RESULTS OF BACTERIA SAMPLING IN THE

WILSON RIVER

Joseph M. Bischoff

and

Timothy J. Sullivan

April 1999

Report Number 97-16-02

E&S Environmental Chemistry, Inc. P.O. Box 609 Corvallis, OR 97339 ABSTRACT Water quality monitoring was conducted at eight sites on the Wilson River during the period late September, 1997 through early March, 1998, from river mile 8.6 to river mile 0.2 near where the river enters Tillamook Bay. Samples were collected approximately weekly by the Tillamook County Creamery Association (TCCA) during the course of the study, plus at more frequent intervals during two storm events in October, 1997 and March, 1998. Samples were analyzed by TCCA for fecal coliform bacteria (FCB) and E. coli. E&S Environmental Chemistry, Inc. provided the data analysis and presentation for this report. FCB concentrations and loads in the Wilson River were higher by a factor of two during the October, 1997 storm than during any of the other five storms monitored by TCCA or E&S. Similar results were found for the Tillamook and Trask Rivers by Sullivan et al. (1998b). Lowest loads in the Wilson River were found during the monitored spring storms in 1997 (by E&S) and 1998 (this study). By far the highest FCB loads were contributed by the land areas that drain into Site 7 (in the mixing zone just below the TCCA outfall) during the October 1997 and March 1998 storms. This site was the only site in the Wilson River basin that has contributing areas occupied by urban land use. Relatively high FCB loads were also found at a variety of other sites. A consistent relationship was not observed between FCB loads and land use among the other sites sampled in the Wilson River watershed. There was little or no relationship observed between measurements of FCB versus E. coli. At FCB concentrations greater than about 200 cfu/100 ml, results of E. coli measurements showed no indication of increasing above 200 colonies/100 ml with increasing FCB, even to FCB measured values > 1,000 cfu/100 ml. The reason for the lack of correspondence between these two bacterial measurements is not known.

2 A. INTRODUCTION Tillamook Bay and its watershed have been the site of intensive water quality monitoring since November, 1996. E&S Environmental Chemistry, Inc. (E&S), under contract to the Tillamook Bay National Estuary Project (TBNEP), Tillamook County, Oregon, has conducted routine water quality monitoring in all of the five rivers that flow into Tillamook Bay. In addition, intensive storm sampling has been conducted by E&S at a variety of sites during six rain storm events between November, 1996 and March, 1998. Additional sampling was also conducted on behalf of TBNEP by other agencies and cooperating institutions during two of those storms. The Tillamook County Creamery Association (TCCA) sampled the Wilson River during the two storms and also performed routine monitoring of the Wilson River, beginning in September, 1997. Results of the work conducted by E&S has mainly been reported in two technical reports to TBNEP (Sullivan et al. 1998a,b). It is well known that several important water quality parameters typically exhibit significant episodic variability. Chief among these in the Tillamook Basin is fecal coliform bacteria (FCB). This study focused on episodic variability in the concentrations of FCB in the Wilson River during storm events that occurred during the rainy season. An additional objective of the storm sampling was to estimate the storm-based loading of FCB to the bay from the Wilson River watershed. Prior to and during the course of the monitoring effort, it became increasingly clear that FCB contamination was a widespread problem throughout the basin, with highest concentration in the Tillamook River, and highest loads in the Trask and Wilson Rivers (c.f., Jackson and Glendening 1982, Sullivan et al. 1998a). The source of this FCB was expected to be variable, with the primary contributors presumed to include dairy operations, septic systems, sewer treatment plants, and urban land use (c.f., Jackson and Glendening 1982). The storm monitoring effort was expanded in the fall of 1997 to include intensive sampling during two storms at about 30 sites on the Tillamook and Trask Rivers by E&S, and at eight sites on the Wilson River by the TCCA. One early fall and one late winter storm were selected for this component of the study. The principal objective of the intensive storm monitoring was to quantify the major contributing areas of bacterial loads along each of these river systems to allow evaluation of land use/bacterial load interactions. An additional objective was to evaluate differences in storm-driven pulses of bacteria at various locations in the watersheds of these three rivers. Most of the bacterial monitoring conducted to date in the Tillamook Basin has involved measurement of FCB. However, the Oregon Department of Environmental Quality (ODEQ) has recently changed its bacterial water quality standard from FCB to Escherichia coli. There is interest, therefore, in determining the extent of agreement between these two measures of bacterial contamination for Tillamook waters. The purpose of this report is to present the results of the intensive storm sampling and long term monitoring in the Wilson River conducted by the TCCA in a similar manner as were the long

