuring POC flux in coastal waters.  Compared more closely, the two approaches indicate that the agreement is better in deeper waters. At 50 m,  234  Th-based fluxes are lower than those measured by sediment traps,  which could in part be due to particle focusing resulting from the narrowing of the fjord with depth. On the other hand, the greater POC fluxes and better agreement between the two methods at 115 m is more consistent with lateral transport of sediment from the sill into the deep basin.  While these results are encouraging and supporting the use of either of these methods to document the sinking flux of POC or any other constituents of marine particles in the coastal zone, their applicability is still limited by the fact that they are both relatively 232  labor intensive and require substantial resources which can limit their application to large scale monitoring programs. The results from this study, however, suggest a possible alternative more readily applicable to larger surveys and monitoring programs, which might be necessary, for instance, to ground-truth satellite data. The simplified method that we propose stems from the observations that dissolved  234  Th activity in coastal waters is  very low and varies little with season and location. Moreover, a large fraction of coastal waters is associated with particles. Because dissolved  Th in  234  Th activities are low, it is  difficult to measure this quantity precisely by subtracting particulate Consequently, there is little point in measuring the total  234  234  Th from total 234Th.  234  Th activities of coastal waters,  which requires co-precipitation by manganese oxide. Instead, simply measuring particulate  234  Th by filtering a few liters of seawater and measuring the associated POC  suffice to estimate of the sinking flux of POC (or any other particle constituents). Although a few measurements of dissolved  234  Th might still be needed to establish its  approximate concentration in the region and/or season of interest, this simplification halves counting times and essentially eliminates the need for processing seawater samples. As such, the method becomes much more amenable to conducting larger scale monitoring or survey programs. In regions where sediment resuspension or lateral transport is important, export fluxes could be obtained by measuring samples collected in the mixed layer or the euphotic zone.  The other advantage afforded by the high scavenging intensity in coastal waters is that the closed, steady state model is generally applicable because there is little to be transported by advection and mixing and because  234  Th in seawater  234  Th activities are generally low 233  year-round. In the present study, the residence time of 234Th in the upper 100 m of the water column in Saanich Inlet varied between 7 and 14 days ([0.5 – 1 dpm/l] / [0.065 dpm/l.d-1]). The spacing between stations needed for a larger scale survey would depend on the lateral eddy diffusion coefficient in the region of study. The extent of lateral mixing varies widely in the coastal zone and it is beyond the scope of this chapter to address this question. However, a model tracking particles at 75 m depth in Georgia Strait suggests that particles would disperse over a radius of about 10 km over a 7 day period (E. Snauffer and S. Allen; pers. comm.). If this holds for Saanich Inlet, the  234  Th method may not be  able to resolve the particle flux gradient between the head and the mouth of Saanich Inlet that has been clearly documented with sediment traps in Timothy et al. (2003). In this case, the good agreement between the fluxes estimated from sediment traps and from  234  Th in  this study may be fortuitous and due to the location of the study site, where the flux measured by the sediment traps happens to be close to the average for the entire inlet. Measuring POC and  234  Th in particles collected along the main axis of the inlet would  answer this question.  234  Chapter 6 Particle fluxes and dynamics in the northeast Pacific Ocean from paired water column measurements of Th-230 and Th-234 activity  6.1 Introduction  Gaining a better understanding of the processes that control the sequestration of atmosphere CO2 to the deep sea is essential to predict the future evolution of atmospheric CO2 in response to human activities (combustion of fossil fuels, cement production, increasing land-use and deforestation). Large international research programs (e.g., JGOFS) have been established to address many aspects of this question and investigate the fluxes of particulate organic carbon (POC), carbonate, and biogenic silica (opal) to the bottom of the ocean. In particular, the deployment of sediment traps greatly improved our understanding of the geographic and vertical variations in particle flux and composition (Honjo et al., 2008). Nevertheless, many questions still remain, especially regarding the factors that control changes in particle flux, re-mineralization of organic matter, and dissolution of biominerals in the mesopelagic zone of the ocean (i.e.. the depth interval 235  between the bottom of the euphotic zone and ~ 1500 m). In this chapter, the activity of two thorium isotopes with widely different half-lives (234Th, 24.1 d;  230  Th, 76.7 ky) are  measured in seawater and particles of different sizes in the upper 1600 m of the water column at station Papa in an attempt to address some of these questions.  Thorium-234 and Thorium-230 are produced uniformly in seawater by decay of their parent uranium isotopes (238U and  234  U, respectively). Thorium is very insoluble in  seawater, and the two isotopes are quickly scavenged from the water column by adsorption on the surface of settling particles, resulting in lower seawater Th activities 234  compared to their parents. Because of its short half-life,  Th is mainly removed from  seawater by radioactive decay. Measureable 234Th deficits with respect to the activity of its parent  238  U require relatively high particle fluxes and  234  Th:238U disequilibria are more  commonly observed in the euphotic zone (e.g., Buesseler et al., 2006 and references therein). In contrast,  230  Th, with its much longer half-life (76.7 ky), is almost entirely  removed from seawater by scavenging and its seawater activity is always much lower than that of 234U. Combining thorium isotopes with different half-lives is commonly used to study particle dynamics and scavenging in the water column (e.g., Bacon and Anderson, 1982; Bacon et al., 1985; Clegg et al., 1991; Cochran et al., 1993; Murname et al., 1994; Marchal and Lam, 2012). Typically, the activities of several isotopes among  230  Th, 228Th, and 234Th are  measured in seawater and between two operationally defined classes of particle: small suspended particles (1 - 53 m), and large sinking particles (> 53 m; Fig. 6.1). 236  Figure 6.1: Model (A) used to describe thorium cycling in seawater. P = Th production in the water column from U decay; D = activity of dissolved Th; F = activity of thorium adsorbed on fine suspended particles; L = activity of thorium adsorbed on large sinking particles; λ = Th decay constant; K1 and K-1 are the adsorption and desorption rate constants between dissolved and fine particles; B 1 and B-1 are the aggregation and disaggregation rate constants between fine and large particles; S = sinking rate of large particles.  