Sei sulla pagina 1di 53

SWAT Calibration Techniques

Calibration, Validation & Verification


F CALIBRATION: model testing with known input
and output used to adjust or estimate factors
F VALIDATION: comparison of model results with
an independent data set (without further
adjustment).
F VERIFICATION: examination of the numerical
technique in the computer code to ascertain that it
truly represents the conceptual model and that
there are no inherent numerical problems

Calibration/Validation Periods

Flow

distinct time period


similar range of
conditions
adequate time period to
simulate conditions

Time
Setup

Calibration

Validation

Model Configuration
F

Land use categories


land use types in watershed, existing and future land
uses, management techniques employed, management
questions

Subwatersheds
location, physical characteristics/soils, gaging station
locations, topographic features, management questions.

Reaches
topographic features, stream morphology, cross-section
data available
Calibration Issues:
individual land use parameter determination
location of gaging station data
location of water quality monitoring information
available information on stream systems

Model Configuration
Calibration Points Example

Calibration/Validation
Procedures
F
F
F

Hydrology - first and foremost


Sediment - next
Water quality - last (nitrogen, phosphorus,
pesticides, DO, bacteria)
Check list for model testing
4 water balance - is it all accounted for?
4 time series
4 annual total - stream flow & base flow
4 monthly/seasonal total
4 frequency duration curve
4 sediment and nutrients balance

Calibration Time Step


F

Calibration sequence
annual water balance
seasonal variability
storm variability
time series plot
u frequency duration curve
u

baseflow
overall time series

Calibration/Validation
Statistics
Mean and standard deviation of the
simulated and measured data
Slope, intercept and regression
coefficient/coefficient of determination
Nash-Suttcliffe Efficiency

Calibration/Validation
Common Problems
F
F

too little data - monitoring period too short


small range of conditions
only small storms
only storms during the spring...

F
F

prediction of future conditions which are


outside the model conditions
calibration/validation does not adequately
test separate pieces of model
accuracy of each land use category prediction

F
F

calibration adjustments destroy physical


representation of system by model
adjustment of the wrong parameters

Calibration/Validation
Suggested References
F

Neitsch, S. L., J. G. Arnold, J. R. Kiniry and J. R. Willams. 2001. Soil and Water
Assessment
Tool

Manual,
USDA-ARS
Publications.
pp:
341-354.
http://www.brc.tamus.edu/swat/manual.

Santhi, C., J. G. Arnold, J. R. Williams, W. A. Dugas, R. Srinivasan and L. M. Hauck.


2001. Validation of the SWAT Model on a Large River Basin with Point and Nonpoint
Sources. J. American Water Resources Association 37(5): 1169-1188.

Srinivasan, R., T. S. Ramanarayanan, J. G. Arnold and S. T. Bednarz. 1997. Large area


hydrologic modeling and assessment: Part II - Model application. J. American Water
Resources Association 34(1): 91-102.

Arnold, J.G., R. S. Muttiah, R. Srinivasan and P. M. Allen. 2000. Regional estimation of


baseflow and groundwater recharge in the upper Mississippi basin. J. Hydrology
227(2000): 21-40.

Hydrology Calibration
Summary
F

Key considerations
Water balance
overall amount
u distribution among hydrologic components
u

Storm sequence
u

time lag or shifts


time of concentration, travel time

shape of hydrograph
peak
recession
consider antecedent conditions

Example Calibration Plot

Example Calibration Plot


Calibration of flow at Hico, Bosque River Watershed, TX
Observed

Simulated

350

250
200
150
100
50

Time

1996

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1972

1970

1968

1966

1964

0
1962

Flow Volume (mm/year)

300

Example Calibration Plot

Hydrologic Calibration
Scenario 1

Flow (cfs)

Simulated
Observed

Time (hours)

Hydrologic Calibration
Model failed to simulate some peak
flows

F
F

Rainfall station is not


representative
Localized storm -no
response
Malfunctioning gages
(precipitation or flow)

Simulated
Observed

Flow (cfs)

Time (hours)

Solutions
F

Use precipitation data from representative


meteorological stations
Carefully review precipitation and flow data for the
particular duration

Hydrologic Calibration
Scenario 2

Flow (cfs)

Simulated
Observed

Time (hours)

Hydrologic Calibration
Model consistently over predicts the
flow
High Surface flow

