Fog prediction using artificial intelligence: A case study in Wamena Airport

Abstract

Fog is one of the atmospheric phenomena that affect airport operations. It can reduce visibility which impacts flight operations (taxiing, take-off, landing). Therefore, fog prediction is needed to support flight safety. The biggest challenge in making weather predictions is the chaotic and complicated process of the atmosphere. This research tries to use artificial intelligence (AI) to predict fog events at Wamena Airport. Design of model prediction using hourly synoptic data set from January 2015 till May 2018. Variables input such as dry ball temperature, wet ball temperature, dew point, relative humidity, cloud cover, wind direction, wind speed, visibility, and present weather for the past six hours ago are used to predict fog or no fog events. We performed a grid search parameter tuning on five algorithms such as Distributed Random Forest (DFR), Deep Learning (DL), Gradient Boosting Machine (GBM), Generalized Linear Model (GLM), and Extreme Randomized Tree (XRT). The best model is obtained from the ensemble model Stacked Ensemble (SE) with an accuracy of above 90% for the fog forecast from one to three hours later.

A review on factors influencing fog formation, classification, forecasting, detection and impacts

“…A similar kind of practice was adopted for predicting visibility for nowcasting fog visibility at Kolkata airport; the study considered pollutants and meteorological data spanning eleven years where the predictors were decided with the help of decision tree algorithm and then used as input in ANN model using the MLP (Multilayer Perceptron) backpropagation learning method for visibility prediction (Dutta and Chaudhuri 2015 ), in Academia da Força Aérea (Colabone et al 2015 ) in International Airport of MACEIÓ and Brazil (Costa et al 2006 ). Another case study of Wamena Airport utilised a bunch of algorithms including Xtreme Randomized Tree (XRT), Generalized Linear Model (GLM), Gradient Boosting Machine (GBM), Deep Learning (DP, Distributed Random Forest (DRF), and stacked Ensemble comprising of all the model’s aforementioned which gave the best result (Dewi and Harsa 2020 ). Performance of various hybrid neural networks in predicting radiation fog events from the fog from the meteorological data collected from the Valladolid airport was compared, the Multiobjective Evolutionary training algorithms performed best, and the importance of each meteorological parameter considered was also presented in terms of prediction (Durán-Rosal et al 2018 ).…”

Improving classification‐based nowcasting of radiation fog with machine learning based on filtered and preprocessed temporal data

“…XGB is used in various fields to achieve state‐of‐the‐art results (Chen & Guestrin, 2016), including weather forecasting (Cai et al ., 2020; Fan et al ., 2021; Kumari & Toshniwal, 2021). It has been shown to outperform other tree‐based models, such as random forest, and can achieve similar forecasting accuracy to neural networks (Dewi et al ., 2020; Kumari & Toshniwal, 2021; Sheridan et al ., 2016). It is robust to smaller datasets (Luckner et al ., 2017) and requires much less computational power (Li et al ., 2022), making large‐scale applicability for fog forecasting realistic.…”

Sofia Airport Visibility Estimation with Two Machine-Learning Techniques

“…They applied an ensemble procedure creating 100 members for each model, which effectively improved the performance in terms of MAE and RMSE compared to individual algorithms. Dewi et al [22] also evaluated five different ML-based algorithms for visibility predictions for Wamena Airport in Indonesia. They additionally use the stacked ensemble model, which combines all the individual models and provides the best performance for every lead time prediction in the study.…”