In this study,  234  Th and  230  Th were measured at several depths in the upper 1600m of the  water column (Table 6.1) at station Papa (50°N, 145°W) to try to estimate changes in the flux of particle constituents (organic carbon, P, Ca, Al) as a function of depth in the mesopelagic zone. The particle dynamics model reported in Fig. 6.1 was also systematically modified to fit the observations. Once the particle dynamics model was established, it was used to refine our understanding of changes in particle flux with depth and to assess possible ways of estimating particle flux and remineralization in the mesopelagic zone from thorium isotope measurements. Station Papa was chosen for this 237  exploratory study on the ground that the profiles of dissolved indicating that the impact of deep water circulation on the  230  Th are nearly linear,  230  Th profiles is small (e.g.,  Francois, 2007).  6.2 Materials and methods  6.2.1 Sample collection and preparation  The samples for this study were collected in June 2010 in the Northeast Pacific at Ocean Station Papa (OSP; 50ºN, 145ºW) in the Alaska Gyre. Thorium-234 and Thorium-230 activities were measured between 10 m and 1600 m in filtered seawater, fine (1-53μm) and large particles (>53μm) collected with large volume in-situ pumps, and a few “extra-large” particles collected with a 236 μm zooplankton MultiNet (Table 6.1). 6.2.1.1 234Th samples  Ten-liter sea water samples were collected at each depth (Table 6.1) using a rosette sampling system and drained into acid-cleaned containers.  238  239  Total 234Th: The methods described by Pike et al. (2005) and modified by Cai et al. (2006) were used. Two liters of sea water from each container were measured with a graduated cylinder and transferred into a Nalgene bottle. After adjusting the pH to 2 with 2N HCl, a pre-weighed and calibrated  230  Th spike (~1g, ~10dpm/g) was added to each sample and left to  equilibrate for 12 hours. Thereafter, the pH was raised back to 8 with concentrated NH4OH, and 100μl KMnO4 (3g/L) and 100μl MnCl2 (8g/L) were added to produce MnO2 for co-precipitation. The Nalgene bottles were shaken vigorously and left to equilibrate for 1hr, before heating in a water bath for an additional 1hr. The samples were then rapidly cooled in an ice bath and the MnO2 collected on tissue-quartz (TQ hereafter) filters (25mm diameter, 1μm pore size), oven-dried at ~50ºC and covered with LDPE film (1.11 mg/cm2) for beta counting. Because of the relatively weak activity of the 2 L samples, the samples were not covered with Al foil to shield possible beta emissions from contaminants and our 234Th counts may be slightly overestimated. Particulate 234Th: Several methods were used to collect particles of different sizes for 234Th analysis:  Small volume filtration (>1μm): After 2L of sea water were transferred to a Nalgene bottle for total  234  Th analysis, the volume of the remaining sea water (~8L) was  measured with a graduated cylinder and filtered on a pre-combusted (350ºC, 4hrs)  240  TQ filter (25mm diameter, 1μm pore size). The samples were then oven-dried at ~50º C and covered with LDPE film for beta counting. Large volume in-situ pump filtration: Large volume, in-situ pumps were used at several depths (Table 6.1) to collect particles on 142mm diameter, 1μm pore size Supor filters for  234  Th (and  230  Th) measurements. A Nylon mesh (53μm pore size)  was placed on top of the Supor filters to collect the larger particles. Generally, over 100 L of sea water was filtered during each deployment. The Supor samples (1μm-53μm) were dried at ~50ºC and a 25mm diameter sub-sample was punched from each filter, covered with LDPE film and beta counted. The large particles collected on the Nylon mesh were recovered by sonication in distilled water and filtration on 25mm diameter Supor filters (1μm pore size). These Supor filters (marked as “mesh” in Table 6.2 and hereafter) were then mounted with LDPE film for 234  Th beta counting (> 53μm).  MultiNet: Because large, rapidly sinking particles are very rare in the water column, it is very difficult to obtain a statistically meaningful sample even with large volume in-situ pumps. An opportunity arose from having on board a HydroBios MultiNet large plankton sampler, which was used to collect zooplankton samples and provided an opportunity to assess this sampling technique as a means of collecting large sinking particles from larger volumes of seawater. The HydroBios MultiNet was thus used in an attempt to collect “extra-large” (XL) sinking particles (>236 μm) at 5 depth intervals: 0-200 m, 200-500 m, 500-800 m, 800-1200 m and 1200-1600 m. This sampling method collected large, rapidly sinking particles from much larger volumes 241  of seawater (61m3, 84m3, 84m3, 111m3 and 119m3, respectively). The rate of ascent of the nets was slowed down to 0.5m/s to allow live zooplankton, which are normally captured by the device, to escape. Visual inspection of the samples confirmed the absence of large living zooplankton in the samples which appeared to consist mostly of passively settling larger aggregates. Even if some smaller zooplankton were caught in the net, they would not significantly affect the Th activity of the sample. Upon recovery, the samples were preserved with 200 l of saturated HgCl2 and stored at 4º C before analysis. Back in the laboratory, the samples were filtered on HCl-cleaned (2N, Seastar) Supor filters (25mm diameter, 1μm pore size). The samples were then oven-dried at ~50ºC and covered with LDPE film for beta counting. 6.2.1.2 POC The fine particulate  234  Th samples obtained by small volume filtration (1μm, on 25mm  TQ filters) were used for POC/PON analysis after completion of the  234  Th counting.  Particulate organic carbon on the LVP samples (1μm-53μm LVP Supor samples and >53 μm, Mesh samples) and XL samples (>236μm, Multinet) were estimated from the phosphorus content measured in the aliquots taken after the total digestion of the Supor filters (see below). 6.2.1.3 230Th samples Dissolved 230Th: Approximately 20L of sea water were collected at each depth (Table 6.1; 40 L were collected at depths shallower than 400m because of their low  230  Th activities)  with a rosette system and drained into acid-cleaned, pre-weighed cubitainers. The samples 242  were immediately weighed and acidified (~pH 2) with 2N HCl before adding 500mg of a 229  Th spike (1.5 dpm/g) in a solution of iron chloride (200mg, 50mg/g). After  equilibration (24 h), concentrated NH4OH was added to adjust the pH to 8–9 and co-precipitate  230  Th and  229  Th with iron hydroxide. The precipitates were redissolved in  the laboratory with HCl and Th was separated by ion exchange chemistry as described by Choi et al. (2001). Particulate  230  Th: Fine particles (1μm-53μm) collected by the LVPs on Supor filters  (142mm, 1μm) were completely digested in concentrated HF, HNO3 and HClO4 acids (Seastar) after addition of 200mg of a  229  Th solution (1.5 dpm/g) as a yield tracer. After  digestion and evaporation to a small drop, the samples were transferred to centrifuge tubes and topped-up to 10ml with 2N seastar HCl. Aliquots were taken to measure  232  Th, P, Ca  and Al, and the remaining of the samples passed through ion exchange resins for Th separation and subsequent measurement of 230Th (Choi et al., 2001).  