Review of TNC Chapter 3 “Climate and Hydrology of the ...

Review of TNC Chapter 3 “Climate and Hydrology of the Upper Gila Basin"

(Garfin et al., Univ. of Arizona)

Reviewer: David S. Gutzler University of New Mexico

Garfin et al. estimate the effects of projected climate change on flows on the upper Gila and San Francisco Rivers Projected flows are derived from coupled climate and hydrologic models Global climate models Regional models Hydrologic model

General conclusion: Decreasing future flows due to climate change

This conclusion is qualitatively consistent with: * Previous projections by US Bureau of Reclamation * My related analysis of upper Gila flows commissioned by ISC

Garfin et al. modeling strategy

Strengths: State of the art models, competently utilized Daily output allows high-resolution analysis

Weaknesses: Known model limitations, e.g. summer precipitation Relatively few simulations (5, instead of dozens or hundreds) Little consideration given to future natural variability

Precipitation Climatologies Global climate models (CMIP3)

Regional atmospheric models (NARCCAP, UA)

Hydrologic model (VIC)

Simulate flows at gages in upper Gila basin: (Gila & Virden on Gila R, Clifton on S. Francisco R)

Garfin et al. principal results: Climate change

Temperature high 3-4°C/century

Snowpack high not shown, but huge decrease

Precipitation total low highly variable

Precipitation variability medium more extreme events

Confidence (my assessment)

Change by 2041-2070

Temperature Change Precipitation Change Wtr Spr Sum Aut Wtr Spr Sum Aut

Spring flow total high

Summer flow total low

Earlier Spring runoff high

Short-term peak flows medium ~10% increase of upper 10%

Garfin et al. principal results: Streamflow change

Average Monthly Flow Change Change in Distribution of Daily Flows

6-19% decrease annual flow }

Confidence (my assessment)

Change by 2041-2070

Low Flows

High Flows

0

CONCLUSIONS: TNC Chapter 3 "Climate and Hydrology of the Upper Gila Basin"

General conclusion: Decreasing future flows on the upper Gila River

Plan for: Significant decrease in average melt season flow .... superimposed on larger natural climatic variability

Uncertain change in average Summer flows

Increase in flow extrema associated with intense storms

Review of TNC Flow Needs Assessment, Chapter 5 and 7

Deborah L. Hathaway November 10, 2014

S.S. Papadopulos & Associates, Inc. www.sspa.com

Chapter 5:

Hydrologic Impacts of CUFA Diversions

• Review of TNC’s use of TNC IHA software to characterize flow under CUFA conditions

• IHA: “Indicators of Hydrologic Alteration”

• Provides descriptive and comparative statistics

• Compare TNC results to similar SSPA evaluation

• Review TNC chapter discussion

1. Set-up CUFA Diversion Model

• Inputs: • Historical daily gaged flows, • Storage assumptions, • CUFA diversion rules and constraints

• Results: Hypothetical daily flow record with CUFA diversions

• Spreadsheet model developed separately by TNC and NMISC

• Versions are different but are sufficiently similar for evaluating environmental impacts under CUFA

2. Analyze Differences in Flow Conditions under CUFA with IHA Model

• General agreement regarding impacts of CUFA (Scn 1):

• Decrease in median flow for March by about 15%, median flow is unchanged in other months

• Reduction of flow peaks in diversion months, typically late winter or spring

• No impact to lower flows, i.e., below 50th percentile (with 150 cfs bypass criteria)

Example of IHA-Generated Comparison

Alternate Scenarios analyzed by TNC

• Scenario 2 – no bypass criteria, with CUFA

• Scenario 3 – climate change, not CUFA

• Scenario 4 – climate change, with CUFA

3. Discussion

• Reduction of Variability

• Wetting of Secondary Channels

• Environmental Impacts

What does IHA analysis tell us about flow differences under CUFA?

• No change to extreme low flows (150 cfs criterion) • No change to low flows (150 cfs criterion) • CUFA does not dry up the river (low flows not diverted) • No change to median monthly flows except in March (March median flow: 189159 cfs) • High pulse events: median peak change 221205 cfs • Small floods: median peak change 28503072 cfs • Large floods: median peak change 12100 11930 cfs

Questions?

