Summary

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 1
 of
 51

Ecological
 Flows
 Science
 Advisory
 Board
 (EFSAB)

Meeting
 Summary

 April
 16,
 2013

Stan
 Adams
 Training
 Facility,
 Jordan
 Lake,
 Chapel
 Hill,
 NC

 X

 Approved
 for
 Distribution
 May
 14,
 2013

Attendance

Members

Hugh
 Barwick,
 Duke
 Energy

Tom
 Cuffney,
 US
 Geological
 Survey

Linda
 Diebolt,
 NC
 League
 of
 Municipalities

Chris
 Goudreau,
 NC
 Wildlife
 Resources
 Commission

Jeff
 Hinshaw,
 North
 Carolina
 State
 University

Sam
 Pearsall,
 Environmental
 Defense
 Fund

Judy
 Ratcliffe,
 NC
 Natural
 Heritage
 Program

Jaime
 Robinson,
 NCAWWA-­‐WEA

Bill
 Swartley,
 NC
 Division
 of
 Forest
 Resources

Fred
 Tarver,
 NC
 Division
 of
 Water
 Resources

Division
 of
 Water
 Resources

Harold
 Brady

Alternates

Rebecca
 Benner,
 The
 Natural
 Conservancy

Kevin
 Hart,
 Division
 of
 Marine
 Fisheries

Ian
 McMillan,
 NC
 Division
 of
 Water
 Resources

Sarah
 McRae,
 US
 Fish
 &
 Wildlife

Vann
 Stancil,
 NC
 Wildlife
 Resources
 Commission

Tom
 Thompson,
 Duke
 Energy

David
 Williams,
 NC
 Division
 of
 SWC

Guests:

Mary
 Davis,
 SARP

Phillip
 Jones,
 RTI

Jim
 Mead,
 Environmental
 Defense
 Fund

Kimberly
 Meitzen,
 The
 Nature
 Conservancy

Jennifer
 Phelan,
 RTI

Haywood
 Phistic,
 LNBA(online)

Marion
 Hopkins
 (online),
 EPA
 Region
 4

Kyle
 Hall
 (online)

NCSU
 Facilitation
 Team

Mary
 Lou
 Addor,
 Natural
 Resource
 Leadership

Institute
 (NRLI)

Nancy
 Sharpless
 (NRLI)

The
 purpose
 of
 the
 Ecological
 Flows
 Science
 Advisory
 Board:
 The
 Ecological
 Flows
 Science
 Advisory
 Board
 (EFSAB)
 will

advise
 NC
 Department
 Environment
 and
 Natural
 Resources
 (NCDENR)
 on
 an
 approach
 to
 characterize
 the
 aquatic
 ecology

of
 different
 river
 basins
 and
 methods
 to
 determine
 the
 flows
 needed
 to
 maintain
 ecological
 integrity.

Presentations,
 reports,
 and
 background
 information
 of
 the
 EFSAB
 are
 available
 at:
 www.nc-­‐water.org/sab

Webinar:
 If
 you
 cannot
 attend
 the
 meeting
 in
 person
 but
 would
 like
 to
 join
 us
 via
 the
 webinar,
 you
 can
 watch
 the
 presentations

and
 listen
 to
 the
 live
 streaming
 audio
 of
 the
 meeting
 by
 going
 to
 https://denr.ncgovconnect.com/sab/
 and
 typing
 your
 name
 in

the
 space
 labeled
 "guest."

NOTE:
 The
 EFSAB
 will
 meet
 May
 14,
 2013,
 9:00am
 until
 4:15pm
 at
 the
 Stan
 Adams

Training
 Facility,
 Jordan
 Lake
 Educational
 State
 Forest
 Center

 Chapel
 Hill,
 NC

(see
 page
 ?
 for
 meeting
 agenda
 topics
 and
 directions
 to
 location).

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013

Meeting
 Summary

Page
 2
 of
 51

April
 16,
 2013:
 Summary
 of
 Decisions/Recommendations
 and
 Proposed
 Actions

Decisions
 and
 Recommendations

1. A
 subcommittee
 is
 set
 up
 to
 organize
 and
 rewrite
 the
 recommendations
 generated
 from
 the

March
 2013
 meeting.
 The
 group
 will
 meet
 May
 1,
 2013.
 An
 invitation
 will
 be
 sent
 to
 those

EFSAB
 members
 who
 were
 not
 in
 attendance
 to
 learn
 about
 the
 subcommittee
 established.
 A

draft
 of
 the
 revisions
 will
 be
 circulated
 before
 the
 May
 14
 EFSAB
 meeting.

Proposed
 Actions

1. Review
 and
 sort
 the
 recommendations
 generated
 March
 2013,
 from
 opinions,
 process
 for

recommendations,
 and
 reword
 partial
 recommendations
 into
 recommendation
 language.

2. A
 range
 of
 acceptable
 deviations
 from
 unaltered
 flow
 would
 be
 used
 if
 the
 sustainable

boundaries
 approach
 were
 used.

3. Seeing
 results
 of
 this
 approach
 on
 some
 smaller
 streams,
 perhaps
 ones
 that
 also
 have
 PHABSIM

sites.

4. Run
 PHABSIM
 in-­‐stream
 flow
 studies
 on
 mountain
 sites.

Table
 of
 Contents

I. Executive
 Summary……………………………………………………………………………………………………….......2

II. Meeting
 Orientation
 and
 March
 19,
 2013
 Meeting
 Summary
 Approval…
 …………………….......7

III. Preliminary
 Results
 of
 the
 BEC—Fish
 Classification,
 Methods
 of
 Optimizing
 Fish
 &

Macro-­‐invertebrate
 Classes……………………………………………………………………………………………………...9

IV. SARP
 River
 Classification
 Framework…………………………………………………………………………………19

V. Results
 of
 Recommendations
 Framework………………………………………………………………………….28

VI. Introduction
 of
 Trial
 Balloons…………………………………………………………………………………………….30

VII.
 Next
 Steps
 and
 Agenda
 for
 May
 14,
 2013……………………………………………………………………………….50

 VIII.
 Information
 on
 the
 Next
 Meeting………………………………………………………………………………………….50

I.
 Executive
 Summary

TITLE:
 Preliminary
 Results
 of
 the
 Biological
 -­‐
 Environmental

 Classification
 (BEC)
 system—Fish

Classification
 and
 Methods
 fo
 Ooptimizing
 Fish
 and
 Macro-­‐intvertebrate
 Classes

Presenters:

 Jennifer
 Phelan,
 Phillip
 Jones,
 RTI;
 Tom
 Cuffney,
 USGS

Jennifer
 introduced
 the
 presentations
 with
 an
 overview
 of
 events
 leading
 to
 the
 BEC
 project,
 and
 what

the
 BEC
 project
 would
 be
 including.

 She
 then
 introduced
 Tom
 Cuffney
 to
 present
 preliminary
 results

on
 developing
 a
 stream
 classification
 system
 based
 on
 geographical
 assemblages
 of
 benthos
 and

associated
 environmental
 (physiographic
 and
 hydrologic)
 attributes.

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 3
 of
 51

Tom
 explained
 the
 statistical
 methods
 that
 he
 tried.

 He
 used
 6
 environmental
 variables
 in
 his
 analyses:

Cumulative
 drainage
 area,
 Sinuosity,
 Precipitation,
 %
 Sand
 in
 soil,
 Elevation,
 and
 NHD
 slope.

 He
 did
 a

cluster
 analysis
 of
 environmental
 variables.

 Results
 showed
 weak
 structure
 in
 the
 data
 (edges
 of

cluster
 are
 fuzzy,
 not
 breaking
 distinctly).

 Using
 existing
 classifications
 (he
 called
 a
 priori)
 and

environmental
 variables
 he
 found
 that
 the
 best
 classifications
 were
 Ecoregion
 3
 and
 Fenneman’s.

 Then

he
 described
 the
 classification
 effort
 using
 invertebrates,
 which
 he
 found
 that
 the
 clusters
 developed

with
 environmental
 data
 were
 not
 good
 at
 explaining
 patterns
 in
 the
 biological
 data
 (using
 partitioning

around
 medoids-­‐clusters
 PAM)).

 Using
 the
 invertebrate
 clusters
 to
 identify
 clusters
 in
 environmental

data
 did
 no
 better.

Next
 steps
 for
 invertebrate
 analyses

• Derive
 invertebrate
 metrics
 (aggregations
 of
 species
 attributes)
 with
 emphasis
 on
 those

sensitive
 to
 flow
 (e.g.,
 filter-­‐feeders,
 collector-­‐gatherers).

• Directly
 related
 invertebrate
 metrics
 to
 environmental
 variables
 (classification
 and
 regression

tree
 (CART))
 to
 develop
 integrated
 classifications

• Relate
 invertebrate
 metrics
 to
 flow
 variables

• Repeat
 analyses
 using
 metrics

 derived
 from
 species
 level
 as
 opposed
 to
 genus

Phillip
 Jones
 then
 presented
 the
 preliminary
 results
 for
 fish
 data.

 He
 used
 similar
 methods
 to
 Tom.

 He

first
 looked
 at
 species
 data
 in
 terms
 of
 environmental
 classification.

 Ecological
 drainage
 unit,
 was
 the

best
 one,
 and
 was
 comparable
 to
 Omernick
 level
 IV.

 The
 results
 for
 PAM
 analysis
 clustering
 were
 not

good.

 Hierarchical
 clustering
 provided
 better
 results.

Their
 recommendations

• Correspondence
 between
 independently
 derived
 environmental
 and
 biological
 classification
 is

weak

• Most
 promising
 approach
 is
 a
 classification
 system
 based
 on
 integrated
 biological
 and

environmental
 attributes
 (e.g.,
 CART
 univariate
 analysis)

• Need
 to
 adjust/optimize
 taxonomic
 resolution
 and
 environmental
 spatial
 scale

• Consider
 the
 purpose
 of
 a
 classification
 system…are
 the
 number
 of
 classes
 workable?

• Use
 an
 existing
 classification
 scheme?
 (one
 of
 these
 existing
 ecoregions,
 etc)

Major
 Discussion
 items/concerns/questions:

Some
 comments
 included:

• Instead
 of
 doing
 clustering
 and
 moving
 on
 to
 flows,
 now
 we’re
 doing
 back
 and
 forth
 between

the
 two
 to
 get
 the
 clusters
 related
 to
 flow
 variables,
 so
 it
 gets
 a
 little
 more
 complicated

• In
 the
 end
 it
 will
 be
 a
 fairly
 simple
 set
 of
 classes,
 like
 2-­‐3

• Its
 important
 to
 capture
 in
 the
 report
 that
 the
 results
 show
 a
 lack
 of
 ability
 to
 have
 very
 defined

categories,
 so
 other
 people
 don’t
 have
 to
 go
 back
 and
 do
 this

• There
 is
 some
 chance
 we’ll
 decide
 we
 can’t
 classify
 rivers
 and
 streams
 in
 NC
 in
 a
 way
 that

provides
 clear
 direction
 on
 how
 to
 set
 ecological
 baseline
 (which
 is
 an
 accomplishment)

• How
 to
 set
 an
 ecological
 baseline
 without
 classification?

Decisions
 Made:

 None

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 4
 of
 51

Proposed
 Actions
 or
 Identified
 Decisions
 to
 be
 made:

 None:

TITLE:
 SARP:
 River
 Classification
 Framework

Presenter:

 Mary
 Davis

Mary
 Davis
 provided
 a
 presentation
 on
 the
 classification
 framework
 developed
 for
 the
 Southern

Atlantic
 Conservation
 Cooperative
 (SALCC).

 We
 are
 not
 proposing
 a
 classification,
 but
 providing
 a

classification
 framework.

 We’re
 giving
 attributes
 that
 can
 be
 used
 in
 any
 number
 of
 ways.

 They

looked
 at
 a
 number
 of
 variables
 to
 use
 for
 classification.

 They
 decided
 on
 size
 class
 and
 flow
 variability.

Size
 class
 is
 based
 on
 basin
 area
 and
 flow
 variability
 (based
 on
 median
 daily
 variability).

 They
 found

using
 these
 that
 it
 follows
 eco
 region.

 She
 provided
 examples
 of
 how
 classifications
 are
 used,
 with

Michigan
 and
 the
 Potomac
 watershed
 as
 examples.

Major
 Discussion
 items/concerns/questions:

Some
 comments
 included:

• Using
 physiographic
 region,
 size
 and
 variability
 gets
 down
 to
 as
 few
 variables
 as
 possible,
 which

may
 help
 us
 move
 forward.

• I
 think
 this
 work
 is
 pointing
 us
 in
 a
 similar
 direction-­‐
 we’re
 not
 that
 different
 from
 where
 other

folks
 have
 trod,
 it
 is
 all
 coming
 back
 to
 basics.

Decisions
 Made:
 None

Proposed
 Actions
 or
 Identified
 Decisions
 to
 be
 made:

 None:

TITLE:
 Review
 of
 March
 19,
 2013
 Recommendations
 List
 Developed
 by
 the
 EFSAB

The
 EFSAB
 reviewed
 the
 recommendations
 brainstormed
 at
 the
 March
 19
 meeting.
 As
 a
 result
 of
 this

review,
 a
 subcommittee
 was
 established
 to
 organize
 and
 rewrite
 the
 recommendations
 list.
 The
 list
 as

brainstormed
 was
 a
 composite
 of
 opinions,
 processes
 for
 getting
 at
 recommendations
 as
 well
 as
 partial

recommendations.
 The
 subcommittee
 is
 scheduled
 to
 meet
 May
 1
 and
 will
 circulate
 their
 discussions

to
 the
 larger
 EFSAB
 before
 the
 May
 14,
 2013
 meeting.

Decisions
 Made:

A
 subcommittee
 is
 set
 up
 to
 organize
 and
 rewrite
 the
 recommendations
 generated
 from
 the
 March

2013
 meeting.
 The
 group
 will
 meet
 May
 1,
 2013.
 An
 invitation
 will
 be
 sent
 to
 those
 EFSAB
 members

who
 were
 not
 in
 attendance.

Proposed
 Actions
 or
 Identified
 Decisions
 to
 be
 made:

Review
 and
 sort
 the
 recommendations
 generated
 from
 opinions,
 process
 for
 recommendations,
 and

reword
 partial
 recommendations
 into
 recommendation
 language.

TITLE:

 Combining
 the
 Concepts
 of
 Eco-­‐deficit
 and
 Sustainability
 Boundaries:

 A
 Trial
 Balloon

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 5
 of
 51

Presenter:
 Chris
 Goudreau

Chris
 Goudreau
 presented
 a
 trial
 balloon
 which
 proposes
 combining
 concepts
 from
 the
 papers
 of

Richter
 et
 al.
 (2011)
 and
 Vogel
 et
 al.
 (2007)
 to
 determine
 ecological
 flows.

 This
 approach
 could
 be
 used

with
 or
 without
 classification.

 Chris
 suggests
 applying
 sustainability
 boundaries
 to
 eco-­‐difference

curves.

 The
 EFSAB
 would
 need
 to
 decide
 what
 the
 acceptable
 deviation
 from
 unaltered
 flow
 would
 be,

but
 when
 the
 eco-­‐difference
 from
 a
 proposed
 withdrawal
 exceeds
 that
 deviation,
 that
 withdrawal

might
 compromise
 the
 ecological
 flow.

 Another
 way
 to
 do
 it,
 which
 would
 be
 more
 stringent,
 is
 to
 say

that
 if
 a
 certain
 part
 of
 the
 curve
 is
 outside
 of
 that
 x%
 band,
 then
 that
 would
 indicate
 that
 the

proposed
 withdrawal
 might
 compromise
 the
 ecological
 flow.

Major
 Discussion
 items/concerns/questions:

• Can
 you
 talk
 a
 little
 bit
 more
 about
 the
 importance
 of
 not
 looking
 at
 seasonality
 versus
 looking

at
 seasonality?

 Because
 one
 of
 the
 concerns
 I
 would
 add
 is
 that
 that
 critical
 period
 maybe
 in

September.

 R:

 Yes,
 the
 example
 I
 have
 up
 here
 is
 January,
 so
 I
 would
 think
 you
 would
 want
 to

do
 it,
 say,
 like
 on
 a
 monthly
 or
 at
 least
 on
 a
 seasonal
 basis,
 not
 on
 an
 annual
 basis.

• This
 is
 a
 very
 sophisticated
 strategy
 for
 combining
 a
 presumptive
 standard
 with
 the
 eco-­‐
difference
 concept,
 and
 I
 really
 like
 it.

 The
 problem
 is
 that
 we’ll
 have
 the
 same
 problem
 with

this
 as
 we
 would
 with
 any
 presumptive
 standard,
 and
 that
 is
 making
 the
 case
 that
 a

presumptive
 standard
 answers
 the
 legal
 question
 of
 how
 the
 eco-­‐system
 will
 react,
 respond.

• As
 part
 of
 the
 BEC,
 is
 this
 one
 of
 the
 metrics
 that
 RTI
 is
 looking
 at
 in
 terms
 of
 trying
 to
 correlate

biological
 condition
 to
 degree
 of
 hydrologic
 alteration
 information.

 I
 don’t
 know
 if
 eco-­‐change

is
 one
 that
 you
 (RTI)
 had
 looked
 at
 yet.
 R:

 We’ve
 calculated
 it
 but,
 yes
 we’re
 using
 it.

 We
 have

eco-­‐surplus,
 eco-­‐deficit,
 eco-­‐change,
 and
 the
 eco-­‐deficits
 and
 the
 eco-­‐surpluses
 are
 expressed

on
 an
 annual
 basis
 and
 on
 a
 seasonal
 basis
 for
 the
 four
 seasons
 that
 you
 used
 for
 PHABSIM.

• What
 if
 you
 relate
 biology
 and
 you
 find
 out
 that
 the
 biology
 itself
 is
 plus
 or
 minus
 30%?

• Can
 you
 show
 us
 some
 of
 the
 smaller
 stream
 systems
 or
 wonder
 what
 those
 implications
 might

be?

 How
 do
 we
 get
 to
 that
 point?

• Could
 RTI
 generate
 these
 graphs
 for
 a
 number
 of
 sites?

 R:
 They
 already
 have
 them.

• There
 are
 different
 sustainable
 boundaries
 for
 the
 different
 drainage
 basin
 sizes.

 So
 depending

on
 how
 much
 water
 is
 actually
 available,
 you’ll
 have
 different
 percentages
 around
 that

sustainable
 boundary
 that
 you
 could
 withdraw
 in
 order
 to
 stay
 within
 a
 range
 that
 would
 still
 in

theory
 protect
 your
 ecological
 integrity.

• The
 eco-­‐deficit
 analysis
 is
 part
 of
 the
 BEC
 analysis
 for
 ecological
 flows.

Decisions
 Made:
 None

Proposed
 Actions
 or
 Identified
 Decisions
 to
 be
 made:

 The
 range
 of
 acceptable
 deviation
 from
 unaltered
 flow
 to
 use
 if
 the
 sustainable
 boundaries

approach
 were
 used.

 Seeing
 results
 of
 this
 approach
 on
 some
 smaller
 streams,
 perhaps
 ones
 that
 also
 have
 PHABSIM

sites.

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 6
 of
 51

TITLE:
 A
 Proposed
 Process
 and
 Report
 Format

Presenter:
 Sam
 Pearsall

Sam
 proposed
 a
 trial
 balloon
 for
 process
 and
 a
 report
 format:

1. Classify
 the
 rivers
 and
 streams
 of
 NC.

 This
 classification
 should
 include,
 at
 the
 absolute

minimum,
 classes
 distinguishing
 hydrologically
 distinct
 physiographic
 regions.
2. For
 each
 class,
 determine
 which
 variables
 are
 the
 most
 important
 (characterize
 the
 class).

 Base

these
 on
 biological
 responses
 to
 variations
 in
 flow.

 Include
 all
 available
 IFIM
 /
 PHABSIM
 data.

3. For
 each
 class,
 determine
 what
 the
 ecological
 baseline
 is.

 This
 should
 be
 based
 on
 the
 most

sensitive
 biota
 for
 which
 we
 have
 adequate
 data.

 The
 ecological
 baseline
 should
 consist
 of:
a. A
 subdivision
 of
 the
 year
 into
 relevant
 "seasons."

 These
 will
 vary
 from
 class
 to
 class
b. A
 target
 flow
 for
 each
 season.

 This
 should
 be
 expressed
 as
 a
 %
 of
 inflow
 defined
 at
 the

catchment
 level;
 the
 most
 powerful
 family
 of
 strategies
 demonstrated
 by
 IFIM
 work.

c. A
 band
 of
 allowable
 variation,
 or
 alternatively,
 specification
 about
 frequency,

magnitude,
 and
 duration
 of
 excursions
 from
 this
 base
 flow
 that
 will
 not
 cumulatively

violate
 the
 definition
 of
 ecological
 integrity.

Major
 Discussion
 items/concerns/questions:

• You’ve
 got
 the
 question
 mark
 beside
 “Classify”
 but
 you
 don’t
 have
 one
 beside
 the
 flow
 and

biological
 relationships,
 and
 that
 seems
 to
 be
 as
 much
 in
 question
 as
 the
 first
 one
 does
 to
 me.

R:

 I
 think
 we’re
 definitely
 seeing
 biological
 responses
 to
 flows.

 Whether
 we’re
 able
 to
 sort

those
 into
 classes
 or
 whether
 we
 have
 to
 treat
 them
 individually,
 remains
 to
 be
 seen.

 But
 we

know
 and
 the
 data
 support
 the
 notion
 that
 organisms
 respond
 to
 flows.

• So,
 when
 are
 those
 results
 coming?

 R:

 As
 the
 result
 of
 an
 additional
 investment
 from
 state

agencies,
 RTI
 should
 be
 able
 to
 produce
 results
 by
 August.

Decisions
 made:
 None

Proposed
 Actions
 or
 Identified
 Decisions
 to
 be
 made:

 None

TITLE:
 DWR’s
 Trial
 Balloons

Presenter:
 Fred
 Tarver

DWR
 proposed
 3
 trial
 balloons:

1. Use
 80%
 of
 inflow
 as
 the
 flow
 by
 requirement,
 statewide
 for
 now.

2. Use
 40%
 of
 annual
 average
 flow,
 statewide
 for
 now.

3. Use
 monthly
 median,
 statewide
 for
 now.

Fred
 noted
 that
 these
 are
 all
 based
 on
 PHABSIM
 in-­‐stream
 flow
 studies,
 which
 is
 a
 widely
 accepted

approach,
 but
 it
 is
 based
 on
 habitat.

 You
 have
 to
 assume
 that
 if
 you
 create
 the
 habitat,
 the
 biota
 will

be
 there.

 He
 also
 noted
 that
 the
 charts
 he
 looked
 at
 are
 in
 the
 Piedmont,
 so
 he
 is
 proposing
 to
 run

PHABSIM
 sites
 in
 the
 mountains
 as
 well.

Using
 80%
 of
 in-­‐flow
 a
 certain
 number
 of
 guilds
 are
 below
 the
 20%
 loss
 of
 habitat.

 Using
 40%
 of
 annual

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 7
 of
 51

average
 and
 monthly
 median
 produce
 similar
 losses
 of
 habitat.

 Monthly
 median
 and
 80%
 of
 in-­‐flow

maintain
 variability
 in
 flows
 mimicking
 a
 natural
 hydrograph,
 although
 40%
 of
 annual
 average
 is
 more

of
 a
 flat
 line.

Major
 Discussion
 items/concerns/questions:

• If
 we’re
 going
 to
 use
 a
 flat
 80%
 of
 flow
 by
 on
 the
 basis
 of
 some
 geographic
 unit
 whether
 it’s
 a

basin
 of
 some
 order,
 it
 makes
 sense
 to
 me
 that
 that
 percent
 of
 flow-­‐by
 perhaps
 should
 vary

according
 to
 season.

 There
 may
 be
 times
 when
 80%
 may
 be
 inadequate.

• But
 isn’t
 that
 percentage,
 when
 you
 run
 it
 through
 the
 model,
 that’s
 a
 daily
 time
 step.

 Say
 it’s

20%
 of
 inflow,
 so
 if
 it’s
 20%
 of
 inflow,
 it
 doesn’t
 matter
 if
 it’s—it’s
 not
 a
 seasonal
 or
 monthly

thing,
 right?

 The
 percentage
 of
 flow-­‐by
 is
 done
 on
 a
 daily
 time
 step.

• It
 seems
 to
 me
 that
 the
 department
 has
 suggested
 the
 best
 member
 or
 a
 member
 of
 each
 of

the
 three
 large
 families
 of
 implementation
 strategies
 and
 in
 each
 case
 they’ve
 recommended

an
 annual
 approach.

