Changes from modelE (AR4) to modelE (AR5)
=========================================

- Higher resolution versions available (2x2.5)

- Cubed Sphere dynamics and grid (C90) resolution

- Multiple changes to cloud schemes

- Incorporation of Menon et al aerosol/cloud microphysics

- Full atmosphere gas phase chemistry

- Ocean biology and full carbon cycle

- Coupler layer for different ocean grids

- Fuller diagnostic output

(note these are a very condensed summary of the differences).

Changes from modelE-2-3-4+ to modelE1 (or 3.0) (AR4) 
====================================================

- big expansion of tracer coding 

- relative humidity sensitivity of aerosol optical thickness

- new cloud micro-physics

- better separation of convective and stratiform cloud

- more interactions between tracer fields and radiation

- NUDGING facility for trying to reproduce field experiments

- full use of atmospheric turbulence 

- improved historical fields for ozone, aerosols

- new conductance scheme for evapo-transpiration

- extended sub-daily diagnostics

- new diag Hdirun for time series of all fields over last month for
  selected grid boxes

- option to run with smaller physics time step 


Changes from modelE-2-3 to modelE-2-3-4+
=======================================

- new conductance scheme option for land surface

- preliminary incorporation of wet and dry deposition for gases and
  aerosols

- angular momentum conservation for UV filter and GW drag

- better modularisation of snow model

- improved option for sea ice albedo (Hansen)

- minor bug fixes and diagnostic improvements

Major changes from version modelE-2-1 to modelE-2-3
===================================================

- Glacial melt (icebergs) now available as input into ocean

- time varying crop fraction possible

- monthly files for the CH4 oxidation into stratosphere

- Further extensions to sub-daily diagnostics to cover all
  geopotential heights, temperatures and velocities, and tropopause
  height

- New scoring statistic for comparsion to observations

- Tested versions at 2x2.5 and 8x10 horizontal resolutions

- More options in stratospheric drag formulation.

- snowice formation now an option 

- zenith angle dependence for all snow/ice albedos

- New ISCCP cloud optical depth/height histograms diagnostics.

- water tracer code in soils

- orbital parameter calculations based on paleo-year

- surface velocities (ice, ocean) no used as input into surface
  boundary conditions.


Major changes from version modelE-2-0-1 to modelE-2-1
=====================================================

- Qflux model now has option to include ice advection.

- new options for controling cloud optical thickness

- Energy conservation now guaranteed

- New physics for deciding on phase of precip (i.e. how quickly
  super-cooled precip turns to ice)

- New coding for ground hydrology

- Small corrections for q-flux/seaice modules

- Improved conservation diagnostics (including new text demonstrating
  conservation properties "conserv.txt")

- Extension to sub-daily diagnostics to include cloud cover (low, mid, high) 
  and relative humidity values.

- Enhanced diagnostics of cloud radiative properties

- Code is now compatible with PGI (Portland Group) compiler.

Major changes from version modelE-2-0 to modelE-2-0-1
=====================================================

 - Minor fixes to almost all components

 - New options for snow-age definitions, Sea-ice and SST cycling.

 - netcdf output now fully functional

 - Changes to makefiles for compilation using Linux compilers (Absoft
   + Lahey/Fujitsu)

Major changes from version modelE-1-3-3 to modelE-2-0
=====================================================

Atmospheric Model:

 - Advection of Q now done outside dynamics loop

 - New sea ice albedo scheme (pudlling and snowage replaced with
   prognostic melt-pond fraction and wet/dry snow flag).

 - solar radiation penetration through sea ice now wavelength dependent

 - the basal fluxes between the ice and ocean are calculated using a new
   turbulence scheme. New routines UNDERICE, iceocean, icelake.

 - parallelisation directives for atmospheric model

 - changes to boundary layer cloud specifications

 - improvements to turbulence calculations (including a fix of a bug
   that caused problems when the first layer was too thin)

 - fix for mass fluxes over topography to prevent unrealistic
   precipitation hotspots

 - Multiple vertical resolutions available

 - new interface for radiation to allow more control of greenhouse
   gases, aerosols etc. from rundeck.