Simulated
Observed

Flow (cfs)

Solutions
F

Time (hours)

Curve number for different land uses-decrease by 10%


(CN in .mgt)
Soil available water - increase upto 0.04
(SOL_AWC in .sol)
Soil evaporation compensation factor increase up to 1.0
(ESCO in *.sub)

Hydrologic Calibration
Model consistently over predicts the
flow
F

High base flow


Too little
evapotranspiration

Solutions
F

Simulated
Observed

Flow (cfs)

Time (hours)

Increase deep percolation loss (Adjust threshold depth of


water in shallow aquifer required for the base flow to
occur) (max 100mm, GWQMN in .gw)
Increase groundwater revap coefficient (max of 0.40,
GW_REVAP in .gw)
Decrease threshold depth of water in shallow aquifer for
revap to occur (min of 0.0, REVAPMN in .gw)

Hydrologic Calibration
Scenario 3

Flow (cfs)

Simulated
Observed

Time (hours)

Hydrologic Calibration
Simulated flow follows the observed pattern
but lags the actual flow consistently
F
F

F
F

Time of concentration is too


long
Less than actual slope for
overland flow
Over estimated surface
roughness
Snow melt parameters
Flood routing coefficients
Solutions
F
F

Simulated
Observed

Flow (cfs)

Time (hours)

Increase slope (up to 20%) for overland flow (SLOPE)


Mannings roughness coefficient- lower it after checking
OV_N tables (OV_N)
The value of overland flow length- lower to 5-10m, if
necessary (SLSUBBSN)

Hydrologic Calibration
Scenario 4

Flow (cfs)

Simulated
Observed

Time (hours)

Hydrologic Calibration
Simulated flow over predicts peak flows but
under predicts all other times

Too little base flow


Too high surface runoff

Solutions
F

F
F

Flow (cfs)

Simulated
Observed

Time (hours)

Adjust surface runoff until reasonable by lowering CN,


increasing SOL_AWC and adjusting ESCO
Adjust base flow until reasonable with GW parameters
Iterate between surface runoff and base flow until both
look reasonable

Sediment Calibration Summary


F

Key considerations
Sources of sediment loadings
Loadings from HRUs/Subbasins
u Channel degradation/deposition

Sediment loading distribution


overall amount
u Seasonal loading
u

distribution by storm sequence


rising and falling limb of hydrograph
peak concentration

Example Calibration Plot

Sediment Calibration
Scenario 1
Sediment
0.60
Observed

Sediment tons/ha

0.50

Simulated

0.40
0.30
0.20
0.10
0.00
1

Time in Months

Sediment Calibration
Model consistently under predicts the
sediment
Sediment

0.60

Observed

Low sediment yield

Sediment tons/ha

0.50

Simulated

0.40
0.30
0.20
0.10

Solutions
F

Calibrate HRU/Subbasin Loading

0.00
1

Time in Months

USLE crop management factor (P) Increase after checking


USLE table for reasonable values (USLE_P in .sub)
USLE slope length factor -- Increase by up to 10m (LS)
(SLSUBBSN in .sub)
Slope of HRUs--Increase by up to 20% (SLOPE in .sub)
Crop practice factor (C) for land use -- Increase by reasonable
amount to account for local conditions (USLE_C in crop.dat)
Verify tillage operations in *.mgt files; increase crop residue
coefficient upto 0.10 ( RSDCO) and increase bio-mixing efficiency
upto to 0.3 for heavy biological activity (BIOMIX in .bsn)

Sediment Calibration
Model consistently under predicts the
sediment
Sediment

0.60

Low sediment yield

Observed
Simulated

0.50
Sediment tons/ha

0.40
0.30
0.20
0.10
0.00
1

Time in Months

Solutions
F

Calibrate Channel degradation/deposition


Linear and exponential parameters used for channel sediment
routing Increase SPCON upto 0.01 & SPEXP to 2.0 (SPCON and
SPEXP in .bsn)
Channel erodibility factor Increase to 0.3 to 0.4 if channel is
erodible (CH_EROD in .rte)
Channel cover factor Increase upto 1.0 if no vegetation exists
on bank/channel bottom (CH_COV in .rte)

