Changes in Arctic sea ice affect atmospheric circulation, ocean current, and polar ecosystems. There have been unprecedented decreases in the amount of Arctic sea ice due to global warming. In this study, a novel 1-month sea ice concentration (SIC) prediction model is proposed, with eight predictors using a deep-learning approach, convolutional neural networks (CNNs). This monthly SIC prediction model based on CNNs is shown to perform better predictions (mean absolute error – MAE – of 2.28 %, anomaly correlation coefficient – ACC – of 0.98, root-mean-square error – RMSE – of 5.76 %, normalized RMSE – nRMSE – of 16.15 %, and NSE – Nash–Sutcliffe efficiency – of 0.97) than a random-forest-based (RF-based) model (MAE of 2.45 %, ACC of 0.98, RMSE of 6.61 %, nRMSE of 18.64 %, and NSE of 0.96) and the persistence model based on the monthly trend (MAE of 4.31 %, ACC of 0.95, RMSE of 10.54 %, nRMSE of 29.17 %, and NSE of 0.89) through hindcast validations. The spatio-temporal analysis also confirmed the superiority of the CNN model. The CNN model showed good SIC prediction results in extreme cases that recorded unforeseen sea ice plummets in 2007 and 2012 with RMSEs of less than 5.0 %. This study also examined the importance of the input variables through a sensitivity analysis. In both the CNN and RF models, the variables of past SICs were identified as the most sensitive factor in predicting SICs. For both models, the SIC-related variables generally contributed more to predict SICs over ice-covered areas, while other meteorological and oceanographic variables were more sensitive to the prediction of SICs in marginal ice zones. The proposed 1-month SIC prediction model provides valuable information which can be used in various applications, such as Arctic shipping-route planning, management of the fishing industry, and long-term sea ice forecasting and dynamics.

Total precipitable water (TPW), a column of water vapor content in the atmosphere, provides information on the spatial distribution of moisture. The high-resolution TPW, together with atmospheric stability indices such as convective available potential energy (CAPE), is an effective indicator of severe weather phenomena in the pre-convective atmospheric condition. With the advent of high performing imaging instrument onboard geostationary satellites such as Advanced Himawari Imager (AHI) onboard Himawari-8 of Japan and Advanced Meteorological Imager (AMI) onboard GeoKompsat-2A of Korea, it is expected that unprecedented spatiotemporal resolution data (e.g., AMI plans to provide 2 km resolution data at every 2 min over the northeast part of East Asia) will be provided. To derive TPW from such high-resolution data in a timely fashion, an efficient algorithm is highly required. Here, machine learning approaches—random forest (RF), extreme gradient boosting (XGB), and deep neural network (DNN)—are assessed for the TPW retrieved from AHI over the clear sky in Northeast Asia area. For the training dataset, the nine infrared brightness temperatures (BT) of AHI (BT8 to 16 centered at 6.2, 6.9, 7.3, 8.6, 9.6, 10.4, 11.2, 12.4, and 13.3 $\mathsf{\mu }\mathrm{m}$ , respectively), six dual channel differences and observation conditions such as time, latitude, longitude, and satellite zenith angle for two years (September 2016 to August 2018) are used. The corresponding TPW is prepared by integrating the water vapor profiles from InterimEuropean Centre for Medium-Range Weather Forecasts Re-Analysis data (ERA-Interim). The algorithm performances are assessed using the ERA-Interim and radiosonde observations (RAOB) as the reference data. The results show that the DNN model performs better than RF and XGB with a correlation coefficient of 0.96, a mean bias of 0.90 mm, and a root mean square error (RMSE) of 4.65 mm when compared to the ERA-Interim. Similarly, DNN results in a correlation coefficient of 0.95, a mean bias of 1.25 mm, and an RMSE of 5.03 mm when compared to RAOB. Contributing variables to retrieve the TPW in each model and the spatial and temporal analysis of the retrieved TPW are carefully examined and discussed.

