20080868 Ver 2_Executive Summary 2020 PCS Creeks Report_20210701

EXECUTIVE SUMMARY
• This report on 2020 data collection is the eighth of the annual PCS creek reports to
present and emphasize summary data, per the ROD Special Condition V on Reporting
where "annual summaries of all data collected" are specified.
• This summary approach is also part of the continued attempt to meet requests of the
reviewers and the Science Panel to reduce the amount of material included in the
report and to simplify the presentation of data. This is the eighth report to utilize
multivariate techniques for multiple monitoring parameters. These techniques will
continue to be used in various combinations until findings warrant a change in
methodology. Analysis of large data sets often results in statistically significant
differences among parameters/variables; however, these differences are not
necessarily also ecologically significant. Detection, measurement, and identification of
ecologically significant issues present a challenge with any complex large-scale
project, especially in coastal environments where the "estuarine paradox" may
confound, or at least obscure, conclusions about flora and fauna which have evolved
resilient adaptations to wide fluctuations in abiotic conditions. This subject is covered
at annual Science Panel meetings.
• For the benefit of those who have never been to any of these creeks, or at least did not
visit them at the beginning of this study, Appendix J was prepared for the 2016 report.
Appendix J included selected photographs of vegetation monitoring plots from 1998
and from the most recent (as of 2016) vegetation monitoring of the original three
creeks (Jacks Creek-2014; Tooley Creek and Huddles Cut-2016). This collection of
photographs provided a visual comparison of physical changes at some of the
wells/plots since 1998. A copy of this photo appendix is available upon request. For
this 2020 report, Appendix G includes photographs of the 1998 and 2020 vegetation
monitoring at Jacks Creek and Tooley Creek as well as photos at Porter Creek and
Duck Creek from the 2012 and 2020 vegetation surveys.
• For all statistical analyses of laboratory results (sediment and water column metals
and ECU water quality parameters), when the limits of detection for a specific method
returned a "< the method detection limit", the "<" was removed and the limit entered as
a value in the spreadsheet. This value is the limit of 100 percent confidence in
detection; while the value may actually be less, if it were present it would not be higher
than this limit. This practice reduces "gaps" in the graphical depictions and increases
the value of the statistical analysis. Nonetheless, some "gaps" do occur due to
environmental conditions (e.g., water too shallow to collect a sample), or laboratory
conditions (e.g., equipment malfunction), or human error. Tabular depictions of metals
results do not remove the "<" and do not include those entries in calculations of
means/standard deviations.
• For all parameters monitored, changes in salinity have been evident: salinity increased
regionally over the early years of the study but steadily declined in the Pamlico River
and tributaries from 2011-2016 (or 2012-2016 for some creeks) until a rise in salinity in
2017 and fluctuations in years after. This short-term trend has been observed at every
data collection instrument (In -Situ Aqua TROLL) regardless of location (upstream or
downstream), creek basin (within South, Durham, Porter, or Duck creeks, or Pamlico
xxx

River drainages), creek type (impact or control), or size of drainage basin.
Correlations with environmental factors showed that salinity was most strongly
associated with Tar River discharge; each Aqua TROLL showed a moderate to strong
negative correlation between salinity and discharge. Salinity seems to be more
influenced by regional and climatic factors such as Tar River discharge, precipitation,
and wind. Similar salinities were observed at all Aqua TROLLs regardless of mine
impact, t-tests showed similar trends in salinity at most impact and control creeks, and
the concordance correlations showed that salinity between impact and control creeks
was similar even post -Mod Alt L. Analysis of salinity within post -Mod Alt L creeks
(Jacks, Jacobs, Drinkwater, Tooley, Huddles Cut, Porter, and DCUT11) showed
similar relationships with four environmental variables (Tar River discharge, rainfall,
wind, and water depth) as the control creeks and pre -Mod Alt L periods. Thus, there is
little evidence to suggest that Mod Alt L activities have impacted creek salinity;
correlations showed that salinity was most associated with Tar River discharge.
