Abstract Meteorological processes such as tides influence ionospheric variability through vertical coupling. For the first time, we have used data from Communication Navigation Outage and Forecasting System (C/NOFS) satellite from 2008–2015 to develop a Neural Network (NN) vertical E × B drift model over the African region, with inclusion of a proxy of tides as one of the inputs together with other physical and geophysical inputs. Two models (with and without tidal proxy input) were developed for both East and West African sectors. To derive the tidal proxy, we first calculate the 60-day running means per year which were subtracted from the actual vertical E × B drift measurements to obtain a set of residuals. The purpose of the subtraction was to remove long-term trends in vertical E × B drift that could potentially alias into tides. A Fast Fourier Transform (FFT) was then performed on the residuals per day to obtain amplitudes and phases which were used as tidal proxy representation in vertical E × B drift modelling. The developed model with tides’ proxy input showed an improvement of 22.2% and 16.7% over the East ( 38 . 8 o E ) and West ( 9 . 2 o W ) African sectors respectively. In most cases, the developed models were able to capture the diurnal patterns of vertical E × B drift including the expected pre-reversal enhancement. The performance of the two models during geomagnetically quiet days and storm periods was comparable but not similar.

中文翻译：

---

努力研究潮汐对非洲地区垂直 E×B 漂移纵向变化的影响

摘要 潮汐等气象过程通过垂直耦合影响电离层变率。我们首次使用来自 2008 年至 2015 年通信导航中断和预测系统 (C/NOFS) 卫星的数据来开发非洲地区的神经网络 (NN) 垂直 E × B 漂移模型，其中包含一个代理潮汐与其他物理和地球物理输入一起作为输入之一。为东非和西非部门开发了两个模型（有和没有潮汐代理输入）。为了导出潮汐代理，我们首先计算每年 60 天的运行平均值，将其从实际垂直 E × B 漂移测量值中减去以获得一组残差。减法的目的是消除可能混入潮汐的垂直 E × B 漂移的长期趋势。然后对每天的残差进行快速傅立叶变换 (FFT) 以获得幅度和相位，这些幅度和相位在垂直 E × B 漂移建模中用作潮汐代理表示。具有潮汐代理输入的开发模型显示，在东部 (38 . 8 o E) 和西部 (9 . 2 o W) 非洲部门分别提高了 22.2% 和 16.7%。在大多数情况下，开发的模型能够捕获垂直 E × B 漂移的昼夜模式，包括预期的反转前增强。两个模型在地磁平静日和风暴期间的性能相当但不相似。与潮汐代理输入的开发模型相比，东部（东经 38 度 8 度）和西非（东经 9 度 2 度）分别提高了 22.2％和 16.7％。在大多数情况下，开发的模型能够捕获垂直 E × B 漂移的昼夜模式，包括预期的反转前增强。两个模型在地磁平静日和风暴期间的性能相当但不相似。具有潮汐代理输入的开发模型显示，在东部 (38 . 8 o E) 和西部 (9 . 2 o W) 非洲部门分别提高了 22.2% 和 16.7%。在大多数情况下，开发的模型能够捕获垂直 E × B 漂移的昼夜模式，包括预期的反转前增强。两个模型在地磁平静日和风暴期间的性能相当但不相似。