From:	WILLIAMS Robert K [robert.k.williams@state.or.us]
Sent:	Thursday, March 02, 2017 9:49 AM
To:	'SEIDEL Paul'
Subject:	Townsend EDB in Groundwater Study

Paul,
I have reviewed the EDB study by Evren Northwest. My comments follow below:
 
Section 2.2:   History of Pesticide Use at Townsend Farms � This section introduces the argument that 
EDB cannot have been sourced from Townsend Farms.  Berry farming is said to have begun in 1961 and 
utilized Telon II rather than the more expensive EDB. (Both fumigants date back to the 1940s.) 
 
Section 4.1:  Surface Water in the Immediate Vicinity of the Subject Property � This section begins with 
the statement that there are no lakes, creeks, or wetland features on the property but goes on to 
note that there is an unnamed �storm water retention� pond in the NE corner.  In the past this pond 
was the result of excessive runoff from spray irrigation of process waste water.  I do not have current 
information about the spray irrigation practices at the site. The site map (Figure 1) shows that �No 
Name Creek� crosses a piece of the property in the south. This creek also parallels the southern 
property boundary and there has been sediment sampling in the creek to assess pesticide runoff 
impacts from the property. 
 
Section 4.2: Site Hydrostratigraphic Units � Under Troutdale Sandstone Aquifer (TSA) there appears to 
be some confusion about unit thickness.  The first sentence reports that the upper 2/3 of the unit is 
sandstone and the lower third is conglomerate. Later in the section it is reported that the sandstone is 
70 feet thick and the conglomerate is 200 feet thick. It is probably best to state that thickness is 
variable.
 
Section 4.3:  Summary � Conceptual Hydrogeological Model and Section 4.4.3:  Groundwater Capture 
and Time of Travel Curves � The first section discusses regional groundwater flow and introduces 
Figure 8. This figure is apparently a compilation of two or more time of travel modeling efforts. The 
DEQ facility profiler is referenced as the source of the figure but there is no explanation of how studies 
were done. (I am not aware that this kind of data is available on the facility profiler; it looks like 
combined GIS shape files). At a minimum it is misleading to combine these studies which have been 
prepared with different modeling methods.  By their elongated shape the green capture zones clearly 
incorporate (an assumed?) groundwater flow direction and gradient (particle tracking analysis) while 
the yellow and purple capture zones (concentric circles typical of a wellhead protection analysis) do 
not. Wells CWV #2 and #3 are shown interacting with each other but not with the other wells. The 
figure is apparently being used to show that EDB contamination in TFI #3 is not related to similar EDB 
contamination in CoF wells.  The CWV wells do not indicate EDB contamination and their capture zones 
do intersect TFI#3, so it is suggested that upgradient pumping is not likely to draw EDB from TFI to 
supply wells. There is no background information on how the modeling was performed or any 
information on the data that was used. If the CoF wells capture zones were portrayed using the same 
kind of modeling used for CWV wells then the capture zones would probably all intersect with the 
nearby TFI well. All wells would likely interact with each other if modeled together.  Capture zones 
depend on pumping rates, pumping duration, flow direction, groundwater gradient, screen interval, 
etc.  None of these variables appear to be considered. 
 
Section 4.4.4: Wells in Geologic Cross Sections - Under the first bullet it is asserted that the NE trending 
groundwater gradient, which aligns with surface topography, is demonstrated by the cross section in 
Figure 5 and that this also explains why TFI�s well capture area does not include CoF well #6. As 
discussed above the capture area of TFI�s well is cylindrical and indicates that it was calculated using a 
different method than that of the CoF wells. Capture area depends on gradient, flow direction, and 
pumping rate. There is no indication from the report that the kind of data collection and analysis 
needed to make the argument was available to the authors. Indeed, the static water levels do suggest 
that flow would be to the NE but high volume supply well pumping can easily overcome static flow 
direction.  Pumping in COF well #6 (likely at a fairly high flow rate) could draw in groundwater from the 
TFI well area. In Figure 5 the confining CU1 unit slopes downward to the SW and could enhance flow in 
a SW direction during CoF supply well pumping. 
 
Discussion under the third bullet depends on analysis of the areal extent of the captures areas in 
Figure 8 which is invalid due to the different methods used to calculate the capture zones.
 
The two cross sections of Figure 5 and 6 suggest that the TSA units in the vicinity of TFI may form a 
mound structure. At the nearby East Multnomah County cleanup site (ECSI #1479) there is a TSA 
mound structure which resulted in radial spread of contamination over a wide area. It is possible that a 
similar structure exists at TFI and could have contributed to upgradient movement of contamination. 
 
Section 5.3.1: Rapid Soil Solutions Letter Report dated August 26, 2015 � DEQ has commented on this 
work in the past (true?). In short the soil sampling from 2-6 feet bgs would be extremely unlikely to 
detect EDB used decades earlier. Due to the physical/chemical properties of EDB its traces might be 
found in deep groundwater below the site but not in near surface soil. Static water levels indicate that 
groundwater flows NE but under pumping conditions local groundwater gradients can be reversed. 
The flawed capture zones analysis plotted in Figure 8 (text appears to erroneously mention Figure 7) 
have been discussed above.  
 
