The importance of land surface processes on atmospheric boundary layer development and larger scale weather has been widely studied for over 30 years. This research has shown that exchanges of energy, moisture and momentum between the atmospheric boundary layer and the land surface are strongly influenced by vegetation and soil moisture. Changes in the atmospheric boundary layer dynamics also flow on to larger scale weather through entrainment with the troposphere and through convective cloud formation.

Given the importance of the land surface on the dynamics of the boundary layer, climate models need to incorporate land surface models, which realistically capture the spatial heterogeneity and temporal dynamics of real landscapes. Dynamic, heterogeneous landscapes are difficult to describe in climate models as the scale at which human and natural processes alter the landscape are often several orders of magnitude finer than the horizontal resolution of the climate model. This scale mismatch between landscape processes and climate models is a major barrier to investigating the climate impacts of landscape heterogeneity and anthropogenic land cover change.

Long term remote sensing and international data integration projects have started to address these scale issues by making global land surface data available at resolutions of the order of 1km. This research project investigates the use of newly available data in generating the land surface parameters of the CSIRO climate models at a common 8km resolution. A Simple Biosphere (SiB) land surface parameterisation has been developed with global fine scale vegetation, soil and satellite mapping. Simple aggregation techniques are used to take the land surface parameters from the 8km resolution to the resolution of the climate model experiments.

This paper details the global land surface data and processing methods used to generate the land surface parameters of the CSIRO Global Circulation Model (GCM) and the CSIRO limited area model (DARLAM). The new parameters are compared at 180 km, 45 km and 15 km resolutions, and with the original land surface parameters generated from coarser vegetation and soil mapping. The paper concludes with an investigation into the impacts of the new land surface parameters on the dynamics of the boundary layer and wider climate simulation over heterogeneous landscapes at the three atmospheric scales.