Abstract
<jats:p>Recent advances in paleoclimate data assimilation (paleo-DA) have produced gridded reconstructions of Indian monsoon hydroclimate variability over the Common Era, yet their consistency with one another and with instrumental records has not been systematically investigated. Here we evaluate the Monsoon Asia Drought Atlas (MADA), a statistical climate field reconstruction (CFR), and three paleo-DA datasets&mdash;the Last Millennium Reanalysis (LMR), the Paleo-Hydrodynamics Data Assimilation (PHYDA), and the Modern Era Reanalysis (ModE-RA)&mdash;against observed precipitation and instrumental Palmer Drought Severity Index (PDSI) datasets over South Asia from 1900&ndash;2000 C.E. We assess interannual and decadal hydroclimate variability across two subdomains, the Core Monsoon Zone (CMZ) and southern India (SI), and compare associated sea-surface temperature (SST) patterns using composite and spatial correlation analyses. ModE-RA precipitation and MADA PDSI align best with observations, though both rely on input instrumental data, which limits the independence of this agreement. LMR, relying on a sparser last-millennium proxy network, shows weak interannual skill but moderate decadal agreement over the CMZ. PHYDA exhibits the weakest agreement with instrumental PDSI and yields decadal monsoon variability and SST correlations opposite in sign to observations, despite a denser regional proxy network. Paleo-DA products perform poorly for the SI region, whereas they capture broad features of SST-monsoon coupling, albeit with biases in magnitude, coherence, and phasing across both timescales. We conclude that the accuracy of monsoon&ndash;hydroclimate reconstructions from paleo-DA products depends on the specific product, variable, and timescale considered, with assimilation methodology playing a primary role relative to regional proxy density.</jats:p>