Modelling earth system processes involve interactions between oceans, atmosphere, cryosphere, land surface dynamics, and human systems. Earth system models encompass all the theses components and more. Regional scale such processes are essential to understand the climate variability and change for regional impact studies. Such models setup will provide a framework for regional assessment of climate change impacts for adaptation and mitigation strategies. Such framework will also provide an opportunity to quantify the uncertainty associated with modelling chain in regional climate change projections. Within the framework of CORDEX a regional earth system model is being developed for CORDEX South-Asia domain for climate change studies. Figure shows the climate modelling chain to dynamically downscale climate change scenario over South-Asia.

Atmosphere (REMO) model
• The atmospheric component of ROM is the Regional atmosphere Model ( REMO), from Germany..
• REMO's prognostic variables are the surface pressure, temperature, water vapor, liquid water, horizontal wind components, and cloud ice.
• REMO uses a rotated grid, with the equator of the rotated system in the middle of the model domain, to avoid the largely different extensions of grid cells close to poles.
Ocean (MPIOM) model

• The oceanic component of ROM is the Max Planck Institute Ocean Model ( MPIOM), developed at the MPI-M, Hamburg, Germany.
• MPIOM is a free surface, primitive equations ocean model, which uses the Boussinesq and incompressibility approximations.
• The model is formulated on an orthogonal curvilinear Arakawa C-grid with z-level vertical discretization.
Hydrological Discharge (HD) model
• Hydrological component of ROM is the Hydrological Discharge ( HD), developed at the MPI-M, Hamburg, Germany.
• It is a state-of-the-art discharge model that is applied and validated on the global scale.
• In the HD model, the lateral water flow is separated into the three flow processes: overland flow, base flow, and river flow.

All the above models are coupled via oasis coupler as shown in the figure. Details of RESM is discussed in Sein et al., 2015.

The proper description of cryospheric processes is essential for simulating the complete terrestrial water cycle in climate models. So far, the terrestrial cryosphere is represented in an extremely simplified way in most state-of-the-art global and regional climate models (GCM and RCM) as they use Static Glacier Masks, i.e. (i) no changes in ice extent, no feedback to atmosphere; (ii) no consideration of water volume stored; (iii) no/simplified runoff generation; (iv) exception: ice sheet models in ESM. Therefore, more sophisticated approach is necessary (i) if contribution of glacial melt-water to runoff, sea level rise is important; (ii) regional climatic changes in glacierized areas are to be assessed focus on discharge in glacierized river basins (Kotslarski et al., 2010).

In REMO a dynamical glacier scheme (DGS) is implemented. DGS have unique ability to simulate the glacier mass balance and dynamically adjusted surface fraction of the each grid box depending on the accumulation and ablation conditions, accounting for direct physical feedback mechanisms. It is applicable for entire mountain ranges and computationally effective, target resolution: RCM grid cell. The coupled model system is referred to as REMO_glacier. The scheme represents surface glacier cover on a subgrid scale and calculates the energy and mass balance of the glacierized part of a grid box. In each time step the full surface energy balance of the glacierized fraction is computed according to:

dQice/snow = K+L+H+LE+G+M

REMO_glacier is first applied over the region using reanalysis data to test the model quality (Kumar et al., 2015). Figure shows the glacier fraction in REMO_glacier gridbox with topography.