Geosci. Model Dev., 11, 3537–3556, 2018 https://doi.org/10.5194/gmd-11-3537-2018 © Author(s) 2018. This work is distributed under the Creative Commons Attribution 4.0 License. A global scavenging and circulation ocean model of thorium-230 and protactinium-231 with improved particle dynamics (NEMO–ProThorP 0.1) Marco van Hulten1,2, Jean-Claude Dutay1, and Matthieu Roy-Barman1 1Laboratoire des Sciences du Climat et de l’Environnement, IPSL, CEA–Orme des Merisiers, 91191 Gif-sur-Yvette, France 2Geophysical Institute, University of Bergen, Bergen, Norway Correspondence: Marco van Hulten ([email protected]) Received: 1 November 2017 – Discussion started: 8 December 2017 Revised: 16 August 2018 – Accepted: 20 August 2018 – Published: 31 August 2018 Abstract. In this paper we set forth a 3-D ocean model of the can be done based on our model as its source code is readily radioactive trace isotopes 230Th and 231Pa. The interest arises available. from the fact that these isotopes are extensively used for in- vestigating particle transport in the ocean and reconstructing past ocean circulation. The tracers are reversibly scavenged 1 Introduction by biogenic and lithogenic particles. Our simulations of 230Th and 231Pa are based on the Oceanic circulation and the carbon cycle play a major role NEMO–PISCES ocean biogeochemistry general circulation in the regulation of the past and present climate. Heat and model, which includes biogenic particles, namely small and carbon dioxide in the atmosphere tend to equilibrate with the big particulate organic carbon, calcium carbonate and bio- ocean surface and are transported down into the deep ocean genic silica. Small and big lithogenic particles from dust de- through the Meridional Overturning Circulation (MOC). position are included in our model as well. Their distributions Biogeochemical cycling also generates organic carbon that generally compare well with the small and big lithogenic par- transfers into the deep ocean through particle sinking. Be- ticle concentrations from recent observations from the GEO- cause of this, the strength of the MOC and particle removal TRACES programme, except for boundary nepheloid layers participate actively in the regulation of the climate on the for which, as of today, there are no non-trivial prognostic Earth. models available on a global scale. Our simulations repro- Trace elements are also affected by these mechanisms duce 230Th and 231Pa dissolved concentrations: they compare and represent useful tools to provide constraints on these well with recent GEOTRACES observations in many parts of processes. The GEOTRACES programme has generated a the ocean. Particulate 230Th and 231Pa concentrations are sig- large unique dataset that can now be used to better under- nificantly improved compared to previous studies, but they stand biogeochemical oceanic processes. Modelling quan- are still too low because of missing particles from nepheloid tifies and provides more information on the processes that layers. Our simulation reproduces the main characteristics of control the oceanic distribution of these new observations. the 231Pa=230Th ratio observed in the sediments and supports In present day climate, it is difficult to measure the MOC a moderate affinity of 231Pa to biogenic silica as suggested by strength, and for past climate there are no measurements recent observations relative to 230Th. available at all. Isotopes and trace elements from sedi- Future model development may further improve under- ment cores are used as proxies to infer past ocean circu- standing, especially when this will include a more com- lation. Several examples include carbon isotopes (Broecker plete representation of all particles, including different size et al., 1988), the cadmium=calcium ratio (Rosenthal et al., classes, manganese hydroxides and nepheloid layers. This 1997), the ratio between protactinium-231 and thorium-230 (231Pa=230Th) (Böhm et al., 2015), and the neodymium iso- Published by Copernicus Publications on behalf of the European Geosciences Union. 3538 M. van Hulten et al.: Global ocean model of 230Th and 231Pa tope ratio 143Nd=144Nd (Piotrowski et al., 2005). These prox- tion models have also been used to simulate these tracers, ies are affected by dynamical and biogeochemical processes. namely by Dutay et al.(2009) (NEMO–PISCES), Siddall Including these proxies in a climate model is a way to better et al.(2007), Rempfer et al.(2017) (Bern3D), and Gu and Liu understand the climatic signal they register. (2017) (CESM). Dutay et al.(2009) demonstrated that the We will focus on 230Th and 231Pa because these isotopes particle concentration simulated by the PISCES model in the are well documented by the international GEOTRACES pro- deep ocean was too low by a factor of 2 in the mesopelagic gramme, and they are particularly suitable to study the trans- zone to 50 in the deepest ocean. This led to overestimated fer of particulate matter since the isotopes’ source in the radionuclide concentrations in the deep ocean. Therefore, it ocean is perfectly known: radioactive decay of uranium iso- is crucial to improve the representation of the particles (Du- topes. Others have modelled 143Nd and 144Nd (Arsouze tay et al., 2009). Rempfer et al.