• Wetland water levels (measured by In -Situ Level TROLLs) at many locations
document the influence in the short-term of large rain events, or several smaller events
in a short amount of time; however, some do not appear to be influenced in the long-
term and their hydroperiods do not always respond logically to rainfall. These
rainfall/hydroperiod variations make it difficult to directly correlate mine activities to a
decrease in wetland hydroperiod in all cases. Aside from rainfall, hydrology at some
wells can also be influenced by large Tar River discharges and by wind tides, which
further complicates interpretation. With addition of 2020 data in the wetland hydrology
analysis, both prongs at Huddles Cut had significantly longer hydroperiods in post -Mod
Alt L years when all well data were combined; Jacks, Jacobs, Tooley, and Porter
creeks also had longer mean combined hydroperiods post -Mod Alt L but not
significantly. On an individual basis, one well in Drinkwater Creek had a significantly
shorter mean hydroperiod post -Mod Alt L, one well in Tooley Creek recorded
significantly longer mean hydroperiod post -Mod Alt L, and one well in Huddles Cut also
showed significantly longer wetland mean hydroperiods post -Mod Alt L. In each creek
affected by Mod Alt L, some wells had longer wetland hydroperiods post -Mod Alt L
while others had shorter ones. At all creeks, annual rainfall in 2020 was the highest
total for all creeks except Tooley and Huddles Cut since monitoring began; however,
annual rainfall post -Mod Alt L was only significantly higher in Porter Creek and
DCUT11. It is difficult to conclude if changes in hydrology at study creeks were due to
Mod -Alt L impacts, after-effects from Hurricane Irene, concomitant regional changes in
salinity and/or climate, or a combination. It does not appear mine activities have
altered the hydrology for Jacks Creek, Jacobs Creek, Drinkwater Creek, Tooley Creek,
Porter Creek, or DCUT11.
In the 2016 report, hydrology data from the Level TROLLs in Huddles Cut and Tooley
Creek were analyzed with regression models of before, during, and after pump
operation in deep wells around the mine perimeter (the deep well pumps around Jacks
Creek were still in operation so it was excluded). The analysis showed there was no
major impact on the response of the Huddles Cut and Tooley Creek Level TROLLs to
pump operation; regression models were similar but not identical. At least a one -day
lag in response to rainfall was noted in the post -pump operation period in Huddles Cut
and Tooley Creek. The analysis of Huddles Cut was extended to include 2017 and
results showed that the one -day lag response to rainfall remained, although shortened
for some wells whose original lag response was more than one -day. The same
analysis was performed for the 2017 report with Jacks Creek hydrology data and the

deep well pump operation times (only three deep wells were used in the Jacks Creek
analysis compared to 12 for Huddles Cut). The 2017 results were similar to Huddles
Cut as the lag number of days for rainfall to become an important predictor changed
from 0 day to 1 day for the periods During and Post pump operation, for most of the
Jacks Creek Level TROLLs. The analysis was run again for the 2018 report for
Huddles Cut and Jacks Creek (six deep wells were used for Jacks in the 2018
analysis) in addition to a new analysis of three Level TROLLs in upper Porter Creek.
Each analysis has shown that the pump operation in the deep water wells had not
changed the positive impact of rainfall on hydrology for nearly every Level TROLL in all
creeks (with the exception of JWSB, HMW9, and HWW7). For most of the Level
TROLLs in Huddles Cut the lag number of days for rainfall to become an important
predictor changed from 0 day to 1 day for the periods During and Post pump
operation. In Jacks Creek the rainfall lags were similar to the results in Huddles Cut in
the 2017 report; however, with the addition of the 2018 data and adjusted During/Post
periods, Jacks Creek important predictor variables were similar among Pre, During,
and Post periods. For the 2019 report, the deep well pump operation analysis was
done for six shallow wells in Porter Creek and three deep wells that were on and then
off for suitable periods to use for analysis. The pump operation in those deep wells did
not seem to have had any major impacts on the hydrologic response to changes to the
external environment (as measured by the Level TROLLs). Regression models
constructed before, during, and after pump operation in the deep wells were similar
(but not identical). The 2020 deep well pump analysis compared six shallow wells in
Porter Creek and five deep wells for on and off periods during corresponding Pre,
During, and Post deep well operation. The regression models showed strong positive
results for the rainfall coefficient which means a hydrologic response is most likely
affected by rainfall.
• Additional years of water quality data may yield different findings; however, the water
quality data analyzed by Dr. David Kimmel of ECU for the 2013 report did include all
years for all water quality monitoring stations and no deleterious effects related to the
mine were detected. For the 2014 report, all years of data were used again but
grouped in two data sets for spatial and temporal analysis and comparison. The first
set included all years between 1999-2011 for Jacks Creek, Tooley Creek, and Huddles
Cut. The second set included all years from 2012 through the current year for all
creeks. For this 2020 report, and presumably going forward unless directed otherwise,
only the second data set will be used for analysis as the information in the first data set
would never change or produce a different result. The 2012 - 2020 analyses and
comparison indicated that variability in water quality was typical of estuarine creeks
and followed a distinct seasonal pattern.
