Modeling Landscape Fluxes

1
Modeling Landscape Fluxes in Open source GIS
Helena Mitasova Dept. of
Marine, Earth and Atmospheric Sciences, NCSU,
Raleigh L. Mitas, J. Hofierka http//www.skagit.m
eas.ncsu.edu/helena/
2
Modeling fluxes in GIS
Two related but different tasks - watershed
analysis - analyze the geometry of DEM to
find the location of streams and watershed
boundaries dry conditions flow pattern -
water flow (hydrologic) simulation - use DEM,
stream, land cover, rainfall and soil data to
simulate the distribution of water flow by
solving the governing equations storm
event flow pattern Example difference in
depressions - the watershed/stream analysis
program will try to find location of a stream
(single cell width) - the process based program
will flood the area
3
Watershed analysis in GIS
A wide range of algorithms available with
different properties - direction of flow D8
(low res.), Dinf (high res.) - handling of
depressions (important for IFSARE, LIDAR)
fill-in, carve-in, shortest path - routing SFD
(1D) along flowline, MFD
fill-in
sh. path
carve-in
D8 Dinf
4
Watershed analysis flow accumulation
Flow accumulation from r.watershed using D8, SFD
D8 6m resolution
D8 20 resolution
H. Mitasova
5
Watershed and flow pattern analysis
Vector-grid, D-infinite 6m resolution
watershed boundaries from r.watershed flow
accumulation and flowlines from r.flow
6
Resolving large depression in IFSARE DEM
most common approach Fill-in (Rivertools,
ArcGIS)
shortest path (r.watershed)
Both are D8, 10m resolution DEM
7
The flat area isn't really flat, it was created
by DEM filling
3D view of the original and filled DEM with
streams derived by Rivertools
trend towards carve-in algorithms for lidar
and IFSAR DEMs
8
Watershed analysis for massive DEMs
Laura Toma, Lars Arge, Duke University r.terraflow
3 hr - 10m resolution entire Panama 10,0002
cells D8, SFD/MFD, fill-in
r.watershed killed after 17 days, 12 hours
3,0002 cells, but more
accurate results Terraflow outperforms also
ArcGRID, TARDEM and other tools
r.watershed r.terraflow rivertools measured sites
9
Water flow over complex terrain using modified
r.flow
Animation illustrates a simple, kinematic wave
overland water flow simulation for uniform
rainfall, soil and cover conditions. Created by
customized version of r.flow by writing output
flowline densities after passing given number of
cells. Depressions are handled as
sinks. Geometry-based flow analysis tools are
solutions of water flow equations for special
cases (e.g., uniform flow velocity)
H. Mitasova
10
Sediment flow at transport capacity
Combines series of kinematic wave overland flow
surfaces with slope using r.mapcalc to illustrate
simulation of sediment flow
H. Mitasova
11
Water flow modeling
Shallow overland flow - St. Venant equation
dynamic kinematic wave
steady state approx. diffusive wave
h - water depth, ie - rainfall excess, v -
velocity given by Mannings eq., ? - diffusion
const
kinematic wave
kinematic wave with predefined channel in the
depression
diffusive wave
H. Mitasova
12
Water flow over complex terrainapproximate
diffusive wave
Water flow depth simulated by approximate
diffusive wave using path sampling
H. Mitasova
13
Path sampling method
- based on duality between particle and field
representation - path
samples represent water or sediment evolving
according to the shallow-water
bivariate continuity equation
- solved by
operator inversion Green's function
representing short time propagation of the
sample points (drift diffusion)
Land cover red - disturbed areas green - forest
Water depth particles continuous
fieldparticle density
Rainfall excess, land cover and topography are
spatially variable
14
Sediment transport
Sediment transport and net erosion/deposition
model is based on 2D generalization of a 1D
hillslope erosion model used in WEPP.
where D (r) ? (r)T (r)-qs(r) is
sources-sinks term and D (r)/Dc(r)
qs(r)/T (r)1 (Foster and Meyer, 1972)
c - sediment concentration, qs - sediment flow,
T - transport capacity, Dc- detachment
capacity, ? first order reaction coef.
Net erosion and deposition is computed as a
divergence of sediment flow.
