Journal Archive

Johnson Creek flows 42 kilometers from the foothills of the Cascade Mountains west into the Willamette River just upstream of Portland, Oregon <Figure 1>. The Johnson Creek watershed has a temperate marine climate typical of the Willamette Valley, on average receiving two-thirds of its 100-150 cm of precipitation a year between the months of November-March, almost exclusively as rain. Because of the seasonal variation in rainfall, as almost no rain falls in the basin between July and October, summer flow in the basin is almost exclusively supplied by groundwater sources, while flow throughout the rest of the year is provided by precipitation runoff. Although mainly agricultural near the headwaters in unincorporated Multnomah and Clackamas Counties, much of the land throughout the 140 square kilometer watershed is urbanized, including 175,000 residents living throughout large areas of Portland and Gresham as well as portions of Milwaukie, Happy Valley, and Damascus (JCWC, 2003). Channelization throughout both agricultural and urban sections of the creek is typical, with the lower 24 kilometers having been lined with rock in the 1930’s as part of a federal flood-prevention project. Ironically, Johnson Creek continues to pose flooding problems to nearby residents, in addition to exhibiting degraded water quality typical of urbanized streams. Despite elevated temperatures and toxics levels, Johnson Creek still supports small populations of salmonids, including native coho and Chinook salmon, as well as steelhead and cutthroat trout (JCWC, 2012a). For this reason, Johnson Creek has been the focus of restoration efforts in recent years aimed at increasing water quality and viable habitat (Levell and Chang, 2008).

<Figure 1>  
Johnson creek watershed (Portland, Oregon)

Kauffman et al. (2010) and ODEQ (2010) argue that riparian vegetation and canopy ratio in watershed and riparian areas have a significant impact on various stream temperature changes within Johnson Creek. ODEQ states that these riparian corridors and riparian buffer zones bring about a microclimate that surrounds the stream, where cooler air temperatures, higher relative humidity, and lower wind speeds are characteristic. Consequently, canopy cover ratio changes in Johnson Creek is one of the most important factors for understanding water quality issues, as vegetation changes in the riparian corridors are associated with water temperature change. The ODEQ has developed a draft total maximum daily load (TMDL) for Johnson Creek’s temperature. The numerous investigations of temperature in Johnson Creek over the years have consistently indicated that elevated temperatures are a problem throughout the watershed (Koehler, 2004).

To sum up, many previous studies on Johnson Creek concluded that its water quality is very poor (JCWC, 2005a). Oregon Water Quality Index by ODEQ ranges from 10 (worst) to 100 (ideal water quality), but Johnson Creek’s score were less than 60 for more than decade (DEQ, 2010). Consequently, urbanization influenced these flow levels because of expansion impervious areas. Peak flows normally occur in winter rain season. Summer low flows in July, August and September reflecting minimal groundwater contributions to streamflow in watershed (JCWC, 2005b). Even though many previous researches studied Johnson Creek, Johnson Creek has been struggling to have a successful restoration bringing good water quality back to it, so our study will try to suggest effective solutions for having improved water quality in Johnson Creek.

Water quality data used in this study consisted of 2000-2010 water temperature data taken from USGS monitoring stations at Regner Road (USGS gage # 14211400), Sycamore (# 14211500), and Milwaukie (#14211550), and TSS data, collected by the Portland Bureau of Environmental Services. TSS data consists of grab samples taken at 3 sites (Regner Road, SE 158th Ave, and Umatilla St. Bridge) from 2003-2010 on an approximately monthly basis, in all seasons and weather conditions. Sites chosen are located in the upper, middle, and lower portions of the watershed and correspond with the USGS gages, which also provided flow data. Daily air temperature data from 2000 to 2005 was collected from National Oceanic and Atmospheric Administration (NOAA) database.

Spatial data includes the watershed boundary, obtained from the State of Oregon, and stormwater basin boundary, developed by Multnomah County in consultation with municipalities in the watershed <Figure 1>. Data on developed land cover, canopy, and floodplain expanses was obtained from the Regional Land Cover Information System (RLIS) database, maintained by the Metro regional government.

The focus area of this study consisted of stormwater basin boundaries, called storm-watersheds, developed for each stream gage. Because of the heterogeneous stormwater management practices employed by jurisdictions within the watershed, it was decided that gage strom-watersheds would represent land areas with the most direct relationship with Johnson Creek.