 I’m
 thinking
 that
 a
 seasonal
 approach
 has
 merit.

• If
 you
 get
 more
 data
 and
 it
 says
 you
 need
 to
 do
 the
 Piedmont
 differently
 from
 Mountains,
 you

can
 potentially
 have
 different
 numbers
 for
 those
 two.

• If
 you
 can’t
 show
 that
 similar
 habitats
 respond
 to
 flow
 in
 the
 same
 way,
 you
 can’t
 talk
 about

statewide
 application
 using
 this
 method,
 in
 my
 mind.

• In
 either
 the
 Piedmont
 or
 the
 Mountains,
 what
 we’ve
 tried
 to
 do
 is
 run
 enough
 different
 habitat

suitability
 curves
 through
 the
 program
 so
 that
 we’ve
 covered
 anything
 that
 would
 be
 there.

Then
 going
 further,
 you
 use
 the
 most
 limiting
 species
 or
 guild
 and
 if
 we’ve
 got
 habitat
 for
 it,

everything
 else
 is
 okay.

• So
 a
 suggestion
 might
 be
 taking
 a
 flow
 by
 approach,
 but
 then
 having
 some
 sort
 of
 threshold

number.

• We
 need
 to
 have
 what
 constitutes
 an
 allowable
 excursion
 and
 how
 often
 can
 you
 have
 one?

• I
 feel
 like
 a
 monthly
 median
 is
 kind
 of
 a
 nice
 number.

 If
 you
 model,
 it
 looks
 like
 you’re
 cuttng
 it

off
 at

 some
 kind
 of
 flat
 line,
 but
 the
 reality
 is
 it
 would
 be
 highly
 variable,
 and
 the
 amount
 of

water
 that
 you
 are
 shooting
 for
 in
 September
 is
 going
 to
 be
 higher
 than
 the
 80%
 flow
 by
 would

be.

Decisions
 made:
 None

Proposed
 Actions
 or
 Identified
 Decisions
 to
 be
 made:

• Run
 PHABSIM
 in-­‐stream
 flow
 studies
 on
 mountain
 sites.

II.

 April
 16,
 2013
 -­‐

 Meeting
 Orientation
 and
 March
 19,
 2013
 -­‐

 Meeting
 Summary

Approval

Members
 and
 alternates
 of
 the
 Ecological
 Board
 Science
 Advisory
 Board
 introduced
 themselves
 and

their
 affiliations.

 Guests
 in
 attendance
 and
 the
 facilitation
 team
 also
 introduced
 themselves.

Everyone
 was
 reminded
 to
 sign-­‐in
 who
 attended
 the
 meeting.

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 8
 of
 51

A
 brief
 orientation
 was
 conducted
 of
 the
 meeting
 facilities
 (restrooms,
 concession)
 and
 available

technology
 (webinar).
 Members
 and
 alternates
 are
 encouraged
 to
 sit
 at
 the
 main
 meeting
 table
 and

guests
 at
 tables
 away
 from
 the
 main
 meeting
 spaces.
 During
 discussions
 of
 the
 members
 and

alternates,
 guests
 may
 comment
 once
 members
 and
 alternates
 have
 completed
 their
 comments
 and

questions.
 During
 small
 group
 work,
 guests
 can
 also
 participate
 in
 small
 group
 discussions
 but
 may

not
 dominate
 the
 time.
 Everyone
 is
 asked
 to
 ensure
 that
 space
 is
 created
 for
 others
 to
 engage.
 From

time
 to
 time,
 the
 facilitators
 will
 conduct
 a
 straw
 poll
 to
 determine
 the
 current
 level
 of
 support
 for
 an

idea
 or
 what
 additional
 information
 is
 needed,
 not
 necessarily
 for
 a
 final
 decision.

The
 EFSAB
 approved
 the
 March
 19,
 2013
 meeting
 summary
 with
 the
 change
 requested
 to
 the

following
 section
 to
 read
 as:

In
 Lieu
 of
 Classification
 Approaches

 -­‐edits
 were
 made
 to
 the
 following
 comment
 to
 ensure
 it

accurately
 conveyed
 the
 author’s
 intent.
 Item
 #20
 (April
 16,
 2013)

If
 we
 get
 to
 a
 specific
 point
 on
 a
 graph
 as
 a
 recommendation,
 we
 need

 to
 provide
 justification
 for

reaching
 that
 conclusion.

 For
 example,
 if
 the
 approach
 uses
 PHabsim,
 I
 would
 need
 to
 see
 validation

that
 the
 habitat
 modeling
 approach
 actually
 reflects
 what
 is
 going
 on
 [with
 the
 biota].

 That’s
 what
 I

need
 to
 get
 some
 comfort
 with
 recommending
 a
 discrete
 flow
 boundary.

The
 agenda
 for
 the
 meeting
 was
 introduced.
 The
 meeting
 objectives
 as
 presented
 included:

• Introduction
 of
 the
 Draft
 Framework
 for
 Recommendations—Results
 of
 the
 March
 Brainstorm

• Presentation:

 Preliminary
 Results
 of
 the
 BEC—Fish
 Classification
 and
 Methods
 for
 Optimizing

Fish
 &
 Macro-­‐invertebrate
 Classes

• Presentation:

 Preliminary
 Results
 of
 the
 BEC—Macro-­‐invertebrate
 Classification

• Presentation:

 SARP:

 River
 Classification
 Framework

• Debriefs
 of
 the
 Presentations

• Results
 of
 the
 Recommendations
 Brainstorm
 at
 the
 March
 19,
 2013
 Meeting

• Introduction
 of
 Trial
 Balloons

• Next
 Steps
 and
 Agenda
 for
 May
 14,
 2013
 Meeting

The
 process
 for
 discussing
 and
 seeking
 consensus
 on
 a
 proposal
 was
 presented
 in
 February
 as
 the

following:

1. Record
 each
 proposal
 separately
 for
 consideration
 and
 discussion

2. Check
 for
 understanding-­‐what
 questions
 do
 we
 have?
 do
 we
 all
 understand
 the
 proposal
 in
 the

same
 way?

3. Revise
 proposal
 as
 needed

4. List
 levels
 of
 support
 from
 charter

a. Level
 1:
 Endorsement
 (I
 like
 it)

b. Level
 2:
 Endorsement
 with
 a
 minor
 point
 of
 contention
 (basically
 I
 like
 it)

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 9
 of
 51

c. Level
 3:
 Agreement
 with
 reservations
 (I
 can
 live
 with
 it)

d. Level
 4:
 Stand
 Aside
 (I
 don’t
 like
 it
 but
 I
 don’t
 want
 to
 hold
 up
 the
 group)

e.Level
 5:
 Block
 (I
 cannot/will
 not
 support
 the
 recommendation,
 decision,
 or
 proposal)

5. Poll
 for
 level
 of
 support
 of
 each
 proposal
 –
 what
 is
 your
 level
 of
 support?

6. Record
 level
 of
 support

III.
 Preliminary
 Results
 of
 the
 BEC—Fish
 Classification,
 Methods
 of
 Optimizing
 Fish
 and

Macro-­‐invertebrate
 Classes

Presenters:

 Jennifer
 Phelan
 and
 Phillip
 Jones,
 RTI;
 Tom
 Cuffney,
 USGS

(Note:
 this
 presentation
 contained
 44
 slides,
 which
 are
 posted
 on
 the
 ncwater.org
 website.

 This

summary
 does
 not
 contain
 all
 of
 the
 slides,
 but
 a
 select
 few
 to
 help
 facilitate
 major
 points).

Jennifer
 Phelan

We’re
 providing
 an
 update
 of
 where
 we
 are
 on
 the
 BEC
 project.

 This
 is
 a
 project
 collaborative
 with

USGS
 and
 RTI.

 Phillip
 Jones
 and
 Lauren
 Patterson,
 myself
 and
 Tom
 Cuffney.

 Its
 producing
 interesting

though
 maybe
 not
 very
 clear
 results.

 We’ll
 present
 some
 preliminary
 findings.

 Also
 funded
 by
 DWR

and
 WRC.

So
 want
 to
 refresh
 memories
 about
 project
 and
 why
 we’re
 doing
 it.
 The
 biofidelity
 analysis-­‐
 purpose

was
 to
 link
 biology
 to
 stream
 classes.

 We
 found
 stream
 classifications
 systems
 based
 on
 flow
 metrics

(EFS
 and
 McManamay)
 could
 not
 be
 extrapolated
 beyond
 catchments
 with
 USGS
 gages

• 49%
 to
 64%
 match
 between
 classifications
 based
 on
 USGS
 gage
 versus
 WaterFALL
 modeled

hydrologic
 data

• ~
 270
 USGS
 gages
 in
 NC

• ~70,000
 NHD+
 catchments
 (that’s
 a
 challenge
 obviously)

• Streams
 class
 can
 change
 depending
 on
 period
 of
 record
 used
 to
 determine
 classes

Even
 if
 you
 used
 same
 data
 source,
 and
 changed
 period
 of
 record,
 you
 can
 change
 stream
 classes.

 That

is
 another
 problem
 of
 using
 flow
 based
 classification.

Se
 we
 concluded
 we
 need
 a
 classification
 system
 that

• Is
 not
 based
 on
 sensitive
 threshold
 values

• Is
 consistent
 and
 reproducible
 using
 USGS
 stream
 gage
 and
 modeled
 data

• Is
 easy
 to
 understand
 and
 implement

• Can
 be
 applied
 throughout
 state

• Captures
 the
 distribution
 of
 aquatic
 biota
 in
 North
 Carolina

Based
 on
 these
 needs
 we
 developed
 the
 Biological-­‐
 Environmental
 Classification
 project
 with
 3

objectives
 (which
 correspond
 to
 3
 steps
 of
 the
 project):

1. Develop
 a
 stream
 classification
 system
 based
 on
 geographical
 assemblages
 of
 aquatic
 biota
 (fish

and
 benthos)
 and
 associated
 environmental
 (physiographic
 and
 hydrologic)
 attributes
 –

Biological-­‐Environmental
 Classification
 (BEC)
 system

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 10
 of
 51

2. Determine
 flow–biology
 response
 relationships
 for
 each
 BEC
 class

3. Determine
 and
 link
 significant
 flow
 metrics
 (and
 associated
 flow–biology
 relationships)
 to
 each

BEC
 class
 to
 support
 determinations
 of
 ecological
 flow

Step
 1:
 Determine
 BEC
 classes
 based
 on
 aquatic
 biota
 assemblages
 and
 environmental
 characteristics

Tom
 and
 Phillip
 have
 been
 working
 on
 Step
 1,
 they
 will
 present
 preliminary
 results
 of
 where
 we
 are

today.

 I’ll
 hand
 it
 over
 to
 Tom
 to
 talk
 about
 where
 we
 are
 from
 the
 environmental
 and
 benthic

perspectives,
 then
 will
 hand
 off
 to
 Phillip
 about
 fish
 and
 recommendations
 about
 where
 we
 go
 from

here.

First
 thing
 is
 we
 looked
 at
 classification
 of
 environmental
 attributes
 and
 how
 those
 divide
 among
 the

states.

 We
 had
 about
 3500
 sites
 across
 the
 country,
 1700
 are
 invertebrate
 sites,
 860
 are
 fish
 sites,
 RTI

put
 in
 another
 900
 random
 points
 to
 ensure
 there
 would
 be
 good
 coverage
 across
 the
 state.

 As
 you

see
 from
 this
 map
 we
 do
 have
 good
 representation
 of
 environmental
 attributes
 across
 the
 state.

In
 terms
 of
 environmental
 factors
 that
 could
 be
 extracted
 by
 GIS
 there
 were
 17
 that
 we
 had
 to
 work

with.

1. NHD
 drainage
 area-­‐
 it
 divides
 basins
 into
 small
 sections
 and
 they
 accumulate
 together

2. Cumulative
 drainage
 area

3. NHD
 slope

4. Slope

5. Elevation

6. Minimum
 elevation

7. Relief
 (max-­‐min

elev)

8. %
 flat
 land
 (<1%

slope)

9. %
 flat
 low
 land

10. %
 flat
 uplands

11. Precipitation

12. Evapotranspiration

13. Precip-­‐Evapotranspiration

14. Temperature

15. Sinuosity

16. Aquifer
 permeability

17. %
 sand
 in
 soils

First
 thing
 we
 did
 was
 look
 at
 it
 in
 terms
 of
 correlation
 matrix.

 These
 are
 just
 the
 correlation
 variables,

the
 highlighted
 ones
 are
 strongly
 correlated
 with
 each
 other.

 One
 of
 the
 dominant
 features
 of
 course

is
 elevation
 which
 dictates
 a
 lot
 of
 the
 values
 for
 other
 parameters.

 In
 terms
 of
 doing
 the
 clustering
 we

wanted
 to
 pull
 out
 a
 subset
 of
 variables
 that
 are
 not
 correlated
 with
 each
 other.
 What
 we
 used
 was

absolute
 value
 of
 >

 .7
 and
 came
 up
 with
 6
 environmental
 variables
 that
 summarize
 this
 environmental

matrix.

• Cumulative
 drainage
 area

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 11
 of
 51

• Sinuosity

• Precipitation

• %
 Sand
 in
 soil

• Elevation

• NHD
 slope

What
 we
 also
 did
 was
 go
 back
 and
 relate
 the
 full
 environmental
 matrix
 to
 reduced
 matrix.
 This
 is
 what

is
 known
 as
 a
 RELATE
 Analysis.

 These
 are
 distributions
 of
 the
 permutations
 test,
 this
 is
 a
 permutations

test.
 This
 is
 the
 correlation
 to
 the
 actual
 data.

 Two
 things
 that
 came
 out
 of
 this
 is
 that
 there
 is
 no

overlap
 so
 these
 are
 significantly
 different
 as
 you
 would
 expect
 and
 also
 you
 have
 a
 pretty
 high

correlation.

 Bottom
 line
 is
 that
 the
 reduced
 matrix
 is
 pretty
 well
 representing
 the
 full
 matrix
 even

though
 we’re
 dealing
 with
 less
 than
 half
 the
 variables.

Cluster
 analysis:
 Environmental
 Variables
 involved:

• Partitioning
 around
 medoids
 (PAM)
 which
 is
 a
 form
 of
 K
 means
 clustering

• Standardized
 data
 (mean
 =
 0,
 sd
 =
 1)

• Euclidean
 distance

• Examined
 2-­‐60
 clusters-­‐
 want
 to
 use
 environmental
 clusters
 and
 relate
 them
 to
 biology

• Average
 silhouette
 width
 used
 to
 determine
 “best”
 clustering
 for
 environmental

• Box
 plots
 of
 variables
 in
 “best”
 clustering
 to
 get
 an
 idea
 of
 what
 is
 really
 different
 among
 these

clusters

This
 is
 output
 from
 PAM
 analysis-­‐
 average
 silhouette
 width,
 it
 tells
 us
 how
 distinct
 the
 clusters
 are
 from

another
 (slide
 12).

 In
 this
 case
 I
 truncated
 it
 to
 30.

 You
 can
 see
 the
 cluster
 values
 peaked
 here
 at
 .34

That
 comes
 out
 to
 being
 our
 best

number
 of
 clusters
 for
 these
 env

variables.

• 0.71-­‐1.00:

 Strong
 structure

• 0.51-­‐0.70:

 Reasonable

structure

• 0.26-­‐0.50:

 Weak
 structure

• <0.25:

 No
 structure

So
 that
 means
 this
 is
 weak
 structure

in
 these
 data.

 This
 isn’t
 unusual
 it

means
 the
 edges
 of
 cluster
 are
 fuzzy.

These
 data
 are
 not
 breaking

distinctly,
 they
 grade
 one
 into

another.

 Not
 unexpected.

 A
 couple

examples
 of
 the
 box
 plots.

 One
 of

the
 things,
 elevation
 which
 is
 driving
 everything
 in
 here
 one
 of
 the
 most
 important
 variables.

 Three
 of

the
 clusters
 (3,4,5)
 have
 high
 average
 or
 median
 elevation,
 and
 2
 that
 are
 intermediate
 and
 2
 that
 are

low.

 This
 is
 about
 the
 best
 differentiation
 that
 we
 had.

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 12
 of
 51

Then
 here
 is
 the
 NHD
 Drainage
 Area
 which

was
 one
 of
 the
 worst
 we
 had.

 Similar
 across

the
 board,
 what
 is
 distinctive
 among
 these

are
 these
 outliers.

You
 can
 go
 back
 to
 the
 data
 and
 look
 at

Characterization
 of
 Clusters,
 look
 at
 low,

medium
 high
 (slide
 15).

 We
 can
 break
 these

things
 out,
 put
 them
 on
 a
 map.

These
 are
 the
 7
 groupings
 here,
 truly
 there

is
 kind
 of
 a
 mountain
 (which
 divides
 into
 2
 groups),
 piedmont,
 coastal
 plain.

 What
 is
 strange
 is
 that
 this

class
 1,
 which
 kind
 of
 occurs
 everywhere.

 That
 doesn’t
 seem
 reasonable.

 I
 don’t
 have
 a
 whole
 lot
 of

confidence
 that
 this
 best
 clustering
 is
 very
 representative.

Q:

 what
 is
 1?

R:

 See
 Characterization
 of
 clusters,
 Cluster
 1

What
 you
 are
 seeing
 is
 a
 catch
 all
 that
 doesn’t
 fit

into
 anything
 else
 falls
 into
 that
 category.

 So

the
 other
 thing
 to
 point
 out
 are
 the
 number
 of
 a

priori
 classifications
 that
 already
 exist
 out
 there:

• U.S.
 EPA
 Omernik
 Ecoregions
 III
 and
 IV

• USFS
 Bailey
 Ecoregions:
 Provinces
 and

Sections

• Fenneman’s
 physiographic
 Provinces
 and

Sections

• USGS
 Wolock’s
 hydrologic
 landscape

regions

• Ecological
 Drainage
 Units

• We
 also
 broke
 into
 stream
 size
 drainage
 area
 in
 km2:

• X
 ≤
 10

• 10
 <
 X
 ≤
 100

• 100
 <
 X
 ≤
 500

• 500
 <
 X
 ≤
 1000

• X
 >
 1000

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 13
 of
 51

We
 have
 16
 a
 priori
 classifications
 we
 can
 look
 at
 as
 well
 as
 the
 classification
 that
 we
 derived.

 The

stream
 size
 is
 arbitrary-­‐
 was
 my
 break
 down
 by
 drainage
 area.

 Its’
 something
 we’d
 like
 to
 play
 with.

 If

anyone
 has
 feedback
 on
 better
 breakdowns

I’d
 like
 to
 hear
 it.

 I
 looked
 at
 an
 Analysis
 of

Similarity,
 which
 is
 a
 multivariate
 ANOVA,

working
 with
 ranked
 classes
 instead
 of

quantitative.

 What
 you
 are
 looking
 at
 is
 R

statistics-­‐
 the
 higher
 the
 R
 the
 higher
 the

difference,
 or
 the
 greater
 the

correspondence
 of
 the
 classes

 (slide
 18).

This
 is
 our
 best
 classification
 for
 the

environmental
 variables
 (PAM
 Cluster
 7?).

The
 ones
 that
 are
 pretty
 high
 are
 Ecoregion

III,
 Fenneman’s,

 Ecological
 drainage
 units

did
 not
 work
 very
 well.

 We
 have
 choices
 of
 a

priori
 classifications,
 PAM
 classifications.

Moving
 on
 to
 classification
 of
 invertebrates.

Classification
 based
 on
 invertebrate
 biota,
 presented
 by
 Tom
 Cuffney

We
 had
 1700
 sites
 to
 begin
 with,
 had
 to
 toss
 some
 out
 got
 about
 1044
 that
 we
 used
 in
 the
 analysis.

Looking
 at
 correspondence
 of
 invertebrates
 with
 a
 priori
 and
 the
 environmental
 clusters
 we
 derived.

Characteristics
 of
 invertebrate
 data
 used:

• Sites
 rated
 by
 DWQ
 as

o Excellent,
 Good,
 or
 Good-­‐Fair

o Standard
 qualitative
 or
 their
 swamp
 methods
 (these
 provide
 best
 representation
 of

entire
 community)

• Most
 recent
 data
 for
 each
 site

• Ordinal
 scale
 data
 (unlike
 the
 fish
 data.

 This
 puts
 some
 limitations
 on
 the
 analysis,
 and
 is
 why

we’re
 using
 ANOSIM)

o Absent
 <
 rare
 <
 common
 <
 abundant

o Coded
 as:
 0,
 1,
 3,
 and
 10

o ANOSIM
 (MANOVA
 for
 ranked
 data)

• Eliminated
 rare
 taxa:
 occur
 <
 5
 sites
 won’t
 influence
 data

• Lowest
 taxa
 level:
 Genus

 (this
 data
 is
 also
 used
 for
 a
 study
 on
 DE
 River,
 allowed
 us
 to
 reduce

down
 to
 490
 taxa)

• Ambiguous
 taxa
 resolved,
 taxa
 harmonized

These
 are
 the
 environmental
 variables,
 PAM
 clusters
 from
 2-­‐
 60.

 Those
 PAM
 clusters
 applied
 to
 the

invertebrate
 data
 and
 seeing
 how
 well
 it
 differentiates.

 What
 we
 find
 here
 is
 that
 it
 doesn’t

differentiate
 very
 well
 at
 all
 (slide
 22).

 So
 there
 is
 not
 a
 very
 good
 correspondence,
 in
 other
 words
 the

cluster
 we
 were
 developing
 with
 environmental
 data,
 are
 not
 good
 at
 explaining
 patterns
 in
 the

biological
 data.
 I
 highlighted
 the
 7th
 one
 here,
 this
 was
 our
 best
 environmental
 differentiation
 and
 its

not
 very
 good
 for
 invertebrates.

 If
 you
 look
 at
 the
 a
 priori
 classifications
 in
 reference
 to
 the

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 14
 of
 51

invertebrates
 rather
 than
 environmental,
 this
 is
 our
 PAM
 clustering
 did
 not
 work
 very
 well.

 There
 was

a
 small
 tendency
 to
 pick
 up
 with
 drainage
 area
 was
 added,
 which
 is
 different
 from
 the
 environmental.

All
 in
 all
 the
 environmental
 clustering
 is
 not
 as
 good
 as
 what
 we
 are
 getting
 with
 the
 a
 priori

classifications.

So
 then
 we
 looked
 at
 invertebrate
 clustering
 itself
 with
 multiple
 possible
 clustering
 methods

• K-­‐means:
 uses
 Euclidean
 distance
 and
 was
 not
 appropriate
 due
 to
 nature
 of
 the
 biological
 data

• PAM:
 Bray-­‐Curtis,
 very
 low
 silhouette
 values
 (low
 structure
 in
 the
 data)

• Hierarchical
 clustering
 (Bray-­‐Curtis):

o Agglomerative:
 you
 start
 with
 sites
 separate
 and
 then
 put
 them
 together,
 it
 had
 many

small
 clusters
 (problematic)

o Divisive
 hierarchical
 clustering:

 what
 are
 “best”
 clustering?

• Examined
 2-­‐60
 clusters

• ANOSIM
 to
 assess
 correspondence
 between
 clusters
 and
 invert
 data
 (similarity
 matrix)

Here
 is
 an
 example
 of
 the
 PAM
 clustering
 and
 why
 we
 dropped
 that
 (slide
 25).

 These
 are
 supposed
 to

peak
 at
 some
 point
 but
 instead
 drops
 down.

 There
 is
 something
 going
 on
 here
 with
 the
 PAM
 clustering

analysis
 and
 the
 introduction
 of
 Bray
 –Curtis
 similarity
 matrix.

 Something
 is
 not
 right,
 this
 says
 there
 is

no
 structure
 in
 the
 data
 but
 if
 you
 do
 a
 multidimensional
 scale
 you’ll
 see
 there
 is
 structure
 in
 the
 data.

We
 did
 take
 the
 divisive
 clustering,
 this
 is
 the
 invertebrate
 clustering
 now
 applied
 to
 the
 environmental

data,
 what
 we
 had
 hoped
 to
 see
 is
 that
 this
 peaked
 somewhere
 you
 and
 would
 get
 a
 peak
 that
 would

identify
 what
 number
 of
 clusters
 you
 should
 use
 to
 match.

 What
 we
 get
 instead
 is
 an
 asymptotic

relationship.
 It
 peaks
 at
 18
 being
 the
 best
 one.

 So
 using
 the
 invertebrate
 clusters
 to
 id
 clusters
 within

the
 environmental
 data
 is
 not
 better
 than
 the
 use
 of
 environmental
 clusters
 to
 identify
 clusters
 within

the
 invertebrate
 data.

