RedAnDA TS, UnivarS and UnivarT are the different versions of RedAnDA, applied respectively to temperature-salinity (coupled), salinity and temperature monthly reconstructions. The coupled approach uses bivariate 3D EOFs with density-weighting, while the univariate version use simple 3D EOFs. In RedAnDA, the system state is described by the reduced set of EOFs, as in reduced-space interpolation. Additionnaly, it uses an Ensemble Kalman Smoother scheme associated with the Analog method, a statistical model based on the Analog method. ISAS20 (an Argo data objective analysis, spanning from 2002 to 2020) is used to estimate the spatial covariability between ocean temperature and salinity, to build the Analog model, which propagates information from the future to the past with a time step of one month, and the monthly climatology from which anomalies are calculated. In RedAnDA, in situ profiles from the EN.4.2.2 dataset are assimilated. It is a product of all types of in situ measurements, with bias corrections and quality assessment. RedAnDA products aim at enhancing the representation of interannual variability in the historical reconstruction of the tropical Pacific.