Inferring Long-Term Tectonic Uplift Patterns from Bayesian Inversion of Fluvially-Incised Landscapes

Earth surface processes encode the combined forcing of tectonics and climate in topography. Separating their contributions is essential for using landscapes as quantitative records of crustal deformation. Here, we develop a method to invert non-dimensionalized, spatially variable fields of long-term rock uplift and rock erodibility from fluvially incised landscapes, using an extended χ-coordinate system that accounts for variability in uplift, erodibility, and precipitation (χ_UKQ) . We invert 170 synthetically-generated landscapes and demonstrate that our method accurately recovers the spatial variability of rock uplift and rock erodibility, even when applied to settings that deviate from the ideal model of equilibrated, detachment-limited channels, which underpins the χ_UKQ framework. We subsequently apply our inversion to six natural landscapes shaped by normal faults (half-grabens), and to a 200-km wide region of the central Himalayas. We show that our inversion can resolve the effect of climate and lithology while extracting uplift fields that are consistent with patterns expected from upper crustal flexure and previous estimates derived from geomorphological markers. The success of our method in recovering rock uplift patterns, isolated from the effects of climate and erodibility, highlights its applicability to settings where long-term uplift trends are unknown, paving the path to deciphering tectonic fingerprints recorded in landscapes over tens of thousands of years.