3 term monitoring and intensive storm sampling results reported by Sullivan et al. (1998a,b). Results are presented for FCB concentrations and loads during the course of the study and comparisons are made of results for FCB versus E. coli measurements.

B. METHODS 1. Site Allocation and Sampling Samples were collected approximately weekly from October 1997 through March 1998 at eight sites on the Wilson River for fecal coliform and E. coli bacteria (Figure 1). Samples were collected and analyzed by the TCCA. All samples were analyzed for E. coli and most were also analyzed for FCB. Storms were selected by the expected duration and intensity of rainfall. The storms were selected in an effort to represent storms of different intensity and differing hydrological response. The early fall storm was preceded by a long dry period when there was little flushing of the watersheds. The winter storm was preceded by wetter antecedent conditions and more continual flushing of the watersheds due to frequent large rainfall events.

Figure 1. Sample site locations for long term monitoring and intensive storm sampling.

4 The Wilson River hydrograph throughout the period of study is depicted in Figure 2. It shows the pattern typical of the Tillamook Basin: low river flows (generally <500 cfs) during summer and frequent storms from October through March. The largest storms generally occur in the period November through January and often achieve peak discharge >10,000 cfs on the Wilson River. Flood stage on the Wilson River is designated as 14,100 cfs. The storms that were monitored by E&S and also those monitored by TCCA during this study are indicated in Figure 2. Two storms (October 1997 and March 1998) were sampled intensively at eight sites on the Wilson River from river mile 8.6 to the mouth. Site selection was determined jointly by TCCA and TBNEP staff. Priorities for site location and sampling included probable point sources of bacteria and nutrients and areas of intensive agriculture. River water samples were analyzed for FCB using the membrane filtration method described in Standard Methods for the Examination of Water and Wastewater (Greenberg et al. 1992). River water samples were also analyzed for E. coli bacteria using the Colilert method. This method reports the most probable value in colonies per 100 ml with a 95% confidence interval.

2. Estimation of River Discharge The USGS maintains gauging stations on the Trask and Wilson Rivers. These data have been gathered and included in the hydrologic data set. Because the Wilson River is large and there are no inflowing tributaries of any size below the forest/agriculture interface, discharge at the gauging station was used unaltered at all of the TCCA sites. The difference between the true discharge at the highest site compared to the lowest site is undoubtedly very small (< 2%). No attempt was made to adjust discharge values to account for the small differences in the sizes of the contributing areas to each site.

3. Watershed Analyses Site locations are listed in Table 1. Subbasins that drain into each sampling site were delineated and digitized into a GIS coverage. FCB loads (cfu/sec) were calculated by multiplying the FCB concentration (cfu/100 ml) by the instantaneous flow (ml/sec). Data were collected over about a four to six day period during each of the storms. Details of transect analyses and calculations can be found in a previous report on storm sampling in the Tillamook Bay watershed (Sullivan et al 1998b). This analysis resulted in the identification of the river segments and their associated subbasins that most frequently contributed the largest loads of FCB to the river during these two storms. Watershed factors thought to influence loading of fecal coliform bacteria to surface waters were quantified using coverages produced by Alsea Geospatial (Corvallis, OR) for the TBNEP. Details on the production of these coverages were reported by Sullivan et al (1998b). Coverages were

5 Wilson River hydrograph throughout the period of study. Numbers are added to six storms sampled in storm monitoring efforts of TCCA and E&S. Figure 2.