After completion of  234  Th counting on the LVP mesh (>53 μm) and multiNet (>236 μm)  samples, they were completely digested in concentrated acid (Seastar HF, HNO3 and HClO4) after addition of 200mg of a  229  Th solution (1.5 dpm/g) as a yield monitor. After  digestion and evaporation to a small drop, the samples were treated as described above for the fine particles.  243  6.2.1.4 232Th, P, Al and Ca samples  After the total digestion of LVP Supor (1-53μm), LVP mesh (>53μm) and MultiNet samples (>236μm), the final solution was transferred to centrifuge tubes and topped-up to 10ml with 2N Seastar HCl. Aliquots (~1ml) were taken from each sample to measure P, Al and Ca. Separate aliquots (1ml) for spiked with  229  Th to correct  232  Th were taken from the LVP Supor samples and  230  Th activities for lithogenic contributions in fine particles  (A(230Thlitho)/A(232Th) = 0.8 dpm/dpm). The Al in LVP mesh samples (>53μm) and MultiNet samples (>236μm) were used to correct  230  Th activities for lithogenic  contributions in large and extra-large particles (A(230Thlitho)/Al = 462dpm/mol).  6.2.2 Sample analysis 6.2.2.1 234Th measurements  The  234  Th samples were counted on a gas-flow proportional five-position beta counter  manufactured by RISØ National Laboratories (Roskilde, Denmark). With approximately 10cm of low-radiation lead surrounding the counter, the background count rates were consistently low for the 5 positions (~0.05–0.09 cpm, Figure a, Appendix E) over the entire period of measurements (June/2010-May/2011).  The detector efficiency for the Mn oxide precipitates was determined by measuring the samples collected below 400m and assuming secular equilibrium with  238  U estimated  from salinity (Table a, Appendix E). The sample at 1000m appears to have been 244  contaminated by a beta emitter and it was thus not used for the calibration. The average obtained from all the samples was used to calculate the efficiency of the detectors and this calibration factor (0.558 ± 0.004) was used to calculate the total  234  Th activity of all  samples. Detector efficiencies for the particulate samples were measured using a standard 238  U/FeO filter following Rutgers van der Loeff and Moore (1999) and tested during the  GEOTRACES inter-calibration exercise (Maiti et al., 2012). All the samples were counted three times (1st counting shortly after sampling, 2nd counting after one month and 3rd counting after 8 months) for periods long enough to reach counting errors below 3%, and decay corrected to estimate the sample activity at the time of sampling (Buesseler et al., 2001). After the third counting, the filters for total yield tracer and  229  Th and  234  Th were digested with  229  Th as a  230  Th (the latter added during the co-precipitation) were  measured by ICP-MS to estimate the  234  Th recovery, following Pike et al. (2005). The  recoveries were generally above 90% (Table b, Appendix E).  6.2.2.2 POC measurements  Following the completion of  234  Th counting, at least 8 months after the initial sample  collection, POC and PON on 25mm TQ filters (fine particles, >1μm) were measured by gas chromatography on a model 1106 Carlo Erba CHN analyzer with a precision of ±1.3% for carbon and ±2% for nitrogen (Verardo et al., 1990). The POC on fine particles (collected on LVP Supor, 1-53μm), large particles (collected on LVP Mesh, >53μm) and XL particles (collected by multinet, >236μm) were estimated from the phosphorus measurements in the aliquots (1ml) taken after the complete digestion of the Supor filters. 245  6.2.2.3 230Th measurements  After separation of Th by anion-exchange, the samples were evaporated to a small drop and fumed with 5ml of concentrated HClO4 for 2 hours to remove organic matter that leached from the anion-exchange resin. After fuming, the samples were taken up in 2N Seastar HNO3 for ICP-MS measurements of 230Th and 229Th. 6.2.2.4 232Th, P, Al, and Ca measurement  The aliquots (1ml, in 2N Seastar HCl) for P, Al and Ca, taken from the final solution (topped-up to 10ml with 2N seastar HCl) after the total digestion of LVP Supor (1-53μm), LVP mesh (>53μm) and multinet samples (>236μm), were evaporated and taken up in 1ml 2N Seastar HNO3. They were then diluted by a factor of 10 to 50 (LVP Supor samples above 300m were diluted 50 times and all other samples were diluted 10 times) with 2N Seastar HNO3 and measured by ICP-MS. Standards for P, Al and Ca were prepared by diluting the 1000ppm standards to 1ppt - 1ppm with 2N Seastar HNO3. The aliquots for 232  Th from the LVP Supor samples were spiked with 100mg  229  Th solution (3950 dpm/g),  transferred to 2N Seastar HNO3 and then measured on an ICP-MS by isotope dilution.  246  247  248  6.3 Results  The  234  Th and  230  Th activities of all seawater and particle samples are reported in Table  6.2. 6.3.1 234Th  The vertical profiles of total  234  Th and  238  U (Figure 6.2) clearly exhibit a  above 80m produced by scavenging. There is also an excess attributed to particle re-mineralization and release of  234  234  Th deficit  234  Th below 80m, which is  Th to seawater.  234  Th:238U  equilibrium is assumed below 350m (within the error of measurements), except at 1000m where the sample appears to have been contaminated by a beta emitter. This sample is thus excluded from further discussion.  Particulate  234  Th activities measured by small volume filtration on 25mm TQ filters  (>1μm) decrease sharply between 60 m and 80 m and more gradually below that depth (Fig. 6.2). There are also minima at ~100m and 200-300m, which coincide with excesses in total 234Th. Fine particulate 234Th activities collected by LVPs on Supor filters (1-53 μm) generally agree with niskin samples but are systematically slightly lower (Figure 6.3). This could, in part, be explained by the collection of larger particles (> 53 μm) during small volume filtration. However, large particles are rare in small volumes, and adding the 234  Th activity measured on the Supor and mesh samples at each depth does not quite 249  account for the values measured when filtering small volumes of seawater on TQ filters. The difference is small and within the error bars of the measurements, but the consistent offset (Fig. 6.3) suggests a small but real difference between the two sampling techniques.  234Th  0.00 0  at OSP June/2010 (dpm/l)  0.50  1.00  1.50  2.00  2.50  200 400  Depth (m)  600 800 1000 U-238  1200 Th-234 Total 1400 Th-234 Particulate (25mm TQ) 1600  Figure 6.2: Total  234  Th, particulate  234  Th (obtained by small volume filtration) and  238  U profiles  (calculated from salinity) at OSP (June/2010).  The  234  Th activities of large particles collected by LVPs on mesh (>53 μm) increase  between 60m and 100m in contrast to the sharp decrease observed with fine particulate 234  Th in this depth interval (Fig. 6.3, 6.4). Below 100m, their  decrease with depth. The  234  Th activities generally  234  Th activities of extra-large (XL) particles collected by  multiNet (>236 μm) are two orders of magnitude lower than the activities measured on the large particles collected with the LVPs (Fig. 6.4). They increase slightly from the 0-200m interval to the 200-500 interval and then decrease with depth. 250  Particulate 234Th at OSP (dpm/l) 0  0.2  0.4  0.6  0.8  1  1.2  0 200 400  Depth (m)  600 800  Th-234 (25mm TQ)  1000  Th-234 (LVP Supor)  1200  Th-234 (LVP Mesh)  1400  Th-234 (Supor+Mesh)  1600  Figure 6.3: Profiles of fine particulate 234Th obtained by small volume filtration on 25mm TQ filtration (>1μm, green triangles), by LVPs on Supor filters (1-53 μm, blue diamonds), and large (>53 μm) particulate  234  particulate  234  Th collected with LVPs on Nylon mesh (brown dots). The sum (>1 μm) of fine  Th from LVP Supor filters (1-53 μm) and large particulate  234  Th from LVP Mesh (>53  μm) are marked as black dots.  