This document discusses extending the Rangeland Hydrology and Erosion Model (RHEM) from hillslopes to watershed and large areas using the KINEROS2 and AGWA hydrology models. It provides an overview of KINEROS2 and AGWA capabilities for modeling hydrology, erosion, and sediment transport at various scales. It also discusses challenges in obtaining RHEM parameters over large areas and potential approaches using data from the National Resources Inventory, ecological site descriptions, remote sensing, and regression relationships. The document concludes with next steps around improving parameterization and integrating state and transition models and remote sensing data.

1. Extending RHEM from Hillslopes to Watersheds
and Large Areas with AGWA/KINEROS2
David Goodrich, Haiyan Wei, Mariano Hernandez, Ken Spaeth, Mary
Nichols, Shea Burns, Phil Guertin, Carl Unkrich
Overview
• KINEROS2 – background
• AGWA Overview and Rangeland
Modeling Tools
• RHEM Parameters
• From NRI
• From remote sensing

2. • Hydrology, erosion, sediment transport
• Event or continuous (< minute time steps)
• Approximate watershed by cascade of overland flow elements
(planar or curvilinear), channels, impoundments
• Physically-based kinematic interactive routing – infiltration,
overland flow, erosion, and sediment transport
Kinematic Runoff and Erosion Model (KINEROS2)
https://www.tucson.ars.ag.gov/kineros/

3. rainfall intensity (i)
infiltration (f)
Finite difference
step length (dx)
fi
x
Q
t
h -=

+


channel element
Excess Runoff From an Overland Flow Model Element
- KINEROS2 /
RHEM
Q
q
Zones of aggregation
and degradation can be
identified in detailed
output

4. New Features and Capabilities
• Incorporation of the Rangeland Hydrology and Erosion Model
(RHEM) and a dynamic (non-steady state) version of the Water
Erosion Prediction Project Model (DWEPP) into KINEROS2.
– Detailed representation of hillslope elements (geometry, soil,
vegetation)
– DWEPP parameters are available for most cultivated agriculture
– RHEM parameters coming with ESD and STM models
– Seamless integration between hillslope erosion modeling and
sediment transport in channels
– Caution: 1-D channel erosion and transport is quite simple
(bank collapse, armoring not treated)
– Make a distinction between
large area modeling (many
hillslopes within a watershed)
and large watershed modeling
(sediment yield at outlet)

5. Southwest Watershed Research Center Tucson - Tombstone, AZ
EXAMPLE: KINEROS2 / RHEM Simulations
Compare sediment yields from different plant
community parameterizations on a small
watershed using K2 - RHEM
Approach
• Parameterize a small watershed for RHEM
• Use RHEM parameterization equations with values
from NRI data for 3 plant communities
• Use observed data from storm events that produced
sizeable sediment yields
• Run RHEM and compare results from different plant
community parameterizations

6. Southwest Watershed Research Center Tucson - Tombstone, AZ
K2 - RHEM Watershed Simulation – 8/6/2006 Event
0
20
40
60
80
100
120
140
1600
10
20
30
40
50
60
70
0 10 20 30 40 50 60
Rainfall(mm/hr)
Outflow(mm/hr)
Time (min)
Precip
Natives
Shrubs
Lehmann
0
20
40
60
80
100
120
140
1600
2
4
6
8
10
12
0 10 20 30 40 50 60
Rainfall(mm/hr)
Sedigraph(kg/s)
Time (min)
Precip
Natives
Shrubs
Lehmann
Lehmann vs Natives Shrubs vs Natives
Runoff (mm / hr) Sediment (kg / s)

7. Watershed ID: 73
Area
Slope Profile (Geometric)
Width
Slope Length
Effective Hyd. Cond.
Rock Volume (Infiltration)
Soil Suction
Porosity
Max saturation
Runoff friction factor (Hydraulic)
Erosion friction factor
Conc. Flow Erosion Coef. (Engelund-
Critical Shear Stress Hansen
Splash & Sheet Erosion Coef. Erosion)
Particle Size Distribution
Interception
N
Watershed - Large Area Application of KINEROS2
・ Topography (USGS 10 m DEM)
・ Soils for texture class – (SSURGO)
・ For RHEM need
 % plant form types
 % Foliar cover
 % Basal plant cover
 % Rock cover
 % Litter cover
 % Cyptogram cover
Parameters needed for each
hillslope model element
Pre – RHEM Formulation
For K2 – RHEM Formulation need:
Motivation for the development of AGWA – Automated
Geospatial Watershed Assessment Tool

8. AGWA – Background - Basics
• An automated GIS interface for watershed modeling (hydrology,
erosion, WQ) designed for resource managers
• Applicable to ungauged / gauged watersheds
• Operates with nationally available data (DEM, Soils, Land Cover)
• Simple, direct method for model parameterization
• Investigate the impacts of land cover change
- Identify sensitive, “at-risk” areas
- Assess impacts of management (e.g. growth, fire, mulch)
• Provide repeatable results for relative change assessments
• Must have good rainfall-runoff observations for quantitative
predictions
• Three established watershed/hillslope models for multiple scales
- KINEROS2
- SWAT
- RHEM / DWEPP (hillslope runoff and erosion within KINEROS2)
- 4000+ Reg. users; 10,500+ downloads, 170 countries; >250
citations

