The Colorado River (CR) is one of the longest rivers of North America (2330 km), with a watershed spanning 640 000 km2 of eight US and Mexican states, making it a critical resource for the arid southwestern part of the continent. This dynamic modern river is imperiled today by overutilization of its water resource and a changing climate. Just as the river system today presents a rapidly unfolding story of environmental change, its past geological history along with the basins receiving its water and sediment present scientists with critical lessons for understanding possible future change.

A key archive of this past history can be found in the late Miocene(?) and early Pliocene Bouse Formation, which crops out and is present in the subsurface along the lower CR valley of western Arizona, southern Nevada, and southeastern California today (Figs. 1, 2) (Metzger, 1968; Buising, 1990; House et al., 2008). The Bouse Formation, which can range from less than 10 m thick in some outcrop exposures to over 250 m in subsurface wells, contains the depositional record of a number of large water bodies, and the scientific investigation of this unit has a legacy of controversy that continues to today. The largest of these water bodies, which occupied the Blythe basin and adjacent areas, covered over 10 000 km2 (Spencer et al., 2013), almost twice the size of the modern Great Salt Lake (Utah). For decades, geologists, paleontologists, and geochemists have been embroiled in a scientific debate as to the paleoenvironmental nature and age of these water bodies (e.g., Buising, 1990; Spencer and Patchett, 1997; McDougall, 2008; Spencer et al., 2008, 2013; McDougall and Miranda Martínez, 2014; Gootee et al., 2016; O'Connell et al., 2017; Bright et al., 2018a, b; Dorsey et al., 2018) and the implications of these variable interpretations for the history of the CR, the integration of the river into the evolving Gulf of California, and regional tectonic history, including both the southern basin and range region and the Colorado Plateau (e.g., Bennett et al., 2016; Karlstrom et al., 2017; Pearthree and House, 2014; Crow et al., 2019a).

The most intense debate has focused on the origin of the basal Bouse Formation and exposures in the Blythe basin. This unit has variably been interpreted as having formed in marine (initially unaffected by CR water, but later estuarine) (Buising, 1990; McDougall and Miranda Martínez, 2014; Dorsey et al., 2018; Crossey et al., 2015) or lacustrine (Spencer and Patchett, 1997; Roskowski et al., 2010; Bright et al., 2016, 2018a, b) conditions, or in some hybrid of these scenarios evolving over time, based on conflicting evidence from fossils, isotope geochemistry, and sedimentology (Fig. 3). The fossil record of the lower Bouse Formation includes a bewildering mixture of apparently marine or estuarine (e.g., planktic foraminiferans, barnacles, and Thalassinoides trace fossils) and lacustrine (ostracodes, Chara, molluscs) fossils (Bright et al., 2018b). Sigmoidal bedding and other depositional features interpreted to be of marine tidal origin (O'Connell et al., 2017) were also interpreted by Spencer et al. (2018) to indicate a lacustrine origin. Isotopic data (C, O, and Sr) from Bouse carbonates uniformly point towards a continental water source, either similar to or modified from CR water, with little indication of a marine influence (Spencer and Patchett, 1997; Bright et al., 2016, 2018a, b). Two broad classes of models have arisen from these interpretations of the Blythe basin deposits, an incursion (or multiple incursions) of marine waters from an evolving proto-Gulf of California (McDougall and Miranda Martínez, 2014; Dorsey et al., 2018) or a “fill and spill” scenario of CR waters making their way downstream below the Grand Canyon by infilling a series of pre-existing basins as large lakes, which eventually infill with sediment and overtop their sill thresholds, allowing the CR to extend its length downstream towards an ultimate interconnection with the gulf (House et al., 2008; Spencer et al., 2013). Most participants in this debate now agree that Bouse Formation deposits in basins upstream from the Blythe basin, while displaying grossly similar stratigraphies and geochemical signatures to that basin, were all fully lacustrine through their history. Thus, it is only the Blythe basin that is currently at the core of this controversy.

In addition to this enigmatic history, the Bouse Formation, as well as the alluvial deposits that overlie it (e.g., Bullhead Alluvium, Howard et al., 2015; Chemehuevi Formation, Malmon et al., 2011; and Blythe Alluvium, Block et al., 2019; Fig. 1), house an outstanding and largely still untapped record of the history of both the CR since the Pliocene and climatic conditions during the early Pliocene. The latter is of particular significance, as this time period represents an initial episode of warming (e.g., Drury et al., 2018) towards the Piacenzian (mid-Pliocene warm period), which may be analogous to our modern anthropogenically driven transition to a warmer planet. Understanding the history of the CR, a critical water resource for desert southwestern North America, through its downstream Bouse Formation paleorecords is thus of value for society as a whole.

