Frequent and prolonged military training maneuvers are an intensive type of land use that may disturb land cover, compact soils, and have lasting effects on adjacent stream hydrology and ecosystems. To better understand the potential effect of military training on hydrologic and environmental processes, the U.S. Geological Survey in cooperation with the U.S. Army established hydrologic and water-quality data-collection networks at the U.S. Army Garrison Fort Carson (AGFC) in 1978 and at the Piñon Canyon Maneuver Site (PCMS) in 1982. The purpose of this report is to present precipitation, streamflow, suspended-sediment, and water-quality data collected by the U.S. Geological Survey at the AGFC and PCMS for water years (WYs) 2016–18 and to evaluate those data in relation to long-term data from the AGFC and PCMS. In WYs 2016–18, the U.S. Geological Survey monitored 26 sites on the AGFC and 17 sites on the PCMS for precipitation amount, streamflow, suspended sediment, and (or) water quality.

On the AGFC, total annual precipitation in WYs 2016–18 was larger than the long-term mean for all 3 years at Rod and Gun Meteorologic Station at Fort Carson, CO (Rod and Gun). There were statistically significant upward trends in annual precipitation at Rod and Gun and Young Hollow Meteorologic Station at Fort Carson, CO (Young Hollow) with slopes of 1.25 and 0.66 inches per year (in/yr), respectively. The precipitation totals for WY 2017 were either the largest on record or in the top three for both sites and at Sullivan Park Meteorologic Station at Fort Carson, CO. On the PCMS, total annual precipitation was larger than the long-term mean in WYs 2016–18 at Brown Sheep Camp Meteorologic Station near Tyrone, CO; CIG Pipeline South Meteorologic Station near Simpson, CO; Bear Springs Hills Meteorologic Station near Houghton, CO (Bear Springs); and Upper Red Rock Canyon Meteorologic Station near Houghton, CO (Upper Red Rock). There were statistically significant upward trends in precipitation at Bear Springs and Upper Red Rock with slopes of 0.16 and 0.19 in/yr, respectively. The precipitation totals for WY 2017 were the largest on record for all sites except for Upper Bent Canyon Meteorological Station near Delhi, CO.

Streamflow was calculated at 18 sites on the AGFC and 7 sites on the PCMS in at least 1 of WYs 2016–18. At AGFC, mean annual (or seasonal) streamflow in WYs 2016–18 was less than the long-term mean at 7 sites and greater than the long-term mean at 3 sites. There were statistically significant downward trends in mean annual or seasonal streamflow at Womack Ditch from Little Fountain Creek near Fort Carson, CO, and Ripley Ditch from Little Fountain Creek at Fort Carson, CO, with slopes of −0.036 and −0.028 cubic feet per second per year (ft³/s/y), respectively; and a significant upward trend in streamflow at Turkey Creek West Seepage below Teller Reservoir near Stone City, CO, with a slope of less than 0.001 ft³/s/y. Unlike for precipitation, the mean annual or seasonal streamflow for WY 2017 was not in the top 3 for any of the 12 sites with measured flow.

At the PCMS, mean annual (or seasonal) streamflow was less than the long-term mean streamflow in WYs 2016–18 at the Taylor Arroyo below Rock Crossing near Thatcher, CO, and Bent Canyon Creek at Mouth near Timpas, CO, sites; and in WYs 2016 and 2018 at the Purgatoire River near Thatcher, CO (Purgatoire Thatcher), and Purgatoire River at Rock Crossing near Timpas, CO (Purgatoire Rock Crossing). There were no statistically significant trends in mean annual (or seasonal) streamflow at sites on the PCMS, and unlike for precipitation, the mean streamflow for WY 2017 was not in the top three for any sites except Purgatoire Rock Crossing. In WYs 2016–18, streamflow from sites on the AGFC and PCMS represented only a small fraction of streamflow in Fountain Creek or the Purgatoire River, and changes in streamflow that resulted from military maneuvers on the AGFC and PCMS were not likely to be detected in the downstream receiving waters.