 So
 what
 next?

Next
 thing
 is
 to
 move
 on
 and
 talk
 with
 Phillip
 about
 this
 approach
 and
 this
 is
 where
 we’re
 headed,
 to

more
 directly
 link
 biology
 with
 environmental
 data
 by
 going
 to
 CART
 (classification
 and
 regression
 tree)

analysis
 and
 regression
 trees
 so
 you
 can
 develop
 models
 to
 try
 to
 define
 clusters.

 So
 these
 sites
 here

(slide
 27)
 have
 <242.9
 meters,
 these
 are
 greater
 than,
 they
 continue
 to
 divide
 so
 you
 are
 directly

linking
 biology
 to
 the
 environmental.

 This
 has
 pluses
 and
 minuses,
 you
 may
 ask
 why
 we
 didn’t
 start

with
 this.

 This
 is
 a
 univariate
 method,
 which
 means
 you
 have
 to
 take
 the
 community
 data
 and
 reduce

it
 to
 a
 univariate
 value,
 there
 are
 hundreds
 of
 ways
 to
 do
 that.

 So
 now
 instead
 of
 looking
 at
 a
 couple

models
 we’re
 looking
 at
 many
 models
 in
 order
 to
 understand
 the
 data
 structure.

 It
 becomes
 more

complicated,
 though
 it
 is
 now
 directly
 relating
 biology
 to
 environmental
 data.

 If
 we
 look
 at
 this
 again,

the
 a
 priori
 clusters
 and
 then
 the
 ones
 we
 get
 from
 CART
 analysis,
 we’re
 getting
 things
 on
 a
 level
 of

what
 we
 got
 before,
 which
 you’d
 expect
 since
 these
 are
 derived
 form
 the
 data.

 In
 other
 words
 these

sites
 are
 based
 on
 the
 biology.

 You
 want
 to
 see
 these
 up
 in
 this
 range
 here
 (slide
 28).

 What’s

interesting
 is
 that
 a
 lot
 of
 the
 other
 a
 priori
 classes
 are
 at
 least
 as
 good
 as
 what
 we
 are
 getting
 out
 of

the
 CART
 analysis.

 If
 we
 look
 now
 at
 the
 number
 of
 classes,
 this
 is
 an
 important
 thing
 because
 the

whole
 effort
 is
 to
 come
 up
 with
 a
 number
 of
 classes,
 then
 use
 to
 categorize
 the
 state
 going
 into
 the

other
 things.

 So
 if
 you
 have
 200
 classes
 that
 is
 not
 a
 useful
 number.

 You
 see

 (Number
 classes
 slide)

that
 if
 you
 add
 in
 drainage
 area
 you
 get
 over
 70,
 a
 much
 better
 job
 here
 of
 controlling
 the
 number
 of

classes
 that
 will
 be
 statistically
 meaningful,
 with
 the
 CART
 analysis.

 The
 other
 thing
 to
 look
 at,
 when

you
 do
 these
 clustering
 and
 come
 up
 with
 overall
 metric
 to
 tell
 globally
 if
 you
 have
 statistical

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 15
 of
 51

significance,
 then
 you
 can
 look
 at

pairwise
 comparisons.

 You
 find

you
 have
 a
 lot
 of
 insignificant

pairwise
 comparison.

 When
 you

do
 CART
 analysis
 it
 means
 you

don’t
 have
 that.

 There
 are

advantages
 with
 going
 with
 CART

analysis.

Next
 steps
 for
 invertebrate

analyses

• Derive
 invertebrate

metrics
 (aggregations
 of

species
 attributes)
 with

emphasis
 on
 those

sensitive
 to
 flow
 (e.g.,

filter-­‐feeders,
 collector-­‐
gatherers).

 Be
 smart
 since
 there
 are
 many
 to
 choose
 from.

• Directly
 related
 invertebrate
 metrics
 to
 environmental
 variables
 (CART)
 to
 develop
 integrated

classifications

• Relate
 invertebrate
 metrics
 to
 flow
 variables:

o Flow
 surplus/deficit
 and
 IHA
 metrics
 (to
 communicate
 with
 other
 people
 doing
 this)

o CART
 analysis
 (does
 good
 job
 identify
 important
 flow
 variables)

o Analyses
 (e.g.,
 quartile
 regression)
 to
 look
 at
 flow
 relationships

 Within
 classes

 State-­‐wide

• Repeat
 analyses
 using
 metrics

 derived
 from
 species
 level
 as
 opposed
 to
 genus

Classification
 based
 on
 Fish,
 presented
 by
 Phillip
 Jones,
 RTI

I’ll
 step
 through
 analysis
 for
 fish
 data,
 very
 similar
 methods
 to
 what
 Tom
 presented,
 some
 differences

because
 data
 was
 different.

 Here
 is
 a
 map
 of
 the
 sample
 sites
 for
 fish.

 Main
 thing
 to
 point
 out
 that
 it
 is

a
 little
 lighter
 in
 the
 coastal
 plain.

Data
 Description
 and
 Formatting

• Took
 most
 recent
 sample
 at
 858
 unique
 XY
 coordinate
 locations

• Count
 data
 at
 species
 level

• Data
 was
 log
 transformed,
 because
 a
 number
 of
 species
 that
 have
 hundreds
 to
 thousands
 of

collections.

• Species
 observed
 at
 <5
 sites
 were
 removed

• Sample
 locations
 with
 no
 fish
 were
 removed

• Bray-­‐Curtis
 method
 used
 to
 calculate
 dissimilarity
 matrix

Analytical
 Approach

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 16
 of
 51

• Environmental
 Classifications

o Associate
 sample
 locations
 and
 community
 data
 with
 (what
 Tom
 called
 a
 priori)
 eco-­‐
region
 level,
 drainage
 class,
 and
 USGS-­‐derived
 environmental
 clusters

o Test
 explanatory
 power
 of
 each
 classification
 (PERMANOVA)-­‐
 similar
 to
 ANOVA
 but

doesn’t
 make
 same
 assumptions

• Biological
 Classification

o Use
 community
 data
 to
 create
 biology-­‐based
 groups
 with
 PAM
 and
 hierarchical

agglomerative
 techniques

o Test
 significance
 and
 explanatory
 power
 using
 a
 number
 of
 metrics
 (Silhouette
 width,

multi-­‐scale
 bootstrap
 re-­‐sampling,
 PERMANOVA)

First,
 this
 looks
 at
 species
 data
 in
 terms
 of
 environmental
 classification
 (slide
 36).

 This
 is
 assigning
 class

based
 on
 x-­‐axis,
 then
 looking
 at
 explanatory
 value.

 You
 can
 see
 they
 are
 low,
 best
 ones
 are
 close
 to
 .4.

The
 best
 one,
 ecological
 drainage
 unit,
 was
 based
 on
 fish
 evolutionary
 lineage,
 so
 its
 not
 too
 surprising

that’s
 the
 best
 for
 fish
 data.

 You
 can
 see
 its
 comparable
 to
 Omernick
 Level

 IV.

 USGS
 is
 on
 the
 right

about
 level
 with
 second
 tier
 (~.2
 value).

 Looked
 at
 some
 drainage
 schemes,
 they
 were
 consistently

very
 low,
 did
 not
 have
 much
 power.

Next
 step
 looked
 at
 just
 using
 biological
 data
 where
 there
 were
 groups
 that
 shook
 out,
 looking
 at

species
 information.

 Here
 are
 results
 for
 the
 PAM
 analysis,
 the
 results
 were
 not
 good
 (slide
 37).

 I
 used

it
 for
 raw
 data
 and
 transformed
 to
 matrix.

 In
 both
 cases,
 these
 values
 tell
 you
 is
 there
 is
 not
 a
 lot
 of

structure
 in
 the
 data
 that
 is
 identified
 by
 this
 approach.

 Similar
 to
 what
 Tom
 found
 with
 same

methodology.

Then
 did
 another
 approach
 called
 Hierarchical
 clustering,
 a
 nice
 feature
 is
 you
 can
 do
 a
 multi-­‐scale

bootstrapping
 that
 gives
 you
 a
 p
 value
 for
 the
 clusters.

 Otherwise
 this
 is
 subjective,
 its
 up
 to
 you
 to

define
 how
 many
 groups
 you
 want
 to
 see,
 you
 can
 put
 any
 number
 you
 want.

 So
 when
 doing

resampling,
 I
 found
 62
 clusters
 out
 of
 858
 possibilities
 that
 met
 standard
 significance
 test
 of
 alpha-­‐0.5.

You
 can
 see
 they
 are
 clustered

pretty
 far
 down
 on
 the
 bottom,
 so

pretty
 high
 resolution
 in
 terms
 of

data,
 I
 mean
 if
 you
 had
 2

groups…this
 is
 saying
 there
 are

some
 significant
 groups,
 but
 they

tend
 to
 be
 small,
 with
 high

numbers.

 This
 is
 an
 example
 of
 8

(slide
 39-­‐
 shown),
 the
 rectangles
 on

each
 of
 the
 cluster
 groups.

This
 is
 an
 example
 of
 what
 you
 find

to
 be
 statistically
 significant,
 to
 give

a
 sense
 of
 scale.

 One
 thing
 to
 get

across,
 even
 though
 there
 is
 a

difference
 between
 statistical

significance
 and
 ecological

significance,
 so
 even
 just
 using

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 17
 of
 51

coarse
 resolution
 number
 of
 8,

 I
 broke
 out
 mean
 and
 median
 environmental
 values
 for
 the
 8
 groups,

there
 is
 nothing
 there
 that
 seems
 unusual.

 You
 can
 see
 how
 elevation
 and
 slope
 would
 influence
 the

biology
 you’d
 find
 at
 these
 sites.

 There
 are
 ways
 you’d
 use
 this
 type
 of
 analysis
 to
 talk
 about
 fish

biology
 and
 get
 results
 that
 are
 reasonable.

 Just
 to
 show
 you
 geographic
 distribution
 of
 the
 groups,

this
 is
 a
 density
 plot
 of
 the
 8
 clusters,
 you
 can
 see
 it
 falls
 out
 how
 you’d
 expect,
 see
 Sandhills
 and

Coastal
 Plain,
 Piedmont,
 then
 type
 clusters
 in
 higher
 resolution.

 Even
 though
 resolution
 was
 quite

above
 statistically
 significant
 threshold,
 the
 results
 are
 pretty
 reasonable.

So
 I
 added
 the
 8
 group
 cluster
 analysis
 to
 the
 graph
 I
 showed
 earlier
 showing
 explanatory
 power.

 This

time
 the
 clusters
 showed
 better
 than
 a
 priori
 methods,
 also
 added
 the
 red
 bar-­‐
 how
 much
 explanatory

power
 the
 number
 of
 clusters
 give
 you.

That’s
 it
 for
 fish.

 There
 are
 a
 couple
 other
 things
 you
 can
 do
 with
 cluster
 analysis.

• Incorporate
 select
 environmental
 variables
 into
 biological
 clustering
 process
 (instead
 of
 using

just
 species
 data
 into
 matrix
 for
 clustering-­‐
 get
 better
 results
 and
 more
 explanatory
 power)

• Assess
 cluster
 p-­‐values
 in
 terms
 of
 centers
 and
 multivariate
 spread
 (its
 not
 hard
 to
 get

significant
 groups,
 but
 when
 you
 look
 at
 what
 groups
 are
 characterizing,
 you
 see
 centers
 of

groups
 are
 very
 close
 together.

 That’s
 typically
 a
 function
 of
 how
 sampling
 works
 and
 number

of
 samples
 per
 site.

 Good
 way
 to
 tease
 out
 if
 clusters
 are
 legitimate
 or
 contain
 the
 same
 info.)

• Classify
 ‘best’
 cluster
 results
 in
 terms
 of
 environmental
 variables;
 assess
 predictive
 power
 using

80/20
 training/test
 regime

 (take
 80%
 of
 data
 and
 test
 how
 well
 it
 does
 using
 the
 20%
 of
 data

you
 didn’t
 use.

 You
 can
 send
 it
 in
 loops
 and
 do
 it
 many
 times,
 figure
 out
 why
 they
 aren’t

characterizing
 very
 well)

Recommendations

• Correspondence
 between
 independently
 derived
 environmental
 and
 biological
 classification

is
 weak

• Most
 promising
 approach
 is
 a
 classification
 system
 based
 on
 integrated
 biological
 and

environmental
 attributes
 (e.g.,
 CART
 univariate
 analysis)

• Need
 to
 adjust/optimize
 taxonomic
 resolution
 and
 environmental
 spatial
 scale
 (for
 instance

with
 fish,
 in
 general
 the
 more
 clusters
 you
 put
 in
 the
 better
 it
 is
 resolved
 as
 you
 are
 dealing
 with

a
 high
 resolution
 level.

 Just
 3-­‐4
 regions
 may
 not
 capture
 that)

• Consider
 the
 purpose
 of
 a
 classification
 system…are
 the
 number
 of
 classes
 workable?

• Use
 an
 existing
 classification
 scheme?
 (one
 of
 these
 existing
 ecoregions,
 etc)

Questions
 (Q),
 Comments
 (C),
 Response
 (R):

Q:

 Why
 wouldn’t
 drainage,
 like
 Neuse
 vs
 Cape
 Fear
 vs
 Broad,
 be
 a
 fundamental
 characteristic
 that

would
 define
 clusters?

R:

 I
 was
 surprised
 as
 well
 that
 stream
 class
 was
 not
 useful
 in
 characterizing
 variability.

 Tom-­‐
 you
 found

something
 similar
 that
 stream
 size
 wasn’t
 useful?

R:

 If
 you
 add
 stream
 size
 for
 the
 invertebrates
 it
 improved
 (combined
 with
 a-­‐priori).

 Added
 another

column
 to
 ecoregion
 that
 included
 stream
 class
 or
 stream
 size.

C:

 Not
 interested
 in
 size-­‐
 basin.

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 18
 of
 51

R:

 was
 interested
 in
 looking
 at
 developing
 in
 the
 region.

 The
 problem
 is
 if
 there
 is
 not
 a
 consistent

relationship
 between
 drainage
 size
 across
 all
 the
 sites,
 you
 won’t
 see
 a
 relationship.
 So
 you
 can
 see
 a

small
 drainage
 that
 has
 been
 developed,
 a
 small
 drainage
 that
 hasn’t
 been,
 and
 you
 wouldn’t
 expect
 a

drainage
 covered
 by
 a
 Wal-­‐Mart
 parking
 lot
 to
 have
 same
 species
 assemblage
 as
 a
 small
 mountain

drainage.

 They
 will
 have
 very
 different
 biological
 assemblage,
 so
 there
 won’t
 be
 a
 relationship

between
 drainage
 size
 and
 species.

 One
 of
 hypotheses
 I
 had
 was
 looking
 at
 land
 use-­‐
 small
 drainage

developed
 against
 small
 drainage
 not
 developed.

C:

 number
 of
 those
 species
 would
 be
 pretty
 low,

C:

 A
 lot
 of
 species
 found
 in
 this
 analysis
 are
 in
 multiple
 basins,
 not
 like
 you’d
 have
 site
 fidelity
 with
 a
 lot

of
 species.

 Most
 occur
 all
 over
 mid-­‐Atlantic
 or
 Tennessee
 side,
 they
 won’t
 fall
 out.

C:

 There
 will
 be
 certain
 species
 that
 have
 different
 characteristics-­‐
 run
 of
 river
 fish
 vs
 those
 that
 spawn

in
 Neuse.

C:

 This
 is
 all
 wadeable
 streams,
 that’s
 a
 major
 component
 that
 may
 be
 an
 R
 factor.

R:

 One
 thing
 that
 surprised
 me
 is
 that
 you
 have
 fairly
 small
 drainage
 classes
 in
 coastal
 plain
 units
 that

were
 sampled.
 Whereas
 rivers
 like
 Roanoke
 and
 Neuse
 are
 not
 going
 to
 be
 sampled
 with
 this
 data
 set.

Q:

 You
 took
 most
 recent
 dates
 of
 sampling,
 did
 you
 go
 back
 and
 look
 at
 what
 years
 those
 were,
 and
 if

there
 were
 variability
 for
 years?

R:

 I
 have
 that
 information,
 I
 hoped
 to
 have
 enough
 sites
 to
 have
 repeated
 sampling,
 but
 we
 would

have
 lost
 a
 large
 number
 of
 sampling
 sites.

 I
 had
 most
 recent
 year
 but
 didn’t
 graphs
 that.

C:

 In
 study
 I
 did,
 I
 noticed
 there
 were
 years
 were
 abundance
 and
 frequency
 were
 higher,
 so
 that
 could

be
 a
 factor.

 It
 might
 be
 useful
 to
 look
 at
 spread.

R:
 It’s
 pretty
 wide.

 Some
 have
 been
 sampled
 3-­‐4
 times.

Q:

 Are
 you
 planning
 on
 doing
 anything
 with
 the
 non-­‐wadeable
 streams?

R:

 I’d
 love
 to
 but
 we
 don’t
 have
 a
 lot
 of
 data.

C:

 There’s
 data
 with
 Marine
 Fisheries,
 years
 of
 data.

R:

 It
 would
 be
 great
 to
 have
 that,
 have
 looked
 at
 that
 as
 part
 of
 relicensing,
 but
 if
 you
 have
 a
 source

you
 can
 provide
 let
 me
 know
 and
 we
 can
 look
 at
 it.

 The
 only
 issue
 that
 may
 come
 up
 is
 if
 there
 are

vastly
 different
 methodologies.

Q:

 Could
 you
 pull
 results
 from
 different
 sampling
 methods
 to
 come
 up
 with
 an
 index?

R:

 Depends,
 if
 you
 had
 consistent
 methods
 for
 large
 streams,
 you
 could
 have
 a
 large
 stream
 group

based
 on
 that
 data,
 and
 a
 wadeable
 stream
 group.

Q:

 Did
 you
 make
 adjustments
 for
 different
 sample
 methods
 for
 fish?

R:
 Wadeable
 streams
 are
 fairly
 consistent,
 so
 we
 didn’t
 do
 that.

 It
 would
 be
 an
 issue
 if
 you
 looked
 at

other
 collection
 methods
 of
 data.

Facilitator:

 we
 invite
 you
 to
 think
 about
 how
 this
 informs
 the
 work
 of
 the
 EFSAB.

 Thought
 to
 break
 it

down
 regarding
 your
 thoughts
 of
 what
 Tom
 presented,
 what
 excites
 you,
 what
 is
 missing
 for
 you,
 and

how
 it
 informs
 your
 work.

 We’ll
 do
 that
 for
 each
 presentation.

 What
 excites
 you?

C:

 It
 doesn’t
 seem
 to
 provide
 any
 real
 discriminatory
 power
 as
 analyzed.

 Take
 home
 message,
 we
 may

be
 plugging
 in
 the
 biological
 data,
 but
 we
 may
 not
 be
 using
 an
 appropriate
 metric,
 using
 genus
 for

invertebrates
 for
 example,
 as
 opposed
 to
 another
 characterization
 of
 benthos.

 Is
 that
 accurate?

Assumption
 is
 if
 you
 plug
 in
 taxonomic
 characteristics
 it
 would
 provide
 that
 discriminatory
 power,
 that

doesn’t
 seem
 to
 be
 the
 case.

R:

 Based
 on
 limited
 environmental
 data
 we
 have
 (pretty
 limited,
 16
 variables,
 down
 to
 6),
 the
 clusters

based
 on
 that
 really
 gives
 you
 now
 more
 ability
 to
 explain
 the
 structure
 of
 the
 invertebrate
 data
 than

the
 existing
 a
 priori
 classifications.

 If
 we
 want
 more
 ability
 to
 differentiate
 we
 need
 to
 go
 to
 different

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 19
 of
 51

methods
 like
 CART
 analysis.

Q:

 Can
 you
 explain
 the
 CART
 stuff
 more?

 You
 would
 collapse
 the
 biological
 data
 for
 each
 site
 into

some
 number,
 then
 use
 that,
 there
 are
 100
 different
 ways
 to
 do
 that?
 Can
 you
 expand
 on
 that?

R:
 All
 the
 analyses
 have
 handled
 the
 community
 as
 an
 entity.

 Now
 we
 will
 break
 it
 down
 and
 look
 at

aspects
 of
 the
 community.

 You
 emphasize
 different
 aspects
 of
 the
 data
 set.

 So
 you
 can
 emphasize

riffle
 guilds,
 or
 filter
 feeders,
 there
 are
 hundreds
 of
 those
 you
 could
 put
 together.
 From
 literature,
 we

know
 there
 are
 other
 sets
 more
 sensitive
 to
 flow.

 That’s
 where
 we’d
 focus
 initially.

 There
 is
 also
 give

and
 take-­‐
 look
 at
 data,
 analysis…could
 bring
 in
 IHA
 metrics,
 to
 see
 how
 they
 relate
 to
 some
 of
 these

flow
 metrics
 as
 well,
 actual
 flows
 we’re
 talking
 about
 are
 eco-­‐deficits.

 If
 we
 start
 screening
 the
 large

number
 of
 metrics
 representing
 communities,
 we
 may
 wind
 up
 with
 different
 aspects
 of
 communities

responding
 to
 flow
 changes
 differently.

 That’s
 what
 you
 want.

 So
 I
 think
 that
 approach
 has
 a
 lot
 of

positive
 aspects,
 so
 instead
 of
 doing
 clustering
 and
 moving
 on
 to
 flows,
 now
 we’re
 doing
 back
 and
 forth

between
 the
 two
 to
 get
 the
 clusters
 related
 to
 the
 flow
 variables
 so
 it
 gets
 a
 little
 more
 complicated.

Q:

 Regarding
 scheduling
 next
 steps,
 where
 do
 you
 envision
 this
 thing
 in
 the
 next
 month
 or
 two?

R:

 Its’
 all
 ongoing.

 Though
 I’ll
 be
 out
 of
 town
 for
 a
 month
 in
 May
 and
 things
 will
 slow
 for
 me
 (Tom).

Earliest
 I
 could
 present
 additional
 work
 to
 the
 group
 will
 be
 in
 June
 (same
 for
 RTI?).

Facilitator:
 Thoughts
 on
 Phillips
 presentation?

 Anything
 that
 is
 encouraging?

Q:

 Re.
 slide
 X,
 I
 was
 thinking
 drainage
 basins
 or
 ecoregions,
 makes
 sense
 to
 me
 biologically
 with
 my

experience
 of
 where
 fish
 are.

 Was
 this
 any
 more
 explanatory?

 It
 gives
 me
 confidence
 that
 it
 may
 not

be
 better
 at
 explaining
 it,
 it
 is
 something
 about
 the
 biology
 that
 we
 know

C:

 One
 take
 home
 is
 that,
 in
 the
 end
 it
 will
 be
 fairly
 simple
 set
 of
 classes.

 It
 may
 be
 real
 simple,
 like
 2

or
 3.

 Maybe
 we’re
 at
 that
 point
 where
 we
 need
 to
 reset
 our
 expectations
 that
 we
 may
 have
 a
 finely

tuned
 system.

 This
 stuff
 and
 what
 Mary
 will
 present
 lead
 me
 to
 think
 that.

C:

 That
 seems
 reasonable
 to
 me.

 Also
 the
 folks
 that
 would
 use
 this
 tool
 are
 going
 to
 appreciate
 that.

C:

 One
 thing
 that
 is
 important
 to
 capture
 in
 summarizing
 for
 report,
 point
 out
 results
 for
 lack
 of
 ability

to
 have
 very
 defined
 categories
 so
 other
 people
 don’t
 have
 to
 go
 back
 and
 do
 it.

C:

 Something
 else
 about
 having
 simple
 classifications,
 BEC
 project
 next
 steps
 are
 to
 develop
 biology-­‐
flow
 relationships.
 If
 you
 have
 too
 many
 stream
 classes,
 it
 defeats
 your
 purpose-­‐
 if
 you
 have
 many

classes
 and
 not
 enough
 data
 it
 is
 difficult.

C:

 There
 is
 at
 least
 some
 chance
 we’re
 going
 to
 arrive
 at
 the
 conclusion
 that
 we
 cannot
 classify
 rivers

and
 streams
 of
 NC
 in
 a
 way
 that
 provides
 clear
 direction
 on
 how
 to
 set
 ecological
 baseline.

 If
 so,
 we’ll

be
 the
 first
 in
 the
 country
 to
 have
 tested
 the
 assumption
 that
 you
 must
 classify
 first,
 as
 rigorously
 as

we
 have,
 and
 arrived
 at
 that
 conclusion.

 It
 may
 be
 characterized
 by
 some
 as
 failure,
 but
 it
 is
 quite
 an

achievement.

