Transcript Saarp_flow-cology2012x

Strikes & Gutterballs
Modeling, monitoring, and
bio-assessment techniques
used in 2 flow ecology studies
in Virginia.
Topics
o
o
Virginia and Instream
Flows
Modeling Approach
• Space for Time
• Resolution/Pour Point
o
Analysis Approach
• Data Source & Quality
• Challenges
o
Takeaways
Total
2000
1500
1000
500
0
1900
1950
2000
2050
Virginia & Instream Flows
o
Unique Regulatory Structure:
• DEQ provides permits with consultation, comment to
State Water Control Board
o
o
o
o
Commenting/Consulting Agencies staffed with
instream flow experts
Public comment (NGO’s, citizens) can necessitate
Water Control Board hearing
Instream flow recommendations in every single
permit
Still a “water rich” state ~10% overall wd/Q, with
isolated high allocated streams
Virginia Goals
o
o
o
Expand & Solidify Scientific Basis for Instream Flow Recs
Provide basin specific impact estimation and resource valuation.
A “3-tiered” approach to developing flow ecology relationships
•
•
•
Tier 1 – Continuous curves describing the incremental relationship between biological
health and flow alteration
Tier 2 – Binary curves ,dividing the alteration spectrum along the line after which
substantial degradation would be expected to occur.
Tier 3 – Best professional judgment, and non-site specific model curves. May be
binary or continuous.
Tier 1: E = f(h)
Tier 2: Binary
Tier 3: Best Prof. Judge.
Modeling: Space for Time
Challenge: Scant before-after data (long-term)
Hypothesis: Areas with low hydro alteration
should have less hydro-biological impacts &
represent the “pre-condition” of altered areas
o Step 1: Create a hydrologic model of existing
Total
conditions
2000
o Step 2: Revert model to some 1500
1000
“pre-development” state
500
• Remove Impoundments
• Remove withdrawals
• Remove Discharges
o
0
1900
1950
Step 3: Calc. % Alteration of Hydro Indices
2000
2050
Hydro Modeling: Under the Hood
o
Rainfall Run-Off Simulation
• HSPF-based
• 26 Land-Uses:

But really, only about 5 that are truly hydrologically
distinct:
•
•
•
•
•
Forest
Impervious
Crop Land
Hay
Pasture Land (similar to Urban Pervious)
• Land Use Can Be Time-Varying/Customized
Flow Routing
o
o
o
Used a physical “storage routing” model
which considers channel slope, crosssectional geometry and roughness.
USGS regression relationships to estimate
channel geometry by physiographic
province and drainage area.
Runs on user-defined time-step, base
model has 1 hour time-step*.
ICPRB Modeling Version
•Very Small Watersheds
•“Sub-Resolution”
•Use unit-area
runoff from larger
scale model
•Route through
small channel
•Performed Well at
original resolution
•At Lower Resolution:
•Low – OK
Median - Good
Excerpted From Potomac River Small Watersheds Study,
ICPRB 2011 (draft)
•High - NSG
HWIModeling Version
•5-200 sqmi watersheds
(mean ~70 sqmi)
• Model calibrated to
USGS gages
•Model Performance:
•Low – over 85% w/in
15% for low 10%
Median – Very Good
•High – Very Good
Excerpted From Virginia HWI Study, Tetratech
2012 (draft)
Flow Alteration Models
o
Beyond land-use: withdrawals,
point source, reservoir operations
Jennings Randolph Flow
Augmentation Reservoir
(above)
Model Assumption & Verification
o
o
The models may
not get the flow
exactly right,
The models will
characterize the
nature of the
alteration.
• D perviousness
• wd/ps
Excerpted From Potomac River Small
Watersheds Study, ICPRB 2011 (draft)
Model Resolution
o
o
Pour Points – How Close is Close
Enough?
Thousands and thousands to 137 
Excerpted From Virginia HWI Study, Tetratech 2012 (draft)
Flow Metrics: ICPRB (Potomac)
Flow Range
Magnitude
High
Annual 3-day
Maximum
Mid
Median
Low
Annual 3-day
Minimum
Duration
High flow
duration DH17
High flow
volume index
MH21
Frequency
Rate of
Change
High Pulse
Count
Flashiness
Low Pulse
Duration
Low Pulse Count
Extreme Low
Flow Frequency
Excerpted From Potomac River Small Watersheds Study, ICPRB 2011 (draft)
Flow Metrics: HWI (Virginia-Wide)
1-Day Maximum (M)
1-Day Minimum (M)
August Low Flow (M)
7Q10 (M)
Number of Reversals (F)
High Flow Rise Rate (R)
Richard-Bakers Index Flashiness
90 Day Maximum (M)
90 Day Minimum (M)
High Flow Timing (T)
Date of Minimum (T)
Base Flow Index (M)
* (M=magnitude, D=duration, F=frequency,
T=timing, R=rate of change)
Biometrics
o
Use what you have, strengths and
limitations:
• Provided good coverage
• “But it’s not made to do that”
o
Devise new methods to overcome
the limitations of old metrics
• Flow Preference
o
Benthics & Fish
Biometrics: ICPRB (Potomac)
Biometrics
Chesapeake Benthic Index of Biotic
Integrity (BIBI)
Shannon-Weiner Index (SW)
% EPT
% Scraper
Hilsenhoff Family Level Biotic Index (FBI)
% Chironomidae
% Clingers
Ecological characteristic
Composite Index
Taxonomic diversity
Composition; generally
pollution sensitive
Feeding group
Pollution Tolerance
Composition; generally
pollution tolerant
Habit group
Biometrics: HWI (Virginia-Wide)
o
Fish
• Number individuals - total
• Number taxa - benthic insectivores, benthic, Centrarchidae,
darters, flow preference, fast, flow preference, moderate, flow
preference, slow, intolerant suckers, native benthic, native
Centrarchidae, native Cyprinidae, native insectivorous
Cyprinidae, native, native round-bodied suckers, native
sunfish, suckers, sunfish, total
• Percent individuals - Cottidae, dace, dominant 01 taxon,
flow preference, fast, flow preference, moderate, flow
preference, slow, game fish, insectivore, insectivorous
Cyprinidae, invertivore and piscivore, lithophils, non native,
omnivores, round-bodied suckers, tolerant, top carnivores
• Index - evenness, Shannon Wiener (log base e)
Biometrics: HWI (Virginia-Wide)
o
Benthic
o
Number individuals - total
o
Number taxa - Bivalvia, collectors, climbers, clingers, Coleoptera,
Diptera, Ephemeroptera, EPT, predators, filterers, Gastropoda,
intolerant, Plecoptera, predators, scrapers, shredders, sprawlers,
swimmers, tolerant, total, Trichoptera
o
Percent individuals - Amphipoda,ratio Baetidae to Ephemeroptera,
Bivalvia, Chironomidae, collectors, climbers, clingers, Coleoptera,
Corbicula, Crustacea, Decapoda, Diptera, dominant 01 taxon,
dominant 02 taxa, Ephemeroptera, EPT, Ephemeroptera & Tricoptera
(no Hydropsychidae), predators, filterers, Gastropoda, ratio
Hydropsychidae to EPT, ratio Hydropsychidae to Trichoptera,
intolerant, Mollusca, non Insecta, Odonata, Oligochaeta, Plecoptera,
predator, Plecoptera & Trichoptera (no Hydropsychidae), scrapers,
shredders, sprawlers, swimmers, tolerant, Trichoptera
o
Index - Beck's, evenness, Gomphidae, Oligochaeta, Diptera,
Hilsenhoff, Shannon Wiener (log base e), Coastal Plain Multimetric
Index (genus), Stream Condition Index (family)
Analysis: Methods, Expectations,
Statistics & Covariates
• Creating a Living System:
“Open-Source” approach to tools, data sets
and deliverables
 Require contractors to deliver analysis
routines, and use Open Source analyssis
systems (“R” is your friend)

