Modelling the Interactions Between Biospheric and Weathering Processes: Towards a Mechanistic Description of the Land Environment

Abstract

Recently, some tentative attempts to calculate the distribution of the rock chemical weathering rate over the continents and its variation from the last glacial maximum to the present have been published in the literature (e.g. Gibbs and Kump, 1994). These investigators used empirical relationships between the weathering rate of most important rock types and water runoff. The reconstruction of past weathering can then be performed from the water budget predicted by general circulation models. Although runoff is one of the most important environmental factors contolling chemical weathering at large spatial scales, many complex processes may influence weathering and its response to climatic forcings. For instance, temperature, orogeny or mechanical weathering (and the associated exposure of fresh mineral surface), and vegetation may exert non negligible controls on weathering, as widely discussed in recent years. In particular the effect of vegetation and soil microbial activity is recognized as important in promoting mineral dissolution (e.g. Schwartzman and Volk, 1989), through enhancement of the soil CO2 pressure and through organic secretion from plant roots and fungi. It is unlikely that simple statistical relationships calibrated on the present-day system can be successful in predicting chemical weathering rates under past climatic conditions when the atmospheric CO2 level and the vegetation distribution were very different from today. For this reason, it is necessary to build more mechanistic models of the coupled vegetation-soilrock system, which should describe in some detail the various processes involved in rock weathering, plant growth and soil microbial activity. However, the drawback of such a process-oriented approach is that it must involve an upscaling methodology from small (site, catchment) to large (river basin, continent) spatial units and from short (day, season) to long (103-106 years) time periods, including a validation at each level. Such an initiative has been pursued over the last few years in the biospheric community and global process-oriented models of the land biosphere are currently available. A similar effort may be conducted within the scientific community studying soil formation and weathering. Here we present a preliminary attempt towards a mechanistic description of rock weathering and its interactions with vegetation and soil biogeochemistry. Our approach is based on the chemical weathering model developed by Gwiazda and Broecker (1994) coupled to the CARAIB (CARbon Assimilation In the Biosphere; Warnant et al., 1994) model of the land biosphere. The CARAIB model couples various modules associated with the budget of soil water, photosynthesis, plant growth and respiration, and soil microbial oxidation of organic matter. The most important outputs are the soil water anaount, evapotranspiration, surface runoff, drainage at the bottom of the root zone, net primary productivity (NPP) of vegetation, autotrophic and heterotrophic respirations, and the contents of vegetation, litter and soil carbon reservoirs. The model is designed to be used at the global scale and, in such applications, a resolution of 1~ ~ in longitude-latitude is usually adopted. Nevertheless, in the case of local applications, the spatial resolution may easily be modified or eventually the calculation may only be performed for one single grid cell (test site). The temporal resolution is 1 day for the calculation of every water and carbon reservoirs. If input climatic data are provided monthly, a weather generator can be used to transform these data into daily inputs. This allows to take into account the high degree of non-linearity of