Sediment Calibration
Model consistently under predicts the
sediment
F

F
F
F

Often only have total sediment yield or


concentration at gage/outlet of watershed
Not sure if source is upland fields or channel
erosion
Visit watershed to see if significant channel
erosion is occurring
Check subbasin yields (t/ha) to make sure they are
reasonable. The remainder must come from the
channels

Sediment Calibration
Scenario 2
Sediment
0.60
Observed

Sediment tons/ha

0.50

Simulated

0.40
0.30
0.20
0.10
0.00
1

Time in Months

Sediment Calibration
Model consistently over predicts the
sediment
Sediment

0.60

Observed

High sediment yield

Solutions

Sediment tons/ha

0.50

Simulated

0.40
0.30
0.20
0.10
0.00
1

Calibrate HRU/Subbasin Loading

Time in Months

USLE crop management factor (P) decrease after checking USLE


table for reasonable values (USLE_P in .sub)
USLE slope length decrease by up to 10m (LS) (SLSUBBSN in .sub)
Slope of HRUs decrease by up to 20% (SLOPE in .sub)
Crop practice factor (C) for land use decrease by reasonable
amount to account for local conditions, check USLE Handbook
(USLE_C in crop.dat)
Verify tillage operations in *.mgt files; crop residue coefficient
decrease down to 0.01 if appropriate for plant (RSDCO) and biomixing efficiency decrease down to 0.01 for lower biological activity
if appropriate (BIOMIX) in .bsn

Sediment Calibration
Model consistently over predicts the
sediment
Sediment

0.60

High sediment yield

Sediment tons/ha

Observed
0.50

Simulated

0.40
0.30
0.20
0.10
0.00
1

Solutions
F

Time in Months

Calibrate Channel degradation/deposition


Linear and exponential parameters used for channel sediment
routing decrease SPCON down to 0.0005 and SPEXP down to
1.0 (SPCON and SPEXP in .bsn)
Channel erodibility factor decrease to 0.01 if bedrock or nonerosive bank material is present (CH_EROD in .rte)
Channel cover factor decrease down to .01 if 100% vegetation
cover exists (CH_COV in .rte)

Nutrients Calibration Summary


F

Key considerations
Sources of nutrient loadings
Loadings from HRUs/Subbasins
u In-stream processes

Nutrient loading distribution


overall amount
u Seasonal loading
u

distribution by storm sequence


rising and falling limb of hydrograph
peak concentration

Example Calibration Plot


Organic N

Org N kg/ha

2.00
1.50
1.00
0.50

57

53

49

45

41

37

33

29

25

21

17

13

0.00
Months(93-97)

Obs Org N kg/ha

Sim org N kg/ha

Mineral N

Min N kg/ha

0.60
0.40
0.20

57

53

49

45

41

37

33

29

25

21

17

13

0.00
Months(93-97)

Obs min N kg/ha

Sim min N kg/ha

Monthly calibration of nitrogen at Hico, Bosque Watershed, TX

Mineral Nitrogen Calibration


Scenario 1
Mineral Nitrogen
0.60
Observed

Mineral N kg/ha

0.50

Simulated

0.40
0.30
0.20
0.10
0.00
1

Time in Months

Mineral Nitrogen Calibration


Model consistently under predicts the
mineral nitrogen
Mineral Nitrogen

0.60

Observed

Low mineral nitrogen loading

Solutions

Mineral N kg/ha

0.50

Simulated

0.40
0.30
0.20
0.10
0.00
1

Calibrate mineral nitrogen loading

Time in Months

Initial concentration of the nutrient in soils increase to realistic


levels (SOL_NO3 in .sol)
Verify fertilizer application rates and make sure if fertilizer
application fraction to surface layer (FRT_LY1) = 1, if there is
surface application of fertilizer or manure (FRT_LY1 in .mgt)
Verify tillage operations in *.mgt files; crop residue coefficient
increase up to 0.10 ( RSDCO); bio-mixing efficiency decrease
down to 0.01 (BIOMIX) in .bsn
Nitrogen percolation coefficient--increase up to 1.0 (NPERCO in
.bsn)
F

Calibrate in-stream mineral nitrogen processes


Fraction of algal biomass that is nitrogen for water
quality -- increase up to 0.10 (AI1 in.wwq)

