1r.terraflow(1) GRASS GIS User's Manual r.terraflow(1)
2
6 r.terraflow - Performs flow computation for massive grids.
7
9 raster, hydrology, flow, accumulation, sink
10
12 r.terraflow
13 r.terraflow --help
14 r.terraflow [-s] elevation=name [filled=name] [direction=name]
15 [swatershed=name] [accumulation=name] [tci=name] [d8cut=float]
16 [memory=memory in MB] [directory=string] [stats=string] [--over‐
17 write] [--help] [--verbose] [--quiet] [--ui]
18 Flags:
20 -s
21 SFD (D8) flow (default is MFD)
22 SFD: single flow direction, MFD: multiple flow direction
23 --overwrite
25 Allow output files to overwrite existing files
26 --help
28 Print usage summary
29 --verbose
31 Verbose module output
32 --quiet
34 Quiet module output
35 --ui
37 Force launching GUI dialog
38 Parameters:
40 elevation=name [required]
41 Name of input elevation raster map
42 filled=name
44 Name for output filled (flooded) elevation raster map
45 direction=name
47 Name for output flow direction raster map
48 swatershed=name
50 Name for output sink-watershed raster map
51 accumulation=name
53 Name for output flow accumulation raster map
54 tci=name
56 Name for output topographic convergence index (tci) raster map
57 d8cut=float
59 Routing using SFD (D8) direction
60 If flow accumulation is larger than this value it is routed using
61 SFD (D8) direction (meaningful only for MFD flow). If no answer is
62 given it defaults to infinity.
63 memory=memory in MB
65 Maximum memory to be used (in MB)
66 Cache size for raster rows
67 Default: 300
68 directory=string
70 Directory to hold temporary files (they can be large)
71 stats=string
73 Name for output file containing runtime statistics
74
76 r.terraflow takes as input a raster digital elevation model (DEM) and
77 computes the flow direction raster and the flow accumulation raster, as
78 well as the flooded elevation raster, sink-watershed raster (partition
79 into watersheds around sinks) and TCI (topographic convergence index)
80 raster maps.
81 r.terraflow computes these rasters using well-known approaches, with
83 the difference that its emphasis is on the computational complexity of
84 the algorithms, rather than on modeling realistic flow. r.terraflow
85 emerged from the necessity of having scalable software able to process
86 efficiently very large terrains. It is based on theoretically optimal
87 algorithms developed in the framework of I/O-efficient algorithms.
88 r.terraflow was designed and optimized especially for massive grids and
89 is able to process terrains which were impractical with similar func‐
90 tions existing in other GIS systems.
91 Flow directions are computed using either the MFD (Multiple Flow Direc‐
93 tion) model or the SFD (Single Flow Direction, or D8) model, illus‐
94 trated below. Both methods compute downslope flow directions by in‐
95 specting the 3-by-3 window around the current cell. The SFD method as‐
96 signs a unique flow direction towards the steepest downslope neighbor.
97 The MFD method assigns multiple flow directions towards all downslope
98 neighbors.
99 Flow direction to steepest downslope neighbor (SFD). Flow direction to all downslope neighbors (MFD).
103 The SFD and the MFD method cannot compute flow directions for cells
106 which have the same height as all their neighbors (flat areas) or cells
107 which do not have downslope neighbors (one-cell pits).
108 • On plateaus (flat areas that spill out) r.terraflow routes flow
110 so that globally the flow goes towards the spill cells of the
111 plateaus.
112 • On sinks (flat areas that do not spill out, including one-cell
114 pits) r.terraflow assigns flow by flooding the terrain until
115 all the sinks are filled and assigning flow directions on the
116 filled terrain.
117 In order to flood the terrain, r.terraflow identifies all sinks and
119 partitions the terrain into sink-watersheds (a sink-watershed contains
120 all the cells that flow into that sink), builds a graph representing
121 the adjacency information of the sink-watersheds, and uses this
122 sink-watershed graph to merge watersheds into each other along their
123 lowest common boundary until all watersheds have a flow path outside
124 the terrain. Flooding produces a sink-less terrain in which every cell
125 has a downslope flow path leading outside the terrain and therefore ev‐
126 ery cell in the terrain can be assigned SFD/MFD flow directions as
127 above.
128 Once flow directions are computed for every cell in the terrain, r.ter‐
130 raflow computes flow accumulation by routing water using the flow di‐
131 rections and keeping track of how much water flows through each cell.
132 If flow accumulation of a cell is larger than the value given by the
134 d8cut option, then the flow of this cell is routed to its neighbors us‐
135 ing the SFD (D8) model. This option affects only the flow accumulation
136 raster and is meaningful only for MFD flow (i.e. if the -s flag is not
137 used); If this option is used for SFD flow it is ignored. The default
138 value of d8cut is infinity.
139 r.terraflow also computes the tci raster (topographic convergence in‐
141 dex, defined as the logarithm of the ratio of flow accumulation and lo‐
142 cal slope).
143 For more details on the algorithms see [1,2,3] below.
145
147 One of the techniques used by r.terraflow is the space-time trade-off.
148 In particular, in order to avoid searches, which are I/O-expensive,
149 r.terraflow computes and works with an augmented elevation raster in
150 which each cell stores relevant information about its 8 neighbors, in
151 total up to 80B per cell. As a result r.terraflow works with interme‐
152 diate temporary files that may be up to 80N bytes, where N is the num‐
153 ber of cells (rows x columns) in the elevation raster (more precisely,
154 80K bytes, where K is the number of valid (not no-data) cells in the
155 input elevation raster).
