海气通量的变化、趋势及其气候效应：观测与模拟研究

发布时间：2018-04-16 09:36

本文选题：海气热通量 + 潜热通量　；参考：《中国科学技术大学》2016年博士论文

【摘要】：海气界面热通量对大气和海洋环流、大气-海洋淡水循环起着至关重要的作用。研究其物理机制和变化形态对于理解和数值化模拟动态的大气-海洋过程有着非常核心的意义。本博士论文使用观测数据、再分析数据和气候模式输出数据研究了包括海表蒸发、潜热通量、净热通量等海气通量的变化、趋势及气候效应,并探讨了对应的模式模拟误差的可能来源。根据伍兹霍尔海洋研究所(WHOI)的客观分析海气通量项目提供的OAFlux数据,全球海表蒸发量在1999-2000年发生年代际趋势转变,从上升趋势转变为下降趋势。2000-2013年海表蒸发量的减少除了由近海表风速减弱所导致的非局地影响外,主要是由海表温度增暖减速导致的海表饱和比湿增速减缓和近海表空气比湿增速加快所导致的。首先,近海表风速的减弱通过直接和间接作用导致全球海表蒸发的减少。另外,海洋上空的大气柱总水汽含量总体增加以及海洋上空的降水减少导致余留在空气中的水分增加,直接导致了近海表空气比湿的增加。海表蒸发率的趋势空间形态都呈现总体减弱,海洋上的降水总体减少,而沃克环流的最强上升支区域则呈现明显的降水增加趋势。陆地上的总降水量增加,总蒸发量上涨。不论是陆地还是海洋,蒸发和降水的关系都更接近局地相关性。这种相关性使我们更加倾向于认为海洋上空的水汽增加是海气界面本身的水汽收支平衡被打破所致。使用OAFlux数据集评估了CMIP5计划中14个耦合气候模式对热带和副热带太平洋海表潜热通量的模拟结果,并分析了可能的模拟偏差的来源。结果表明模式模拟的潜热通量的空间分布形态与观测结果匹配的很好,不过模拟值一般比观测值偏高达到20-30 Wm-2。模式在空气比湿和海表温度气候态的模拟上表现较好,近海表风速气候态的模拟偏差很可能就是模式对潜热通量气候态模拟偏差的来源。使用模式间EOF分析方法表明,模式偏离平均态的空间一致性主要是由模式对海表温度的模拟差异造成的。对于潜热通量气候态的季节变化,模拟结果和观测数据吻合地很好,主要的模拟偏差与近海表风速季节变化的模拟偏差有关。而对潜热通量线性趋势的模拟评估显示,模式对海表温度和近海表空气比湿的模拟整体上偏大,由此导致的海气比湿差△q的趋势对潜热通量的反馈作用将会非常微弱,同时模式对近海表风速趋势的模拟偏小许多,两个因素可能就共同导致了模式对潜热通量增加趋势的模拟值极大地偏低。使用由中国科学院大气物理研究所大气科学和地球流体力学数值模拟国家重点实验室(LASG/IAP)研发的全球海洋-大气-陆地系统模式FGOALS-g2和FGOALS-s2,对比分析了它们对海气界面热收支的模拟结果。在热带印度洋海域,FGOALS两个模式对海表净热通量都存在大约30 W m-2的海盆尺度的低估,潜热通量的大幅高估被认为是净热误差的最主要来源。两个模式共同体现了印度洋偶极子状的误差分布。通过分析年际序列,FGOALS-s2对净热通量的模拟误差主要来自于向下的净短波辐射和向七的潜热通量的高估,而FGOALS-g2的模拟误差主要来自于其对湍流热通量的高估。在观测数据中,净热通量在热带印度洋中心海域存在极大的降低趋势,而这个特征在模式中未能体现。潜热通量的趋势主导了净热通量的趋势,因此模式对于净热通量趋势的模拟误差主要来源于其对潜热通量的模拟误差,即低估了潜热通量的增加趋势。在太平洋和大西洋海域,总体上对净热通量也存在海盆尺度的低估。对特征区域(西部边界流如黑潮、墨西哥湾流：赤道太平洋和大西洋冷舌)进行单独的分析可以知道,虽然在极值区模拟值偏大,但是模拟值的递减幅度较大,导致其高值区域相对较小。模式模拟的各分量趋势分布均不能和OAFlux显示的趋势分布形态相匹配,但净热通量趋势分布主要由湍流热通量尤其是潜热通量的趋势分布所支配,这与印度洋海域的结果一致。
[Abstract]:The air sea heat fluxes of atmospheric and oceanic circulation, Air Sea freshwater circulation plays a vital role. To study the physical mechanism and the changes of morphology for dynamic simulation and numerical understanding of the atmosphere has a very significance of the core process. This dissertation use ocean observation data, reanalysis data and climate model output data research including sea surface evaporation, latent heat flux, net heat flux and flux changes, trends and climate effects, and to explore the possible sources of the corresponding model error. According to the Woods Holzer Institute of Oceanography (WHOI) OAFlux data objective analysis of air sea flux project, global sea surface evaporation occurred in 1999-2000 years the interdecadal change trend, from the rise to.2000-2013 year downward trend in sea surface evaporation reduction in addition to non local effects from the offshore wind speed caused by weak, Lord If by sea surface temperature warming led to the reduction of sea surface specific humidity slows and saturated humidity of air to accelerate the growth of offshore surface caused. First, weakening offshore surface wind speed due to the global sea surface evaporation reduction through direct and indirect effects. In addition, the total atmospheric column water vapor content over the ocean and over the ocean the precipitation reduction in remaining in the water in the air increases, directly led to the increase in offshore surface humidity of air. The sea surface evaporation rate trend has weakened the overall spatial form, the overall precipitation on the ocean circulation and reduce the strongest Walker rising branch area showed obvious increasing trend of precipitation. The total precipitation on land increased and the total amount of evaporation increases. Whether on land or sea, the relationship between evaporation and precipitation are more close to the local correlation. This correlation makes us more inclined to think that the ocean water Steam air sea interface itself is water