Mineral Nitrogen Calibration


Scenario 2
Mineral Nitrogen
0.60
Observed

Mineral N kg/ha

0.50

Simulated

0.40
0.30
0.20
0.10
0.00
1

Time in Months

Mineral Nitrogen Calibration


Model consistently over predicts the
mineral nitrogen

Mineral Nitrogen

0.60

Observed
Simulated

High mineral nitrogen loading

Mineral N kg/ha

0.50
0.40
0.30
0.20
0.10

Solutions
F

0.00
1

Calibrate mineral nitrogen loading

Time in Months

Initial concentration of the nutrient in soils decrease to near


zero if appropriate (SOL_NO3 in .sol)
Verify fertilizer application rates and if fertilizer is incorporated or
knifed in to lower layers, adjust FRT_LY1 down to 0.01 (FRT_LY1
in .mgt)
Verify tillage operations in *.mgt files; crop residue coefficient
decrease down to 0.01 ( RSDCO) and bio-mixing efficiency
increase to 0.4 (BIOMIX) in .bsn
Nitrogen percolation coefficient (NPERCO in .bsn)--decrease
down to 0.01
F

Calibrate in-stream mineral nitrogen processes


Fraction of algal biomass that is nitrogen for water quality
decrease to 0.06 (AI1 in.wwq)

Organic Nitrogen Calibration


Scenario 1
Organic Nitrogen
0.60
Observed

Organic N kg/ha

0.50

Simulated

0.40
0.30
0.20
0.10
0.00
1

Time in Months

Organic Nitrogen Calibration


Model consistently under predicts the
organic nitrogen

Organic Nitrogen

0.60

Low Organic nitrogen


loading

Observed

0.50
Organic N kg/ha

Simulated

0.40
0.30
0.20
0.10

Solutions
F

0.00
1

Time in Months

Calibrate organic nitrogen loading


Initial concentration of the nutrient in soils (SOL_ORGN in .sol) -increase to a reasonable level
Verify fertilizer application rates; make sure if fertilizer
application fraction to surface layer = 1, if there is surface
application of fertilizer or manure (FRT_LY1 in .mgt)

Calibrate in-stream organic nitrogen processes


Fraction of algal biomass that is nitrogen for water quality (AI1
in.wwq) -- increase upto 0.10

Organic Nitrogen Calibration


Scenario 2
Organic Nitrogen
0.60
Observed

Organic N kg/ha

0.50

Simulated

0.40
0.30
0.20
0.10
0.00
1

Time in Months

Organic Nitrogen Calibration


Model consistently over predicts the
organic nitrogen

Organic Nitrogen

0.60

High Organic nitrogen


loading

Solutions

Organic N kg/ha

Observed
0.50

Simulated

0.40
0.30
0.20
0.10
0.00
1

Time in Months

Calibrate organic nitrogen loading


Initial concentration of the nutrient in soils (SOL_ORGN in .sol)-decrease to a reasonable level
Verify fertilizer application rates and if fertilizer is incorporated or
knifed in to lower layers, adjust FRT_LY1 down to 0.01 (FRT_LY1
in .mgt)

Calibrate in-stream organic nitrogen processes


Fraction of algal biomass that is nitrogen for water quality (AI1
in.wwq)-- decrease to 0.06.

Soluble Phosphorus Calibration


Scenario 1
Soluble Phosphorus
0.60
Observed

Soluble P kg/ha

0.50

Simulated

0.40
0.30
0.20
0.10
0.00
1

Time in Months

Soluble Phosphorus Calibration


Model consistently under predicts the
soluble phosphorus

Soluble Phosphorus

0.60

Observed

Low soluble phosphorus loading

Solutions

Soluble P kg/ha

0.50

Simulated

0.40
0.30
0.20
0.10
0.00

Calibrate soluble phosphorus loading

Time in Months

Initial concentration of the nutrient in soils increase to reasonable


levels, up to 250-300ppm if manure has been applied for several
years (SOL_MINP in .sol)
Verify fertilizer application rates; make sure if fertilizer application
fraction to surface layer = 1, if there is surface application of
fertilizer or manure (FRT_LY1 in .mgt)
Verify tillage operations in *.mgt files; crop residue coefficient
increase up to 0.10 ( RSDCO); bio-mixing efficiency decrease
down to 0.01 (BIOMIX in .bsn)
Phosphorus percolation coefficient decrease down to 10 (PPERCO
in .bsn)
Phosphorus soil partitioning coefficient decrease down to 100
(PHOSKD in .bsn)

