WinTR-20 SensitivityMarch 20091 WinTR-20 Sensitivity to Input Parameters.

WinTR-20 Sensitivity March 2009 1

WinTR-20 Sensitivity to Input Parameters


Lesson Objectives

1. Identify the various WinTR-20 Input Parameters that affect the volume of runoff and peak discharge predictions.

2. Identify the relative sensitivity of WinTR-20 to its input parameters in predicting the peak and/or volume of runoff.

3. Identify the relative sensitivity of WinTR-20 to its input parameters in relation to channel routing.


WinTR-20 Hydrology Model

Predicts Volume of Runoff Predicts Peak Rate of Runoff Predicts Entire Hydrograph of Runoff Based on Watershed and Rainfall

Characteristics Modeled as Input Parameters Changes to Input Parameters Will Change

the Volume and Rate of Runoff Predicted


WinTR-20 Watershed Input Variables

Drainage Area Runoff Curve Number (RCN) Time of Concentration (tc)

Unit Peak Factor (UPF) of Dimensionless Unit Hydrograph (DUH)

Antecedent Runoff Condition (ARC)


WinTR-20 Rainfall Input Variables Depth of Rainfall Rainfall Distribution (includes duration)


Effects of Variation in Drainage Area

Sensitivity to Drainage Area

0.70

0.80

0.90

1.00

1.10

1.20

1.30

0.6 0.8 1 1.2

% of Runoff Vol or Peak Compared to Unit

% o

f Dra

inag

e A

rea

Com

pare

d to

Uni

t

Volume Ratio

Peak Ratio

Linear Correlation

% Change in DA results in comparable change to predicted volume and peak of runoff.

Be sure DA is being properly identified (be aware of non-contributing areas).


Effects of Variation in RCN

% Change in RCN results in exaggerated change to predicted volume and peak of runoff.

RCN can be influenced by stage of vegetal growth and/or antecedent rainfall at time of storm event.

Sensitivity to RCN

0.00

0.50

1.00

1.50

2.00

2.50

60% 80% 100% 120% 140%% of RCN Compared to Unit (RCN 70)

% o

f R

un

off

Vo

l o

r P

ea

k C

om

pa

red

to U

nit Volume Ratio

Peak Ratio

LinearCorrelation


Effects of Variation in tc

% Change in tc results in decreased change to predicted peak rate of runoff (no change in volume).

A decrease in tc results in an increase in predicted peak discharge.

Sensitivity to tc

0.500.600.700.800.901.001.101.201.30

60% 80% 100% 120% 140%

% of tc Compared to Unit (1.0 hr)

% o

f R

un

off

Vo

l or

Pe

ak

Co

mp

are

d t

o U

nit

Volume Ratio

Peak Ratio

Linear Correlation


Effects of Variation in Unit Peak Factor

% Change in UPF results in nearly similar change to predicted peak rate of runoff (no change in volume).

UPF is a watershed based response to excess rainfall assumed to be similar per inch of runoff.

Sensitivity to Unit Peak Factor

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

0.00 0.50 1.00 1.50

% of Unit Peak Factor Compared to 484

% o

f R

un

off

Vo

l or

Pe

ak

Co

mp

are

d t

o U

nit

Volume Ratio

Peak Ratio

Linear Correlation


Effects of Variation inAntecedent Runoff Condition (ARC) ARC values of 1 or 3 alter the RCN selected

for assumed ARC 2 conditions. ARC 2 is normally assumed for design. ARC 1 can be used to help calibrate for a

known “drought” condition prior to the target storm event (not necessarily accurate).

ARC 3 can be used to help calibrate for a known “saturated soil” condition prior to the target storm event (not necessarily accurate).


ARC Adjustments (Continued)

For this example: DA = 1.0 mi2, tc = 1 hr, RCN = 70, 4.0 inch 24 hr Type II Rainfall

ARC 2 – (RCN 70), Qv = 1.33”, Qp = 437 cfs

ARC 1 – (RCN 51), Qv = 0.37”, Qp = 65 cfs

ARC 3 – (RCN 85), Qv = 2.46”, Qp = 874 cfs

WinTR-20 results are very sensitive to changes in ARC. Be sure that assumed change is appropriate or alter RCN within ARC 2 conditions for finer adjustment.


Effects of Variation in Rainfall Depth

% Change in Rainfall Depth results in exaggerated change to predicted volume and peak of runoff.

Be sure that the actual Rainfall that has occurred and is being calibrated to is properly identified for the entire watershed.

