Site Descriptions

Seven tributaries of the Tualatin River with beaver activity were selected for water-quality monitoring (fig. 1). The stream reaches along Fanno and Bronson Creeks were intensively studied with multiparameter water-quality data collection. These two urban stream reaches were the focus of recent habitat restoration projects but differ in their upstream drainage areas, percent impervious area, and number and types of beaver dams (table 1; app. 1). Most of the results from this study were derived from data gathered from these intensively studied reaches along Fanno and Bronson Creeks. Additional reaches (referred to as “auxiliary reaches”), were monitored for water temperature to determine if patterns measured along the intensive reaches were consistent with patterns measured in other reaches in the basin. Auxiliary reaches differ in site characteristics and provide a broader understanding of the effects of beaver activity on water temperature in urban streams.

Fanno Creek at Greenway Park Intensive Study Reach

Fanno Creek originates in the northeastern hills of the basin on the west side of the City of Portland and is a major tributary to the Tualatin River (fig. 1). The upstream contributing area to the study reach is approximately 27 km² of mostly low-gradient, urban lands (table 1). The study reach is in Greenway Park, a popular urban park. Greenway Park was the focus of restoration efforts from 2008 to 2012 by Clean Water Services. During that restoration, the channel, once straightened for agricultural purposes, was re-meandered, large wood was placed in the channel to provide habitat complexity, and native vegetation was planted throughout the park (Smith and Ory, 2005). Keith and others (2014) noted that the Greenway Park segment of Fanno Creek was one of the largest potential sinks for fine sediment along Fanno Creek, due to its low and wide floodplain and vegetation that decreases stream velocity by increasing surface roughness within the channel. Beavers colonized Fanno Creek in the park in 2012 and built multiple dams and a lodge within the study reach. One dam extending onto the floodplain was 30 m long and created a large pond (referred to as the “south pond”; fig. 2) that flooded the newly meandered channel, old channel, nearby paths, and riparian areas. Other floodplain features in the study reach included the north pond and adjacent Koll Center Wetland, which were both off-channel features that were present prior to the construction of beaver dams.

The 940-m study reach in Greenway Park decreases in elevation by approximately 1.9 m from the upstream to downstream monitoring sites, and the reach varies in channel characteristics and riparian shade along its length. The stream between Southwest Hall Boulevard and the south pond is channelized, as deep as 2.5 m, and often is shaded by riparian vegetation (fig. 2). Within the south pond, the water spreads onto the floodplain because the channel is blocked by the 30-m-long beaver dam. In this section of the study reach, the main channel is narrow and deep but is surrounded by a ponded area with a surface area of approximately 7,940 m². This ponded area is generally shallow (0.5 m deep on average) with little riparian shade. Downstream from the south pond and long beaver dam, the channel is narrow, deep, and shaded by riparian vegetation. The old channel is shaded, narrow, and wadable except for deep areas upstream from the two beaver dams. Water velocity throughout the reach is slow during low flows (<0.01 m/s; White and others, 2025b).

Continuous multiparameter water-quality monitors were deployed in the following three locations: (1) Upstream at Fanno Creek (UF), (2) Ponded at Fanno Creek (PF), and (3) Downstream at Fanno Creek (DF; fig. 2). Similar conditions were measured at the UF and PF sites during this study, likely because channel characteristics and riparian shade were similar at the two sites and because the water remained within the channel in this section of the stream. The DF continuous monitor was downstream from the long dam, and measurements reflected processes occurring in the large south pond.

Bronson Creek Intensive Study Reach

Bronson Creek begins in the Tualatin Mountains, an area characterized by steep terrain, and flows in a southwesterly direction (fig. 1; table 1). The upstream contributing area to the Bronson Creek study reach is approximately 8 km² of land that is rapidly transitioning from rural to urban land use. The 1,610-m-long reach has a wide, flat-valley floodplain that is surrounded by residential properties. Clean Water Services added large wood and planted riparian vegetation at the site from 2004 to 2019 in an effort to improve water quality and enhance habitat diversity of the floodplain and stream (Smith and Ory, 2005). The upstream part of the study reach is near the foothills in a forested area where the stream is confined to a narrow channel with various depths and ponded by multiple beaver dams (fig. 3). Downstream from the forested area, Bronson Creek splits into multiple side channels and has three large dams; the side channels flow through dense stands of reed canary grass (Phalaris arundinacea). Throughout the rest of the reach, Bronson Creek continues to flow diffusely in several channels through reed canary grass and forested areas and has an unknown number of beaver dams. Water velocity throughout the reach is slow during low flows (0.01 m/s; White and others, 2025b), and the banks of the stream often are full even in summer. The effects of subsurface water inputs and exchange on water quality are indicated at Bronson Creek; in this report, “subsurface water” refers to groundwater and (or) hyporheic water.

