tevolutionary-vs-diagnostic.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
---
tevolutionary-vs-diagnostic.rst (1977B)
---
1 .. include:: ../../global.txt
2
3 .. _sec-basicmodes:
4
5 Evolutionary versus diagnostic modeling
6 ---------------------------------------
7
8 The main goal of a numerical ice sheet model like PISM is to be a dynamical system which
9 evolves as similarly as possible to the modeled ice sheet. Such a goal assumes one has the
10 "right" climate inputs and parameter choices at each time step. It also assumes one has
11 the "right" initial conditions, such as an adequate description of the present state of
12 the ice sheet, but this assumption is rarely satisfied. Instead a variety of heuristics
13 must be used to minimally-initialize an ice sheet model. For options associated to
14 establishing mathematical initial conditions when first starting PISM, see section
15 :ref:`sec-initboot`.
16
17 Inside PISM are evolution-in-time partial differential equations which are solved by
18 taking small time steps. "Small" may vary from thousandths to tens of model years, in
19 practice, depending primarily on grid resolution, but also on modeled ice geometry and
20 flow speed. Time steps are chosen adaptively in PISM, according to the stability criteria
21 of the combined numerical methods :cite:`BBssasliding`, :cite:`BBL`.
22
23 However, especially for ice streams and shelves, non-time-stepping "diagnostic" solution
24 of the stress balance partial differential equations might be the desired computation, and
25 PISM can also produce such "diagnostic" velocity fields. Such computations necessarily
26 assume that the ice geometry, viscosity, and boundary stresses are known. Because of the
27 slowness of the ice, in the sense that inertia can be neglected in the stress balance
28 :cite:`Fowler`, such computations can determine the ice velocity.
29
30 Sections :ref:`sec-start` and :ref:`sec-ross` give examples illustrating evolutionary and
31 diagnostic modes of PISM, respectively. The first describes time-stepping evolution models
32 for the Greenland ice sheet, while the second describes a diagnostic SSA model for the
33 Ross ice shelf.
34