Article pubs.acs.org/est
Effects of an Extreme Flood on Trace Elements in River WaterFrom Urban Stream to Major River Basin Larry B. Barber,*,† Suzanne S. Paschke,‡ William A. Battaglin,‡ Chris Douville,§ Kevin C. Fitzgerald,†,∥ Steffanie H. Keefe,† David A. Roth,† and Alan M. Vajda⊥ †
U.S. Geological Survey, 3215 Marine Street, Boulder, Colorado 80303, United States U.S. Geological Survey, Denver Federal Center, Denver, Colorado 80225, United States § City of Boulder, 4049 75th Street, Boulder, Colorado 80301, United States ∥ Carollo Engineers, Inc., 12592 W Explorer Drive, Boise, Idaho 83713, United States ⊥ University of Colorado Denver, CB171, Denver, Colorado 80217, United States ‡
S Supporting Information *
ABSTRACT: Major floods adversely affect water quality through surface runoff, groundwater discharge, and damage to municipal water infrastructure. Despite their importance, it can be difficult to assess the effects of floods on streamwater chemistry because of challenges collecting samples and the absence of baseline data. This study documents water quality during the September 2013 extreme flood in the South Platte River, Colorado, USA. Weekly time-series water samples were collected from 3 urban source waters (municipal tap water, streamwater, and wastewater treatment facility effluent) under normal-flow and flood conditions. In addition, water samples were collected during the flood at 5 locations along the South Platte River and from 7 tributaries along the Colorado Front Range. Samples were analyzed for 54 major and trace elements. Specific chemical tracers, representing different natural and anthropogenic sources and geochemical behaviors, were used to compare streamwater composition before and during the flood. The results differentiate hydrological processes that affected water quality: (1) in the upper watershed, runoff diluted most dissolved constituents, (2) in the urban corridor and lower watershed, runoff mobilized soluble constituents accumulated on the landscape and contributed to stream loading, and (3) floodinduced groundwater discharge mobilized soluble constituents stored in the vadose zone.
■
INTRODUCTION The effects of floods on the composition of surface water and groundwater are complex.1 Because of the unpredictable nature of floods (magnitude and frequency of recurrence), there usually is an absence of baseline chemical data from immediately before the event to provide insight into environmental processes controlling water quality.2 Occasionally, baseline data are established3 that allows comparison of preand post-event conditions, as was done with bed sediments following the 1993 Upper Mississippi River flood.4 The need for a better understanding of the relations between flooding, water quality, and urbanization is a topic of global interest,5 particularly effects on municipal infrastructure, such as drinking water treatment facilities (DWTFs) and wastewater treatment facilities (WWTFs).6,7 Extreme flooding occurred (Table S1) in the South Platte River Basin (Figure 1) in response to record amounts of precipitation from September 9 to 16, 2013 across the Colorado Front Range.8−11 This event provided an opportunity to assess the effect of flooding on streamwater quality. One of the headwater streams receiving significant rainfall was in the © XXXX American Chemical Society
Boulder Creek watershed, where 43 cm of precipitation fell from September 10 to 16 (annual average is 48 cm yr−1).12 Flooding of Boulder Creek began on September 11, and discharge at the U.S. Geological Survey (USGS) streamgage on Boulder Creek at North 75th Street near Boulder, CO [0673020; https://waterdata.usgs.gov/usa/nwis/uv?06730200] increased from 1.5 to about 240 m3 s−1.8 Similarly, Saint Vrain Creek had a sharp rise in discharge to a peak of about 900 m3 s−1, 3 times the 1941 record (about 300 m3 s−1).8 Other streams along the Front Range had similar rainfalls and contributed to the South Platte River flooding from Denver to the Colorado−Nebraska state line, and beyond. Downstream from the confluence with the Cache la Poudre River, flooding in the South Platte River resulted from progression of the upstream flood crest. The maximum flood discharge of about 1700 m3 s−1 occurred on September 15 at Fort Morgan, and Received: April 19, 2017 Revised: July 7, 2017 Accepted: August 4, 2017
A
DOI: 10.1021/acs.est.7b01767 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
Figure 1. Map showing locations of the Boulder urban-water study and the South Platte River main stem and tributary sampling. [See Table S2 for detailed information on South Platte River sampling locations; TP, Boulder tap water; BC, Boulder Creek water; WWTF, wastewater treatment facility; EF, wastewater treatment facility effluent.]
