Evaluation of asymmetric Oxygen Minimum Zones in the tropical Pacific: a basin-scale OGCM-DMEC V1.0

The tropical Pacific Ocean holds the world’s two largest Oxygen Minimum Zones (OMZs), showing a prominent hemispheric asymmetry, with a much stronger and broader OMZ north of the equator. However, there is a lack of 10 quantitative assessments of physical and biological regulations on the asymmetry of tropical Pacific OMZs. Here, we apply a fully coupled basin-scale model (OGCM-DMEC V1.0) to investigate the impacts of physical supply and biological consumption on the dynamics of OMZs in the tropical Pacific. We first utilize observational data to evaluate and improve our model simulation, and find that mid-depth DO is more sensitive to the parameterization of background diffusion. Enhanced background diffusion results in higher DO concentrations at mid-depth, leading to significant improvement of our 15 model capability to reproduce the asymmetric OMZs. Our study shows that while physical supply of DO is increased in majority of the tropical Pacific due to enhanced background diffusion, there is little increase in the largest OMZ to the north. Interestingly, enhanced background diffusion results in lower rates of biological consumption over ~300-1000 m in the entire basin, which is associated with redistribution of dissolved organic matter (DOM). Our analyses demonstrate that weaker physical supply in the ETNP is the dominant process responsible for the asymmetric DO in the core OMZs (~200-600 m) 20 while higher biological consumption to the north plays a larger role in regulating DO concentration beneath the OMZs (~600-800 m), with implication for the asymmetric OMZs. This study highlights the roles of physical supply and biological consumption in shaping the asymmetric OMZs in the tropical Pacific, underscoring the need to understand both physical and biological processes for accurate projections of DO variability.

Dissolved oxygen (DO) is a sensitive indicator of physical and biogeochemical processes in the ocean thus a key parameter for understanding the ocean's role in the climate system (Stramma et al., 2010). In addition to photosynthesis and respiration, the distribution of DO in the world's oceans is also regulated by air-sea gas exchange, ocean circulation and ventilation (Bopp et al., 2002;Brandt et al., 2015;Levin, 2018). Unlike most dissolved nutrients that display an increase in 35 concentration with depth, DO concentration is generally low at mid-depth of the ocean. The most remarkable feature in the oceanic oxygen dynamics is the so-called Oxygen Minimum Zone (OMZ) that is often present below 200 m in the open oceans (Karstensen et al., 2008;, where DO concentration is less than 20 mmol m -3 (Gilly et al., 2013;Paulmier and Ruiz-Pino, 2009).

40
The world's two largest OMZs are observed at ~200-700 m in the Eastern Tropical North Pacific (ETNP) and South Pacific (ETSP), showing a peculiar asymmetric structure across the equator, i.e., a much larger volume of suboxic water (<20 mmol m -3 ) to the north than to the south (Karstensen et al., 2008;Paulmier and Ruiz-Pino, 2009;. It is known that OMZs are caused by the biological consumption associated with remineralization of organic matter (OM), and weak physical supply of DO due to sluggish subsurface ocean circulation and ventilation (Brandt et al., 2015;Kalvelage et al., 45 2015). Although there have been a number of observation based analyses addressing the dynamics of OMZs in the tropical Pacific during the past decade (Czeschel et al., 2012;Garç on et al., 2019;Schmidtko et al., 2017;Stramma et al., 2010), our understanding is uncompleted in terms of the underlying mechanisms that regulate DO dynamics at mid-depth due to the limitation of available data Stramma et al., 2012).

50
Large-scale physical-biogeochemical models have become a useful tool for advance understanding of the oceanic oxygen cycle (Cabré et al., 2015;Ward et al., 2018;Williams et al., 2014), and have demonstrated that physical processes such as ventilation and circulation play a dominant role in regulating the dynamics of tropical OMZs (Deutsch et al., 2006;Ito et al., 2019;Montes et al., 2014;Stramma et al., 2010). However, many current global and regional models could not consistently reproduce the dynamics of tropical Pacific OMZs (Cabré et al., 2015;Stramma et al., 2012), which may be related to 55 unresolved transport process, unaccounted biological consumption or missing biogeochemical feedbacks in the models (Keller et al., 2016;Palter and Trossman, 2018;Shigemitsu et al., 2017). Moreover, there has been a lack of quantitative assessments on the relative roles of physical supply and biological consumption in driving the tropical asymmetric OMZs using validated models (Keller et al., 2016;.

