CEE 795

Water Resources Modeling and GIS

Lecture 8: Hydrologic Modeling and AGWA
April 3, 2006

Learning Objectives:
• Describe the steps in hydrologic modeling
• Evaluate different types of hydrologic models
• Summarize the components of AGWA

Handouts: Assignments: 1
Design Point

3
5

6
Hydrologic/Watershed Modeling

Design Point
1

3
5

Thomas Piechota, Ph.D., P.E.

Department of Civil and Environmental Engineering
University of Nevada, Las Vegas
piechota@[Link]
 Definitions

• Watershed: area that topographically

contributes to the drainage to a point of
interest
• Streamflow: runoff (rate or volume) at a
specified point in a watershed.
• Hydrologic budget: accounting of water
in a system.
Conceptual Model of Watershed Modeling
Typical Input
• Topography
• Soil Characteristics
• Land cover
• Land use
• Meteorological data

Typical Output
• Streamflow
• Subsurface Flow
• Depth to water table
 Steps to Hydrologic Modeling
1. Delineate watershed
2. Obtain hydrologic and geographic data
3. Select modeling approach
4. Calibrate/Verify model
5. Use model for assessment/prediction/design
 What is a Watershed?
• Area that topographically contributes to the
drainage to a point of interest
Natural Watershed

Points of Interest
• Road crossing
• Stream gage
• Reservoir inlet
• Wastewater treatment plant
• Location of stream restoration
Urban Watershed
98
100

98

100 98
103
100

108

110

105

98
99

97

100

103
100 Design Point
USGS Quad Map
 Digital Elevation Model (DEM)
• Digital file that stores the elevation of the land
surface a specified grid cell size (e.g., 30 meters)
 Steps to Hydrologic Modeling
1. Delineate watershed
2. Obtain hydrologic and geographic data
3. Select modeling approach
4. Calibrate/Verify model
5. Use model for assessment/prediction/design
 Geographic Data
• Land cover • Land use
Land Cover
Forest
Oak Woodlands
Mesquite Woodlands
Grasslands
Desertscrub
Riparian
Agriculture
Urban
Water
Barren / Clouds

0 5 10 km

1992 NALC Hillshade DEM STATSGO

 Geographic Data
• Soil type/classification

EM STATSGO
 Hydrologic Data
• Meteorological Data
– Temperature
– Precipitation
– Wind speed
– Humidity
• Extrapolation of point
measurements
– Theissen Polygons
– Inverse distance weighting
 Hydrologic Data Streamflow
• Hydrologic Data Design Point
1

– Streamflow 2

• Peak discharge
• Daily flow volume 3
5

• Annual flow volume

– Soil moisture 6

– Groundwater level
 Steps to Hydrologic Modeling
1. Delineate watershed
2. Obtain hydrologic and geographic data
3. Select modeling approach
4. Calibrate/Verify model
5. Use model for assessment/prediction/design
 Modeling Approaches (examples)
TIME SCALE
Event-based Continuous Simulation
(minute to day) (days – years)
Empirical
Rational Method
Regression equ’s
SCS Unit Hydrograph Simple Model
Transfer Functions
Simple models
Physically-based
KINEROS
Based on physical
Stanford Watershed SWAT
processes
Model VIC-3L
Complicated
TOPMODEL TOPMODEL
Many parameters
Basis for Many Hydrologic Models

• Hydrologic Budget
Transpiration (T) (In – Out = Streamflow (Q)
ΔStorage)
Evaporation (E)
Groundwater out (GW out )
Reservoir
Precipitation (P)
Groundwater in (GWin)

Infiltration (I)

Watershed

(P + GWin) – (E + T + I + GWout + Q) = ΔStoragereservoir

Which Model Should be Used?

