20080868 Ver 2_POTENTIAL FOR TRAWL BIAS IN THE PCS CREEKS STUDY_20180222

To: Jeff Furness — Nutrien
From: Randall Pulley — CZR Incorporated
Date: February 22, 2018
Re: Memo to address the potential for trawl bias in the Nutrien creeks study
In discussion of 2016 fish data analysis and results at the August 2017 Science Panel meeting, the
potential for bias in the trawl data was raised based on the time of day the creeks were trawled. This
memo has been prepared to address the issue of potential bias.
As background, the fish collection protocol was approved in the first creeks monitoring plan to meet a
'condition of the 1996 USACE mine permit for activities at the then Texasgulf mine operation in Aurora,
NC. Three creeks to be impacted by the 1996 permitted activities under Alternative E were sampled with
one additional creek as a control. Of those four creeks, three were suitable for juvenile fish collection via
trawl (including the control) and the fourth was only suitable for juvenile fish collection via fyke net due
to its shallow narrow configuration. While fish data have been collected in these four creeks almost every
year since 1998, a second USACE permit for continued mine operations was issued to PCS Phosphate
Company, Inc. (now Nutrien) in 2009 and required the addition of several new creeks to be impacted by
the 2009 permitted activities for Mod Alt L, as well as several new control creeks. An updated creeks
monitoring plan was approved by regulatory agencies with no change to fish collection methods or
protocols, and fish collection began in seven new creeks in 2011, two additional creeks in 2013, and the
final two creeks identified in the plan to begin fish collection in 2018.
Desktop research of easily accessible articles using several search terms revealed numerous articles which
discussed differences in fish use of shallow, deeper, and micro tidal estuarine habitats but only resulted in
one study particular to NC which displayed significance of finfish catch between day/night while testing
bycatch reduction devices (BRD) in shrimp otter trawls (Brown et al. 2017). (Other search terms, more
time devoted to the search, or paid subscriptions or access to specific journals or periodicals may result in
other pertinent NC studies.) Thus far, the most important common finding across the accessed articles or
abstracts appears to be that regardless of location (temperate, equatorial, shallow, or deeper water)
differences in nekton abundance, composition, and sometimes diversity was most affected by day/night
patterns. The day/night differences were suspected to be closely tied to predator/prey relationships. For
example, most predatory birds fly in the daytime, so some juvenile fish may be more inclined to be
reclusive during daylight hours, or that larger more predatory fish were more active in some environments
in either day or night.
2151 Alternate AIA South • SUITE 2000 • JUPITER, FLORIDA 33477-3902
TEL [phone] • FAX [phone] • [e-mail] • www.CZRINC.com
4709 COLLEGE ACRES DRIVE
SUITE 2
I N C O R P O R A T E D
WILMINGTON, NORTH CAROLINA 28403-1725
ti
ENVIRONMENTAL CONSULTANTS
TEL [phone]
FAX [phone]
[e-mail]
To: Jeff Furness — Nutrien
From: Randall Pulley — CZR Incorporated
Date: February 22, 2018
Re: Memo to address the potential for trawl bias in the Nutrien creeks study
In discussion of 2016 fish data analysis and results at the August 2017 Science Panel meeting, the
potential for bias in the trawl data was raised based on the time of day the creeks were trawled. This
memo has been prepared to address the issue of potential bias.
As background, the fish collection protocol was approved in the first creeks monitoring plan to meet a
'condition of the 1996 USACE mine permit for activities at the then Texasgulf mine operation in Aurora,
NC. Three creeks to be impacted by the 1996 permitted activities under Alternative E were sampled with
one additional creek as a control. Of those four creeks, three were suitable for juvenile fish collection via
trawl (including the control) and the fourth was only suitable for juvenile fish collection via fyke net due
to its shallow narrow configuration. While fish data have been collected in these four creeks almost every
year since 1998, a second USACE permit for continued mine operations was issued to PCS Phosphate
Company, Inc. (now Nutrien) in 2009 and required the addition of several new creeks to be impacted by
the 2009 permitted activities for Mod Alt L, as well as several new control creeks. An updated creeks
monitoring plan was approved by regulatory agencies with no change to fish collection methods or
protocols, and fish collection began in seven new creeks in 2011, two additional creeks in 2013, and the
final two creeks identified in the plan to begin fish collection in 2018.
Desktop research of easily accessible articles using several search terms revealed numerous articles which
discussed differences in fish use of shallow, deeper, and micro tidal estuarine habitats but only resulted in
one study particular to NC which displayed significance of finfish catch between day/night while testing
bycatch reduction devices (BRD) in shrimp otter trawls (Brown et al. 2017). (Other search terms, more
time devoted to the search, or paid subscriptions or access to specific journals or periodicals may result in
other pertinent NC studies.) Thus far, the most important common finding across the accessed articles or
abstracts appears to be that regardless of location (temperate, equatorial, shallow, or deeper water)
differences in nekton abundance, composition, and sometimes diversity was most affected by day/night
patterns. The day/night differences were suspected to be closely tied to predator/prey relationships. For
example, most predatory birds fly in the daytime, so some juvenile fish may be more inclined to be
reclusive during daylight hours, or that larger more predatory fish were more active in some environments
in either day or night.
2151 Alternate AIA South • SUITE 2000 • JUPITER, FLORIDA 33477-3902
TEL [phone] • FAX [phone] • [e-mail] • www.CZRINC.com

