Land Surface Energy Exchange Over Mainland China And Preliminary Analysis Of Its Responses To Large-scale Climate Change

The water and energy exchange between the land and the atmosphere is an important process in the climate system. The land surface heat fluxes are determined by the local atmospheric and surface conditions, but their tempo-spatial variations may be affected by the weather and climate factors on regional and even larger scales.In turn, the tempo-spatial variations of surface heat fluxes possibly have a feedback on the weather and climate evolution.Based on the accurate construction of land surface flux datasets on regional and global scales, the paper investigated the relationship between land surface flux and large-scale climate change characterized by ENSO、PDO and global averaged temperature and possible physical mechanism.An inter-comparison and evaluation of land surface heat fluxes over China in seven reanalysis/modeling products is performed in this study to assess the capacity of existing flux datasets in presenting the tempo-spatial variations of heat fluxes on regional scale.To further improve the accuracy of land surface flux datasets, this study attempts to reconstruct a dataset with higher accuracy over northern China from 1979 to 2010 by synthesizing multiple data sources including ground measurements, reanalysis products, and remote sensing products.The constructed heat flux dataset was then applied to analyze the variability of land surface fluxes over northern China in response to large-scale climate change such as ENSO and PDO which play leading roles in the climate system.To further understand the response of the global land surface fluxes to ENSO and other large-scale climate change, the paper regarded the ensemble average of five global gridded flux dataset as real land surface flux so as to avoid the errors of a certain dataset, and then investigated the possible linkage between the global land surface fluxes and large-scale climate change. In this paper, the main conclusions are summarized as follows:(1) In general, the expected seasonal variation and spatial patterns related to the major climatic regimes and geographical features in China are well reproduced by all products. However, large differences in four reanalysis products are identified, while three off-line land surface model’s products are correlated well with each other, perhaps partly due to the common atmospheric forcing fields used for these models. From the perspective of Bowen ratio, off-line land surface models convert larger fraction of surface available energy into sensible heat flux than the reanalysis under all climatic regimes. In the field of sensible heat flux, there are three centers with higher inter-annual variability located in west of Xinjiang and Tibet, Northeast China, and eastern region of Inner Mongolia, respectively. With respects of types of underlying surfaces, the flux products behave better at grassland sites than at forest sites. They have relatively poor performances in describing temporal variation of sensible heat flux(Hs) at forest sites. In the mean time, the latent heat flux(LE)and net radiation(Rn) are significantly overestimated. Besides, the mean square errors of flux products are decomposed to describe the contributions of the bias, correlation and difference in standard deviation. The bias terms are more significant compared to the standard deviation term, particularly for the terms of LE and Rn. Finally a ranking system is adopted to compare the performance of all data sets for the given quantities. For Hs and Rn, ERA-Interim gets the highest scores. NOAH and CLM have the best performance in simulating LE, while ERA-Interim remains relatively higher score in LE simulation as well.(2) Assuming that in-site observations are closer to the real state of land surface energy exchange than the reanalysis/modeling datasets, a comparison of gridded flux data with the observations suggests that the gridded flux data interpolated to the measurement sites can be seen as a linear function of the observations. For one certain gridded flux dataset, there is little difference in linear relationships between the gridded data and the observations over different types of surface, which is an important statistical regularity for the dataset construction.Besides,cross-validation results showed that the accuracy of the constructed dataset is improved compared with these original datasets, largely because of contributions of reduced mean bias errors and standard deviation errors.The error of the constructed data introduced by the resample scale is also an emergent issue in the case of integrating multiple data sets at various spatial resolutions. The impact of the resample scale has been assessed to be small by applying the proposed data construction method to different grid sizes to compare the sensitivity to the resample scale among the constructed and original data sets. It is found that the constructed flux dataset can reasonably reproduce the spatial pattern and inter-annual variations of surface heat flux with the maxmum interannual variability over semi-arid regions. The land heat fluxes over northern China are consistent with the summer and annual averages of spatial characteristics of dry-wet condition. The sensible heat flux also shows a significant decrease in the west part of northern China accompanying its gradual wetter trend and an obvious increase in east part of northern China from 1960 to 2010.Meanwhile, the spatial distributions of correlations between turbulent heat fluxes and possible impacting factors such as the precipitation and net radiation indicate that the interannual variability of sensible and latent heat fluxes over northern China is dominated by dry-wet condition rather than net radiation.(3) Through the analysis of land surface fluxes over northern China in the next year when EN SO events occur compared with other years, we obtained the regularity how the interannual variability of land surface fluxes responds to ENSO events. The overall impacts of three types of ENSO events on heat flux exchange over northern China are also investigated. La Nina has no significant influence on interannual variability of land surface fluxes. EP El Nino significantly affects the variations of turbulent heat fluxes in northern part of Northeast China, while CP El Nino has an opposite but weak influence compared with EP El Nino. To find the maximum area where land surface fluxes significantly change due to impacts of ENSO, the spatial pattern of the land surface heat fluxes in extreme ENSO events (i.e., the year 1998) was further analyzed. In addition, through a comparison of the area where the variations of heat fluxes are usually affected by ENSO events with the area with the flux variations controlled by extreme ENSO events, the area where the variations of heat fluxes affected by ENSO events with different frequency was identified. In addition, the analysis of land surface fluxes in different PDO phases shows that the responses of land surface fluxes to PDO vary with regions of northern China. Generally, ENSO, PDO and other large-scale climatic factors through changing land dry-wet conditions significantly influence the turbulent energy partitioning, but have a weak impact on net radiation.(4) In addition, the relationships between the global average land surface fluxes and ENSO indices such as Nino 3.4 sst and SOI were investigated.There are obvious correlations between interannual variability of the global average land surface fluxes particular for sensible heat flux and ENSO indices, while the global average land surface precipitation is more closely associated with the degree of ENSO index.In the world, regions such as eastern Australia are hotspots where interannual variability of land surface fluxes is significantly influenced by ENSO. Responses of variability of land surface heat fluxes worldwide to ENSO events usually accompany significant change of precipitation.There is an obvious regularity in the spatial pattern of correlation coefficient of Land surface fluxes and global average temperature, indicating global warming is likely to be a background which causes variations of global land surface fluxes.