• It Depends on:
– What time scale are you working at?
– What hydrologic quantity are you trying to
obtain?
– What data do you have for your watershed?
– How fast of a computer do you have?
 Spatial Scaling of
Semi-Distributed
Models
Parameters assigned to
Lumped each grid cell, but cells Fully-Distributed
Parameters assigned to with same parameters are Parameters assigned to
each subbasin grouped each grid cell

A3

A1
A2
 Stanford Watershed Model
(HSPF)
• Physically-based and continuous simulation
STANFORD WATERSHED MODEL
Process

Input
Potential ET
Precipitation Output
Actual ET Temperature Storage
Radiation Decision
Wind,Dewpoint ET - Evapotranspiration
* Parameters
n Order taken to
Snowmelt meet ET demand

CEPSC*
2 LSUR*
ET Interception Delayed Infiltration SLSUR*
Storage NSUR*
INFILT* Overland Flow
Direct
3 ET Infiltration UZSN* INTFW*
IRC*
LZSN* Upper Zone
Interflow
Lower Zone Storage
5 ET Storage PERC
LZETP* DEEPFR* AGWRC*
Groundwater
4 ET Storage
Deep or Inactive AGWETP*
1 ET Groundwater BASETP*
To Stream
 Kinematic Runoff and Erosion Model
(KINEROS)
• Developed by USDA
• [Link]
• Event oriented & physically based
• Describes the processes of
interception, infiltration, surface
runoff and erosion
 TOPMODEL
• Semi-distributed &
physically-based Evapotranspiration

• Relates hydrologic Precipitation

processes (e.g., overland Source Area

flow, subsurface flow) to
topographic characteristics Infiltration
of watershed Macropore
Overland
Flow

• Efficiency of lumped model Flow

and physical theory of a Drainage

Subsurface Flow
distributed model Total
Flow
TOPMODEL Example

Pacific Ocean
 Variable Infiltration Capacity (VIC-3L)
• Continuous simulation and physically-based
• Macroscale hydrologic model that solves full water
and energy balances
VIC-3L Example
Anamoly Three Layers Soil Moisture
( Upper Mississippi Basin)

200
100

Moisture (inch)
Anomaly Soil
0
Jan-50
-100 Sep-63 May-77 Feb-91 Oct-04

-200
Time (Month)
layer1 layer2 layer3
 Steps to Hydrologic Modeling
1. Delineate watershed
2. Obtain hydrologic and geographic data
3. Select modeling approach
4. Calibrate/Verify model
5. Use model for assessment/prediction/design
 Calibrating a Model
• Typically the model is calibrated against
observed streamflow data
• Depending on the model complexity,
parameters are adjusted until observed
streamflow equals model streamflow
• Which observed value to use:
– Qpeak
Qpeak
– Qvolume
Q tpeak Qvolume
– tpeak

t
 Sensitive Parameters

• Precipitation
• Soil parameters
– Hydraulic conductivity
– Soil water holding capacity
• Evaporation (for continuous simulation)
• Flow routing parameters (for event-
based)
 Uncertainties
• Precipitation
– Extrapolation of point to other areas
– Temporal resolution of data
• Soils information
– Surveys are based on site visits and then
extrapolated
• Routing parameters
– Usually assigned based on empirical studies
 Steps to Hydrologic Modeling
1. Delineate watershed
2. Obtain hydrologic and geographic data
3. Select modeling approach
4. Calibrate/Verify model
5. Use model for assessment/prediction/design
 Use of Models

• Assessment
– What happens if land use/land cover is
changed?
• Prediction
– Flood forecasting
• Design
– How much flow will occur in a 100 year
storm?
AUTOMATED GEOSPATIAL WATERSHED ASSESSMENT
A GIS-BASED WATERSHED MODELING TOOL

William Kepner and Darius Semmens

US – EPA Landscape Ecology Branch Las Vegas, NV

David Goodrich, Mariano Hernandez, Shea Burns,

Averill Cate, Soren Scott, and Lainie Levick
USDA-ARS Southwest Watershed Research Center, Tucson, AZ

