Abstract:

An eddy-covariance system and a micrometeorological station were installed at south periphery of Gurbantonggut Desert in April of 2002. Evaluation of the data showed that the measurements of sensible heat (H), latent heat (LE) and CO₂ fluxes were reliable. However, a special phenomenon was revealed from detailed analysis on the diurnal and seasonal changes of the fluxes: when LE fluxes signified a strong water shortage and hence severe water stress should have been applied on the plants, CO₂ fluxes indicated that the photosynthesis of the plant community was not affected. This was obviously against generally-accepted theory on plant-water relations. No concrete evidence available so far can explain this phenomenon, but we may speculate that local shrub plant might not benefit from the rainfall. Being a salinized desert, its upper soil layer contains high salt content. Thus shrub roots could not develop and survive at upper layer of the soil. On the contrary, roots might develop mainly at deep soil layers near the groundwater table. Therefore, whether rainfall wetted the upper soil layer or not, plant water condition was not affected, but the evapotranspiration (LE) was. The measured change of LE fluxes mainly came from evaporation of soil surface and CO₂ fluxes were mainly determined by plant photosynthesis. Since the former is mainly influenced by soil water condition near soil surface and the latter determined by soil water condition near the groundwater table, it is just reasonable that the two aspects are not correlated. Thus, the phenomena reported here occurred under the combined local climatic, botanical and soil conditions and it is not really against the principle of plant-water relations in general.

Key words: desert, sensible heat, latent heat, net photosynthesis, groundwater table, Junggar