replying to A. W. Jacobel et al. Nature Geoscience (2019)

We disagree with the arguments put forth by Jacobel et al.1 and stand by our original interpretations2. Jacobel et al.1 assert that no evidence of Fe fertilization by dust is provided by Loveley et al.2 and claim the same for their dataset from nearby Ocean Drilling Program (ODP) Site 1240. However, we interpret their dataset differently: there is a moderate, statistically strong positive correlation between the excess Ba (xsBa) and 232Th fluxes (coefficient of correlation, r = + 0.49, P < 0.01; fig. 1a in ref. 1). The relationship between their 232Th and Fe fluxes is even stronger (r = + 0.62, P < 0.01; Fig. 1b in ref. 1), indicating a significant influence of dust on Fe availability.

Fluxes of xsBa scale nonlinearly with dust flux at site MV1014-02-17JC. We stated that, and noted a clear relationship between xsBa and dust “during or near Heinrich Stadials (HS) 1, 2, 5, 6 and 7”2. For HS6 and 7 there are statistically significant positive correlations between xsBa and dust fluxes (r = + 0.47, P < 0.05 and r = + 0.96, P < 0.01, respectively). For HS1, 2 and 5, xsBa is likely affected by bottom-water hypoxia, defined by authigenic U values that are the highest in the entire record (>10 ppm)2. The low-O2 bottom waters probably released Ba3, which diffused downwards slightly and re-precipitated, causing the xsBa flux to lead the dust flux signal by only ~1 kyr.

We see no evidence in Jacobel et al.1 in support of their claim that upwelling of equatorial undercurrent waters provides all of the Fe and is the sole cause of fertilization in the easternmost eastern equatorial Pacific (EEP). During past cold events, similar to boreal winter conditions today, a southward-shifted intertropical convergence zone probably reduced equatorial upwelling4. Regardless, Fe concentrations in the eastward-flowing equatorial undercurrent, sourced from continental inputs in the western Pacific, are nearly zero by 110° W due to scavenging and the short residence time of Fe (0.8–1.0 nM at 140° W to ≤0.09 nM at 110° W)5. Even farther east, the Galapagos Islands obstruct, deflect and weaken the equatorial undercurrent6, making the role it plays in supplying Fe to our site at 86° W even less likely. It is apparent that our site would be more sensitive to dust fertilization due to continental proximity, which leads to dust fluxes that are about 5–10 times greater than those at all 11 locations considered by Jacobel and colleagues1. In fact, east of the Galapagos Islands2,7,8,9,10,11,12, changes in aeolian delivery and its role in export production and the global CO2 cycle cannot be ruled out. We therefore find the assertions1 that there is unquestionably no dust fertilization at our site or theirs to be unwarranted.

The EEP is the greatest oceanic source of CO2 to the atmosphere today13, yet an abundance of data show that the easternmost EEP may have been a net sink of CO2 at times during the last deglacial and glacial periods7,9,10,11,12. The argument1 against net changes in biological pump efficiency and atmospheric \(p_{{\mathrm {CO}}_2}\) drawdown relies on nutrients sourced from equatorial upwelling, resulting in surface waters with elevated CO2 concentrations. This assumption fails to consider that Fe fertilization from dust can increase the efficiency of the biological pump and carbon sequestration in a high nutrient–low chlorophyll region, even with no net changes in upwelling.