This study investigates the impact of sea surface temperature (SST) describing the lower boundary of Weather Research and Forecasting (WRF) model on the extreme weather events that occurred on the Mediterranean (MED) and the Eastern Black Sea (EBS) regions of Turkey. For each region, one extreme event case is selected and characterized as a summer convective system in EBS and as a winter synoptic system in MED region. The SST impact on the WRF model forecasts of these events is accomplished in two ways. First, the SST analysis is conducted by making a 10-days simulation for each event with and without activating the SST options (update and skin options) available in the model configuration. In these runs, the selected initial and boundary condition datasets, Global Forecasting System (GFS) for the MED region and the ERA5 for the EBS regions, provide corresponding internal SST data. Second, the prediction performance of the model is evaluated among the simulations of non-external, time-varying (GFS and ERA5 SST) and external, time–varying, high-resolution SST products (The Group for High Resolution Sea Surface Temperature; GHRSST, Medspiration, and NCEP) for each event. The results show that the WRF model simulations, even for short-term event predictions, are highly sensitive to time-variant SST options. The overestimation and high spreading feature of maximum and total precipitation are realistically reduced by time-variant SST products. The Medspiration and the NCEP sources for the MED region and the GHRSST and the Medspiration sources for the EBS region yielded warmer SSTs among the simulations. By establishing the amplified air-sea interactions, these sources provide more appropriate peak precipitation distributions in both regions. Furthermore, based on this study, the effect of using high-resolution SST (GHRSST and Medspiration) is more prominent in predicting convective event developed over complex topography of the EBS while the resolution effect is not much critical for synoptic system simulation even under the complex topography in MED. Lastly, the highest mean correlation of conditional hourly rain is calculated as 0.65 with the Medspiration product on MED and 0.48 with the GHRSST product on the EBS region.

本研究调查了描述天气研究和预报 (WRF) 模型下边界的海面温度 (SST) 对发生在土耳其地中海 (MED) 和黑海东部 (EBS) 地区的极端天气事件的影响。对于每个区域，选择一个极端事件案例并将其表征为 EBS 中的夏季对流系统和 MED 区域中的冬季天气系统。SST 对这些事件的 WRF 模型预测的影响是通过两种方式实现的。首先，在激活和不激活模型配置中可用的 SST 选项（更新和皮肤选项）的情况下，通过对每个事件进行 10 天的模拟来进行 SST 分析。在这些运行中，选定的初始和边界条件数据集、MED 区域的全球预测系统 (GFS) 和 EBS 区域的 ERA5、提供相应的内部SST数据。其次，在非外部时变（GFS 和 ERA5 SST）和外部时变高分辨率 SST 产品（The Group for High Resolution Sea Surface Temperature；GHRSST）的模拟中评估模型的预测性能、Medspiration 和 NCEP）用于每个事件。结果表明，即使是短期事件预测，WRF 模型模拟也对时变 SST 选项高度敏感。时变 SST 产品实际上减少了最大和总降水量的高估和高传播特征。MED 区域的 Medspiration 和 NCEP 源以及 EBS 区域的 GHRSST 和 Medspiration 源在模拟中产生了更暖的 SST。通过建立放大的海气相互作用，这些来源提供了两个地区更合适的峰值降水分布。此外，基于这项研究，使用高分辨率 SST（GHRSST 和 Medspiration）在预测 EBS 复杂地形上发展的对流事件方面的效果更加突出，而分辨率效果对于天气系统模拟即使在复杂的环境下也不是很关键。 MED 中的地形。最后，条件每小时降雨的最高平均相关性计算为 MED 的 Medspiration 产品为 0.65，EBS 区域的 GHRSST 产品为 0.48。使用高分辨率 SST（GHRSST 和 Medspiration）在预测 EBS 复杂地形上发展的对流事件方面的效果更为突出，而分辨率效果对于天气系统模拟甚至在 MED 复杂地形下也不是很重要。最后，条件每小时降雨的最高平均相关性计算为 MED 的 Medspiration 产品为 0.65，EBS 区域的 GHRSST 产品为 0.48。使用高分辨率 SST（GHRSST 和 Medspiration）在预测 EBS 复杂地形上发展的对流事件方面的效果更为突出，而分辨率效果对于天气系统模拟甚至在 MED 复杂地形下也不是很重要。最后，条件每小时降雨的最高平均相关性计算为 MED 的 Medspiration 产品为 0.65，EBS 区域的 GHRSST 产品为 0.48。