15
Sediment flow for different soils
sand ? ? 1
clay ? ? 0.1
detachment capacity limited case DCL
transport capacity limited case TCL
16
Path sampling method properties
- robust and flexible based on stochastic
representation - multi-dimensional - scalable
accuracy determined by the number of samples,
ideal for distributed computing
17
Erosion/deposition 2D flow
ARKLSCP
Observed depth of colluvial deposits
USLE/RUSLE erosion based on slope length
erosion deposition
1D change in sediment flow
2D change in sediment flow
Modeled distribution of erosion and deposition
18
Impact of detachment capacity coef. on pattern
of sediment flow and erosion/deposition
Changing detachment capacity while transport
capacity stays constant with increasing
detachment deposition increases and sediment
output remains the same (DCL -gt TCL case)
19
Impact of transport capacity coef. on pattern of
sediment flow and erosion/deposition
Changing transport capacity while detachment
capacity stays constant with increasing
transport capacity deposition decreases and
sediment output increases (TCL -gt DCL case)
20
Impact of parameterscritical shear stress
21
Multiscale simulations
Simulation of overland water flow 10m
resolution combined with 2m resolution
22
Implementation in GRASS and Applications
Original FORTRAN program SIMWE was implemented in
GRASS as r.sim.water r.sim.sediment
Applications focus on spatial pattern of water
and sediment flow and net erosion/deposition
under spatially variable conditions studies of
land use change impact, design of conservation
measures.
H. Mitasova
23
Water flow with depressions application to
drainage planning
Spatial distribution of water depth after 1 hour
rainfall. Based on 6m resolution DEM derived from
RTK-GPS survey
Spatial distribution of water depth 24 hours
after rainfall with and without drainage
H. Mitasova
24
Optimizing conservation measures
Original LU
sediment flow
net erosion/deposition Optimized
LU
25
Modeling impact of hedges
Observed slope change
big difference in roughness small diff.
in rougness, large diffusion, small diff. in
roughness, small diffusion
Seth M. Dabney et al. 1999 Lanscape Benching from
Tillage Erosion Between Grass Hedges
26
Evolution of hedge slope without tillage
Change in erosion and deposition pattern
Seth M. Dabney et al. 1999 Lanscape Benching from
Tillage Erosion Between Grass Hedges
27
Change in land use SW Cent. Campus
1993
future
2001
28
1993 overland flow and location of sediment
control structures
constructed wetland
check dam/spreader
detention area
H. Mitasova
29
Overland flow finished golf course
H. Mitasova
30
Sediment flow and erosion/deposition
Current
0.002m3/s
0.001kg/ms
erosion87kg/s
Construction
0.69
0.01-6.0kg/ms
erosion968kg/s
31
Control measures
Check dams and buffers
future
"Rolling construction"
H. Mitasova
32
Impact of extended buffers
cft/s 0.02 0.2 2.0 20
Added grass
Erosion
359kg/s
Added forest
Deposition
142kg/s
Buffers must have excellent infiltration , they
protect well against erosion from sheet flow but
are ineffective for concentrated flow
33
Impact of terrain and land use change on water
depthwith possible consequences for ecosystems
pre-and post-development
pre- and during construction
redistribution of water flow increase from C,
decrease from A,B for small events, spikes for
extreme events
increase in water depth on construction site and
within the stream buffer
H. Mitasova
34
Watershed analysis and hydrologic modeling in
GRASS
active r.watershed watershed analysis, D8, SFD,
shortest path r.flow hillslope erosion, Dinf,
SFD r.flowmd hillslope erosion, Dinf,
MFD r.terraflow watershed analysis, D8,
SFD/MFD r.sim.water hydrologic
simulation r.sim.sediment sediment transport,
erosion/deposition no new development r.topmodel
hydrologic simulation r.water.fea hydrologic
simulation, kinematic wave r.hydro.casc2d
hydrologic simulation, diffusive wave SWAT,
answers, agnps, kineros http//grass.baylor.edu/mo
delintegration.html
H. Mitasova
35
Open source GIS GRASS
http//grass.itc.it/
General purpose GIS for raster, vector, site and
image data processing, analysis, modeling and
visualization. Developed at US Army CERL
(1982-1995). General Public License (GPL) in 1999
free to run, modify, distribute, and release
(but it cannot be modified and released as a
proprietary system)
Coordination Markus Neteler, ITC-Trento
council type approach automated web-based
infrastructure supported by Intevation GmbH,
Free GIS Project
36
GRASS linked to other Free Software
Interoperability and data accessibility
Map production
(Geo)statistics
Online data dissemination
vis5d, povray
PROJ
Databases
Markus Neteler
37
Conclusions I
Mapping and monitoring - new generation of
georeferenced data (lidar, imagery) high
spatial and temporal resolution, massive data
sets - new methods for processing and
analysis Modeling - new generation of
spatially distributed, process-based models -
coupling with GIS - increasing importance of
stochastic techniques, path sampling very
promising - multi-scale, multi-model approaches
H. Mitasova
38
Conclusions II
Open Source Software - key aspect
flexibility, easy to adjust and combine -
appropriate both for education (students can
explore and develop the code) and for
cutting-edge applications Computational -
important isssues robustness, scaling,
scalability Applications - sustainable
development and environmental management -
optimized conservation measures (eg, construction
sites)
H. Mitasova
39
Acknowledgment
This project is funded by the NRC/ARO fellowship
and
North Carolina sediment control commission for
erosion and sediment control study
H. Mitasova