Water depth was also a component of differences in trawl catch but differences seemed to vary mostly
when the water column was deep enough to allow for stratification and different assemblages of nekton in
the water column. Dr. Frederick Scharf concedes bias between day/night otter trawl catch; however,
believes finfish catch in otter trawls is not significantly affected by daylight time of day bias (am/mid/pm)
(UNCW Fisheries Professor — personal communication).
The otter trawl net dimension and the shallow nature of the creek mouths in the Nutrien study allows for
almost the entire water column in each creek to be sampled within each 75 -yard 1 minute trawl. In terms
of logistics (e.g., distance from boat launch site), there is some variation in the time of day that individual
creeks may be sampled, but by specific design, no trawl creeks have ever been sampled at night.
Therefore, it is recognized that day time only trawl data does not fully represent all the fish that may be
active in the creek mouths during the night. It must be noted that fyke nets are left overnight, so while
fyke net data may include the additional night nekton, this creek type is very shallow (sometimes the
entire fyke net is not beneath the water surface) and by its nature may not have as much variation between
night and day as a deeper system.
A brief analysis of data collected since 2011 was performed on the trawl creeks. These seven years
include all creeks currently sampled but do not include all years of earlier data collected in the four
original creeks. Figure 1 displays the 10 trawl creeks from 2011-2017 divided into three arbitrary
categories (AM, MID, and PM); each category includes several hours of the day as shown on the graphs.
Figure 2 divides the seven years into pre and post categories for each creek impacted by the 2009
permitted activities for Mod Alt L and shows the percent by category for the combined pre and post years.
Jacks Creek is closest to the boat launch and Duck Creek and Porter Creek are the farthest. Not only do
logistics such as distance from the boat launch play a part in the time of day a creek may be sampled as a
general rule, but other deviations (e.g., number of creeks trawled in a given day) also varies across the
data set. These other variations (e.g., numbers of fish in the trawl from day-to-day, week -to -week or
year-to-year; amount of submerged vegetation in the net; weather which may determine time of day to
most safely cross the Pamlico river to Duck Creek) also may determine whether a particular creek is the
"last" creek of the day or the "first" creek of the next day.
Figure 1 leads to several observations: 1) no one creek is strictly sampled at a particular time of day; 2)
Jacks and Duck are the creeks most likely to be sampled in the morning; 3) Porter is most often sampled
in the middle of the day; and 4) Porter and Duck are least likely to be sampled in the afternoon.
In conclusion, we do not believe that the lack of randomized fish collection has resulted in biased trawl
data related to daylight time of day sampling (am/mid/pm).
Brown, K., B. Price, L. Lee, S. Baker, S. Mirabilio. 2017. Technical Solutions to Reduce Bycatch in the
North Carolina Shrimp Trawl Fishery. Final Report to the NCMFC and NOAA Bycatch Reduction
Engineering Porgram for the study period June -July 2015 and July -September 2016. NCDEQ, DMF,
Morehead City, NC.
Page 2 of 4
February 22, 2018

Cumulative Trawl Time of Day
2011-2017
41.76 39.56
46.15 50.55
51.65 54.95 56.04
T 59.34
O
0
v
E
c
v
U
L
a 43.96
48.35 57.14
43.96 41.76
40.66
PM = 1401 and later
Mid = 1001-1400
AM = 1000 and earlier
8.79
15.38
75.82
43.96 38.46
. 28.57
.29
6.59 440 8.79
0.00 3.30
Jacks Little Jacobs PA2 Drinkwater Tooley Long Muddy Porter Duck
Creek
Figure 1. Trawl times by percent of the day in April, May, and June 2011-2017 in 10 creeks in the Nutrien creeks study.
Page 3 of 4
February 22, 2018

23.08
66.67
15.38
Impact Creeks - Trawl Time of Day
2011-2017 Pre and Post Years
51.28
43.59
50.00
44.23
69.23
49.23
44.62
61.54
55.13
44.87
PM = 1401 and later
■ Mid = 1001-1400
■ AM = 1000 and earlier
11.54
16.92
E
75.38
76.92
7 � 38.46
30.77
5.13 � 7.69
11.54
0.00 `
Pre (4) Post (3) Pre (3) Post (4) Pre (2) Post (5) Pre (1) Post (6) Pre (5) Post (2)
Jacks Jacobs Drinkwater Tooley Porter
Creek
Figure 2. Trawl times by percent of the day in April, May, and June 2011-2017 in five impact creeks in the Nutrien creeks study
divided into pre- and post- years for each creek.
Page 4 of 4
February 22, 2018