GMT和Pygmt提供了一个远程数据功能,可以使用函数datasets远程下载多种在线数据,并进行处理和绘图。这里以pygmt为例绘制海底地壳年龄、陆地地形。 地壳数据[1]包含了不同的分辨率,对应不同文件大小,最粗为1d,全球数据仅125K,最大分辨率1m,全球数据188M。绘图5 ]% `# j" a3 U: Y9 O9 ~+ `/ |5 \/ i% p
[C] 纯文本查看 复制代码 import pygmt
grid_globe = pygmt.datasets.load_earth_age(resolution='06m', region="-180/180/-90/90", registration=None)
fig = pygmt.Figure()
fig.grdimage(grid=grid_globe, projection="R15c", region="0/360/-89/89", frame=True,cmap="crustal_age.cpt")
fig.colorbar(frame=["af", "x+lage", "y+lMyr"],cmap="crustal_age.cpt")
fig.show() 9 Y9 | T- d8 D( k9 [( }
8 n- U- t9 g: v
' ]2 c) Q+ t1 `+ g# X+ ]* `4 r; [
上面的调色板crustal_age可以在.gmt/cache/下找到,而远程数据也下载到了./gmt/server/下面。
: u* s1 X5 i0 t9 G( s地形数据地形数据[2]包含多种不同分辨率,对应不同的文件大小,最粗为1d,文件大小128k,最高分辨率为1s,文件大小达41G: SRTM绘图[C] 纯文本查看 复制代码 # 雅鲁藏布江大峡谷[/b]grid = pygmt.datasets.load_earth_relief(
"03s",
region=[94, 95.5, 29, 30],
registration="gridline",
use_srtm=True,
)
# calculate the reflection of a light source projecting from west to east
# (azimuth of 270 degrees) and at a latitude of 30 degrees from the horizon
dgrid = pygmt.grdgradient(grid=grid, radiance=[270, 30])
fig = pygmt.Figure()
fig.grdimage(grid=grid, projection="M15c", region=[94, 95.5, 29, 30], frame=['WSrt+t"Original Data Elevation Model"',"xa", "ya"],cmap="dem1")
fig.colorbar(position="JML+o1.8c/0c+w10c/0.9c",frame=["af", "y+lmeter"])
fig.coast(rivers="a/1p",borders="2/5,red")
# Shift plot origin of the second map by 12.5 cm in x direction
fig.shift_origin(xshift="20c")
pygmt.makecpt(cmap="gray", series=[-1.5, 0.3, 0.01])
fig.grdimage(
grid=dgrid,
projection="M15c",
frame=['lSEt+t"Hillshade Map"', "xa0.1", "ya0.1"],
cmap=True,
)
fig.coast(rivers="a/1p",borders="2/5,red")
# Shift plot origin of the second map by 12.5 cm in x direction
fig.shift_origin(xshift="20c")
fig.grdimage(
grid=grid,
shading=dgrid,
projection="M15c",
frame=['lSEt+t"Hillshade Map"', "xa0.1", "ya0.1"],
cmap="dem1",
)
fig.coast(rivers="a/1p",borders="2/5,red")
fig.show(width="20c")
fig.savefig("srtm.png") 3 B' @+ e( T6 S/ V. _4 S3 k( n
Q: ?8 ], L0 w
! d& Y9 P2 t* H7 R7 E% Z0 s: v) B9 t/ P9 S
3D地形图
: O& Z; P% M- u1 ~[C] 纯文本查看 复制代码 fig = pygmt.Figure()
fig.grdview(
grid=grid,
region=[94.7, 95.2, 29.5, 30],
perspective=[250, 60],
frame=["xa", "ya", "WSNE"],
projection="M15c",
zsize="15c",
surftype="s",
cmap="dem1",
# Set the plane elevation to 1,000 meters and make the fill "gray"
plane="000+ggray",
)
fig.show()
- U5 y5 [5 ^) u) p
f) O+ B" B& Y W3 V
同样,我们还可以使用pygmt.grdview绘制三维地形图。下面是我曾经到过山脚下,但是在云中的南迦巴瓦峰。 6 c: |, d+ h G5 V' p
* V6 O6 Y- |! e8 Y2 |
" Z( b8 f& c. k" u, X. G5 A( F+ T: I/ u% ` W) _) K3 U6 `& k/ |' Z
附:遥感影像和地形的结合在github存在一个30Day*****的系列代码库,其中包含绘图领域的30DayMapChallenge2021,恰好已经使用GMT完成了这项工作,作者是Pygmt的核心开发者Weiji。 这里有两个遥感影像和地形结合的例子(17和18),可以作为很好的学习材料.
, ~) a$ M. o! u
3 T, }( ~# m0 H% u' C8 E( u
% |% V, N1 O" }! t8 E" K9 n
References[1] 地壳数据: https://www.generic-mapping-tools.org/remote-datasets/earth-age.html
, G1 b6 |8 T2 c5 U6 d[2] 地形数据: https://www.generic-mapping-tools.org/remote-datasets/earth-relief.html
7 X: S5 ~4 L8 p2 }0 {; p g/ ]) b
来源:海洋遥感数据共享' D! s8 |1 y; g; V" Z6 M5 K" O: s' @! }
; E3 Q ?; \* P2 G' L4 L |