60
A fully coupled basin-scale physical-biogeochemical model (OGCM-DMEC V1.0) was developed for the tropical Pacific (Wang et al., 2008;Wang et al., 2015;Wang et al., 2009), which shows capability of reproducing observed spatial and temporal variations in physical, nutrient and carbon fields. In this study, we use the validated basin-scale model to evaluate the simulation of oxygen cycle, focusing on the dynamics of mid-depth DO. We first carry out model evaluation and https://doi.org/10.5194/gmd-2020-105 Preprint. Discussion started: 29 July 2020 c Author(s) 2020. CC BY 4.0 License. validation by using observational DO data to improve the simulation of OMZs in the tropical Pacific. Then, we use the 65 improved model to quantify how physical supply and biological consumption regulate the dynamics of DO at mid-depth.
The objective of this study is to advance our model capacity to simulate the oceanic oxygen cycle, and to explore the mechanisms driving the asymmetric OMZs in the tropical Pacific.

Model description
The OGCM-DMEC V1.0 is a fully coupled basin-scale physical-biogeochemical model that has shown a good model-data 70 agreement in the carbon cycle for the tropical Pacific Ocean (Wang et al., 2015). A brief description is presented in this section.

Ocean physical model
The basin-scale OGCM, a reduced-gravity, primitive-equation, sigma-coordinate model, is coupled to an advective 75 atmospheric model (Murtugudde et al., 1996). There are 20 layers with variable thicknesses in the OGCM. The mixed layer (the upper-most layer) is determined by the Chen mixing scheme (Chen et al., 1994), which varies from 10 m to 50 m on the equator. The remaining layers in the euphotic zone are approximately 10 m in thickness. The model domain is between 30º S and 30º N, and zonal resolution is 1º . Meridional resolution varies between 0.3º and 0.6º over 15º S-15º N (1/3° over 10°ºS -10ºN ), and increases to 2º in the southern and northern "sponge layers" (the 20º -30º bands) where temperature, salinity, 80 nitrate and iron are gradually relaxed back towards the observed climatological seasonal means from the World Ocean Atlas, 2013 (WOA2013: http://www.nodc.noaa.gov/OC5/woa13/pubwoa13.html).
The model is forced by atmospheric conditions: climatological monthly means of solar radiation and cloudiness, and interannual, 6-day means of precipitation and surface wind stress. Precipitation is from ftp://ftp.cdc.noaa.gov/Datasets/gpcp, 85 and surface wind stresses from the National Centers for Environmental Prediction (NCEP) reanalysis (Kalnay et al., 1996).
Air temperature and humidity above the ocean surface are computed by the atmospheric mixed layer model. Initial conditions were obtained from outputs of an interannual hindcast simulation over 1948-1978, which  https://doi.org/10.5194/gmd-2020-105 Preprint. Discussion started: 29 July 2020 c Author(s) 2020. CC BY 4.0 License.

Ocean biogeochemical model
The dynamic marine ecosystem-carbon (DMEC) model is the main part of the biogeochemical model that is embedded in the basin-scale OGCM. The DMEC model consists of eleven components: small (S) and large (L) sizes of phytoplankton (PS 95 and PL), zooplankton (ZS and ZL) and detritus (DS and DL), DON, ammonium, nitrate, dissolved iron, and DO ( Figure 1). All biological components use nitrogen as their unit, and are computed in a manner similar to physical variables.
In this model, net community production (NCP) is computed as: where 6.625 is the C:N ratio, µ the rate of phytoplankton growth, r the rate of zooplankton respiration, c the rates of detritus 100 decomposition and DON remineralization. The equations for biogeochemical processes and model parameters have been reported in details previously (Wang et al., 2008;Wang et al., 2015).