Chapter 7:

Groundwater and Surface Water Interaction

• Review of contributing data – surface water, groundwater

• Review of TNC calculations and analyses

• Further exploration of data where relevant to analyses

• Review of TNC discussion

• Analysis of TNC concerns

TNC Concern: Increased Groundwater Rates of Recession will Impact Riparian Vegetation

• Concern based on opinion that surface water flow rates will fall more rapidly under CUFA conditions (Fall Rate)

IHA Analysis – Fall Rates decrease, not increase, under CUFA

• Concern that groundwater levels will drop too fast for vegetation under CUFA conditions (Recession Rate)

Groundwater levels won’t drop at faster rates under CUFA because surface flows don’t fall at faster rates

TNC’s calculated groundwater recession rates are generally not relevant because they isolate a portion of the hydrograph without considering impacts of the preceding peak:

TNC Concern: Reduction in flow will reduce seepage and recharge causing substantial declines

in groundwater levels

• Concern not supported by data:

Data show that extreme low flow and channel drying is the primary cause substantial declines

Data show that small increases in flow will initiate and sustain recharge

Effects of CUFA on Groundwater Levels

• CUFA diversions typically occur in periods where groundwater is rising, alluvial recharge continues to occur but at lower rate over period of diversion

• Groundwater rises less rapidly, and falls less rapidly

• CUFA effects on groundwater in range of 0 – 1 foot

• CUFA diversions will not occur under low flow conditions nor cause steep groundwater declines

TNC Concern: Rapid groundwater recession and steep groundwater declines will impact riparian

vegetation with cascading impacts to birds, mammals, etc.

Groundwater differences occur during higher water periods and do not constitute rapid or steep declines

Questions?

© 2014 HDR Architecture, Inc., all rights reserved. © 2014 HDR Architecture, Inc., all rights reserved. © 2014 HDR Architecture, Inc., all rights reserved. © 2014 HDR, Inc., all rights reserved. © 2014 HDR, Inc., all rights reserved. © 2014 HDR, Inc., all rights reserved. © 2014 HDR, Inc., all rights reserved. © 2014 HDR, all rights reserved.

© 2014 HDR, all rights reserved.

Chapter 6 Review

Gila River Flow Needs Assessment, TNC 2014

Andy McCoy, PhD, PE, HDR Jarvis Caldwell, HDR

Summary and critique Chapter 6 - Hydrodynamic Modeling and

Ecohydraulic Relationships by Dr. Mark Stone and Dr. Ryan Morrison. Primary Reviewers: Andy McCoy, PhD, PE, HDR. Jarvis Caldwell, Aquatic Sciences Manager, SW HDR

Expert Review - Purpose

The stated effect on riparian inundation is unreliable for the following reasons:

Summary

• Do not present relevant background information to support or document the development process, limitations or uncertainty of the: topographic surface, computational mesh or boundary conditions.

• Do not present calibration or validation methods or results

Summary cont. • Do not explain the significance of model

limitations regarding sediment movement.

• Do not provide adequate discussion on Bayesian Network model process, data inputs or how they relate to annual recruitment results.

• Do not provide scientific reference to explain whether reported changes in annual inundation is significant for seed recruitment.

Pg 141.Techniques were applied to correct inferior LiDAR data. Not described. No topography < 750 cfs.

there could be significant

error in the modeled channel, conveyance and results.

Topographic Uncertainty

No description or

quantification of computational mesh adequacy to the LiDAR resolution, areas of complex bathymetry, and high velocity gradients.

Computation Mesh

Accurate rating curve fundamental to simulation results oHEC-RAS 2010 o rating curves/boundary

conditions not presented o their development and

manipulations (Pg. 121) oSpatial relation to sites

Boundary Conditions

1397.50

1397.60

1397.70

1397.80

1397.90

1398.00

1398.10

1398.20

1398.30

0 200 400 600 800 1000

Wat

er S

urfa

ce E

levat

ion

(m)

Discharge (cfs)

There is no documentation

of the SRH2D model calibration or validation process. The degree of accuracy of

results cannot be placed into context.

Calibration & Validation Pg 121. Calibration was achieved by

making minor adjustments to Manning’s n values in order to more closely match observed water surface elevations during high flow events. Water surface elevations were estimated from piezometer data located at the TNC transects.