Phillip Guertin
University of Arizona, Tucson, AZ

Scott Miller
University of Wyoming, Laramie, WY
 Project Background & Acknowledgements
• Long-Term Research Project
– Landscape Ecology Branch USDA-ARS
– 5 years David Goodrich
• Interdisciplinary Mariano Hernandez
– Watershed management
– Landscape ecology Averill Cate
– Atmospheric modeling Ian Burns
– Remote sensing
– GIS
Casey Tifft
• Multi-Agency Soren Scott
– USDA – ARS US-EPA
– US – EPA
– University of Arizona Bill Kepner
– University of Wyoming Darius Semmens
– USGS
•
Dan Heggem
Student Support
– 2 Post-Doc Bruce Jones
– 2 PhD Don Ebert
– 2 Masters
University of Arizona
Phil Guertin
University of Wyoming
Scott Miller
 Introduction
• PC-based GIS tool for watershed modeling
– KINEROS & SWAT (modular)
• Investigate the impacts of land-use/cover change on
runoff, erosion, and water quality at multiple scales
• Compare and visualize results
• Targeted for use by research
scientists and management
specialists
• Useful in conducting TMDL
analyses
• Widely applicable
Soil and Water Assessment Tool
(SWAT)
• Daily time step
• Distributed: empirical and physically-based model
• Hydrology, sediment, nutrient, and pesticide yields
• Larger watersheds (> 1,000 km2)
• Similar effort used by BASINS
Abstract Routing Representation
71
73
to next channel
73 71
73 channel 73
pseudo-
channel 71
Kinematic Runoff and Erosion Model
(KINEROS2)
• Event-based (< minute time steps)
• Distributed: physically-based model with
dynamic routing
• Hydrology, erosion, sediment transport
• Smaller watersheds (< 100 km2) Abstract Routing Representation

71 71
72

74 73 73

72
74
AGWA ArcView Interface
 AGWA Conceptual Design:
Inputs and Outputs
Watershed Delineation
using Digital Elevation Watershed Discretization Intersect model
Model (DEM) (model elements)
+ elements with

Soils

Land
Cover

Output results that can be displayed in AGWA

KINEROS Outputs SWAT Outputs
Channel Infiltration (m3/km) Precipitation (mm) Rain
(Observed or
Plane Infiltration (mm) ET (mm) Design Storm) Run model and
Runoff (mm or m3) Percolation (mm) import results

Sediment yield (kg) Channel Disch. (m3/day)

Peak flow (m3/s or mm/hr) Transmission loss (mm)
Channel Scour (mm) Water yield (mm) Results

Sediment discharge (kg/s) Sediment yield (t/ha)

 Navigating Through AGWA
Generate Watershed Outline grid

Subdivide Watershed Into Model Elements polygon

SWAT Choose the model to run KINEROS

Intersect Soils & Land Cover look-up tables

Generate rainfall input files

Daily Rainfall from… Storm Event from…
Gauge locations NOAA Atlas-II
Thiessen map Pre-defined return-period / magnitude
Pre-defined continuous record “Create-your-own”
 Navigating Through AGWA, Cont’d…
Prepare input data

Subwatersheds & Channels Channel & Plane Elements

Continuous Rainfall Records Event (Return Period) Rainfall

external to
Run The Hydrologic Model & Import Results AGWA

Display/Compare Results

SWAT outputs: KINEROS outputs:

•Runoff, water yield (mm) •Runoff (mm,m3)
Visualization for
•Channel Discharge (m3/day) •Sediment Yield (kg/ha)
•Evapotranspiration (mm) each model
•Infiltration (mm)
•Percolation (mm) element
•Transmission losses (m3/km)
•Transmission Losses (mm) •Peak runoff rate (m3/s)
•Sediment Yields (mm) •Peak sediment discharge (kg/s)
SWAT Parameter Estimation
- Example: Curve Number from NLCD land cover
 Higher numbers result in higher runoff

