Simple mass balance models are applied to a high resolution 234Th profile from the northwest Pacific to examine the magnitude, rate, and depth distribution of particle remineralization processes below the euphotic zone (Ez). Here, excess 234Th (234Th > 238U) below the Ez is attributed to fragmentation processes that result in the conversion of sinking to non-sinking particles. By considering particulate organic carbon (POC) to 234Th ratios on particles, we show that POC flux attenuation is larger than for 234Th, which we attribute to bacterial and zooplankton consumption of sinking POC. Three case studies are used to demonstrate how different combinations of particle fragmentation and POC respiration impact flux attenuation below the Ez. When sampled with high vertical resolution and precision, 234Th and POC/234Th ratios provide insights into both export from the Ez and the extent to which sinking particle fluxes and associated minerals are attenuated with depth.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 Within the ocean, particle formation and dissolution plays a critical role in the cycling of many natural and anthropogenically produced elements that effect biological production, human health and even Earth's climate. One way of understanding the marine particle dynamics in the upper ocean is to utilize thorium-234 as a natural tracer of these processes. 234Th is a naturally occurring, short-lived radionuclide (t1/2 = 24.1 d) produced by the radioactive decay of 238U (t1/2 = 4.47 × 109 y). Unlike its conservative parent, 234Th is highly particle reactive in seawater, making it an ideal tracer of particle dynamics within marine systems. In essence, when 234Th is in disequilibrium with 238U (total activities 234Th < 238U), the loss of 234Th on sinking particles is large relative to its production rate and rapid relative to its half life [Waples et al., 2006]. 234Th adsorbed on to sinking particles is subsequently transported into deeper waters where particle formation is reduced and remineralization processes result in non-sinking particles or conversion to dissolved phases. These remineralization processes effectively act as an additional source of 234Th to subsurface waters, thereby potentially causing activities of 234Th in excess of that supplied by in situ 238U decay (234Th > 238U).
 Most studies dealing with 234Th have focused on the upper ocean and the role of 234Th as a tracer of surface ocean particle export derived from biological activity. These studies use the ratio of carbon (or other element or compound) to 234Th on sinking particles to empirically convert 234Th derived fluxes into elements of more interest, such as particulate organic carbon (POC), biogenic silica or trace metals [e.g., Buesseler, 1998]. Although the concept of 234Th excess at depth is not new, many earlier studies were limited by methodology, resulting in low vertical sampling resolution [Bacon et al., 1996; Usbeck et al., 2002].
 The development of a small volume technique for 234Th measurements has increased both ease in sampling and precision (≤5%) [Benitez-Nelson et al., 2001a, 2001b; Pike et al., 2005] leading to its increasingly widespread application [Waples et al., 2006]. As a result, there are now a number of high resolution vertical 234Th profiles where excess 234Th in subsurface waters is detected [Savoye et al., 2004; Buesseler et al., 2008; Maiti et al., 2008; Buesseler et al., 2009]. In many cases, excess 234Th activities are found immediately below the euphotic zone (Ez = 0.1% light) and mixed layer. This excess 234Th peak is contained within a rather narrow layer of water where remineralization of sinking particles via fragmentation and respiration of POC by bacteria and/or zooplankton occurs at rates that are sufficient to cause a temporal excess in the subsurface 234Th activity (note we use remineralization to refer to the combined biological and physicochemical processes that causes particle flux attenuation).
 Here, we utilize a high resolution 234Th profile collected from the NW Pacific to examine how excess 234Th activities at depth, coupled with a simple mass balance model, can be used to examine the magnitude, rate, and depth distribution of particle remineralization processes below the Ez. These concepts are further expanded to the application of 234Th as a tracer of both surface POC export and subsequent remineralization within the twilight zone (Ez to 1000 m). The ideas presented here set the foundation for future in depth studies of 234Th and POC fluxes and remineralization.
2. Model Development and Discussion
 The magnitude and depth of excess 234Th activities are controlled by a number of different processes: the absolute flux of sinking particles (e.g., development and decline of blooms within the Ez); the nature of the sinking particles (i.e., lability and sinking rate); subsurface remineralization (e.g., changes in zooplankton or bacterial respiration rates, and/or their depth distributions); by the physical environment (e.g. density discontinuities, changes to temperature or geochemistry); or some combination of the above.
