Numerical Assessment of the Practical and Intrinsic Predictability of Warm-Season Convection Initiation Using Mesoscale Predictability Experiment (MPEX) Data

深层湿对流的形成，或对流的启动，对它所处的大气状态非常敏感。因此，准确地预测对流起始事件的时间和位置对风暴尺度的数值模拟来说是一个巨大的挑战。这项研究的最终目标是提高预测对流启动的能力，从而减少往往伴随着强烈对流事件的财产和生命损失。为此，本研究通过评估中尺度可预报性试验(MPEX)在对流前环境中获得的定向天气到中α尺度观测对允许对流的部分对流起始事件的真实数据数值模拟的影响，检验了对流起始的实际和内在的可预报性。研究预对流环境的更密集采样表示的可预测性是如何演变的，这将增加对对流启动时天气到中尺度控制的基本理解。智力上的优点：利用两套集合卡尔曼滤波-初始化的、允许对流的真实数据数值模拟，一组包括MPEX观测，另一组不包括MPEX观测，这项研究检验了这样的假设，即对预对流大气状态的更密集采样表示足以改善MPEX采样的事件范围内原始对流启动时间和位置的实际可预测性。概率空间验证方法将被用来检验这一假设。利用“完美模式”和“完美观测”方法研究了三个MPEX密集观测期的初始对流起始事件，每个周期具有不同的盛行天气尺度流型，考察了亚网格尺度行星边界层过程数值表示中初始条件的不确定性和可变性对对流起始的内在可预报性的影响。在这样做的过程中，这项研究将通过提供迫切需要的洞察，了解当前观测限制对对流启动的可预测性施加的限制，从而扩大理解，对流启动是一种内在的多尺度、非线性物理过程。它将进一步阐明天气尺度和中尺度α尺度对对流启动的可预测性的控制，并确定影响其内在可预测性的更大尺度吸引子的存在。广泛影响：深层潮湿对流通常对财产和生命都造成重大影响。通过开发更准确、更长期的深层潮湿对流预报来缓解这些影响的能力取决于我们更好地预测其启动的能力。这项研究提供了对对流启动的基本洞察，这将使我们预测深层潮湿对流的时间、位置和发生的能力有所进步。这些进展有望减少每年发生的深度、潮湿对流和相关现象造成的重大财产和生命损失。关于可预测性、概率和不确定性的交叉研究主题将通过一个本科生、非专业荣誉研讨会的发展而被传达给非专业人员，该研讨会的标题是“理解和交流大气科学中的概率和不确定性”。参与研究的研究生将被指导其研究的内在社会意义，提供机会向不同的非专业受众交流研究成果，并鼓励他们接受将物理科学和社会科学相结合的培训。

The formation of deep, moist convection, or convection initiation, is highly sensitive to the atmospheric state in which it occurs. Consequently, accurately predicting the timing and location of convection initiation events poses a formidable challenge for storm-scale numerical simulations. The ultimate goal of this research is to improve the ability to predict convection initiation and, subsequently, mitigate the loss of property and life that often accompany intense convective events. To that end, this study examines the practical and intrinsic predictability of convection initiation by assessing the impact of targeted synoptic- to meso-alpha-scale observations obtained in the pre-convective environment by the Mesoscale Predictability Experiment (MPEX) upon convection-permitting real-data numerical simulations of selected convection initiation events. Investigating how predictability evolves in response to a more intensively-sampled representation of the pre-convective environment will increase fundamental understanding regarding the synoptic- to meso-alpha-scale controls upon convection initiation.Intellectual Merit: Utilizing two ensembles of Ensemble Kalman filter-initialized, convection-permitting real-data numerical simulations, one incorporating MPEX observations and one not, this research tests the hypothesis that a more intensively-sampled representation of the pre-convective atmospheric state is sufficient to improve the practical predictability of pristine convection initiation timing and location over the range of events sampled by MPEX. Probabilistic, temporally-binned spatial verification methods will be utilized to test this hypothesis. Utilizing "perfect model" and "perfect observations" approaches applied to the study of the initial convection initiation event from three MPEX intensive observation periods, each characterized by a different prevailing synoptic-scale flow pattern, the influences of initial condition uncertainty and variability in the numerical representation of sub-grid-scale planetary boundary layer processes upon the intrinsic predictability of convection initiation are examined. In so doing, this research will expand understanding through the provision of critically-needed insight into the limits imposed by current observational constraints upon the predictability of convection initiation, an inherently multi-scale, non-linear physical process. It will further illuminate the controls upon the predictability of convection initiation exerted by the synoptic- and meso-alpha-scales and identify the presence of larger-scale attractors that influence its intrinsic predictability.Broader Impacts: Deep, moist convection routinely poses significant impacts to both property and life. The ability to mitigate these impacts through the development of more accurate, longer-lead forecasts of deep, moist convection hinges upon our ability to better predict its initiation. Basic insight into convection initiation provided by the research will lead to advances in our ability to predict the timing, location, and occurrence of deep, moist convection. Such advances offer the promise of reducing the substantial losses of property and life due to deep, moist convection and associated phenomena incurred annually. The cross-cutting research themes of predictability, probability, and uncertainty will be communicated to non-specialists through the development of an undergraduate, non-majors Honors seminar titled "Understanding and Communicating Probability and Uncertainty in the Atmospheric Sciences." Graduate students involved with the research will be mentored as to the inherent societal significance of their research, afforded opportunities to communicate research findings to diverse, non-specialist audiences, and encouraged to acquire training in integrating the physical and social sciences.