There was no attempt to quantify the types and numbers of areas that will be overpredicted or underpredicted soley because of the assumptions Riparian vegetation Single substrate grain size

used for each study subreach – size not stated, significance not stated

Sediment Motion

Authors state that changes

in inundation frequency (<750 cfs) are not shown due to poor topography Authors state some areas

reduced inundation up to 50%

Inundation Frequency

Pg 133. In the spring (b), much of the intermediate zones are expected to be inundated between 20% and 40 % less frequently

Bayesian network input parameters – compounded errors and uncertainty Figure 10 – 50% of area,

inundated 0.2% annually – 18 hrs inundation frequency

changes is on order of 2 to 9 total hours in a year. Is two to nine hours is

significant?

Recruitment

Conclusion • There are serious shortcomings in model development

documentation, background information and assumptions.

• This lack of transparency undermines confidence in reported results used to develop the conclusions.

• To qualify results, many of the real or perceived model shortcomings should be cleared up with a comprehensive model development report.

© 2014 HDR Architecture, Inc., all rights reserved. © 2014 HDR Architecture, Inc., all rights reserved. © 2014 HDR Architecture, Inc., all rights reserved. © 2014 HDR, Inc., all rights reserved. © 2014 HDR, Inc., all rights reserved. © 2014 HDR, Inc., all rights reserved. © 2014 HDR, Inc., all rights reserved. © 2014 HDR, all rights reserved.

© 2014 HDR, all rights reserved.

Chapter 10 Review

Gila River Flow Needs Assessment, TNC 2014

Dave Ward, HDR Jarvis Caldwell, HDR

Summarize and critique Chapter 10 – Effects of Altered Flow Regimes and habitat Fragmentation on Gila River Fishes by Dr. David Propst and Dr. Thomas Turner Primary Reviewer:

Dave Ward, National Fisheries Science Lead, HDR

Expert Review - Purpose

Magnitude of effect of diversion project overstated because the authors:

Summary

• Infer too much from analyses with limited predictive value

• Imply relationships between fish density and discharge for which no analyses are presented

• Provide contradictory statements regarding fish life history and effects of withdrawal

• Rely on data from only one site to represent fish populations for the entire reach

• Incorrectly imply reach is presently un-altered from “natural” state

Characterization of reach as ‘un-altered’ is misleading. Native fish have been

extirpated, and non-native species have been introduced. Water diverted from river

Upstream and Downstream

Mischaracterization of Cliff-Gila Valley

Page 233 – Gila chub was extirpated from New Mexico by the 1950s and roundtail chub has not been found in the Cliff-Gila Valley reach since 1991.

Reliance on data from one site to characterize entire reach imprudent and not reliable for broad-scale implications. Lack of recognition that

existing diversions have historically reduced river flow to near no-flow conditions just upstream of study site.

One Study Site – Limited Sample Size

Page 255 – At Riverside, most Sonora suckers collected in autumn were age-0 individuals and adults were uncommon.

Coefficients of determination (r2) are consistently low o1.0 - 0.0 oEg. 0.25 – explains 25% of

the variation

Low values for r2 limit the predictive value of the models

Limited Predictive Value

Page 243 – Table 5. Regression Analysis.

No statistical analyses presented for relationships between fish density and flows less than and greater than 185 cfs. Contrary to what the authors

state or imply, Figures 12 (spikedace) and 20 (loach minnow) indicate that the relationship between spring discharge and fish density may be negative or non-existent at flows >185 cfs.

Fish Density and Flow

Page 257 – Among the flow attributes considered, both spikedace and loach minnow responded positively to increasing spring mean daily discharge through about 300 cfs, indicating this spring discharge level represents a threshold.

Figure 12 - Spikedace

Authors often imply that a new obstacle to upstream migration will fragment fish populations. To the contrary, Figure 31

shows that for spikedace, allelic richness already decreases in an upstream direction, indicating upstream movement limited.

Upstream Migration Support limited upstream movement on Page 257 – While drift contributes to maintenance of downstream populations, it cannot be so massive that it depopulates source reaches.

Authors state that there was no relation between monsoon flows and fish density, but later claim that withdrawals during monsoon season under climate change scenario would further compromise conditions for native fishes.

Monsoon Season Page 258 – By and large, there was no relation between monsoon or winter flows and autumn density of Cliff-Gila Valley native fishes.

