Zhang et al. [
1] argue that changes in northern hemisphere summer insolation (NHSI) caused by the precession of Earth’s orbit have controlled the timing and pace of the late Pleistocene East Asian summer monsoon (EASM) and Indian summer monsoon (ISM). Since the early 2000s, several high-resolution cave
δ18O records from China dated with unprecedented precision have been published [
2,
3,
4,
5] and are widely interpreted as proxies for East Asian summer monsoon (EASM) intensity [
5]. Modern observations, however, do not demonstrate a clear relationship between the amount of precipitation and the
δ18O signature in cave drip water [
6,
7,
8]. There is no modern analogue for the spatially homogenous ‘rainfall’ patterns inferred from the cave
δ18O records [
6], and mass-balance calculations demonstrate that the range of cave
δ18O variations requires unfeasibly high changes in rainfall [
9]. The cave
δ18O records are also inconsistent with other published proxy records of EASM rainfall based on Chinese Loess magnetic records [
9] and beryllium isotopes [
10]. The interpretation that negative excursions in the cave
δ18O records reflect changes in the amount of monsoon rainfall alone is only partially supported [
11,
12] and remains widely contested [
13]. The Zhang et al. study also does not address evidence from recently published marine and terrestrial proxy records of monsoon precipitation from the EASM and ISM domains [
14,
15], which are at odds with the major conclusions drawn therein [
1].
Early reconstructions of Asian summer monsoon intensity were based on proxies of wind strength over the Arabian Sea [
16,
17,
18]. These results were interpreted to infer changes in the intensity of monsoon precipitation over land given that strong summer monsoon winds promote upwelling-driven productivity in the Arabian Sea [
19], and transport large amounts of moisture to the Asian continent [
20]. More recently, seawater
δ18O (
δ18Osw) records from important monsoon sink regions such as the East China Sea [
14] and the Andaman Sea [
15] have been published (
Figure 1). These reconstructions reflect changes in regional rainfall integrated over large river basins and the open ocean. Gebregiorgis et al., [
15] showed that Pleistocene Andaman Sea
δ18Osw variability has the same late precession-band phase as the Arabian Sea wind records [
16,
17,
18], confirming the link between the summer monsoon winds in the Arabian Sea and rainfall in the larger Bay of Bengal region. Unlike speleothem
δ18O, the Arabian Sea and Andaman ISM records lag NHSI maxima by ~8 kyrs, which delineates ISM sensitivity to global ice volume as well as Southern Hemisphere moisture and energy export [
16,
17,
18]. The phase wheel diagram in
Figure 2 illustrates the precession phasing of marine and speleothem based EASM and ISM reconstructions.
The conclusion of Zhang et al. [
1] that EASM and ISM variability on orbital timescales has been driven directly by changes in Northern Hemisphere summer insolation without significant temporal lags is entirely inconsistent with the EASM
δ18Osw reconstruction from the East China Sea [
14]. This record shows no concentration of variance in the precession band once the temperature and ice-volume influence on foraminiferal-calcite
δ18O are removed and, hence, cannot be plotted on
Figure 2. It is also inconsistent with the Andaman Sea
δ18Osw record, which closely follows the timing of the Arabian Sea precession band phase [
15]. Zhang et al. do not present a discussion of these strongly divergent results and interpretations, which were both derived from the same isotopic system and archive (
δ18O of CaCO
3) in the same monsoon regions. Given that large-scale atmospheric circulation controls the transport of moisture between sources and sinks, the resultant speleothem
δ18O signal likely reflects changes in moisture sources and pathways [
9,
21] or land-surface temperature or other processes [
22], whereas the seawater
δ18O signature in marginal basins reflects the amount and timing of regional monsoon rainfall.