• Understanding the System

Managing the Expectations of contractors,
scientists and policy makers
• Understanding the use of statistics, and
making sure that analysts do as well
Ecological Health Modeling System
o
Main Drivers of
Ecological Health:
1. Native/Naturalized
Community (stream
class/location
dependent)
2. Extent of detrimental
flow alteration
3. Water Quality
o
Without knowing all
three of the above,
we face greater
(sometimes
unacceptable)
uncertainty
Flow
Alteration
Ecology
= f (Flow)
Ecological
Health
Community
= f (Class)
Stream
Class
Ecology
= f(Quality)
Water
Quality
Expectations & Covariates
o
“A” not “The”
• The expectation of flow being a sole
cause and effect is only valid in streams
without any other controlling factors
Land Use/Habitat
 Water Quality

o
Covariate analysis is essential to:
• Verify that relationships demonstrate
causation, not just correlation
• Provide cleaner graphs
Tempering Expectations
o
o
Sounds like "it’s not that great”
But it IS that great, its just not that straightforward
It Might not look like Figure A – but At least
Figure B
Abundance vs. Alteration
Fraction of Maximum
Population
o
1
0.75
0.5
0.25
0
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Fraction of Mean Daily Flow REMOVED
1
Establishing Flow-Ecology
Hypotheses (FE-Hype)
o
Seemed to be disagreement in
process, or perhaps
miscommunications/semantic
misunderstandings:
• Do we just mine for significant stats?
• Do we ONLY check for flow metrics and
bio indicators that we think SHOULD
have merit?
• Is the reality actually somewhere in
between?
FE-Hype: Points of View
o
Points of View:
• Our use of IHA metrics is an implicit
flow-ecology hypothesis: these are
ecologically important flows, so…
• But, just because a bio-metric shows
some correlation with a ecological-flow
metric doesn’t mean there is any causal
relationship
o
Ultimately, both POV are true
FE-Hype: But Wait, There’s More
o
o
Sometimes, it is just as important to
evaluate situations where you
thought there should be a
relationship that failed to
materialize
Both flow regimes, and ecological
indices are models – we might
actually have some error here.
Act Now and Get This Bonus
o
In the end, we cannot make a ruling
about resource allocation based on a
relationship that seems to have
statistical significance, but for which
we have no flow-ecology hypothesis
to explain.
Ways of Looking at Data
o
o
o
o
Linear Regression
Quantile Regression
Pearson Ranking
Probability of
“Adverse Impact”
How Significant is the
Relationship?
o
The use and mis-use
of R2
• R2 shows us % variation
explained by x-y
o
The p-value
• p tells us probability of
being illusory
o
What % of health
change do we expect
a single flow metric to
control?
Excerpted From Virginia HWI Study, Tetratech
2012 (draft)
How Much Alteration is
Enough?
o
o
o
Too much of the
"wrong kind"
(urbanization)
Not enough of the
"right kind" (things we
have regulatory
control over)
But this is usable:
• Maybe +/-20% is not
the kiss of death?
• Beyond +/-20% we
start to scrutinize
heavily.
Excerpted From Virginia HWI Study, Tetratech
2012 (draft)
Flow-Preference Metrics
Excerpted From Virginia HWI Study, Tetratech 2012 (draft)
The Y-axis: A Glass Half Full
o
o
o
o
Metrics Aren’t
Always Numerical
ICPRB Team Used
“Probability of Fair
or Better”
Good, but alas,
Urban Signature
“Risk Management”
approach that
works for managers
Excerpted From Potomac River Small
Watersheds Study, ICPRB 2011 (draft)
Our Takeaways
o
o
Highly urbanized systems provide a great
challenge for developing F-E relationships
Space for Time shows great promise
• Choosing hydrologic resolution is very important in
maximizing use of data
• The corollary: you must have a resolution that provides
data coverage that fulfills statistical assumptions
o
o
o
Operational rules, withdrawals and discharges are
all very potent sources of alteration
Flow-preference metrics show promise for “yaxis”
Some traditional metrics are not a “no-go”