Continuous multiparameter water-quality monitors were deployed at the following three locations: (1) Upstream at Bronson Creek (UB), (2) Ponded at Bronson Creek (PB), and (3) Downstream at Bronson Creek (DB) (fig. 3). The UB site was free flowing for the first 3 months of the study, after which a beaver dam was built 20 m downstream from the water-quality monitor resulting in ponding at the site. During low flows, the PB site measured channelized, ponded water with minimal riparian shade. The DB site was approximately 980 m downstream from the PB site and measured channelized, ponded water that was partly shaded. The beaver-affected area continued downstream from the study reach, with additional beaver dams and ponding.

Water Temperature

Comparison of water temperatures from the three multiparameter monitors and one water-temperature sensor deployed in Fanno Creek (fig. 4) shows longitudinal patterns in water temperature along the study reach. Continuous water-temperature analyses used the 7-day moving average of the daily maximum (7dADM) in accordance with the criterion used in the State of Oregon water temperature standard (Oregon Department of Environmental Quality, 2019a). During summer 2016 (May 18–September 30), 7dADM measurements at the DF site were significantly (Tukey post-hoc p values <0.01) warmer than at the PF and UF sites (fig. 4), suggesting that the water in the unshaded and shallow areas of the south pond likely was warmed by solar radiation. The 7dADM water temperature at the DF site was greater than 18 °C—the criterion for the protection of fish rearing and migration uses in summer—during most of summers 2016 and 2017. However, hourly water-temperature measurements indicate that water exiting the intensively monitored beaver-affected reach (measured at DF) was sometimes cooler than 18 °C (from 7 to 11 a.m.; fig. 5). The 7dADM values were not significantly different among the three sites during the winter months. The pattern likely is driven by lower air temperatures, less sunlight, and increased cloud cover during the shorter winter days, along with increased streamflow and decreased residence time; however, separation of the DF site from UF and PF sites began to appear as early as March 2017 as warming occurred during spring (fig. 4)

Downstream from the wide, unshaded, and warm beaver pond in Fanno Creek, the stream returned to a narrow, deep, and shaded channel. A water-temperature sensor placed 600 m downstream from the DF monitor (referred to as “Farther downstream at Fanno Creek” [FDF] [USGS site 452704122474200]; in free-flowing water not affected by beaver activity) indicated cooler water temperatures than those measured at the DF site (fig. 4). The cooling effect downstream from the large beaver pond likely was due to a shift in the heat budget of the stream, with a decreased total flux of incoming solar radiation in the narrower and more-shaded channel, thus allowing the stream to release a part of the heat it absorbed while in the shallow and unshaded beaver pond upstream.

The five continuous water-temperature sensors deployed in and around the large, south pond in Fanno Creek at Greenway Park showed diverse diel patterns. During summer, minimum daily water temperatures at the five sites were similar, but the daily maxima varied in their magnitude and timing (fig. 6). On July 26, 2017, at 1700 hours, the shallow, unshaded site (Site 2) measured 30.6 °C, whereas the deep, unshaded site (Site 3) measured 22.0 °C. Water temperatures at deep locations reached the daily maximum at around 2300 hours, whereas water temperatures at shallow locations reached the daily maximum at around 1700 hours each day. The 7dADM water temperatures for the five ponded sites were >20 °C during July and August 2017. Comparing water temperatures measured in a ponded reach to relevant water-quality standards can require identifying a representative sampling location, among other considerations; however, the water-temperature ranges indicate that on warm summer days, all sites within the ponded area may have thermally stressed sensitive aquatic life that prefer cool water.

