Estimating Streamflow Channel
Losses with the Green-Ampt
Model
Neil Hutten
Ag Eng 558
April 20, 2001
Presentation Outline
• Introduction and Motivation
• Channel Loss Estimation Methods
• Rawls & Brakensiek (1983) Determinations
of Green-Ampt Parameters
• Application to a stream site
• Uncertainties and Conclusions
Motivation
• RCRA requires TSDs to determine whether
they are located in 100-year floodplains
• Additional engineering studies are required
if TSD is located in a floodplain ($$)
• Floodplain extent is influenced by losses
• Stream “loss” is groundwater “gain”
• Stream channel losses can be pathways for
subsurface contaminant transport
Commonly-used methods to
estimate stream channel losses
• Assume infiltration losses are balanced by
local precipitation gains
• Representative Reach Loss
– Stream Gage A minus Stream Gage B
• Adjust/extrapolate a “known” rate
• Adjust peak flows from regression
equations to equal peak flows obtained from
flood frequency analyses of gaged data.
Limitations of stream loss
methods
•
•
•
•
•
•
No data at ungaged sites
Ephemeral or intermittent streams
Extent of groundwater contributions
Evapotranspiration vs. Infiltration
Channel wetting and drying cycles
Single value for an entire stream reach
Green-Ampt Overview
•Ponded or unponded infiltration
•Deep homogeneous soil
•Water infiltrates as “piston flow”
•Sharply defined wetting front
• Green-Ampt Overview
• Rate Form (f) of G-A Equation assumes a
ponded surface so the infiltration rate equals
infiltration capacity of the media.
• Depth of Ponding can be neglected.
Green-Ampt Model
•Green-Ampt Parameters
• Effective Suction at Wetting Front
• Effective Hydraulic Conductivity
• Soil Porosity
Green-Ampt Depth to Wetting Front
2 K Sf t
Lf 
N
2 ( Hyd . Cond .) * (Cap. Suction) * (time)
Lf 
Porosity
Green-Ampt Infiltration Rate
( Sf  Lf )
f K*
Lf
Cap. Suction  Depth to Wetting Front
f  Hyd Cond . *
Depth to Wetting Front
•Rawls and Brakensiek (1980s)
• Determined ranges of values for:
– Wetting Front Suction
– Hydraulic Conductivity
– Soil Porosity
• For eleven USDA Soil Textures
• 1200 Soils, 5000 Horizons, 34 States
• Methods described in standard references
Wetting Front Suction
Sf  exp[ 6.53  7.326()  0.00158(C 2 )  3.809( 2 )  0.000344(S)(C)  0.04989(S)()
 0.0016(S 2 )( 2 )  0.0016(C 2 )( 2 )  0.0000136(S 2 )(C)  0.00348(C 2 )()
 0.000799(S 2 )()]
Where :
S  Percent Sand
C  Percent Clay
  Porosity
Wetting Front Suction
with Texture
Porosity with Texture
Hydraulic Conductivity
with Texture
Green-Ampt Parameters as
Determined by Rawls/Brakensiek
Meanwhile, back at the river…
• Streamflow Losses on Big Lost River were
determined from stream gage station data
by Bennett (1990)
• Average annual streamflow: 1965 to 1987
• Sixteen (16) streamflow measuring sites and
stations
Big Lost River Losses
•
•
•
•
1.5 cfs/mi: west bndry, INEEL to div. dam
2.5 cfs/mi: div. dam to Hwy 26
5 cfs/mi: Hwy 26 to Lincoln Blvd (ICPP)
1 cfs/mi: Lincoln Blvd (ICPP) to Lincoln Blvd
(NRF)
• 4 cfs/mi: Lincoln Blvd (NRF) to BLR Sinks
• 2 cfs/mi: above BLR Sinks
• 18 cfs/mi: in the Big Lost River Sinks.
Measured Channel Loss
• Stream gaging station 11
• Bennett’s measured loss = 2 cfs/mile
• Channel Width varies from 40 to 60 feet;
– 40 feet was used
• Measured infiltration rate = 1.04 cm/hr.
Specifics at Station 11
• Coarse pebble to cobble gravel above
gaging station 11
• Sediment grades to sand and “sandy silt”
below station 11
• Sandy Loam set of Green-Ampt parameters
was used for “sandy silt”
Selection of G-A Parameters
•
•
•
•
•
Sandy Loam (R-B) Parameters
Porosity (phi) = 0.453cc/cc
Wetting front Suction head Sf = 11.01 cm
Hydraulic conductivity Ks = 1.09 cm/hr
“Modelled” Sandy Loam Infiltration Rate
after ten hours was 1.61 cm/hr
• Measured rate was 1.04 cm/hr (2 cfs/mi)
Depth to Wetting Front and
Infiltration Rate for Sandy Loam
35
30
25
20
Depth to
Wetting (cm)
15
Infiltration Rate
cm/hour
10
5
0
1 3 5
7 9 11 13 15 17 19
Depth to Wetting Front and
Infiltration Rate for Sand
80
70
60
Depth to
Wetting (cm)
50
40
Infiltration Rate
cm/hour
30
20
10
0
1 3 5
7 9 11 13 15 17 19
Ten-hour Wetting Depth and Infiltration
Rates with Green-Ampt Parameters
After Ten (10) Hours
Wet Front Hydraulic Wet. Dept. Infiltr. Rate
Porosity Suction Conduct.
(cm)
(cm/hr)
Sand
Loamy Sand
Sandy Loam
Loam
Silt Loam
Sa Cl Loam
Clay Loam
Si Cl Loam
Sandy Clay
0.437
0.437
0.453
0.463
0.501
0.398
0.464
0.471
0.43
4.95
6.13
11.01
8.89
16.68
21.85
20.88
27.30
23.90
11.78
2.99
1.09
0.66
0.34
0.15
0.10
0.10
0.06
51.66
28.90
23.02
15.92
15.05
12.83
9.49
10.77
8.17
12.91
3.62
1.61
1.03
0.72
0.41
0.32
0.35
0.24
Uncertainties
• Pre-existing initial moisture contents were not considered
• R-B Porosity, Wetting Front Suction, and Hydraulic
Conductivities for media larger than sand not available
• Infiltration characteristics of ephemeral channel bottoms
compared to infiltration of upland soil sites
• Layering, textural changes, surface crusts, etc.
Conclusions
• Green-Ampt parameters developed by Rawls and
Brakensiek may be a useful tool to determine stream
channel infiltration loss rates.
• Ten-hour Modeled Infiltration Rate (1.6 cm/hr)
approximated the measured infiltration rate (1.0 cm/hr)
• Time frame of measured infiltration rate was not specified.
• Compare field samples with R-K parameters
• This is a research area worth further investigation.