This study analyzed water temperature data for the years 2000-2010 and used the Spearman rank correlation model to determine the relationship between air temperature and water temperature patterns at gage stations during summer seasons from 2000-2005. In a previous study, Chang et al. (2005) used Spearman’s rank correlation model to understand influence of land cover to water quality. The seven-day average of daily maximum temperature (hereafter called as 7DADM) was calculated to compare summer water temperatures over temporal and spatial scales, from 2000-2010 at each gage, then compared to canopy cover rates developed for each strom-watershed. The significance of all statistical analyses was tested by SPSS 26.

To analyze the impact of stormwater on TSS within the Johnson Creek Watershed, data for TSS and flow were related on both spatial and temporal scales in order to analyze differences within upper, middle, and lower portions of the watershed as well as the dynamics of change between the 3 different time periods for which continuous data was available, 2003-2005, 2006-2007, and 2008-2010.

Temperature and TSS/flow results were then related to land cover, using two metrics, canopy cover and development. In order to compare temperature and canopy, the area of canopy cover within a 100 meter buffer around the creek and its tributaries was calculated within each strom-watershed. For TSS analysis, the area of developed land was calculated both within gage storm-watersheds as well as within a 100 meter buffer around the creek and its tributaries within each strom-watershed.

Restoration efforts were also related to temperature analysis by reviewing restoration activities focused on increasing canopy cover around each of the three stream gages used in this study. Analysis of restoration plans was also undertaken, reviewing changes in priorities set forth in plans from 1991-2012.

This researcher examines a hypothesis that the number of exceeded seven-day average of daily maximum temperature (7DADM) and averaged daily maximum water temperature during the summer increase with urbanization influences, due to canopy ratio changes in the riparian near-area of Johnson Creek. Many previous studies have reviewed water quality issues in Johnson Creek between the 1900s and the 1990s, however much restoration work has been undertaken along the creek in the last decade in an attempt to control temperature. Therefore, this paper will focus on the impact of restoration work on stream temperature in the period 2000-2010.

Finally, this author discusses the relationship between water quality and stream flow under the urbanization pressure within the watershed from 2003-2010, based upon the results of data analysis with regard to hydroclimate, land-use, and restoration. Levels of sediment transport are expected to be highest during high flow periods, especially during storm events. This study also assess changes in this relationship over time, comparing 3 successive time periods (2003-2005, 2006-2007, 2008-2010), and space, comparing data from gauges in the upper, middle, and lower portions of Johnson Creek. Differences in stormwater management infrastructure and land cover within the watershed are expected to result in the majority of sediment being sourced from upper, agricultural portions of the creek.

Flow level change has an inverse proportion correlation between water temperature since 1999 (<Figures 2, 3, 4>). One of the serious issues in Johnson Creek is low stream flow during summer, which has caused high water temperatures. Furthermore, this high stream temperature (water temperature) issue is associated with air temperature change in Johnson Creek. Sinokrot and Stefan (1993) state that air temperature directly affects water increases in water temperature. Thus, our study collected air temperature data observed downstream (Milwaukie) from 2000 to 2005 <Figure 6>.

<Figure 2>  
Averaged daily streamflow and water temperature (°C) in Milwaukie, Johnson creek from 1999 to 2010 (cms: cubic meter per sec.)

<Figure 3>  
Averaged daily streamflow and water temperature (°C) in Sycamore, Johnson creek from 1999 to 2010 (cms: cubic meter per sec.)

<Figure 4>  
Averaged daily streamflow and water temperature (°C) in Regner, Johnson creek from 1999 to 2010 (cms: cubic meter per sec.)

<Figure 5>  
Correlation coefficient boxplot graph between daily water temperature and air temperature

<Figure 6>  
Air and water temperature change pattern at Milwaukie of 2000 (above) and 2005 (bottom) (°C)

The results of our research indicated the importance of land cover factors in understanding stream temperature in the Johnson Creek watershed in Oregon (Lalli and Parsons, 1997). The effects of human modification in Johnson Creek have caused an increase in stream temperature, and elevated water temperatures are very harmful to aquatic organisms. In addition, water temperature partially determines the concentration of dissolved oxygen and carbon dioxide, which are profoundly linked with biological processes.