 So
 the
 thing
 we
 must
 cope
 with,
 is
 how
 to
 set
 ecological
 baseline
 without
 classification,

which
 as
 far
 as
 I
 know
 has
 not
 been
 attempted.

 I
 think
 we’re
 approaching
 a
 good
 place-­‐
 either
 we’ll

have
 a
 class
 system
 we
 believe
 in,
 or
 we
 would
 have
 altered
 national
 paradigm,
 and
 will
 need
 a

conclusion
 about
 how
 to
 move
 forward
 without
 classification.

 One
 or
 the
 other
 will
 happen
 in
 the
 next

60
 days.

IV.
 Presentation:

 SARP:

 River
 Classification
 Framework

Presenter:

 Mary
 Davis,
 SARP

We’ve
 been
 working
 in
 OK,
 LA,
 doing
 much
 the
 same
 thing.

 Today
 I’m
 reporting
 on
 a
 product
 we’ve

been
 developing
 for
 SALCC.

 It
 fits
 in
 nicely
 and
 it’s
 not
 a
 bad
 time
 following
 the
 work
 just
 reported
 on

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 20
 of
 51

to
 step
 back
 and
 look
 at
 classifications
 again.

 I’ve
 been
 coordinating
 the
 Southern
 Instream
 Flow

Network
 since
 2007.

 We’ve
 come
 together
 as
 part
 of
 the
 Southern
 Atlantic
 Regional
 Partnership
 (Fish

and
 wildlife
 organizations
 in
 14
 states,
 leveraging
 resources).

 Purpose
 of
 the
 network
 is
 to
 facilitate

protective
 instream
 flow
 policies
 and
 practices
 in
 15
 southern
 states
 by
 providing
 science-­‐based

resources
 and
 opening
 lines
 of
 communication.

 Today
 I’ll
 talk
 about
 the
 SARP
 River
 Classification

Framework
 and
 then
 spend
 a
 few
 minutes
 about
 how
 classification
 is
 used.

We
 are
 not
 proposing
 a
 classification
 -­‐
 we’re
 providing
 a
 classification
 framework.

 There
 are
 lots
 of

attributes
 that
 we
 can
 use
 to
 characterize
 a
 particular
 point
 in
 a
 stream.

 We’re
 giving
 you
 attributes

that
 can
 be
 used
 in
 any
 number
 of
 ways.

SARP
 River
 Classification
 Framework
 Objectives:

• Characterize

 streams
 by
 ecologically
 relevant
 characteristics

• Provide
 common
 terms
 for
 describing
 rivers
 across
 the
 region
 (trying
 to
 get
 beyond
 drawing

lines
 around
 a
 river,
 but
 saying
 data
 for
 fish,
 macro-­‐invertebrates
 can
 be
 used
 from
 other

states.)

• Support
 development
 of
 flow-­‐ecology
 relationships

My
 work
 is

guided
 by
 the

ELOHA

(Ecological

Limits
 of

Hydrologic

Alteration)

framework

(see
 slide
 9)-­‐

you’ve
 been

introduced
 to

this
 so
 will
 not

go
 into
 a
 lot
 of

detail.

 We’re

talking
 about

characterizatio
n.

 Experts
 that

developed

ELOHA
 felt

that

characterization
 would
 be
 a
 necessary
 step
 before
 describing
 eco-­‐responses
 to
 alteration.

 We’re

finding
 that
 may
 or
 may
 not
 be
 the
 case.

The
 question
 is
 –
 what
 does
 this
 characterization
 look
 like?

 For
 SARP
 we’ve
 done
 the
 hydrologic

classification
 and
 the
 geomorphic
 sub-­‐classification.

I
 want
 to
 point
 out
 our
 classification
 system
 is
 standardized
 with
 NHD
 plus
 dataset,
 which
 RTI
 is
 doing

hydrologic
 modeling
 based
 on.

 It
 was
 a
 community
 based
 process-­‐
 we
 called
 together
 experts
 from

around
 and
 outside
 the
 region
 to
 discuss
 what
 were
 the
 ecologically
 significant
 components
 that
 we

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 21
 of
 51

wanted
 to
 classify,
 and
 what
 were
 the
 reasonable
 breaks
 for
 those
 classes
 using
 a
 quantitative
 metric.

The
 Nature
 Conservancy
 out
 of
 Boston
 did
 our
 modeling,
 so
 Kim
 and
 Becca
 are
 familiar
 with
 the
 work.

Geomorphic
 subclassification

Stream
 gradient:

 each
 one
 NHD
 plus
 river
 segment
 is
 given
 an
 attribute
 of
 stream
 gradient,
 we
 had

five
 levels
 of
 classes.

 Very
 flat
 up
 to
 high
 gradient
 in
 the
 mountains-­‐
 we
 had
 quite
 a
 bit
 of
 variability

throughout
 the
 SARP
 region.

Ecoregions:

 Each
 segment
 is
 attributed
 to
 different
 ecoregions,
 EPA
 III
 and
 IV,
 freshwater
 and

ecological
 drainage
 unit.

Stream
 temperature:
 We
 tried
 doing
 a
 stream
 temperature
 classification,
 but
 we
 could
 not
 find
 breaks

for
 what
 was
 largely
 a
 warm
 class
 in
 the
 southeast.

 TNC
 were
 able
 to
 put
 breaks
 for
 their
 stream

classes
 based
 on
 biological
 associations,
 we
 applied
 that
 stream
 temp
 model
 to
 SE,
 that’s
 what

assigned
 most
 of
 SE
 to
 a
 warm
 class.

 We
 thought
 that’s
 not
 right,
 there
 are
 more
 divisions
 than
 that-­‐

we
 know
 that
 south
 Florida
 should
 have
 a
 hot
 class.

 We
 weren’t
 able
 to
 distinguish
 temp
 to
 break
 out

anything
 other
 than
 warm
 class
 for
 the
 SE.

 For
 the
 fish
 assemblages
 we
 were
 able
 to
 pull
 together

from
 similar
 data
 that
 you’ve
 been
 working
 with.

 Elevation,
 eco-­‐region
 ,
 slope
 were
 better
 predictors

of
 assemblage
 than
 temperature.

 Though
 NC
 does
 get
 into
 the
 cool
 classification.

Stream
 size:

 based
 on
 basin
 area
 and
 mean
 annual
 flow.

Baseflow
 index

• characterization
 of
 the
 dominance
 of
 groundwater
 versus
 surface
 water
 flow
 in
 the
 river.

Soils-­‐
 much
 new
 data
 from
 NRCS

• Available
 water
 capacity

• Soil
 organic
 carbon

• %
 Sand,
 Silt,
 and
 Clay

Bedrock
 Geology:

 broken
 largely
 by
 influence
 on
 stream
 pH.

 We
 didn’t
 get
 far
 on
 PH
 because
 we

didn’t
 have
 good
 validation
 samples

Land
 forms:

 TNC
 has
 done
 cool
 modeling-­‐
 hills,
 valley,
 aspect.

 Now
 we
 have
 an
 interesting
 component

called
 land
 forms

Land
 use:

 urban,
 agriculture

We
 have
 a
 tremendous
 amount
 of

 information.

 The
 way
 I
 visualize
 it
 we
 have
 each
 one
 of
 these

NHD+
 segments
 in
 a
 matrix,
 we’ve
 got
 columns
 for
 each
 of
 the
 attributes.

 We
 use
 that
 to
 get
 towards

our
 hydrologic
 classification.

 Before
 I
 get
 to
 that
 let
 me
 talk
 about
 the
 classification
 that
 we
 chose
 to

use
 in
 our
 framework.

 Again,
 this
 is
 meant
 to
 provide
 info
 it
 is
 not
 meant
 to
 override
 or
 exclude
 other

classifications
 that
 are
 in
 place,
 this
 is
 meant
 to
 add
 information.

We
 ran
 into
 same
 problems
 NC
 did
 by
 trying
 to
 use
 cluster
 analysis
 of
 hydro
 metrics
 to
 describe
 hydro

classes
 of
 rivers.

 You
 get
 good
 correspondence
 with
 some
 of
 the
 metrics
 like
 drainage
 area
 or
 base

flow
 index
 driving
 the
 formation
 of
 different
 clusters,
 but
 then
 you
 start
 looking
 at
 how
 the
 members

of
 each
 cluster
 are
 classified.

 You
 may
 call
 some
 of
 those
 members
 perennial,
 but
 then
 there
 are

members
 of
 that
 cluster
 that
 have
 many
 days
 of
 low
 flow
 and
 are
 really
 intermittent.

 So
 you
 can’t
 go

out
 to
 a
 river
 and
 say
 “that’s
 an
 x”
 because
 the
 clusters
 are
 multi-­‐dimensional
 and
 the
 membership
 is

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 22
 of
 51

not
 distinct.

 So
 we
 gave
 up
 on
 the
 multi-­‐
 vriant
 cluster
 approach
 to
 develop
 hydrologic
 classifications.

That’s
 the
 original
 EFS,
 McManamay,
 we
 had
 someone
 named
 Conrad
 working
 on
 this-­‐
 we
 gave
 up
 on

it.

Since
 then,
 Henrickson
 was
 doing
 a
 national
 hydrologic
 classification.

 He
 also
 saw
 that
 the
 cluster

analysis
 wasn’t
 working-­‐
 he
 did
 it
 in
 many
 places

 (PA,
 TX)
 but
 nobody
 was
 using
 it.

 He
 came
 up
 with
 a

more
 ecologically
 relevant
 logical
 approach
 to
 classifying
 rivers.

 This
 is
 his
 approach.

It’s
 a
 hierarchy
 of
 classification
 based
 on
 what
 experts
 think
 are
 the
 most
 important
 factors
 describing

a
 river.

 You
 can
 walk
 out
 and
 say-­‐
 that
 is
 a
 perennially
 baseflow
 fed
 river
 and
 people
 know
 what
 you

are
 talking
 about.

Perennial
 vs
 intermittent

Size
 (median
 daily
 flows)

Variability
 of
 flows
 (groundwater
 vs
 surface
 water)

Flood
 frequency-­‐
 median
 number
 of
 days
 water
 out
 of
 the
 channel

Flood
 timing(Across
 nation
 there
 is
 a
 seasonality
 shift
 for
 high
 flows)

This
 committee
 went
 through
 the
 variables.

 We
 had
 to
 use
 available
 information-­‐
 gage
 data
 at

reference
 gages
 or
 minimally
 disturbed
 gages.

 We
 had
 to
 make
 some
 adjustments.

 The
 classification

I’ll
 show
 you
 follows
 this
 the
 best
 we
 could.

 The
 beauty
 of
 this
 system
 is
 it
 forms
 a
 national

classification,
 but
 you
 can
 subclass
 within
 these
 classes
 that
 he
 assigned,
 so
 you
 can
 stay
 consistent

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 23
 of
 51

with
 the
 national
 but
 we
 found
 we
 really
 needed
 a
 regional.

First
 thing
 we
 found,
 we
 could
 not
 distinguish
 perennial
 vs
 intermittent
 streams.

 Its
 defined
 by

whether
 there
 are
 any
 zero
 flow
 days.

 There
 are
 problems
 with
 defining
 zero
 flow
 days,
 such
 as
 gage

variability.

 We
 found
 a
 lot
 of
 low
 flow
 rivers
 that
 get
 down
 to
 1-­‐2
 cfs,
 but
 within
 error
 of
 gage
 that

could
 be
 zero.

 Trying
 to
 model
 zero
 flow
 days
 was
 a
 non-­‐starter.

C:

 At
 a
 national
 level
 it
 makes
 sense
 talking
 about
 desert
 streams
 vs
 perennial
 streams,
 but
 in
 the

southeast
 it
 makes
 much
 less
 sense.

In
 SE,
 we
 can
 define
 intermittent
 based
 on
 a
 stream
 going
 dry,
 or
 it’s
 just
 so
 flat
 and
 slow
 you
 can’t

detect
 flow.

 Not
 to
 say
 intermittency
 is
 not
 important,
 but
 in
 our
 classification
 we
 weren’t
 able
 to
 offer

that
 distinction.

Size
 attribute-­‐
 they
 did
 it
 by
 median
 daily
 flow.

 We
 didn’t
 have
 that
 information,
 plus
 we
 had
 already

finished
 a
 size
 classification
 for
 the
 South
 Atlantic
 region.

 We
 have
 size
 attributes
 and
 most
 of
 it
 fits

into
 their
 small
 size
 (which
 is
 a
 sizable
 river)

• Headwater,

• Creek,

• Small
 R.,

• Medium
 R.,

• Large
 R.,

• Great
 R.

Variability
 attribute
 (median
 annual
 std
 deviation/mean
 flow)

• Very
 low,

• Low,

• Medium-­‐High

Flood
 frequency
 attribute-­‐
 we
 didn’t
 have
 the
 data
 to
 analyze
 it.

 Somebody
 could
 do
 that
 in
 the

future.

Flood
 timing-­‐
 we
 don’t
 have
 any
 differentiation
 in
 the
 east.

 As
 you
 move
 west
 you
 pick
 up
 high
 flows

later
 in
 the
 year.

What
 we
 wind
 up
 with:
 the
 hydrologic
 classification
 came
 up
 to
 be
 very
 simple.

We
 can
 offer
 size
 class
 and
 variability.

 We
 have
 stream
 size
 class,
 using
 basin
 area
 –
 can
 use
 flow
 or

basin
 area.

 We
 found
 people
 more
 comfortable
 with
 basin
 area
 than
 flow.

 That
 comes
 with
 NHD+

database.

 It
 was
 no
 problem
 to
 assign
 size
 for
 all
 NHD+
 reaches.

For
 Flow
 variability
 class
 we
 only
 had
 that
 at
 gage
 stations.

Predict
 Flow
 Variability
 Class
 for
 Ungaged
 Locations
 (The
 modeling
 work
 consisted
 of
 four
 major
 steps)

• Compile
 set
 of
 USGS
 gages,
 assign
 hydrologic
 class,
 and
 link
 them
 to
 the
 appropriate
 NHD+

reach

• Attribute
 each
 stream
 reach
 and
 gage
 with
 GIS
 predictor
 variables

• Build
 random
 forest
 (RF)
 classification
 models
 using
 the
 random
 forest
 package
 in
 R
 (statistical

package)

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 24
 of
 51

• Apply
 the
 best
 RF
 model
 to
 each
 stream
 reach
 and
 map
 each
 stream
 reach
 according
 to
 the

“highest
 probability”
 class

Of
 75
 predictor
 variables
 we
 used,

consistently
 the
 most
 important
 variables

were:

• mean
 baseflow
 index
 –

relative
 importance
 of

ground
 and
 surface
 water

• stream
 size

• cumulative
 drainage
 area

and

• run-­‐off
 coefficient.

This
 is
 what
 we
 came
 up
 with

 -­‐
 Flow

Variability.

 The
 dots
 are
 USGS
 gages,
 blue

are
 very
 low
 variability,
 tend
 to
 be
 in
 SC

and
 southern
 NC
 and
 large
 rivers.

 Green-­‐

low
 variability
 (mostly
 coastal
 plain
 and
 up

to
 base
 of
 mountains.

 This
 is
 the
 SLCC
 boundary.

 Medium
 variability,
 red
 is
 high
 variability.

 There

weren’t
 many
 high
 variability
 sites
 so
 we
 combined
 with
 medium
 variability.

 Orange
 is
 medium–high

variability.

 If
 you
 zoom
 in
 you
 can
 see
 there
 is
 some
 distinction
 at
 the
 reach
 level
 that
 doesn’t
 show
 up

on
 a
 map
 like
 this.

 As
 you’d
 expect,
 larger
 rivers
 are
 low
 variability.

Blue
 is
 coastal
 plain.

I’ve
 got
 3
 slides
 coming
 up
 showing
 this
 classification
 framework
 is
 meant
 to
 mix
 and
 match
 what

attributes
 might
 be
 useful.

This
 slide
 shows
 Ecoregion
 III
 but

Ecoregion
 IV
 was
 more
 important

for
 fish
 and
 macro-­‐invertebrates.

Then
 you
 can
 put
 size
 on
 top
 of

that,
 and
 flow
 variability
 on
 top
 of

that.

 So
 then
 you
 can
 talk
 about

size
 of
 river,
 how
 much

groundwater
 it
 is,
 and
 what

ecoregion.

 That
 tends
 to
 be
 what

ecologist
 do
 when
 they
 describe

rivers,
 and
 we’ve
 found
 that
 fish

assemblages
 in
 groundwater

dominated
 systems
 tend
 to
 be

more
 sensitive
 to
 water

withdrawals
 than
 those
 in
 highly

variable
 systems.

 They’ve
 already

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 25
 of
 51

adapted
 to
 highly
 variable
 systems.

 Whereas
 groundwater
 driven
 rivers
 assemblages
 aren’t
 as
 well

adapted.

I
 also
 want
 to
 show
 when
 I

listed
 variables
 important
 for

describing
 variability
 classes-­‐

• mean
 baseflow

index
 –
 relative

importance
 of

ground
 and

surface
 water

• stream
 size

• cumulative

drainage
 area

and

• run-­‐off

coefficient.

Ecoregion
 is
 not
 in
 there.

 What
 I
 love
 about
 this
 is
 it
 follows
 the
 ecoregion.

 It
 just
 fell
 out.

You
 have
 a
 lot
 of
 attributes,
 what
 do
 you
 do
 with
 them?

 Having
 followed
 your
 process,
 I
 wanted
 to

review
 how
 these
 classifications
 are
 used.

 Basic
 question
 is
 when
 do
 you
 classify?
 I’ll
 use
 Michigan
 and

the
 Potomac
 watershed
 as
 examples.

Authors
 of
 ELOHA
 conceived
 of
 the
 classification
 before
 analysis
 which
 is
 what
 you
 have
 been
 working

through
 for
 a
 while.

The
 concept
 is
 the
 flow-­‐ecology
 relationships
 are
 determined
 a
 priori.

 The
 assumption
 is
 the

classification
 is
 going
 to
 some
 way
 be
 sensitive
 to
 flow
 alteration.

 So
 far
 you’ve
 got
 things
 grouped
 but

haven’t
 answered
 if
 they
 are
 predictive
 of
 flow
 responses.

So,
 the
 question
 is,
 classify
 before
 the
 analysis
 or
 to
 inform
 the
 analysis?

Michigan-­‐
 their
 experience
 for
 setting
 instream
 flow
 has
 been
 brought
 as
 a
 good
 example
 of
 ELOHA.

MI
 had
 a
 lot
 of
 data,
 they
 used
 a
 lot
 of
 expert
 knowledge
 about
 fish
 and
 their
 reason.

 They
 came
 to
 an

a
 priori
 approach
 to
 classification
 using
 stream
 temp
 and
 stream
 size.

 Two
 parameters
 that
 fish
 are

sensitive
 to
 and
 were
 sensitive
 to
 water
 withdrawals.
 If
 you
 withdraw
 water
 you
 slow
 a
 stream
 down

and
 warm
 it
 up,
 change
 classes.

 And
 the
 fishes
 respond
 to
 this.

 They
 used
 their
 fish
 data
 from
 17,000

fish
 sample
 locations.

 Those
 locations
 dominated
 by
 cold
 fishes
 were
 classified
 as
 cold
 water,
 etc.

 Its

very
 simple,
 logical.

 They
 used
 a
 modeling
 approach-­‐
 IFIM
 type
 habitat
 suitability
 of
 fish
 tolerance
 to

stream
 temp
 and
 assemblages
 in
 different
 sizes
 to
 develop
 a
 complex
 and
 interesting
 approach
 to
 say

how
 fish
 in
 different
 streams
 responded
 when
 you
 withdrew
 flows…

The
 point
 is
 their
 classification
 was
 done
 beforehand,
 logical,
 but
 has
 never
 been
 tested.

 They
 didn’t

go
 through
 the
 rigor
 you
 did
 to
 see
 if
 classification
 had
 any
 relation
 to
 flow
 sensitivity-­‐
 they
 are
 using

models.

That
 is
 an
 example
 of
 successful
 classification
 early
 on.

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 26
 of
 51

Potomac
 watershed
 used
 a
 different

method-­‐
 they
 used
 hydrologic
 modeling
 like

you,
 pulled
 in
 macro-­‐invertebrate
 data

from
 states
 that
 intersect
 with
 that

watershed.

 Rather
 than
 classify
 first,
 they

threw
 the
 data
 into
 a
 regression
 analyses.

X
 axis-­‐
 measure
 of
 flow
 alteration

Y
 axis-­‐
 measure
 of
 macro-­‐invertebrate

condition

They
 had
 a
 wide
 range
 of
 variability
 of
 the

condition
 of
 macro-­‐invertebrates.

So
 at
 this
 point
 the
 question
 is
 will

classification
 help
 improve
 the
 strength
 of

these
 relationships?

 What
 they
 found,
 was

some
 biological
 metrics
 do
 not
 need
 to
 be

classified.

With
 Family
 level
 taxa-­‐
 these
 did
 not
 need

to
 be
 classified
 (they
 didn’t
 make
 any

difference):

• watershed
 size

• season

• bioregion

On
 the
 other
 hand,
 there
 were
 other
 variables
 that
 did
 need
 to
 be
 classified.

• %
 shredders-­‐
 related
 to
 watershed
 size

• %net
 caddisfly-­‐
 related
 to
 season

• %
 Chironomidae-­‐
 related
 to
 bioregion

So
 what
 these
 other
 experiences
 tell
 you
 about
 when
 to
 classify-­‐
 it’s
 not
 clear.

 TNC
 released
 a
 report

of
 9
 case
 studies
 of
 ELOHA.

 Sometimes
 classification
 worked
 and
 sometimes
 it
 didn’t.

Regarding
 who
 has
 ever
 gone
 without
 classification,
 it
 would
 be
 in
 the
 report.

Questions
 (Q),
 Comments
 (C),
 Response
 (R):

C:
 I
 think
 the
 link
 to
 that
 report
 went
 to
 EFSAB

I
 was
 glad
 to
 see
 Tom
 and
 Philip
 working
 on
 letting
 response
 to
 flow
 alteration
 begin
 to
 inform
 the

appropriate
 class.

 It
 may
 be
 as
 simple
 as
 ecoregion,
 or
 size
 of
 river.

 The
 analyses
 seem
 to
 be
 moving
 in

that
 direction.

Q:
 I
 started
 thinking
 about
 how
 we
 might
 use
 this
 report-­‐-­‐the
 3
 slides
 of
 using
 physiographic
 region,

size,
 variability.

 It
 gets
 down
 to
 as
 few
 variables
 as
 possible,
 may
 help
 inform
 us
 moving
 forward.

Rather
 than
 gradient,
 temperature-­‐physiographic
 province
 will
 essentially
 do
 the
 same
 thing.

 Why
 use

2-­‐3
 variables
 when
 you
 can
 use
 one.

R:
 If
 you
 include
 biologic
 classification
 before
 you
 set
 standards,
 going
 back
 to
 the
 slide
 Phillip
 had
 (8

centroids
 in
 NC)
 I
 wasn’t
 sure
 how
 you’d
 assign
 biological
 class
 to
 unsampled
 locations?

C:

 When
 you
 were
 showing
 various
 parameters
 that
 didn’t
 show
 correlation,
 I
 thought
 seasonality
 was

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 27
 of
 51

interesting
 given
 all
 the
 seasonal
 cues
 for
 spawning,
 etc.

C:
 On
 national
 level,
 snow
 runoff
 is
 important.

R:

 Not
 to
 say
 that
 in
 a
 river
 with
 a
 pattern,
 that
 you
 don’t
 want
 to
 try
 to
 maintain
 it.

Q:

 This
 is
 just
 for
 the
 Potomac,
 it
 may
 be
 scale
 related.

R:

 This
 is
 season
 of
 sample,
 you’d
 expect
 taxa
 to
 shift.

Q:
 Regarding
 intermittent
 vs
 perennial,
 my
 gut
 tells
 me
 it’s
 not
 zero
 flow
 days,
 you
 have
 7Q10
 of
 a

certain
 level.

 You
 can
 classify
 streams
 by
 what
 their
 7Q10
 is,
 and
 it
 explains
 some
 vulnerability
 to
 the

biota.

 If
 you
 have
 rivers
 with
 really
 low
 7Q10s,
 they
 are
 probably
 not
 great
 candidates
 for
 water

withdrawa,
 if
 you’re
 talking
 about
 a
 screening
 tool
 for
 water
 withdrawals.

R:

 That
 logic
 is
 what
 led
 us
 to
 realize
 that
 in
 the
 east
 we
 see
 zero
 flow
 days
 differently
 than
 in
 the

western
 U.S.

 In
 west,
 they
 are
 drying
 the
 river
 out.

