随着观测手段和气候模式的发展，全球平均气温变化趋势的检测与归因已经取了巨大进步，温室气体浓度增加已经被认为是导致全球变暖的主要原因。但是目前对区域气候变化的检测与归因却存在着十分显著的差异。探究区域气候变化空间格局的形成原因以及区域气候对全球升温的响应机制，对提升气候模式在区域尺度的模拟能力，降低区域尺度气候检测与归因的不确定性，提高区域气候未来趋势的预估能力有着十分重要的意义。本研究利用我国1960-2014年间逐日的气象观测数据，对地表温度观测序列进行了严格的均一化检测与订正，并在此基础上分析了我国1960-2003年间地气温的长期趋势与空间格局，然后基于地气能量平衡，定量评估了地表太阳辐射（Rs）和降水(P)变化对区域地气温升温格局的影响；此外，我们还对比分析了五种常用的再分析数据对我国气温日变化特征的模拟能力，包括CFSR、ERA-Interim、JRA55、MERRA2与NCEP-R2, 并且探究了入射太阳辐射和降水频次(PF)偏差对再分析中气温日变化的气候态、年际变率以及长期趋势的影响，主要研究结果如下： （1）在1960-2003年间，我国日最高地表温度（Ts-max）、日最低地表温度（Ts-min）与日平均地表温度（Ts-mean）中分别有53.1%、56.2%与50.2%的站点观测序列存在显著的非均一问题，共检测出断点2958、3458与2452个；在2004-2014年间，我国Ts-max、Ts-min与Ts-mean的时间序列中分别有62.2%、82.6%与70.0%的站点观测序列存在显著的非均一问题，断点数量分别为2948、3062与2948个。Ts-min的非均一性站点的数量以及断点的数量都明显高于Ts-max和Ts-mean，表明Ts-min对于观测环境的变化较为敏感且其序列波动较小，导致在观测环境发生变化时，Ts-min的观测序列中最容易产生较为显著的断点。在长期趋势上，1960-2003年间，订正后Ts-max和Ts-min的升温速率都出现了一定的降低0.01~0.02oC/10yr，而订正后Ts-mean的升温速率没有出现明显变化。在空间上，订正后我国南方地区Ts-max以及华北平原地区Ts-min的长期趋势明显增加，升温速率的区域一致性增强。1960-2014年间，年平均地表温度升温速率经过均值法和quantile-matching（QM）法订正后的升温速率分别降低了0.04~0.18 oC/10yr与0.03~0.14oC/10yr。订正后我国东北地区，西北地区和云贵高原地区的Ts-max升温速率明显降低，而我国华北平原，四川盆地以及东南沿海地区的Ts-max的升温速率显著增加；Ts-min在我国的升温速率普遍降低，降低幅度北方明显高于南方地区，尤其是年降雪量较大的区域。 （2）1960-2003年间，我国地表温度和气温总体呈显著上升趋势，全国平均Ts-max、Ts-min、气温日最高温（Ta-max）与气温日最低温（Ta-min）的升温速率分别为0.23±0.09、0,32±0.06、0.17±0.07与0.36±0.06 oC/10yr。但升温速率存在明显的区域差异，南方和华北平原地区Ts-max与Ta-max的升温速率明显低于其他区域。通过归因分析表明，入射太阳辐射（Rs）和降水量（P）是导致我国地气升温速率产生区域差异的主要原因。在这一时期，我国Rs呈显著的下降趋势，全国平均下降速率为-1.50±0.42(W/m2)/10yr。Rs下降导致我国Ts-max与Ta-max的升温速率分别下降了0.14与0.05 oC/10yr。更为重要的是，我国南方和华北平原地区Rs下降速率明显高于我国其他地区。在调整Rs和P的影响后,我国Ts-max和Ta-max升温速率的空间差异基本消失。例如，在调整Rs与P变化的影响后，我国华北平原和黄土高原地区Ts-max与Ta-max的升温速率差异分别下降了97.8%与68.3%；Ts-max和Ta-max升温速率的季节差异分别下降了45.0%与17.2%；而地气温的升温速率日较差分别下降了33.0%与29.1%。 （3）1980-2014年间，在订正海拔差异导致的偏差后，五种再分析产品能够很好地模拟我国Ta-min的多年平均值，但是对Ta-max与气温日较差（DTR）多年平均值的模拟存在显著的偏差。除MERRA2之外，其他再分析数据明显低估了Ta-max与DTR的多年平均值，低估幅度为-1.21~-6.64 oC；而MERRA2高估了Ta-max与DTR的多年平均值，高估幅度分别为0.35 oC和0.81 oC，但MERRA2中Ta-max与DTR的多年平均值与观测最为接近。对于年际变率，五种再分析产品都能很好地模拟出Ta-max（r=0.71~0.85）与Ta-min（r=0.68~0.79）的年际变率，但是对DTR（r=0.51~0.63）的年际变率的模拟较差，JRA55对气温日变化年际变率的模拟效果最好。再分析产品普遍低估了我国Ta-min的升温速率，低估幅度为-0.13~0.17 oC/10yr。我国西北地区Ta-max的升温速率被低估了-0.24~0.40 oC/10yr，而我国东南地区的Ta-max升温速率被高估了0.18~0.33 oC/10yr。DTR的长期趋势被高估了0.01~0.26 oC/10yr，尤其是我国东南地区和华北平原地区。再分析产品中Rs与降水频次（PF）的偏差是导致Ta-max与DTR长期趋势偏差的重要原因，其中Rs偏差能分别解释31.0~38.7%与31.9~37.8%的Ta-max与DTR长期趋势的偏差；PF偏差能分别解释9.8~22.2%与7.4~15.3%的Ta-max与DTR的长期趋势偏差。

Increases in observational data and rapid developments in simulation capacity of climate models have provided evidence for the phenomenon of global warming, and the increases in anthropogenic greenhouse gases and other anthropogenic effects are considered as the primary causes. However, significant spatial heterogeneities in climate warming have been observed and became an ongoing challenge for climate change research. Investigations of the spatial and temporal patterns of regional climate changes and regional climatic response mechanisms to global change are crucial for increasing the accuracy of models designed to detect and explain the causes of global climate change and predictions of future regional climate change. Herein, we explored the effect of land-atmosphere energy budget on the regional climate warming by the observations recorded at 2474 stations in China from 1960 to 2014. Firstly, we detected the homogenization of observations of the land surface temperature and adjusted the shift in time series in land surface records by RHtest V4. Secondly, based on the homogenized observations, we analyzed the surface and warming in China during 1960-2003 and quantified the impact of surface solar radiation (Rs) and precipitation (P) on them. Additionally, we evaluated the diurnal cycle of air temperature in five popular reanalyses, including Climate Forecast System Reanalysis (CFSR), the ERA-Interim reanalysis, the Japanese 55-year Reanalysis (JRA55), veRsion 2 of the Modern Era Reanalysis for Research and Application (MERRA2), and the National CenteRs for Environmental Prediction (NCEP)/Department of Energy (DOE) Global Reanalysis 2 (NCEP-R2), and analyzed the contribution of Rs and precipitation frequency (PF) to the bias of diurnal cycle of air temperatures in multiyear mean, interannual variability, and long-term trend. During 1960-2003, there are 53.1%, 56.2%, 