teismint-2.rst - pism - [fork] customized build of PISM, the parallel ice sheet model (tillflux branch)
(HTM) git clone git://src.adamsgaard.dk/pism
(DIR) Log
(DIR) Files
(DIR) Refs
(DIR) LICENSE
---
teismint-2.rst (6314B)
---
1 .. include:: ../../global.txt
2
3 .. _sec-EISMINTII:
4
5 EISMINT II
6 ----------
7
8 There are seven experiments described in the published EISMINT II writeup :cite:`EISMINT00`.
9 They are named A, B, C, D, F, G, and H. They have these common features:
10
11 - runs are for `2\times 10^5` years, with no prescribed time step;
12 - a `61\times 61` horizontal grid on a square domain (`1500` km side length) is prescribed;
13 - surface inputs (temperature and mass balance) have angular symmetry around the grid center;
14 - the bed is flat and does not move (no isostasy);
15 - the temperature in the bedrock is not modeled;
16 - only the cold (not polythermal) thermomechanically-coupled SIA is used :cite:`EISMINT00`; and
17 - basal melt rates do not affect the evolution of the ice sheet.
18
19 The experiments differ from each other in their various combinations of surface
20 temperature and mass balance parameterizations. Experiments H and G involve basal sliding,
21 under the physically-dubious SIA sliding rubric (:cite:`BBssasliding`, Appendix B), while the
22 others don't. Four experiments start with zero ice (A,F,G,H), while the other experiments
23 (B,C,D) start from the final state of experiment A.
24
25 In addition to the seven experiments published in :cite:`EISMINT00`, there were an additional
26 five experiments described in the EISMINT II intercomparison description :cite:`EISIIdescribe`,
27 labeled E, I, J, K, and L. These experiments share most features listed above, but with
28 the following differences. Experiment E is the same as experiment A except that the peak
29 of the accumulation, and also the low point of the surface temperature, are shifted by 100
30 km in both `x` and `y` directions; also experiment E starts with the final state of
31 experiment A. Experiments I and J are similar to experiment A but with non-flat "trough"
32 bed topography. Experiments K and L are similar to experiment C but with non-flat "mound"
33 bed topography.
34
35 See :numref:`tab-eisII` for how to run all EISMINT II experiments in PISM.
36 Experiments E -- L are only documented in :cite:`EISIIdescribe`.
37
38 .. list-table:: Running the EISMINT II experiments in PISM. Use ``-skip -skip_max 5``, on
39 the `61\times 61` default grid, for significant speedup.
40 :name: tab-eisII
41 :header-rows: 1
42 :widths: 5,2
43
44 * - Command: "``pisms +``"
45 - Relation to experiment A
46
47 * - ``-eisII A -Mx 61 -My 61 -Mz 61 -Lz 5000 -y 2e5 -o eisIIA.nc``
48 -
49 * - ``-eisII B -i eisIIA.nc -y 2e5 -o eisIIB.nc``
50 - warmer
51 * - ``-eisII C -i eisIIA.nc -y 2e5 -o eisIIC.nc``
52 - less snow (lower accumulation)
53 * - ``-eisII D -i eisIIA.nc -y 2e5 -o eisIID.nc``
54 - smaller area of accumulation
55 * - ``-eisII F -Mx 61 -My 61 -Mz 81 -Lz 6000 -y 2e5 -o eisIIF.nc``
56 - colder; famous spokes :cite:`BBL`
57 * - ``-eisII G -Mx 61 -My 61 -Mz 201 -Lz 5000 -y 2e5 -o eisIIG.nc``
58 - sliding (regardless of temperature)
59 * - ``-eisII H -Mx 61 -My 61 -Mz 201 -Lz 5000 -y 2e5 -o eisIIH.nc``
60 - melt-temperature activated sliding
61 * - ``-eisII E -i eisIIA.nc -y 2e5 -o eisIIE.nc``
62 - shifted climate maps
63 * - ``-eisII I -Mx 61 -My 61 -Mz 61 -Lz 5000 -y 2e5 -o eisIII.nc``
64 - trough topography
65 * - ``-eisII J -i eisIII.nc -y 2e5 -o eisIIJ.nc``
66 - trough topography and less snow
67 * - ``-eisII K -Mx 61 -My 61 -Mz 61 -Lz 5000 -y 2e5 -o eisIIK.nc``
68 - mound topography
69 * - ``-eisII L -i eisIIK.nc -y 2e5 -o eisIIL.nc``
70 - mound topography and less snow
71
72 The vertical grid is not specified in EISMINT II, but a good simulation of the
73 thermomechanically-coupled conditions near the base of the ice requires relatively-fine
74 resolution there. We suggest using the default unequally-spaced grid. With 61 levels it
75 gives a grid spacing of `\sim 20 m` in the ice layer closest to the bed, but more vertical
76 levels are generally better. Alternatively these experiments can be done with an
77 equally-spaced grid; in this case we suggest using enough vertical levels to give 20 m
78 spacing, for example. When there is sliding, even more vertical resolution is recommended
79 (see :numref:`tab-eisII`). Also, the vertical extent must be sufficient so that when
80 the ice thickness grows large, especially before thermo-softening brings it back down, the
81 vertical grid is tall enough to include all the ice. :numref:`tab-eisII` therefore
82 includes suggested settings of ``-Lz``; experiment F is different because ice thickness
83 increases with colder temperatures.
84
85 These SIA-only simulations parallelize well. Very roughly, for the standard `61\times 61`
86 horizontal grid, wall-clock-time speedups will occur up to about 30 processors. Runs on
87 finer (horizontal) grids will benefit from even more processors. Also, the "skip"
88 mechanism which avoids updating the temperature at each time step is effective, so options
89 like ``-skip -skip_max 5`` are recommended.
90
91 The EISMINT II experiments can be run with various modifications of the default settings.
92 For instance, a twice-finer grid in the horizontal is "``-Mx 121 -My 121``".
93 :numref:`tab-eisIIoptions` lists some optional settings which are particular to the
94 EISMINT II experiments.
95
96 .. list-table:: Changing the default settings for EISMINT II
97 :name: tab-eisIIoptions
98 :header-rows: 1
99 :widths: 1,2,1,2
100
101 * - Option
102 - Default values [experiments]
103 - Units
104 - Meaning
105
106 * - :opt:`-eisII`
107 - A
108 -
109 - Choose single character name of EISMINT II :cite:`EISMINT00` simplified geometry
110 experiment. See :numref:`tab-eisII`.
111
112 * - :opt:`-Mmax`
113 - 0.5 [ABDEFGHIK],
114
115 0.25 [CJL]
116 - `m / year`
117 - max value of accumulation rate
118
119 * - :opt:`-Rel`
120 - 450 [ABEFGHIK],
121
122 425 [CDJL]
123 - km
124 - radial distance to equilibrium line
125
126 * - :opt:`-Sb`
127 - `10^{-2}` [*all*]
128 - `(m/year)/km`
129 - radial gradient of accumulation rate
130
131 * - :opt:`-ST`
132 - `1.67 \times 10^{-2}` [*all*]
133 - K/km
134 - radial gradient of surface temperature
135
136 * - :opt:`-Tmin`
137 - 238.15 [ACDEGHIJKL],
138
139 243.15 [B],
140
141 223.15 [F]
142 - K
143 - max of surface temperature
144
145 * - :opt:`-bmr_in_cont`
146 -
147 -
148 - Include the basal melt rate in the mass continuity computation; overrides EISMINT
149 II default.
150
151 See subdirectory ``examples/eismintII/`` for a simple helper script ``runexp.sh``.