 Here
 you
 can
 dry
 the
 river
 out
 or
 have
 negligible

flow.

 So
 we
 needed
 to
 get
 a
 different
 parameter
 than
 zero
 flow
 days.

 Size
 is
 directly
 correlated
 with

mean
 annual
 flow.

 Instead
 of
 7Q10
 you
 can
 use
 size
 class
 as
 a
 proxy
 for
 very
 low
 flow.

C:
 That’s
 what
 Susquehanna
 approach,
 where
 they
 had
 more
 stringent
 requirements
 for
 smaller

watersheds.

 Same
 idea.

C:

 if
 we
 go
 that
 direction,
 it
 seems
 like
 we’d
 want
 to
 draw
 the
 connection.

 We’re
 using
 stream
 size

which
 is
 highly
 correlated
 with
 x,
 y,
 z.

 It
 would
 explain
 these
 concerns.

R:

 John
 Kauffman
 was
 working
 on
 a
 report
 in
 VA
 looking
 at
 effect
 of
 withdrawals
 on
 small
 streams
 vs

large
 streams.

 Size
 is
 important.

Facilitator:

 Questions
 of
 clarification?

 Or
 thoughts
 on
 how
 this
 informs
 your
 work?

C:

 I
 think
 its
 pointing
 us
 in
 a
 similar
 direction-­‐
 we’re
 not
 that
 different
 from
 where
 other
 folks
 have

trod.

 Folks
 have
 tried
 different
 things
 but
 it
 is
 all
 coming
 back
 to
 some
 basics.

R:

 Talking
 about
 taking
 fish
 and
 macroinverts
 forward,
 what
 the
 Potomac
 found
 was
 that
 it
 was
 flow

alteration
 that
 sometimes
 classification
 helped,
 sometimes
 it
 didn’t.

 It’s
 a
 little
 different
 trajectory

than
 what
 you
 are
 on.

Q:

 Where
 are
 we
 with
 flow
 alterations?

 We
 have
 streams
 that
 are
 in
 a
 current
 state,
 and
 we
 need
 to

have
 measures
 of
 ecological
 integrity
 related
 to
 current
 state.

 How
 are
 we
 going
 to
 capture
 what
 the

current
 state
 is?

 For
 unsampled
 reaches?

C:

 If
 you
 successfully
 model
 biological
 with
 your
 flow
 or
 your
 physical,
 if
 you
 have
 information
 from

ungaged
 sites,
 that’s
 your
 prediction.

Q:

 Wouldn’t
 it
 be
 a
 reference
 though,
 we
 have
 to
 throw
 in
 land
 use
 somewhere.

C:

 RTI
 is
 developing
 that,
 right?

C:

 Yes,
 and
 we’ll
 be
 presenting
 that.

 We
 have
 the
 IHA
 metrics,
 and
 correlation
 between
 some
 of
 the

metrics,
 have
 since
 increased
 and
 revised
 metrics
 to
 those
 we’re
 currently
 using.

 I’ll
 present
 those
 at

future
 meeting.

 Characterizing
 current
 condition
 –
 there
 is
 a
 possibility
 to
 use
 modeling
 to
 characterize

that
 regardless
 of
 biology
 data…you
 could
 almost
 classify
 rivers
 based
 on
 current
 conditions
 now.
 How

altered
 are
 they?

 (this
 was
 difficult
 to
 hear
 in
 the
 recording,
 but
 this
 information
 will
 be
 presented
 by

RTI
 in
 future)

Q:

 This
 project
 has
 essentially
 been
 completed?

 The
 SALCC
 effort?

R:

 Yes,
 we’re
 done.

 It
 will
 be
 distributed
 to
 SALCC.

Q:

 Is
 it
 a
 state
 where
 you
 could
 take
 a
 GIS
 system
 and
 do
 an
 analysis
 on
 NC?

R:

 Most
 of
 state
 has
 been
 done,
 just
 need
 to
 get
 flow
 variability
 for
 the
 mountains.

Q:

 Regarding
 the
 work
 on
 the
 ungaged
 streams?

 What
 was
 error?

R:

 20%
 error.

C:

 What
 they
 did
 in
 the
 NE
 classification
 that
 could
 be
 done
 with
 ours,
 even
 stream
 size
 can
 be

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 28
 of
 51

collapsed
 down
 to
 small/medium/large.

R:

 One
 advantage
 of
 hierarchal
 classification
 approach,
 you
 don’t
 need
 to
 be
 worried
 about
 sensitivity

of
 class
 breaks.

 We
 have
 the
 number
 and
 we
 have
 the
 category.

 I
 would
 imagine
 Tom
 would
 use
 basin

area
 rather
 than
 class
 of
 stream
 size
 for
 example,
 then
 you
 don’t
 have
 to
 worry
 about
 class
 breaks-­‐
 the

data
 will
 tell
 you
 the
 important
 break.

V.
 Presentation:

 Results
 of
 Recommendations
 Brainstorm
 in
 March

Presenter:
 Mary
 Lou
 Addor,
 Facilitator

Given
 the
 recommendation
 framework
 developed
 at
 the
 March
 19,
 2013
 meeting,
 what
 do
 you
 believe

are
 the
 next
 steps
 to
 move
 you
 and
 the
 rest
 of
 the
 EFSAB
 closer
 toward
 recommendations
 for
 a

planning
 tool
 and
 writing
 a
 recommended
 report?
 The
 EFSAB
 was
 asked
 to
 respond
 to
 4
 questions

about
 the
 brainstormed
 list
 of
 recommendations:

1. General
 reactions
 including
 how
 it
 is
 organized?
 Categorized?

2. What
 changes/additions
 could
 be
 added
 to
 the
 recommendation
 list
 based
 on
 what
 you
 heard

today
 about:

a. BEC

b. SALCC

c. Anything
 else
 missing?
 If
 so,
 which
 category?

General
 Reactions:

1. Organize
 parts-­‐
 condense

2. Page
 3:

 charge/parameters
 –
 place
 in
 a
 preamble/preface
 to
 the
 report
 (examples:
 flow

requirements
 and
 other
 aspects
 of
 the
 law
 that
 must
 be
 made
 explicit)

3. Refined
 charge
 –
 a
 report
 will
 read
 by
 a
 larger
 audience
 other
 than
 DWR.
 Convey
 context
 about

the
 report
 so
 that
 it
 is
 not
 misinterpretative
 or
 raised
 unnecessary
 flags
 for
 constituents,

planners,
 etc.

4. Include
 the
 EFSAB’s
 reactions
 and
 concerns
 regarding
 the
 original
 charge
 (ex:

defining/characterizing
 the
 ecology)

5. Philosophical
 points
 of
 view
 many
 need
 to
 be
 edited.
 Example
 page
 6,
 necessity
 of
 model.

Separate
 out
 statements
 of
 concepts,
 opinion
 from
 recommendations

6. Page
 10:

 pre/post
 monitoring
 –
 outside
 of
 EFSAB
 charge.

7. Pre/post
 monitoring
 withdrawals
 discussion:
 caution,
 if
 the
 recommendation
 report
 includes

future
 considerations,
 remember
 that
 the
 work
 of
 the
 EFSAB
 is
 a
 planning
 tool
 and
 is
 not
 being

applied
 to
 policy/permitting
 decisions.
 Make
 sure
 this
 section
 is
 expressed
 as
 future
 learning

and
 understanding.

8. Page
 9:
 Triggers
 for
 future
 site
 (DWR
 already
 has
 established
 triggers)
 –
 how
 that
 might
 relate

to
 adaptive
 management.

9. Allow
 the
 EFSAB
 to
 include
 value
 added
 recommendations,
 insights,
 concerns
 as
 a
 result
 of
 3

years
 of
 work
 to
 share
 with
 the
 broader
 audience.

10. Include
 definition
 of
 consensus
 principals,
 recommendations,
 trial
 balloons.

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 29
 of
 51

11. Page
 4:

 In
 lieu
 of
 classification
 approaches,
 #
 17,
 18,
 20
 –
 fall
 under
 determining
 ecological

flows.
 Page
 5:

 other
 aspects
 –
 points
 1-­‐9,
 except
 3
 –
 determining
 ecological
 flows.

12. Acknowledge
 the
 EFSAB
 used
 models
 to
 respond
 to
 charge
 –
 but
 is
 not
 advising
 DWR
 on
 models

to
 use
 nor
 that
 DWR
 has
 to
 use
 models.

13. Use
 actual
 gage
 data
 where
 it
 is
 available/OASIS
 calibrated
 for
 gage
 data.

14. Treatment
 of
 Coastal
 areas
 (question:
 why
 is
 it
 separated
 in
 the
 recommendations
 framework)

when
 what
 happens
 in
 Piedmont
 affects
 the
 Coast.

a. Limited
 coastal
 data

b. None
 of
 the
 models
 accessible
 to
 EFSAB
 could
 analyze
 coastal
 variables
 (flows
 difficult
 to

predict)

c. Withdrawal
 pressures
 slight
 in
 coastal
 area;
 EFSAB
 more
 focused
 on
 Piedmont
 followed

with
 Mountains
 given
 more
 data,
 models
 to
 analyze
 variables
 and
 withdrawal
 pressures

are
 greater.

16.
 Approaches
 to
 flows

a. In
 the
 recommendations
 framework
 –
 it
 includes
 how
 to
 estimate
 flows
 but
 it
 is
 missing

how
 to
 link
 ecological
 flows.
 Pages
 (2,
 3,
 6,
 7)
 –
 Tom
 Cuffney
 recommendation

17. Page
 5:
 In
 lieu
 of
 PHABSIM
 –
 habitat
 to
 biota
 (there
 is
 a
 data
 gap).

18. Develop
 a
 Work
 Team
 to
 sort
 through
 the
 brainstormed
 list
 of
 recommendations

What
 changes/additions
 could
 be
 added
 to
 the
 recommendation
 list
 based
 on
 what
 you
 heard
 today

about
 BEC
 and
 SALCC:

BEC

19. With
 BEC
 –
 there
 is
 high
 uncertainty.
 We
 need
 to
 be
 afraid
 to
 classify
 “happily”
 and
 sufficiently.

Be
 prepared
 to
 move
 forward
 in
 the
 absence
 of
 a
 suitable
 system.
 There
 is
 the
 possibility
 of

identifying
 ecological
 integrity
 without
 good
 classes.
 Be
 mindful,
 some
 are
 unable
 to
 conceive

how
 to
 move
 forward
 without
 some
 sort
 of
 classification
 system.

20. BEC
 has
 2
 parts
 –
 Classification
 and
 Biological
 Response
 Curve.

 Just
 because
 classification

doesn’t
 work
 doesn’t
 mean
 Biological
 Response
 Curve
 will
 not.

21. Have
 a
 number
 of
 good
 classifications
 we
 can
 try
 to
 use
 that
 could
 be
 useful.

SALCC
 –
 Mary
 Davis

22. IHA
 –
 other
 aspects
 to
 define
 ecological
 integrity
 using
 CART
 analysis
 (wide
 use
 of
 this
 method).

23. How
 do
 we
 feel
 about
 the
 national
 approach?
 Mary
 was
 not
 recommending
 the
 national

approach
 since
 southeast
 systems
 differ
 from
 western
 systems
 but
 did
 suggest
 use
 of
 the

Northeast
 approach
 given
 drainage,
 etc
 for
 the
 SE
 region

24. Confusion
 around
 if
 she
 was
 using
 a
 prior
 vs.
 alternative
 prior
 classification;
 she
 said
 some

things
 will
 drop
 out
 naturally
 as
 a
 result
 of
 their
 system.

25. What
 systems
 have
 been
 tested
 with
 biological
 composition?
 Fall
 back
 classification
 may
 be

topoedaphic
 (though
 it
 remains
 untested;
 other
 states
 have
 used
 classifications
 that
 remain

untested
 such
 as
 Michigan’s).

26. Doubts
 on
 getting
 to
 biology
 (to
 determine
 flow
 alteration
 to
 biology).
 Perhaps
 look
 to
 role
 of

habitat…In
 order,
 consider:

1. Is
 unaltered
 flow
 the
 best
 metric?

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 30
 of
 51

2. Next
 Surrogate
 level
 is
 habitat

3. Biology/be
 great
 to
 find
 linkages

4. Literature

In
 Lieu
 of
 Classification
 Approaches

 -­‐edits
 were
 made
 to
 the
 following
 comment
 to
 ensure
 it

accurately
 conveyed
 the
 author’s
 intent.
 Item
 #20
 (April
 16,
 2013)

If
 we
 get
 to
 a
 specific
 point
 on
 a
 graph
 as
 a
 recommendation,
 we
 need
 to
 provide
 justification
 for

reaching
 that
 conclusion.

 For
 example,
 if
 the
 approach
 uses
 PHabsim,
 I
 would
 need
 to
 see
 validation

that
 the
 habitat
 modeling
 approach
 actually
 reflects
 what
 is
 going
 on
 [with
 the
 biota].

 That’s
 what
 I

need
 to
 get
 some
 comfort
 with
 recommending
 a
 discrete
 flow
 boundary.

Develop
 Work
 Team:
 Fred
 Tarver,
 Sam
 Pearsall,
 Chris
 Goudreau,
 Amy
 Pickle,
 facilitation
 team
 where

possible

2. Organize

3. Rewrite

4. Circulate
 before
 the
 May
 14
 EFSAB
 meeting

VI.

 Introduction
 of
 Trial
 Balloons

The
 facilitator
 next
 invited
 the
 presentation
 of
 several
 trial
 balloons
 that
 some
 EFSAB
 members
 have

proposed:

Eco-­‐difference
 Trial
 Balloon,
 presented
 by
 Chris
 Goudreau

This
 is
 a
 concept
 that
 I
 think
 can
 be
 applied
 whether
 we
 have
 classes
 or
 not.

 You
 might
 tweak
 it,
 if
 we

do
 have
 classes.

 For
 certain
 classes,
 you
 could
 make
 some
 adjustments
 to
 some
 numbers
 in
 it
 and

different
 for
 other
 classes,
 so
 it’s
 not
 prescriptive.

 It’s
 really
 for
 the
 basic
 planning
 tool.

 So,
 what
 I
 did

was
 look
 at
 the
 Vogel
 paper
 again,
 which
 describes
 this
 concept
 of
 eco-­‐deficit,
 eco-­‐surplus.

 RTI
 had

also
 been
 using
 this
 in
 their
 internal
 research.
 Essentially
 what
 you
 do
 is
 for
 a
 point
 on
 a
 river,
 you

create
 flow
 duration
 curves
 for
 multiple
 years
 and
 then
 take
 the
 median
 of
 those
 curves.

 You
 do
 that

for
 an
 unaltered
 condition
 and
 for
 an
 altered
 condition
 and
 then
 the
 difference
 between
 those
 curves

is
 what
 is
 either
 called
 eco-­‐surplus
 or
 eco-­‐deficit.

 If
 you
 have
 less
 flow
 in
 this
 river,
 in
 an
 altered

condition,
 then
 that
 would
 be
 an
 eco-­‐deficit.

 So
 it’s
 really
 this
 median
 of
 a
 flow
 deficit
 or
 surplus.

 This

really
 is
 just
 dealing
 with
 this
 relatively
 simple
 concept
 focusing
 on
 flow
 magnitude.

 So,
 the
 reason

why
 this
 came
 up,
 Vogel
 looked
 at
 this,
 was
 when
 they
 looked
 at
 all
 the
 IHA
 statistics
 and
 all
 this
 stuff

was
 correlated,
 they
 asked
 which
 one
 is
 more
 important?
 This
 was
 kind
 of
 a
 simplified
 metric
 to
 do

away
 with
 all
 that
 correlation.

 It
 focuses
 on
 magnitude
 and
 if
 you
 shorten
 the
 time
 period,
 you
 can
 do

this
 on
 an
 annual
 basis
 or
 on
 a
 monthly
 basis.

 You
 do
 get
 some
 seasonality
 or
 timing
 in
 it
 if
 you
 do
 it
 on

a
 shorter
 time
 frame,
 but
 it
 doesn’t
 really
 get
 at
 things
 like
 duration
 or
 frequency
 aspects
 of
 all
 your

flow
 statistics,
 so
 just
 keep
 that
 in
 mind.

So
 here’s
 an
 example
 of
 an
 altered
 flow
 curve
 and
 an
 unaltered
 curve
 and
 the
 difference
 between

them.

 On
 the
 top
 part
 of
 that
 flow
 duration
 curve
 would
 be
 eco-­‐deficit
 and
 the
 bottom
 part
 is
 eco-­‐
Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 31
 of
 51

surplus.

 You
 can
 have
 an
 overall
 no
 net
 change
 in
 volume,
 but
 you’ve
 shifted
 it
 from
 one
 part
 of
 the

year
 to
 another,
 for
 example,
 so
 you
 still
 have
 deficits
 and
 surpluses.

This
 next
 slide
 is
 an
 example
 of
 a
 reservoir
 where
 you
 would
 have
 lower
 high
 flows
 but
 higher
 low

flows
 and
 so
 you
 shifted
 total
 volume
 around.

 I
 took
 some
 data
 from
 Tar
 River
 at
 Rocky
 Mount;
 there’s

real
 data
 there,
 but
 some
 of
 the
 altered
 flows
 I
 just
 kind
 of
 made
 up
 or
 took
 a
 certain
 month
 out
 of
 the

flow
 record
 to
 give
 you
 an
 example.

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 32
 of
 51

 The
 next
 slide
 is
 another
 example
 where
 you
 have
 a
 flow
 deficit
 all
 year
 long,
 for
 example,
 from
 water

withdrawal.

So
 what
 I’d
 call
 eco
 difference
 is
 really
 the
 absolute
 value
 of
 a
 deficit
 or
 surplus.

 When
 calculated
 on

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 33
 of
 51

an
 overall
 percentage
 basis,
 eco-­‐difference
 gives
 measure
 of
 relative
 change
 from
 the
 unaltered

condition.

 Let’s
 use
 an
 example.

So
 the
 bottom
 curve
 there
 would
 be
 the
 eco-­‐difference
 of
 those
 two
 curves,
 just
 subtracting
 one
 from

the
 other,
 the
 absolute
 value
 of
 it.

 So
 when
 they’re
 right
 on
 top
 of
 each
 other,
 it’s
 zero
 and
 when

they’re
 far
 apart,
 you
 get
 some
 higher
 value.

 When
 you
 do
 that
 on
 a
 percentage
 basis,

(that’s
 what
 that
 orange
 box
 is
 in
 the
 middle),
 you
 see,
 overall,
 22%
 difference
 between
 those
 two

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 34
 of
 51

curves,
 in
 an
 absolute
 sense.

 And
 you
 can
 also
 look
 at
 it
 as
 which
 parts
 of
 the
 curve
 are
 the
 most

divergent,
 which
 is
 the
 green
 line.

The
 other
 thing
 I
 was
 thinking
 about
 is
 Richter’s
 paper
 about
 sustainability
 boundaries—which
 includes

the
 presumptive
 standard
 discussion
 as
 well.

 Richter’s
 paper
 discusses
 deviation
 from
 inflow
 on
 a

daily
 basis,
 suggesting
 that

 ±11-­‐20%
 provides
 moderate
 protection,
 while
 ±0-­‐10%
 provides
 high

protection.
 The
 paper
 talks
 about
 how,
 as
 long
 as
 you’re
 not
 too
 far
 away
 from
 that
 central
 value,

you’re
 okay.

 They
 talked
 about
 it
 on
 a
 daily
 flow
 basis,
 so
 you
 withdraw
 10%,
 that’s
 okay.

 We
 talked

about
 it
 here,
 concluding
 that
 10
 or
 20
 percent
 withdrawal
 on
 a
 daily
 basis
 would
 be
 an
 acceptable

approach.

 And
 then,
 if
 you
 wanted
 to
 be
 more
 conservative,
 you
 could
 have
 less
 than
 10%
 or
 10-­‐20%.

This
 approach
 maintains
 the
 flow
 hydrology
 curve.

So
 I
 was
 thinking
 about
 how
 to
 combine
 these
 two
 things.

 I’m
 not
 breaking
 new
 ground
 here.
 For
 the

Susquehanna
 study
 part
 of
 their
 recommendations
 recommend
 the
 same
 kind
 of
 deal.

 I
 am
 suggesting

applying
 those
 sustainability
 boundaries
 to
 the
 eco
 difference
 curves.

 So
 I
 have
 some
 examples.

 First
 I

just
 chose
 15%
 just
 to
 show
 you
 how
 it
 looks,
 but
 that
 would
 be
 one
 of
 those
 decisions
 we
 would
 need

to
 make:
 what
 should
 that
 number
 be?

So
 here
 is
 a
 flow
 duration
 curve
 that
 has
 a
 15%
 plus
 and
 15%
 minus
 boundary
 around
 it.

For
 each
 point
 on
 here,
 since
 this
 big
 curve
 is
 made
 up
 of—I
 forget
 how
 many
 years
 I
 did,
 say
 60
 years

of
 data—this
 width
 here
 is
 derived
 from
 the
 variation
 in
 the
 data
 itself.

 It’s
 not
 just
 plus
 or
 minus
 15%

of
 the
 flow
 value,
 it’s
 based
 on
 whatever
 the
 cumulative
 flow
 duration
 curves
 you
 have
 in
 your
 data

set.

 So
 when
 I
 was
 thinking
 about
 how
 it
 would
 be
 applied,
 say
 by
 DWR,
 is
 if
 you
 compare
 a
 new
 flow

regime
 proposal
 to
 the
 existing
 condition,
 one
 way
 would
 be
 if
 the
 eco-­‐difference
 is
 15%
 then
 it
 would

indicate
 that
 the
 proposed
 withdrawal
 might
 compromise
 the
 ecological
 flow.

 And
 so
 here’s
 an

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 35
 of
 51

example
 where
 the
 overall
 flow
 difference
 is
 22%.

 That
 would
 be
 one
 way
 you
 could
 apply

this.

 The
 computer
 model
 says
 we’ve
 got
 a
 problem
 at
 this
 point
 in
 the
 model.

Another
 way
 to
 do
 it
 is
 to
 say
 that
 if
 a
 certain
 part
 of
 the
 curve
 is
 outside
 of
 that
 15%
 band,
 then
 that

would
 indicate
 that
 the
 proposed
 withdrawal
 might
 compromise
 the
 ecological
 flow.

 This
 approach
 is

actually
 more
 stringent
 because
 only
 one
 part
 of
 it
 has
 to
 be
 more
 than
 15%
 out,
 not
 the
 overall
 total

being
 more
 than
 15%.

 So
 in
 this
 case
 here,
 the
 altered
 flow
 is
 the
 green
 curve
 and
 it
 lies
 within
 the

bounds
 so
 the
 withdrawal
 doesn’t
 trip
 the
 flag.

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 36
 of
 51

In
 this
 next
 example,
 it
 would.

 Part
 of
 it
 is
 in,
 but
 part
 of
 it
 is
 out.

 And
 then
 part
 of
 it
 is
 back
 in.

 So
 this

would
 be
 a
 more
 stringent
 application
 of
 it,that
 would
 trip
 the
 flag.

Questions
 (Q),
 Comments
 (C),
 Response
 (R):

Q:

 One
 observation
 and
 one
 question.

 The
 observation
 is
 that
 it’s
 important
 to
 remember
 that
 you’re

not
 weighing
 surplus
 and
 deficit
 differently.

 So
 surplus
 and
 deficit
 count
 the
 same,
 even
 though
 one
 of

them
 may
 be
 much
 more
 troublesome
 than
 the
 other
 for
 any
 given
 stream.

 The
 question
 is
 when
 you

measure
 a
 current
 flow
 regime
 against
 the
 70-­‐year
 average,
 one
 of
 the
 things
 you’re
 always
 going
 to

find
 is
 that
 there’s
 some
 eco-­‐difference
 that
 is
 entirely
 attributable
 to
 climate
 change
 or
 some
 other

natural
 evolution
 of
 the
 system.

 For
 example,
 the
 Conservancy
 for
 many,
 many
 years
 liked
 to
 show

pre-­‐dam
 and
 post-­‐dam
 flow
 regimes
 on
 the
 Roanoke,
 and
 as
 it
 happened
 when
 the
 dams
 were
 built,

the
 climate
 changed.

R:

 Yes,
 but
 the
 way
 I
 would
 think
 this
 would
 work,
 though,
 is
 when
 they
 run
 it
 through
 the
 model,

those
 conditions
 would
 be
 the
 same.

 The
 only
 thing
 that
 would
 be
 different
 in
 the
 model
 would
 be
 the

change
 due
 to
 withdrawals,
 not
 changes
 due
 to
 climate.

 It’s
 modeled.

 You
 could
 account
 for
 that.

