1r.spread(1) Grass User's Manual r.spread(1)
2
3
4
6 r.spread - Simulates elliptically anisotropic spread on a graphics
7 window and generates a raster map of the cumulative time of spread,
8 given raster maps containing the rates of spread (ROS), the ROS direc‐
9 tions and the spread origins. It optionally produces raster maps to
10 contain backlink UTM coordinates for tracing spread paths.
11
13 raster
14
16 r.spread
17 r.spread help
18 r.spread [-vds] max=string dir=string base=string start=string
19 [spot_dist=string] [w_speed=string] [f_mois=string]
20 [least_size=odd int] [comp_dens=decimal] [init_time=int (>= 0)]
21 [lag=int (>= 0)] [backdrop=string] output=string [x_output=string]
22 [y_output=string] [--overwrite]
23
24 Flags:
25 -v Run VERBOSELY
26
27 -d DISPLAY 'live' spread process on screen
28
29 -s For wildfires: consider SPOTTING effect
30
31 --overwrite
32
33 Parameters:
34 max=string
35 Name of raster map containing MAX rate of spread (ROS) (cm/min)
36
37 dir=string
38 Name of raster map containing DIRections of max ROS (degree)
39
40 base=string
41 Name of raster map containing BASE ROS (cm/min)
42
43 start=string
44 Name of raster map containing STARTing sources
45
46 spot_dist=string
47 Name of raster map containing max SPOTting DISTance (m) (required
48 w/ -s)
49
50 w_speed=string
51 Name of raster map containing midflame Wind SPEED (ft/min)
52 (required w/ -s)
53
54 f_mois=string
55 Name of raster map containing fine Fuel MOISture of the cell
56 receiving a spotting firebrand (%) (required w/ -s)
57
58 least_size=odd int
59 Basic sampling window SIZE needed to meet certain accuracy (3)
60 Options: 3,5,7,9,11,13,15
61
62 comp_dens=decimal
63 Sampling DENSity for additional COMPutin (range: 0.0 - 1.0 (0.5))
64
65 init_time=int (>= 0)
66 INITial TIME for current simulation (0) (min)
67
68 lag=int (>= 0)
69 Simulating time duration LAG (fill the region) (min)
70
71 backdrop=string
72 Name of raster map as a display backdrop
73
74 output=string
75 Name of raster map to contain OUTPUT spread time (min)
76
77 x_output=string
78 Name of raster map to contain X_BACK coordiates
79
80 y_output=string
81 Name of raster map to contain Y_BACK coordiates
82
84 Spread phenomena usually show uneven movement over space. Such uneven‐
85 ness is due to two reasons:
86 1) the uneven conditions from location to location, which can be called
87 SPATIAL HETEROGENEITY, and
88 2) the uneven conditions in different directions, which can be called
89 ANISOTROPY.
90 The anisotropy of spread occurs when any of the determining factors
91 have directional components. For example, wind and topography cause an‐
92 isotropic spread of wildfires.
93
94 One of the simplest spatial heterogeneous and anisotropic spread is
95 elliptical spread, in which, each local spread shape can be thought as
96 an ellipse. In a raster setting, cell centers are foci of the spread
97 ellipses, and the spread phenomenon moves fastest toward apogees and
98 slowest to perigees. The sizes and shapes of spread ellipses may vary
99 cell by cell. So the overall spread shape is commonly not an ellipse.
100
101 r.spread simulates elliptically anisotropic spread phenomena, given
102 three raster map layers about ROS (base ROS, maximum ROS and direction
103 of the maximum ROS) plus a raster map layer showing the starting
104 sources. These ROS layers define unique ellipses for all cell loca‐
105 tions in the current geographic region as if each cell center was a
106 potential spread origin. For some wildfire spread, these ROS layers
107 can be generated by another GRASS raster program r.ros. The actual
108 locations reached by a spread event are constrained by the actual
109 spread origins and the elapsed spread time.
110
111 r.spread optionally produces raster maps to contain backlink UTM coor‐
112 dinates for each raster cell of the spread time map. The spread paths
113 can be accurately traced based on the backlink information by another
114 GRASS raster program r.spreadpath.
115
116 Part of the spotting function in r.spread is based on Chase (1984) and
117 Rothermel (1983). More information on r.spread, r.ros and r.spreadpath
118 can be found in Xu (1994).
119
121 -v Run verbosely, printing information about program progress to
122 standard output.
123
124 -d Display the "live" simulation on screen. A graphics window must
125 be opened and selected before using this option.
126
127 -s For wildfires, also consider spotting.
128
130 max=name
131 Name of an existing raster map layer in the user's current
132 mapset search path containing the maximum ROS values
133 (cm/minute).
134
135 dir=name
136 Name of an existing raster map layer in the user's current
137 mapset search path containing directions of the maximum ROSes,
138 clockwise from north (degree).
139
140 base=name
141 Name of an existing raster map layer in the user's current
142 mapset search path containing the ROS values in the directions
143 perpendicular to maximum ROSes' (cm/minute). These ROSes are
144 also the ones without the effect of directional factors.
145
146 start=name
147 Name of an existing raster map layer in the user's current
148 mapset search path containing starting locations of the spread
149 phenomenon. Any positive integers in this map are recognized as
150 starting sources.
151
152 spot_dist=name
153 Name of an existing raster map layer in the user's current
154 mapset search path containing the maximum potential spotting
155 distances (meters).
156
157 w_speed=name
158 Name of an existing raster map layer in the user's current
159 mapset search path containing wind velocities at half of the
160 average flame height (feet/minute).
