Quantifying the factors that influence flows in headwater

streams

By: Les Stanfield

Southern Science Information Section, Ministry of Natural Resources

In a d

rought

year.

...

Historical Perspective• 1930’s deforested landscape • Hurricane Hazel

1980 -2000 Defining Productive Capacity in streams

• Habitat models (Bowlby and Roff (1986), Steedman (1988), Stoneman and Jones (2000), etc.

Correlate specific habitat features with fish biomass

Guides site specific mitigation..... But planning?

Strategic Implementation (mid 80’s to present)

Rigorous application of the no-net- loss (net gain) policy on direct fish

habitat.....

Did I mention this is Les’s perspective.....

Mid 90’s Watershed planning (requires a landscape perspective)

2006: Three papers Quantifying landscape influences on biophysical properties of streams

Since 1995 a coalition of partners has been working to understand how disturbance influences stream health

(MNR, DFO, EC, MOE, TRCA, GRCA, CLOCA, HRCA, Michigan DNR, GLC, UofT, UofG)

Background

Key findings

-6-4

-2

02

4

6

0 5 10 15 20

LDI

Fish

Can

onic

al A

xis

1

Strong threshold response in fish community to development

LDI = ∑% area land cover type*LDI rating

LDI Ratings range from 0.0 for water to 0.9 for roads

18 land cover categories.....

126,000 brook trout < 21% of historic range

Regression model for Brook Trout

Take home message

1. Landscape models better predictors than instream habitat

2. Only 2-3% of residual variation from landscape models explained by competition, habitat or riparian condition

3. Catchment is the best predictor of stream conditions

Protecting direct fish habitat may not be enough

Need to protect processes important to fisheries in areas currently not protected..... Headwaters

1. Hydrology (baseflow/peak flows)2. geomorphology3. Connectivity

Factors that influence flows

• Baseflow Stanfield . L. W., B. Kilgour, K. Todd, S. Holysh A. Piggott and M. Baker. In press. Estimating Summer Low-flow in Streams in a Morainal Landscape

using Spatial Hydrologic Models. Canadian Water Resources Journal.

• Peak flowsStanfield, L. W. 2009. Understanding the factors that influence headwater stream flows in response to storm events. University of Toronto, Masters Thesis. 75 pp

0

200

400

600

800

1,000

1,200

1,400

1,600

0 1,000,000 2,000,000 3,000,000 4,000,000

BFI*Area

Sta

ndar

d Lo

w F

low r2 = 0.81

0

200

400

600

800

1,000

1,200

1,400

0 500 1,000 1,500 2,000 2,500 3,000 3,500

MODFLOW-modified

Stan

dard

ized

Low

Flo

w r2 = 0.89

0

200

400

600

800

1,000

1,200

1,400

1,600

0 1 2 3 4

MODFLOW

Sta

ndar

d Lo

w F

low

r2 = 0.83

0200400600800

1,0001,2001,4001,6001,800

0 300,000,000 600,000,000 900,000,000

Area (m2)

Stan

dard

ized

Low

Flo

w r2 = 0.79

0

200

400

600

800

1,000

1,200

1,400

0 500,000 1,000,000 1,500,000 2,000,000 2,500,000

POS_B_SUM

Stan

dard

ized

Low

Flo

w

r2 = 0.76

Poor correlationin small catchments

Minimum size for a permanent

flowing stream17,800 ha!

Check your watershed day

Volunteer agency collaboration

Baseflow at most road Crossings

Inventory perched culverts!

Hypothesis Generation

Study Area Determination/ Site Selection

Data Collection

MeasureStage Response

Estimate VelocityMeasure Wetted Area

of ResponseCalculate Event

Discharge

MeasureRainfall

Attribute Catchment Geology/land cover/slope

Interpolate Mean Catchment rainfall

Summarize Catchment attributes

Correlations(bi-plots, RDA)

Technique Validation

Peak flow Methods

North-south land use/land cover gradient.....

Measuring stream response

Crest Stage Gauges and Manning’s equatio

Estimating Q: (Area*Velocity)

• Area good...• Manning’s needs

validation

Manning’s Equation..... V =1/n*R2/3*S1/2

Defining and Attributing site catchments

• Arc-Hydro define catchments• GIS used to attribute

– Geology (Quaternary)– Land use/land cover (SOLRIS) (LDI)– Slope (valley)

Rain event• Volunteer and agency network to

measure rainfall in each catchment • Extract maximum value for period

between sampling (12 events) and rainfall 2 days prior (rb)

110

Select sites at furthest upstream accessible area, contrast in geology and land use

Rain gauge network (volunteer and agency)

1811

Results

The search for patterns…..

