1. HiGEM A national UK programme in ‘Grand Challenge’ high resolution modelling of the global environment between NERC and the Hadley Centre

2. HiGEM Goals Science Aims:
To achieve a major advance in the fidelity of simulations of the Global Environment by developing an Earth System Model with unprecedented resolution, comprehensively evaluated and verified.
To perform a ‘Grand Challenge’ multi-century simulation with the new Earth System Model, HiGEM
To study mechanisms of climate variability and change on timescales of days to centuries
To improve our understanding of non-linear processes that lead to interactions between small and large spatial scales, and between high and low frequencies, within the Earth system.

3. HiGEM Goals 2. Strategic Aims:
To integrate the global environment modelling activities across NERC
To support the formation of a national partnership between NERC and the Hadley Centre
To exploit the major increases in computing power afforded by HPCx (UK) and the Earth Simulator (Japan)

4. HiGEM Timeline
A three year programme ( ) to build on the HadGEM1 model:
Hadley Centre: ~3 atmosphere ~1 ocean
HiGEM 2004:  atmosphere 1/3 ocean
HiGEM 2006: 2/3 atmosphere 1/6 ocean

5. Why a consortium is needed:
Complexity of the Hadley Centre Global Environment Model
No single group has the combined expertise to develop, evaluate and exploit HiGEM
‘Grand Challenge’ science requires highly coordinated approach 6

6. HiGEM Partners
Hadley Centre: Provision of HadGEM1, expert advice on climate modelling
NCAS/CGAM: Atmospheric processes, coordination, HPC expertise
British Antarctic Survey: Polar processes, modelling the cryosphere
Centre for Ecology and Hydrology: Land surface processes and modelling
Environmental Systems Science Centre : Clouds and radiation processes, model evaluation against satellite data
Southampton Oceanography Centre: Ocean processes and modelling, remote sensing
University of East Anglia: Ocean processes and modelling
British Atmospheric Data Centre: Data management

7. Why a move to higher resolution is necessary:
Non-linear feedbacks in space and time in the climate system
Improvements in parametrisations can only reduce the systematic errors in models so far, some processes need to be explicitly resolved
Organisation and coupling in atmospheric and oceanic processes requires representation of nonlocal interactions (waves need to be adequately resolved)
Local variations in land surface types and orogaphy can have very large impacts on local climate (e.g. precipitation)
Better understanding and quantification of physical processes in models that look more like the real world

8. Complexity in the atmosphere
Water vapour and window channel radiances from Meteosat-7
Same fields from HadAM3 at climate resolution

9. Better weather systems in a high resolution model
Precipitation in HadAM3 with 150km (left) and 300km (right) atmosphere
……and stronger storm tracks
High pass (<10 days) 500hPa geopotential height variance

10. Northern Hemisphere Blocking – based on theta on PV=2
Blocking frequency increases with horizontal resolution of model
Gill Martin (Hadley Centre)

11. Complexity in the ocean10
…. and as simulated by OCCAM
1/40
1/120
SSTs in the Gulf stream from infrared measurements aboard MODIS

12. Complexity in the land surface
How the Rockies appear in a standard resolution (~400km) model
… and in a high resolution (~60km) model.
Large variations in vegetation are not resolved but have impacts on local climate

13. Scales of orography that are credibly resolved
Alps 500 km (~ 2 gridpoints 3 resolution model)
Andes 300 km
Rockies 1200 km
Davies and Brown (2001)

14. Complexity in the cryosphere
AVHRR image of the southern Weddell Sea showing coastal polynyas (typically 100km x 20km).
Polynyas are sites of intense air-sea interaction and play a major role in production of Antarctic bottom water

15. Progress Update: Port of HadGEM1 to HPCx to start very soon
Defining ancillary files to work with 1 atmosphere 1/3 ocean (land/sea mask from GEBCO dataset)
Development and testing of HiGEM model due to start May 2004 to coincide with upgrate to HPCx 6

16. General Bathymetry Chart of the Oceans - GEBCO
New 1-minute grid available
Will be used to define 1/3 HiGEM ocean model
Needs careful checking as to representation of islands, straights and sills appropriate for this resolution

17. HPCx 40 IBM pSeries 690 Regatta nodes
Each node has 32 POWER4 processors
1280 processors total
April-June 2004 upgrade to 2X performance with faster switches