Long-term monitoring of sandbars on the Colorado River in Grand Canyon using remote sensing

Abstract

Closure of Glen Canyon Dam in 1963 dramatically changed discharge and sediment supply to the downstream Colorado River in Marble and Grand Canyons. Magnitudes of seasonal flow variation have been suppressed, while daily fluctuations have increased because of hydropower generation. Lake Powell, the upstream reservoir, traps all sediment, leaving the Paria and Little Colorado Rivers as the main suppliers of fine sediment to the system below Glen Canyon Dam. The reduction in sediment supply, along with changes in discharge, have resulted in finesediment deficit (Topping et al., 2000), leading to a decrease in the size and number of alluvial sandbars (Schmidt and Graf, 1990; Schmidt et al., 2004). However, the understanding of these important spatial and temporal changes in sandbars located along the banks of the river have been limited to infrequent measurements mostly made by direct visitation and topographic surveying (Hazel et al., 2010). Aerial photographs are the only data available from which it is possible to evaluate changes in alluvial deposits at a large number of sites and compare recent conditions with those that existed prior to the initiation of ground-based monitoring in the early 1990s. Previous studies have evaluated the effects of Glen Canyon Dam on sandbars by analysis of comprehensive maps of surficial geology that are based on seven sets of aerial imagery taken between 1935 and 1996 for selected reaches in the first 120 km downstream from Lees Ferry, Arizona (Figure 1). These studies showed that the area of exposed sand in eddy-deposition zones was less in the post-dam period than in the pre-dam period (Leschin and Schmidt, 1995; Schmidt et al., 1999b; Sondossi, 2001, Sondossi and Schmidt, 2001, Schmidt et al., 2004). In this study, we extend these analyses to encompass a 74-year period by including maps of sand deposits visible in aerial imagery taken in 2002, 2005, and 2009 for the same reaches that were mapped in the earlier studies. Results are analyzed for two post-dam periods, based on the implementation of the first controlled flood in March 1996. The period from 1965 to March 1996 is the pre-controlled flood period and was dominated by flows that fluctuated up to the maximum capacity of the Glen Canyon Dam powerplant. Beginning in 1991, fluctuations were constrained such that maximum daily flows were typically less than 65 percent of powerplant capacity. Thus, the pre-controlled flood period also includes five years of restricted dam operations. This period also included unplanned spills from the reservoir in 1983, 1984, and 1986. We refer to the period from April 1996 to 2009 as the controlled-flood period. This period consisted entirely of restricted dam operations and included three controlled floods conducted as sandbar-building experiments. We show that the areal extent of exposed sand was greater in the images taken in the controlled-flood period than in the pre-controlled flood period. We also show that in the controlled-flood period, the area of exposed sand is negatively correlated with the elapsed time since the most recent controlled flood.