Standardized for each catchment per day(Q*sec/day/catchment area)

0 10 20 30 40 50 60

020

4060

8010

0

rain mm

disc

harg

e m

m

15 ,148 ^

HAR1_03LDI = 0.72 and BFI = 0.15

0 10 20 30 40 50 600

2040

6080

100

rain mm

disc

harg

e m

m

3 ,139 ^11, 150 ^

HAR1_04LDI = 0.73 and BFI = 0.15

0 10 20 30 40 50 60

020

4060

8010

0

rain mm

disc

harg

e m

m

HAR2_01LDI = 0.57 and BFI = 0.19

0 10 20 30 40 50 60

020

4060

8010

0

rain mm

disc

harg

e m

m

D-006LDI = 0.17 and BFI = 0.25

0 10 20 30 40 50 60

020

4060

8010

0

rain mm

disc

harg

e m

m

DA18-17LDI = 0.28 and BFI = 0.23

0 10 20 30 40 50 60

020

4060

8010

0rain mm

disc

harg

e m

m

GAN18-3LDI = 0.34 and BFI = 0.16

Poorly drained soils.....

consistently higher response in highest development areas,

0 10 20 30 40 50 60

020

4060

8010

0

rain mm

disc

harg

e m

m

A01_01LDI = 0.75 and BFI = 0.4

2 ,106^3 ,112 ^

0 10 20 30 40 50 600

2040

6080

100

rain mm

disc

harg

e m

m

A04_15LDI = 0.38 and BFI = 0.35

0 10 20 30 40 50 60

020

4060

8010

0

rain mm

disc

harg

e m

m

A04_14LDI = 0.33 and BFI = 0.35

0 10 20 30 40 50 60

020

4060

8010

0

rain mm

disc

harg

e m

m

GAN6-10LDI = 0.04 and BFI = 0.35

0 10 20 30 40 50 60

020

4060

8010

0

rain mm

disc

harg

e m

m

D-026LDI = 0.07 and BFI = 0.35

0 10 20 30 40 50 60

020

4060

8010

0rain mm

disc

harg

e m

m

D-027LDI = 0.1 and BFI = 0.35

Moderately well drained soils

High response if really high LDI,.... the rest not perfect!

0 10 20 30 40 50 60

020

4060

8010

0

rain mm

disc

harg

e m

m

GAN20-7LDI = 0.8 and BFI = 0.69

0 10 20 30 40 50 60

020

4060

8010

0rain mm

disc

harg

e m

m

GLENHOLMLDI = 0.66 and BFI = 0.69

0 10 20 30 40 50 60

020

4060

8010

0

rain mm

disc

harg

e m

m

DA18_hydLDI = 0.7 and BFI = 0.69

0 10 20 30 40 50 60

020

4060

8010

0

rain mm

disc

harg

e m

m

GAN3-11LDI = 0.03 and BFI = 0.71

0 10 20 30 40 50 60

020

4060

8010

0

rain mm

disc

harg

e m

m

GAN4-17LDI = 0.04 and BFI = 0.71

0 10 20 30 40 50 60

020

4060

8010

0

rain mmdi

scha

rge

mm

GAN3-8LDI = 0.04 and BFI = 0.71

Well drained soils

Poor response in all catchments.... geology rules

Redundancy and Partial Redundancy Analysis

• Used to quantify patterns and evaluate significance

water

tile

PoorModerate

Well

Slope

LDI

rb12

rb11

rb7rn12

rn11

rn8

rn7

-0.600

-0.400

-0.200

0.000

0.200

0.400

0.600

0.800

1.000

-1.000 -0.800 -0.600 -0.400 -0.200 0.000 0.200 0.400 0.600

PCA 1

PCA

2

dis7

dis8

dis12

dis11

1. Higher poor and well drained soils and LDI > higher discharge2. Higher rain and moderate drained soils > lower discharge3. Discharge scores very similar

Partial RDA

P=0.02

P=0.002

P=0.33

Both Land use and Geology affect discharges.. Rain is not important.... Too messy!

Residual R2=0.76!

a= -0.03

Geology and slopeall rainfalld= -0.01`

g= 0.01

f= 0.05

b= 0.11

c= 0.11

LDI

e= -0.03

Conclusions

• Geology and land use are both important..... but lots of variability

• Responses in drought years are not highly correlated with rainfall

- soil moisture - crop condition ?

Implications for “Planning”1. Headwater systems can generate large volumes of

storm water2. Existing predictive models (means) will not protect

headwater stream low or peak flow conditions .... and biota!

3. Most vulnerable areas to change is likely poorly drained soils.... Where development is concentrated or planned!

4. At present cannot predict specific impacts to flows from land use with the possible exception of poorly drained soils

Implications for monitoring

• Crest Stage approaches work well for broad scale studies – need to validate Manning’s n

• Existing rain gauge network is inadequate• Need to invest in both low flow and peak

flow monitoring programs.... Can’t rely on predictive models!

CLIMATE CHANGE!Spec

ies

at R

isk!

U.S. supreme court decision to protect headwater systems under Endangered species Act.

Implications for monitoring

• Crest Stage approaches work well for broad scale studies – need to validate Manning’s n

• Existing rain gauge network is inadequate

• Need to invest in both low flow and peak flow monitoring programs.... Can’t rely on predictive models!

Acknowledgements• Laura Del Giudice, Scott Jarive (TRCA)• Jeff McNeice, Ian Kelsey(CLOCA)• Pam Lancaster (GRCA)• Oak Ridge Moraine Foundation• Toronto Region Conservation

Authority (Laura DelGuidice, S. Jarvie)• Ministry of Natural Resources

(esp. Silvia Strobl, Will Johnson, SSIS)• Lake Ontario Modeling Team• Ministry of Environment• Ontario Ministry of Agriculture/Food

• Grand River Conservation Authority (Graham Smith)

Sarah Ross, Mike Berenz, Amie Cousins and Kristina Abengoza