Chapter 1:

A Mechanistic Investigation of the Coral Mn/Ca-based Trade-wind Proxy at Kiritimati

For the full manuscript at Geochimica et Cosmochimica Acta:


Trade winds help store heat and slow global warming

Trade-wind strength across the Pacific Ocean can change in accordance with decadal climate variability such as Interdecadal Pacific Oscillation (IPO), shown here as zonal wind stress trends. Trade winds are weaker during positive IPO phases (+IPO) and stronger during negative IPO phases (-IPO). Stronger trade winds during -IPO lead to greater heat storage in the ocean, resulting in a slow-down of global warming. Weaker trade winds during +IPO lead to less heat storage, resulting in accelerated warming.

Here you can see the periods of accelerated and stalled global warming resulting from the changes in trade-wind strength during +IPO and -IPO phases, respectively. Changes in the direction and speed of trade winds in the Pacific occur not only on decadal timescales like we see here, but also on interannual timescales due to El Niño-Southern Oscillation. This highlights the key role that winds play in our climate system, and sets the stage for why it is critical to study past wind behavior.

El Niño: initiated & sustained by Westerly Wind Events (WWE)

As an El Niño event develops and Walker circulation weakens, Pacific trade winds (which normally blow from east to west) also weaken and can even reverse in their direction.

This reversal of easterly trade winds, or Westerly Wind Events (WWEs), is shown in this animation as positive zonal winds. The splashes of colors indicate WWEs that occurred before and during the very strong El Niño event of 1997-1998. These WWEs initiate the movement of warm water from the West Pacific Warm Pool towards the east, and the continued propagation of these westerly winds sustain the anomalously warm sea surface conditions of an established El Niño event.

This link between Pacific trade winds and climate phenomena such as El Niño and IPO drives interest in studying trade-wind patterns in the equatorial Pacific and finding ways to fill in the temporal and spatial gaps in the historical wind record.


Coral skeletons are archives of past climate

Coral skeletons act as high-resolution geochemical archives for surface ocean conditions because they grow quickly and continuously. As the coral grows, it incorporates certain trace elements into its calcium carbonate skeleton, whose concentrations reflect changes in climate parameters. A few recent studies have shown that the manganese-to-calcium ratio (Mn/Ca) of reef-building corals located at remote, low-lying equatorial Pacific atolls with west-facing lagoons can be used to reconstruct trade-wind patterns.

These studies were conducted at Tarawa, Butaritari, and Kiritimati – all atolls with west-facing lagoons (Fig. 1).

Figure 1: Google satellite imagery of (a) Butaritari, (b) Tarawa, and (c) Kiritimati atolls, with (d) a map of the equatorial Pacific Ocean for context. Sites of previously analyzed samples are numbered according to their associated study. Lagoon (blue; LAG20-C02) and lake (green; C22A, C22B) sediment cores analyzed in this study are marked by colored triangles, with colors corresponding to sediment and porewater Mn records in Fig. 5. The site from which lagoon water was collected weekly from December 2018 - April 2019 is indicated by a white star. Transect A-B is linked to the cross-section schematic in Fig. 2. Map data: GoogleMaps and Terrametrics (2020).

Coral Mn/Ca spikes signal reversals in trade-wind direction

Figure 3 from Shen et al. (1992): A chemical indicator of trade wind reversal in corals from the western tropical Pacific.

Shen et al. (1992) found that spikes in Mn/Ca measured in a coral at Tarawa atoll (Fig. 1b) were concurrent with positive zonal winds (westerly winds) associated with the El Niño events of 1965-1966, 1968-1969, 1972-1973, and 1976-1977.

Thompson et al. (2015) extended this record at Tarawa to span a century (1890-1990), and continued to see a link between coral Mn/Ca and westerly wind events (WWEs) back in time.

Figure 2 from Thompson et al. (2015): Early twentieth-century warming linked to tropical Pacific wind strength

The proposed mechanism for this coral-based trade-wind indicator at sites with large west-facing lagoon consists of 5 reservoirs of manganese (Fig. 2):

  1. dust
  2. sediment
  3. porewater
  4. seawater
  5. coral

Figure 2: Schematic detailing the mechanism behind the coral Mn/Ca-based trade-wind proxy at Kiritimati, adapted from the mechanism originally proposed by Shen et al. (1992) at Tarawa. This cross-section is linked to transect A-B in Fig. 1.