 The models developed here (see Text S1 for details) use data from site K2 (47°N, 161°E) in the northwestern Pacific Ocean in order to illustrate a conceptual framework for interpreting the shape and size of excess 234Th features observed in the water column. Using the K2 234Th activity profile and steady state derived 234Th fluxes (assuming physical processes to be negligible), we model three processes that likely influence 234Th:238U disequilibria (both 234Th excess and deficiency) at any site.
2.1. Changes in the Flux of 234Th
 In order to have a significant excess of 234Th in the ocean, there must be a deficiency of 234Th somewhere else. In the open ocean, variations in the magnitude of the 234Th deficiency, controls the maximum extent of excess 234Th at depth. This is illustrated in our K2 example by maintaining constant depth and remineralization rates while increasing or decreasing the originally measured 234Th export flux at 60 m by 50%. When increasing the flux by 50%, the 234Th deficit within the upper 60 m increases to account for the higher 234Th flux (green line Figure 1a). Using a higher flux but the same remineralization rates between 60–120 m, results in greater excess 234Th. The reverse is true when the flux is decreased by 50% (blue line Figure 1a), resulting in a less prominent 234Th excess zone that can be easily missed by low resolution 234Th profiles.
2.2. Width of Remineralization Zone
 In the K2 profile, the most prominent zone of remineralization is just below the Ez between 60 and 120 m, where ∼220 dpm m−2 d−1 of the 234Th flux is remineralized. However, the relative contribution of the different depth horizons to the total 234Th remineralized is not uniform. If the same amount of 234Th remineralization were to take place with uniform intensity between 60–120 m, it would result in a plateau of 234Th excess (Figure 1b, brown line) and this is not observed. On the other hand, if remineralization was limited to within a 10 m (i.e. between 60–70 m) or a 30 m (i.e. between 60–90 m) wide layer, the peak in 234Th excess becomes more intense and clearly identifiable (Figure 1b). Unfortunately, these narrow peaks also become relatively easier to miss without high resolution sampling, since they are essentially defined by only one or two 234Th data points.
2.3. Variable Rates of Remineralization
 In order to understand the sensitivity of the 234Th profiles to changes in remineralization rates between 60 and 120 m, the fraction of 234Th remineralized in this zone is increased by 2–3 fold or decreased by 4 fold from that observed in the original data (Figure 1c, red line). When the remineralization rate increases, the 234Th excess peak becomes increasingly prominent and is more easily identified (Figure 1c). In contrast, when the remineralization rate is decreased, it becomes more difficult to identify any 234Th excess in the profile. In the current example (Figure 1c, brown line) even with samples collected at 10 m sampling intervals, it would be difficult to detect excess 234Th if the remineralization rate decreased by more than two fold since the 234Th activities in the remineralization zone become indistinguishable from 238U beyond analytical uncertainty (Figure 1c).
 Using the model calculations, we demonstrate how simple changes in 234Th flux, the depth over which that remineralization occurs, and remineralization rate, influences the magnitude and location of excess 234Th within the water column. In the ocean, the manifestation of excess 234Th in the water column is likely controlled by multiple processes. For example, the absence of a clear 234Th excess layer even within a high resolution profile does not necessarily mean that there is no significant remineralization. Rather, it could simply be spread out over a wider depth zone that does not allow for clear identification of a 234Th excess peak within analytical uncertainties (similar to brown line in Figure 1b). What can be said, however, is that if an excess 234Th peak is observed, it generally indicates a rather narrow layer of more intense remineralization (10's of meters) that is larger than 10% of the overlying 234Th particle flux. For any given profile, the models described here allow one to set limits on the extent of 234Th remineralization and depth over which these higher remineralization rates occur.
3. 234Th as a Proxy for POC Remineralization
 In this section, we combine the 234Th flux and remineralization models with POC/234Th data, to examine POC flux attenuation, focusing on the 100 m layer immediately below the Ez (Ez+100). It is within this layer that regional and temporal differences in POC flux attenuation are typically largest [Buesseler and Boyd, 2009]. Using the K2 example, 234Th fluxes would reach a maximum around 60 m, and then decrease such that the net 234Th flux at Ez+100 is about 10% lower than at Ez (Figure 1d). As a sensitivity analysis, we also plotted a wider range of possible 234Th remineralization scenarios (as shown in Figure 1c) where the loss in 234Th flux between Ez and Ez+100 varies between 0 and 40% (Figure 1d). For example, when the remineralization rate increases by three fold (Figure 1c, green line) or two fold (Figure 1c, blue line) or decreases by four fold (Figure 1c, brown line), there is a corresponding decrease in 234Th flux at Ez+100 by 40% (Figure 1d, green line), 25% (Figure 1d, blue line) and 0% (Figure 1d, brown line) respectively.