6 Table 1. Wilson River site locations. Site Number Site Name River Mile WR1 Mills Bridge 8.6 WR2 DS RV Camp 7.4 WR3 Boat Ramp 6.2 WR4 Josi Farm 4.7 WR5 Sollie Smith Bridge 3.9 WR6 Highway 101 1.8 WR7 TCCA Outfall 1.3 WR8 Geinger Farms 0.2

projected in UTM zone 10 for this analysis. Subbasins draining to each sampling location were delineated and digitized. Since dikes are a prominent feature of the lower Wilson River, these delineations included the effects of dikes and drainage ditches on waterflow patterns so that only areas that appear to actually drain into the Wilson River were included as contributing areas in the land use analysis. Information regarding the location of dikes and their influence on the discharge to the river from various contributing areas was provided by Bruce Follansby (TBNEP). Land use type and development type were quantified from these coverages for each subbasin that drained into a particular sampling site, including area used for pastureland or agriculture, area of riparian zone, and area of urban or rural residential land use. Some of these classifications overlapped, such as rural residential and agriculture and were therefore quantified as both types. For example, if a polygon was coded as both rural residential and agricultural, then the area of the polygon was used in calculating both the total agricultural area and the total rural residential area. Centroids were produced for the development types designated as farm building clusters and rural residential clusters. Each represented a discrete cluster of residential homes or farm buildings. The total numbers of centroids and type for each sub-basin were then quantified.

4. Load Calculations Calculation of FCB load for each sampling site and occasion was accomplished by multiplying the bacterial concentration times the discharge. Because Wilson River discharge is essentially constant across space at a given time in the lower portion of the watershed where this study was conducted, changes in load values between sites is a result of changes in bacterial concentrations,

7 and not changes in discharge. However, temporal changes in load at a given site can be the result of both changes in discharge and changes in bacterial concentrations. Total storm loads for FCB were calculated for each storm event sampled. This was accomplished by calculating the area under the curve for the hydrograph of each storm (Sullivan et al. 1998b), in discrete segments corresponding to the available FCB measurements. For each segment, the FCB measurement taken at the beginning of the time segment was averaged with the FCB concentration measured at the end of the time segment. This average was then multiplied by the cumulative discharge during the time segment. Load estimates should be viewed as first approximations. More rigorous quantification of storm-based, and especially annual, loads would require additional monitoring data and the application of one or more non-point source pollution models.

C. RESULTS 1. Routine Monitoring Monitoring results for FCB and E. coli at Sollie Smith bridge (Site WR5) are presented in Figures 3 and 4, respectively. This is the primary monitoring site used in the studies by Sullivan et al. (1998a,b). Both measured bacterial concentrations and estimated loads are presented. Most samples had FCB concentrations in the range of 20 to 800 cfu/100 ml, although four samples exhibited higher concentrations (2,000-6,000 cfu/100 ml), all of which were collected during the October, 1997 storm. Samples having concentrations > 200 cfu/100 ml tended to be associated with increasing or high flow conditions (Figure 3, top). FCB loads were generally < 0.5 x 106 cfu/sec except during storm periods in October, November, and January. The highest measured load was 19 x 106 cfu/sec at peak discharge during the largest storm of the season (Figure 3, bottom). Monitoring results for E. coli are presented in Figure 4. Most samples had measured concentrations < 0.1 x 106 colonies/100 ml and estimated loads < 0.1 x 106 colonies/sec. As was the case for the FCB measurements, the highest loads of E. coli occurred during the large storm at the beginning of November. The peak load of E. coli measured during that storm was 22 x 106 colonies/sec. In general, there was little or no agreement between measured values for FCB and E. coli for samples collected at the same place and time (Figure 5). For relatively low concentrations (< 200 colonies/100 ml), there was not a clear pattern indicating that one measurement yielded higher results than the other (Figure 5, top). However, at higher concentrations of FCB (> 200 cfu/100 ml), there was not a concomitant increase in E. coli measured values as FCB increased to greater than 1,000 cfu/100 ml (Figure 5, bottom).