Particulate 0.02  234Th 0.04  Particulate 234Th at OSP (dpm/l)  at OSP (dpm/l) 0.06  0.08  0  0.1  0  0  200  200  400  400  600  600  Depth (m)  Depth (m)  0  800  1000  0.0001  0.0002  0.0003  0.0004  0.0005  800  1000  Th-234 XL MultiNet  Th-234 (LVP Mesh) 1200  1200  1400  1400  1600  1600  Figure 6.4: Profiles of 234Th in large particles (>53μm, collected on LVP Mesh) and “extra-large” (XL) particles (>236μm, collected by MultiNet). 251  252  6.3.2 POC  The concentrations of POC in fine particles were measured directly on the 25mm TQ filters (>1μm). Particulate organic carbon on the LVP Supor filters (1-53 μm), the LVP mesh samples (> 53 m) and the multiNet samples (> 236 m) were estimated from their P content, assuming a C/P ratio of 106.  POC concentrations obtained from the 25mm TQ samples decrease with depth and drop rapidly from 60m to 100m (Table 6.3, Fig. 6.5), which coincides with a total 234Th excess (Fig. 6.3), further indicating intensive remineralization below the euphotic zone. Particulate organic carbon concentrations estimated from the P content of the LVP Supor samples and a Redfield ratio of 106 are similar to those measured on the 25mm TQ filters below 600m but significantly higher at shallower depths, especially above 100m (Fig. 6.5). If the two filtration methods sample the same pool of particles, this discrepancy suggests that the C/P of organic matter at shallow depth at station Papa is significantly lower than the canonical Redfield ratio of 106, and that C/P increases with depth. If we assume that the particles collected on the 25mm TQ filters and the LVP Supor have the same POC/234Th, we can also estimate POC concentration in the LVP-Supor samples from their 234  Th activities and estimate the C/P ratio of the particles (Fig. 6.6). These estimates show  the increasing trend in C/P with depth, which is consistent with the preferential remineralization of P over C. However, the C/P estimated for shallow waters appears unrealistically low.  253  POC (μg/l) 0  20  40  60  POC (μg/l) 80  100  0.00 0  200  200  400  400  600  600  800  25mm TQ  1000 1200 1400  Depth (m)  Depth (m)  0  2.00  4.00  6.00  8.00  800  25mm TQ 1000  LVP Supor  1200  LVP Supor  LVP Mesh  1400  LVP Mesh  1600  1600  Multinet  Multinet 1800  1800  POC/234Th (μg/dpm) 0  20  40  60  80  100  120  0 200 400  Depth (m)  600  25mm TQ  800 LVP Supor 1000  LVP Mesh 1200 1400 1600 1800  Figure 6.5: POC profiles and POC/234Th for the particles collected on 25mm TQ filters, the fine particles collected by LVP on Supor filters (1-53 μm), and Nitex mesh (>53μm). POC for LVP samples were estimated from the P concentrations multiplied by the Redfield ratio (106).  The POC concentrations estimated from the P content of the large particles (>53 μm) collected on Nylon mesh (Table 6.4), assuming C/P = 106, are lower than the POC associated with fine particles but show a similar decreasing trend with depth (Fig. 6.5, Table 6.3). The POC concentrations estimated from the P content of the extra-large (XL) 254  particles (>236 μm) collected with the MultiNet (Table 6.4), assuming C/P = 106, decrease with depth gradually from 100m to 1000m (Fig.6.5). They are similar to the POC concentration associated with large particles (mesh samples) below 1000m but significantly higher above 1000m.  C/P (mol/mol) 0  50  100  150  0 200 400  Depth (m)  600 800 1000 1200 1400 1600 1800  Figure 6.6: Redfield ratio derived from the 25mm TQ and LVP Supor data by assuming the POC/ 234Th are the same on those two different types of fine particles  6.3.3 POC/234Th The POC/234Th ratios measured directly on the 25mm TQ filters decrease from a maximum of ~ 50 g C/dpm in surface water and quickly drop below 20 g C/dpm in the upper 100 m of the water column (Fig. 6.5, Table 6.3). These results are similar to those obtained by Charette et al. (1999).  255  To estimate the POC/234Th ratios of the LVP-Supor samples, we can either assume that they are similar to the ratios measured directly on the 25mm TQ filters (this is the assumption used to estimate C/P in Fig. 6.6) or we can assume that C/P = 106 at all depth, in which case, we find a good agreement with the TQ filter measurements below 600 m but significantly higher values in shallower water (Fig. 6.5). The POC/234Th ratios of the large particles (>53μm, LVP Mesh) estimated from their P content and C/P = 106 are similar to the values measured on the TQ samples in the upper 400 m and only slightly lower in deeper waters (Fig. 6.5). On the other hand, they are lower than the POC/234Th on the LVP Supor samples estimated from P concentration and C/P = 106. The POC/234Th ratios of the XL particles (>236μm, Multinet) estimated from their P content and C/P = 106 are hundreds of times higher than ratios of fine and large particles (Table 6.3). 6.3.4 230Th Both dissolved and fine particulate (1-53μm) 230Th activities increase almost linearly with depth (Fig. 6.7). The fraction of total 230Th associated with fine particles increases linearly from about 10% at the surface to over 20% at 1600m (Fig. 6.8), with a higher ratio at 60 m. The large particulate (>53μm)  230  Th activities, however, don't follow this linear  increase trend but display a maximum at 400m (Fig. 6.9). The XL particulate (>236 μm) 230  Th activities are three orders of magnitude lower and below detection limit for the  256  samples taken from the 0-200m and 200-500m depth intervals. Below 500 m, they become measureable and increase with depth (Fig. 6.9).  230Th  0  0.05  0.1  at OSP June/2010 (dpm/T) 0.15  0.2  0.25  0.3  0.35  0 Th-230 Dissolved  200  Th-230 Fine particulate (LVP Supor)  400  Th-230 Large particulate (LVP Mesh)  Depth (m)  600 800 1000 1200 1400 1600  Figure 6.7: Profiles of dissolved  230  Th, 230Th on fine particles (1-53 μm, collected on LVP Supor) and  large particles (>53μm, collected on LVP mesh).  Particulate 0.0%  230Th/Total 230Th  20.0%  40.0%  0 200 400  Depth (m)  600 800 1000 1200 1400 1600 1800  Figure 6.8: The fraction of total 230Th associated with fine particles. 257  Particulate  230Th  Particulate 230Th at OSP June/2010 (dpm/T)  at OSP June/2010 (dpm/T)  0  0.05  0  0.1  0  0  200  200  0.00002  0.00004  400  400  Fine (LVP Supor) 600 Large (LVP Mesh) 800 1000  Depth (m)  Depth (m)  600  800 1000  1200  1200  1400  1400  1600  1600  XL (multinet)  Figure 6.9: Profiles of 230Th on fine (1-53 μm collected on LVP Supor filters), large (>53 μm, collected on LVP mesh) and XL particles (>236 μm collected on MultiNet).  POC/230Th (mmol/dpm)  POC/230Th (mmol/dpm) 0  500  0  1000  0  0  200  200  400  400  LVP Supor LVP Mesh  Depth (m)  Depth (m)  20  30  40  600  600 800  10  LVP Supor 800  1000  1000  1200  1200  1400  1400  1600  1600  LVP Mesh  Figure 6.10: POC/230Th for particles collected by LVP on Supor filters (1-53 μm), and Nitex mesh (>53μm). POC for LVP samples were estimated from their P concentration and the Redfield ratio (106).  