Dissolved Oxygen

Dissolved-oxygen concentrations in streams are affected by (1) the oxygen concentrations of incoming waters (such as creek water, runoff, and subsurface water), (2) inputs from photosynthesis, (3) losses to respiration and organic-matter decomposition, and (4) exchange with the atmosphere across the air/water interface (reaeration). In slow-moving or ponded reaches, reaeration tends to be slow. Photosynthesis requires sunlight as well as sufficient nutrients and favorable temperatures, and the plant species that photosynthesize—phytoplankton, macrophytes, or periphyton (which can have differential effects on primary production)— are themselves determined by other habitat variables such as flow velocity and substrate. Decomposition of organic matter occurs in suspension as well as at the sediment/water interface, and the rate increases with increasing temperature. Oxygen can exceed its water solubility when photosynthetic production is high, and concentrations can decrease to zero if consumption through respiration or decomposition is large; these effects are exaggerated when gas exchange with the atmosphere is slow, such as when water velocities or wind are low.

All three continuous-monitoring sites along Fanno Creek showed low dissolved-oxygen concentrations for parts of summers 2016 and 2017, less than the three-tiered cool-water dissolved-oxygen standard (Oregon Administrative Rules, 340-041-0016(3); Oregon Department of Environmental Quality, 2019b; table 4). Concentrations less than the standard are relevant because they may be indicative of stressful conditions for sensitive aquatic organisms. The water-quality standard sets the minimum instantaneous dissolved-oxygen concentration for cool-water aquatic life at 4 mg/L. Instantaneous dissolved-oxygen values measured at the UF site (and representing concentrations entering the study area) reached concentrations less than 3 mg/L on multiple days in August 2017 (fig. 7) but also reached as high as 10 mg/L that month. Measurements at the PF site showed a prolonged period of dissolved-oxygen concentrations less than 2 mg/L during August 2017, which may have been a result of relatively stagnant water (low streamflow), thermal stratification, and a proportionately high oxygen consumption from the decomposition of organic matter in surficial sediment (sediment oxygen demand; Rounds and Doyle, 1997; fig. 7). Dissolved-oxygen concentrations at the DF site varied substantially in the summer months, sometimes fluctuating more than 4 mg/L in a day. Water exiting the ponded area (measured at the DF site) was either oxygenated (50–70 percent oxygen saturation) or hypoxic (oxygen percentages and concentrations near zero) because of the primary productivity, respiration, and sediment oxygen demand occurring in the shallow waters of the south pond just upstream.

The low dissolved-oxygen concentrations measured in the Fanno Creek reach were not unexpected because long-term monitoring stations throughout the Tualatin River Basin have established that dissolved-oxygen concentrations in valley-bottom streams can reach low concentrations during summer even in the absence of beaver dams and beaver activity (Sobieszczyk and others, 2014). In the hot, dry summer months (June–September) from 2009 through 2019, dissolved-oxygen concentrations measured in Fanno Creek at Durham (USGS site 14206950) typically ranged from 6 to 8 mg/L, with a few deviations <4 mg/L (U.S. Geological Survey, 2019b). Concentrations for the same dates in a nearby stream (Beaverton Creek; USGS site 453004122510301) mostly ranged from 3 to 6 mg/L and dissolved-oxygen concentrations <2 mg/L were measured for multiple weeks in some years (U.S. Geological Survey, 2019b). Although site characteristics differ between Beaverton and Fanno Creeks, the low gradient and large accumulation of decomposing organic matter in the sediments of valley-bottom streams in the Tualatin River Basin can cause low dissolved-oxygen concentrations in the absence of beaver dams.

pH

An important water-quality indicator is pH because it affects biological processes, organisms, and the solubility of many chemical constituents. A pH of 7.0 is neutral, whereas typical rainwater in northwestern Oregon is slightly acidic (pH=5.65) because atmospheric carbon dioxide reacts with rainwater to form carbonic acid. Photosynthesis causes pH to increase because that process removes dissolved carbon dioxide (and carbonic acid) from the water; respiration reverses the process and causes pH to decrease.