Section 5.3.2: Extent of EDB in Ground Water - Under the second sub-bullet for TFI #3 it is noted that 
of all the wells in the study area the maximum concentration of EDB was detected in this well. The text 
goes on to assert that this fact does not indicate that TFI is the source but rather that this particular 
well may be in more of a direct line than CoF wells to the source area. The text attempts to forestall 
the conclusion that the highest concentration points to the source but does not present a convincing 
argument. Another fact that is not mentioned anywhere in the study is that the pesticide mixing area 
for the facility, during active farming years, was located next to the two TFI wells under discussion. 
There are no known reports of spills and it is asserted that EDB was not used at the facility. None of 
the concentrations in any of the wells are particularly high so it is hard to use concentration to pinpoint 
a source. No specific source has been identified. The low concentrations detected so far suggest that 
multiple and/or diffuse sources may exist. 
 
Fourth sub-bullet, city of Fairview Well #6: It is reported that this well has had the highest EDB 
concentrations of the CoF wells and more detections than all the other supply wells. It should also be 
pointed out that this supply well is the nearest well to the TFI wells and thus more likely to be 
impacted by any contamination at TFI. 
 
Section 6.0 Discussion: It is pointed out that sampling intervals are not consistent and pumping rate 
and intervals for supply wells are unknown.  Indeed, all these factors make the analysis more difficult 
and uncertain. 
 
Figure 10 is used in a data modeling attempt to show that the source for EDB must be upgradient or 
south of the CoF wells. The model and figure assumptions are difficult to follow and not well 
explained. This effort appears to suffer from the fact (from Table 1) that except for one TFI well #3 
sampling event on 2/11/05 (result=ND) no TFI wells were sampled between 2001 and 2010. The first 
EDB detections in the area were in TFI well #2 in 1994. This well extends to 136 feet below ground 
surface (bgs) but was abandoned by 2003.  TFI well #3 is screened from 205-245 bgs so it�s much 
deeper screened interval would be expected to detect contamination at a later date. Fairview wells #6 
and #5 first detected EDB in 2005 but the next earliest sampling event for these wells was in 2002. The 
effort seeks to find patterns in the temporal variability of detects vs non-detects for very low 
concentrations near the limit of detection. At these levels, detection vs non-detection could easily be 
related to sampling technique, lab error, etc. The lack of consistently collected sampling data and the 
great extrapolation from limited data makes this kind of modeling effort implausible. The many other 
variables such as groundwater pumping rate, pumping intervals, and geologic heterogeneity also make 
this kind of analysis difficult. The fact that TFI well #2 was the first well to detect EDB by over 10 years 
while nearby CoF wells were being sampled every 2-3 years may be important. 
 
7.0:  Summary and Conclusions: In the long experience of the DEQ Tanks program, EDB from gas 
station sources is very rare in the Portland area. An agricultural source seems likely for this area. The 
capture zone/wellhead protection analysis of Figure 8 is improperly constructed and not suitable for its 
intended use without revision.
 
The text states that if the EDB detects in TFI #3 reflect EDB detects in CoF well #6 then there should be 
early detects in CoF #6  at the same early date of detection in TFI #2. This might be the case if there 
was an upgradient source but the fact that EDB was not detected in CoF #6 until 2005 suggests a 
possible downgradient source. In any case the much greater depth for the screen interval in CoF #6 
also supports later detection.
 
DEQ Hydrogeologist summary: There are several reasons to think that Townsend Farms is not the 
source for EDB in study area wells, including statements that EDB was not used at the farm, the farms 
generally downgradient position, and the likely presence of other berry farms in the area.  However, 
there are also several reasons to think Townsend Farms could have been a source including: the first 
and highest EDB detections occurred at the farm and the possibility of the CoF supply well pumping 
drawing contaminants upgradient to the well, possibly aided by a TSA mound. Considering the 
information available to date TFI cannot be ruled out as a potential source for EDB contamination.  
 
DEQ was not given the opportunity to review the work plan for the Evren Northwest EDB Study. As 
suggested in DEQs �2010 Groundwater Contamination Source Evaluation � Fairview City Wells� for ECSI 
# 5279, a more useful approach for further investigation might have been to search for additional 
undocumented deep wells at older residences in the study area which could be sampled to better 
define the plume extent. There is another TFI well (MULT 82081) that has not yet been sampled. This 
is a likely next step. New deep monitoring wells would be expensive to install and hard to locate. 
Another useful avenue would be to research the locations of historic area berry farms. Townsend 
Farms staff likely have this kind of historic information. 
 
As mentioned above, the low EDB concentrations in all the study area wells make it difficult to use 
concentration to pinpoint a source and no specific sources have been identified. The low 
concentrations detected suggest that multiple and diffuse sources may exist. A rigorous study to 
pinpoint a source is likely to be a significant undertaking and considering the age and likely diffuse 
nature of releases is unlikely to be successful.  Unless a significant source area is found then remedial 
efforts would be impractical. Supply well wellhead treatment might be an economical solution for 
reinstating the use of Fairview groundwater.
 
Robert Williams
Oregon Department of Environmental Quality
700 NE Multnomah St., Suite #600
Portland, Oregon  97232
503-229-6802
williams.robert.k@deq.state.or.us