The following graphs indicate air temperature and water temperature change patterns during summer days (low-stream flow season) in Johnson Creek. The day with the highest recorded air temperature had the highest water temperature in the graphs. Thus, this researcher can conclude that summer water temperature of Johnson Creek is largely controlled by air temperature (Correlation coefficient value=.829).

Johnson Creek indicated many of usual problems of urban streams, such as poor water quality. This researcher observed the poor water quality through our analysis of summer water temperatures in Johnson Creek. Sub-basin land cover variables, such as canopy cover rate, can be associated with water temperature change. The author conclude that existing land cover patterns will continue to decrease water quality with higher water temperatures if one does not provide any mitigation solutions for this issue in Johnson Creek.

<Figure 7>  
Boxplot graph in water temperature change at Johnson creek (2000 and 2009) (°C)

In <Figure 8>, the average daily maximum water temperature at most Johnson Creek gauging stations during the summer were observed to be higher than the state standard. The canopy cover rate in Johnson Creek was measured for evaluating land cover change between 2006 and 2010. Isaac et al. (2010) argue that solar radiation within 1 Km of watershed buffer areas is a major factor in stream temperature change. Thus, this researcher collected canopy cover rate data at Milwaukie, Sycamore, and Regner in 2002 and 2006 <Figure 9>.

<Figure 8>  
Average daily maximum water temperature during summer in Johnson creek

<Figure 9>  
Canopy cover ratio in Johnson creek 2000 and 2009

The canopy cover rate at Milwaukie in 2006 did not change much compared with 2002. On the other hand, Regner showed an obvious decrease, from 30% to 26%, while the canopy cover rates at Milwaukie and Sycamore in 2006 indicated small increases. The trend of canopy cover ratio change indicates a decrease in overall canopy cover rate in Johnson Creek. These rate changes are associated with riparian vegetation change and development in the upper stream of Johnson Creek. After development upstream, riparian vegetation no longer provides enough shade to keep stream temperatures from rising in the Johnson Creek watershed area, indicating that water temperature during the summer is influenced by canopy cover ratio. This researcher analyzed 7DADM during the summer, from 2000 to 2010, to examine our assumption that decreased canopy cover rate brings about increased water temperature, using the Johnson Creek case. The number of weeks that stream temperatures exceeded the standard temperature of 18.0°C stated by ODEQ was 14 in 2000 and 16 in 2010 at Milwaukie, and 10 in 2000 and 16 in 2010 at Sycamore. Both Milwaukie and Sycamore show increased values in 7DADM evaluation. Regner had 5 weeks of temperatures above the standard in 2000 and 13 in 2010, an increase of 160%.

<Figure 10>  
7DADM in Johnson creek of 2000, 2005, and 2010

Consequently, this researcher was able to conclude that canopy cover rate influenced water temperature in Johnson Creek, because Regner had a reduced canopy cover rate and remarkably increased 7DADM value in the analysis (<Table 2> and <Figure 11>).

<Figure 11>  
Number of exceeded 7DADM in JC2 (Milwaukie), JC6 (Sycamore) and JC15 (Regner)

For floodplains, this researcher compared data from 2003 and 2006 and noted, while some areas of Johnson Creek had seen a significant decrease in the floodplains others saw expansion <Figure 12>.

<Figure 12>  
Change in the Johnson creek floodplain between 2003 and 2006

For vegetation the focus was on canopy cover. Forest canopy in particular is important as it has the biggest effect on stream temperature and thus, water quality (Roth et al., 2010). In examining the overall area of the sub-basins, the lower sub-basins, Milwaukie and Sycamore, the forest canopy increased. In the upper sub-basins forest canopy decreased. Sycamore showed the biggest increase with a 3% gain while Milwaukie had a gain of 1%. Regner decreased by 4%<Figure 13>.

<Figure 13>  
Percentage change of tree canopy within the sub-basins

In examining the buffer data for canopy cover <Figure 14>, changes were more pronounced within the buffer than in the watershed as a whole. For Sycamore, there was a 14% increase in canopy cover in a mere 4 years. Milwaukie had a 4% gain and Regner notched a 2% gain, a change from when canopy cover was analyzed in the entire Regner basin rather than having a 100 meter parameter.