Computation of oxygen sources and sinks
The time evolution of DO is regulated by physical, biological and chemical processes: where u, v, and w are zonal, meridional, and vertical velocity, respectively. Omix is the vertical mixing term that is calculated by three subroutines (Chen et al., 1994). Briefly, the first one computes convection to remove instabilities in the water column, and the second one determines the mixed layer depth. The third one computes partial mixing (Pm) between two adjacent layers to relieve gradient Richardson number instability, which is calculated as follows: where the mixing parameter λ is set to 1.
The biological source/sink term Obio is computed as follows: The flux of O2 from the atmosphere to the surface ocean is computed as: where Osat is the O2 saturation, a function of temperature and salinity (Weiss, 1970), and K0 the gas transfer velocity that is a 120 function of wind speed (us) and SST according to Wanninkhof (1992): https://doi.org/10.5194/gmd-2020-105 Preprint. Discussion started: 29 July 2020 c Author(s) 2020. CC BY 4.0 License.

Sensitive experiments
The underestimated DO at mid-depth may be a result of overestimation of removal (primarily the consumption associated 140 with remineralization of organic materials) and/or underestimation of supply (e.g., vertical mixing). We conduct a sensitivity study that consists of four new simulations. The first two experiments are designed to reduce the biological consumption of DO: by reducing DON remineralization constant (cDON in equation 1) by 50% (Exp1), or changing O:C utilization ratio (ROC in equation 5) from 1.3 to 1.0 (Exp2); the other two are to increase supply of DO by changing the partial mixing parameter (Pm in equation 4 ) from 0 to 10 -4 m -2 s -1 (Exp3), or applying a variable Pm (Table 1). 145 The comparisons between simulated DO and WOA2013 climatology data are shown in Figure 3  of 120°W (Figure 3e and 3f), although the simulated volume of suboxic water is slightly larger than from the observation data ( Figure 3a). Figure 4a illustrates a larger volume of suboxic water located north of ~4°N and a smaller volume of suboxic water over 155 8°S-4°S, which are separated by relatively higher DO (>30 mmol m -3 ) along the equator. Both Exp1 and Exp2 (Figure 4c and 4d), as well as the reference run (Figure 4b), produce much larger volumes of suboxic water that are extend to the equatorial region, and even merge into one. Exp3 and Exp4 (Figure 4e and 4f) capture the observed basin-scale spatial distribution of DO, particularly the asymmetric feature (i.e., a larger volume of suboxic water to the north but a smaller size of suboxic water to the south), and relatively higher DO (~40 mmol m -3 ) over 4°S-2°N. Many ESMs are unable to reproduce 160 the positions and sizes of the OMZs in the eastern tropical Pacific Ocean (Bao and Li, 2016;Cabré et al., 2015); some models overestimated the extent of suboxic water, which might be due to over-estimated productivity in the euphotic zone (Moore et al., 2013;Williams et al., 2014).

Model skill assessment 165
To further evaluate the performance of experiments on OMZs, four statistical measures are applied over 300-500 m in the ETNP (150°W-90°W, 0°-10°N) and ETSP (150°W-90°W, 10°S-0°). As showed in Table 2, compared to the reference run, both bias and RMSE decrease in all four experiments, with the smallest bias (-8.59 mmol m -3 in ETNP and 1.05 mmol m -3 in ETSP) and RMSE (12.36 mmol m -3 in ETNP and 8.43 mmol m -3 in ETSP) found in Exp3. Many current models show a maximum of RMSE (~20-80 mmol m -3 ) with respect to observed DO from mixed layer to 1000 m (Bao and Li, 2016;Cabré 170 et al., 2015). Table 1 and Figure 5 illustrate that Exp3 produces the best outputs, with the highest correlation coefficient (0.72 in ETNP and 0.893 in ETSP) and also close-to-1 NSD (1.03 in ETNP and 1.71 in ETSP).

Results and discussions
In this section, we further compare the improved model simulation (EXP3) and reference run to diagnose the relative contributions of physical supply and biological consumption to the asymmetric OMZs in the tropical Pacific, aiming to 175 identify the underlying mechanisms regulating the dynamics of mid-depth DO.