8 Figure 3. Concentration (top; cfu/100 ml) and load (bottom; cfu/sec) of fecal coliform bacteria and river discharge (cfs)at the primary monitoring site (Sollie Smith Bridge) on the Wilson River.

9 Figure 4. Concentration (top; colonies/100 ml) and load (bottom; colonies/sec) of E. coli bacteria and river discharge (cfs) at the primary monitoring site (Sollie Smith Bridge) on the Wilson River.

10 Figure 5. Relationship between E. coli bacteria concentrations (colonies/100 ml) and fecal coliform bacteria concentrations (cfu/100 ml). The top panel shows the relationship for values below 500 cfu/100 ml. The bottom panel shows the relationship for all values except for three outliers.

11 2. Storm Monitoring The first storm that was sampled in an intensive fashion, both spatially and temporally, was the first sizable storm of the fall season in 1997. It occurred during early October and resulted in 6.5 in (17cm) of precipitation at Tillamook. The fall storm was unique in a number of respects. Other than a very small two-day rain event three to four days previously (which did not increase flows much, especially in the larger rivers) and a moderate storm in mid-September, conditions had been dry for months prior to the storm (Figure 2). In addition, this was a fairly large storm, even by winter-storm standards. Discharge in the Wilson River reached nearly 9,000 cfs (Figure 2). Peak FCB concentrations corresponded temporally with large precipitation inputs, but decreased as the hydrograph began to rise significantly (Figure 6), following the same patterns found in the Tillamook and Trask Rivers (Sullivan et al. 1998b). The peak discharge occurred a little less than two days later, when bacterial concentrations had already decreased to about one tenth of the peak FCB concentrations. FCB concentrations achieved during the fall storm were very high, up to nearly 5,800 cfu/100 ml in the Wilson River. Very high concentrations were also reached in the Trask ( 2,800cfu/100 ml) and Tillamook Rivers (3,700 cfu/100 ml; Sullivan et al. 1998b). FCB loads in the Wilson River exceeded 1.0 x 106 cfu/sec throughout much of the storm, with peak loads over 1.5 x 106 cfu/sec (Figure 6, bottom). Peak loads were also high during this storm (by Tillamook River standards) in the Tillamook River, in the range of 0.5 to 0.6 x 106 cfu/sec, and about 1.2 x106 cfu/sec in the Trask River (Sullivan et al. 1998b). The second intensively-sampled storm occurred in early March, 1998. It was preceded by a four day period of generally decreasing discharge and little rainfall (Figure 7). The storm was moderate in size and increased Wilson River discharge to over 3,500 cfs, from a pre-storm baseline of just over 1,000 cfs. A total of 4.3 in (11cm) of precipitation was recorded in Tillamook. FCB concentrations demonstrated little response to changes in Wilson River flows or increased precipitation. Peak concentrations reached only 60 cfu/100 ml and loads did not increase above 100 cfu/sec. The Tillamook and Trask Rivers demonstrated a greater response to the spring storm, with concentrations reaching 400 cfu/100 ml in the Tillamook and 900 cfu/100 ml in the Trask River. FCB loads above 0.2 x 106 cfu/sec were only seen in the Trask River during this storm. Results of these storm load calculations are summarized in Table 2. This table includes the results of storm load calculations presented by Sullivan et al. (1998b) for storms 1-3 and storm 5, as well as calculations made using TCCA data for storms 4 and 6. The October, 1997 storm (storm 4) had by far the highest estimated storm load of the six storms during which the Wilson River was monitored. An estimated 353 x 1012 cfu of FCB was discharged to the bay by the Wilson River during that storm (Table 2), as compared with estimates for the Trask and Tillamook Rivers of 225 x 1012 and 108 x 1012, respectively (Sullivan et al. 1998b). Because the Wilson River primary

12 Figure 6. Concentration (top; cfu/100 ml) and load (bottom; cfu/sec) of fecal coliform bacteria, precipitation, and river discharge (cfs) at the primary monitoring site (Sollie Smith Bridge) on the Wilson River for the October 1997 storm.

