E. B. Safran

Spatially Variable Exhumation Rates in Orogenic Belts: An Andean Example

The Cordillera Real of the Bolivian Andes is a large, tectonically active mountain range that dominates sediment influx into the Amazon Basin, but rates of exhumation in the orogen are poorly known. We present 20 new apatite fission track ages from two valleys in the Cordillera Real to constrain patterns of mountain range exhumation over 106–107 yr. We interpret these and previously published data from a third valley using a 2‐D thermal model that accounts for topographic and advective influences on measured cooling ages. Exhumation rates in the Cordillera Real are ∼0.2–0.6 mm/yr, comparable to rates in parts of Denali, the Washington Cascades, the Olympic Mountains, and the European Alps and an order of magnitude slower than rates in Taiwan, Nanga Parbat, the Greater Himalaya of Nepal, and the Southern Alps of New Zealand. Three‐ to fourfold cooling age variations in the Cordillera Real imply at least twofold exhumation rate variations within and between valleys over distances of only tens of kilometers. Topography in the cross‐valley dimension affects exhumation rate estimates by 20%–30% in the downstream portions of two sample transects. Along‐valley topographic effects are less significant in this setting, affecting exhumation rate estimates by <15%. The most significant along‐valley topographic effects are associated with long‐wavelength mountain shape, including both retreat of the closure temperature isotherm near the mountain crest and compression of low‐temperature isotherms farther down the mountain flank. Locally varying phenomena (e.g., subregional structural history or transient patterns of local channel incision) must exert important controls on long‐term erosion patterns in order to produce observed short‐wavelength exhumation rate variations. Comparison of exhumation rate estimates with modern erosion rates suggests that long‐term and short‐term average erosion rates likely vary by less than twofold.

Owyhee River intracanyon lava flows: Does the river give a dam?

Rivers carved into uplifted plateaus are commonly disrupted by discrete events from the surrounding landscape, such as lava flows or large mass movements. These disruptions are independent of slope, basin area, or channel discharge, and can dominate aspects of valley morphology and channel behavior for many kilometers. We document and assess the effects of one type of disruptive event, lava dams, on river valley morphology and incision rates at a variety of time scales, using examples from the Owyhee River in southeastern Oregon. Six sets of basaltic lava flows entered and dammed the river canyon during two periods in the late Cenozoic ca. 2 Ma–780 ka and 250–70 ka. The dams are strongly asymmetric, with steep, blunt escarpments facing up valley and long, low slopes down valley. None of the dams shows evidence of catastrophic failure; all blocked the river and diverted water over or around the dam crest. The net effect of the dams was therefore to inhibit rather than promote incision. Once incision resumed, most of the intracanyon flows were incised relatively rapidly and therefore did not exert a lasting impact on the river valley profile over time scales >10 6 yr. The net long-term incision rate from the time of the oldest documented lava dam, the Bogus Rim lava dam (≤1.7 Ma), to present was 0.18 mm/yr, but incision rates through or around individual lava dams were up to an order of magnitude greater. At least three lava dams (Bogus Rim, Saddle Butte, and West Crater) show evidence that incision initiated only after the impounded lakes filled completely with sediment and there was gravel transport across the dams. The most recent lava dam, formed by the West Crater lava flow around 70 ka, persisted for at least 25 k.y. before incision began, and the dam was largely removed within another 35 k.y. The time scale over which the lava dams inhibit incision is therefore directly affected by both the volume of lava forming the dam and the time required for sediment to fill the blocked valley. Variations in this primary process of incision through the lava dams could be influenced by additional independent factors such as regional uplift, drainage integration, or climate that affect the relative base level, discharge, and sediment yield within the watershed. By redirecting the river, tributaries, and subsequent lava flows to different parts of the canyon, lava dams create a distinct valley morphology of flat, broad basalt shelves capping steep cliffs of Tertiary sediment. This stratigraphy is conducive to landsliding and extends the effects of intracanyon lava flows on channel geomorphology beyond the lifetime of the dams.

Controls on large landslide distribution and implications for the geomorphic evolution of the southern interior Columbia River basin

Large landslides (>0.1 km 2 ) are important agents of geomorphic change. While most common in rugged mountain ranges, large landslides can also be widespread in relatively low-relief (several 100 m) terrain, where their distribution has been relatively little studied. A fuller understanding of the role of large landslides in landscape evolution requires addressing this gap, since the distribution of large landslides may affect broad regions through interactions with channel processes, and since the dominant controls on landslide distribution might be expected to vary with tectonic setting. We documented >400 landslides between 0.1 and ∼40 km 2 across ∼140,000 km 2 of eastern Oregon, in the semiarid, southern interior Columbia River basin. The mapped landslides cluster in a NW-SE–trending band that is 50–100 km wide. Landslides predominantly occur where even modest local relief (∼100 m) exists near key contacts between weak sedimentary or volcaniclastic rock and coherent cap rock. Fault density exerts no control on landslide distribution, while ∼10% of mapped landslides cluster within 3–10 km of mapped fold axes. Landslide occurrence is curtailed to the NE by thick packages of coherent basalt and to the SW by limited local relief. Our results suggest that future mass movements will localize in areas stratigraphically preconditioned for landsliding by a geologic history of fluviolacustrine and volcaniclastic sedimentation and episodic capping by coherent lava flows. In such areas, episodic landsliding may persist for hundreds of thousands of years or more, producing valley wall slopes of ∼7°–13° and impacting local channels with an evolving array of mass movement styles.