Distribution of DO and physical supple at mid-depth
We first compare the distribution of DO over 300-500 m between reference run and Exp3. The reference run produces much large volume of suboxic waters (<20 mmol m -3 ) in both the ETNP and ETSP where the two OMZs are merged (Figure 6a). 180 On the other hand, Exp3 performs well in reproducing the sizes and locations of two asymmetric OMZs (Figure 6b). Mid-https://doi.org/10.5194/gmd-2020-105 Preprint. Discussion started: 29 July 2020 c Author(s) 2020. CC BY 4.0 License.
depth DO concentration is 1-20 mmol m -3 higher in the Exp3 than in the reference run, with the largest difference in the western equatorial Pacific (Figure 6c). However, despite of the small absolute increase (1-5 mmol m -3 ) in the ETNP, the relative increase of DO is approximately 200-400%, indicating that mid-depth DO in the north OMZ is more sensitive to the parameterization of physical processes. 185 The overall increased DO at mid-depth is likely a result of stronger physical supply, as seen below, due to enhanced background diffusion in Exp3. Numerous studies have indicated that physical mixing is the only source of DO for OMZs (Brandt et al., 2015;Czeschel et al., 2012;Talley et al., 2016), particularly, turbulent diffusion accounted for 89% of the net DO supply for the core OMZ layer (Llanillo et al., 2018). The reference run in our study shows a range of 4-11 mmol m -3 yr -190 1 in physical supply over the 300-500 m, with higher rates (>11 mmol m -3 yr -1 ) found mainly in the northern central tropical Pacific, which is located inside of the suboxic waters ( Figure 7a). Exp3 produces generally higher rates (5-12 mmol m -3 yr -1 ) of physical supply, with a larger size of high-rate (>11 mmol m -3 yr -1 ) water near the edge of northern suboxic water ( Figure   7b). There have been a handful of modelling studies of physical supply rate for the mid-depth waters of tropical Pacific, which show similar magnitudes (from ~4 to ~8 mmol m -3 yr -1 ) (Llanillo et al., 2018;Shigemitsu et al., 2017). Figure 7c  195 illustrates that physical supply is increased by ~0.5-1 mmol m -3 yr -1 in most of the mid-depth, with the highest increase in the southern part of central equatorial Pacific. Interestingly, there is somehow a small decrease (by ~0.1 mmol m -3 yr -1 ) of physical supply in the ETNP-OMZ, implying that changes in other processes such as biological consumption may also contribute to the enhanced DO at mid-depth.

Biological consumption of DO at mid-depth
We further compare biological consumption of DO between Exp3 and the reference run ( Figure 8). For the reference run, biological consumption rate is <6 mmol m -3 yr -1 off the equator and >10 mmol m -3 yr -1 in most of the upwelling region, with higher rates (>11 mmol m -3 yr -1 ) mainly located in the northern central tropical Pacific, and in the suboxic waters ( Figure 8a).
Clearly, Exp3 produces lower rates of biological consumption in the entire basin, showing smaller size of >11 mmol m -3 yr -1 205 relative to the reference run, which is located on the southern edge of northern suboxic waters (Figure 8b). Limited modelling studies reported slightly lower rates (<5 mmol m -3 yr -1 ) in the mid-waters of tropical Pacific (Llanillo et al., 2018;Shigemitsu et al., 2017). The decrease in biological consumption ranges from ~0.1 to 0.3 mmol m -3 yr -1 (Figure 8c). For the northern OMZ, biological consumption decreases by ~0.15-0.2 mmol m -3 yr -1 , which is greater than the decreased rate (~0.1 mmol m -3 yr -1 ) of physical supply (see Figure 7c). 210 Biological consumption of DO in the mid-depth water largely results from remineralization of dissolved organic matter (DOM) and decomposition of particulate organic matter (POM). The reference run simulates a range of ~6-12 mmol m -3 yr -1 and ~0.6-1.5 mmol m -3 yr -1 for DOM remineralization and POM decomposition, respectively, with the largest values inside https://doi.org/10.5194/gmd-2020-105 Preprint. Discussion started: 29 July 2020 c Author(s) 2020. CC BY 4.0 License. of the suboxic waters (Figure 9a and 9b). Clearly, Exp3 produces lower rates for both DOM remineralization and POM 215 decomposition over the entire basin (Figure 9c and 9d), with a decrease of ~0.1-0.2 mmol m -3 yr -1 and ~0.01-0.02 mmol m -3 yr -1 , respectively (Figure 9e and 9f). Apparently, the consumption rate contributed by remineralization of DOM is much greater than that by decomposition of POM at mid-depth, indicating that remineralization of DOM is the dominant process consuming DO at mid-depth. Many other studies also illustrated that the extent of OMZs was not only sensitive to physical mixing but also to remineralization of DOM (Kriest et al., 2010;Loginova et al., 2019). 220