Values of pH in Fanno Creek generally remained within the 6.5 (minimum) to 8.5 (maximum) standard unit range specified by the State of Oregon water-quality standard (Oregon Department of Environmental Quality, 2019c; fig. 8). Values less than 6.5 standard units measured at the UF site corresponded with large streamflow events, indicating a larger proportion of water from slightly acidic rainfall during those storms. During the synoptic survey on May 23, 2017, an extensive algal bloom was observed and photographed in the south pond. Large amounts of photosynthesizing organisms in shallow water could increase the pH because of the uptake of dissolved carbon dioxide (and subsequent decrease in carbonic acid). High pH conditions are not uncommon during large algal blooms and typically are temporary. Values of pH were not recorded during the synoptic survey in the south pond on that date; however, elevated pH values were measured at the DF monitor, indicating that the affected water was transported downstream. During the synoptic survey on September 12, 2017, instantaneous pH values at multiple locations within the south pond exceeded the 8.5 maximum criterion (Poor, 2020), and another algal bloom was observed at that time. However, elevated pH values were not measured at the DF monitor.

Specific Conductance

The specific conductance of a waterbody is related to the concentration of charged solutes that it contains (such as bicarbonate [HCO₃^-], sodium ion [Na⁺], chloride ion [Cl^-], potassium ion [K⁺], magnesium ion [Mg²⁺], etc.). Specific conductance often increases in a downstream manner, and patterns in the data may indicate the presence and locations of point sources or diffuse inputs such as subsurface water. For example, rainwater contains few solutes, and rainwater inputs result in localized decreases in specific conductance. Conversely, subsurface, agricultural, and municipal water inputs can contain high concentrations of solutes, increasing the specific conductance of the stream.

Specific conductance values measured in Fanno Creek at Greenway Park were high in summers 2016 and 2017, relative to historical measurements recorded 8.5 km downstream at Fanno Creek at Durham (USGS site 14206950). Values reached >1,000 µS/cm in summer 2016 in the study reach, whereas values reached a maximum of only 400 µS/cm that summer at Fanno Creek at Durham (fig. 9); a tributary entering Fanno Creek between Greenway Park and Durham diluted the specific conductance. A source of elevated conductance water upstream from Greenway Park likely caused the localized elevated specific conductance values in the study reach, which created a clear signal to trace from UF to DF.

Five discrete time periods in summer (low flows) and winter (high flows) were chosen to analyze the travel time of water between each continuous water-quality monitor in the Greenway Park study reach. For each period, the date and time were recorded when the measured specific conductance at each monitoring location (UF, PF, and DF) reached a designated threshold in the data (for example, ≥700 µS/cm in summer [fig. 9] and ≤140 µS/cm in winter). The time differences showed the travel time required for the pulse of upstream water with high specific conductance (summer) or the diluted rainwater (winter) to move through the study reach.

The average travel time from the UF to the DF sites (a distance of 940 m) during summer (low flow; n=5) was approximately 146 hours, or about 6 days (fig. 10). Travel time was longer through the large south pond (from PF to DF sites) compared to the channelized part of the reach from UF to PF sites, likely because of decreased water velocity in the beaver pond. In winter (high flow; n=5), the average travel time from UF to DF sites was approximately 3 hours, and the average travel times from UF to PF sites and from PF to DF sites were similar (fig. 10). These analyses were not conducive to determining the overall effect of the beaver dams on water velocity in the reach during high flows; however, hydraulic models were used to examine the effects of beaver dams in the Fanno Creek at Greenway Park study reach on water storage during storm events (see White and others, 2025b).

Synoptic Measurements

Synoptic measurements were used to characterize the spatial variability in water temperature, dissolved-oxygen concentration, and pH in the south pond of the Greenway Park study reach during four hot summer afternoons (Poor, 2020). Data collected at the same depth (10 cm below the water surface) showed substantial spatial variation in water temperature and dissolved oxygen on the measurement dates in May, August, and September 2016–17 (fig. 11). During a synoptic sampling on September 12, 2017, dissolved-oxygen concentrations at a depth of 10 cm ranged from nearly anoxic (0.3 mg/L, 3 percent oxygen saturation) in the shaded old channel to supersaturated conditions (23.6 mg/L, 304 percent oxygen saturation) in the shallow, unshaded part of the pond (fig. 11).