CURVE NUMBER
NLCD Hydrologic Soil Group
Land cover A B C D Cover (%)
High intensity residential (22) 81 88 91 93 15
Bare rock/sand/clay (31) 96 96 96 96 2
Forest (41) 55 75 80 50
Shrubland (51) 63 77 85 88 25
Grasslands/herbaceous (71) 80 87 93 70
Small grains (83) 65 76 84 88 80
KINEROS Parameter Estimation
 Parameters based on soil texture (STATSGO, SSURGO, FAO)
Texture Ksat Suction Porosity Smax CV Sand Silt Clay Dist Kff
Clay 0.6 407.0 0.475 0.81 0.50 27 23 50 0.16 0.34
Fractured Bedrock 0.6 407.0 0.475 0.81 0.50 27 23 50 0.16 0.05
Clay Loam 2.3 259.0 0.464 0.84 0.94 32 34 34 0.24 0.39
Sandy Clay Loam 4.3 263.0 0.398 0.83 0.60 59 11 30 0.40 0.36
Silt 6.8 203.0 0.501 0.97 0.50 23 61 16 0.23 0.49
Loam 13.0 108.0 0.463 0.94 0.40 42 39 19 0.25 0.42
Sandy Loam 26.0 127.0 0.453 0.91 1.90 65 23 12 0.38 0.32
Gravel 210.0 46.0 0.437 0.95 0.69 27 23 50 0.16 0.15

 Parameters based on land-cover classification (e.g. NLCD)

Land Cover Type Interception (mm/hr) Canopy (%) Manning's n

Forest 1.15 30 0.070
Oak Woodland 1.15 20 0.040
Mesquite Woodland 1.15 20 0.040
Grassland 2.0 25 0.050
Desertscrub 3.0 10 0.055
Riparian 1.15 70 0.060
Agriculture 0.75 50 0.040
Urba n 0.0 0.0 0.010
AGWA Soil Weighting (KINEROS)
Intersection of model
• Area and depth weighting of soil element with soils map
parameters
• Area weighting of averaged AZ061
MUID values for each watershed
element

AZ076
Components for MUID AZ061
Component 2
AZ067
45% Component 3
35%

Component 1 Layers for component 3

20%
Layer 1 2

Layer 2 2 9 inches

Layer 3 5
Parameter Manipulation (optional)

Stream channel attributes

Ksat

Can manually
change parameters
for each channel
and plane element

Upland plane attributes Ksat

Visualization of Results
Multiple simulation runs Calculate and view
for a given watershed differences between
model runs

Color-ramping of
results for each
element to show Automated tracking of
spatial variability simulation inputs
 Spatial and Temporal Scaling of Results
 Using SWAT and KINEROS for integrated watershed assessment
 Land cover change analysis and impact on hydrologic response

Upper San Pedro High urban growth Sierra Vista Subwatershed

River Basin 1973-1997
KINEROS Results
Concentrated urbanization

ARIZONA

Phoenix
#

Tucson #

SONORA
N

Water yield change Forest

between 1973 and 1997 Oak Woodland
Mesquite
<<WY >>WY Desertscrub
Grassland
SWAT Results 1997 Land Cover Urban
 Urbanization Effects (KINEROS2)

Pre-urbanization Post-urbanization
1973 Land cover 1997 Land cover

• Results from pre- and post-urbanization simulations

using the 10-year, 1-hour design storm event
Limitations of GIS - Model Linkage
• Model Parameters are based on look-up tables
- need for local calibration for accuracy
- FIELD WORK!
• Subdivision of the watershed is based on topography
- prefer it be based on intersection of soil, lc, topography
• No sub-pixel variability in source (GIS) data
- condition, temporal (seasonal, annual) variability
- MRLC created over multi-year data capture
• No model element variability in model input
- averaging due to upscaling

Most useful for relative assessment unless calibrated

 Land-Cover Modification Tool
 Allows users to build management scenarios
 Location of land-cover alterations specified by either drawing a polygon on
the display, or specifying a polygon map

Types of Land-Cover Changes:

• Change entire user-defined area to new land cover
• Change one land-cover type to another in user-defined area
• Change land-cover type within user-supplied polygon map
• Create a random land-cover pattern
• e.g. to simulate burn pattern, change to 64% barren, 31% desert scrub,
and 5% mesquite woodland
 Alternative Futures: Base Change Scenarios
1. CONSTRAINED – Assumes population increase less than 2020 forecast
(78,500). Development in existing areas, e.g. 90% urban.
2. PLANS – Assumes population increase as forecast for 2020 (95,000).
Development in mostly existing areas, e.g. 80% urban and 15% suburban.
3. OPEN – Assumes population increase more than 2020 forecast (111,500).
Most constraints on land development removed. Development occurs mostly
into rural areas (60%) and less in existing urban areas (15%).
 Percent Change in Runoff under Future Scenarios
• There is considerable variation – particularly between extremes
produced by constrained and open scenarios (Kepner et al., 2004)

• Surface runoff will

increase in all three
scenarios

• Sediment yield will

increase especially
as new surfaces are
disturbed and
surface runoff
increases

(Derived from using

future land covers
and AGWA)
Plans
 Applications of National & International
Significance
National
• NYCDEP – Catskill/Delaware watershed
assessment 

• Upper San Pedro Partnership – watershed
planning, cost-benefit analysis
 
• EMAP – Oregon (AGWA-ATtILA)  integrated
alternative futures assessment  
• ReVA – SEQL alternative futures 
• EPA Region 9 – CWA 404 and NEPA
• EPA Region 10 – 404/NEPA, transportation planning
• NWS – Real-time flood warning
• USFS – Post-fire assessment & rehabilitation planning
• AZ – State is using AGWA for TMDL planning and education of municipal officials

International
• NATO Committee on the Challenges to Modern Society (CCMS) – Integrated
hydrologic/ecological landscape change assessment
• Southwest Consortium for Environmental Research and Policy (SCERP) –
U.S./Mexico trans-border watershed management
• UNESCO Global Network for Water and Development Information (G-WADI) –
International arid-region hydrologic modeling
AGWA Milestones
• AGWA 1.1 released at the Fed. Interagency
Hydrologic Modeling Conference, July 2002
• Externally peer-evaluated through two separate federal review
processes (EPA/600/R-02/046 & ARS/137460)
• AGWA added to
• EPA Council for Regulatory Environmental Modeling (CREM) database
• NASA Applied Sciences Directorate model and analysis systems
• USGS Surface-water Modeling Interest Group archives

• AGWA 1.4 released in July, 2004

• AGWA integrated into BASINS 3.1 release, August 2004
• Training – national and international
• Free public download and full documentation via parallel EPA
and ARS web sites
• 1200+ registered users (excluding BASINS users)
 AGWA Support & Distribution
• Fact Sheets, Product Announcement, Brochures
• Documentation and User Manual
• Quality Assurance Report

Research Plan

Code Structure (Avenue
Scripts, Dialogs, System
Calls)

EPA and USDA/ARS
companion Websites

Journal Publications
(Hernandez et al. 2000, Miller
et al. 2002a, Miller et al.
2002b, Kepner et al., 2004)

Training: Las Vegas (2001);
Reston (2002); Tucson
(2003); San Diego (2004)

• AGWA Web Sites

[Link]
[Link]
 Future Directions
Migrating to ArcGIS (AGWA 2.0) and the Internet (DotAGWA)
• Final ArcView version (AGWA 1.5) release at FIHMC (April 2006)
• Detailed, peer reviewed design plan for AGWA migration to ArcGIS
and Internet completed April, 2005
• Beta-release of ArcGIS and Internet versions, 2006
• Final ArcGIS and Internet release with full documentation, 2007

Integration of additional models

• Opus – USDA-ARS integrated simulation model for transport of non-
point source pollutants (2007)
• MODFLOW – USGS ground-water model will be coupled with AGWA-
KINEROS surface-water model (planning meeting 2006)
• GAP habitat models – integrated hydrologic and ecological
assessments (proposal pending)