Spatial variations of DOM and its remineralization rate
Remineralization rate of DOM in the ocean is determined by the size of DOM pool and temperature (Brewer and Peltzer, 2016;Wang et al., 2008). Given that there is little difference (<10 -5 °C) in seawater temperature between different model experiments, the reduced consumption rates due to DOM remineralization would be a result of a smaller size of DOM. Here, 225 we evaluate the zonal and meridional distributions distribution of dissolved organic nitrogen (DON) together with remineralization rate of DON. As shown in Figure 10a-10d, modelled DON decreases from 6-8 mmol N m -3 near the surface to 2-4 mmol N m -3 over 400-1000 m. Notably, our modelled remineralization rate of DON decreases from ~20 mmol m -3 yr -1 in the euphotic zone to ~4-8 mmol m -3 yr -1 below 400 m, which is similar to the vertical variation of DON. Enhanced background diffusion leads to a decrease in DON concentration (up to 0.2 mmol N m -3 ) over 300-900 m, but an increase 230 below 1000 m in the eastern tropical Pacific (Figure 10e and 10f).
Our previous study showed that the basin-scale model reproduced observed meridional distribution of surface DON (fall 1992) in the eastern tropical Pacific (Wang et al., 2008). A recent field study reported that surface DON concentration was ~5-7 mmol N m -3 in the ETSP (Loginova et al., 2019), which is similar to our model results. While there was little 235 information on measured rates of DOM remineralization and oxygen consumption at mid-depth of tropical Pacific, there were some field studies conducted in other ocean/sections (e.g., subtropical Atlantic and North Pacific), which showed comparable or slightly lower rates of oxygen consumption (Sonnerup et al., 2013;Stanley et al., 2012). For instance, cruise data within the subtropical North Pacific showed that consumption rate ranged from 8.3 mmol m -3 yr -1 at ~200 m to <3.1 mmol m -3 yr -1 below 500 m (Sonnerup et al., 2013). 240

Impacts of physical supply and biological consumption on asymmetry of OMZs
Previous studies have demonstrated meridional asymmetric features in many physical and biological fields in the tropical Pacific, e.g., temperature and salinity (Fiedler and Talley, 2006), circulation and ventilation (Kessler, 2006;Kuntz and Schrag, 2018), nitrogen and carbon cycles (Libby and Wheeler, 1997;Wang et al., 2009), which may be largely associated with the asymmetries in water mass exchange between the equatorial and off-equator Pacific Ocean (Kug et al., 2003). 245 https://doi.org/10.5194/gmd-2020-105 Preprint. Discussion started: 29 July 2020 c Author(s) 2020. CC BY 4.0 License.
Accordingly, one may assume that the hemisphere asymmetry of OMZs could be related to the differences in physical supply and biological consumption between the ETNP and ETSP.
While the dynamics of the tropical Pacific's OMZ is sensitive to changes in ocean circulation (Czeschel et al., 2011;Montes et al., 2014;Stramma et al., 2010), small scale processes such as background diffusion also have a significant influence on 250 the supply of DO into the core OMZ and the mean state of DO distribution (Duteil and Oschlies, 2011;Getzlaff and Dietze, 2013;Llanillo et al., 2018). Clearly, our model experiment shows that enhanced background diffusion leads to a significant increase in DO concentration below 200 m (Figure 11a). The increase of DO is similar below 600 m in the ETNP and ETSP, but differs largely between the two regions, with much greater values over 200-600 m in the ETSP.