vapour balance is disturbed. Using the OAFlux dataset to evaluate the simulation results of tropical and sub tropical Pacific sea surface latent heat flux of 14 coupled climate model CMIP5 plan, and analyzes the sources of possible errors of the simulation. The results show that the latent heat flux model to simulate the spatial distribution and morphology observation, good results, but the simulation value is generally higher than observed value high up to 20-30 Wm-2. model is better in the air humidity and sea surface temperature climatology simulation, simulation of offshore wind speed deviation of climatological probably is the source of model deviation on the latent heat flux. The use of climatological analysis methods showed that EOF mode and the consistency model deviates from the mean state space is mainly caused by the differences between the simulation model of sea surface temperature. For the seasonal variation of latent heat flux climatology, simulation The results agree well with observed data, the main simulation deviation and offshore simulation deviation of wind seasonal changes. While the simulation evaluation of latent heat flux linear trend shows that the pattern of sea surface temperature and sea surface specific humidity simulation on the whole is too large, resulting in the sea air humidity Q trend of latent heat the flux feedback effect will be very weak, and the simulation model of offshore wind speed trend of the small lot, two factors may lead to extremely low values of common simulation model of latent heat flux increased. By the use of Chinese Department of the Institute of Atmospheric Physics, atmospheric sciences and geophysical fluid dynamics numerical simulation (State Key Laboratory of LASG/IAP d) global ocean atmosphere land system model FGOALS-g2 and FGOALS-s2, analyzes their simulation results of air sea heat balance in tropical India ocean. The sea, two models have underestimated FGOALS about 30 W m-2 on the net surface heat flux of the basin scale, the latent heat flux significantly overvalued is believed to be the main source of net thermal error. The two mode reflects the error distribution of India Ocean Dipole shape. Through the analysis of interannual sequence simulation error on the net the heat flux in FGOALS-s2 is mainly from the net shortwave radiation and downward to the latent heat flux of seven FGOALS-g2 and overestimated, simulation error is mainly from the turbulent heat flux overestimated. In observational data, the net heat flux in tropical India ocean sea center domain greatly reduce the trend, and this feature in the model failed to reflect. The latent heat flux dominates the trend of net heat flux trend, so the simulation error model for the net heat flux trend comes mainly from the simulation error of the latent heat flux, latent heat flux is underestimated by With the trend. In the Pacific and the Atlantic ocean, there are also underestimate the overall basin scale of net heat flux. The feature region (western boundary current such as the Kuroshio, the Gulf of Mexico and the Atlantic: equatorial Pacific cold tongue) separately analysis can know, although the value is too large in the quasi extreme area model, but the simulation value decline obviously, due to the high value area is relatively small. The distribution of each component model are not the trend and OAFlux showed a trend of distribution pattern matching, but the net heat flux distribution trend is mainly composed of the turbulent heat flux especially dominated the distribution tendency of latent heat flux, the India ocean and the consistency of the results.

【学位授予单位】：中国科学技术大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：P732.6

【相似文献】