Conditions and processes affecting sand resources at archeological sites in the Colorado River corridor below Glen Canyon Dam, Arizona

Abstract

This study examined links among fluvial, aeolian, and hillslope geomorphic processes that affect archeological sites and surrounding landscapes in the Colorado River corridor downstream from Glen Canyon Dam, Arizona. We assessed the potential for Colorado River sediment to enhance the preservation of river-corridor archeological resources through aeolian sand deposition or mitigation of gully erosion. By identifying locally prevailing wind directions, locations of modern sandbars, and likely aeolian-transport barriers, we determined that relatively few archeological sites are now ideally situated to receive aeolian sand supply from sandbars deposited by recent controlled floods. Whereas three-fourths of the 358 river-corridor archeological sites we examined include Colorado River sediment as an integral component of their geomorphic context, only 32 sites currently appear to have a high degree of connectivity (coupled interactions) between modern fluvial sandbars and sand-dominated landscapes downwind. This represents a substantial decrease from past decades, as determined by aerial-photograph analysis. Thus, we infer that recent controlled floods have had a limited, and declining, influence on archeological-site preservation. Within the study area, overland-flow (gully) erosion is less severe in sand landscapes with active aeolian sand than in landscapes that lack aeolian transport; gullies terminate more commonly in active sand (sand that is mobile by wind rather than stabilized by biologic soil crust). We infer that these characteristics largely result from aeolian sand transport being an effective gully-limiting and gully-annealing mechanism. Aeolian sand activity in the river corridor varies substantially as a function of reach morphology and dominant wind direction relative to the river-corridor orientation, factors that control accommodation space for river-derived sand and the modern sand supply to aeolian dunes. These attributes, together with an inverse correlation between aeolian sand activity and gully occurrence, define varying degrees of net long-term gully-erosion risk for sediment deposits and associated archeological sites in different regions of the river corridor. Over most of the river corridor, including some of the archeologically richest regions, sand is too inactive with respect to aeolian transport to anneal gullies effectively. At eight selected archeological sites that we studied with high-resolution terrestrial lidar scans for more than a year, sand loss by overland flow (gully erosion) and aeolian deflation generally exceeded deposition, such that erosion dominated over most monitoring intervals—even at four sites with strong connectivity to modern sand supply. The Glen Canyon reach of the river corridor appears especially vulnerable to gully erosion. Among the sites that we monitored in detail, erosion generally dominated over deposition to a greater degree at four Glen Canyon sites with no modern sand supply than at four Marble–Grand Canyon sites with aeolian sand supply from controlled-flood sandbars. Although gross annual-scale erosion rates were similar among the Glen Canyon sites and among the Marble–Grand Canyon sites, a relative lack of depositional processes led to greater net erosion at the Glen Canyon sites. Having found no differences in weather patterns to suggest greater erosive forcing in Glen Canyon, and no conclusively influential differences in the slope or watershed area contributing to gully formation, we attribute the greater erosion at the Glen Canyon sites to a combination of inherent geomorphic context (high terraces that do not receive modern sediment supply) and pronounced effects of postdam sediment-supply limitation. We conclude that most of the river-corridor archeological sites are at elevated risk of net erosion under present dam operations. In the present flow regime, controlled floods do not simulate the magnitude or frequency of natural floods, and are not large enough to deposit sand at elevations that were flooded at annual to decadal intervals in predam time. For archeological sites that depend upon river-derived sand, we infer elevated erosion risk owing to a combination of reduced sand supply (both fluvial and aeolian) through (1) the lower-than-natural flood magnitude, frequency, and sediment supply of the controlled-flooding protocol; (2) reduction of open, dry sand area available for wind redistribution under current normal (nonflood) dam operations, which do not include flows as low as natural seasonal low flows and do include substantial daily flow fluctuations; and (3) impeded aeolian sand entrainment and transport owing to increased riparian vegetation growth in the absence of larger, more-frequent floods. If dam operations were to increase the supply of sand available for windblown transport—for example, through larger floods, sediment augmentation, or increased fluvial sandbar exposure by low flows—and also decrease riparian vegetation, the prevalence of active aeolian sand could increase over time, and the propensity for unmitigated gully erosion could decrease. Although the evolution of river-corridor landscapes and archeological sites has been altered fundamentally by the lack of large, sediment-rich floods (flows on the order of 5,000 m 3 /s), some combination of sediment-rich flows above 1,270 m 3 /s, seasonal flows below 226 m 3 /s, and riparian-vegetation removal might increase the preservation potential for sand-dependent archeological resources in the Colorado River corridor.