 - Adjustable input of CH4 related stratospheric water vapour

 - Numerous parameters for the gravity wave drag  adjustable from the rundeck 

Coupled Model:

 - Gary's ocean model now fully coupled (plug and play option)

 - Ice dynamics now a completely separate module

 - Ocean heat transport diagnostic in printout

 - much more extensive ocean diagnostics

Tracer code:

 - tracer code for air mass and water mass tracers (uses preprocessing
   directives set from the run deck)

 - wet/dry deposition code for general tracers (aerosols, particles,
   water and soluble gases)

 - multiple example tracers pre-coded

Diagnostics: 

 - Numerous errors in diagnostic printout fixed

 - Re-arrangement of budget pages to provide more relevant info.

 - Conservation diagnostics for AM and KE actually add up.

 - Configurable daily and sub-daily diagnsotic output for selected quantities

Run Envrionment:

 - .modelErc file in home directory to enable a configuration to set
   up a local distribution on any machine

 - automatic compilation of online documentation for any particular
   configuration. 

 - new 'lock' file method for preventing un-intended writing over of files.

 - MUST USE COMMANDS 'runE' and 'sswE' to start and stop model

 - automatic re-starting with smaller timestep for some crashes (set from 
   initial rundeck)

Major Changes from Model II' to modelE-1-3-3
============================================

Variable name changes/re-definitions:

  - FDATA replaced with FLAND/FOCEAN/FLICE/FLAKE/FEARTH

  - GDATA replaced with:
    SNOWI,HSI  (sea ice)
    SNOWE,TEARTH,WEARTH,AIEARTH,SNOAGE (earth)
    TLANDI, SNOWLI (land ice)

  - ODATA replaced with: TOCEAN, RSI and MSI (TLAKE over land)

  - Second order moments are condensed to one array. Whether the
    quadratic or linear upstream scheme is required now only requires
    a change in the QUSCOM/QUSDEF modules. Multiple advection routines
    replaced with one general purpose routine.

  - Arrays that are used for column physics only are now written with
    the column index first (i.e. L,I,J) for performance improvement
 
  - Ground temperatures for SURFCE/RADIA and PBL are now saved in the
    GTEMP array.

  - parameters for both the model and all sub modules are now saved in a
    parameter block that is of variable length and self-documenting
    (replacing JC/RC in the old code).

Structural and Fortran 90 related changes

  - Common blocks replaced with physics-related MODULEs

  - Model common block (i.e. BB399xyz.COM) now MODEL_COM MODULE

  - physical constants all defined in CONSTANT module

  - Each physical submodule (ie. ice, snow, ocean, land, land ice,
    clouds,etc.) has it's own fortran MODULE. As far as possible,
    different modules interact only through fluxes.

  - initialisation of each sub-module is now in init_PHYS (if required)

  - daily processes moved into daily_PHYS (if needed)

  - Many errors in diagnostics and in the seaice/Qflux model corrected

  - lakes are seperated from ocean variables/code.

  - sea ice now seperated from ocean/lakes code. All fluxes are now passed.

  - Explicit use of B-grid velocity points is now implicit.

Diagnostics

  - pdE (used to print diagnostics) is now just an alternate entry to
    the model, and thus cannot be out of date.

  - diagnostic arrays now addressed by name, not number

Input/Output

  - Named files in rundeck instead of unit numbers

  - FILEMANAGER module to allocate unit numbers

  - input/output decentralized to io_PHYS routines for each PHYS submodule

Runtime environment

  - compiling/linking/setup now under control of an automatic makefile process

  - NAMELIST input:
      TAUI/E replaced with YEARI/E,MONTHI/E,DATEI/E,HOURI/E (alt. IHOURE)
      NDPRNT/NDZERO replaced with NIPRNT and NMONAV
      DT related variables such as NRAD etc, now relate to DTsrc
      (=1hour)
      SIGE setting replaced with PLTOP (pressure at the top of
      the level) - SIGE is now computed.
      U00 replaced with U00wtr and U00ice (separate numbers for
      ice and water clouds)

  - Local parameters are now stored in the parameter database which mainly 
    supplants the NAMELIST input. 


New physics

  - Sea ice over the ocean is now formed at TFO (=-1.8 deg).

  - Solar radiation now penetrates into (and through) the sea ice.

  - Sea ice salinity is now a prognostic variable (though not yet
    fully coupled to the thermodynamics).

  - lake temperature/ice cover are now predicted using a two layer
    energy/mass conserving model.

  - Turbulence can now be applied throughtout the atmosphere (ATURB.f).
