2D Daisy
========

Thoughts and time estimates of the developing of support for 2
dimensional water transport in the soil matrix and in macropores.  The
support must coexist and integrate with the existing 1D model.

Prototypical applications.
==========================

There are many potential uses, the two that will be used for design
are: 

1. Crops growing in rows.
2. Preferential flow directed to pipe drains.

Model thoughts.
==============

For the first version, we will reuse the existing 1D bioclimnate
model, even though we know it is inadequate for crops in rows.

The remaining work can be divided into upper boundary (soil surface),
lower boundary (groundwater), content (including a soil description),
transport, roots, and turnover.

Turnover and content.
---------------------

We how the current models for turnover (mineralization, nitrification,
etc.) can be reused, but the turnover code has to be divided from the
code managing the state (content).  We hope initially to be able to
reuse the 1D static description of the soil profile.

Roots.
------

A 2D root growth and uptake model must be developed.  

Surface.
--------

We don't actually know what to do here, in particular for a surface
with a slope, nor how to connect the 1D upper boundary of the soil
with the 0D lower boundary of the bioclimate.

Groundwater model.
------------------

Need to be updated to serve as a 1D boundary, but we assume all
boundaries will be uniform, except the pipe drain boundary which will
be reimplemented as a (very large) macropore.

Matrix.
-------

A finite-volume based model with a rectangular grid and a generic
source/sink term.  Should exist both for water and for solute.

Macropores.
-----------

We need to develop a 2D model which interact with the matrix through
the source/sink term.  Should exist both for water and solute.  The
macropores also need to act as a container, unlike the current 1D
model.  

Time Estimate:
==============

Mikkel: Total 2 years: 1 year matrix flow prototype in mathlab, 1/2
year C++ reimplementation, 1/2 year developing macropore model and
code.

Per: Total 2 years: 1/2 year on each of turnover and content
reorganization, root model implementation, the new boundaries, and
help with C++ implementation.

Total time estimate 4 years, which will give a working (non-prototype)
integrated tool that can solve one of the initial problems, and serve
as a base for the other.

We still need help developing the macropore model and root model.