13 Figure 7. Concentration (top; cfu/100 ml) and load (bottom; cfu/sec) of fecal coliform bacteria, precipitation, and river discharge (cfs) at the primary monitoring site (Sollie Smith Bridge) on the Wilson River for the March 1998 storm.

14 Table 2. Results of storm discharge and fecal coliform bacteria load calculations at the Wilson River primary monitoring site for the intensive storms monitored by the TCCA and E&S.

Cumulative Storm Discharge and FCB Loads Peak River Peak FCB Storm Cumulative Discharge Concentration Water Volume Total FCB Load No. Storm Dates Precip. (in) (cfs x 103) (cfu/100 ml) (m3 x 106) (cfu x 1012)

12 12/4/96- 12/8/96 5.2 10.4 596 57.4 59.9

22 1/16/97- 1/21/97 5.5 4.6 2720 33.1 174

32 3/19/97- 3/25/97 4.5 10.7 21 72.7 6.9

41 9/30/97- 10/8/97 6.5 8.8 5800 50.4 353.0

52 2/09/98- 2/18/98 5.4 5.1 220 54.9 42.9

61 2/27/98- 3/08/98 4.3 3.7 60 47.2 9.2

1 sampled by TCCA 2 sampled by E&S Environmental Chemistry, Inc.

monitoring site at Sollie Smith Bridge is located above sites WR6 and WR7, these load estimates do not include the large loads attributed to the Wilson River at these downriver sites. In the case of all three of these rivers, the total estimated FCB load during the October, 1997 storm was more than double the load of any other monitored storm. The smallest loads in the Wilson River were contributed by the two smallest storms (4.3 in and 4.5 in of precipitation respectively), which occurred in the spring of 1997 and in the spring of 1998 (Storms 3 and 6, Table 2).

D. DISCUSSION The bacterial fluxes measured at each site on the Wilson River throughout the duration of the first intensive monitoring effort (Storm 4) are shown sequentially in Figure 8. Data are presented from river mile (RM) 8.6 at Mills Bridge (top panel) downriver to RM 0.2, at Geinger Farms near where the Wilson River enters the bay. FCB loads in the Wilson River during the October, 1997 storm showed pronounced variability among sample occasions at a given site and among sites within a given time period (Figure 8). There was one peak in FCB load at three of the four upper sites on the Wilson River (RM 7.4, 6.2, 4.7); however there was no peak in the uppermost site, only a slight increase in loads associated with increased flows (Figure 8). The peak in FCB load during the later stages of the storm at two of the upper sampling sites (RM 6.2 and RM 4.7) corresponded temporally with increased flows. At the next two downriver sites, RM 3.9 and RM 1.8, the FCB loads decreased with the increase in flows in contrast to the two upriver sites (RM 6.2 and RM 4.7) where loads increased dramatically with

15 Figure 8. Changes in fecal coliform bacteria load (cfu/sec) at each river mile and discharge (cfs) at the monitoring sites on the Wilson River for the October 1997 storm.