 In

other
 words,
 those
 curves
 would
 be
 derived—if
 you
 were
 going
 60
 years
 of
 unaltered,
 then
 you
 would

run
 60
 years
 of
 altered
 with
 everything
 else
 being
 the
 same.

R:

 Got
 it.

 Okay.

 That’s
 how
 to
 solve
 that
 problem.

Q:

 So
 you’re
 using
 this
 to
 guide
 a
 use
 decision
 on
 an
 impoundment
 or
 something
 like
 that,
 not

necessarily
 long-­‐term
 changes,
 but
 maybe
 incremental
 or
 slightly
 incremental
 changes
 that
 over
 time

add
 up
 to
 a
 substantial
 eco-­‐deficit?

R:

 Well,
 it
 would
 all
 go
 down
 to
 how
 DWR
 uses
 the
 models.

 The
 way
 I
 would
 understand
 it,
 for

example
 like
 using
 OASIS,
 you
 would
 run
 a
 period
 of
 record
 under
 unaltered
 and
 then
 a
 period
 of

record
 under
 altered,
 with
 whatever
 those
 total
 alterations
 are,
 30
 and
 50
 years
 out,
 whatever
 the

time
 step—whatever
 the
 projected
 future
 conditions
 are.

 So
 they
 would
 incorporate
 cumulative

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 37
 of
 51

impacts
 to
 that
 point
 in
 the
 model.

C:

 You
 can
 model
 any
 future
 conditions
 over
 the
 full
 period
 of
 record
 and
 come
 up
 with
 a
 difference.

R:

 And
 you’re
 looking
 at
 what’s
 the
 central
 difference.

Q:

 Can
 you
 talk
 a
 little
 bit
 more
 about
 the
 importance
 of
 not
 looking
 at
 seasonality
 versus
 looking
 at

seasonality?

 Because
 one
 of
 the
 concerns
 I
 would
 add
 is
 that
 that
 critical
 period
 maybe
 in
 September.

R:

 Yes,
 the
 example
 I
 have
 up
 here
 is
 January,
 so
 I
 would
 think
 you
 would
 want
 to
 do
 it,
 say,
 like
 on
 a

monthly
 or
 at
 least
 on
 a
 seasonal
 basis,
 not
 on
 an
 annual
 basis.

 Because
 then
 you
 would
 pick
 up—
okay,
 well
 it
 looks
 good
 in
 January
 through
 May
 but
 oh,
 we’ve
 got
 this
 problem
 in
 October
 or

something.

 And
 it
 would
 tell
 you
 what
 kind
 of
 a
 problem
 you’ve
 got.

 Is
 it
 a
 surplus,
 is
 it
 a
 deficit,
 is
 it
 a

certain
 part
 of
 the
 flow
 duration
 curve?

 If
 the
 problem
 is
 in
 the
 high
 flows,
 that
 might
 lead
 you
 to
 one

kind
 of
 recommendation
 and
 if
 the
 problem
 is
 down
 here,
 in
 the
 low
 flows,
 it
 might
 need
 another
 kind

of
 recommendation.

C:

 This
 is
 a
 very
 sophisticated
 strategy
 for
 combining
 a
 presumptive
 standard
 with
 the
 eco-­‐difference

concept,
 and
 I
 really
 like
 it.

 The
 problem
 is
 that
 we’ll
 have
 the
 same
 problem
 with
 this
 as
 we
 would

with
 any
 presumptive
 standard,
 and
 that
 is
 making
 the
 case
 that
 a
 presumptive
 standard
 answers
 the

legal
 question
 of
 how
 the
 ecosystem
 will
 react,
 respond.

R:

 And
 I
 think
 it
 goes
 back
 to
 some
 of
 what
 Tom’s
 question
 was,
 when
 we
 were
 going
 through
 the

exercises,
 what
 is
 that
 link?

 And
 it
 might
 be
 that
 the
 answer
 is
 there
 are
 linkages
 in
 the
 literature,
 but

we
 don’t
 have
 anything
 specific
 to
 this
 particular
 case
 and
 that’s
 where
 we
 have
 to
 hang
 our
 hat.

Q:

 But
 as
 part
 of
 the
 BEC,
 is
 this
 one
 of
 the
 metrics
 that
 RTI
 is
 looking
 at
 in
 terms
 of
 trying
 to
 correlate

biological
 condition
 to
 degree
 of
 hydrologic
 alteration
 information.

 I
 don’t
 know
 if
 eco-­‐change
 is
 one

that
 you
 guys
 had
 looked
 at
 yet.

R:

 We’ve
 calculated
 it
 but,
 yes
 we’re
 using
 it.

 We
 have
 eco-­‐surplus,
 eco-­‐deficit,
 eco-­‐change,
 and
 the

eco-­‐deficits
 and
 the
 eco-­‐surpluses
 are
 expressed
 on
 an
 annual
 basis
 and
 on
 a
 seasonal
 basis
 for
 the

four
 seasons
 that
 you
 used
 for
 PHABSim.

R:

 So
 I
 think
 that
 we’ll
 hopefully
 know.

C:

 So
 one
 recommendation
 with
 this—going
 with
 this
 method—is
 as
 well
 as
 having
 the
 sustainable

boundary,
 you
 might
 also
 want
 to
 have
 a
 lower
 threshold.

 Say
 for
 example
 you
 have
 a
 particularly
 low

flow
 year,
 and
 that’s
 something
 that
 could
 take
 you
 down
 close
 to
 and
 below
 your
 10th
 percentile

flow.

 So
 it’s
 often
 important
 to
 put
 some
 sort
 of
 low
 flow
 threshold
 when
 using
 this
 approach
 that,

you
 know,
 you
 want
 to
 be
 careful
 not
 to
 fall
 below.

 And
 then,
 often
 times,
 what
 you
 can
 do
 as
 well
 as

looking
 at
 the
 percent
 eco-­‐deficit,
 you
 can
 translate
 that
 back
 out
 to
 the
 number
 of
 days
 that
 during
 a

low
 flow
 year
 or
 low
 flow
 month,
 that
 water
 wouldn’t
 be
 available
 to
 be
 withdrawn
 to
 get
 an
 idea
 of

what
 kind
 of
 limits
 might
 be
 applied
 to
 different
 permits.

 So,
 if
 you
 have
 a
 hydrograph
 that
 is
 of
 a

higher
 flow
 year
 than
 that
 year,
 those
 months,
 you
 might
 not
 have
 any
 days
 where
 you
 have
 a

restriction,
 but
 if
 we
 happen
 to
 go
 into
 a
 drought
 year,
 then
 you
 might
 be
 able
 to
 predict
 if
 we’re
 in
 a

drought
 year,
 then
 yeah,
 there
 might
 be
 10
 days
 in
 August
 where
 withdrawals
 would
 be
 problematic.

So
 I
 think
 it’s
 important
 to
 try
 to
 look
 at
 the
 boundaries
 of
 high
 flow
 years
 and
 drought
 years
 with
 the

median
 approach.

R:

 And
 so
 again,
 as
 a
 planning
 tool,
 this
 would
 just
 say
 there’s
 a
 problem
 and
 then
 what
 do
 you
 do

about
 it?

C:

 I
 was
 saying
 everything
 Fred
 was
 going
 to
 say—we
 need
 to
 not
 drift
 into
 permitting
 language.

 But

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 38
 of
 51

as
 a
 tool
 for
 identifying
 when
 ecological
 integrity
 is
 at
 risk,
 this
 looks
 like
 a
 useful
 tool.

Q:

 The
 flow
 for
 the
 Tar
 River
 at
 Rocky
 Mount
 is
 pretty
 substantial;
 you
 are
 above
 1,000
 CFS
 a

considerable
 part
 of
 the
 year.

 But
 if
 you
 were
 working
 on
 a
 much
 smaller
 system,
 would
 you
 use
 the

same
 boundary?

 I’m
 just
 wondering
 if
 that
 15%
 at
 a
 stream
 that
 never
 exceeds
 1,000
 CFS
 might
 be

more
 ecologically
 significant.

R:

 And
 I
 think
 that’s
 really
 where
 the
 discussion
 would
 have
 to
 happen
 here-­‐-­‐is
 that
 percentage.

 If
 we

don’t
 have
 empirical
 data
 to
 help
 us
 set
 that,
 then
 it’s
 going
 to
 be
 literature
 and
 professional
 opinion

and
 do
 you
 make
 adjustments
 for
 the
 things
 like
 Kimberly’s
 talking
 about
 or
 you’re
 talking
 about.

 This

isn’t
 a
 silver
 bullet,
 it’s
 just
 an
 approach
 that
 still
 requires
 some
 thoughtful
 input
 on
 what
 those

numbers
 are.

Q:

 What
 if
 you
 relate
 biology
 and
 you
 find
 out
 that
 the
 biology
 itself
 is
 plus
 or
 minus
 30%,
 which
 would

be
 pretty
 low,
 actually.

 How
 do
 you
 handle
 the
 risk,
 then?
 Because
 you’re
 going
 to
 have
 to
 have
 it

much
 higher
 than
 your
 lower
 level
 there
 and
 a
 less
 than
 50%
 chance
 that
 you’re
 actually
 going
 to
 crash

the
 system.

 So
 how
 do
 we,
 I
 mean,
 how
 do
 you
 use
 this
 without
 being
 able
 to
 incorporate
 into
 it
 some

estimate
 of
 the
 variability
 in
 the
 biological
 response?

R:

 I
 think
 it’s
 the
 same
 question
 that
 could
 be
 asked
 about
 everything
 that
 we’re
 talking
 about.

 We’re

struggling
 with
 what
 that
 is.

R:

 I
 think
 we’re
 fortunate
 here
 in
 that
 we
 have
 a
 number
 of
 biological
 sites
 where
 we
 can
 address
 that.

R:

 I
 think
 for
 certain
 situations
 and
 systems,
 we
 can
 probably
 have
 a
 better
 guess
 at
 that.

R:

 You
 know
 that
 variability’s
 going
 to
 be
 enormous.

R:

 Yes,
 and
 it
 goes
 back
 to
 what
 I
 was
 saying
 earlier
 is
 how
 much
 do
 we
 trust
 in
 kind
 of
 the
 underlying

paradigm
 that
 flow
 is
 the
 master
 variable,
 that
 we
 just
 trust
 it
 as
 long
 as
 we’re
 keeping—and
 that’s

kind
 of
 what
 Richter’s
 papers
 are
 all
 about,
 that
 sustainability
 boundaries
 thing.

 Not
 trying
 to
 dig
 too

far
 down
 into
 what
 really
 happens
 with
 the
 biology—it’s
 just
 under
 the
 assumption
 that
 if
 you
 maintain

the
 flows,
 you
 maintain
 the
 biology.

R:

 So
 I
 see
 it
 kind
 of
 like
 as
 a
 recipe.

 You
 have
 this
 recipe
 for
 ecological
 integrity,
 right?

 And
 you
 need

to
 know
 how
 much
 water
 goes
 into
 that
 recipe.

 There
 are
 a
 lot
 of
 other
 ingredients
 to
 ecological

integrity,
 but
 if
 you
 get
 the
 water
 right,
 you
 may
 or
 may
 not
 get
 ecological
 integrity
 because
 you
 might

use
 salt
 instead
 of
 sugar.

 We
 are
 not
 measuring
 that,
 nobody
 is
 measuring
 whether
 we’re—you
 know,

determining
 whether
 we’re
 using
 salt
 or
 sugar
 in
 the
 recipe.

 But
 we
 do
 want
 to
 try
 to
 make
 sure
 we

get
 that
 quantity
 of
 water
 right.

 And
 so,
 if
 we
 are
 constantly
 assuming—if
 we’re
 saying
 we’re
 not

getting
 it
 right
 if
 the
 recipe—if
 the
 integrity
 doesn’t
 come
 out
 right,
 that’s
 not
 the
 way
 to
 measure
 it.

You
 can’t
 measure
 whether
 you’re
 getting
 the
 water
 right
 by
 whether
 you’re
 getting
 ecological

integrity
 right
 because
 there
 are
 too
 many
 other
 ingredients
 in
 the
 recipe.

 But
 how
 do
 we
 come
 back

to—do
 we
 have
 enough
 water
 going
 in.

 And
 I
 think
 that’s
 where
 Richter’s
 saying
 that
 if
 you’re
 getting

as
 close
 as
 you
 can
 to
 this
 historic,
 natural,
 unaltered
 flow
 then
 the
 odds
 are
 better
 that
 you’re
 getting

enough
 in.

 That’s
 really
 all
 they’re
 ever
 going
 to
 be
 able
 to
 achieve.

 So,
 I
 mean
 I
 just
 feel
 like
 if
 we

hang
 it
 on
 whether
 we
 can
 catch
 that
 biological
 or
 ecological
 integrity
 piece,
 we’ll
 undercut
 our
 ability

to
 actually
 use
 flow.

C:

 Well,
 the
 law
 says
 that
 we’re
 supposed
 to
 advise
 DENR
 on
 how
 to
 determine
 how
 much
 water
 is

required
 to
 maintain
 ecological
 integrity.

 I
 understand
 that
 there
 are
 lots
 of
 other
 things
 required
 to

maintain
 ecological
 integrity,
 but
 it
 seems
 to
 me
 that
 we
 have
 to
 address
 the
 question
 in
 the
 law.

 If

we’re
 unable
 to
 connect
 how
 much
 water
 there
 is
 to
 ecological
 integrity
 because
 the
 variables
 are
 too

complex,
 then
 we
 just
 are
 going
 to
 have
 to
 report
 that
 out
 and
 we
 will
 have
 failed
 with
 our
 mission.

 I

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 39
 of
 51

hope
 that
 we
 can,
 in
 fact,
 come
 up
 with
 some
 answer,
 however
 rough,
 for
 how
 much
 water
 is
 required

to
 maintain
 ecological
 integrity.

 Am
 I
 misunderstanding
 your
 concern?

R:

 My
 concern
 is
 when
 we
 come
 back
 to
 whether
 or
 not
 we
 have
 a
 reference
 situation
 at
 Tar
 River
 in

Rocky
 Mount
 or
 that
 with
 the
 adaptive
 management
 that
 we
 go
 out
 and
 measure
 and
 validate
 whether

we
 got
 our
 guess
 or
 our
 well-­‐reasoned
 flow
 correct,
 we
 don’t
 have
 any
 control
 over
 how
 many
 NPDES

permits
 are
 going
 in
 at
 that
 point,
 that
 will
 also
 influence
 ecological
 integrity.

 So,
 I
 feel
 like
 that
 the

most
 straight
 forward
 approach
 is
 to
 say,
 how
 close
 are
 we
 to
 the
 unaltered
 flow?

 And
 the
 closer
 we

can
 achieve
 that,
 that’s
 the
 whole
 contribution
 that
 DWR
 can
 make
 toward
 ecological
 integrity.

R:

 I
 think
 what
 we’re
 talking
 about
 is
 something
 that’s
 been
 going
 through
 my
 mind
 for
 a
 while.

 You

look
 at
 the
 charge
 in
 the
 statute;
 it
 is
 what
 it
 is,
 but
 my
 response
 to
 that
 would
 be,
 in
 whatever
 we’re

writing
 up,
 say
 yes,
 but
 we
 really
 can’t
 answer
 that
 head-­‐on
 for
 the
 reasons
 you’re
 pointing
 out
 and
 the

best
 we
 can
 do—we
 tried
 to
 find
 these
 links,
 we
 went
 way
 beyond
 any
 other
 state
 that’s
 ever
 done
 this

to
 try
 to
 find
 those
 links,
 and
 here
 are
 the
 few
 links
 that
 we
 found
 and
 the
 other
 ones
 we
 don’t
 and
 so,

we
 have
 to
 just
 step
 back
 and
 use
 the
 assumption,
 the
 ecological
 theory
 that
 this
 is
 what
 is
 driving
 the

system
 and
 so
 on
 and
 so
 forth
 and
 make
 the
 justification
 that
 way.

 It
 doesn’t
 say
 that—I
 don’t
 say
 we

would
 fail,
 it’s
 just
 that’s
 where
 we’re
 at
 now.

R:

 If
 we
 go
 back
 to
 the
 department
 and
 say
 our
 best
 shot
 at
 the
 answer
 for
 your
 question
 is
 stick
 to

natural
 flows,
 we’ve
 probably
 failed.

 If
 we
 go
 back
 to
 the
 department
 and
 say
 our
 best
 shot
 is
 stick

within
 this
 range
 around
 the
 natural
 flows
 for
 the
 following
 reasons,
 that’s
 not
 failure.

 If
 we
 can
 go

back
 to
 the
 department
 and
 say
 stick
 within
 this
 range
 around
 the
 natural
 flow
 because
 here’s
 how
 the

ecological
 integrity
 of
 the
 system
 will
 respond
 if
 you
 don’t,
 then
 we
 have
 succeeded
 in
 a
 very
 big
 way.

While
 I
 think
 that
 we’re
 going
 to
 come
 in
 somewhere
 between
 B
 and
 C,
 I
 sincerely
 hope
 we
 come
 in

somewhere
 B
 and
 C,
 I
 hope
 we
 don’t
 go
 back
 with
 A.

R:

 I
 would
 expect
 that
 we
 would
 be
 able
 to
 show
 that
 there
 are
 relationships.

 They
 are
 not
 going
 to
 be

tight,
 but
 I
 don’t
 think
 you
 need
 them
 to
 be
 tight
 if
 you
 can
 show
 that.

 The
 first
 thing
 to
 show
 is
 that

this
 ecological
 difference
 is
 related
 to
 biological
 responses
 and
 what
 components
 of
 biological

responses
 are
 they
 related
 to
 and
 if
 they’re
 ecologically
 significant,
 then
 that’s
 building
 the
 justification

for
 that.

 It’s
 not
 going
 to
 be
 perfect
 relationships.

 You’re
 not
 going
 to
 have
 real
 tight
 bands
 around
 it,

but
 you
 will
 be
 able
 to
 say
 the
 probability
 of
 getting
 a
 good
 community
 is
 decreasing
 as
 this
 eco-­‐deficit

is
 increasing.

R:

 Of
 the
 work
 that
 I’ve
 seen
 so
 far
 from
 RTI
 there
 are
 some
 relationships
 there,
 they’re
 going
 to
 be
 for

certain
 streams,
 certainly
 related
 to
 or
 are
 constrained
 by
 the
 data
 sets
 that
 we
 have-­‐-­‐wadeable

streams
 that
 have
 either
 bug
 or
 fish
 data.

 So,
 outside
 of
 that,
 we
 still
 have
 a
 black
 box
 and
 so
 we
 might

see
 what
 Sam’s
 talking
 about
 for
 certain
 streams,
 and
 the
 other
 ones
 it’s
 going
 to
 be
 more
 generic.

 But

I
 think
 this
 kind
 of
 approach,
 and
 other
 approaches,
 would
 be
 useful
 in
 either
 case.

C:

 I
 like
 it.

After
 the
 facilitator
 prompted
 for
 further
 questions
 and
 there
 were
 none,
 the
 facilitator
 asked
 if
 the

Board
 wanted
 to
 add
 any
 recommendations
 to
 the
 list
 as
 a
 result
 of
 this
 presentation.

C:

 When
 I
 listened
 to
 Chris
 talk
 and
 look
 at
 his
 presentation,
 I
 see
 no
 conflict
 at
 all
 between
 what
 he

has
 proposed
 and
 what
 I’m
 about
 to
 propose.

 As
 a
 matter
 of
 fact,
 I
 see
 them
 as
 potentially
 merging

into
 some
 common
 system.

 I
 think
 we
 keep
 it
 on
 the
 table.

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 40
 of
 51

Q:

 Just
 a
 point
 of
 clarification
 for
 me
 then
 is
 are
 the
 things
 we’re
 keeping
 on
 the
 table
 going
 to
 be
 part

of
 that
 one
 recommendations
 document
 that
 Sam
 et
 al
 are
 sort
 of
 shrinking
 down?

 If
 so,
 then
 yes,
 I

agree
 that
 this
 should
 be
 a
 part
 of
 that.

 Or
 are
 we
 having
 other
 mechanisms
 for
 keeping
 things
 on
 the

table
 and
 that
 whole
 generation
 of
 recommendations
 lists
 is
 one
 thing,
 and
 there’s
 this
 trial
 balloon,

and
 there’s
 another
 trial
 balloon,
 and
 then
 we
 consider
 those,
 or
 what?

Facilitator:

 That’s
 sort
 of
 what
 I’m
 asking
 you
 all,
 as
 a
 group,
 how
 you
 want
 to
 approach
 this.

 I
 think

one
 way
 to
 go
 is
 to
 add
 proposals
 to
 this
 list
 that
 essentially
 are
 staying
 in
 front
 of
 the
 group.

 And
 then

the
 Board
 can
 hone
 down
 from
 that
 list;
 that’s
 what
 I
 understood
 the
 group
 wanted
 to
 do.

 So,
 that’s

essentially
 what
 I’m
 asking
 is
 what
 you
 want
 to
 do,
 how
 do
 you
 want
 to
 work
 this
 into
 where
 we
 are?

R:

 I
 think
 at
 some
 point
 this
 is
 going
 to
 devolve
 to
 a
 writing
 committee.

 When
 I
 do
 my
 presentation
 on

my
 trial
 balloon,
 I’m
 going
 to
 show
 how
 Chris’
 trial
 balloon
 fits
 right
 in.

 I’m
 going
 to
 show
 how
 Fred’s

happy
 face
 number
 one
 fits
 right
 in
 and
 it
 seems
 to
 me
 that
 what
 we’re
 actually
 doing
 with
 these
 trial

balloons
 is
 framing
 the
 solution
 in
 different
 ways
 but
 I
 think
 we’re
 all—I
 think
 we’re
 actually
 working

toward
 what
 is
 appearing
 to
 be
 more
 and
 more
 like
 a
 single
 solution.

Facilitator:

 That
 is
 the
 value
 of
 the
 trial
 balloons…

C:

 I
 mean,
 I
 think
 what
 these
 trial
 balloons
 are
 going
 to
 represent
 is
 a
 suite
 of
 recommendations
 that

are
 presented
 in
 this
 format
 but
 maybe
 we
 could
 have
 you
 present
 them
 in
 the
 type
 of
 format
 that

we’re
 going
 to
 go
 after
 with
 that
 work
 group.

 And
 I
 guess
 it
 would
 be—and
 that
 way
 it
 could
 be

included
 in
 that.

 But
 these
 are
 clearly
 like,
 not
 only
 having
 them
 individually
 laid
 out,
 like
 sort
 of

separated—this
 is
 actually
 in
 a
 total
 package
 direction.

 So
 I
 mean
 I
 just
 think—I
 see
 that
 they
 need
 to

be
 integrated
 somehow.

R:

 What
 I
 see
 is
 that
 that
 other
 big
 long
 list
 of
 stuff
 is
 the
 framework
 within
 which
 a
 specific
 trial

balloon
 needs
 to
 fit
 within.

 If
 it
 meets
 all
 those
 other
 things
 on
 our
 list,
 then,
 it
 can
 move
 forward.

Q:

 So
 would
 you
 have
 to
 break
 this
 apart
 into
 those
 component
 pieces?

R:

 Oh
 God.

R:

 I
 know;
 it’s
 sort
 of
 fully
 developed,
 or
 at
 least,
 you
 know,
 pretty
 far
 developed
 into
 that.

R:

 I
 think
 as
 long
 as
 we
 can
 show
 that
 it
 meets
 the
 framework,
 I
 don’t
 know
 that
 we
 need
 to
 tear
 it

apart.

Q:

 Presenting
 the
 flow
 data,
 you
 give
 it
 in
 a
 flow
 duration
 curve.

 Richter
 presented
 his
 in
 an
 annual

hydrograph,
 I
 don’t
 know
 if—

R:

 Well
 the
 Vogel
 paper
 talks
 about
 eco-­‐deficits;
 Richter’s
 uses
 the
 flow
 duration
 curves.

C:

 So
 I
 guess
 one
 uses
 flow
 duration
 and
 one
 uses
 annual
 hydrographs,
 I
 don’t
 know
 which
 would
 be—

R:

 Well
 because
 Richter’s
 approach
 is
 on
 a
 daily
 time
 step,
 you
 can
 take,
 as
 long
 as
 you’re
 within
 plus

or
 minus
 10
 or
 20
 percent
 each
 day.

 So
 it
 really
 goes
 back
 to
 some
 of
 the
 stuff
 that
 you
 were

proposing,
 the
 percentage
 withdrawal
 or
 flow
 by.

 That’s
 really
 what
 his
 thing
 is,
 is
 just
 putting
 a

percentage
 on
 a
 daily
 time
 step;
 that’s
 why
 he
 can
 use
 annual
 hydrographs.

