This is a well written manuscript that appears to have two principal goals. First, there is presentation of a rich new dataset incorporating stratigraphic, sedimentologic, and provenance data for part of the Magdalena Valley basin in Colombia—this is a valuable and welcome contribution. Second, there is a provocative proposal involving speculation of a correlation between the Mag Valley and the Llanos foreland basin, with implications for the areal extent of the Miocene “Pebas” marine ingression—in my opinion, this is highly speculative and not necessary for the success of the manuscript.

My review below focuses on the more provocative elements of the study, where I am skeptical of the proposed regional correlations. However, please note that I find the datasets for the Mag Valley to be very convincing, valuable to understanding of the northern Andes, and worthy of publication. So please consider these two components individually.

(1) Regional correlations, Pebas seaway, marine ingression(s).

A key component in the regional correlations involves the assumption that the mid Miocene deposits in the Magdalena Valley intraorogenic basin were originally in topographic and depositional continuity with the Llanos foreland (and broader foreland of Ecuador, northern Peru, Brazil), such that the Pebas seaway was a continuous feature spanning the Eastern Cordillera (EC) of Colombia. This is challenging, because there seems no compelling reason to believe there was no EC topography and therefore believe the low-elevation Mag Valley was connected to the low-elevation Llanos basin at ~16-12 Ma. Of course, this speculation is possible, but there are several potential arguments to pursue. First, some evidence showing that there was not significant shortening prior to 12 Ma in the EC would be helpful. Second, and more importantly, evidence is required to demonstrate that the mid Miocene facies of the Mag Valley were contiguous with those of the Llanos basin. These lines of evidence would be critical, because most studies suggest the EC has been emergent from the latest Oligocene onward. As a suggestion, perhaps the authors can make a compelling case for the topographic development and basin connectivity across the southernmost part of EC and its interactions with Garzon Massif.

This study relies heavily on De la Parra et al. 2019 (Palaios), but that study finds no evidence for marine conditions in the Mag Valley, and no evidence for a direct link to possible seaways of the Llanos basin. The chronology from Jaramillo et al. 2017 (Science) shows a very broad 18-12 Ma timeframe for marine conditions in the Llanos. Therefore, the existence in the Mag Valley of lacustrine facies broadly of middle Miocene age is not very compelling evidence. The presence of lacustrine deposits in the Mag Valley, Llanos, and intervening EC (Floresta basin and similar areas) only seems to demonstrate a common situation with closed nonmarine drainage systems. There is no strong reason to associate these separate lacustrine facies with marine ingressions in other parts of South America.

Please note that these issue pertain to the Neogene record. The correlation of Cretaceous, Paleocene, and possibly Eocene units from the Putumayo / Llanos basin to the Upper Magdalena Valley (UMV) is well established by past studies (Alejandro Mora et al., 2010 AAPG Search & Discovery
https://www.searchanddiscovery.com/pdfz/documents/2010/50246mora/ndx_mora.pdf.html
and Wolaver et al. 2015 AAPG Bulletin).

(2) Neogene evolution of the Magdalena Valley basin

The authors emphasize that the westernmost basin does not contain the Honda Group, but rather must be re-defined as a new stratigraphic unit, which they refer to as the Rovira Formation. In fact, although this locality represents a proximal clastic wedge derived from a western source (Central Cordillera), it seems quite possible that it would be continuous with the Honda Formation farther to the east. The new data, in the form of precise depositional (stratigraphic) ages may well resolve this issue. So, although I would expect lateral continuity between the newly defined Rovira Formation (west) and Honda Group (east), I am willing to be convinced otherwise by the U-Pb geochronological data, in the form of maximum depositional ages (MDAs).

The presentation would benefit greatly from a regional stratigraphic profile, to support the time-stratigraphic statements and interpretations. This could be a cross-sectional profile with distance (X axis) vs. time or thickness (Y axis). This would provide critical context for the series of measured sections (Figs. 4-6).

Similarly, some discussion of the time resolution for Neogene palynomorphs is warranted. Globally, most nonmarine palynomorphs of Neogene age are not very age diagnostic, so please add some statements or details regarding the palynology age situation for Colombia. This is most critical for age interpretations and correlations / non-correlations among the Barzalosa, Santa Teresa, La Cira, and Honda Formations, along with the newly proposed Rovira Formation.

Similarly, for these 5 stratigraphic units, consideration of the lateral facies variations remains an important issue. Most of these units have local sources, suggesting that there could be very rapid lateral facies changes (i.e., lithologic changes over short lateral distances, say <10-50 km).

Finally, the low-temperature thermochronological data are not very definitive. Perhaps there is some alternative way to present the data, including a plot of the thermochron ages directly above a cross section, in order to show spatial similarities / differences along structural profiles. In addition, perhaps there is room for a more-direct integration of the cooling record with the stratigraphic and provenance histories.

Brian Horton
University of Texas at Austin
Austin, TX, USA