land cover effects on watershed hydrologic …land cover effects on runoff land cover can have...

LAND COVER EFFECTS ONWATERSHED HYDROLOGIC

MEMORY

Jason P. JulianRobert H. Gardner

Oct 14, 2013

Dai

ly D

isch

arge

(m3 /s

)

Runoff = Precip – ET – Soil Moisture – Groundwater

Watershed Hydrologic Memory

Morphometric variablesArea (A)

Stream order (OHS)

Drainage density (Dd)

Mean channel slope (Sc)

Basin shape (Rf)

Geologic variablesSilt-clay percentage (SC%)

Soil depth to bedrock (Zbr)

Hydrologic variablesReservoir storage percentage (RS%)

Precipitation effectiveness ratio (Rpe)

Land-cover variablesPercent water-wetland (%WW)

Percent urban (%UR)

Percent forest (%FO)

Percent agriculture (%AG)

Runoff affected by …

Eastern Piedmont – 87 watersheds• Similar morphometry

• Pear- or oval-shaped• Dendritic drainage

• Moderate relief • neither topographic or

subsurface controls dominate

• Similar geology• Thick clay-rich soils• Deeply weathered bedrock

• Similar climate• Mid-latitude, humid subtropical• No dry season

• Many flow gages with long continuous daily records

Urban

Urban

Agriculture

Forest

Grassland

Barren

Wetland

Water

Morphometric variablesArea (A)

Stream order (OHS)

Drainage density (Dd)

Mean channel slope (Sc)

Basin form ratio (Rf)

Geologic variablesSilt-clay percentage (SC%)

Depth to bedrock (Zbr)

Hydrologic variablesReservoir storage percentage (RS%)

Precipitation effectiveness ratio (Rpe)

Land-cover variablesPercent water-wetland (%WW)

Percent urban (%UR)

Percent forest (%FO)

Percent agriculture (%AG)

Runoff affected by …

Which period or frequency will you analyze?

All of them

What about stationarity with respect to climate?

We use the same 40-y (1968-2007) records for all watersheds

1968 2008

decadal

yearly

monthly

dailyIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

Pd

Pcp

Pa

annual peak

cross point

βlf

βhf

-4

-6

-12

-10

-8

dailyweeklymonthlyyearlydecadal

frequency

spec

tral p

ower

(or v

aria

nce)

Power Spectral Analysis(think temporal correlation using a moving window)

(long-term memory)

(short-term memory)

Rainfall - (white-noise) Runoff - (red-noise)

fcp = 5.6 ± 0.6-βhf = 0.42 ± 0.05-βlf = 0.02 ± 0.02

fcp = 6.0 ± 1.3-βhf = 1.84 ± 1.15-βlf = 0.45 ± 0.16

cross-pointShort-term memoryLong-term memory

Spectral VariableBest landscape predictor

(r)Stepwise model r2

(α = 0.05)Daily power, Pd %Wetland (-0.49) 0.57

Annual power, Pa %Wetland (0.41) 0.21*

Spectral slope, -β0 %Wetland (0.57) 0.62

Cross point power, Pcp %Urban (0.50) 0.39

Cross point frequency, fcp Slope (-0.32) 0.20*

Long-term memory, -βlf %Urban (-0.47) 0.43

Short-term memory, -βhf %Wetland (0.55) 0.58

Do landscape attributes dictate a catchment’s hydrologic memory?

* Low correlation

Annual power, Pa Daily power, Pd Cross-point power, Pcp

Den

sity

Cross point frequency, fcp

Den

sity

Den

sity

Spectral slope, -β0 Low-frequency slope, -βlf High-frequency slope, -βhf

Watershed Memory

Runoff Variance

Urban coverage, %UR

Dai

ly P

ower

P dr = -0.20

r = 0.67

r = -0.24

r = 0.68

cp = 11%

cp = 13%

Lon

g-te

rm m

emor

y

-βlf

cp = 14%

r = 0.37r = -0.65

Cro

ss-p

oint

pow

er

P cp

Urban Thresholds

Stream biota studies with 10-15% threshold• Paul and Meyer, 2001, Ann Rev Ecology,

Evolution, and Systematics• Utz et al., 2009, Ecological Indicators• Roy et al., 2003, Freshwater Biology

Urban coverage, %UR

Dai

ly P

ower

P dr = -0.20

r = 0.67

r = -0.24

r = 0.68

cp = 11%

cp = 13%

Lon

g-te

rm m

emor

y

-βlf

cp = 14%

r = 0.37r = -0.65

Cro

ss-p

oint

pow

er

P cp

Urban Thresholds

Affects hydrologic drought?

adapted fromVannote et al., 1980

Longitudinal Spatial Patterns in Spectral Variables

Stream Order12

3

4

5

6

7

Precipitation

Runoff ‐ variance

Runoff ‐memory

Climate-influenced

Landscape-influenced

Low frequency Pa -βlf

High frequency fcp -βhf

A matrix for characterizing Hydrologic Signatures?

Land Cover Effects on Runoff Land cover can have considerable and predictable

effects on runoff patterns (aka watershed memory)

10-15% urban threshold above which urban coverage becomes the dominant control on runoff patterns

Downstream threshold (after 3rd-order) where watershed processes become dominant over precipitation in determining runoff patterns in Eastern Piedmont

Matrix for hydrologic signatures:[climate vs. landscape effects] [low vs. high frequency events]

Questions?

Jason.Julian@txstate.edu