Lunar evolution models predict that fractional crystallization of a lunar magma ocean (MO) produces a melt layer enriched in incompatible elements (KREEP) and that that the global surficial distribution of Th is a proxy for the distribution of KREEP. Th and FeO distributions were used to divide the crust into three main Terranes: Procellarum KREEP (PKT), Feldspathic Highland (FHT) and South Pole-Aitken (SPAT). There are obvious high-Th locations in PKT, but almost none in FHT. The lateral extent/distribution of the residual MO “KREEP” layer is currently a matter of debate. Asymmetric Th distribution suggests that the Procellarum region may be the oldest impact basin, causing accumulation of KREEP-rich residual liquid on the nearside; the lunar farside lacks the same high Th abundances and other KREEPy elements. Other mechanisms have been proposed to explain the asymmetry of KREEP materials (inhomogeneous MO differentiation; antipodal impact effects). We used superposed impact craters/basins to document the 3-D extent of Th in PKT/SPA to establish KREEP geometry/evolution: These relationships suggest that the high Th distribution in PKT/SPA is related to sequential oblique and near-vertical impacts. 1) An oblique impact formed a Procellarum basin, removing much of the upper/middle crust and bringing the residual KREEP-rich layer closer to the surface. 2) Formation of SPA by oblique impact similarly removed much of the upper/middle crust; the KREEP-rich lower crust was covered by SPA impact melt. 3) The more-vertical, but still oblique, Imbrium impact penetrated into the now-shallow KREEP layer, ejecting it to form the FMF and related Th-rich ejecta/impact melt deposits. 4) Later mare volcanism buried these KREEP-rich deposits. 5) In SPA, vertical impacts Brikeland and Oresme V excavated Th-rich material below the melt sheet to create locally high Th anomalies. In summary, this interpretive scenario attributes the origin of the PKT and SPA Th anomalies to sequential impact processes, requiring no additional processes to concentrate KREEP in the SPA or PKT terrain at depth. Analysis of the young Procellarum mare basalts returned by Chang’E 5 will provide a test of this hypothesis.