OR/17/068 Introduction
| Collins, S. 2017. Incorporating groundwater flow in land surface models: literature review and recommendations for further work. British Geological Survey. (OR/17/068). |
The land surface model Joint UK Land Environment Simulation (JULES) (Best et al., 2011; Clark et al., 2011) is a key component of NERC’s Earth System Modelling Strategy. It is used in global and kilometre-scale weather forecasting, global climate prediction and earth system modelling. Despite being at the cutting edge of international land surface modelling, particularly with regard to mass and energy exchanges with the atmosphere, its soil–hydrological components are highly constrained. HydroJULES is a NERC-funded project that brings together NERC Centre-Surveys — specifically the Centre for Ecology and Hydrology, BGS and the National Centre for Atmospheric Science – to investigate how to improve the simulation of the whole hydrological cycle in models. BGS’ role is to inform the inclusion of groundwater in both JULES and the hydrological model Grid to Grid (e.g. Bell at al, 2009). To facilitate this a literature review has been undertaken of the current methods for inclusion of groundwater in land surface models.
If we want to project the impacts of future climates, the way in which vegetation, soil and snow exchange water, energy and carbon with the atmosphere must be considered (Pitman, 2003). This is achieved with land surface models (LSMs), which provide physics-based descriptions of the processes involved.
Traditionally, LSMs focused on near-surface hydrology using the 1D Richards equation to calculate vertical flow in the soil (e.g. Gedney and Cox, 2003; Yeh and Eltahir, 2005a). The first LSMs did not include any simulation of groundwater, but instead applied a free drainage boundary condition to the bottom of a fixed soil column. However, in the last 20 years, the number of LSMs and their capabilities have increased significantly and groundwater simulation is now included in LSMs of all scales, from single basins to the globe. This report explores the range of methods used, from simplified lumped models (Section 2) to complex and computationally expensive distributed models (Section 3). In Section 3.3, the parameterisation of these models is detailed.