Sensitivity to Rainfall Depth

0.50

0.75

1.00

1.25

1.50

70% 90% 110% 130%

% of Rainfall Compared to Unit (4in)

% o

f R

un

off

Vo

lum

e,

Pe

ak

Co

mp

are

d t

o U

nit

Volume Ratio

Peak Ratio

Linear Correlation


Effects of Variation in Rainfall Distribution

Design rainfall distributions normally set by criteria (e.g. Type I, IA, II, or III).

Can attempt to calibrate to a historical rainfall event of known varying intensity (recording rain gage).

Rainfall distribution alone (not depth) only effects the rate of runoff, not the volume.


Effects of Variation in Rainfall Distribution(Continued)

For this example: DA = 1.0 mi2, tc = 1 hr, RCN = 70, 4.0 inch 24 hr Type II Rainfall

Type II - Qp = 437 cfs

Type I - Qp = 221 cfs

Type IA - Qp = 106 cfs

Type III - Qp = 383 cfs

WinTR-20 peaks are very sensitive to selection of rainfall distribution. Calibrate with the best known rainfall distribution.


Parameter Selection for Desired Change in WinTR-20 Runoff Volume

WinTR-20 Parameter to be Changed, Independent of Others

Desired Change in Runoff Volume (%)

-50% -25% -10% -5% +5% +10% +25% +50%

Required Change in Drainage Area -50% -25% -10% -5% +5% +10% +25% +50%

Required Change in Rainfall -26% -13% -5% -2.5% +2.5% +5% +12.5% +23%

Required Change in RCN -17% -8% -2% -1% +1% +2% +7% +13%

Required Change in Time of

Concentration N/C N/C N/C N/C N/C N/C N/C N/C

Required Change in Unit Peak

Factor N/C N/C N/C N/C N/C N/C N/C N/C

N/C signifies, No Change possible to alter volume. This parameter does not effect volume prediction.


Parameter Selection for Desired Change in WinTR-20 Peak Runoff

WinTR-20 Parameter to be Changed, Independent of Others

Desired Change in Runoff Peak (%)

-50% -25% -10% -5% +5% +10% +25% +50%

Required Change in Drainage Area -50% -25% -10% -5% +5% +10% +25% +50%

Required Change in Rainfall -24% -12% -5% -2.50% +2.5% +5% +11% +21%

Required Change in RCN -13.5% -6% -2% -1% +1% +2% +5.5% +11%

Required Change in Time of

Concentration +150% +50% +15% +7% -6% -12% -26.5% -44%

Required Change in Unit Peak

Factor -54% -29% -12% -6% +6% +13% +33% +72%


Combined Parameter Impacts

Assumed Normal Run DA = 1 mi2, RCN =70, tc = 1.0 hr, UPF = 484

Runoff Volume = 1.33”, Peak Rate = 437 cfs

Low Run DA = 1 mi2, RCN =63, tc = 1.25 hr, UPF = 300


High Run DA = 1 mi2, RCN =77, tc = 0.75 hr, UPF = 600



WinTR-20 Channel Routing Model Predicts hydrograph (including peak) at

downstream end of reach. Based on cross section and reach

characteristics modeled as input parameters. Changes to input parameters will change the

peak discharge and hydrograph shape predicted at the end of the reach.


WinTR-20 Channel and Reach Input Variables Selection of representative cross section Cross section rating table (slope and “n”) Channel length Flood plain length Shape of inflow hydrograph Base flow (if significant)


WinTR-20 Channel Routing Sensitivity Test Trapezoidal cross section, BW = 15, SS = 2:1 Slope = 0.001 and 0.004 Manning n = 0.03, 0.04, 0.05 Channel length, 0.8 to 1.2 mile Inflow hydrograph, DA = 1, CN = 80, Tc = 0.5

and 1.0, RF = 3.2 inches, Type II storm Base flow = 0.0 60 WinTR-20 runs


Two Inflow hydrographs

Red (higher) is the hydrograph for Tc = 0.5 hour. Green (lower) is the hydrograph for Tc = 1.0 hour.


Effects of Variation in Length and “n”

% Change in length results in less change to predicted peak outflow.

% Change in Manning “n” results in less change to predicted peak outflow.



% Change in length and “n” results in less change to predicted peak outflow.

Length and “n” less sensitive for Tc = 1.0 hydrograph.




Results for steep slope are less sensitive.




Results for Tc = 1.0 hydrograph are even less.


Porcupine Mountains State Park, Michigan


Questions???


The End

WinTR-20 SensitivityMarch 20091 WinTR-20 Sensitivity to Input Parameters.

Documents

peak of runoff

runoff peak

runoff volume

peak andor volume of

comparable change

exaggerated change

inches of runoff

unit rcn