C:

 I
 was
 saying
 that
 it
 could
 be
 presented
 either
 way.

R:

 I’ve
 heard
 Sam
 talk
 previously
 about
 flow
 duration
 curves
 and
 how
 you
 kind
 of
 lose
 some
 of
 the

resolution
 and…

R:

 The
 reason
 why
 you
 do
 it
 this
 way
 is
 so
 you
 can
 do
 the
 calculation.

C:

 An
 annual
 flow
 duration
 curve
 would
 hide
 important
 data,
 but
 you’re
 proposing
 monthly
 flow

duration
 curves.

R:

 Or
 seasonal.

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 41
 of
 51

Facilitator:

 So
 back
 to
 the
 question
 of
 how
 we
 work
 this
 into
 moving
 forward—I’m
 wondering
 if
 maybe

what
 might
 be
 useful,
 Chris,
 would
 be
 to
 put
 together
 the
 questions
 that
 you
 feel
 this
 raises
 and
 need

to
 be
 addressed
 to
 be,
 like
 you
 were
 saying,
 determining
 what
 would
 the
 value
 be
 or
 what
 would
 the

range
 be,
 that
 sort
 of
 thing
 perhaps,
 that
 could
 go
 on
 our
 list
 for
 further
 discussion.

R:

 Can
 we
 also—I
 don’t
 know
 who
 does
 this
 stuff
 but
 I
 mean,
 can
 you
 look
 at—can
 you
 show
 us
 some

of
 the
 smaller
 stream
 systems
 or
 wonder
 what
 those
 implications
 might
 be?

 How
 do
 we
 get
 to
 that

point?

R:

 I
 just
 did
 this-­‐-­‐pulled
 gage
 data
 and
 generated
 spreadsheets.

 I
 mean
 really,
 this
 could
 be
 done
 and
 I

assume
 automated,
 and
 maybe
 someone’s
 already
 done
 it,
 with
 model
 data
 and
 spit
 this
 stuff
 out.

 I

mean,
 that’s
 certainly
 what
 RTI’s
 doing
 so,
 to
 get
 that—it
 took
 me
 awhile
 to
 do
 this.

 I
 mean
 if
 they

could
 spit
 it
 out
 and
 you
 know,
 a
 bunch
 of
 them…

C:

 They’ve
 already
 got
 it.

R:

 But
 do
 we
 want
 to
 see
 that
 for
 then—streams
 that
 we
 have
 PHABSIM
 data
 for?

 I
 mean,
 just
 some

way
 to
 tie
 in
 all
 these
 different
 sets
 of
 data
 so
 that
 we’re
 looking
 at
 apples
 and
 apples.

 I
 just
 have
 a

hard
 time
 with
 a
 lot
 of
 different…

C:

 So
 one
 of
 the
 things,
 Judy,
 with
 the
 analysis
 that
 I’m
 doing
 at
 the
 44
 gages
 is
 really
 similar
 to
 this

instead
 of
 looking
 at
 the
 entire
 curve,
 though,
 I
 just
 pulled
 out
 the
 5
 percentiles
 so
 that
 it
 would

capture
 like,
 changes
 in
 high—the
 75th
 and
 50th,
 the
 25th,
 and
 the
 10th
 -­‐-­‐and
 one
 of
 the
 things
 I’m

looking
 at
 right
 now,
 is
 how
 those
 change
 across
 basins
 and
 with
 different
 drainage
 basin
 areas.

 So

looking
 at
 what,
 for
 example,
 the
 discharge
 is
 at
 the
 10th
 percentile
 flow
 across
 the
 different
 drainage

basin
 areas,
 and
 then
 what
 that—and
 if
 you
 look
 at
 a
 sustainable
 boundary
 around
 that,
 how
 many

million
 gallons
 per
 day
 would
 be
 available
 to
 withdraw
 and
 stay
 within
 a
 sustainable
 boundary.

 And

I’m
 doing
 it
 by
 variation
 by
 drainage
 basin
 and
 on
 the
 monthly
 time
 steps.

 But
 instead
 of
 looking
 at
 the

entire
 curve,
 I’m
 just
 pulling
 out
 5
 of
 the
 percentiles
 from
 the
 curve.

 So
 that
 might
 be
 useful
 and
 I

mean
 that
 stuff
 will
 be
 available
 in
 June.

 But
 it
 looks
 at
 the
 monthly
 variation
 and
 variation
 by
 drainage

basin
 area.

 But
 it’s
 really
 similar,
 it’s
 really
 similar
 to
 that.

 And
 one
 of
 the
 things
 is
 there’s
 different

sustainable
 boundaries
 for
 the
 different
 drainage
 basin
 sizes.

 So
 depending
 on
 how
 much
 water
 is

actually
 available,
 you’ll
 have
 different
 percentages
 around
 that
 sustainable
 boundary
 that
 you
 could

withdraw
 in
 order
 to
 stay
 within
 a
 range
 that
 would
 still
 in
 theory
 protect
 your
 ecological
 integrity.

 But

I’m
 also
 translating
 back
 out
 the
 million
 gallons
 per
 day
 so
 you’ve
 got
 to
 see
 what
 how
 much
 water

would
 be
 available
 as
 opposed
 to
 just
 a
 CFS
 reading.

C:

 Good
 points.

Q:

 It’s
 also
 correct—right,
 Jennifer?—that
 the
 eco-­‐deficit
 stuff
 is
 part
 of
 the
 BEC
 analysis
 for
 ecological

flows?

R:

 Yes.

C:

 We’re
 not
 testing
 different
 scenarios,
 but
 we
 are
 comparing
 or
 looking
 at
 ecological
 responses
 to

these
 deficits.

R:

 And
 I
 was
 saying
 before,
 it
 was
 like
 on
 the
 annual
 basis
 and
 then
 by
 seasonal
 eco-­‐deficits
 in
 addition

to
 try
 to
 also
 [inaudible]
 percentiles,
 as
 well.

Facilitator:

 So
 we’ve
 come
 up
 with
 a
 couple
 of
 things
 to
 discuss
 further:
 1)
 the
 range
 to
 use;
 2)
 perhaps

seeing
 it
 on
 some
 smaller
 streams,
 perhaps
 ones
 that
 also
 have
 PHABSIM.

 And
 then
 what
 I
 would

suggest
 in
 the
 interest
 of
 time
 today,
 is
 if
 you
 have
 additional
 thoughts
 that
 come
 from
 this
 that
 you

would
 like
 to
 have
 added
 to
 the
 list
 for
 consensus
 principles
 or
 identified
 areas
 for
 discussion
 that
 you

e-­‐mail
 Sam,
 or
 us,
 and
 we
 can
 add
 that
 into
 the
 list
 to
 be
 included
 in
 the
 work
 that
 the
 subcommittee

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 42
 of
 51

does.

 Does
 that
 make
 sense?

 Does
 that
 work?
 [group
 members
 nodded,
 and
 there
 were
 no

objections].

 The
 facilitator
 then
 invited
 Sam
 Pearsall
 to
 present
 his
 proposal.

Sam
 Pearsall’s
 Trial
 Balloon

Was
 it
 Herman
 Cain,
 the
 presidential
 candidate
 who
 had
 the
 9-­‐9-­‐9
 plan?

 I
 have
 the
 3-­‐3-­‐3
 plan:

1. Classify
 the
 rivers
 and
 streams
 of
 NC.

 This
 classification
 should
 include,
 at
 the
 absolute

minimum,
 classes
 distinguishing
 hydrologically
 distinct
 physiographic
 regions.
2. For
 each
 class,
 determine
 which
 variables
 are
 the
 most
 important
 (characterize
 the
 class).

 Base

these
 on
 biological
 responses
 to
 variations
 in
 flow.

 Include
 all
 available
 IFIM
 /
 PHabSim
 data.

3. For
 each
 class,
 determine
 what
 the
 ecological
 baseline
 is.

 This
 should
 be
 based
 on
 the
 most

sensitive
 biota
 for
 which
 we
 have
 adequate
 data.

 The
 ecological
 baseline
 should
 consist
 of:
a. A
 subdivision
 of
 the
 year
 into
 relevant
 "seasons."

 These
 will
 vary
 from
 class
 to
 class
b. A
 target
 flow
 for
 each
 season:

 this
 should
 be
 expressed
 as
 a
 %
 of
 inflow
 defined
 at
 the

catchment
 level;
 the
 most
 powerful
 family
 of
 strategies
 demonstrated
 by
 IFIM
 work.

c. A
 band
 of
 allowable
 variation,
 or
 alternatively,
 specification
 about
 frequency,

magnitude,
 and
 duration
 of
 excursions
 from
 this
 base
 flow
 that
 will
 not
 cumulatively

violate
 the
 definition
 of
 ecological
 integrity.

I’m
 keeping
 in
 mind
 as
 well
 as
 I
 can
 our
 mandate
 and
 suggesting
 that
 our
 recommendation
 to
 the

Department
 should
 consist
 of
 three
 parts.

 One
 is
 our
 actual
 recommendation
 to
 the
 Department
 on

how
 it
 should
 go
 about
 deciding
 how
 much
 water
 is
 required
 to
 maintain
 ecological
 integrity
 as
 defined

in
 the
 Act.

 Number
 one
 would
 be
 our
 recommendation.

 Number
 two
 would
 be
 our
 experience,

reporting
 on
 how
 we
 came
 up
 with
 that
 recommendation:

 What
 research
 we
 did,
 what
 strategies
 we

tried,
 we
 can
 talk
 about
 how
 we
 looked
 at
 a
 couple
 of
 purely
 flow
 based
 strategies,
 we
 had
 a
 hard
 time

doing
 biological
 integrity
 with
 those,
 there
 are
 other
 issues
 with
 those.

 Finally
 Part
 3
 would
 pretty

much
 be
 attachments;
 Number
 3
 would
 basically
 be
 a
 bibliography
 including
 all
 the
 presentations
 and

all
 of
 the
 papers
 that
 were
 incorporated
 into
 our
 deliberations
 and
 a
 URL
 so
 that
 a
 reader
 of
 this

document
 can
 get
 to
 each
 of
 them
 if
 they
 chose
 to
 do
 it.

 This
 would
 be
 a
 page
 or
 two
 of
 references,

not
 the
 references
 themselves.

 So
 that’s
 the
 first
 three.

 Within
 that,
 we’ve
 got
 three
 more
 under

number
 one
 and
 these
 are
 the
 same
 three
 things
 that
 were
 on
 Jennifer’s
 what,
 third
 or
 fourth
 slides

this
 morning?

 One
 is
 classify.

 Two
 is
 develop
 flow
 biology
 relationships.

 And
 three
 is
 determine

ecological
 baselines.

 I
 had
 originally
 called
 this
 classification,
 characterization,
 and
 prescription
 but

prescription
 sounds
 a
 little
 bit
 like
 permitting
 so
 I
 thought
 we
 shouldn’t
 use
 that.

 And
 then
 three
 more,

and
 these
 are
 one,
 come
 up
 with
 seasons,
 assuming
 that
 they
 won’t
 be
 the
 same
 for
 every
 class.

Number
 two,
 target
 flows.

 And
 number
 three
 is
 allowable
 variance,
 and
 by
 allowable
 I
 don’t
 mean

legally
 allowable,
 I
 mean
 some
 other
 kind
 of
 allowable.

 And
 that’s
 it.

 That’s
 my
 whole
 trial
 balloon.

The
 thing
 that
 Chris
 was
 proposing
 is
 a
 strategy
 for
 solving
 those
 two
 problems,
 and
 it’s
 entirely

possible
 that
 this
 one
 isn’t
 going
 to
 work
 out.

 I
 still
 have
 high
 hopes
 that
 it
 will.

 And
 Fred’s
 strategy
 is

also
 there.

 So,
 I
 mean
 that’s
 it.

 There’s
 nothing
 particularly
 unusual
 or
 innovative
 or
 remarkable

about
 that—we’ve
 been
 talking
 along
 these
 lines
 for
 a
 couple
 of
 years
 and
 all
 I
 hope
 to
 do
 today
 is
 just

remind
 you
 that
 that
 is
 sort
 of
 along
 the
 lines
 of
 what
 we’ve
 been
 talking
 for
 awhile
 and
 keep
 it
 alive
 as

a
 potential
 outline
 for
 how
 we
 frame
 our
 report.

 End
 of
 presentation.

Q:

 You’ve
 got
 the
 question
 mark
 beside
 “Classify”
 but
 you
 don’t
 have
 one
 beside
 the
 flow
 and

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 43
 of
 51

biological
 relationships,
 and
 that
 seems
 to
 be
 as
 much
 in
 question
 as
 the
 first
 one
 does
 to
 me.

R:

 I
 think
 we’re
 definitely
 seeing
 biological
 responses
 to
 flows.

 Whether
 we’re
 able
 to
 sort
 those
 into

classes
 or
 whether
 we
 have
 to
 treat
 them
 individually,
 remains
 to
 be
 seen.

 But
 we
 know
 and
 the
 data

support
 the
 notion
 that
 organisms
 respond
 to
 flows.

Q:

 So,
 when
 are
 those
 results
 coming?

Q:

 May/June?

R:

 Well,
 as
 the
 result
 of
 an
 additional
 investment
 from
 state
 agencies,
 RTI
 should
 be
 able
 to
 produce

results
 by
 August.

 Now
 if
 we
 can
 get
 some
 preliminary
 results
 or
 partial
 results
 before
 then,
 we
 will.

And
 Tom
 I
 think
 you’re
 chugging
 right
 along
 and
 probably
 will
 have
 some
 updates
 right
 along,
 except
 in

May
 when
 you’ll
 be
 out
 of
 town.

R:

 There
 will
 be
 somebody
 else
 here
 representing
 the
 USGS
 and
 will
 bring
 whatever
 I
 have
 done
 by

that
 time.

Q:

 Okay,
 this
 is
 another
 clarification
 question
 on
 my
 part.

 So
 this
 recommendations
 document
 is
 being

shriveled?

 And
 now
 for
 me
 to
 understand
 is
 going
 to
 be
 a
 set
 of—these
 are
 our
 recommendations
 of

how
 to
 proceed
 and
 then
 the
 trial
 balloons
 are
 examples
 of
 how
 you
 could
 proceed?

 Is
 that
 basically

right?

Facilitator:

 Yes.

R:

 I
 didn’t
 really
 understand
 what
 a
 trial
 balloon
 was
 so
 I
 made
 something
 up
 and
 answered
 it.

R:

 But
 to
 me
 this
 is
 similar
 to
 the
 set
 of
 recommendations
 that
 we
 have
 this
 long
 list
 that
 you’ll
 help

simplify,
 but
 Chris’
 is
 something
 very
 different,
 it’s
 actually
 quantitative,
 it
 fills
 in
 the
 framework.

 And

to
 me
 a
 lot
 of
 these
 recommendations
 were
 just
 that,
 a
 framework,
 not
 really
 a
 recommendation.

 I

would
 just
 say
 this
 is
 a
 methodology,
 or
 a
 process,
 or
 things
 you
 must
 consider
 to
 be
 able
 to
 do

anything
 and
 then
 that’s
 an
 example
 of
 how
 you
 could
 actually
 do
 it.

 Is
 that
 right
 in
 my
 thinking?

C:

 I
 see
 a
 trial
 balloon,
 personally,
 as
 a
 recommendation
 in
 process.

 I
 see
 the
 trial
 balloon
 as

something
 that
 we
 have
 as
 a
 placeholder
 right
 now
 until
 we
 can
 develop
 them
 into
 recommendations,

at
 which
 time
 the
 trial
 balloons
 will
 cease
 to
 be
 trial
 balloons—they’ll
 become
 recommendations.

R:

 Right,
 but
 there’s
 still
 a
 recommendation
 for
 how
 to
 fill
 in
 some
 component
 of
 this
 broader

framework
 that’s
 a
 little
 bit
 more
 quality
 and
 a
 little
 less
 quantity.

Facilitator:

 you
 could
 have
 five
 of
 them
 in
 that
 one
 place.

 Or
 six,
 or
 whatever.

R:

 Okay.

 I
 was
 trying
 to
 sort
 out
 the
 difference
 between
 this
 big
 list
 of
 recommendations
 and

something
 that’s
 much
 more
 specific
 and
 focused
 in
 quantity
 that
 Chris
 presented.

 Sam's
 framework

helps
 me
 understand
 how
 it
 all
 comes
 together.

 It’s
 just
 the
 terms
 that
 we’re
 using,
 calling
 these

recommendations
 and
 those
 trial
 balloons,
 means
 very
 little
 to
 me.

 But
 I’m
 understanding
 now
 what

we
 actually
 mean.

R:

 I
 had
 that
 same
 semantic
 difficulty.

 I
 didn't
 know
 what
 a
 trial
 balloon
 is.

R:

 But
 I’m
 understanding
 what
 they
 are
 now.

C:

 They’re
 floating
 proposals.

C:

 Straw
 men.

R:

 Straw
 men,
 yes.

Sam
 also

 offered
 a
 report
 format
 outline:

 1.
 Preface

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 44
 of
 51

 2.
 Recommendations

 Classifications

 Flow/Biology
 Relationships

 Determine
 Eco-­‐system
 baseline

 3.

 Experience

 What
 the
 EFSAB
 did
 to
 arrive
 at
 these
 recommendations

 What
 was
 tested?
 What
 was
 the
 level
 of
 rigor?

 What
 was
 reviwed?
 How
 did
 the
 review
 contribute
 to
 overall
 understanding?

 4.
 Attachments

 References
 of
 Articles

 Publications

Fred
 Tarver’s
 Trial
 Balloon

The
 facilitator
 then
 invited
 Fred
 Tarver
 to
 present
 his
 trial
 balloons.

 Fred
 presented
 his
 dual
 matrix,

which
 he
 sent
 out
 before
 the
 March
 meeting.

 On
 the
 top
 header
 the
 right
 side
 shows
 the
 potential

ecological
 flow
 methodologies
 we’ve
 been
 looking
 at.

 And
 then
 on
 the
 left
 side
 are
 the

characterizations/classifications
 that
 we’ve
 been
 attempting.

 Down
 the
 center
 are
 some
 of
 the
 flow

strategies
 that
 we’ve
 been
 looking
 at
 when
 Jim
 was
 producing
 the
 charts
 from
 PHabSim.

 When
 you
 try

to
 construct
 a
 flow
 strategy,
 you
 can
 say
 well,
 based
 on
 the
 ecological
 flow
 strategy,
 which
 ever
 way

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 45
 of
 51

you
 choose
 to
 pick,
 whether
 it’s
 a
 yearly
 flow
 strategy,
 a
 seasonal,
 or
 a
 monthly
 flow
 strategy,
 you
 can

come
 up
 with—is
 it
 a
 percent
 inflow,
 a
 percent
 of
 yearly
 average
 flow
 or
 one
 of
 the
 more
 traditional

flow-­‐by
 requirements.

 These
 are
 all
 based
 on
 PHABSIM
 in-­‐stream
 flow
 studies
 and
 there’s
 a
 plethora

of
 papers
 evaluating
 the
 IFIM
 process,
 so
 we’re
 actually
 looking
 at
 responses
 to
 flows
 but
 this
 is

habitat.

 So
 you
 have
 to
 make
 that
 leap
 of
 faith
 between
 habitat
 and
 the
 biota;
 when
 you
 buy
 into
 this

process,
 you
 kind
 of
 buy
 into
 the
 fact
 that
 if
 you
 create
 the
 habitat,
 the
 biota
 will
 be
 there.

 If
 you
 look

at
 these
 habitat
 responses,
 obviously
 as
 you
 move
 to
 the
 left
 side
 of
 these
 three
 suites
 of
 flow

requirements,
 you
 have
 less
 loss
 of
 habitat
 with
 monthly
 median
 as
 opposed
 to
 the
 traditional

7Q10,
 for
 example.

 You
 have
 fewer
 guilds
 having
 a
 loss
 of
 habitat
 exceeding
 20%
 at
 the
 left
 side
 of

these
 flow
 regimes,
 so
 monthly
 median,
 60%
 of
 annual
 average,
 and
 then
 90%
 of
 inflow
 as
 your
 flow
 by

requirement.

 As
 a
 trial
 balloon,
 DWR
 proposes
 using
 80%
 of
 inflow
 as
 the
 flow
 by
 requirement.

 And

since
 we
 haven’t
 come
 up
 with
 a
 classification/characterization
 process
 yet,
 we
 figured
 well,
 we’ll
 just

go
 with
 a
 statewide
 requirement
 for
 right
 now
 because
 it
 kind
 of
 gets
 back
 to
 what
 Chris
 was
 talking

about
 with
 the
 sustainability
 boundary,
 or
 the
 presumptive
 standard.

 If
 you
 maintain
 the
 natural

hydrograph,
 minus
 a
 certain
 percent,
 then
 in
 theory
 across
 the
 state
 it
 should
 protect
 the
 biota
 there.

The
 happy
 face
 represents
 80%
 of
 inflow
 as
 your
 flow-­‐by
 statewide
 application.

 You
 could
 argue,
 if
 you

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 46
 of
 51

look
 at
 these
 charts,
 these
 are
 all
 based
 on
 in-­‐stream
 flow
 study
 sites
 that
 are
 in
 the
 Piedmont,
 so

technically,
 an
 astute
 observer
 of
 these
 bar
 charts
 might
 say
 well,
 actually
 your
 smiley
 face
 should
 be

over
 here
 under
 physiographic
 regions
 for
 Piedmont
 because
 these
 are
 all
 based
 on
 Piedmont
 studies

so
 the
 proposal
 is
 to
 look
 at
 Mountain
 PHABSIM
 sites
 so
 hopefully
 the
 responses
 might
 compare
 to

these
 similarly.

 We’re
 scheduled
 to
 do
 those
 in
 the
 next
 month
 or
 so.

 We
 picked
 this
 80%
 inflow
 as

your
 flow
 by.

 At
 80%
 a
 certain
 number
 of
 guilds
 that
 are
 below
 the
 20%
 loss
 of
 habitat
 so,
 when
 we’re

putting
 forth
 trial
 balloons
 we
 said
 well,
 rather
 than
 just
 putting
 forth
 one,
 give
 people
 something
 else

to
 consider
 so
 we
 picked
 40%
 of
 annual
 average
 and
 also
 the
 monthly
 median.

 And
 as
 you
 look
 at

these
 various
 bar
 charts,
 you’ll
 notice
 that
 they
 have
 different
 habitat
 responses
 so,
 some
 will
 be

greater
 than
 others
 but
 for
 the
 most
 part
 it
 seemed
 like
 they’re
 pretty
 much
 on
 the
 same
 line
 of
 the

number
 of
 guilds
 that
 were
 below
 the
 20%
 loss
 of
 habitat.

 So
 that’s
 kind
 of
 our
 philosophy
 for
 why
 we

picked
 these
 three.

 Also,
 maintaining
 some
 of
 the
 natural
 variability-­‐-­‐if
 you
 have
 the
 monthly
 median,

that
 does
 maintain
 a
 certain
 amount
 of
 variability
 in
 your
 flows,
 as
 does
 the
 80%
 of
 inflow
 as
 your
 flow

by
 requirement.

 That
 still
 maintains
 some
 kin
 to
 natural
 variability.

 The
 40%
 of
 annual
 average
 is
 sort

of
 a
 flat
 line
 since
 it
 is,
 you’re
 taking
 a
 percent
 of
 a
 single
 value.

Q:

 Why
 no
 monthly
 or
 seasonal
 happy
 faces?

[00:45:15]

R:

 Well,
 that’s
 a
 good
 question.

 Jim
 did
 look
 at
 some
 seasonality
 in
 terms
 of
 some
 of
 these
 studies,

but
 as
 some
 of
 it
 shows
 up
 there,
 I
 guess
 we
 could
 have
 gone
 deeper
 into
 the
 weeds
 and
 pulled
 up

some
 seasonal
 ones,
 but
 I
 guess
 at
 the
 time
 we
 didn’t,
 so…

R:

 Well,
 I
 understand
 why
 you
 didn’t
 but
 as
 a
 strategy
 for
 consideration
 if
 we’re
 going
 to
 use
 a
 flat
 80%

of
 flow
 by
 on
 the
 basis
 of
 some
 geographic
 unit
 whether
 it’s
 a
 basin
 of
 some
 order,
 it
 makes
 sense
 to

me
 that
 that
 percent
 of
 flow-­‐by
 perhaps
 should
 vary
 according
 to
 season.

R:

 I
 mean
 I
 thought
 that
 the
 point
 was
 that
 the
 80%
 seemed
 to
 be
 protective,
 even
 at
 the
 low
 flow

seasons
 and
 that’s
 if
 you
 match
 that
 even
 at
 low
 flow
 seasons,
 we
 wouldn’t
 hesitate
 to
 allow.