 The POC/234Th ratio at K2 (hereafter abbreviated as C/Th) decreases from >15–20 μM dpm−1 to <3–5 μmol dpm−1 below 100 m (Figure 1e). As in other studies, variability in C/Th is greatest in the Ez and decreases with depth (see review by Buesseler et al. ). Here C/Th is the same in both sediment traps and on the larger particle size classes collected on 20 and 51 μm nominal pore size screens. We calculate the POC flux by multiplying the 234Th flux by a straight line fit to the C/Th data (Figure 1f). The decrease in 234Th and POC flux from Ez to Ez+100 is 10% and 54% respectively for the K2 profile. However for the same changes in remineralization rate as shown in Figure 1c, the 234Th flux decreases by 0–40% where as the POC flux decreases from 48–70% between Ez and Ez+100 (Figures 1d and 1f). Interestingly in these two figures, the flux attenuation between Ez and Ez+100 is much larger for POC (48–70%) than for 234Th (0–40%) due to the decrease in C/Th with depth (Figure 1e). However it must be noted that changes in remineralization rate may also affect the C/Th which is not taken into account in our model formulation.
 Conceptually, we hypothesize that differences in the 234Th and POC flux profiles provide important insights into the processes that control flux attenuation. The 234Th excess at depth can be attributed to any process that results in the conversion of sinking to non-sinking particles, to which 234Th is attached. We will call this process fragmentation, as fragmentation of fecal pellets or marine snow by zooplankton (coprohexy) is a common example of this type of process [Lampitt et al., 1990]. However, other processes such as particle disaggregation due to physical-chemical processes, or even a decrease in particle sinking speed to the point where the flux per day is slower than the ingrowth and decay of 234Th, would all decrease the 234Th flux. Sinking POC on the other hand, is a food source for mesopelagic bacteria and zooplankton. Thus, in addition to fragmentation processes, POC flux will be attenuated due to respiration within the mesopelagic. This second process would not necessarily alter the flux of 234Th (though it may indirectly by changing particle properties). In this example, fragmentation processes decrease both 234Th and POC fluxes by 10%, while the POC flux will be further attenuated by an additional 44% due to C respiration processes (0 to 10% change due to fragmentation for 234Th, and 44% to 54% for POC due to respiration). Please note that the possible loss of POC via formation of dissolved organic carbon is not taken into consideration in this paper.
 In the following section, we examine three other settings where the vertical 234Th and C/Th resolution is not as high, but still sufficient to illustrate regional and temporal variability in flux attenuation. These studies help to illustrate the relative importance of C respiration and particle fragmentation processes in controlling overall particle flux remineralization in the open ocean.
4. 234Th Derived Particle Remineralization: Comparison of Three Study Sites
4.1. Case 1
 In a study of the impact of mesoscale eddies on upper ocean biogeochemistry in the Sargasso Sea [McGillicuddy et al., 2007], an excess 234Th peak was observed at all stations, but especially at the center of a cyclonic eddy, where prominent excess 234Th peaks were found immediately below the euphotic zone [Buesseler et al., 2008]. The excess 234Th activities between the base of the Ez and Ez+100 (Figure 2a) leads to decreases in the 234Th flux by 67 and 31% (Figure 2b) and even larger decreases in POC flux, by 77 and 53% (Figure 2c) as C/Th decreases as well. At K2 a 10% decrease in 234Th flux resulted in a 54% decrease in POC flux. The reduced sensitivity of 234Th flux to changes in the POC flux within these eddies suggests that fragmentation processes, as evidenced by higher 234Th flux losses), are more dominant than respiration in comparison to the K2 site. The greater POC loss due to respiration at K2 is supported by upper 150 m water column data which indicate much higher zooplankton and bacterial production at K2 (zooplankton biomass ≈ 3500 mg m−2 [Steinberg et al., 2008]; BP ≈ 62 mg C m−2 d−1 [Boyd et al., 2008]) relative to inside the eddies (zooplankton biomass ≈ 700 mg m−2 [Goldthwait and Steinberg, 2008]; BP ≈ 17 mg C m−2 d−1 [Ewart et al., 2008]).
