Spatial and temporal patterns in streamflow, water chemistry, and aquatic macroinvertebrates of selected streams in Fairfax County, Virginia, 2007–18

Abstract

Urbanization substantially alters the landscape in ways that can impact stream hydrology, water chemistry, and the health of aquatic communities. Stormwater best management practices (BMPs) are the primary tools used to mitigate the effects of urban stressors such as increased runoff, decreased baseflow, and increased nutrient and sediment transport. To date, Fairfax County Virginia’s stormwater management program has made substantial investments into the implementation of both structural and nonstructural BMPs aimed at restoring and protecting watersheds. The U.S. Geological Survey (USGS), in cooperation with Fairfax County, Virginia, established a long-term water-resources monitoring program to evaluate the watershed-scale effects of these investments. Monitoring began at 14 stations in 2007 and was expanded to 20 stations in 2013. This report utilized the first 10 years of data collection to (1) assess water quantity and quality, as well as ecological condition; (2) compute annual nutrient and sediment loads; and (3) evaluate trends in streamflow, water quality, and ecological condition. Efforts are underway to link the biotic and abiotic patterns described herein to watershed management practices as well as factors such as land use change, public works infrastructure, and climate. Hydrologic, chemical, and benthic macroinvertebrate community conditions in the streams monitored were similar to those observed in other studies of urban streams. Multidecadal trends in baseflow indices and runoff ratios at long-term Chesapeake Bay Non-tidal Network streamgages (CB-NTN) indicate a decrease in groundwater recharge and increase in storm runoff as a result of urbanization. Streamflow yields varied spatially with land cover, geology, and soil characteristics, whereas flashiness was positively related to impervious area. Dissolved oxygen typically was lowest in the Coastal Plain and across all Triassic Lowlands streams, and highest in the Piedmont. Dissolved oxygen concentrations generally were above Virginia’s minimum criterion of 4.0 milligrams per liter (mg/L), most violations occurred at Paul Spring Branch in the Coastal Plain during the warmest months of the year owing to increased chemical and biological oxygen demand. Typical pH values of the monitored streams centered on neutrality (pH = 7); however, diurnal fluctuations were most prevalent in the continuous pH data at Flatlick Branch (FLAT; a Triassic Lowlands station), as a result of increased photosynthesis catalyzed by phosphorus-rich geology. Specific conductance (SC) varied spatially owing to geology (highest at Triassic Lowlands stations) and anthropogenic disturbance (watersheds with high impervious land cover). Specific conductance typically was inversely related to streamflow except in winter months following deicing road salt applications, when values increased by several orders of magnitude. A significant increase in SC of about 2 percent per year was observed from the combined trend result of all monitoring stations over the 10-year period. Significant SC increases occurred at nearly all monitoring stations. Increasing trends were observed during winter and nonwinter months, which suggests that salts applied to deice roadways and other impervious surfaces are stored in the environment and released year-round. Suspended-sediment (SS) concentrations in monthly samples did not vary significantly between most stations, but typically were highest in the spring and lowest in the fall as a result of seasonal differences in streamflow and climate. Suspended-sediment yields ranged from 62 to 1,428 tons per square mile (ton/mi 2 ), with a median of 302 ton/mi 2 . Annual loads were greatest during the wettest water years (October 1-September 30; 2008, 2011, and 2014), with the greatest interannual variability occurring at Difficult Run above Fox Lake (DIFF) and South Fork Little Difficult Run (SFLIL). Suspended sediment was primarily composed of silts and clays; however, the proportion of sand in suspended sediment was related positively to streamflow. Cross-correlation analyses suggested the dominant sources of SS were streambank erosion and resuspension of in-channel material at DIFF and FLAT; whereas, upland sources and erosion of upper streambanks were more common at Dead Run (DEAD), Long Branch (LONG), and SFLIL. Median total phosphorus (TP) concentrations ranged from 0.016 to 0.077 mg/L, with a networkwide median of 0.022 mg/L, were highest in the warm season (April-September), and were composed primarily of dissolved phosphorous. Although TP concentrations were relatively low across the network, the highest concentrations were consistently at stations located in the Triassic Lowlands, owing to phosphorous-rich geology, and in the Coastal Plain, owing to the low-phosphorous sorptive capacity of those soils. A significant increase in TP concentration occurred in a few stations, but the combined trend results from all stations demonstrated a significant increase of about 4 percent per year. Networkwide increases were also observed in total dissolved phosphorus, orthophosphate, and total particulate phosphorus. The composition of TP shifted from dissolved to particulate as streamflow increased and for this reason loads primarily were composed of particulate phosphorous. Median annual TP loads were highest at FLAT and DEAD and ranged from 247 to 642 pounds per square mile (lbs/mi 2 ) networkwide. Interannual variability in phosphorous yields was apparent at most stations; the highest loading years were also the wettest years during the study period and coincident with the highest peak annual flows. Total nitrogen (TN) concentrations typically were low throughout the network with exceptions occurring at stations located in watersheds with a high density of septic infrastructure. Elevated TN concentrations also were observed in some watersheds without a high density of septic systems and may be attributable to geologic and soil properties that limit denitrification as well as other unknown anthropogenic inputs. Total nitrogen typically was dominated by nitrate during baseflows; however, the proportion of particulate nitrogen increased during stormflows. Total nitrogen yields were similar across stations, with medians ranging from about 3,600 to 6,300 lbs/mi 2 and were related to annual streamflow volume. Total nitrogen concentrations and flow-normalized concentrations decreased over the 10-year period at 7 stations, with median reductions of about 2.5 percent. Increasing trends were observed at the two stations with the highest median TN concentration (Captain Hickory Run and SFLIL, 3–5 mg/L), both watersheds contain a high density of septic infrastructure. The combined trend results from all stations revealed no trend in TN and a declining trend in nitrate of about 2 percent per year. Overall, benthic community metrics indicated that streams throughout Fairfax County were initially of poor health; however, many metrics show an improving trend (from poor to fair based on the Fairfax County Index of Biological Integrity [IBI]). Significant increasing trends in IBI occurred at the network-scale and at 4 individual stations; additionally, scores improved by at least 1 qualitative category (for example, poor to fair, fair to good) at 11 of the 14 stations between 2009 (the first year all 14 stations were sampled) and 2017. Changes in all metrics suggest that the biodiversity, function, and condition of streams in Fairfax County are improving, but some of these improvements are driven by increased diversity and percent composition of organisms that are tolerant of the urban environment.