258  259  6.3.5 POC/230Th The POC/230Th ratios on the fine (LVP-Supor), large (LVP-Mesh) and extra-large (MultiNet) samples (Fig. 6.10) were calculated by assuming C/P = 106 at all depths (Table 6.5). The POC/230Th ratios in fine particles decrease with depth from 1300mmol/dpm at 20 m to 1.8 mmol/dpm at 1600 m (Fig. 6.10), consistent with increasing POC degradation and 230Th scavenging with depth. The POC/230Th ratios on large particles increase sharply between 60 m and 100 m and then generally decrease with depth. They are substantially lower than the ratios measured in fine particles at shallower depth and then converge in deeper waters. This is in contrast to POC/234Th ratios, which are more similar for the two particle sizes (Fig. 6.5). One way to explain this trend in POC/230Th is by invoking formation of the large particles in the upper 1000 m by aggregation of fine particles and consumption of a larger fraction of their organic matter by heterotrophs before sinking, compared to the fine particles at the same depth. The size of the larger particles may be why they tend to be preferentially consumed. The similarity in the trend of POC/234Th for the two particle pools could be explained by decomposition/consumption rates of POC in these large particles of the same order as the decay rate of  234  Th. On the other hand, the POC/230Th  ratios on the extra-large particles are several hundred times higher than POC/230Th ratios on fine particles (Table 6.5). They also decrease sharply with depth, from ~1300 nmol/dpm at 600 m to ten times lower at 1400 m. Their much higher POC/ 230Th ratios  260  compared to the smaller particles indicate that these extra-large particles are not formed by aggregation of smaller ones from the same depth horizon, even in the upper 200 m. The total amount of POC collected with the MultiNet between 0 – 200 m was ~ 10 mmoles (1.772 mgC/m3 x 61 m3). If these extra-large particles were formed by aggregation of small particles with a POC/230Th ratio of ~ 200 nmoles/dpm (Table 6.5), their  230  Th activity should have been easily measureable, and not below detection limits,  as we found. Instead, the extra-large particles we collected with the MultiNet must have been produced directly by marine organisms and their low POC/230Th must reflect their low surface area to volume ratio (Buesseler et al., 2006). This conclusion is however in part based on the assumption that the slow rate of ascent of the MultiNet allowed all zooplankton to escape and the material recovered consisted only of passively sinking large particles. The technique was used without pre-planning and prompted by serendipity. Although a cursory examination of the samples seems to indicate that living zooplankton were largely absent, a much more systematic study of the technique is still required to fully evaluate this relatively simple sampling approach and the validity of this conclusion.  6.3.6 P and Ca The vertical profiles of fine (1- 53 m) particulate P and Ca are very similar. Both decrease sharply in the upper 100 m of the water column (Fig. 6.11, Table 6.4). While particulate P concentration decreases sharply between 20 m and 60 m, however, the largest decrease in particulate Ca occurs between 60 m and 100 m. Likewise, the 261  concentration of both elements decreases with depth in the large particles (> 53 m), particularly down to 400 m depth (Fig. 6.11). In contrast, the concentration profiles of the two elements in the XL particles (> 236 m) collected with the multiNet are very different. The P concentrations associated with XL particles are higher than the P concentrations associated with large particles, and decrease gradually down to 1000 m depth. On the other hand, the Ca concentrations associated with XL particles is much lower than the concentrations associated with the large particles (Fig. 6.11).  6.3.7 Al  The Al concentrations associated with fine particles show a systematic trend with higher concentration towards the surface, a minimum at 300 m, and a slight increase toward 1600m. The Al concentrations associated with the XL particles are ten times lower (Fig. 6.11), and display a similar trend as for the fine particles, but with a deeper minimum at ~ 1000 m. In contrast, the Al concentrations in the large particles collected by LVP do not show a clear trend with depth, and they are generally slightly higher than for the XL particles (Fig. 6.11).  262  Fine particulate P (mmol/m3) 0.02  0.04  0.06  0.08  L and XL particulate P (mmol/m3) 0.1  0  0  0  200  200  400  400  600  600  800  LVP Supor  1000  Depth (m)  Depth (m)  0  1400  1400  1600  1600  1800  1800  Fine particulate Ca (mmol/m3)  L and XL particulate Ca (mmol/m3) 0  1.5 0  200  200  400  400  600  600  800  LVP Supor  1000  1000  1200 1400  1400  1600  1600  1800  1800  Fine particulate Al (mmol/m3) 0.002  0.004  0.006  0.06  0.08  0.1  Multinet  0  200  200  400  400  600  600  LVP Supor  Depth (m)  0  1000  0.04  L and XL particulate Al (mmol/m3) 0.008  0  800  0.02  LVP Mesh  800  1200  0  Depth (m)  Multinet  0  Depth (m)  Depth (m)  1  0.0015  LVP Mesh  1000 1200  0.5  0.001  800  1200  0  0.0005  800 1000  1200  1200  1400  1400  1600  1600  1800  1800  0.0001 0.0002 0.0003 0.0004 0.0005 0.0006  LVP Mesh  Multinet  Figure 6.11: P, Ca, and Al concentrations associated with fine, large and XL particles. 263  6.4 Discussion  6.4.1 Th-234 fluxes Integrating the 234Th deficit over the depth where it is measureable gives the removal rate of 234Th from the upper water column, expressed in dpm/m2.d (see Chapter 5): 𝑥  FluxThz=λ234∫0 (A238 − A234) dz  (6.1)  where z is the depth of integration. Applying equation 6.1 to the 234Th activities measured on the TQ samples at station P (Fig. 6.2) provides the flux of (Fig. 6.12, Table 6.6). The  234  Th as a function of depth  234  Th flux increases systematically in the upper 80m, where a  deficit is found, then decreases gradually to a depth of 350 m where particle remineralization releases  234  Th back to the water column, and stays seemingly constant  below. The latter observation stems from the fact that the beta counter was calibrated based on the assumption that the samples below 400m are in secular equilibrium. However, there is also a limit on how deep in the water column this calculation can be applied meaningfully. Given the precision of our  234  Th measurements, we cannot  distinguish secular equilibrium from 234Th excesses or deficits of ±0.05 dpm/L. Therefore, below the depth of measurable deficit or excess (in the presence case, below 300 m), the uncertainty added to the flux of  234  Th to the seafloor is 50 dpm/m3 x 0.0288 y-1 for every  vertical meter, i.e. an added uncertainty of 144 dpm/m2.d per 100m. Since the 234Th flux at 300 m is 500 dpm/m2.d ± 50%, the uncertainty on the flux scavenged from or released to  264  deeper waters very quickly overwhelms the  234  Th flux originating from the upper water  column, as illustrated by the increasing error bars with depth reported on Fig. 6.12.  Equation 6.1 also assumes that the surface water at station OSP is a closed system at steady-state. Since there is no distinct spring bloom at OSP (Harrison, 2002), this assumption should be reasonably valid but cannot be verified due to lack of  234  Th data  from the previous month.  