16 Figure 8. Continued.

17 Figure 8. Continued.

18 increased flows. At site 7 (RM 1.3), TCCA outfall, FCB loads showed large amounts of variability with decreasing loads as flows increased. The largest loads were frequently in the mixing zone of the TCCA outfall ( RM 1.3). The FCB concentration at the most downriver site (RM 0.2) remained low (< 0.2 cfu/sec) throughout the storm except for one peak associated with intense rainfall in the earliest portion of the storm. The results for site 7 (RM 1.3) need to be interpreted with caution. Samples at this site are collected within the mixing zone of the TCCA outfall pipe. High concentrations of FCB in this point source discharge are not necessarily well-mixed with Wilson River waters. Thus, the load estimates at site 7 could be biased high. The cause of the decrease in estimated FCB loads from RM 1.8 to RM 0.2 is unknown. The only reasonable explanation that we know of for such consistency and such low values at site 8 is the possibility that samples collected at site 8, which is very close to the bay, were collected from a salt water wedge beneath the Wilson River discharge. Because specific conductance and temperature were not measured, we cannot verify if this actually occurred. The bacterial fluxes measured at each site on the Wilson River throughout the duration of the second intensive monitoring effort (Storm 6) are shown sequentially in Figure 9. At all of the upper sites (RM 8.6 to RM 3.9), the FCB load remained low throughout the storm. However each of these sites did show a slight increase in loads during the storm. The highest loads were observed at site 7, in the mixing zone of the TCCA outfall. The most downriver site (RM 0.2) had higher loads (> 0.2 x 106 cfu/sec) towards the end of the sampling period near peak discharge. These relatively high loads corresponded to the highest FCB loads at the adjacent upriver site, which also occurred at that time. During the March storm, however, loads at the upriver sites on the Wilson River were lower by about a factor of two as compared with Trask River sites. The very high loads observed at Wilson River sites 3, 4, and 7 during the October storm and at Wilson River site 7 during the March storm were much higher than at any of the Trask or Tillamook River sites. Although the TCCA outfall may constitute a part of this loading source, it is clearly not the only significant source because large FCB loads were also estimated for sites 3 and 4, at the boat ramp and Josie Farm, which are upriver from the TCCA outfall. The principal sites on the Wilson River that contributed FCB loads to the river during the October 1997 storm and the March 1998 storm are listed in Tables 3 and 4. The major load contributing sites (compared to their neighboring upriver sites) were scattered across the portion of the basin (lowlands) that was sampled. High load contributing areas were found as high upriver as WR1 (RM 8.6) and as low as WR 7 (RM 1.3; Table 3). For the October 1997 storm, the largest FCB loads were consistently found at WR7, followed by WR5 and WR4 midway through the portion of the basin that was sampled. The largest loads of FCB contributed during the March 1998 storm to the

19 Figure 9. Changes in fecal coliform bacteria load (cfu/sec) at each river mile and discharge (cfs) at the monitoring sites on the Wilson River for the March 1998 storm.

20 Figure 9. Continued.

21 Table 3. Rank order bacteria load contributions at monitoring sites on the Wilson River during all 12-hr time slices sampled during the October, 1997 storm. The top four load contributing sites are listed. Ordered by Load Per Ordered by Load Per River Ordered by Load1 Contributing Area2 Miles to Upriver Site3 Site R.M. Total Score Site R.M. Total Score Site R.M. Total Score WR7 1.3 43 WR7 1.3 43 WR7 1.3 44 WR5 3.9 20 WR5 3.9 22 WR5 3.9 21 WR4 4.7 18 WR4 4.7 19 WR2 7.4 20 WR2 7.4 16 WR3 6.2 16 WR4 4.7 17 1 Scores for each time slice were calculated as the load within that time slice at a given site, minus the load at the site immediately upriver. Total score is the sum of scores for each site across the time slices sampled during the storm. The site which contributed the largest load within a time slice was assigned a score of 5; the site contributing the second largest load was assigned a score of 4; and so on to the fifth largest load (score=1). The maximum possible total score would be 45 (largest load in each of nine time slices). 2 Before calculating the score, the load difference between each pair of sites was divided by the contributing area (ha) of the watershed that drained into the river between those two sites. 3 Calculated as above, except the loads were divided by the number of river miles between each pair of sites.

Table 4. Rank order bacteria load contributions at monitoring sites on the Wilson River during all 12-hr time slices sampled during the March, 1998 storm. The top four load contributing sites are listed. Ordered by Load Per Ordered by Load Per River Ordered by Load1 Contributing Area2 Miles to Upriver Site3 Site R.M. Total Score Site R.M. Total Score Site R.M. Total Score WR7 1.3 25 WR7 1.3 25 WR7 1.3 25 WR1 8.6 10 WR5 3.9 10 WR5 3.9 11 WR5 3.9 9 WR3 6.2 7 WR2 7.4 7 WR4 4.7 8 WR4 4.7 7 WR3 6.2 7 1 Scores for each time slice were calculated as the load within that time slice at a given site, minus the load at the site immediately upriver. Total score is the sum of scores for each site across the time slices sampled during the storm. The site which contributed the largest load within a time slice was assigned a score of 5; the site contributing the second largest load was assigned a score of 4; and so on to the fifth largest load (score=1). The maximum possible total score would be 25 (largest load in each of five time slices). 2 Before calculating the score, the load difference between each pair of sites was divided by the contributing area (ha) of the watershed that drained into the river between those two sites. 3 Calculated as above, except the loads were divided by the number of river miles between each pair of sites.