255
Recent studies have emphasized the role of changes in physical processes for the observed asymmetric OMZs in the tropical oceans. For instance, larger-scale mass transport related to circulation and ventilation in the southern hemisphere is more efficient than in the northern hemisphere (Kuntz and Schrag, 2018), and the transit time from the surface to the OMZ is much longer in the ETNP than in the ETSP (Fu et al., 2018;Sonnerup et al., 2013). Our study shows that enhanced background diffusion increases the physical supply of DO over most of the water column, with significant increases (>0.4 260 mmol m -3 yr -1 ) below 200 m in the ETSP (Figure 11b). However, the ETNP shows a much smaller increase in physical supply, particular over ~300-700 m (<0.2 mmol m -3 yr -1 ) (Figure 11a), indicating that physical supply may be largely responsible for the asymmetry of OMZs. On the other hand, increased DO is greatest at ~700 m in the ETNP (Figure 11a), which cannot be explained by the small increase in physical supply (<0.4 mmol m -3 yr -1 ), implying that biological consumption may also play a role in regulating the dynamics of mid-depth DO. 265 There is evidence that the size of tropical OMZs is largely influenced by biological processes, such as organic matter export, remineralization and oxygen consumption (Cavan et al., 2017;Keller et al., 2016). Our further analyses demonstrate that biological consumption generally decreases in majority of the water column under enhanced background diffusion, with the largest decrease (~0.6-0.8 mmol m -3 yr -1 ) over ~600-800 m (Figure 11c), which is greater than the increased rate (~0.3-0.5 270 mmol m -3 yr -1 ) of physical supply (Figure 11b). Enhanced background diffusion leads to remarkable decrease of DON over ~300-900 m, with a larger decrease in the ETNP (Figure 11d). Earlier field studies revealed that DON concentration in the euphotic zone is much higher to the north than to the south in the central-eastern tropical Pacific (Libby and Wheeler, 1997;Raimbault et al., 1999). Later studies showed that rates of DOM remineralization and oxygen consumption are also greater at mid-depth in the ETNP than in the ETSP (Feely et al., 2004;Kalvelage et al., 2015;Tiano et al., 2014). Apparently, the 275 asymmetric OMZs in the tropical Pacific are attributed to not only physical processes, but also biological processes.

Conclusion
This paper describes an evaluation and validation of a fully coupled basin-scale model (OGCM-DMEC V1.0), focusing on the sensitivity of the asymmetric OMZs in the tropical Pacific to different parameterizations. Our results show that the 280 improved model with enhanced background diffusion successfully reproduces the observed asymmetric OMZs in tropical Pacific, indicating that distribution of DO at mid-depth is more sensitive to physical mixing (e.g., enhanced background diffusion) rather than biological processes (e.g., reduced DO utilization rate).
Our analyses suggest that DO concentration at mid-depth is significantly increased due to enhanced background diffusion, 285 resulting from changes in both physical supply and biological consumption. On the one hand, physical supply of DO is largely increased in the majority of the tropical Pacific, but shows little change at mid-depth in the ETNP. On the other hand, biological consumption shows a significant decline over 300-1000 m and an increase below 1000 m, as a result of the redistribution of DOM in the water column.

290
Our further analyses demonstrate that the asymmetric OMZs in the tropical Pacific are largely associated with asymmetry in both physical supply and biological consumption. In particular, physical supply plays a dominant role for the asymmetry of core OMZs (~200-600 m) while biological consumption has impacts on the asymmetric DO below the OMZs, with implications for the hemisphere asymmetric OMZs. Future studies should consider both physical and biological aspects to deliver comprehensive assessments on the interactions between DO and DOM, which is critical for a better understanding of 295 variability and drivers of the tropical asymmetric OMZs.
Author contributions. X.J.W. and K.W. designed the study, performed the simulations and prepared the manuscript. R.M., D.X.Z. and R.H.Z. contributed to analysis, interpretation of results and writing.

305
Competing interests. The authors declare that they have no conflict of interest.