Water quality in creeks impacted by Mod Alt L followed a temporal trend that was also
influenced by the spatial location of the water quality monitoring station. Creeks with
the most recent impact (Porter Creek and DCUT11) showed higher variability in water
quality parameters than before, but due to their proximity to the Pamlico River estuary
this variability was less pronounced than the previously impacted creeks further from
the river. On a temporal basis, most stations responded to biological (i.e., chlorophyll
a and DO) and non -conservative biochemical concentrations (e.g., nutrient
concentrations) during the warmer months when the phytoplankton community is most
active and in contrast, during colder months responded mostly to conservative
elements (e.g., salt and pH). On a spatial basis, cluster analysis showed most creeks
group together on geomorphic similarities which influence characteristics of the water

(e.g., distance from river, upstream stations vs downstream stations, and/or water
depths). Creeks that have been impacted further in the past combined both post -Mod
Alt-L conditions and weather, and water quality appeared to present a persistent trend
in terms of intra-annual variability. This indicated that stabilization of the water quality
parameters is likely to continue. While most all post -Mod Alt L creeks had statistically
significant differences with at least several of the 17 water quality parameters pre -Mod
Alt L values, the majority of water quality changes pre- and post -Mod Alt L were not
ecologically significant, as no changes in ecosystem structure or function were
detected. Continued monitoring will determine if these changes persist; however
analysis to date suggests that water quality conditions become more consistent over
time.
• For sediment and water column metals, no obvious trends have been found to indicate
that mine continuation has changed either sediment or water column metal
concentrations relative to changes observed in control creeks or relative to changes
observed pre -Mod Alt L. Metal concentrations in some creeks have had a statistically
significant increase (e.g., Ag, Cd, Mo, and Se) or decrease (e.g., Al, Cr, and Cu) post -
Mod Alt L; however, changes in laboratory methods and laboratory equipment
contribute to most of the differences when results are compared to the earlier studies.
With the exception of one anomalous measurement of Zn in Jacks Creek in 2015,
concentrations of sediment metals or concentrations of water column metals in the
studied creeks are not likely to be associated with detectable biological effects.
• There have been some changes to canopy cover and vegetation in monitored plots in
portions of some study creeks and variability to wetland hydroperiods at some
locations. Most of the dominant species surveyed in 2020 were dominants in previous
years. There was two new dominants in Jacks Creek transects (JW2 and JW9) and
one each in Porter Creek and DCUT11: the one in DCUT11 was a non -wetland
species. When pre- and post -Mod Alt L percentages of brackish intolerant dominants
were compared for creeks overall, Huddles Cut and Jacks Creek decreased. Through
a SIMPROF, based on the presence/absence of species, temporally closer years were
more similar to one another than less recent years for all creeks, including control
creeks. For the creeks with many years of data (Jacks Creek, Tooley Creek, and
Huddles Cut), some of the differences seen between pre- and post -Mod Alt L may be
related to changes in the ecosystem over the years due to natural processes,
disturbances from hurricanes, and sea level rise.
However, no changes detected over the course of the study can be confidently
connected to mine activities, with the exception of the upper end of the west prong of
Huddles Cut. Part of one of the most upstream vegetation transects on the west prong
underwent physical changes in 2012 driven by surface depressions related to an
unstable near -surface discontinuous lithologic unit known as the Croatan Clay. These
depressions seemed to stem from proximity to the mine perimeter canal which caused
dewatering and/or piping in this unit. This is the only known occurrence and review of
previous boring records from the South Creek side of Mod Alt L found no other
locations of concern regarding the Croatan Clay. As of the end of 2016, the mine had
progressed around all of the study creeks along South Creek and no other surface
depressions were noted. Wetland hydroperiods in two nearby wells in Huddles Cut
west prong seem to have been affected by either proximity to the dewatering of the
Croatan Clay, or a reduction in drainage basin, or a combination of both factors since
their wetland hydroperiods were decreasing before the depressions formed; however,

since 2016 all wells recorded hydroperiods for the entire growing season.
• No change in fish forage base or assemblages of managed species due to mine
activity were apparent. Multivariate cluster analysis of managed fish for all creeks and
all collection years revealed some differences based on gear type (fyke net vs trawl)
and also separated some pre -Mod Alt L and post -Mod Alt L years within clusters;
however, the multivariate cluster analysis did not reveal distinct changes in managed
fish assemblages due to mine activities within the drainage basins of Jacks Creek,
Jacobs Creek, Drinkwater Creek, Tooley Creek, Huddles Cut, Porter Creek, or
DCUT11. The fish guild dendrograms also show no clear trend among the pre- and
post -Mod Alt L fish assemblages that could indicate potential effects from mine
activities; only one cluster of the 16 consisted of solely post -Mod Alt L data. As with
richness and abundance data, most other post -Mod Alt L guild creek -years were
distributed into clusters which also contained a corresponding pre -Mod Alt L year for
the same creek or for a control creek. Comparison of interannual variability detected
spatial differences of statistical significance between pre- and post -Mod Alt L years of
managed fish in Jacks and Tooley creeks; however, the differences were likely
affected by regional low CPUE for all species for 2016-2017, which are two of the post -
Mod Alt L years for these two creeks.