Page 276 - Because sufficient water cannot be diverted during spring, water is also withdrawn during the monsoon season, further compromising conditions for native fishes.

Misunderstanding of AWSA diversion proposal. Authors imply that proposed

diversion and storage project control or even release “multiple flow pulses”.

Flow Pulses Page 276 – The most troublesome attribute of the projected flow regime with CUFA diversion is the effect of multiple flow pulses during spawning and early ontogeny..

Conclusion

• Chapter provides useful information on status and ecology of fishes in the Cliff-Gila Valley, and effectively describes the types of effects that a new diversion may have.

• However, authors infer too much from analyses of limited predictive value and sometimes use artificial discharge “thresholds” to enforce questionable points.

• Authors provide contradictory statements regarding the effects of withdrawals and rely on a single site to characterize the entire reach.

• Authors imply throughout the report that a diversion would alter the Gila from its “natural” state, although the river has undergone substantial changes already.

Report Reviewer: Michael Hatch

SWCA Environmental Consultants

November 10, 2014

Report Review “Effects of Altered Flow Regimes and Habitat

Fragmentation on Gila Fishes” by Thomas F. Turner and David L. Propst

University of New Mexico

A July 2014 publication of the Nature Conservancy

Purpose of this Review

Advise ISC of significant findings,

Identify problematic elements of study design and execution,

Identify problematic elements of study analysis and conclusions,

Identify measures needed to rectify problematic elements of experimental protocol,

Offer alternative conclusions if warranted.

Turner-Propst Study Scope

Study focus − effects of altered flow under AWSA: community composition, trends in abundance, density effects, and age-specific effects.

An additional study topic concerns the effects of river fragmentation.

Problematic Elements of Turner-Propst Study Design and Execution

Lack of fidelity between remotely measured hydrologic predictor values with conditions at the fish sampling site.

Investigated “averaged” hydrologic variables do not adequately represent regional environmental extremes.

Response variables are poorly suited for a hydrology-based effects analysis.

Problems that stem from small biased samples.

Flawed procedures of age determinations.

Failures to meet underlying assumptions of statistical analyses.

Problematic Elements of Turner-Propst Analysis and Conclusions

Only a small portion of the total variation in species density is explained by its linear relationship with the individual flow attributes that were studied.

Site conditions vary over time so it is unclear which factors are responsible for any perceived patterns.

Relationships between environmental and response variables may only be valid over the range of values found at the surveyed site.

Some analyses comingle categorical and continuous variables.

Regression analyses rely on static measures and are poor at elucidating dynamic relationships between factors (e.g., those involving feedbacks and time lags).

Loach Minnow Autumn Density − February-April Discharge (Source: Turner and Propst 2014)

SWCA 2012

Desert Sucker Autumn Density − Days < 30 cfs (Source: Turner and Propst 2014)

Review Findings

Age-specific Effects Linked to Flow Flawed basis for interpreting and modeling species

growth with advancing age.

0

50

100

150

200

250

300

350

400

38 42 46 50 54 58 62 66 70 74 78 82 86

Nu

mbe

r of

Fis

h

Standard Length (2 mm intervals)

Age (year class)

1 2 3 4 5 6 7

Stan

dard

Len

gth

(mm

)

20

30

40

50

60

70

Observedvon Bertalanffy modeled lengths-at-age

(2008) (2007) (2006) (2005) (2004) (2003) (2002)

lt= 80.1699 (1-EXP(-0.1989 (Age - -1.5029)))

Findings − Effects of River Fragmentation

Whereas the discussion of river fragmentation is important to the aquatic species conservation, it is not specifically relevant to an effects assessment for current AWSA proposals. Current AWSA proposals − avoid impediments to fish movement, avoid depletions of flow that fragment

aquatic habitats, restrict water storage facilities to off-channel

catchments, and preserves ecological functionality.

Summary and Conclusions

The study design and statistical analyses are flawed.

Evidence is lacking to affirm or refute assertions of flow effects on fish in the Gila River Basin of New Mexico.

The discussion of river fragmentation is considered important and generally relevant to the management of all natural resources in the Gila River Basin that may directly or indirectly affect aquatic habitats in the Gila River Basin. Certainly this section of the report is relevant to endangered species recovery planning and the formulation of management policy.