Sample collection and analysis
In May 2016, we identified 14 sampling sites along the South Platte River, spanning the length of the Denver metropolitan area (Figure 1). The first point was located at the mouth of Waterton Canyon, where the South Platte River emerges from the foothills. From preliminary work, we knew that water quality at this location was similar to points higher in the watershed and for this reason, we used it as our upstream control point (Heilman & Schliemann, 2015). From Waterton Canyon, water flows into Chatfield Reservoir. We placed our second sampling point immediately downstream of the reservoir, 14 km from the first point. The remaining points were spaced approximately 4.8 km (3 mi) apart with negligible variation due to sampling accessibility. Additional sampling sites were included above and below two of the main tributaries to the South Platte River: Cherry Creek and Clear Creek (Figure 1).
From May 2016 – May 2019, every two weeks (March – November) and every four weeks (December – February), we assessed a range of water quality parameters at every site. We measured water temperature, pH, dissolved oxygen (DO) in situ; and collected grab samples to measure the concentrations of nitrate-N (NO3-N), ammonia-N (NH3-N), and orthophosphate (PO43-) in the lab within 3 hours of collection. In addition, we performed a BOD5 test using a DO probe and standard methods (Eaton, Clesceri, Rice, Greenberg, & Franson, 2005). Temperature, pH, DO, BOD, and the concentration of NO3-N were measured using Hach probes and a Hach HQ40d meter (Hach CO, Longmont, Colorado, USA). The concentrations of NH3-N and orthophosphate were measured colorimetrically using a Hach DR900 colorimeter (Hach CO, 2020).
From November 2017 – February 2018, we systematically inventoried all visible storm drains, effluent points, and natural tributaries that join the South Platte River from the 14 km site to the 63 km site. We did not sample between the 0 km and 14 km sites due to access limitations. Furthermore, this area is a state park, with a little urban development. Daily, we walked segments of the river, marking each point using a handheld Garmin etrex 20x GPS unit (Garmin Ltd., Olathe, Kansas, USA) . We also recorded whether the outfall point had flowing water, the type of outfall (storm drain, effluent point, or natural), the outfall structure (i.e., pipe or culvert), and the relative size of the outfall.
After collecting 18 months of data, we noticed that two of the areas we were sampling (28 km – 37 km and 47 – 48 km) seemed to have consistently high concentrations of nutrients. Consequently, in November 2017, we initiated an intensive study of these two “hotspot” areas. Over six months (November 2017 – April 2018), we sampled each flowing outfall from the beginning of the hotspot (i.e. 28 km and 47 km, respectively) to 5 km upstream (i.e. 23 km and 42 km, respectively). In addition, we sampled the river in this same section intensively, taking samples every 100 m. At each sampling point, we measured the parameters discussed above.