• Refinements and modifications to taxonomic information often change over time based
on additional information collected about particular species or professional
consensus/agreement; in addition there are times when species are "lumped together"
and then later "split". With the PCS creeks benthic data, numerous name changes
have occurred over the years (e.g., Hobsonia florida became Amphicteus floridus,
Tubificidae became Naididae), often between one year's report and the next.
Identification guides published since the early years of the 22-year study also provide
confidence that an Apochorophium unidentified to species in a given earlier year was
either A. lacustre or A. louisianum, species which are now grouped in the multivariate
database and treated as one, A. lacustre/louisianum; for taxa richness and calculation
of EBI, they are separated. Every year, the creeks benthic database was updated to
reflect the taxonomy for each taxon collected that year; but not until multivariate
analysis required the entire benthic database across all years for the 2013 report was
the entire database updated taxonomically for the dendrograms produced in each
subsequent report.
For five of the seven creeks whose basins have been reduced by Mod Alt L activities,
comparison of pre- and post -Mod Alt L interannual variability at upstream and
downstream benthic stations of both sweep and ponar collections within each creek
detected spatial differences of statistical significance between pre- and post —Mod Alt
L macroinvertebrate communities (Sweeps- upstream and downstream Jacobs and
Tooley creeks, upstream Huddles Cut, and downstream Porter Creek. Ponar-
downstream Tooley Creek). This result suggests that close to 75 percent of the 2020
comparisons (21 out of 28) in the impacted creeks showed the benthic
macroinvertebrate community of post -Mod Alt L years as similar to that of the pre -Mod
Alt L years; thus, it is difficult to discern any mine -related patterns in benthic
macroinvertebrate communities. In 2020, benthic abundance and/or taxa count for 17
out of 40 creek sites were record lows for both ponars and/or sweeps. Since these
record lows were in both impact and control creeks, it's unlikely these changes in
benthic abundance and taxa count were due to mine -related activities. The sweeps
and ponar richness and abundance dendrograms show no clear trend among the pre-
xxxiv

and post -Mod Alt L macrobenthic data that could indicate potential effects from mine
activities and, importantly, no clusters consisted of solely post -Mod Alt L data. Except
for Huddles Cut, sweep and ponar years/locations richness and abundance are
distributed into clusters represented by similar years/locations for the control creeks
and/or other pre -Mod Alt L creek -years. For Huddles Cut, the analysis showed that
most pre- and post -Mod Alt L years commonly clustered together, although usually
with no other creek. The clusters continue to point to the uniqueness of Huddles Cut
compared to other creeks. The macrobenthic guild dendrograms also show no clear
trend among the pre- and post -Mod Alt L data that could indicate potential effects from
mine activities and no cluster consisted solely of post -Mod Alt L data this year. As with
richness and abundance data, most other post -Mod Alt L years for the guilds were
distributed into clusters which also contained a pre -Mod Alt L year for the same creek
or a control creek. Scraper was a predominant trophic guild among most of the creek
sites (e.g., Tooley upstream, DCUT 11 upstream, PA2 downstream, and DCUT19
upstream), while shredder and filter feeder trophic guild was predominant in Porter
Creek downstream. Common scraper or filter -feeder trophic guilds were Littordinops
tenuipes (scraper) and Mediomastus sp. (collector filter -feeder, as well as collector -
gatherer). The mixed model ANOVA on guild composition showed that changes in the
benthic communities of four impact creeks post -Mod Alt L did significantly differ than
the changes in benthic communities of their respective control creeks during the same
time periods. In Jacks Creek upstream there were less herbivore and detritivore
trophic level post -Mod Alt L when compared to Muddy Creek (control). The opposite
trend was found in downstream Drinkwater Creek, a nearby impact creek, when
compared to PA2 (control), with herbivore and detritivore greater post -Mod Alt L in
Drinkwater Creek. However, downstream Tooley Creek post -Mod Alt L, another
nearby impact creek, had less herbivore and detritivore trophic level compared to two
control creeks (Muddy and Long). In Porter Creek upstream post -Mod Alt, shredder,
scraper, and grazer were less compared to Duck Creek (control), but slightly greater
when compared to Little Creek (control).
xxxv