Timing and magnitude of coral Mn/Ca spikes can be different

Despite the link between coral Mn/Ca and trade-wind behavior existing at multiple islands, Sayani et al. (2021) found that the properties of the Mn/Ca spikes are not always the same.

  • Tarawa: Mn/Ca spikes ~60 nmol/mol; concurrent with WWEs
  • Butaritari: Mn/Ca spikes ~400 nmol/mol; lag WWEs by ~1 year
  • Kiritimati: Mn/Ca spikes ~300 nmol/mol; lag WWEs by ~1 year

The proposed mechanism has not been tested beyond Tarawa...


... In this study, by analyzing water, sediment, and porewater samples from the main lagoon and lakes of Kiritimati, we trace the step-wise transformation of manganese as it migrates between seawater, sediment, porewater, and coral.

Key objectives:

1) Compare the Mn reservoirs that link coral Mn/Ca to westerly wind activity (i.e., lagoon sediments, porewater and seawater) at Kiritimati and Tarawa atoll.

2) Deconstruct the mechanism of Mn(II) accumulation in sediment porewater by analyzing the geochemistry of a Kiritimati lake as a case study.

3) Examine the spatiotemporal variability of lagoon and lake water Mn(II).

Why Kiritimati?

  • located at 1.9°N, 157.5°W, Kiritimati is in the path of trade winds and WWEs
  • has a large (190 km 2 ) west-facing lagoon
  • has an extensive network of interconnected, shallow lakes on its eastern side that are adjacent to the main lagoon

Lakes act as a case study -- lack of WWE-driven mixing allows certain steps of the mechanism to be isolated in the lakes, providing a window into Mn behavior in the lagoon.


Kiritimati vs. Tarawa: Water, Sediment, Porewater reservoirs

At Kiritimati, coral Mn/Ca background values are 2-3x greater and spikes are 14x greater than those at Tarawa. This suggests that there's a difference in the magnitude of Mn reservoirs of the two islands.

Figure 3: Depth profiles of (a) water, (b) lagoon sediment, and (c) corresponding sediment porewater Mn concentration at Kiritimati (solid line) and Tarawa (dotted line; Shen et al., 1992).

Kiritimati and Tarawa's water and lagoon sediment [Mn] are comparable (Fig. 3). However, Tarawa's porewater [Mn] is ~18x that of Kiritimati. Why? After evaluating a few potential explanations, the main driver of this difference is most likely the porewater "recharge time", or the amount of time that elapsed between when the core was sampled and when the most recent porewater resuspension event occurred. This "recharge time" is determined by the vertical reach of wind-driven mixing, which is dependent upon a combination of:

  • WWE strength
  • depth of the lagoon


Kiritimati's lagoon water [Mn] is spatiotemporally variable

Lagoon water [Mn] is elevated directly following a period of observed WWEs associated with the 2015-2016 El Niño event, whereas [Mn] is 7x lower during a period that does not overlap with or follow any WWEs (Fig. 7).

Figure 7: Time series of lagoon water Mn(II) concentration from March-August 2016 and December 2018 - April 2019 with zonal wind speed observations and precipitation data from Kiritimati. Lagoon water collected in 2018-2019 was sampled from site indicated by a white star in Fig. 1.

WWE activity and resulting porewater resuspension are the main drivers of the temporal variability of [Mn] in Kiritimati's lagoon.

The water [Mn] of Kiritimati's lakes are positively correlated with their salinity, likely due to both being controlled by the same hydrologic inputs (precipitation, groundwater intrusion) and outputs (evaporation). The most saline lakes and those with greatest water [Mn] tend to be the farthest inland and most isolated from the seawater flux from the lagoon/coast.

Figure 6: Map of Kiritimati with lakes shaded (December 2018, this study) and outlined (July 2005, Saenger et al., 2006) according to salinity category. White (black) circles mark locations where water was sampled from the lakes (lagoon) in December 2018, with sizes corresponding to Mn(II) concentration measured, ranging from 0.01-0.8 ppb. Bathymetric DEM of lagoon shows a depth range of 0.2-10.5 m.

