Land-atmosphere interaction and climate variability
MetadataShow full item record
Land-atmosphere interaction includes complex feedbacks among radiative, hydrological, and ecological processes, and the understanding of it is hindered by many factors such as the heterogeneity of land surface properties, the chaotic nature of the atmosphere, and the lack of observational data. In this study, several different methods are used to investigate the land-atmosphere interaction processes and their relationship with climate variability. Firstly, a simple one-dimensional model is developed to simulate the dominant soil-vegetation-atmosphere interaction processes in the warm climate. Although the physical processes are described coarsely, the model can be more easily used to find some relationships which may be drown out or distorted by noise. The influence of land on climate variability mainly lies in it memory, which is greatly related with the atmospheric forcing, so this model is used to investigate the influence of different forcing strengths on land-atmosphere interaction and its difference at different land covers. The findings from the simple model can provide guidance for other studies. The second part of the study compares a lagged soil moisture-precipitation (S-P) correlation (soil moisture in current day and precipitation in future 30 days) in three atmospheric reanalysis products (ERA-40, NCEP/DOE reanalysis-2, and North American Regional Reanalysis (NARR)), Global Soil Wetness Project Phase 2 (GSWP-2) data, and NCAR CAM3 simulations. Different datasets and model simulations come to a similar negative-dominant S-P correlation pattern. This is different from the traditional view that the soil moisture should have positive influence on future precipitation. Further analysis shows that this correlation pattern is not caused by the soil moisture feedback but due to the combined effect of the precipitation oscillation and the memory of soil moisture. Theoretical analysis confirms the above results and finds that the precipitation time series with the strongest oscillation at 32-60 day period is most likely to induce a significantly negative S-P correlation, and regions with longer soil water retention time are more likely to have a significantly negative S-P correlation. This study illustrates that a lagged correlation does not always indicate a causal relation.