Weather forecasting is predicting the state of the atmosphere for a given location and time, and it depends on many meteorological parameters, which is a challenging task. In recent years, the use of artificial intelligence (AI) technology has grown leaps and bounds in understating data-driven Earth systems. The main objective of the present paper is to predict the zonal wind using machine learning (ML) techniques. Traditionally, numerical weather prediction (NWP) models are used to predict the weather, but they have their constraints and limitations. Rapid advances in AI- and ML-based models could learn quickly to predict high-impact weather events directly from the observed data. Their computations are also faster than the conventional models. In the present study, we have exploited Extreme Gradient Boosting (XGBoost), and Random Forest Regression approaches on a long time series of data to predict zonal winds at different pressure levels. The data of monthly mean values of zonal wind on different pressure levels for a period of 20 years from 2001 to 2020 over Thiruvananthapuram (8.5° N; 76.9° E), is used. The model results are validated with the observations from the unseen data. The models have proven to provide satisfactory results with XGBoost and Random Forest Regression with a mean absolute error of 1.84 and 2.62, respectively. Based on the results obtained, the ML-based models can be sufficiently reliable in predicting the vertical structure of zonal winds on monthly scales. Evaluated the performance of these models and also indicates the XGBoost is the best fit model compared to Random Forest Regression for predicting the vertical structure of zonal wind throughout the year over a given region. Thus the significance of the present study lies in the ability to predict the zonal wind speed at different altitude levels, which is an essential factor for sustainable wind power generation along the western coast of India. Also useful for future climate projections are the Indian summer monsoon low-level jet and tropical-easterly jet at 850 and 200 hPa levels.

天气预报是预测给定地点和时间的大气状态，它取决于许多气象参数，这是一项具有挑战性的任务。近年来，人工智能 (AI) 技术的使用在理解数据驱动的地球系统方面取得了突飞猛进的发展。本文的主要目标是使用机器学习（ML）技术预测纬向风。传统上，数值天气预报（NWP）模型用于预测天气，但它们有其局限性。基于人工智能和机器学习的模型的快速发展可以快速学习，直接根据观测数据预测高影响的天气事件。他们的计算也比传统模型更快。在本研究中，我们利用极端梯度提升（XGBoost）和随机森林回归方法对长期数据进行预测，以预测不同压力水平下的纬向风。使用2001年至2020年20年间特里凡得琅（北纬8.5°；东经76.9°）20年间不同气压水平纬向风的月平均值数据。模型结果通过未见数据的观察得到验证。事实证明，这些模型可以通过 XGBoost 和随机森林回归提供令人满意的结果，平均绝对误差分别为 1.84 和 2.62。根据获得的结果，基于机器学习的模型可以足够可靠地预测月尺度上纬向风的垂直结构。评估了这些模型的性能，并表明与随机森林回归相比，XGBoost 是预测给定区域全年纬向风垂直结构的最佳拟合模型。因此，本研究的意义在于能够预测不同海拔高度的纬向风速，这是印度西海岸可持续风力发电的重要因素。对于未来气候预测也有用的是印度夏季季风低空急流和 850 和 200 hPa 水平的热带-东风急流。