Suspended-sediment concentrations, loads, and yields for WYs 2016–18, were analyzed at two sites on the AGFC and five sites on the PCMS. On the AGFC, mean seasonal suspended-sediment concentrations ranged from 3.10 to 155 milligrams per liter (mg/L), mean seasonal suspended-sediment loads ranged from 0.04 to 27.1 tons per day (t/d), and seasonal suspended-sediment yields ranged from 0.28 to 216 tons per season per square mile (t/s/mi²). Suspended-sediment yields at the two AGFC sites in WYs 2016–18 were all less than the long-term means. On the PCMS, mean seasonal suspended-sediment concentrations (at sites with some streamflow during a WY) ranged from 1.12 to 41.8 mg/L, mean suspended-sediment loads ranged from 0.01 to 13.1 t/d, and seasonal suspended-sediment yields ranged from 0.06 to 57.4 t/s/mi². Suspended-sediment yields at the five PCMS sites in WYs 2016–18 were all less than the long-term means. In WYs 2016–18, mean daily suspended-sediment loads at Little Fountain were 1.3, 2.5, and 7.6 percent, respectively, of the mean daily suspended-sediment load at Fountain Creek at Security, Colorado. Likewise, the total of mean daily suspended-sediment loads from the five tributary sites to the Purgatoire River in WYs 2016–18 were about 0.25, 0.17, and 3.2 percent, respectively, of the historical mean daily suspended-sediment load at Purgatoire Thatcher.

Spearman’s rank correlation coefficient was used to evaluate the strength and form of the relations between daily total precipitation and daily mean streamflow and between daily mean streamflow and suspended-sediment concentration and load for WYs 2016–18. For the sites on the AGFC and PCMS, there were weak or statistically insignificant positive correlations between precipitation and streamflow at nearby streamgauges, but strong statistically significant positive correlations between streamflow and suspended-sediment concentration and load. The ephemeral nature of the streams, quantity and timing of precipitation, air temperature, seasonal soil-moisture deficits, and effective runoff detention in erosion-control ponds could all contribute to inconsistent conversion of precipitation to streamflow.

Water-quality data were analyzed for as many as 43 parameters from 9 samples collected from 3 sites on the AGFC and from 37 samples collected from 4 sites on PCMS during WYs 2016–18. The concentrations of selected water-quality parameters were compared to regulatory standards for aquatic life from the Colorado Department of Public Health and Environment (CDPHE) or aquatic-life criteria from the U.S. Environmental Protection Agency (EPA). There is at least 1 CDPHE standard or EPA criterion for 30 of the 43 water-quality parameters.

For all samples from both the AGFC and the PCMS in WYs 2016–18, the concentrations of most water-quality parameters were compliant with the associated standards or criteria. However, there were some exceedances of standards or criteria: 11 samples exceeded the CDPHE recreational class standard for Escherichia coli concentration, 9 samples exceeded the CDPHE chronic unfiltered phosphorus aquatic-life standard, 36 samples exceeded the CDPHE chronic sulfate aquatic-life standard, 5 samples exceeded the EPA criterion for selenium, 7 samples exceeded the EPA criterion for aluminum, 2 samples exceeded the CDPHE chronic standard for iron, and 15 samples exceeded the CDPHE chronic standard for manganese.

Identifying potential effects of military training on water quality in adjacent streams on the AGFC and PCMS is difficult due to the ephemeral nature of streamflow, limited number of sampling locations and samples, and limited access to the study areas. At the PCMS, pairs of water-quality samples were collected in March and May 2017 before and after an April–May 2017 military training event. At the Purgatoire Rock Crossing site, streamflow at the time of the May sample was approximately 35 times larger than streamflow for the March sample. The absolute percent differences of concentrations for 27 parameters ranged from −71.7 to 183 percent, and 7 parameters had increases in concentration whereas 22 parameters had no change or decreases in concentrations. The absolute percent differences of loads for 24 parameters ranged from 141 to 198 percent. The generally lower concentrations and higher loads were expected given the higher streamflows at the time of collection of the May compared to the March samples.