It was through this lens of multiple lines of scientific interest in the Bouse Formation that a workshop was held 28 February–3 March 2019 in Parker, AZ, USA, to discuss the possibility of developing a scientific drilling project targeting this formation and its overlying CR-related strata. Over the course of 4 d, 33 participants debated whether, where, and how a drilling and coring campaign might proceed that could address (and perhaps resolve) the fundamental controversies surrounding the origin of the Bouse Formation in the Blythe basin and how high-resolution drill core records might improve our understanding of the paleoclimate and paleohydrology of this region.

The workshop opened with a series of background talks intended to lay out our current understanding of the Bouse Formation, its tectonic setting, stratigraphy and sedimentology, paleontology, age, and geochemistry. Karl Karlstrom and Jacob Thacker (University of New Mexico) and Vicki Langenheim (USGS) discussed our current understanding from isostatic modeling of syn- and post-Bouse deformation, structural mapping, and geophysical surveys. Isostatic modeling is useful for estimating the original vertical position of key features related to the Bouse Formation, such as shorelines. Understanding the depth and shape of the Bouse depocenters (particularly those lying below sea level today) and their deformation history is critical for decision-making on appropriate drilling targets based on variable predictions of the marine incursion versus lake-fill and spill models. Understanding Bouse basinal strain history from outcrop fault geometries is key to understanding how tectonics has influenced lower CR deposits before, during, and after Bouse deposition. Regional gravity and magnetics data can also help inform us about potentially favorable depocenters for drilling thicker Bouse and post-Bouse stratigraphic sections.

The stratigraphy and sedimentology of the Bouse Formation have been a major focus of work for several research groups represented at the workshop, including Kyle House and Ryan Crow (USGS); Phil Pearthree and Brian Gootee (Arizona Geological Survey); Becky Dorsey, Brennan O'Connell, and Kevin Gardner (University of Oregon); and Colleen Cassidy (University of Arizona). House, Pearthree, and Gootee discussed the general stratigraphic framework of the early Pliocene (standing water and deltaic) Bouse Formation and post-Bouse (Pliocene–Holocene) alluvial deposits formed by the through-flowing Colorado River. They reviewed evidence from a series of N–S basins for the serial decantation of solute-rich CR waters with sediments stored in each basin until the delta wedge surmounts the outlet sill, at which point the outlet converted each successive lake into a valley, removing the drainage divide and allowing rapid incision in the upstream basin. The basal deposits of the Bouse in each successive basin are carbonates, sitting over buttressed unconformities at variable elevations (inferred by Pearthree and others to represent the rapid transgression of a series of lakes). This process was repeated multiple times as the CR worked its way southwards. This “decant–deposit delta–degrade–repeat” model (Fig. 3) provides a testable hypothesis for high-resolution drill core records. Above each basal carbonate is a sequence of siliciclastic fine-grained deposits, overlain by coarser-grained gravels and conglomerates that record the successive CR deltaic and fluvial aggradation. Dorsey, O'Connell, and Gardner presented the case for a marine (and specifically marine macrotidal) depositional setting for the lower Bouse carbonate member in the Blythe basin, based on the presence of sigmoidal and compound dune bedforms, rhythmites inferred to have tidal periodicities, flaser bedding, and marine-like cements, as well as trace and body fossils discussed below. Cassidy discussed the joint USGS–UA project to digitize all existing log data (extensive but of variable quality) from research and water wells throughout the lower CR valley. These data are being used to develop realistic isopachs and structure maps of the Bouse, locating wells where the controversial basal carbonate was encountered and, from that information, predict optimal drilling targets for thick Bouse and post-Bouse sections (Fig. 4). Natalia Zakharova (Central Michigan University) discussed how downhole logging might be integrated into a future drilling campaign for the Bouse with drill core measurements of stratigraphy, and Anders Noren (University of Minnesota Continental Scientific Drilling Coordination Office) provided a “Drilling 101” primer for the many outcrop-focused geoscientists in the room.

Evidence from fossils and geochemistry weighs heavily in the Bouse debate but, like much else about the Bouse, is highly controversial. Kris McDougall-Reid (USGS) and Steve Hasiotis (University of Kansas) presented paleontological evidence from foraminiferans (some planktic species unknown from lakes) and trace fossils (e.g., burrowing shrimp typical of macrotidal flats) in support of a marine interpretation for the lower Bouse. McDougall also argued, based on biostratigraphic evidence, for a late Miocene age of the lowermost carbonates predating the arrival of CR waters. Scott Staratt discussed the potential for using diatoms to resolve the Bouse controversy. The one outcrop locality of Bouse which has been studied in detail for diatoms (well west of the main outcrop belt) contains a mix of marine and freshwater species. Similarly, Jessica Tierney (University of Arizona) discussed the potential for organic geochemical biomarkers to resolve the origins of the lower Bouse, as well as for paleoclimate studies discussed below. In contrast, Jordon Bright (Northern Arizona University) presented both stable isotope and faunal evidence from ostracodes indicating deposition of the lower Bouse in a consistently lacustrine water body, which was at times stratified and saline, but without detectable marine influence. Similarly, Laurie Crossey (University of New Mexico) studied the lowest Bouse carbonate unit, a “traver-tufa”-present ubiquitously draping pre-Bouse bedrock, which isotopically, morphologically (apparently deposited subaqueously), and paleontologically (presence of carbonate Chara algal casts) is consistent with a lacustrine origin interpretation.

