Improvement Of Surface Temperature And Warming Analysis

Global surface temperature data is key foundation of climate change research.Only few leading institutes worldwide have generated unique global land surface air temperature(LSAT),sea surface temperature(SST),and global surface temperature(ST)datasets.This paper discusses the newly developed global surface temperature dataset CMST(China Merged Surface Temperature),obtained by merging the China's homogenized and integrated global LSAT(C-LSAT 1.3)with SST(ERSST.v5 released by NOAA/NCEI).C-LSAT is firstly updated to increase data coverage over land area,and generate C-LSAT1.3.Compared to other LSAT datasets,such as GHCNv3,CRUTEM4 and Berkeley,C-LSAT1.3 shows improved performance in spatial and temporal distributions,and more data available grids.Compared with other datasets,CMST shows higher global ocean/land data coverage and consistency in terms of interannual-decadal variability and long-term trends.Moreover,the CMST captures well the regional climate change characteristics over Asia and the high-latitude zones.Furthermore,this paper uses results from 27 historical global climate model simulations of the International Coupling Model Comparison Phase 5(CMIP5),integrated from 1900 to 2005.Compared to CMIP5 and other global surface temperature datasets,CMST shows good consistency with the multi-model ensemble average(MAM)over Asia and at global scale.This paper also used CMST to systematically evaluate CMIP5 historical simulations over Asia in the period 1900 to 2005.Results from statistical analysis show that CMI5 reproduce reasonably the spatio-temporal characteristic of CMST.We select the 9 best model simulations(MT9)to generate a new ensemble.MT9 shows significant similarity to observations(MAM overestimates the temperature change)in both time variation and spatial distribution.In Asia,the warming rate is much smaller in MT9 than in MAM.Finally,the slowing down of climate change rate between 1998 and 2012(during the so-called "hiatus")and the trend since 1998 have been reinvestigated.Trend of C-LSAT is estimated to 0.247 ± 0.049 °C/decade for the period of 1998-2017,which is larger than for 1951-2017 and 1900-2017.Trend of CMST is estimated to 0.190 ± 0.036 for the periods of 1998-2017.The long-term global ST warming trend during the past century(1900–2017)remains essentially unchanged in most datasets,while the recent warming trend since 1998 increases slightly.As for the trend during 1998-2012,all the other global surface temperature data sets have underestimated the warming rate for both land area and the whole globe.However,the CMST global mean ST warming trend is close to those of the other global observational datasets which include satellite/Buoy infilled datasets,and the ERA-Interim reanalysis dataset.The weakened warming trend associated with global "warming hiatus" is clearly a conditional statistical artifact influenced by observation spatial coverages.