22 Wilson River were also at site WR7, followed by the highest site in the watershed, WR1 (RM 8.6; Table 4). The site located just below the TCCA outfall was consistently the principal load contributing site during the fall and spring storms, with site WR5 also being consistently a high load contributor. All sample locations except site 8 (RM 0.2) scored high as load contributors during one or both storms. Jackson and Glendening (1982) hypothesized that the primary sources of fecal coliform bacteria to the rivers may require saturated ground. Source types that fit this pattern include field application of manure, barnyards located some distance from the river, and inadequate on-site sewage disposal systems. We evaluated the relationship between the segments of the Wilson River that contributed the largest FCB loads and the land uses that were associated with the lands that drained into those river segments. The objective was to look for land use patterns associated with high FCB loading rates. One important note concerning the Wilson River is the series of dikes along the lower portions of the river for flood control. These dikes may have a significant role in the routing of overland flow to surface waters and maintain subsequent controls on the loading of FCB from the watershed. The percentage of each drainage area within agricultural, riparian, rural residential and urban land use categories was tabulated and is shown in Table 5. The frequency of farm building clusters and rural residential building clusters within each drainage area is listed in Table 6. These same land use analyses were repeated for a set of drainage areas (subbasins) defined in a different way. For this second set of land use analyses, the drainage areas contributing to each sampling site were restricted to those within 100m on either side of waterways (river, tributary streams and/or drainage ditches). The percentages of the area within the various land use types that occurred in the buffered streamside portion of the sample site drainage area subbasins are listed in Table 7. The numbers of farm building and rural residential building clusters within each of the buffered subbasins are listed in Table 8.

Table 5. Land use types by sample site subbasin in the Wilson River watershed.

%Rural %Forested Upland Sub-basin Area (ha) %Agricultural %Riparian Residential1 %Urban1 and Other WR8 274 47.5 4.0 14.9 0.0 48.5 WR3 477 32.0 4.1 9.5 0.0 63.9 WR6 267 27.5 3.5 10.0 0.0 67.5 WR4 594 26.4 3.1 5.0 0.0 70.5 WR7 346 17.2 4.7 3.5 2.2 76.2 WR5 314 12.4 1.9 6.8 0.0 85.8 WR2 5379 0.2 0.0 0.1 0.0 99.8 WR1 42000 0.0 0.0 0.1 0.0 100.0 1 Some areas were classified as both rural residential and as agriculture or urban and agriculture.

23 Table 6. Number of farm and rural residential building clusters within each of the sample site drainage areas included in the October 1997 and March 1998 intensively-sampled storms.

Rural Residential Building Farm Building Clusters Clusters Site #Centroids Site #Centroids WR4 6 WR1 37 WR8 5 WR3 22 WR5 2 WR6 16 WR6 2 WR4 15 WR3 2 WR8 13 WR1 0 WR2 9 WR2 0 WR7 8 WR7 0 WR5 1

Table 7. Land use types by sample site subbasin within a 100 m buffer zone on either side of the river plus all tributary streams and drainage ditches.

%Rural %Forested Upland Sub-basin Area (ha) %Agricultural %Riparian Residential1 %Urban1 and Other WR8 96 58.0 9.2 19.4 0.0 32.8 WR3 199 41.7 6.3 11.2 0.0 51.9 WR4 254 30.0 6.1 6.7 0.0 63.9 WR6 111 29.5 8.4 5.0 0.0 58.8 WR7 97 28.9 14.1 1.7 5.4 52.6 WR5 111 18.6 5.0 2.2 0.0 76.4 WR2 727 1.2 0.0 0.6 0.0 98.7 WR1 6378 0.0 0.0 0.4 0.0 100.0 1 Some areas were classified as both rural residential and as agriculture or urban and agriculture.