Unlike the lakes, the lagoon's water [Mn] is not significantly correlated with salinity. Lagoon water [Mn] is generally lower than that of the lakes, and is much more homogeneous. One exception to this is that at the mouth of the lagoon, where two channels north and south of a small island connect the lagoon to the open ocean (labeled "Channels" in Fig. 6). In particular, sampling sites near the southern channel have greater [Mn] compared to those near the northern channel. Past records of dredging indicate that the northern channel experiences a faster sedimentation rate compared to the southern channel, suggesting a relatively slower flow rate. A faster flow rate through the southern channel implies greater turbulence and could possibly lead to a greater potential for resuspension of Mn-enriched sediment porewater.

Therefore, lagoon morphology may play a role in controlling the [Mn] of the water column, and thus how much reaches nearby coral colonies.


Now that we have a better understanding of how WWEs translate to porewater flux and ultimately to spikes in coral Mn/Ca, we can apply this proxy to corals from other sites across the equatorial Pacific (with west-facing lagoons, of course) to reliably reconstruct trade-wind behavior in the past, and to get a better sense of wind signatures west and east of the date line.

Stay tuned for more!


This work was published in Geochimica et Cosmochimica Acta with the following co-authors: Diane M. Thompson, Stephan R. Hlohowskyj, Jessica E. Carilli, Gwyneth Gordon, Tyler J. Goepfert, Hussein R. Sayani, Thomas M. Marchitto, Kim M. Cobb

Funding and support were provided by NSF OCE-1702130. Fieldwork in Kiritimati was conducted with the support of the Government of Kiribati’s Ministry of Environment Lands and Agricultural Development and the Ministry of Fisheries and Marine Resources Development, under Environmental Licenses 011/18 and 008/19 and Research Consent Certificates. We thank Mr. Teratau, Peter Kaitama, and Ann Burentia for their assistance in expediting permits, and we thank Tiito Teabi for collecting seawater samples in 2016, and Otea Ioteba for collecting weekly seawater samples from 2018-2019, and again in 2020-2021. 

Figure 1: Google satellite imagery of (a) Butaritari, (b) Tarawa, and (c) Kiritimati atolls, with (d) a map of the equatorial Pacific Ocean for context. Sites of previously analyzed samples are numbered according to their associated study. Lagoon (blue; LAG20-C02) and lake (green; C22A, C22B) sediment cores analyzed in this study are marked by colored triangles, with colors corresponding to sediment and porewater Mn records in Fig. 5. The site from which lagoon water was collected weekly from December 2018 - April 2019 is indicated by a white star. Transect A-B is linked to the cross-section schematic in Fig. 2. Map data: GoogleMaps and Terrametrics (2020).

Figure 3 from Shen et al. (1992): A chemical indicator of trade wind reversal in corals from the western tropical Pacific.

Figure 2 from Thompson et al. (2015): Early twentieth-century warming linked to tropical Pacific wind strength

Figure 2: Schematic detailing the mechanism behind the coral Mn/Ca-based trade-wind proxy at Kiritimati, adapted from the mechanism originally proposed by Shen et al. (1992) at Tarawa. This cross-section is linked to transect A-B in Fig. 1.

Figure 3: Depth profiles of (a) water, (b) lagoon sediment, and (c) corresponding sediment porewater Mn concentration at Kiritimati (solid line) and Tarawa (dotted line; Shen et al., 1992).

Figure 7: Time series of lagoon water Mn(II) concentration from March-August 2016 and December 2018 - April 2019 with zonal wind speed observations and precipitation data from Kiritimati. Lagoon water collected in 2018-2019 was sampled from site indicated by a white star in Fig. 1.

Figure 6: Map of Kiritimati with lakes shaded (December 2018, this study) and outlined (July 2005, Saenger et al., 2006) according to salinity category. White (black) circles mark locations where water was sampled from the lakes (lagoon) in December 2018, with sizes corresponding to Mn(II) concentration measured, ranging from 0.01-0.8 ppb. Bathymetric DEM of lagoon shows a depth range of 0.2-10.5 m.