9. Watershed Characterization
(model elements)
+
Land
Cover
Soils
Rain
(Observed or
Design Storm)
Results
Run model
and import
results
Intersect model
elements with
Watershed Delineation
using Digital Elevation
Model (DEM)
Sediment yield (t/ha)Sediment discharge (kg/s)
Water yield (mm)Channel Scour (mm)
Transmission loss (mm)Peak flow (m3/s or mm/hr)
Channel Disch. (m3/day)Sediment yield (kg)
Percolation (mm)Runoff (mm or m3)
ET (mm)Plane Infiltration (mm)
Precipitation (mm)Channel Infiltration (m3/km)
SWAT OutputsKINEROS Outputs
AGWA Conceptual Design:
Inputs and Outputs
Output results that can be displayed in AGWA
Hillslope

10. Visualization of Results
Color-ramping of results
for each element to
show spatial variability
Calculate and view
differences between
model runs
Multiple simulation runs
for a given watershed
Channel simulation
differences also
displayed
8
Hydrograph/Sedigra
ph for overland and
channel elements

11. Data for AGWA Parameterization
• Digital Elevation Model
- Usually USGS 10m – 30m DEM will work
fine in western terrains in large watersheds
- LIDAR can be used
• Soils
- USDA STATSGO – nationally available;
SSURGO where available
- FAO soils globally
• Land Use - Land Cover (NLCD, ReGAP, ESD)
• Weather
- If not using design storms - “good” rainfall
data is essential in time/space (more later)
• Management Information
- Where and what
- Information must be provided by user!
(i.e. burn severity map)
Topography
Land Cover
Soils

12. o
Representative Slope Profile and Flow Length
• Calculate a weighted length for each flow path
• Lr is the representative flow length
• Calculate a weight grid
for every cell on the
hillslope
• Calculate a weighted
slope for each cell - Si
NRI Pt
(Flanagan et al. 2011)

13. o
Representative Slope Profile and Flow Length
NRI Pt
• AGWA/RHEM
Calculates a
weighted length
and slope for each
flow path using
DEM
(Flanagan et al. ‘11)
1374
1376
1378
1380
1382
1384
1386
1388
1390
1392
0 100 200 300 400 500 600 700 800 900 1000
Hillslope Profile (Element 61)
AGWA/Flanagan (S=1.6%)
Visual Slope Profile (S=1.5%)
Uniform Slope (S=1.8%) L = 50m
web RHEM Gen. Profile (S=1.7%)
Ele. 61 At NRI point
estimate uniform
slope, L = 50 m
(used in RCA)
Most natural
hillslope proflie
have lower slopes
at ridge and toe.
Digitize 3 flow
paths from ridge
Using web RHEM
to match general
hillslope profile

14. Sediment Yield vs Soil Loss (total for all elements)
Uniform S - HRC Unif. S - ATL Complex S - HRC Comp. S - ALT
SedimentYieldandSoilLoss(tons/ha/year)
0
3
6
Sediment Yield
Soil Loss
Point: Differences in soil loss & sediment yield due to slope shape
(uniform vs complex) dominate differences due to change in ecosite
cover condition

15. AGWA Rangeland Management Toolkit
• Use of vegetation data for model parameterization
– Vegetation Monitoring Data
– Rangeland Health Data
– NRCS Ecological Site Descriptions
– Management treatment effects
• Land cover modification
– Use to incorporate rangeland improvements
• Stock ponds / reservoirs
• Buffer strips
• Post-Fire effects
• Multi-watershed analysis
Post-Fire % Change
} Or a Combination

16. Land-Cover Modification Tool
Allows user to specify type and location of land-cover
alterations by either drawing a polygon on the display, or
specifying selected features from a polygon map (i.e. a
pasture).
Types of Land-Cover Changes:
• Change entire user-defined area to new land cover (uniform)
• Change land-cover type to another (random or
patchy/fractal)
• Can specify % success of change due to practice (e.g.
Shrub management and removal)
Large Area Application of RHEM
• AGWA will –
• Delineate drainage area
• Discretize drainage area into hillslopes
• Derive the complex slope profile of the hillslopes
• AGWA tools will –
• Allow incorporation of BMPs
• Assess management scenarios
• Transition to different ESD state

17. Key to Extending RHEM to Watershed Scales
・Obtaining RHEM parameters over large areas:
- % plant form present - % Foliar cover - % Basal cover
- % Rock cover - % Litter cover - % Cyptogram cover
・ HOW? (NRI, ESD & STM, ReGAP, remote sensing, ??)
HUC – 12 Boundary
0 – 10 % INCREASE
25 – 30 % decrease
40 – 50 %
decrease
0 – 10 % DECREASE
n = 3577 hillslope
model elements
Example: Cienega Ck. CEAP Assessment (198 mi2)
% Change in Sediment Yield from Pre-Conservation Condition (‘92) to Post-Cons. (2006)
RHEM derived from NRI & PLANTS database by Ecological Site
for three time periods: Historic Ref Cond., Pre- & Post Cons. Spending