Th-234 fluxes (dpm/m2/day) 0  500  1000  0 200 400  Depth (m)  600 800 1000 1200 1400 1600 1800  Figure 6.12: Th-234 fluxes based on 234Th:238U deficit  265  266  6.4.2 Fluxes of POC, Ca and Al estimated from the 234Th:238U deficit  The fluxes of any particle constituent X can be estimated by multiplying the  234  Th fluxes  obtained from equation 6.1 by the X/234Th ratio of the particles that sink from the surface and scavenge 234Th: FluxXz= FluxThz *(CXz/CThz)  (6.2)  Where FluxXz and FluxThz are the sinking fluxes of X and  234  Th at depth z, CXz and CThz  are the concentrations of X and the activities of 234Th in the sinking particles at depth z.  6.4.2.1 POC fluxes  Applying equation 6.2 to POC measured on the 25mm TQ samples (Fig. 6.5) provides the flux of POC as a function of depth (Fig. 6.13, Table 6.6). POC fluxes at station Papa gradually increase from the surface to 60 m and then decrease with depth due to organic matter remineralization, a pattern that has been observed before (e.g., Bacon et al., 1996). The flux estimates from this study are similar to those obtained at OSP in May 1996 by Charette et al. (1999).  267  POC fluxes (mg/m2/day) 0  10  20  30  40  0 200 400 600 Depth (m)  25mm TQ  800 Trap Winter  1000 Trap Spring  1200 Trap Summer  1400 Trap Fall  1600 Trap Annual  1800  Figure 6.13: POC fluxes derived from the 234Th flux and POC/234Th ratio on the 25mm TQ filters. The triangles show the POC fluxes from sediment traps reported by Timothy et al., (submitted) at the same location. The smaller black, yellow, purple and pink triangles stand for the winter (20/November-19/February), spring (20/February-20/May), summer (21/May-19/August) and fall (20/Augest-19/November) trap fluxes, respectively. The flux for each season is averaged over the corresponding sampling period. The annual flux (larger blue triangle) is the average of the flux over their entire sampling period (from May/1989 to June/2006 for the trap at 200m and from March/1983 to June/2006 for the trap at 1000m).  POC fluxes were also measured at OSP with sediment traps deployed at 200m and 1000m (Timothy et al. submitted) between March/1986 and Jun/2006. The  234  Th-based POC  fluxes agree reasonably well with these sediment trap results (Fig. 6.13). At 200 m, the 268  234  Th-derived flux is somewhat lower than the sediment trap flux collected in spring  (20/Feb –20/May, Timothy et al., submitted). The POC flux estimated by the  234  Th  method at 1000 m is very close to the flux measured by the sediment trap in spring. This agreement is however fortuitous, considering the large uncertainties in the fluxes derived from  234  Th at this depth (Fig. 6.12). Nonetheless, the  234  Th method provides a maximum  POC flux at 1000 m consistent with the sediment trap data.  The POC fluxes estimated from the LVP-Supor  234  Th activities and P concentrations with  a constant Redfield ratio of 106 are larger in the upper 100m than those obtained from direct POC and 234Th measurements in 25mm TQ filters but converge at 200m (Fig. 6.14). This difference arises from uncertainties regarding the C/P ratio of particulate organic matter. Because the POC/234Th ratios of the large particles collected by LVP on Nylon mesh (>53μm) are very similar to those measured on the TQ samples (Fig. 6.5), the large particles necessarily provide flux estimates that are also very similar to the estimation based on 25mm TQ (>1μm) (Fig. 6.14).  269  POC fluxes (mg/m2/day) 0  10  20  30  40  50  60  70  0 50 100  Depth (m)  150 200  25mm TQ  250 LVP Supor  300 350  LVP Mesh  400  Figure 6.14: POC fluxes derived from the POC/234Th ratio on all three samples.  6.4.2.2 Ca fluxes The flux of biogenic Ca is calculated based on equation 6.2 and the Ca bio/234Th ratio on both fine (LVP Supor) and large (LVP Mesh) samples (Tables 6.7, 6.8). Cabio was calculated by estimating lithogenic Ca from the Al content (Ca/Al = 0.3mol/mol) and subtracting it from the total particulate Ca. Biogenic Ca flux estimated from the fine particles shows a maximum at 60m, as for the POC flux. It quickly decreases from 60 m to 300m and stays almost constant (within large error bars) below 300m (Fig. 6.15).  270  Biogenic Ca flux (mmol/m2/d) 0  0.5  1  1.5  2  2.5  3  0 200 400  Depth (m)  600 800  LVP Supor  1000  LVP Mesh  1200  Trap Spring  1400  Trap Summer 1600 1800  Figure 6.15: Biogenic Ca fluxes derived from the biogenic Ca/234Th ratio on LVP-Supor and mesh samples. Triangles show the fluxes measured with sediment trap (Timothy et al., submitted).  The sharp decrease in biogenic Ca flux between 60 m and 300 m can be attributed to biologically-mediated carbonate dissolution resulting from CO2 released by organic matter remineralization in zooplankton guts or particles aggregates (e.g. Milliman et al., 1999). Biogenic Ca flux obtained from the large particles (LVP Mesh) shows a similar trend but with somewhat higher fluxes above 400m (Fig. 6.14). The fluxes of biogenic Ca derived from  234  Th are similar to fluxes measured with sediment traps (Timothy et al.,  submitted). The latter also indicate a nearly constant carbonate flux between 200m and 1000m. 6.4.2.3 Al fluxes The flux of Al is calculated based on equation 6.2 and the Al/234Th ratio on both fine (LVP Supor) and large (LVP Mesh) samples (Tables 6.7, 6.8). The Al fluxes estimated from both 271  fine and large particles are very low (Fig. 6.16) and consistent with aeolian dust flux estimated for this region (Mahowald et al., 1999; 0.1-0.2 g dust / m2.y; ~ 10% Al in dust; aeolian Al flux = ~ 0.001 – 0.002 mmol/m2.d). The Al fluxes associated with fine particles increase between 20 m and 100 m suggesting lateral transport of fine particles from the shelf (Lam and Bishop, 2008). Below 200 m, Al fluxes decrease and appear to remain constant within the large error bars of the method at these depths. This would suggest removal of fine Al-containing particles by aggregation below 100m. The Al fluxes obtained from the large particles are similar and uniformly low. If the decrease in Al fluxes associated with small particles below 100 m is due to aggregation, the aggregates involved must have been very large and not effectively sampled by the LVP.  The very  low Al fluxes estimated here are also consistent with the sediment trap study at station Papa, which report “negligible” lithogenic fluxes (Timothy et al., submitted).  Al flux (mmol/m2/d) 0  0.002  0.004  0.006  0.008  0.01  0.012  0  Depth (m)  200 400  LVP Supor  600  LVP Mesh  800 1000 1200 1400 1600 1800  Figure 6.16: Al fluxes derived from the Al/234Th ratio on LVP samples. 272  273  274  275  6.4.3 230Thxs fluxes The 234Th:238U method provides an effective means for estimating POC or other elemental sinking fluxes in the water column, as long as the scavenging intensity is large enough to generate a 234Th deficit that can be measured precisely. As discussed above, because  234  Th  has a very short half-life, a measureable deficit is often restricted to the upper water column. Thorium-234 is thus a powerful tool to estimate export flux from surface water (Buesseler et al., 2006) but has very limited applicability to deeper water (with the exception of the bottom nepheloid layer, e.g. Rutgers van der Loeff and Bacon, 1989). With its much longer half-life (~76.6 ky) and short residence time in the water column (ca. 30 years), 230Th decay in seawater is very limited and most of the 230Th produced from the decay of 234U must be removed by particle scavenging. As a result, there is always a very large deficit between the activity of total  230  Th in seawater and that of its parent  fact, the deficit is so large that, to a very good approximation, the decay of neglected and therefore the flux of  234  U. In  230  Th can be  230  Th scavenged from the water column (FluxThxsz) is  very nearly equal to the rate of production in the water column: 𝑥  FluxThxsz=λ230∫0 (A234 − 0) dz  where A234 is the activity of activity of  (6.3)  234  U in seawater and z is the depth of integration. Since the  234  U is essentially constant with depth (2910 dpm/m 3), the flux of  230  Thxs is  thus simply proportional to water depth:  FluxThxsz=λ230 A234 z = 0.0267 z dpm/m2.y  (6.4) 276  Some deviations from this simple equation could arise as a result of lateral transport of dissolved  230  Th by advection or turbulent mixing (e.g., Francois, 2007). However,  observations and modeling studies indicate that equation 6.4 is accurate to within 30% over most of the ocean (Henderson et al., 1999; Yu et al., 2001, Luo et al., 2010). This equation has been widely used in sediments to estimate past changes in the rain rate of particles reaching the seafloor (e.g., Francois, 2004) but never applied systematically to water column particles. 