50.2% observed records of daily maximum surface temperature (Ts-max), daily minimum temperature (Ts-min), and daily mean surface temperature (Ts-mean) being diagnosed as the heterogeneous time series, which have 2958, 3458, and 2452 shifts respectively. During 2004-2014, there are 62.2%, 82.6%, 70.0% observed records of Ts-max, Ts-min, and Ts-mean being diagnosed as the heterogeneous time series, which have 2958, 3458, and 2452 shifts respectively. The number of heterogeneous stations and total shifts for Ts-min is much larger than that for Ts-max and Ts-mean, which indicated the observed records of Ts-min is more sensitive to changes in the observed surrounding. In long-term trend, after adjustment, the warming rates of Ts-max and Ts-min decrease by 0.01~0.02 oC/10yr while the warming rate of Ts-mean have no significant change during 1960-2003. In spatial, after adjustment, the warming rate of Ts-max in South China and that of Ts-min in North China Plain increased, thus the spatial pattern of their warming trend becomes more consistency. During 1960-2014, after the Mean adjustment and quartile-matching (QM) adjustment, the warming rate of surface temperatures decrease by 0.04~0.18 and 0.03~0.14oC/10yr respectively. After adjustment, the warming rate of Ts-max in Northeast China, Northwest China and Yungui Plateau decreases, while that in North China Plain, Sichuan Basin, and Southwest China increases significantly. In this period, the warming rate of Ts-min increase over China and increase amplitude in North China is much higher than that in South China. During 1960-2003, the warming of daily maximum surface (Ts-max) and air (Ta-max) temperatures showed a significant spatial pattern, stronger in the northwest China and weaker in South China and the North China Plain. These warming spatial patterns are attributed to surface solar radiation (Rs) and precipitation (P), the key parameteRs of surface energy budget. During the study period, Rs decreased by ?1.50 (W/m2) /10yr in China and caused the trends of Ts-max and Ta-max decreased by 0.139 and 0.053 °C/10yr, respectively. More importantly, South China and the North China Plain had an extremely higher dimming rate than other regions. The spatial contrasts of trends of Ts-max and Ta-max in China are significantly reduced after adjusting for the impact of Rs and P. For example, the difference in warming rates between North China Plain and Loess Plateau reduce by 97.8% and 68.3% for Ts-max and Ta-max respectively. After adjusting for the impact of Rs and P, the seasonal contrast of Ts-max and Ta-max decreased by 45.0% and 17.2%, and the daily contrast of warming rates of surface and air temperature decreased by 33.0% and 29.1% over China. This study shows an essential role of land energy budget in determining regional warming. After adjusting the elevation difference between observation and reanalyses, we find that the reanalyses reproduce Ta-min very well during 1980-2014 over China; however, with the exception of MERRA2, they substantially underestimate Ta-max and DTR by -1.21~-6.84 °C. MERRA2 overestimates Ta-max and DTR by 0.35 and 0.81 °C, respectively, and yields the best reproduction of the climatologies of these variables. The five reanalyses display skill in reproducing the interannual variability of Ta-max (r=0.71~0.85), followed by Ta-min (r=0.68~0.79); however, they show relatively poor performance in reproducing DTR (0.51~0.63), and JRA55 displays the best performance of the reanalyses. The reanalyses consistently underestimate the warming trend in Ta-min by -0.13~-0.17 °C/10yr throughout China. The reanalyses underestimate the warming trend in Ta-max by -0.24~-0.40 °C/10yr in Northwest China and overestimate this quantity by 0.18~0.33 °C/10yr in Southeast China. These biases in the trends in Ta-max and Ta-min introduce a positive bias in the trend in DTR of 0.01~0.26 °C/10yr within China, especially in the North China Plain and Southeast China. The part biases in the trends in Ta-max and DTR are attributed to biases in the trends in Rs and PF in the reanalyses. The biases in the trends in Rs and PF explain 31.0~38.7% and 9.8~22.2% of the bias in the trend in Ta-max in the reanalyses and 31.9~37.8% and 7.4~15.3% of the bias in the trend in DTR in the reanalyses, respectively.