Soluble Phosphorus Calibration


Model consistently under predicts the
soluble phosphorus

Soluble Phosphorus

0.60

Low soluble phosphorus loading

Observed
0.50

Soluble P kg/ha

Simulated

0.40
0.30
0.20
0.10
0.00
1

Time in Months

Solutions
F

Calibrate in-stream soluble phosphorus processes


Fraction of algal biomass that is phosphorus for water quality
increase up to 0.03 (AI2 in.wwq)

Soluble Phosphorus Calibration


Scenario 2
Soluble Phosphorus
0.60
Observed
0.50

Simulated

Sol P kg/ha

0.40
0.30
0.20
0.10
0.00
1

Time in Months

Soluble Phosphorus Calibration


Model consistently over predicts the
soluble phosphorus

Soluble Phosphorus

0.60

Observed

0.50

Sol P kg/ha

High soluble phosphorus


loading
Solutions

Simulated

0.40
0.30
0.20
0.10
0.00
1

Time in Months

Calibrate soluble phosphorus loading


Initial concentration of the nutrient in soils decrease to near
zero if appropriate (SOL_MINP in .sol)
Verify fertilizer application rates and if fertilizer is incorporated or
knifed in to lower layers, adjust FRT_LY1 down to 0.01 (FRT_LY1
in .mgt)
Verify tillage operations in *.mgt files; crop residue coefficient
decrease down to 0.01 ( RSDCO) and bio-mixing efficiency
increase to 0.4 (BIOMIX in .bsn)
Phosphorus percolation coefficient increase up to 20 (PPERCO
in .bsn)
Phosphorus soil partitioning coefficient increase up to 200
(PHOSKD in .bsn)

Soluble Phosphorus Calibration


Model consistently over predicts the
soluble phosphorus
Soluble Phosphorus

High soluble phosphorus


loading

0.60
Observed
0.50

Simulated

0.40
Sol P kg/ha

0.30
0.20
0.10
0.00
1

Time in Months

Solutions
F

Calibrate in-stream soluble phosphorus processes


Fraction of algal biomass that is phosphorus for water quality
decrease down to 0.01 (AI2 in.wwq)

Organic Phosphorus Calibration


Scenario 1
Organic Phosphorus
0.60
Observed

Organic P kg/ha

0.50

Simulated

0.40
0.30
0.20
0.10
0.00
1

Time in Months

Organic Phosphorus Calibration


Model consistently under predicts the
organic phosphorus

Organic Phosphorus

0.60

Low organic phosphorus


loading

Observed
0.50

Organic P kg/ha

Simulated

0.40
0.30
0.20
0.10

Solutions
F

0.00
1

Time in Months

Calibrate organic phosphorus loading


Initial concentration of the nutrient in soils (SOL_ORGP in .sol) -increase to a reasonable level
Verify fertilizer application rates; make sure if fertilizer
application fraction to surface layer = 1, if there is surface
application of fertilizer or manure (FRT_LY1 in .mgt)

Calibrate in-stream organic phosphorus processes


Fraction of algal biomass that is phosphorus for water quality
(AI2 in.wwq) -- increase up to 0.03 (AI2 in.wwq)

Organic Phosphorus Calibration


Scenario 2
Organic Phosphorus
0.60
Observed

Organic P kg/ha

0.50

Simulated

0.40
0.30
0.20
0.10
0.00
1

Time in Months

Organic Phosphorus Calibration


Model consistently over predicts the
organic phosphorus

Organic Phosphorus

0.60

High organic phosphorus


loading

Organic P kg/ha

Observed
0.50

Simulated

0.40
0.30
0.20
0.10
0.00

Solutions
F

Time in Months

Calibrate organic phosphorus loading


Initial concentration of the nutrient in soils (SOL_ORGP in .sol)-decrease to a reasonable level
Verify fertilizer application rates and if fertilizer is incorporated or
knifed in to lower layers, adjust FRT_LY1 down to 0.01 (FRT_LY1
in .mgt)

Calibrate in-stream organic phosphorus processes


Fraction of algal biomass that is phosphorus for water quality -decrease down to 0.01 (AI2 in.wwq)

Calibration/Validation
Suggestion
F

Information given here are for


guidance. It is suggested to take care
to adjust the various parameters
appropriately depending on the local
watershed conditions.