6.4.4 Fluxes of POC, Ca and Al estimated by normalization to 230Thxs  Because the  230  Th deficit is always very large and easy to quantify, the principles applied  to estimate particle flux in the upper water column from the  234  Th deficit could be  similarly applied to 230Th over the entire water column. FluxXz= FluxxsTh-230z *(CXz/CxsTh-230z)  (6.5)  where FluxXz and FluxxsTh-230z are the sinking fluxes of X (X=P, Ca, Al, etc.) and scavenged flux of  230  Th at depth z, CXz and CxsTh-230z are the concentrations of X and the  activities of “excess”  230  Th in the sinking particles at depth z. “Excess”  230  Th is the total  activity of 230Th in marine particles corrected for 230Th in secular equilibrium with 234U in the lithogenic fraction of the particles (Table 6.9).  The main difficulty in applying this method is to measure the very low  230  Th  concentrations associated with particles (CxsTh-230z) particularly at shallow depths, which 277  requires the use of large volume in-situ pumps and, as for  234  Th, establishing the  appropriate (CXz/CxsTh-230z) to estimate the flux of X.  6.4.4.1 POC fluxes  As the first step, the POC flux was calculated by multiplying the  230  Th fluxes by the  POC/230Thxs ratio estimated from the P content and the Redfield ratio (106), and the  230  Th  activities measured on the fine particles collected by LVP on Supor filters (Fig. 6.17): FluxPOCz= FluxThxsz *(CPz/CThxsz)*106  (6.6)  POC fluxes (mg/m2/day) 0  5  10  15  20  25  30  35  0 200 400 600 Depth (m)  Th-230 norm LVP Supor  800  Th-230 norm LVP Mesh Th-234 25mm TQ  1000  Trap Winter  Trap Spring  1200  Trap Summer  1400  Trap Fall Trap Annual  1600 1800  Figure 6.17: POC fluxes derived by  230  Th normalization on fine particles compared with the POC  fluxes estimated by other means as in Figure 6.12. 278  The POC fluxes derived by this approach are similar to the POC fluxes estimated from 234  Th (Fig. 6.5). The  230  Th-derived POC fluxes also display a maximum at 60m and a  rapid decrease to 200m. The fluxes derived from this method also indicate a gradually decreasing POC flux from 300m to 1600m and at 200 m and 1000 m they are in reasonable agreement with the annually-average POC sediment trap flux measured at 1000 m. At 200 m depth, the residence time of  230  Th in seawater is approximately 2 years  (0.05 dpm/m3 / 0.0267 dpm/m3.y-1). The somewhat higher annually averaged fluxes measured by the sediment trap at 200 m may thus reflect inter-annual variability. At 1000 m, the residence time of 230Th increases to 0.25 dpm/m3 / 0.0267 dpm/m3.y-1 = ~ 10 years, consistent with the better agreement between trap flux (averaged between 1982 and 2006; Timothy et al., submitted) and suggest that  230  Th-normalization at this depth (Fig. 6.17). These results  230  Th normalization on fine particles may provide an effective means to  estimate the POC fluxes in the deep water column where the  234  Th method cannot be  applied.  If the fluxes thus calculated are accurate, the extent of POC remineralization between sampling depths could also be readily calculated by difference, providing crucial data on the efficacy of the biological pump. However, particles with larger size are usually considered to be the major contributor of the POC flux to the deep sea because they have much faster sinking velocity compared to the fine particles. With our data, we can also test 230  Th  normalization  on  the  large  (LVP  Mesh,  >53μm)  and  XL  particles  (MultiNet, >236μm) to estimate POC flux. The POC fluxes obtained from the large particles, however, are more than 5 times lower than that from the fine particles and are 279  evidently lower than the  234  Th-derived estimates or the sediment trap records (Table 6.9,  Fig. 6.13). This is because their POC/230Th ratios are lower than the POC/230Th of fine particles (Fig. 6.10). On the other hand, the POC fluxes obtained from the XL particles are more than 200 times higher and much greater than the trap records (Table 6.9). Clearly,  230  234  Th estimation or the sediment  Th-normalization cannot be applied to the extra-large  particles. In fact, the POC/230Th ratio that must be applied to correctly estimate the flux of POC by 230  Th-normalization (and also the export fluxes estimated from the  234  Th deficit) is the  flux-weighted average of all particle sizes that contribute to the sinking flux. Such an average could be obtained from the analysis of material collected by sediment traps that would work perfectly, i.e., traps that would intercept without biases the settling flux in all class sizes. The present observation that  230  Th-normalization on fine particles seems to  work (or at least agree with sediment traps and the  234  Th method) is thus likely fortuitous  and reflects the fact that the POC/230Th of these fine particles is similar to the POC/230Th flux-weighted average for the large (low POC/230Th) and extra-large (high POC/230Th) particles intercepted by the traps.  6.4.4.2 Ca fluxes  The biogenic Ca fluxes were also calculated by multiplying the  230  Th fluxes by the  biogenic Cabio/230Thxs ratio on both fine (LVP Supor) and large (LVP Mesh) particles: FluxCabioz= FluxThxsz *(CCabioz/CThxsz)  (6.7) 280  The biogenic Ca fluxes calculated from the fine (LVP Supor) particles increase from 20m to reach the maximum at 60 to 100 m. They decrease quickly from 100m to 200m and then much more slowly from 200m to 1600m (Fig. 6.18). Taken at face value, these fluxes would indicate about 50% carbonate dissolution between 100 m and 200 m depth, consistent with other lines of evidence (Milliman et al., 1999, Feely et al., 2002, Antia et al., 2008). The  230  Th-normalized biogenic Ca fluxes are also reasonably consistent with  biogenic Ca fluxes derived from  234  Th:238U (Table 6.8; although the flux estimated at 60  m is lower, possibly reflecting the difference in the integration time of the two proxies) and sediment traps (Timothy et al., submitted). The fluxes of biogenic Ca calculated from the large (LVP Mesh) particles are smaller. On the other hand, the Ca fluxes obtained by 230  Th-normalization on the extra-large particles are much larger, as was the case for POC  fluxes. They are also very variable because of the very low carbonate concentration in these particles (Fig. 6.11). Again, at first sight, these results suggest that  230  Th  normalization on fine particles may provide an effective means of estimating the biogenic Ca fluxes in the water column, and carbonate dissolution in the mesopelagic zone but, as for POC, this apparent success is likely fortuitous.  