Investigating crop-climate interactions within a General Circulation Model (GCM)

There is increasing evidence that the land surface can significantly affect the overlying atmosphere and that changes in land use may have a substantial impact on the local climate. Global crop area has increased dramatically in the last few decades and now occupies a major part of the Earth's land surface. However, the increase in area appears to have levelled off recently suggesting that future increases in crop production may come from increases in the intensification of current cropping systems rather than extensification. The impacts of such changes on the local climate and hydrology are, as yet, unknown. Therefore, a realistic representation of crop growth and development as part of the land surface of atmospheric GCMs is required.

Tropical crop production is highly vulnerable to variations in seasonal precipitation. To better understand this relationship under current and future climates GCMs are ideal tools to provide the estimates of climate. Research has also shown that crop yield can be significantly impacted by sub-daily temperature variations suggesting that a fully coupled crop-climate model is required in preference to using daily GCM output to run crop models `offline'.

The disparity between the operational spatial scales of GCMs and crop models is vast. Most crop models are developed to work at the plot or field scales while GCMs generally operate with a horizontal resolution of the order 100km. The development of a crop model that works at a spatial scale similar to that of GCMs is a necessary first step in the development of a combined crop-climate model. The General Large Area Model for annual crops (GLAM) is a process-based crop model that has been developed to work at the regional scale, where a statistical relationship exists between the weather and crop yield.

In order to investigate the two-way interaction, the crop model (GLAM) must firstly be able to respond to the weather and climate of the atmospheric model (the UK Met Office atmosphere-only GCM, HadAM3). The simulated crop growth must then feed back into the surface characteristics of the land surface scheme of HadAM3 (MOSES2: the Met Office Surface Exchange Scheme). MOSES2, however, models the physiology of its chosen plant functional types and their interaction with soil water. This makes the coupling a non-trivial exercise.

Initial simulations of the coupled GLAM-HadAM3 model have been completed. The distribution of the tropical crop was determined based on an agro-ecological zones methodology considering the water availability and temperature requirements of the crop. The model results show that realistic crop growth is simulated in response to the climate of the atmospheric model. For example, in the seasonally arid tropics of India, only one growing season is simulated, while in the humid tropics of Papua New Guinea the climate is such that crop growth is possible throughout the year, resulting in the simulation of two full growing seasons per year.

This paper will describe the methodology adopted to couple GLAM and HadAM3 and present some results of coupled model simulations.