trun-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
---
trun-2.rst (6958B)
---
1 .. include:: ../../global.txt
2
3 .. _sec-ssarun:
4
5 Second run: a better ice-dynamics model
6 ---------------------------------------
7
8 It is widely-understood that ice sheets slide on their bases, especially when liquid water
9 is present at the base (see :cite:`Joughinetal2001`, :cite:`MacAyeal`, among others). An
10 important aspect of modeling such sliding is the inclusion of membrane or "longitudinal"
11 stresses into the stress balance :cite:`BBssasliding`. The basic stress balance in PISM
12 which involves membrane stresses is the Shallow Shelf Approximation (SSA)
13 :cite:`WeisGreveHutter`. The stress balance used in the previous section was, by contrast,
14 the (thermomechanically-coupled) non-sliding, non-membrane-stress Shallow Ice
15 Approximation (SIA) :cite:`BBL`, :cite:`EISMINT00`. The preferred ice dynamics model
16 within PISM, that allows both sliding balanced by membrane stresses and shear flow as
17 described by the SIA, is the SIA+SSA "hybrid" model :cite:`BBssasliding`,
18 :cite:`Winkelmannetal2011`. For more on stress balance theories see section
19 :ref:`sec-dynamics` of this Manual.
20
21 The practical issue with models of sliding is that a distinctly-uncertain parameter space
22 must be introduced. This especially involves parameters controlling the amount and
23 pressure of subglacial water (see :cite:`AschwandenAdalgeirsdottirKhroulev`,
24 :cite:`Clarke05`, :cite:`Tulaczyketal2000`, :cite:`vanPeltOerlemans2012`, among others). In
25 this regard, PISM uses the concept of a saturated and pressurized subglacial till with a
26 modeled distribution of yield stress :cite:`BBssasliding`, :cite:`SchoofStream`. The yield
27 stress arises from the PISM model of the production of subglacial water, which is itself
28 computed through the conservation of energy model :cite:`AschwandenBuelerKhroulevBlatter`.
29 We use such models in the rest of this Getting Started section.
30
31 While the ``spinup.sh`` script has default sliding-related parameters, for demonstration
32 purposes we change one parameter. We replace the default power `q=0.25` in the
33 sliding law (the equation which relates both the subglacial sliding velocity and the till
34 yield stress to the basal shear stress which appears in the SSA stress balance) by a less
35 "plastic" and more "linear" choice `q=0.5`. See section :ref:`sec-basestrength`
36 for more on sliding laws. To see the run we propose, do
37
38 .. literalinclude:: scripts/run-2-echo.sh
39 :language: bash
40 :lines: 3-
41
42 Now remove "``PISM_DO=echo``" and redirect the text output into a file to start the run:
43
44 .. literalinclude:: scripts/run-2.sh
45 :language: bash
46 :lines: 3-
47
48 This run should take 10 minutes or less.\ [#]_
49
50 When this run is finished it produces ``g20km_10ka_hy.nc``. As before do
51
52 .. code-block:: none
53
54 ncdump -h g20km_10ka_hy.nc |grep history
55
56 to see performance results for your machine.
57
58 The results of this run are shown in :numref:`fig-secondoutputcoarse`. We show the basal
59 sliding speed field ``velbase_mag`` in this Figure, where :numref:`fig-firstoutput` had
60 the ``mask``, but the reader can check that ``velbase_mag`` is zero in the nonsliding
61 SIA-only result ``g20km_10ka.nc``.
62
63 .. figure:: figures/g20km-10ka-hy-usurf-csurf-cbase.png
64 :name: fig-secondoutputcoarse
65
66 Fields from output file ``g20km_10ka_hy.nc``.
67
68 :Left: :var:`usurf`, the ice sheet surface elevation in meters.
69 :Middle: :var:`velsurf_mag`, the surface speed in m/year, including the 100 m/year
70 contour (solid black).
71 :Right: the sliding speed :var:`velbase_mag`, shown the same way as :var:`velsurf_mag`.
72
73 The hybrid model includes sliding, and it is important to evaluate that aspect of the
74 output. However, though it is critical to the response of the ice to changes in climate,
75 basal sliding velocity is essentially unobservable in real ice sheets. On the other hand,
76 because of relatively-recent advances in radar and image technology and processing
77 :cite:`Joughin2002`, the surface velocity of an ice sheet can be measured.
78
79 So, how good is our model result ``velsurf_mag``? :numref:`fig-csurfvsobserved` compares
80 the radar-observed ``surfvelmag`` field in the downloaded SeaRISE-Greenland data file
81 ``Greenland_5km_v1.1.nc`` with the just-computed PISM result. The reader might agree with
82 these broad qualitative judgements:
83
84 - the model results and the observed surface velocity look similar, and
85 - slow near-divide flow is generally in the right areas and of generally the right
86 magnitude, but
87 - the observed Northeast Greenland ice stream is more distinct than in the model.
88
89 .. figure:: figures/g-insar-20km-10km-comparison.png
90 :name: fig-csurfvsobserved
91
92 Comparing observed and modeled surface speed.
93
94 All figures have a common scale (m/year), with 100 m/year contour shown (solid black).
95
96 :Left: :var:`surfvelmag`, the observed values from SeaRISE data file
97 ``Greenland_5km_v1.1.nc``.
98 :Middle: :var:`velsurf_mag` from ``g20km_10ka_hy.nc``.
99 :Right: :var:`velsurf_mag` from ``g10km_10ka_hy.nc``.
100
101 We can compare these PISM results to other observed-vs-model comparisons of surface
102 velocity maps, for example Figure 1 in :cite:`Priceetal2011` and Figure 8 in
103 :cite:`Larouretal2012`. Only ice-sheet-wide parameters and models were used here in PISM,
104 that is, each location in the ice sheet was modeled by the same physics. By comparison,
105 those published comparisons involved tuning a large number of spatially-variable
106 subglacial parameters to values which would yield close match to observations of the
107 surface velocity. Such tuning techniques, called "inversion" or "assimilation" of the
108 surface velocity data, are also possible in PISM,\ [#]_ but the advantage of having few
109 parameters in a model is well-known: the results reflect the underlying model, not the
110 flexibility of many parameters.
111
112 We have only tried two of the many models possible in PISM, and we are free to identify
113 and adjust important parameters. The first parameter change we consider, in the next
114 subsection, is one of the most important: grid resolution.
115
116 .. rubric:: Footnotes
117
118 .. [#] Regarding the relative speeds of the runs that produce ``g20km_10ka.nc`` and
119 ``g20km_10ka_hy.nc``, note that the computation of the SSA stress balance is
120 substantially more expensive than the SIA in a per-step sense. However, the SSA
121 stress balance in combination with the mass continuity equation causes the maximum
122 diffusivity in the ice sheet to be substantially lower during the run. Because the
123 maximum diffusivity controls the time-step in the PISM adaptive time-stepping
124 scheme :cite:`BBL`, the number of time steps is reduced in the hybrid run. To see
125 this contrast use ``ncview ts_g20km_10ka*nc`` to view variables ``max_diffusivity``
126 and ``dt``.
127
128 .. [#] See :cite:`vanPeltetal2013` (inversion of DEMs for basal topography) and
129 :cite:`Habermannetal2013` (inversion surface velocities for basal shear stress) for
130 PISM-based inversion methods and analysis.