24 Table 8. Farm buildings and rural residential centroids occurring within 100 meter buffer of streams and drainage ditches.

Rural Residential Building Farm Building Clusters Clusters Site #Centroids Site #Centroids WR4 2 WR1 23 WR8 2 WR3 10 WR5 2 WR4 9 WR6 1 WR6 8 WR7 0 WR8 7 WR1 0 WR2 6 WR2 0 WR7 4 WR3 0 WR5 0

Evaluation of the spatial land use patterns within the contributing drainage areas to each of the monitoring sites generally did not reveal consistent patterns. However, site WR7 was consistently the highest FCB load contributor (Tables 3 and 4) and was the only site that had a contributing area which contained urban land use. Urban land use emerged as the most striking feature associated with high FCB loads in the Trask River (Sullivan et al. 1998b) and our results for the Wilson River were consistent with these findings. Site WR5 emerged as the second most frequent high-load contributing area during the two storms. We noted nothing significant about the land use within the contributing area to Site WR5, compared to other sites. During the spring storm, site WR1 was identified as a frequent load contributor (Table 4). The contributing area to site 1 contained by far the largest number of rural residential building clusters (Table 6). Additionally, 62% (23 centroids; Table 8) of these rural residential building clusters are within 100 meters of a stream or drainage ditch. Because the contributing area to this site includes essentially the entire upper Wilson River which is predominantly forested, percentages of land use for this site are meaningless. Nevertheless, the large number of rural residential building clusters and their proximity to surface waters suggests an important potential influence of rural residential land use on Wilson River water quality at site WR1.

E. CONCLUSIONS A number of conclusions can be drawn from the storm sampling that was conducted in the Wilson River. Key conclusions include the following: • FCB loads in the Wilson River differed from storm to storm, with the highest FCB loads contributed during the early fall storm.

25 • Major load contributing sites were scattered throughout the Wilson River basin, with high load contributing sites as high upriver as RM 8.6 and as low as RM 1.3.

• The land area that consistently contributed the highest loads to the Wilson River was the area that contained urban land use and the TCCA outfall.

26 F. REFERENCES

USDA-SCS. 1978. Tillamook Bay Drainage Basin Erosion and Sediment Study, Portland, OR.

Bernert, J.A. and T. J. Sullivan. 1997. Quality Assurance Plan for the Tillamook Bay National Estuary Project, River Water Quality Monitoring Project, E&S Environmental Chemistry, Inc., Corvallis, OR. Report Number 96-20-01.

Bernert, J.A. and T.J. Sullivan. 1998. Bathymetric Analysis of Tillamook Bay, E&S Environmental Chemistry, Inc., Corvallis, OR.

Greenberg, A.E., L.S. Clesceri, and A.D. Eaton, Eds. 1992. Standard Methods for the Examination of Water and Wastewater. 18th ed. American Public Health Assoc./American Water Works Assoc./Water Environment Federation, Washington, DC.

Jackson, J. and E. Glendening. 1982. Tillamook Bay bacteria study fecal source summary report. Oregon Department of Environmental Quality, Portland, OR.

Sullivan, T.J., J.M. Bischoff, K.B. Vaché, M. Wustenberg, J. Moore. 1998a. Water Quality Monitoring in the Tillamook Watershed, Final Report to the Tillamook Bay National Estuary Project, E&S Environmental Chemistry, Inc., Corvallis, OR.

Sullivan, T.J., J.M. Bischoff, and K.B. Vaché. 1998b. Results of Storm Sampling in the Tillamook Bay Watershed, Final Report to the Tillamook Bay National Estuary Project, E&S Environmental Chemistry, Inc., Corvallis, OR.

Tillamook County SWCD and Tillamook Bay Water Quality Committee. 1981. Tillamook Bay Drainage Basin Agricultural Non-point Source Pollution Abatement Plan, Tillamook, OR.

27