281  Biogenic Ca flux (mmol/m2/d) 0  0.2  0.4  0.6  0.8  1  1.2  1.4  0 200 400  230Th (LVP Supor)  Depth (m)  600  230Th (LVP Mesh)  800  234Th (LVP Supor) 1000  234Th (LVP Mesh) 1200  Trap Spring  1400  Trap Summer  1600  Trap Annual  1800  Figure 6.18: Biogenic Ca fluxes derived by  230  Thxs normalization compared to those derived by  234  Th:238U and sediment traps.  6.4.4.3 Al fluxes The Al fluxes calculated with equation 6.5 and the Al/230Thxs ratio on fine particles (LVP Supor) are very similar to those estimated with  234  Th (Fig. 6.19) and also record a  prominent maximum at 100m and a nearly constant flux below. These deeper fluxes can now be estimated with much smaller uncertainties than from  234  Th and indicate the Al  fluxes are essentially constant with depth to 1,600m. Fluxes obtained from the large particles (LVP Mesh) are generally lower.  282  Al flux (mmol/m2/d) 0  0.002  0.004  0.006  0.008  0.01  0.012  0.014  0 200 400  Depth (m)  600 800  230Th (LVP Supor)  1000  230Th (LVP Mesh) 1200  234Th (LVP Supor) 1400  234Th (LVP Mesh)  1600 1800  Figure 6.19: Al fluxes derived by 230Thxs normalization compared to those derived by 234Th:238U.  6.4.5 Particle dynamics  The above observations suggest that  230  Thxs normalization on fine suspended particles  could provide accurate flux estimates down the water column over the entire depth range of the mesopelagic zone and deeper. The fluxes derived from the present data set are consistent with the fluxes of POC, biogenic Ca and Al derived from  234  Th:238U  disequilibrium in the upper water column and the fluxes measured with sediment traps at 200m and 1000 m. This is, however, surprising since small suspended particles are generally not viewed as the particles that are exported to deeper waters. To better evaluate the validity of this apparent success and establish whether this approach could be generally applicable, it is necessary to gain a better understanding of the influence of particle dynamics on flux estimates derived from 230Th-normalization. 283  The strategy followed to test  230  Th-normalization is to develop a particle dynamics model  that would approximate the profiles of 234Th, 230Th and POC measured at OSP. The goal is not to obtain a perfect fit and estimate the optimal particle dynamic parameters (adsorption and desorption rate constants; aggregation and disaggregation rate constants, sinking rates) that best describe particle dynamics at station Papa, but to develop a reasonable model which could then be used to compare the “true” sinking flux of POC (i.e., the sum of [POC] multiplied by sinking rates for each class of particles) and 230  Th-normalization on fine particles. Each parameter of the model can then be modified  in sensitivity tests to better understand how they affect  230  Th-normalization and under  what circumstances it could provide a good estimate of the “true” flux.  6.4.5.1 Model A  As a first step, the parameters of the particle dynamic model (model A) depicted on Fig. 6.1 (K1, K-1, B1, B-1, S) were varied in an attempt to reproduce the profiles of 230  234  Th and  Th dissolved, and associated with fine and large particles measured at OSP. The  following equations were used: P + K-1 F = (K1 + ) D  (6.8)  K1 D + B-1 L = (K-1 + B1 +) F  (6.9)  B1 F = (B-1 +) L + S L/z  (6.10)  284  where P = production rate of Th in seawater,  is its decay constant (negligible for 230Th); D = activity of dissolved Th; F = activity of thorium adsorbed on fine suspended particles; L = activity of thorium adsorbed on large sinking particles; λ = Th decay constant; K 1 and K-1 are the adsorption and desorption rate constants between dissolved and fine particles; B1 and B-1 are the aggregation and disaggregation rate constants between fine and large particles; S = sinking rate of large particles.  We started with K1 = 0.5/y and K-1 = 1.6/y (Bacon and Anderson, 1982), B1 = 3/y and B-1 = 150/y (Clegg et al., 1991), and S = 150m/day (Berelson et al., 2002). With these parametric values kept constant with depth, model A generates profiles of dissolved and fine particulate Th that are different from the observations but in the range of measured values. On the other hand, the model grossly underestimates the concentration of  234  Th  and 230Th associated with large particles (>53μm) (Fig. 6.20). The model is also unable to reproduce the low dissolved  234  Th and high fine particulate  234  Th measured in surface  waters, where K1 must be >>0.5 y-1, reflecting higher scavenging intensity due to higher particle concentrations and fluxes. Because we are mostly interested in measuring flux in the mesopelagic zone, only the data obtained between 60 m and 1600 m are considered in the following discussion and the sensitivity tests reported in Appendix E.  285  500  1000 1500 0  0.1  0.2 0.3 (dpm/m3) F230  1000 1500 2000  0.4  0  0  500  500  Depth(m)  Depth(m)  Depth(m)  500  2000  1000 1500 2000  Depth(m)  D234 0  0  0.05 (dpm/m3) L230  0  500  500  1500 2000  0  0.01 0.02 (dpm/m3)  0.03  500 1000 1500 2000 2500 (dpm/m3) F234  1500  0  1000  0  1000  2000  0.1  Depth(m)  Depth(m)  D230 0  0  200  400 600 (dpm/m3) L234  800  50 (dpm/m3)  100  1000 1500 2000  0  Figure 6.20: Output from model A with K1 = 0.5/y and K-1 = 1.6/y (Bacon and Anderson, 1982); B1 = 3/y and B-1 = 150/y (Clegg et al., 1991); and S = 150m/day (Berelson et al., 2002). Sensitivity test are reported in Appendix E.  286  In an attempt to improve the fit between observations and model A, each of the 5 parameters was systematically varied to assess its impact on the Th profiles. The results of these tests are shown in Appendix E.  Taking Fig. 6.20 as the starting point, increasing the adsorption rate constant K1 decreases 234  dissolved and increases the fine and large particulate decreases the dissolved  Th concentrations. It also  230  Th activity but does not affect the  particles. While this improves the fit for  230  Th associated with  234  Th, it worsens the fit for  230  Th (Fig. c;  Appendix E). Increasing the desorption rate constant K-1 has an opposite but smaller effect 234  on dissolved and particulate  Th (Fig. d). It increases dissolved  230  Th activity but does  not change particulate 230Th. Changing K1 and K-1 has no impact on the activities of 230Th associated with large particles, which always remain much too low (Figs c, d; Appendix E). Increasing the aggregation rate constant B1 decreases dissolved and fine particulate 230  Th activities but lowers much less the activities of  Th in these two pools. While  234  increasing B1 increases the increase  234  Th activities associated with large particles, it fails to  230  Th activities significantly (Fig. e; appendix E). Increasing the disaggregation  rate constant has the opposite effect. Decreasing the sinking rate of particles (S) can substantially increase the their  230  Th activities of large sinking particles but fails to increase  234  Th activities significantly. It also produces exceedingly high dissolved and fine  particulate 230Th (Fig. f; appendix E). Clearly, the distribution of 230Th and 234Th activities measured at station Papa cannot be explained by the model A described in Fig. 6.1. In particular, it is incompatible with the relatively large activities m