R:

 There
 may
 be
 times
 when
 80%
 of
 flow-­‐by
 is
 inadequate.

C:

 You
 would
 have
 to
 go
 back
 and
 look
 through
 each
 of
 those
 curves
 again
 to
 find
 out
 whether
 that

was
 the
 case
 or
 not.

R:

 As
 a
 straw
 man,
 it
 seems
 to
 me
 that
 the
 department
 has
 suggested
 the
 best
 member
 or
 a
 member

of
 each
 of
 the
 three
 large
 families
 of
 implementation
 strategies
 and
 in
 each
 case
 they’ve

recommended
 an
 annual
 approach
 and
 I’m
 thinking
 that
 a
 seasonal
 approach
 has
 merit.

R:

 Monthly
 median
 would
 be.

C:
 Monthly
 median
 would
 capture
 some
 of
 the
 seasonalities.

R:

 A
 monthly
 median
 is
 even
 harder
 than
 seasonal
 median,
 but
 I’m
 just
 not
 positive
 that
 this
 will
 solve

the
 problem
 as
 implemented
 on
 an
 annual
 basis.

 I’m
 reasonably
 comfortable
 that
 it
 will
 do
 a
 better

job
 of
 solving
 a
 problem
 if
 it’s
 implemented
 on
 a
 shorter
 time
 step.

Q:

 But
 isn’t
 that
 percentage,
 when
 you
 run
 it
 through
 the
 model,
 that’s
 a
 daily
 time
 step.

 Say
 it’s
 20%

of
 inflow,
 so
 if
 it’s
 20%
 of
 inflow,
 it
 doesn’t
 matter
 if
 it’s—it’s
 not
 a
 seasonal
 or
 monthly
 thing,
 right?

The
 percentage
 of
 flow-­‐by
 is
 done
 on
 a
 daily
 time
 step.

R:

 For
 very
 low
 flow
 streams,
 100%
 of
 in-­‐flow
 may
 be
 required
 at
 certain
 seasons.

R:

 Well,
 so
 you’d
 have
 to
 go
 back
 through
 all
 of
 those
 curves
 and
 unless
 one
 of
 them
 just
 stands
 out
 as

crashing
 it.

 That’s
 really,
 I
 guess,
 the
 assignment
 was
 to—is
 there
 a
 difference.

R:

 My
 recommendation
 is
 that
 if
 it’s
 possible
 and
 practical,
 we
 should-­‐-­‐whether
 we
 have
 classes
 or

not-­‐-­‐we
 should
 think
 about
 seasonal
 flows
 and
 targets
 for
 those
 seasons
 and
 allow
 the
 variations
 from

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 47
 of
 51

those
 targets
 and
 if
 it
 turns
 out
 that
 the
 target
 is
 80%
 of
 instantaneous
 flow,
 or
 80%
 of
 daily
 mean,
 or

whatever
 flow
 that
 geographic
 unit,
 whether
 it’s
 a
 ridge
 or
 a
 basin
 of
 some
 order,
 it
 should
 to
 the

extent
 possible,
 reflect
 reality.

 It
 should
 reflect
 something
 we
 know
 about
 that
 kind
 of
 stream.

Q:

 And
 are
 you
 saying,
 in
 terms
 of
 seasonality,
 go
 from
 80
 to
 90
 some
 other
 variable
 at
 certain
 times
 of

the
 season?

R:

 If
 you’re
 defining
 that
 your
 strategy
 is
 percent
 of
 inflow,
 for
 some
 order
 of
 basin,
 you
 have
 the

option
 of
 saying
 it
 should
 be
 80%
 all
 the
 time.

 But
 you
 may
 have
 the
 option
 of
 saying
 it
 should
 be
 80%

above
 a
 certain
 low
 flow
 threshold,
 below
 which
 it
 should
 be
 90%,
 and
 below
 that
 it
 should
 be
 100%.

We
 talked
 about
 that
 earlier
 today.

 You
 may
 not
 have
 the
 information
 that
 you
 need
 to
 make
 that—to

write
 that
 kind
 of
 prescription,
 or
 it
 may
 be
 that
 it’s
 too
 complex
 and
 cumbersome
 to
 implement.

 But

ideally,
 whether
 we
 have
 classes
 or
 not,
 I
 think
 we
 should
 do
 our
 best
 with
 seasonality
 and
 we
 should

do
 our
 best
 with
 making
 our
 target
 flows
 fit
 reality.

R:That’s
 tricky.

C:

 The
 node
 in
 OASIS
 is
 going
 to
 have
 to
 be
 classified,
 potentially,
 or
 the
 node
 is
 going
 to
 have
 a
 goal
 or

an
 objective-­‐-­‐

 80%
 flow-­‐by
 or
 something
 like
 that?

 So
 if
 the
 node
 is
 placed
 on
 a
 stream
 that’s
 class
 X,

it
 will
 end
 up
 with
 this
 classification
 and
 it
 would
 get
 a
 prescription.

 So
 when
 we’re
 thinking
 about
 how

finely
 we
 have
 to
 understand
 it,
 we
 just
 have
 to
 understand
 where
 the
 nodes
 are
 probably
 going
 to
 go-­‐
-­‐
 what
 types
 of
 streams
 are
 going
 to
 get
 requests
 for
 withdrawal
 and
 if
 those
 are
 streams
 that
 have
 so

little
 total
 discharge
 at
 low
 flow
 season
 that
 they
 meet
 100%,
 that
 would
 be
 a
 prescription.

 But
 I
 would

think
 that
 a
 lot
 of
 the
 withdrawals
 aren’t
 probably
 going
 to
 be
 targeted
 to
 those
 streams,
 that
 they’re

going
 to
 be
 targeted
 as
 streams
 that
 with
 this
 80%
 would
 be
 potentially
 a
 reasonable
 daily
 target.

 I

mean
 is
 that
 fair?

 Is
 that
 what
 we’re
 supposed
 to
 be
 thinking
 about,
 is
 when
 you’re
 actually
 classifying

the
 node,
 once
 you
 place
 it
 into
 OASIS,
 that
 is
 going
 to
 have
 a
 class
 and
 it
 will
 either
 be
 sufficient
 to

have
 this
 statewide
 80%
 or
 as
 Sam’s
 pointed
 out,
 there
 may
 be
 streams
 of
 a
 particular
 size
 or
 their

ability
 to
 have
 base
 flow,
 for
 example,
 that
 may
 be
 more
 vulnerable
 and
 they
 need
 to
 go
 in
 a
 different

class.

R:

 Yes,
 when
 you
 set
 those
 flowbys
 for
 the
 particular
 node,
 if
 they
 are
 going
 to
 differ
 statewide,
 to

some
 extent
 you
 have
 to
 know
 how
 to
 set
 up
 those
 various
 nodes
 in
 terms
 of
 when
 are
 those
 flowbys

when
 do
 you
 have
 the
 red
 flags
 come
 up.

 So
 when
 you
 place
 those
 nodes,
 if
 we’re
 going
 to
 have

something
 other
 than
 statewide,
 you
 have
 to
 know
 what
 that
 flow
 scenario
 is
 in
 order
 to
 have
 those

red
 flags
 pop
 up,
 if
 they
 do.

Q:

 The
 framework
 though
 that
 you
 have
 in
 your
 matrix
 can
 handle
 those
 more
 specific
 approaches,

right?

 You
 move
 the
 smiley
 face
 a
 couple
 columns
 to
 the
 left
 or
 whatever,
 and
 you’re
 still
 within
 your

conceptual
 framework.

R:

 Yes.

C:

 If
 you
 get
 more
 data
 and
 it
 says
 you
 need
 to
 do
 the
 Piedmont
 differently
 from
 Mountains,
 you
 can

potentially
 have
 different
 numbers
 for
 those
 two.

R:
 This
 matrix
 is
 just
 my
 concept;
 the
 trial
 balloon
 is
 where
 the
 smiley
 faces
 are.

 The
 matrix
 can
 be

modified.

 Like
 I
 did
 say,
 since
 these
 are
 based
 on
 Piedmont,
 then
 perhaps,
 it
 may
 not
 be
 a
 statewide

concept,
 it
 might
 be
 Piedmont
 and
 when
 we
 look
 at
 the
 PHABSIMs
 for
 the
 Mountains,
 it
 might
 vary
 for

the
 Mountain
 sites.

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 48
 of
 51

C:

 What
 I
 want
 to
 point
 out
 about
 the
 way
 you
 guys
 are
 going
 to
 run
 the
 Mountains
 is
 it
 seems
 like

when
 we
 were
 in
 the
 Piedmont,
 we
 decided
 to
 have
 at
 least
 a
 suite
 of
 guilds
 that
 were
 shared
 amongst

those
 and
 it
 seems
 like
 we
 would
 want
 to
 have,
 to
 the
 best
 of
 our
 ability
 even
 though
 I
 understand
 that

some
 of
 the
 Mountain
 streams
 will
 have
 unique
 guilds,
 that
 there
 would
 be
 comparable
 guilds,
 so
 that

we
 can
 compare
 the
 results
 from
 the
 Mountains
 to
 the
 Piedmont
 to
 understand
 if
 that
 same
 guild

responds
 the
 same
 in
 the
 Mountains
 versus
 the
 Piedmont.

 Is
 that
 reasonable?

 Is
 that
 part
 of
 the

objective?

R:

 I
 sent
 an
 e-­‐mail
 about
 this.

R:

 Recently,
 you
 mean?

R:

 A
 week
 or
 so
 ago.

Q:
 That
 spreadsheet
 about
 the
 various
 sites,
 or
 whatever?

R:Yes,
 and
 then
 before
 that,
 the
 guilds.
 There
 will
 be
 some
 differences.

 But
 there
 will
 be
 some

similarities.
 But
 they’re
 not
 going
 to
 be
 a
 one-­‐to-­‐one
 match.

R:

 I
 don’t
 think
 it
 would
 a
 one
 and
 one
 match,
 but
 at
 the
 same,
 I
 feel
 like
 there’s—that
 the
 advantage

of
 including
 the
 responses
 to
 the
 same
 guilds
 is
 that
 the
 guild
 is
 that
 biology
 link
 to
 the
 habitat;
 you’re

really
 describing
 habitat.

 It’s
 really,
 purely,
 a
 description
 of
 habitat
 so
 if
 you
 took
 the
 guild

conversation
 out
 of
 it
 for
 the
 moment
 and
 you
 just
 said,
 how
 well
 are
 we
 maintaining
 fast,
 shallow

habitat?

 Do
 we
 maintain
 that
 in
 the
 Mountains?

 Do
 we
 maintain
 it
 in
 the
 Piedmont,
 and
 in
 the
 Coastal

Plain
 at
 80%
 flow
 by
 and
 strictly
 pull
 ourselves
 back
 to
 a
 habitat
 conversation.

 Flow
 versus
 habitat.

 I

think
 that
 is
 the
 simplest
 way
 to
 assess
 it
 and
 then
 the
 next
 step
 would
 be
 lining
 the
 guilds
 back
 up
 and

then
 have
 that
 conversation
 about
 how
 the
 biology
 responds.

 What
 I
 really
 want
 to
 know
 is,
 does
 the

French
 Broad
 act
 exactly
 the
 same
 way
 as
 the
 Tar
 River
 at
 Tarboro
 under
 certain
 flow
 conditions?

 And

it’s
 very
 likely
 they
 could—the
 shallow
 increases
 and
 the
 deep
 decreases
 at
 80%
 in
 a
 predictable
 way.

Is
 it
 equally
 predictable
 for
 the
 French
 Broad
 as
 it
 is
 for
 the
 Tar
 River?

 And
 then
 afterward,
 you
 come

back
 to
 how
 the
 guilds
 match
 up
 and
 whether
 in
 the
 mountains,
 you
 really
 are
 trying
 to
 tell
 with
 the

chubs
 if
 the
 Mountain
 suite
 of
 chubs
 act
 the
 same
 way
 as
 the
 Piedmont
 suite
 of
 chubs.

 You
 see
 what

I'm
 saying?

 If
 we
 don’t
 have
 any
 one
 to
 one
 matches,
 this
 conversation
 is
 just
 purely
 hypothetical.

 If

you
 can’t
 show
 that
 similar
 habitats
 respond
 to
 flow
 in
 the
 same
 way,
 you
 can’t
 talk
 about
 statewide

application
 using
 this
 method,
 in
 my
 mind.

 We
 really
 are
 talking
 about
 habitat,
 right?

R:

 Yes,
 I
 think
 saying
 it
 that
 way,
 my
 response
 would
 be,
 even
 if
 the
 guilds
 are
 different,
 the
 question
 is

are
 we
 protecting
 the
 habitat.

R:

 That’s
 the
 question,
 but
 if
 we
 use
 the
 terms
 of
 the
 guild,
 just
 like
 when
 we
 called
 our
 old

classification
 when
 we
 say
 the
 Coastal
 Plain
 streams,
 it
 kind
 of
 messes
 you
 up—it
 messes
 up
 your

ability
 to
 speak
 about
 how
 well
 we
 are
 maintaining
 habitat
 availability.

R:

 Yes,
 because
 in
 either
 the
 Piedmont
 or
 the
 Mountains,
 what
 we’ve
 tried
 to
 do
 is
 run
 enough

different
 habitat
 suitability
 curves
 through
 the
 program
 so
 that
 we’ve
 covered
 anything
 that
 would
 be

there.

 Then
 going
 further,
 you
 use
 the
 most
 limiting
 species
 or
 guild
 and
 if
 we’ve
 got
 habitat
 for
 it,

everything
 else
 is
 okay.

R:

 Right.

 The
 only
 thing
 that
 worries
 me
 about
 that
 is
 when
 there
 are
 those
 ecological
 communities

that
 don’t
 have
 that
 component.

 If
 it
 was
 zero
 in
 the
 natural
 hydrograph
 and
 it
 goes
 to
 200%,
 are
 we

capturing
 that
 change
 and
 is
 it
 meaningful
 to
 the
 ecological
 integrity.

 I
 think
 that
 is
 where
 there’s
 a

shift
 we
 might
 not
 be
 capturing.

 So,
 anyway,
 I
 think
 that
 this
 Mountain
 work
 is
 critical
 in
 trying
 to

make
 this
 work,
 unless
 you
 just
 really
 want
 to
 have
 a
 Piedmont
 prescription.

Q:

 When
 you
 have
 the
 flowbys,
 is
 there
 a
 threshold
 other
 than
 just
 like
 a
 7Q10
 for
 WaterFALL
 that
 you

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 49
 of
 51

try
 to
 present
 in
 that
 region?

R:

 The
 way
 that
 was
 analyzed
 is
 there
 is
 no
 threshold
 period.

 Just
 flow
 equals
 0.8
 times
 flow.

R:

 So
 it
 could,
 in
 theory,
 then
 produce
 some
 really
 low
 flows
 during
 drought
 times?

R:

 The
 way
 this
 was
 modeled-­‐-­‐
 now
 in
 reality,
 we
 can
 implement
 it
 in
 a
 different
 way.

 I
 mean
 this
 also

assumed
 that
 flow
 would
 be
 modified
 every
 day.

 And
 again,
 in
 reality,
 that
 would
 be
 twice
 a
 week

C:

 I
 guess
 the
 only
 baseline
 would
 be
 whatever
 the
 lowest
 value
 in
 the
 flow
 record
 would
 be?

 Don’t

you
 default
 to
 whatever
 the
 model
 flows
 in
 the
 deck?

R:

 Yes.

C:

 So
 a
 suggestion
 might
 be
 taking
 a
 flow-­‐by
 approach,
 but
 then
 having
 some
 sort
 of
 threshold

number.

R:

 That’s
 kind
 of
 like
 what
 Sam
 was
 talking
 about—what
 triggers
 a
 flag?

 Is
 it
 frequency
 or
 duration
 or

some
 sort
 of
 threshold
 that
 you’d
 have
 to-­‐-­‐

R:

 We
 need
 to
 have
 what
 constitutes
 an
 allowable
 excursion
 and
 how
 often
 can
 you
 have
 one?

R:

 To
 meet
 the
 definition
 in
 the
 statute.

 Still
 be
 able
 to
 recover.

R:

 But
 it’s
 still—I
 guess
 that
 still
 wouldn’t
 necessarily
 protect
 much
 natural
 variability
 because
 you

could
 always
 take
 it
 right
 down
 to
 that
 threshold
 continuously.

C:

 I
 have
 a
 feeling
 that
 nature
 will
 keep
 variation
 in
 there
 in
 spite
 of
 our
 best
 efforts.

 But
 I
 agree
 that

targeting
 a
 percentage
 of
 a
 period
 of
 record
 flow
 is
 going
 to
 be
 very
 challenging
 and
 it’s
 going
 to
 be

based
 on
 the
 use
 of
 models.

 There’s
 just
 nowhere
 where,
 or
 at
 least
 there
 are
 not
 very
 many
 places,

where
 someone
 who
 wants
 to
 take
 water
 out
 of
 the
 stream
 has
 a
 convenient
 gage
 right
 upstream
 and

a
 convenient
 gage
 right
 downstream
 and
 can
 report
 on
 what
 they
 have
 actually
 done
 to
 instantaneous

flow.

R:

 But
 again
 it’s
 a
 planning
 thing.

R:

 Right.

 Those
 are
 site-­‐specific
 kind
 of
 questions,
 like
 whether—

R:

 It
 just
 raises
 the
 flag.

R:

 Yes,
 it’s
 basically
 a
 model
 activity
 anyway.

C:

 A
 couple
 observations,
 as
 well
 as
 just
 reminders,
 is
 that
 this
 was
 set
 up
 with
 no
 floor,
 no
 safety
 net

beyond
 which
 you
 would
 not
 go.

 It’s
 always
 some
 percentage
 skimmed
 off,
 whether
 it’s
 10,
 20,
 or
 30

percent
 of
 ambient
 flow
 removed.

 One
 threshold
 that
 gets
 thrown
 around
 sometimes
 would
 be
 okay,

well
 yes,
 but
 you
 can’t
 go
 below
 the
 7Q10.

 Well,
 there’s
 nothing
 magic
 about
 7Q10,
 in
 fact,
 flow
 often

will
 go
 below
 the
 7Q10
 for
 a
 couple
 of
 scattered
 days,
 ever
 year.

 The
 7Q10
 is
 a
 flow
 that
 occurs
 for

seven
 consecutive
 days,
 once
 every
 ten
 years.

 But
 say
 7Q10
 is
 10,
 you
 may
 get
 a
 flow
 of
 8
 here
 and

there
 during
 a
 dry
 time
 depending
 on
 the
 kind
 of
 year
 it
 is.

 So,
 if
 you
 let
 ambient
 flows
 go
 to
 8

occasionally,
 well,
 so
 does
 Mother
 Nature.

 That’s
 one
 thought
 we
 had
 in
 mind.

 Another
 is,
 remember

this
 is
 a
 percentage
 reduction,
 so
 at
 low
 flows
 the
 magnitude
 of
 that
 reduction
 is
 small.

 The
 7Q10
 at

Rocky
 River
 at
 Highway
 64
 near
 Siler
 City,
 which
 is
 apparently-­‐-­‐it’s
 a
 small-­‐ish
 stream,
 but
 it’s
 still
 got
 a

drainage
 area
 of
 over
 50
 square
 miles.

 The
 7Q10
 there
 is
 a
 very
 flashy,
 slate
 belt
 stream-­‐-­‐7Q10
 is
 0.5.

Eighty
 percent
 would
 give
 a
 flow
 of
 0.4
 and
 I
 defy
 you
 to
 measure
 0.1
 CFS
 in
 a
 stream.

R:

 That’s
 where
 I
 feel
 like
 a
 monthly
 median
 is
 kind
 of
 a
 nice
 number.

R:

 But
 that
 is
 much
 less
 variable,
 that
 is—

R:

 That’s
 when
 you
 model
 it.

 The
 natural
 variation
 is
 significant
 around
 that
 median
 and
 you
 don’t

have
 instantaneous
 withdrawal,
 common
 sense,
 instantaneous
 withdrawal.

 You
 have
 intermittent

withdrawal
 up
 to
 a
 certain
 amount
 per
 day
 and
 it
 can
 be
 on
 average,
 right?

 So
 it
 could
 be
 more
 on
 one

day
 and
 less
 on
 the
 next
 day.

 If
 you
 model,
 it
 looks
 like
 you’re
 cutting
 it
 off
 at

 some
 kind
 of
 flat
 line,

but
 the
 reality
 is
 it
 would
 be
 highly
 variable,
 and
 the
 amount
 of
 water
 that
 you
 are
 shooting
 for
 in

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 50
 of
 51

September
 is
 going
 to
 be
 higher
 than
 the
 80%
 flow
 by
 would
 be.

R:

 At
 times,
 yes.

 You
 should
 still
 have
 to
 have
 a
 discussion,
 like
 Sam
 talked
 about,
 in
 terms
 of
 what’s

an
 acceptable
 excursion
 beyond
 your
 threshold
 because
 if
 you
 set
 your
 target
 as
 a
 median
 flow,
 well,

again
 half
 the
 time
 under
 natural
 circumstances
 the
 flow
 is
 less
 than
 that
 threshold.

C:

 And
 the
 other
 half
 the
 time,
 it’s
 more
 than.

R:

 Right.

C:

 The
 eco-­‐difference
 is
 going
 to
 be
 substantial
 most
 of
 the
 time.

 Another
 trial
 balloon
 goes

unpopped.

R:

 Despite
 everyone’s
 best
 efforts.

VIII.
 May
 14
 Meeting
 Agenda
 and
 the
 Timeline

The
 EFSAB
 proposed
 the
 following
 items
 for
 the
 May
 14
 agenda:

• BEC
 update
 on
 flow/biological
 response
 relationships

• Board’s
 Input
 on
 concerns
 about
 endangered
 species

• Complete
 Report
 Outline
 and
 section-­‐writing
 assignment

The
 EFSAB
 also
 updated
 the
 timeline
 as
 follows:

June
 18:

• PHABSIMs
 using
 Mountain
 sites
 –
 how
 it
 correlates
 with
 class

• Update
 on
 BEC?

• What’s
 missing

• Determine
 numbers

July
 16:

• Coastal
 Discussion

o Contributions
 from

 Coastal
 Coordination
 Group

• Endangered
 Species

• Draft
 report
 strategy
 agreed-­‐upon

August
 20:

 To
 be
 determined

September
 24:

 To
 be
 determined

October
 22:
 To
 be
 determined

December
 3:
 Final
 Meeting

IX.
 Information
 on
 Next
 Meeting

The
 draft
 agenda
 for
 the
 May
 14,
 2013
 meeting
 includes:

• Biological-­‐Environmental
 Classification
 (BEC)
 preliminary

 results
 for
 classification

• Flow
 /
 biological
 response
 relationships
 proof
 of
 concept

• Endangered
 Species

• Discussion
 of
 Framework
 for
 Recommendations

Ecological
 Flows
 Science
 Advisory
 Board

 April
 16,
 2013
 Meeting
 Summary

Page
 51
 of
 51

The
 next
 meeting
 of
 the
 EFSAB
 is
 scheduled
 for
 May
 14,
 2013
 at
 the
 Stan
 Adams
 Educational
 Center
 from

9:0oam
 until
 4:15pm.
 Please
 remember
 to
 bring
 lunch
 and
 refreshments
 with
 you.
 Coffee
 will
 be
 available
 on

site
 and
 soft
 drinks
 are
 ($1).
 Webinar:
 If
 you
 cannot
 attend
 the
 meeting
 in
 person
 but
 would
 like
 to
 join
 us
 via

the
 webinar,
 you
 can
 watch
 the
 presentations
 and
 listen
 to
 the
 live
 streaming
 audio
 of
 the
 meeting
 by
 accessing

the
 link
 and
 typing
 your
 name
 in
 the
 space
 labeled
 “guest”:

 https://denr.ncgovconnect.com/sab/

Meeting
 Location
 &
 Directions:
 The
 meeting
 location
 is
 the
 Stanford
 M.
 Adams
 Training
 Facility
 at
 Jordan
 Lake

Educational
 State
 Forest.
 Directions
 are:

 2832
 Big
 Woods
 Road,
 Chapel
 Hill,
 NC

 27517.

 From
 Rt
 64
 and
 Big

Woods
 Road,
 it
 will
 be
 the
 first
 Forest
 Service
 sign
 on
 the
 right.

 Pass
 the
 office
 building
 and
 continue
 on
 through

the
 gate
 to
 the
 education
 center.
 For
 Map
 link:

 http://go.ncsu.edu/stanadams