Given the fill and spill framework (which is uncontested from the upstream basins) and evidence for marine incursions in the southern Blythe basin, it is clear that good geochronological control in multiple basins will be critical for evaluating the various models of how the Bouse Formation was formed. Ryan Crow and Keith Howard (USGS), Shannon Dulin (University of Oklahoma), and Steve Kuehn (Concord University) discussed the most promising dating approaches based on Ar/Ar, magnetostratigraphy, and tephrostratigraphic data collected in basin margin outcrops to date. For example, recent advances in the combined application of 40Ar/39Ar dating of tephras and detrital sanidine with magnetostratigraphy were discussed that suggest the CR arrived in upstream basins between 5.24 and 4.6 Ma and at the Gulf of California between 4.8 and 4.63 Ma (Crow et al., 2019b); this is about half a million years later than previously suggested. Expanded sections, which could be provided by basin center drill cores of the Bouse and post-Bouse sediments, could provide an increased probability of collecting tephras and detrital sanidines in a stratigraphic context and a more complete and accurate paleomagnetic record. The fill-and-spill model for the CR all the way south to the uncontested marine-influenced region south of the Chocolate Mountain divide makes clearly differentiated geochronological predictions for the basal Bouse Formation and onset of CR-derived siliciclastics in each basin relative to the marine incursion model. Additionally, establishing the duration of Bouse deposition is critical to determining the timescale over which the lower Colorado River, a continental-scale river, became integrated with the ocean.

The potential of the Bouse Formation to provide a detailed paleoclimate record for the southwest during the early Pliocene had not been explored prior to this workshop. However, the age of the unit coupled with the potentially time-rich record (∼100–500 kyr) during an intriguing period in Earth's history, make the Bouse a particularly compelling target for such studies. In addition to Tierney's suggested biomarker studies (in particular glycerol dialkyl glycerol tetraether (GDGT) investigations of paleotemperatures and compound specific isotopic studies of leaf waxes in the fine-grained Bouse units), the potential for fossil pollen studies of paleoclimate was discussed by Vania Stefanova (University of Minnesota), and clumped/triple oxygen isotopes were discussed by Karl Lang (Queens College) and Dan Ibarra (Stanford University), all of which seem promising approaches. Alison Smith (Kent State University) discussed other initiatives for obtaining Pliocene paleoclimate records from drill cores situated further north in western North America and how they might complement results that could arise from drill cores from the lower CR valley.

A Bouse–CR drilling project could have substantial and societally significant education and outreach dimensions. The lower CR is at the nexus of significant water rights controversies, following years of extended drought conditions in the CR basin, and both paleorecords of the CR and new subsurface information about post-Bouse aquifers could be of considerable interest. Local stakeholders from the Colorado River Indian Tribes and the Cibola National Wildlife Refuge (both of whose lands include potential drilling targets) and local municipalities were either present at the meeting and field trip or have been involved in discussions about the project to date. Amy Myrbo (University of Minnesota) discussed education and outreach programs (some derived from other successful past drilling and coring projects) that could serve as models for a future Bouse–CR-related project, focusing around training or display opportunities for local stakeholders' underserved communities which are chronically under-represented in STEM fields and especially the geosciences.

Following the presentations discussed above, breakout sessions, plenary discussions, and a field trip to visit key outcrop and potential drilling sites allowed workshop participants to consider whether and where moving forward with a drilling campaign in the lower CR valley was warranted (Fig. 5). There was a strong consensus that a series of drill cores from depocenter sites along a N–S axial transect of the Blythe and adjacent basins could help address both the key outstanding paleoenvironmental controversies surrounding the Bouse Formation. They could also provide a seminal record of both early Pliocene paleoclimate for southwestern North America and the Colorado River's paleohydrologic history (Fig. 4). In all cases the objective would be to obtain cores where the Bouse Formation is both thick (maximum probable duration and temporal resolution) and fine-grained and where the controversial basal carbonate units are both present and fossiliferous. Fortunately, indications of all of these conditions can be obtained from the compiled water or geotechnical well driller's cuttings logs (no continuous cores are available). The specific targets which seem most promising (and to provide data unlikely to be recorded in basin marginal outcrops alone) are as follows.

Southern Blythe basin. This is probably the most critical drill site, as it would be in close proximity to many of the outcrops (e.g., Fig. 5b) that are most central to the Bouse debate. Because of its probable lower paleo-elevation, an expanded section in the depocenter here would provide a record of the entire transgressive infilling of the basin by whatever type of standing water body was present.