In addition to the synoptic measurements collected at a depth of 10 cm, measurements were collected at several other depths. In general, warmer water temperatures and higher dissolved-oxygen concentrations were measured in shallow and unshaded areas, and cooler water with lower dissolved oxygen was found in deeper channels and in areas with abundant riparian cover (fig. 12; Poor, 2020). Sunlight and warm water temperatures in the shallow, unshaded pond promoted algal growth and photosynthesis, resulting in supersaturated dissolved-oxygen concentrations and high-pH conditions. The abundant sunlight and warm-weather conditions caused the water to stratify, with less-dense warmer water floating on top of cooler water in the deeper areas, thus providing a separation of conditions from top to bottom. Water at greater depths showed less spatial variation in both water temperature and dissolved oxygen compared to surface measurements (fig. 12). In all areas of the pond, an accumulation of decomposing organic matter in the sediments, combined with warm water temperatures that increase the rate of organic-matter decomposition, likely resulted in a substantial rate of sediment oxygen demand (Rounds and Doyle, 1997). In the absence of photosynthesis in the shaded areas or near the bottom of a stratified water column, a high rate of sediment oxygen demand could explain the near anoxic conditions measured in deeper water during these synoptic samplings.

Water Temperature

Water-temperature patterns in the Bronson Creek study reach (fig. 3) varied over the course of the year in response to changes in weather and flow conditions. Among the longitudinal monitoring sites (UB, PB, DB), the UB site was often the warmest location in winter and the coolest location in summer, probably because the creek’s source was subsurface water in the nearby foothills, and subsurface water temperatures are buffered compared to surface-water temperatures (Arrigoni and others, 2008). The upstream site also is partly shaded, which helps the water remain cooler during summer. Summertime temperatures were highest at the PB site, followed by the DB and UB sites (figs. 13 and 14), probably because of solar heating in unshaded reaches and surface/subsurface-water exchange processes occurring in the water-saturated areas of the valley bottom in the lower sections of the study reach.

In the Bronson Creek study reach, the 7dADM water temperature at the PB site was significantly (Tukey post-hoc p values <0.01) warmer than at the UB and DB sites during summers 2016 (May 19–September 30) and 2017 (May 16–September 30). From July through September in 2016 and 2017, the instantaneous water-temperature patterns at the DB site were markedly different from those at the PB site. The DB site diel fluctuation became more muted, and its daily maxima were often multiple degrees cooler than those at the PB site (fig. 14). Water in the creek often was near bank-full in summer, and the floodplain seemed to be water-saturated, indicating that the water table was near the land surface. Ponding due to multiple beaver dams likely influenced that elevated water table and may have increased flow through the substrate of the floodplain. Subsurface flow is often characterized by dampened diel temperature variation, lag times in daily peak temperatures, low dissolved-oxygen concentrations, and elevated specific conductance (Arrigoni and others, 2008; Hinkle and others, 2014). During all seasons, water at the DB site had the most time and opportunity to interact with the subsurface system of the floodplain, compared to the PB and UB sites; the UB site was close to the subsurface water source in the steeper headwaters of Bronson Creek. The temperatures shown in figure 14 correspond to low streamflow conditions of late summer, a time period when the influx of subsurface water likely affected water temperature at the DB site proportionately more than at other times of the year because of the long residence time.

In addition to the three continuous multiparameter instruments, five continuous water-temperature sensors were deployed at locations with varying depths and shade in the ponded area near the PB continuous water-quality monitor (fig. 13; table 2). From October 2016 to May 2017, water-temperature sensors at Sites 1–4, which were placed in interconnected channels along Bronson Creek, collected similar temperature measurements. Temperatures recorded at Bronson Creek Site 5, in mostly shaded and shallow water, were multiple degrees Celsius warmer than those from the other sites from October to May, suggesting that this particular body of water may not have been part of the main channelized flow of the stream. Minimum instantaneous water temperatures often were coolest at Site 5 compared to other sites (June–September), with large diel fluctuations (not shown). In this case, separation from the main channel flow during summer may have allowed other factors (such as air temperature fluctuations and shading) to influence water temperatures in this small body of water, creating relatively cooler temperatures at certain times of day compared to other locations in the ponded area. Complex channel morphology and localized factors created local areas with a range of temperatures, potentially providing aquatic habitats for a diverse set of aquatic organisms.

