Volltext (11 MB pdf)
Literaturverzeichnis (References)
Download ArcView Extension Topocrop (link to zip-file)
Download my Avenue scripts
Verfügbarkeit in Universitätsbibliotheken
(Suche nach "Hochgenaue Digitale Geländemodelle"):
- Gemeinsamer Verbundkatalog (GVK)
- Südwestdeutscher
Bibliotheksverbund (SWB)
-
Universitätsbibliothek Rostock
Topocrop Documentation
Topocrop is a
ArcView Extension (avx) that requires a full version of ArcView 3.X
and the SpatialAnalyst.
Topocrop is a compilation of Avenue scripts that can be useful
for terrain analysis. A DEM is required as Grid.
The algorithms were optimised for a grid size of 1 - 10 m and a
corresponding elevation quality of plus minus 5 - 20 cm. This is
reached by RTK-GPS surveying campaigns and Laserscanning DEMs.
The main purpose of the extension was to help me to derive and
analyse Topographic Wetness Index grids for a Precision
Agriculture project (www.preagro.de). Thus, the goal was the
assessment of the pattern of potential soil moisture on an
agricultural plot (field size 10 - 200 hectares) under Central
European soil and climatic conditions. More about the results can
be found in
The scripts available with topocrop are:
1) Topographic
Wetness Index
The script calculates Beven & Kirkby's topographic wetness
index TWI = ln(As/tan beta) based on the ArcView flow algorithm (D8).
In addition to the actual TWI grid, a mean filtered grid is
caluclated (neighbourhood statistics with 3 cells radius).
There are known problems with the flow algorithm in flat areas.
Open the script (load system script terrain.twi) and add the line
"slopeGrid = slopeGrid+1". Thus, there are no flat
areas and flow will be transported to a neighbouring cell. This
introduces some arbitrary miscalulations but enables the script
to run even in flat areas. Those parts of the grid should be
clipped after the calculation.
For more references about Wetness Indices see Wilson and Gallant
(2000) and others, Link References above.
2) Streampower
Index
Calculates Moore's stream power index ln(As*tan beta). See MOORE
and BURCH for references. The script is very similar to the TWI
above. The difference is the stronger linear flow line appearence
of the result. The stream power and thus the water erosion
potential along preferred surface flow lines rises with slope and
catchment area of a cell. The most likely water erosion paths on
hilly fields can be detected. The script also calcultates the
mean filtered grid in order to blur the static, unrealistic
pattern caused by the D8/flow algorithm.
3) Slope-Length-Factor
Calculates the slope-length factor for USLE based on work by
Moore & Wilson 1992. This "process based" model is
adapted to the empirical factor of the soil loss erosion equation.
The pattern of soil loss assessment is reduced to two factors:
catchment area of a cell (flow accumulation) and slope.
The equation and parameters used here are: LSGrid = (((flowaccGrid/22.13)^0.4)*1.4*(((slopeGrid.sin)/0.0896)^1.3))
4) Landfrom Elements scale 1:5000
This script is still rather unconvenient to work with -
improvement by experienced Avenue users is appreciated (terrain.form)!
At the moment zou have to follow the info boxes at each step very
carefully.
It works well with DEM grids of 2 - 5 m raster size. The
curvature threshold values have to be adapted for different
scales. The goal of the script is to derive 9 landform elements
from a DEM based on combinations of profile and tangential
curvature. A grid with the following elements is calculated (grid
value): convex-diverging (5), convex-parallel (4), convex-converging
(1), straight-diverging (10), straight-parallel (8), straight-converging
(2), concave-diverging (15), concave-parallel (12), concave-converging
(3).
In detail, the script conducts the following steps:
- Calculating the terrain curvature
-- elevGrid.curvature("pro_curv".asfilename, "horiz_curv".asfilename,
nil, nil)
-- tangentGrid = horiz_curv*slope_rad
- Smoothing the curvature grids (mean filter)
-- r = tanG.FocalStats(#GRID_STATYPE_MEAN, nbrhd, FALSE)
and pro?curv resp.
- Reclassifying the continuous grids to three curvature classes
each (threshold values are proposed and have to be entered)
-- MsgBox.Info("New classes: (1) negative..-0.05 (2)-0.05..0.05
(3) 0.05..positive ","Reclassification Info").
For units see ArcView help, discussion for curvature
-- r = pro_rclG.ReclassByClassList(fieldName, classList,
toNoData)
-- MsgBox.Info("New classes: (1) negative..-0.002 (4) -0.002..0.001
(5) 0.001..positive","Reclassification Info").
- Mind the
class names in brackets! This is important for the following step
in order to get 9 different landform classes!
- The classified grids of profile and tangential curvature are
then multiplied, the three times three values result in 9 classes
-- formgrid=proGthm*tanGthm
- In the last step, an overlay theme is calculated to hide the
flat (meaningless) areas. You need to load the legend file (avl).
The nine
landform classes correspond strongly to the TWI pattern since
both depend on curvature of the terrain. Converging and concave
classes represent potentially wet areas, diverging and convex
classes are usually drier parts of the fields and correspond to
topslopes and peaks.
References: Dikau, R. (1993):
The application of a digital relief model to landform analysis in
geomorphology. In: Raper, J. [Ed.]: Three-Dimensional
Applications in GIS. Taylor & Francis. London. pp. 51-77.
5) Map Overlay (Multiply)
6) Local Correlation Map (written by Robin Gebbers)
7) RMSE for points
Dr. Frank Schmidt, 2004-02-22
contact: schmidt at beak.de
