Principal Investigator Tony Jakeman Project r52

Centre for Resource and Environmental Studies Machine VP

Co-Investigators Robert J Oglesby, David P Hansen and David A Post,

Department of Earth and Atmospheric Science, Purdue University and Centre for Resource and Environmental Studies

Coupling Surface Hydrology Models to the CCM

The project aims to incorporate a surface hydrology model, IHACRES, into a Global Climate Model (GCM) to improve the description of water exchange between the land surface and the atmosphere. IHACRES is an acronym for Identification of Unit Hydrographs And Component flows from Rainfall, Evaporation and Streamflow data. CCM2, the lastest version of the USA National Center for Atmospheric Research GCM, will be used as has a much higher spatial resolution than earlier models and hence can be better coupled to IHACRES, which is most applicable at the basin or catchment scale. A strategy has been developed whereby IHACRES will be tuned to selected basins and then used with CCM2 to generate stream runoff and evaporation for all (model) basins. The modelled stream discharges will be compared to available observations, while the IHACRES-supplied evaporation will be compared to that computed directly by CCM2. The goal is to predict runoff and evapotranspiration responses to climatic inputs (changes or variability) on a diurnal basis. The model should also be applicable to assessing effects of prescribed vegetation changes on these water balance response terms.

What are the basic questions addressed?

Current models of the water exchange between the land surface and the atmosphere are overly simplistic in terms of their physics (e.g. bucket models) or overly complicated and parameterised (e.g. Biosphere-Atmosphere Transfer Schemes). IHACRES has proven to be a relatively simple but effective model, incorporating the basic physics and having around six parameters depending on the climatological regime, that has been shown to be applicable in a wide variety of hydroclimatologies. Its computational simplicity means it does not hamper the performance speed of GCMs and its accuracy compares extremely favourably with all other models, including more complex models. The aim of the project is to make this development available and to show the improvement attainable with its use.

What are the results to date and future of the work?

Extensive experiments are being made in which IHACRES is driven off-line by CCM2 results. A number of large basins (over 500 km2), which have been well studied and have higher confidence in previous studies using IHACRES, are selected as regions of interest. The first experiment is for Upper Goulburn Basin, located upstream of Lake Eildon and the east corner of Goulburn Basin, Victoria. The CCM2 simulated daily temperature and rainfall are used as inputs to drive IHACRES to yield simulated runoff explicitly and evapotranspiration implicitly. Comparison of CCM2 simulated and observed precipitation shows that the magnitude of CCM2 precipitation is close to observation for winter and becomes way too much during summer. Although significant rainfall occurs in the control CCM2 simulation, virtually no runoff occurs during the entire year (in other words, Australia's largest river system does not exist). On the other hand, IHACRES produces runoff when driven by the above CCM2 results, and the runoff generally follows seasonal variation pattern of the rainfall, and therefore is much more reasonable in context of rainfall- runoff processes.

What computational techniques are used and why is a supercomputer required?

In climate models a range of computational techniques are used including forward numerical solutions of differential equations at grid points over the entire globe. The hydrological model uses a transfer function or discrete difference equation at each grid cell. A supercomputer is needed because of the large amounts of computation, in time and space, and the mass storage which is required.