156 All these intermediate temporary files are stored in the path specified
158 by the directory option. Note: directory must contain enough free disk
159 space in order to store up to 2 x 80N bytes.
160 The memory option can be used to set the maximum amount of main memory
162 (RAM) the module will use during processing. In practice its value
163 should be an underestimate of the amount of available (free) main mem‐
164 ory on the machine. r.terraflow will use at all times at most this much
165 memory, and the virtual memory system (swap space) will never be used.
166 The default value is 300 MB.
167 The stats option defines the name of the file that contains the statis‐
169 tics (stats) of the run.
170 r.terraflow has a limit on the number of rows and columns (max 32,767
172 each).
173 The internal type used by r.terraflow to store elevations can be de‐
175 fined at compile-time. By default, r.terraflow is compiled to store el‐
176 evations internally as floats. Other versions can be created by the
177 user if needed.
178 Hints concerning compilation with storage of elevations internally as
180 shorts: such a version uses less space (up to 60B per cell, up to 60N
181 intermediate file) and therefore is more space and time efficient.
182 r.terraflow is intended for use with floating point raster data
183 (FCELL), and r.terraflow (short) with integer raster data (CELL) in
184 which the maximum elevation does not exceed the value of a short
185 SHRT_MAX=32767 (this is not a constraint for any terrain data of the
186 Earth, if elevation is stored in meters). Both r.terraflow and r.ter‐
187 raflow (short) work with input elevation rasters which can be either
188 integer, floating point or double (CELL, FCELL, DCELL). If the input
189 raster contains a value that exceeds the allowed internal range (short
190 for r.terraflow (short), float for r.terraflow), the program exits with
191 a warning message. Otherwise, if all values in the input elevation
192 raster are in range, they will be converted (truncated) to the internal
193 elevation type (short for r.terraflow (short), float for r.terraflow).
194 In this case precision may be lost and artificial flat areas may be
195 created. For instance, if r.terraflow (short) is used with floating
196 point raster data (FCELL or DCELL), the values of the elevation will be
197 truncated as shorts. This may create artificial flat areas, and the
198 output of r.terraflow (short) may be less realistic than those of
199 r.terraflow on floating point raster data. The outputs of r.terraflow
200 (short) and r.terraflow are identical for integer raster data (CELL
201 maps).
202
204 Example for small area in North Carolina sample dataset to calculate
205 flow accumulation:
206 g.region raster=elev_lid792_1m
207 r.terraflow elevation=elev_lid792_1m accumulation=elev_lid792_1m_accumulation
208 Flow accumulation
209 Spearfish sample data set:
211 g.region raster=elevation.10m -p
212 r.terraflow elev=elevation.10m filled=elevation10m.filled \
213 dir=elevation10m.mfdir swatershed=elevation10m.watershed \
214 accumulation=elevation10m.accu tci=elevation10m.tci
215 g.region raster=elevation.10m -p
216 r.terraflow elev=elevation.10m filled=elevation10m.filled \
217 dir=elevation10m.mfdir swatershed=elevation10m.watershed \
218 accumulation=elevation10m.accu tci=elevation10m.tci d8cut=500 memory=800 \
219 stats=elevation10mstats.txt
220
222 1 The TerraFlow project at Duke University
223 2 I/O-efficient algorithms for problems on grid-based terrains.
225 Lars Arge, Laura Toma, and Jeffrey S. Vitter. In Proc. Workshop
226 on Algorithm Engineering and Experimentation, 2000. To appear in
227 Journal of Experimental Algorithms.
228 3 Flow computation on massive grids. Lars Arge, Jeffrey S. Chase,
230 Patrick N. Halpin, Laura Toma, Jeffrey S. Vitter, Dean Urban and
231 Rajiv Wickremesinghe. In Proc. ACM Symposium on Advances in Geo‐
232 graphic Information Systems, 2001.
233 4 Flow computation on massive grid terrains. Lars Arge, Jeffrey
235 S. Chase, Patrick N. Halpin, Laura Toma, Jeffrey S. Vitter, Dean
236 Urban and Rajiv Wickremesinghe. In GeoInformatica, Interna‐
237 tional Journal on Advances of Computer Science for Geographic
238 Information Systems, 7(4):283-313, December 2003.
239
241 r.flow, r.basins.fill, r.drain, r.topidx, r.topmodel, r.water.outlet,
242 r.watershed
243
245 Original version of program: The TerraFlow
246 project, 1999, Duke University.
247 Lars Arge, Jeff Chase, Pat Halpin, Laura Toma, Dean Urban, Jeff
248 Vitter, Rajiv Wickremesinghe.
249 Porting to GRASS GIS, 2002:
251 Lars Arge, Helena Mitasova, Laura Toma.
252 Contact: Laura Toma
254
256 Available at: r.terraflow source code (history)
257 Accessed: Mon Jun 20 16:46:47 2022
259 Main index | Raster index | Topics index | Keywords index | Graphical
261 index | Full index
262 © 2003-2022 GRASS Development Team, GRASS GIS 8.2.0 Reference Manual
264GRASS 8.2.0 r.terraflow(1)