<Figure 14>  
Percentage change of tree canopy within the 100m buffers

Vegetated riparian zones are narrow, minimal or lacking throughout much of the watershed. Johnson creek was calculated to have only 40% of its cover still intact (The Conservation Registry, 2002). Analysis has shown that for salmon and trout to thrive in Johnson Creek, the stream channels shaded area needs to reach 80% (Dent et al., 2005). To reach this goal they are trying to plant trees at all points possible along the stream banks. The amount of potential shading condition is shown in <Figure 15>.

<Figure 15>  
Potential canopy coverage in Johnson creek watershed source: Conservation registry 2002

The reference condition for the riparian zone is a continuous corridor of 91 meters, composed of mature native vegetation, with <2 stream crossings/ KM of stream length (The Conservation Registry, 2012a). Of the 179 acres within the riparian corridor, approximately 61 acres (34%) are developed urban areas including streets, yards, buildings, and parking lots. So far approximately 40 acres have been planted in the past 5 years. Unfortunately our analysis shows that this warming trend continues during the period of our study, 2000-2010, and that restoration efforts have yet to lower the creeks temperature.

The first site referenced was the Sycamore Gauge Restoration <Figure 16>. There were two separate projects at this site, both formerly private residences. The objective for the first site was to remove a home, while enhancing and restoring the functional ecological values of the site (The Conservation Registry, 2012b). Once the structures were removed from the site, the Watershed Revegetation Program (WRP) began work in 2001. 2520 native trees and shrubs were installed in the project area, along 76 meters of stream-bank. After three years of maintenance and monitoring, the desired plant density levels had been reached. As of 2006, Japanese knotweed and garlic mustard were present along the stream-bank and were treated. As of 2011, no knotweed was detected though garlic mustard was still apparent on the property to the east. Himalayan blackberry was a minor problem, but due to lack of treatment, by 2011 blackberry had encroached significantly into the planted area, with vines growing over plantings. Even with the invasive species overshadowing the plantings, the survival of plantings has been high (The Conservation Registry, 2012b). Although the plants are surviving and a thick patch of young alders are present, our analysis shows no correlation between the project and the temperature of the creek.

<Figure 16>  
Restoration site at Sycamore gage

The project on the adjacent “C” shaped property to the east has a successful establishment of native vegetation. The WRP planted 1500 native trees and 500 native shrubs along 150 meters of stream-bank in 1998. As of 2006, since the plants were well-established and the five-year commitment to maintenance and monitoring had transpired, the project was no longer maintained by the WRP. Portland Public Parks and Recreation took over maintenance. As of 2011, native vegetation was still present, although the untreated area inside of the “C” was covered by 3 m tall blackberry that was aggressively encroaching into the planted area (The Conservation Registry, 2012c). The Johnson Creek Watershed Council (JCWC) again needs to intervene to help guide the maintenance of the restored area so the natural plantings are able to grow large enough to establish their own sustainable ecosystem.

The second reference site is located just up from and including the confluence of Johnson creek with the Willamette River <Figure 17>. The project reach extends 411.5 meters from the 17th Street Bridge downstream to the mouth of Johnson Creek (The Conservation Registry, 2012d). This confluence area supports juvenile rearing and adult migration of federally listed Chinook, coho, and steelhead. However, due to decades of urban development, the reach is heavily degraded. Stream banks have been armored and confined, large wood and riparian forests have been removed, former floodplains have been filled and are crisscrossed with sewer and electric infrastructure, and until the 1960’s, the Portland-Milwaukie Streetcar crossed Johnson Creek at this site. Over the past 10 years, JCWC has worked with landowners along the project reach to conduct riparian restoration projects. In 2003 they installed 1,600 native trees and shrubs and planted 200 shrub cuttings in the stream bank, and 520 native trees and shrubs the next year (The Conservation Registry, 2012e). In winter 2012, community volunteers planted native trees, shrubs, and seeded wild flowers in disturbed areas within the six acres of adjacent riparian forests. The goal of this restoration is to inspire and facilitate community investment in the Johnson Creek Watershed for the protection and enhancement of its natural resources (Larson et al., 2010; JCWC, 2012b).

<Figure 17>  
Restoration site at Klein point/Johnson creek confluence

The site has had mixed levels of plant survival. In the second year of monitoring it was observed that beaver had significantly damaged the plantings in an area that extended 7.62 m. up the slope from the creek. Despite the varying levels of survival within the project area, the WRP’s plant density targets have been achieved for this project. Invasive species treatment and control has been a constant part of the restoration efforts at this site. Appendix photographs 1-3 show the restoration results.