the flood crest exited Colorado on September 18 after reaching a peak of 620 m3 s−1 at Julesburg.8 The 2013 flood was an extreme event for the Rocky Mountain region because of its expansive extent and long duration of high flows. This region typically experiences localized, short-duration flash floods. As the result of the 2013 flooding, municipal infrastructure (including DWTFs and WWTFs) was damaged at multiple locations along the Colorado Front Range.9 The geochemistry of natural waters has been the topic of research for many decades and has focused on major elements and a few heavy metals (lead, copper, cadmium, and zinc).13 More recently, attention has focused on the influence of anthropogenic activities on the geochemistry of urban environments.14,15 Urban geochemistry is distinctive from natural geochemistry, because the built infrastructure, industrial and commercial activities, point-source discharges, lawn maintenance, atmospheric emissions, and transportation networks contribute concentrated loading of many elements to the land surface. Anthropogenic activities associated with urbanization have a long history of pollution by elements, such as lead.16 In the past, most studies of urban pollution have focused on heavy metals in soils and sediments17 and water-soluble elements, such a boron.18 More recently, new classes of synthetic organometallic compounds and nanoparticles have resulted in global pollution of the aquatic environment that threatens to disrupt the natural elemental distributions in rivers.19 This study uses the distinct source characteristics and geochemical behaviors of a suite of trace elements to address (1) limited knowledge of sources and mobilization of trace elements in the urban environment by floods, (2) difficulty in capturing information during unpredictable flood events, and (3) evaluation of dynamic processes that occur during the course of floods. Chemical tracers were used to characterize
processes controlling water composition during an extreme flood at different scales (urban stream to major river) and provide a framework for assessing contributions of natural and anthropogenic contaminants.
■
METHODS Study Locations. The urban-stream component of this study focused on Boulder Creek (Figure 1), which originates at the Rocky Mountain Continental Divide, flows through the City of Boulder, and joins Saint Vrain Creek downstream from Longmont. Saint Vrain Creek joins the South Platte River downstream from Fort Lupton. Boulder Creek has been the subject of long-term research on contaminants, major water sources, changes in water quality over time, and dominant landscape controls.20−25 The South Platte River drains central and eastern Colorado, flows 720 km from its headwaters to its confluence with the North Platte River in Nebraska, and spans a range of climates, geologic characteristics, and land uses.26 The mountainous headwaters of the South Platte River basin are underlain by Precambrian igneous and metamorphic rocks, and the eastern plains are underlain by Cretaceous and Tertiary sedimentary rocks (shale, siltstone, and sandstone) and Quaternary deposits.27,28 Sampling. The effect of the 2013 flood on water quality was captured as part of an ongoing series of on-site experiments on infrastructure upgrades conducted at the Boulder WWTF.29,30 Time-series water samples from 3 components of the urbanwater system were collected at the Boulder WWTF (Figure 1): municipal tap water (TP), Boulder Creek water (BC) from approximately 0.5 km upstream from the WWTF outfall, and WWTF effluent (EF). Weekly samples were collected under normal-flow conditions in 2012 (September 20 and 25 and October 3, 9, and 16). In 2013, samples were collected under B
DOI: 10.1021/acs.est.7b01767 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
Figure 2. Hydrographs for the U.S. Geological Survey, Boulder Creek at North 75th Street near Boulder, CO streamgage [06730200; http:// waterdata.usgs.gov/usa/nwis/uv?06730200] during (A) 2012 and (B) 2013. City of Boulder Wastewater Treatment Facility (WWTF) sewage inflow during (C) 2012 and (D) 2013. [Insets show flow conditions during the sampling events (shown by gray bars) and specific sampling dates (shown by red stars).]
trations of the various constituents were evaluated against quality-assurance samples including field blanks (n = 4) and duplicate samples (n = 5). Each sample was analyzed in triplicate. Ammonia, nitrate, total suspended solids, and Escherichia coli were measured by standard methods at the Boulder WWTF laboratory. Statistical analyses were performed in R 3.2.1 using the Mann−Whitney U-test.34 “Dissolved” concentrations in the 3 Boulder source waters were evaluated for significant differences by comparing 2012 and 2013 normal-flow samples and 2013 normal-flow and flood samples. Results with p < 0.05 are reported as significant.
normal-flow conditions (August 13, 20, and 27 and September 3 and 10) and flood conditions (September 17, 24, and 30). The source water for TP is withdrawn from the upper Boulder Creek watershed (underlain by igneous and metamorphic rocks) and treated for municipal water supply using coagulation, filtration, and disinfection, with a typical residence time from intake to tap of about 12 h [https:// bouldercolorado.gov/water]. Because much of the DWTF infrastructure is in the upper watershed, the 2013 flood had limited effects on operations. Water from the BC site also originates in the upper watershed, and flows through the City of Boulder and across sedimentary rocks for about 10 km. The EF water originates as TP, is used in domestic and commercial activities, and is disposed of down the drain. The wastewater is conveyed by sewer lines to the WWTF where it undergoes activated sludge secondary treatment before being discharged back into Boulder Creek. In addition, a basin-wide sampling of the South Platte River during the flood was conducted from September 18 to 22 at 5 main stem and 7 tributary sites (Figure 1 and Table S2), at times ranging from 4 to 7 d after peak flows. Water-quality sampling followed USGS field procedures.31 Grab samples were collected directly from the tap (TP) or from continuously operating stainless steel pumps equipped with Teflon tubing (BC and EF). South Platte River water samples were collected using flow-proportional techniques where possible; otherwise grab samples were collected. Unfiltered and filtered water samples were collected in polyethylene bottles, and acidified to pH < 2.0 with ultrahigh purity nitric acid. Filtered samples were passed through 0.45-μm nucleopore membranes to determine the operationally defined “dissolved” phase, which includes colloids