4.2. Case 2
 Station ALOHA in the subtropical Pacific Ocean, is characterized by very low 234Th disequilibria, resulting in 234Th fluxes averaging only 400 ± 400 dpm m−2 d−1 at the base of the Ez [Buesseler et al., 2009]. The system is characterized by small picoplankton and grazers, with a deep chlorophyll maxima and extensive recycling of POC in the upper 125 m leading to low export ratios of 7% [Buesseler and Boyd, 2009]. A few of these profiles show a slight 234Th excess below the Ez (Figure 2d) and illustrate our earlier model discussions that with low surface 234Th fluxes, excess 234Th peaks are very difficult to discern (blue line in Figure 1a). In the first profile (solid triangle in Figure 2e) there is 0% loss in 234Th flux between Ez and Ez+100 with a corresponding 39% loss in POC flux (solid triangle in Figure 2f). This suggests a dominant role of respiration over fragmentation. However the other profile (solid circle in Figures 2e and 2f) collected 10 days latter shows very similar loss in both 234Th flux (80%) and POC (83%) indicating a shift towards fragmentation. Sufficient biological data was not collected during these two profiles to understand what caused the shift in the fragmentation and respiration processes. The presence of a detectable 234Th excess peak in this profile and from both the profiles from Sargasso eddies may suggest that prominent 234Th excess peaks are more often associated with fragmentation processes.
4.3. Case 3
 During this iron enrichment experiment, biological production was enhanced in HNLC waters south of the Polar Front [Buesseler et al., 2004]. The 234Th profiles shown here are from days 6 and 27 during which they changed significantly, showing a large decrease in total 234Th in the upper 70 m and elimination of the 234Th excess below the Ez (Figure 2g). The result is an increase in overall export flux from the Ez for both 234Th and POC (Figures 2h and 2i). Of importance to this discussion is the narrow layer of significant POC flux attenuation below the Ez early in the bloom where 234Th flux and POC flux decreases by 40 and 65% respectively between Ez and Ez+100 (solid triangles in Figures 2h and 2i). During this stage both fragmentation and respiration played an important role in remineralization. However, latter in the experiment there is an increase in the POC export below the Ez resulting in no net loss of 234Th or POC between Ez and Ez+100 (circles in Figures 2h and 2i). The changes below Ez suggests either more efficient transport of sinking POC due to differences in sinking particle properties (such as sinking rate, particle lability), and/or a change in remineralization possibly due to shift in the depth of maximal zooplankton feeding [Buesseler et al., 2005]. This may have resulted in very little fragmentation and lower respiration in this zone leading to negligible attenuation of 234Th and POC flux. Biological data were not collected with sufficient resolution to resolve the causes, but this is another example of where 234Th was able to document not only a change in POC export from the Ez, but a very dynamic shift in POC attenuation below the Ez over a span of three weeks.
5. Summary and Future Work
 Throughout the water column, attached and free living bacteria recycle organic matter, breaking down POC into colloidal and dissolved organic and inorganic forms of C, thereby changing particle properties and stickiness (e.g. TEP) [Passow et al., 2001]. Thus, bacteria can potentially increase, or more generally, decrease POC fluxes [Taylor et al., 2001; Azam et al., 1992]. Zooplankton grazing can create or aggregate particles during feeding, resulting in the production of rapidly sinking fecal material, and/or fragment sinking particles by their feeding activities. As they consume POC they produce dissolved organic matter (DOM) and respire some C as dissolved inorganic C. Zooplankton can also actively transport DOM and POC from the surface to depth by vertical migration, though this is highly variable [Longhurst et al., 1990; Wilson et al., 2008].
234Th provides a novel mechanism by which the above processes may be further examined and understood, by pinpointing the depths and extent to which the flux of sinking particles and associated minerals are attenuated. This is clearly evident in the K2 example and case studies, where layers immediately below the Ez contain sufficient rates of remineralization to result in POC flux attenuations even when 234Th flux sometimes remain relatively unchanged. Such knowledge can only be gained through high resolution vertical studies, and this is a major advantage of 234Th over other methods like sediment traps. Ultimately, we would like to use this approach to better constrain the bacterial and zooplankton processes responsible for the fragmentation and respiration of sinking particles as they sink through the water column. This will require not only high resolution sampling of 234Th activities, but also the C/Th activity on sinking particles. Furthermore, biological community structure and activity need to be sampled at similar resolution. In this manner, it will be possible to not only examine particle remineralization processes relative to 234Th and C, but other biologically relevant elements, such as nitrogen, phosphorus, biogenic silica, and trace metals.
 The data presented here are from a number of studies funded by the National Science Foundation. Special thanks to Steve Pike for collection and analyses of many of the 234Th samples.