The headwaters of Johnson Creek are in agricultural areas in the Cascade foothills, and along its 26-mile course the lands surrounding the creek change quickly from agricultural, to residential, to densely urbanized industrial. Management of stormwater runoff differs both between and within land use types. For example, in the eastern portion of the watershed, although mainly agricultural, many of these lands manage runoff through a series of ditches and drains, which act similarly to the networks of storm drains common in urban areas in their efficiency of stormwater delivery to streams. In urban areas, stormwater management ranges from storm drains directly connecting runoff from urbanized impervious land cover into Johnson Creek to storm drain networks that either route runoff out of the basin entirely (e.g. Portland’s combined sewer) or into management systems designed to allow for infiltration (such as dry wells, “green streets”, and pervious pavement).

In the course of urban development, sources of sediment for entering a stream decline after the construction phase (Paul and Meyer, 2001); however the history of the Johnson Creek watershed is such that most of the infrastructure from urban and agricultural development was in existence before hydrologic data collection began in 1941. For this reason, hydrologic trends associated with urbanization (more frequent, higher volume storm events, flashier hydrograph) were not detected from 1941-2006 in Johnson Creek (Lee and Synder, 2009). Major sources of sediment within Johnson Creek can then be assumed to originate in agricultural areas in the upper watershed.

Sedimentation is recognized by the EPA as one of the most pervasive water quality problems faced by streams in the United States (EPA, 2002). The major concern regarding sedimentation in Johnson Creek is the effect it has on salmon spawning, as the presence of fine sediments can reduce the survival rate of salmon spawn. Pollutants can also attach themselves to soil particles, and in the case of Johnson Creek, research has shown that DDT continues to be present in the stream decades after its abolition (JCWC, 2005a). The Oregon Department of Environmental Quality has judged that the best way to manage DDT in Johnson Creek is to reduce the level of Total Suspended Solids (TSS) in the stream (Brown and Caldwell, 2009).

Sediment transport in Johnson Creek is driven by wet-season rains (defined here as November 1-April 30), which is demonstrated by a strong correlation between TSS and Flow during the wet season <Figure 18>. Low TSS and flow amounts during the dry-season (May 1-October 31) reveal that total sediment transport during the wet-season also dwarfs dry-season transport <Figure 19>. These findings are supported by recent USGS analysis (Stonewall and Bragg, 2012) which found that from 2007-2010, 70% of suspended sediment in Johnson Creek was transported in the months of November, December, and January. For this reason, analysis of TSS in Johnson Creek was based on wet-season TSS data.

<Figure 18>  
Spatial difference in wet-season TSS/flow relationship 2003-2010

<Figure 19>  
Spatial difference in dry season TSS/flow relationship 2003-2010

Spatial analysis shows that flow has a greater impact on suspended sediment at upstream gages, which support the hypothesis that agricultural lands in the upper watershed are the major source of suspended sediment in Johnson Creek <Figure 18>.

Temporal analysis yielded the finding that between the years of 2003-2010, flow is clearly having an increasingly greater impact on suspended sediment within Johnson Creek (<Figures 20-22>). At each of the three gages, equal flows are resulting in higher TSS readings in the years 2006-2010 than in 2003-2005. At both the Sycamore and Milwaukie gages, this relationship increases with each subsequent time period. At Regner Rd., located in the upper watershed, this trend reversed itself during the 2008-2010 period, however was still significantly greater than 2003-2005. This finding would indicate that between 2007-2010, sources for sediment decreased in the upper watershed.

<Figure 20>  
Wet season TSS/flow relationship, Milwaukie gage 2003-2010

<Figure 21>  
Wet season TSS/flow relationship, Sycamore gage 2003-2010

<Figure 22>  
Wet season TSS/Flow relationship, Regner gage 2003-2010

Analysis of development within the Johnson Creek strom-watershed demonstrates that significant areas of both the upper and lower strom-watersheds were developed between 2003-2010, while development occurred at a slower rate in the middle strom-watershed<Figure 23>. Within 100 meters of the creek and its tributaries, analysis showed that rates of development were again highest in the upper and lower strom-watersheds and slower within the middle strom-watershed <Figure 24>.

<Figure 23>  
Changes in development within Johnson creek strom-watersheds 2003-2010

<Figure 24>  
Changes in development within 100 m stream buffer within Johnson Creek strom-watersheds