161
162 f_mois=name
163 Name of an existing raster map layer in the user's current
164 mapset search path containing the 1-hour (<.25") fuel moisture
165 (percentage content multiplied by 100).
166
167 least_size=odd int An odd integer ranging 3 - 15 indicating
168 the basic sampling window size within which all cells will be
169 considered to see whether they will be reached by the current
170 spread cell. The default number is 3 which means a 3x3 window.
171
172 comp_dens=decimal A decimal number ranging 0.0 - 1.0 indicating
173 additional sampling cells will be considered to see whether they
174 will be reached by the current spread cell. The closer to 1.0
175 the decimal number is, the longer the program will run and the
176 higher the simulation accuracy will be. The default number is
177 0.5.
178
179 init_time=int A non-negative number specifying the initial
180 time for the current spread simulation (minutes). This is useful
181 when multiple phase simulation is conducted. The default time is
182 0.
183
184 lag=int A non-negative integer specifying the simulating
185 duration time lag (minutes). The default is infinite, but the
186 program will terminate when the current geographic region/mask
187 has been filled. It also controls the computational time, the
188 shorter the time lag, the faster the program will run.
189
190 backdrop=name
191 Name of an existing raster map layer in the user's current
192 mapset search path to be used as the background on which the
193 "live" movement will be shown.
194
195 output=name
196 Name of the new raster map layer to contain the results of the
197 cumulative spread time needed for a phenomenon to reach each
198 cell from the starting sources (minutes).
199
200 x_output=name
201 Name of the new raster map layer to contain the results of back‐
202 link information in UTM easting coordinates for each cell.
203
204 y_output=name
205 Name of the new raster map layer to contain the results of back‐
206 link information in UTM northing coordinates for each cell.
207
209 The user can run r.spread either interactively or non- interactively.
210 The program is run interactively if the user types r.spread without
211 specifying flag settings and parameter values on the command line. In
212 this case, the user will be prompted for input.
213
214 Alternately, the user can run r.spread non-interactively, by specifying
215 the names of raster map layers and desired options on the command line,
216 using the form:
217
218 r.spread [-vds] max=name dir=name base=name start=name [spot_dist=name]
219 [w_speed=name] [f_mois=name] [least_size=odds int] [comp_dens=decimal]
220 [init_time=int (>=0)] [lag=int (>= 0)] [backdrop=name] output=name
221 [x_output=name] [y_output=name] The -d option can only be used after a
222 graphics window is opened and selected.
223
224 Options spot_dist=name, w_speed=name and f_mois=name must all be given
225 if the -s option is used.
226
228 Assume we have inputs, the following simulates a spotting- involved
229 wildfire on the graphics window and generates three raster maps to con‐
230 tain spread time, backlink information in UTM northing and easting
231 coordinates:
232
233 r.spread -ds max=my_ros.max dir=my_ros.maxdir base=my_ros.base
234 start=fire_origin spot_dist=my_ros.spotdist w_speed=wind_speed
235 f_mois=1hour_moisture backdrop=image_burned output=my_spread x_out‐
236 put=my_spread.x y_output=my_spread.y
237
239 1. r.spread is a specific implementation of the shortest path algo‐
240 rithm. r.cost GRASS program served as the starting point for the
241 development of r.spread. One of the major differences between the two
242 programs is that r.cost only simulates ISOTROPIC spread while r.spread
243 can simulate ELLIPTICALLY ANISOTROPIC spread, including isotropic
244 spread as a special case.
245
246 2. Before running r.spread, the user should prepare the ROS (base, max
247 and direction) maps using appropriate models. For some wildfire spread,
248 a separate GRASS program r.ros based on Rothermel's fire equation does
249 such work. The combination of the two forms a simulation of wildfire
250 spread.
251
252 3. The relationship of the start map and ROS maps should be logically
253 correct, i.e. a starting source (a positive value in the start map)
254 should not be located in a spread BARRIER (zero value in the ROS maps).
255 Otherwise the program refuses to run.
256
257 4. r.spread uses the current geographic region settings. The output map
258 layer will not go outside the boundaries set in the region, and will
259 not be influenced by starting sources outside. So any change of the
260 current region may influence the output. The recommendation is to use
261 slightly larger region than needed. Refer to g.region to set an appro‐
262 priate geographic region.
263
264 5. The inputs to r.spread should be in proper units.
265
266 6. r.spread is a computationally intensive program. The user may need
267 to choose appropriate size of the geographic region and resolution.
268
269 7. A low and medium (i.e. <= 0.5) sampling density can improve accuracy
270 for elliptical simulation significantly, without adding significantly
271 extra running time. Further increasing the sample density will not gain
272 much accuracy while requiring greatly additional running time.
273
275 g.region, r.cost, r.spreadpath, r.ros
276
278 Chase, Carolyn, H., 1984, Spotting distance from wind-driven surface
279 fires -- extensions of equations for pocket calculators, US Forest Ser‐
280 vice, Res. Note INT-346, Ogden, Utah.
281
282 Rothermel, R. C., 1983, How to predict the spread and intensity of for‐
283 est and range fires. US Forest Service, Gen. Tech. Rep. INT-143.
284 Ogden, Utah.
285
286 Xu, Jianping, 1994, Simulating the spread of wildfires using a geo‐
287 graphic information system and remote sensing, Ph. D. Dissertation,
288 Rutgers University, New Brunswick, New Jersey.
289
291 Jianping Xu and Richard G. Lathrop, Jr., Center for Remote Sensing and
292 Spatial Analysis, Rutgers University.
293
294 Last changed: $Date: 2006/04/13 19:25:42 $
295
296 Full index
297
298
299
300GRASS 6.2.2 r.spread(1)