Temperature sensors placed upstream (Site 3) and downstream (Site 4), but at similar distances, from one mud-packed dam in the Bronson Creek study reach measured substantially different 7dADM water temperatures during summer 2017 (May 16–September 30). The upstream site, located in an unshaded part of the channel, was significantly cooler (Wilcoxon signed-rank p value <0.05) than the downstream site that had approximately 40 percent canopy cover during summer. The 7dADM water temperature generally was more than 1 °C higher at the downstream site during summer 2017 (fig. 13). The measured temperature difference may be a result of the water depths and characteristics of the dam. This dam was packed with mud such that little to no water was observed leaking through the dam. During summer, water in the pond upstream from the dam became stagnant and probably was stratified as solar radiation heated the uppermost layer of the water column. The mud-packed dam likely prevented any deeper, cooler water in the lower layers of the water column from flowing downstream. Instead, the warmer upper layer of the water column flowed over the top of the dam, transporting warm water to the sensor located downstream from the dam. This result suggests that dam-construction materials and characteristics can have an important localized effect on stream temperature.

Dissolved Oxygen

Dissolved-oxygen measurements from the continuous multiparameter instruments varied among the three sites in the Bronson Creek study reach, with summertime oxygen concentrations decreasing from the UB to the DB site (fig. 15). Patterns in these data may be partially explained through an understanding of the physical characteristics of this reach. For example, water entering the reach from the upstream foothills probably was well oxygenated by the reaeration of water in the steeper slopes, despite the initially low dissolved oxygen of its subsurface source. Decreasing dissolved-oxygen concentrations at downstream sites likely were caused by (1) increased oxygen demands from decomposing organic matter deposited in the lower-velocity ponded areas, (2) decreased reaeration in those ponded areas, and (3) increased interactions with hypoxic subsurface water.

Subsurface inputs and exchange can decrease the dissolved-oxygen concentration because of organic-matter decomposition and the general presence of a reducing environment in the subsurface. From May through September 2017, the PB site had a pronounced diel fluctuation—the effects of photosynthesis and respiration from the growth of algae and an accumulation of aquatic plants (fig. 15). However, the DB site did not have a defined diel pattern. The minimal diel fluctuations at the DB site were likely the result of little to no photosynthesis along with subsurface inputs containing low dissolved oxygen.

Specific Conductance

The hypothesis that subsurface flow interactions affected water quality at the DB site is supported by an examination of measurements of specific conductance. During summer, specific conductance increased by about 10 µS/cm from the UB to PB sites and increased by about 40 µS/cm from the PB to DB sites (fig. 16). A substantially higher specific conductance at the DB site is consistent with the movement of a fraction of the stream water through soil pore spaces because a greater interaction with soils tends to mobilize and transport dissolved solutes from the subsurface to the stream. Because the specific conductance was affected by subsurface exchange, a clear signal could not be traced from the UB to DB sites and the travel time of water through the reach could not be calculated as simply as it was in the Fanno Creek study reach; more explicit tracer tests, possibly combined with modeling efforts, would be needed to better estimate travel time and characterize interactions between subsurface and surface water at this location.

Effects of Dam Building

The effects of beaver activity on water quality also were captured through an unintentional case study that occurred at the UB site of the Bronson Creek study reach. A continuous water-quality monitor was deployed in a free-flowing section in May 2016. Around August 11, 2016, beavers built a dam 20 m downstream from the monitor, turning the original free-flowing section into a slow-moving, narrow pond that began to accumulate sediment and organic matter, and causing the dissolved-oxygen concentration to decrease and become erratic (fig. 17; app. 3). During the 7 days prior to the dam construction, instantaneous dissolved-oxygen concentrations ranged from a minimum of 7.5 mg/L to a maximum of 9.1 mg/L, whereas instantaneous concentrations ranged from a minimum of 1.8 mg/L to a maximum of 7.0 mg/L during the 7 days following the dam construction. Field observations and data collected over the following months confirmed monitor readings and indicated that the site characteristics were changing. The dam caused the site to transition from free flowing to almost stagnant and ponded water during low streamflow. The sensors were closer to the streambed than before dam construction because of the accumulation of sediment behind the new beaver dam, and the sensors were temporarily buried by sediment during a large storm in November 2016. The erratic readings may have been caused by heterogeneous conditions in the newly ponded water column or by beavers stirring up sediment that affected oxygen measurements near the bed. Suspension of sediment that contains organic matter can temporarily increase the biochemical oxygen demand of a stream (Rounds and Doyle, 1997), a process that would be consistent with the erratic readings at this site. The monitor was repositioned to measure the middle of the new water column in December 2016. Dissolved-oxygen readings from December 2016 through August 2017 were less irregular, indicating that the site may have reached a new equilibrium with the beaver dam still in place.