(2017) showed that taking et al., 2008, 2009; Ayache et al., 2016), 13C(Tagliabue et al., into account additional sinks at the sea floor and at the ocean 2009), and 14C(Mouchet, 2013). The ratio between 230Th margins yields an improved agreement with observations, es- and 231Pa is used as a proxy for past ocean conditions, but pecially for the dissolved phases of 230Th and 231Pa. Partic- this signal is potentially affected by both circulation and bio- ulate ratios improved to a lesser extent; the authors have not geochemical changes. Therefore, a correct understanding of presented an evaluation of their simulated particulate concen- the scavenging and underlying particle dynamics is essential trations. Gu and Liu(2017) had similar goals as this study. in order to better simulate these tracers (Dutay et al., 2015). In this study, we try to improve on previous studies that Protactinium-231 and thorium-230 are produced in the simulate the distribution of 230Th and 231Pa. Our approach ocean by the α decay of uranium-235 and uranium-234, re- is to improve the mechanistic description of the particle and spectively. Because the activity of uranium is approximately radionuclide cycling, as well as on the side of system de- uniform in the ocean, 231Pa and 230Th are produced at a rel- sign and reusability of the model. We evaluate how well the −3 −3 −1 atively constant rate (βPa D 2:33 × 10 dpmm yr and model fits with observations, but tuning is not one of our −2 −3 −1 βTh D 2:52 × 10 dpmm yr ; dpm: disintegrations per main goals. Just like Dutay et al.(2009), we use the NEMO– minute) (Henderson and Anderson, 2003). They are both OPA (Nucleus for European Modelling of the Ocean – Océan scavenged rapidly by the many particles that reside in the PAralelisé) ocean general circulation model (Madec and the ocean and settle towards the sea floor. 231Pa is less sensitive NEMO team, 2016) and the PISCES biogeochemical model to particle scavenging than 230Th, which is reflected in the (Aumont et al., 2015). Aumont et al.(2017) showed that dis- longer residence time of 231Pa (80–200 yr) compared to that solution rates of particulate organic carbon (POC) in PISCES of 230Th (20–40 yr) (Yu et al., 1996). 231Pa and 230Th are were overestimated, and they improved on this by introduc- radioactive, decaying to radium isotopes and having a half- ing a spectrum of different labilities. This improved the simu- life of 32.76 and 75.40 kyr, respectively. Each combination lation of both small and big POC significantly, so we use this of particle–radionuclide adsorption has a different reactiv- same model for our simulations. There are several improve- ity. Other factors may affect the adsorption of radionuclides ments since Dutay et al.(2009); namely, the new model onto particles. For example, smaller particles have larger sur- face area to volume ratios, which results in an increase in – includes lithogenic particles from dust deposition; the number of these radionuclides adsorbed per particle. Ad- ditionally, while the adsorption rate is expected to increase – has improved the biogeochemistry, affecting the bio- with particle concentration, the adsorption rate may be partly genic particle distributions (Aumont et al., 2015; Au- limited by the coagulation of non-filtered particles (or col- mont et al., 2017); loids) to filtered particles (Honeyman et al., 1988; Honeyman – includes three different phases per nuclide (dissolved, and Santschi, 1988). The vertical distributions of natural ra- and adsorbed onto big (and small) particles), whereas dionuclides, such as 230Th and 231Pa, are hence sensitive to Dutay et al.(2009) include only a single compart- the distribution and mixture of particles. As a consequence ment (total concentration) for each nuclide from which of the different particle reactivities of 230Th and 231Pa, the they calculated the respective phases based on chemical dissolved concentration ratio 231Pa=230Th deviates from the equilibrium; production activity ratio of 0.093 (Anderson et al., 1983; An- derson, 2003; Rutgers van der Loeff et al., 2016). – is more precise and explicit on the mathematical formal- As both particle dynamics and circulation of the ocean ism; affect 230Th and 231Pa, numerical biogeochemical general circulation models are used to study the relative contribu- – is written in Fortran 95 instead of Fortran 77, making tion of these mechanisms. The isotopes have been simulated the extension of its use to other modern Fortran models in models of intermediate complexity, for instance by Hen- easier; and derson et al.(1999) (LSG-OGCM), Marchal et al.(2000) (EMIC 2.5D), Heinze et al.(2006, 2018) (HAMOCC) and – is part of a modern model framework, NEMO–TOP, fa- Luo et al.(2010).