18. Ecological Sites
Loamy Upland 12-16” p.z.
Allotments
Until machine readable ESD and STM are available nationally what
can be done to obtain RHEM parameters over large areas?
 How can NRI and local NRCS observations be utilized ?
 Could a expansion factor be developed using EcoSites and visual
interpretation of aerial/satellite imagery + AGWA cover change
tool?
 Can multi-band remote sensing assist ?
**
*
NRCS Field Pt. #1
NRCS Field Pt. #2

19. From Matt Jones
2018 SRM meeting

20. Remotely Sensed (Landsat) Veg. Cover
 Provides cover estimates
throughout the year (~16-day
intervals, 30 x 30 m) from 1984
to present
 For plant phenology in SW can
use early summer image for
shrub (mesquite canopy cover-
CC) using SAVI and fall image
for total CC post monsoon with
SATVI
 Can then subtract the two
images to estimate non-shrub
green and senescent
cciTNC = 0.69(ccg) + 9.24
R2
= 0.40*
cciARS = 0.37(ccg) + 45.08
R2
= 0.47*
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
Ground-Measured Foliar Canopy Cover (%)
Image-BasedFoliarCanopyCover(%)
TNC
ARS
Linear
(ARS)* Value is statistically significant (P ≤ 0.05).
1-to-1
(All data)
(ARS data only)
1991 2007 2008 2011
CC (%)
1991 2007 2008 2011
CC (%)
Shrub Canopy CoverHollifield-Collins

21. Foliar Cover to Litter and Basal
 For some ES,
defensible
regression
between foliar
cover and % litter
and % basal have
been found using
NRI data
 Allows higher
temporal
frequency of
RHEM parameter
than NRI collects
% change in SY from ’92 to ‘07
% change in SY from ’92 to ‘09
% change in SY from ’92 to ‘11
Negative % change in SY => a decrease
-80 - 75
-5 - 0
50 - 250

22. Southwest Watershed Research Center Tucson - Tombstone, AZ
Conclusions and Next Steps
• AGWA is ready to ingest national EcoSite and State & Transition
Model data to automatically derive RHEM parameters for modeling
multiple hillslopes over large areas when it becomes available
• In the meantime -
• The AGWA Land Cover Change tool can be used to evaluate the
hydrologic/erosion effects of management scenarios over large areas
• Remote sensing (RS) can be used to effectively
• Track the changes in foliar cover over time
• Quantify shrub and non-shrub foliar cover changes from
• Mechanical brush removal and fire
• Next steps
• With more NRI points attempt to build better regressions relating
foliar cover to litter and basal cover for more ecosites.
• Explore how RS & ES could be used to develop an image of states
within a simplified STM model for that ES

23. THANK YOU - Questions
I’d like to hear your thoughts and suggestions on obtaining
RHEM parameters over large areas
Dave Goodrich
Dave.Goodrich@ars.usda.gov
AGWA Training (free) at Univ. of Arizona
(Typically Oct. and March)
https://www.tucson.ars.ag.gov/kineros/
https://www.tucson.ars.ag.gov/agwa/

25. 1991 2007 2009 2011
CC (%)
1991 2007 2009 2011
CC (%)
exclosure
grazed
pasture
Exclosure Burn - 2009 Grazed
1991
2007
2009
2011
 Prescribed burn in 2009
outlined by blue
polygon
 Drought in summer
monsoon in 2009
 No grazing in exclosure
 If letters above bars are
different they are
statistically different at
the (P=0.05) level

26. Ecosites &
SWReGAP
(lumped)
classes

27.  Mechanical
grubbing in 2010 in
5 areas
 2011 wildfires in
green outline
covering the corral
and NE portion of
the mule treatment
areas
1991
2007
2009
2011

29. Remotely Sensed (Landsat) Veg. Cover
 Provides cover estimates
throughout the year (~16-day
intervals, 30 x 30 m) from 1984
to present
 For plant phenology in SW can
use early summer image for
shrub (mesquite canopy cover-
CC) using SAVI and fall image
for total CC post monsoon with
SATVI
 Can then subtract the two
images to estimate non-shrub
green and senescent
cciTNC = 0.69(ccg) + 9.24
R2
= 0.40*
cciARS = 0.37(ccg) + 45.08
R2
= 0.47*
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
Ground-Measured Foliar Canopy Cover (%)
Image-BasedFoliarCanopyCover(%)
TNC
ARS
Linear
(ARS)* Value is statistically significant (P ≤ 0.05).
1-to-1
(All data)
(ARS data only)
1991 2007 2008 2011
CC (%)
1991 2007 2008 2011
CC (%)
Shrub Canopy Cover

30. Stock Tanks
• A common rangeland management is the installation of
stock ponds to provide water to livestock. Stock ponds
can also be viewed as storm water retention structures
and sediment basins, common best management
practices for flood and water quality mitigation.