Zonal Indian Ocean Variability Drives Millennial‐Scale Precipitation Changes in Northern Madagascar

Abstract

The low latitude Indian Ocean is warming faster than other tropical basins, and its interannual climate variability is projected to become more extreme under future emissions scenarios with substantial impacts on developing Indian Ocean rim countries. Therefore, it has become increasingly important to understand the drivers of regional precipitation in a changing climate. Here we present a new speleothem record from Anjohibe, a cave in northwest (NW) Madagascar well situated to record past changes in the Intertropical Convergence Zone (ITCZ). U‐Th ages date speleothem growth from 27 to 14 ka. δ18O, δ13C, and trace metal proxies reconstruct drier conditions during Heinrich Stadials 1 and 2, and wetter conditions during the Last Glacial Maximum and Bølling–Allerød. This is surprising considering hypotheses arguing for southward (northward) ITCZ shifts during North Atlantic cooling (warming) events, which would be expected to result in wetter (drier) conditions at Anjohibe in the Southern Hemisphere tropics. The reconstructed Indian Ocean zonal (west‐east) sea surface temperature (SST) gradient is in close agreement with hydroclimate proxies in NW Madagascar, with periods of increased precipitation correlating with relatively warmer conditions in the western Indian Ocean and cooler conditions in the eastern Indian Ocean. Such gradients could drive long‐term shifts in the strength of the Walker circulation with widespread effects on hydroclimate across East Africa. These results suggest that during abrupt millennial‐scale climate changes, it is not meridional ITCZ shifts, but the tropical Indian Ocean SST gradient and Walker circulation driving East African hydroclimate variability.

Zonal control on Holocene precipitation in northwestern Madagascar based on a stalagmite from Anjohibe

“…Holocene paleoclimate in northwestern Madagascar cannot, therefore, be explained by southward, meridional migration of the ITCZ and tropical rain belt as expected in the Global Paleomonsoon paradigm. Studies in Madagascar at multi-decadal scale in the late Holocene 6 and during Heinrich Event 1 66 have also noted in phase climate variations between Madagascar and Northern Hemisphere records and suggested that east to west SST gradients in the Indian Ocean, with warming in the west relative to the east, are associated with greater rainfall in the study area. In addition, studies in East Africa 13 , 67 , 68 have found that, on orbital to centennial timescales, western Indian Ocean SSTs play a dominant role in enhancing moisture transport and rainfall over the region, similar to what is observed during the annual cycle today 69 .…”

Opposite Hydrological Conditions between the Younger Dryas and the 8.2 ka Event Revealed by Stalagmite from Northwest Madagascar in East Africa

“…This suggests a similar climatic response over the study area to those of the HS1 and YD, both of which were triggered by large amounts of freshwater influx into the North Atlantic. Following the HS1 [91] and before the YD (this study), although the records are not consecutive based on this study and Ref. [91], a substantial negative rebound was documented, corresponding to a wet B/A period.…”

“…Following the HS1 [91] and before the YD (this study), although the records are not consecutive based on this study and Ref. [91], a substantial negative rebound was documented, corresponding to a wet B/A period. This dry HS1 and YD and wet B/A record is identical with the expectation in Northern Hemisphere, but antiphase with the interhemispheric "see-saw" pattern on the millennial scale [30,93,94].…”

“…Indeed, the hydroclimatic variability over tropical East Africa is remarkably dominated by the Norther Hemisphere climate signal during the last deglaciation. In addition to the YD, stalagmite records from the same cave reveal prominently positive δ 18 O excursions during the Heinrich Stadial 1 (HS1) [91], which was believed to result from the weakening/shutdown AMOC, consistent with the large-scale drought conditions in the entire Southeast Africa [92]. This suggests a similar climatic response over the study area to those of the HS1 and YD, both of which were triggered by large amounts of freshwater influx into the North Atlantic.…”

“…(l) U K ′ 37 derived SSTs from the marine sediment core in the Mozambique Channel [97]. (m) Tropical western Indian Ocean SSTs using the Last Glacial Maximum reanalysis [91]. (n) Marine core (CD154-10-06P) elemental ratios of Fe/K, KwaZulu-Natal [96].…”