| The precipitable water (PW) is one of the most abundant and important greenhouse gases, and is the link between the substance and energy in the earth system, also plays the most important role in the energy balance and water cycle in the world, and the PW has the important effect on the weather and climate change. The feedback of water vapor is one of the biggest effects which is sensitive to the climate system, on the other hand, the feedback of water vapor can enlarge the heating effect of other greenhouse gases, and could lead to the occurrence of extreme weather and climate events become the more and stronger. Therefore, analyzing the spatial and temporal distribution characteristics and long-term change trend of atmospheric water vapor systematically and evaluating the regional and global climate effect of water vapor feedback are of great significance for our further understanding of the global warming and the response mechanism of the water vapor.In this paper, we made full use of the primary and homogenized air humidity observation data, and analyzed the spatial and temporal distribution and long-term trend of water vapor systematically combining with a variety of global atmospheric reanalysis data and the results of numerical simulation. Then we forecast its future changes and assess the relationship between the moisture change and the climate warming in China, all of the above results which can provide scientific basis of the impact assessment for China’s climate change. Therefore, we first assessed the applicability of 9 kinds of global atmospheric reanalysis products which are NCEP/NCAR, NCEP/DOE, MERRA, JRA-55, JRA-25, ERA-Interim, ERA-40, CFSR and 20CR in the atmospheric water vapor changes of China by applying the homogenized air humidity observation data. Then we evaluated the ability of simulating of the international Coupled Model Inter-comparison Project stage 5(CMIP5) multi-model results (BCC-CSM1.1, BCC-CSM1.1(m), CanESM2, CNRM-CM5, MTROC5, MRI-CGCM3 and NorESM1-m) for the atmospheric water vapor changes in China systematically based on the observation results, and analyzed the future scenarios of the water vapor changes under the different concentration of the different typical concentration path (RCPs). Finally we further estimated the quantitative relationship between the surface temperature changes and water vapor changes, and clearly understand the climate effect of water vapor feedback and its uncertainty of China.Due to the observation data have been continuous since the 1970s, this study focuses on the research of water vapor changes and feedback effects after 1970s in China. The main research methods include:The analysis of percentage of error, long-term trend, time series, Taylor, EOF and so on.The main contents and conclusions are as follows:(1) In most areas of China, most of reanalysis data can better describe the climatological characteristics of the observations, especially in the east of China. The PW differences between the reanalyses and the observations are within-20% for most of northern and eastern China, but the reanalyses underestimate the observed PW by 20%-40% for western China, in particular by-50% for the western Tibetan Plateau. In most of the eastern and northern China, most reanalysis data can describe seasonal and inter-annual variations of water vapor which observations show, but have poor describing ability to observations in most regions of Western China, especially in the Tibet plateau. An Empirical Orthogonal Function (EOF) analysis suggests that the reanalysis products, especially the newer-generation ones, are capable of depicting the spatio-temporal variations of the leading PW modes from observations in China during 1979-2012. For the second mode of EOF which is related to ENSO change, the 20CR can reflect the spatial and temporal evolution of the observation best among all of reanalysis products.(2) In all of the reanalysis data, the newer-generation reanalyses (such as: ERA-40, ERA-Interim, MERRA, JRA-55 and JRA-25) reproduce well the observed PW climatology and inter-annual variations with smaller root mean-square (RMS) error and bias than the older-generation ones (NCEP/NCAR and NCEP/DOE) over most China. Although the reanalysis data of water vapor have a certain ability to describe the variation of water vapor in China, it obtained from integration model which can influenced by the error of the model itself system, assimilation technique and observation system error and other factors, and can not completely replace the observation data to reflect the real state of the atmospheric state and objectively reflect the characteristics of climate change. Therefore the reanalysis data of water vapor should be used with cautious for climate change research.(3) Most of the coupled models and their ensemble can well depict the spatial variation and increased long-term change of water vapor which the observation showed, especially in the southern China, but the variation and the trend pattern weaker than the observation value, and there shows a clear dependence between temporal and spatial variation to a certain extent. The CanESM2 and MIROC5 can better performance long-term changes of water vapor which the observations show in China than other models, In addition, the performance of ensemble of all models is better than them with individual.(4) In different typical concentration path (RCPs) scenario we can see there is an increasing trend of water vapor in the future, especially in the RCP8.5 scenario whose the growth rate is 2 times than the RCP4.5 scenario. The water vapor may be decreasing over a dry region, and increasing over a wet region in the future showed by high emission scenarios, the change is more obvious in the RCP8.5 scenario. The change of water vapor also exists obvious regional and seasonal dependence, we predicted that North of China would be one of increasing significantly areas about water vapor in future.(5) Trends in long term of the PW were suffered from not only the temperature but also the relative humidity. The results of this paper showed the trends in surface temperature and PW were well correlated between observations, reanalyses, and the results of CMIP5, and the significant positive correlation was found between the surface temperature and PW, which the correlation from observation is 0.65, from the reanalyses ranged from 0.5 to 0.6, from CIMP5 results is more than 0.6, and the value of the ensemble of multi-models even is 0.92. In conclusion, most of the reanalyses and model results could well capture the trends of warming and wetting which showed by observation data over China. The correlation between the surface temperature and the PW (dPW/dT) for the observation is 4.9%K-1, which means the water vapor will change about 4.9% when the surface temperature rising 1℃, and the result is less than the theoretical results of Clausius-Clapeyron equation (C-C). For the reanalyses, like MERRA, JRA-55, JRA-25 and ERA-40, the values of which are slightly more than 4.9% K-1 which implied that the reanalyses overestimated the correlation of the observation, and less than values calculated from C-C. For the CMIP5 results, most of models overestimated the correlation of the observation, and the values of dPW/dT for most of models are 6%K"1. Moreover, the results of dPW/dT from CNRM-CM5, MRI-CGCM3, NorESM1-M and their ensemble are more than the value of C-C showing. |
PDF Full Text Request