This paper proposes a complete framework of a machine learning-based model that detects convective initiation (CI) from geostationary meteorological satellite data. The suggested framework consists of three main processes: (1) An automated sampling tool; (2) machine learning-based CI detection modelling; (3) repeated model tuning through validation. In this study, the automated sampling tool was able to track the CI objects iteratively, even without ancillary data such as an atmospheric motion vector (AMV). The collected samples were used to train the machine learning model for CI detection. Random forest (RF) was used to classify the CI and non-CI. To enhance the advantages of the machine learning approach, we adopted model tuning to iteratively update the training dataset from each validation result by adding hits and misses to the CI samples, and false alarms and correct negatives to the non-CI samples. Using 12 interest fields from the Himawari-8 Advanced Himawari Imager (AHI) over the Korean Peninsula, this simple and intuitive tuning process increased the overall probability of detection (POD) from 0.79 to 0.82 and decreased the overall false alarm rate (FAR) from 0.46 to 0.37 with around 40 min of the lead-time. Amongst the 12 interest fields, $T_b$ (11.2) µm was identified as the most significant predictor in the RF model, followed by $T_b$ (8.6—11.2) µm, and $T_b$ (6.2–7.3) µm. The effect of model tuning on the CI detection performance was also analyzed using spatiotemporal validation maps. By automatically collecting and updating the machine learning training dataset, the suggested framework is expected to help the maintenance of the CI detection model from an operational perspective.

Recent studies have enhanced the mapping performance of the local climate zone (LCZ), a standard framework for evaluating urban form and function for urban heat island research, through remote sensing (RS) images and deep learning classifiers such as convolutional neural networks (CNNs). The accuracy in the urban-type LCZ (LCZ1-10), however, remains relatively low because RS data cannot provide vertical or horizontal building components in detail. Geographic information system (GIS)-based building datasets can be used as primary sources in LCZ classification, but there is a limit to using them as input data for CNN due to their incompleteness. This study proposes novel methods to classify LCZ using Sentinel 2 images and incomplete building data based on a CNN classifier. We designed three schemes (S1, S2, and a scheme fusion; SF) for mapping 50 m LCZs in two megacities: Berlin and Seoul. S1 used only RS images, and S2 used RS and building components such as area and height (or the number of stories). SF combined two schemes (S1 and S2) based on three conditions, mainly focusing on the confidence level of the CNN classifier. When compared to S1, the overall accuracies for all LCZ classes (OA) and the urban-type LCZ (OA_urb) of SF increased by about 4% and 7–9%, respectively, for the two study areas. This study shows that SF can compensate for the imperfections in the building data, which causes misclassifications in S2. The suggested approach can be excellent guidance to produce a high accuracy LCZ map for cities where building databases can be obtained, even if they are incomplete.

Skillful quantitative precipitation nowcasting (QPN) is important for predicting precipitation in the upcoming few hours and thus avoiding significant socioeconomic damage. Recent QPN studies have actively adopted deep learning (DL) to generate precipitation maps using sequences of ground radar data. Although high skill scores in forecasting precipitation areas of weak intensity (~1 mm/h) have been achieved, the horizontal movement of precipitation areas could not be accurately simulated, exhibiting poor forecasting skills for stronger intensities. For lead times up to 120 min, this study suggests using an improved radar-based QPN model that utilizes a state-of-the-art DL model termed simpler yet better video prediction (SimVP). An independent evaluation using ground radar data in South Korea from June to September 2022 demonstrated that the proposed model outperformed the existing DL models in terms of critical score index (CSI) with a lead time of 120 min (0.46, 0.23, and 0.09 for 1, 5, and 10 mm/h thresholds, respectively). Three case analyses were conducted to reflect various precipitation conditions: heavy rainfall, typhoons, and fast-moving narrow convection events. The proposed SimVP-based QPN model yielded robust performance for all cases, producing a comparable or highest CSI at the lead time of 120 min with a 1 mm/h threshold (0.49, 0.69, and 0.29 for heavy rainfall, typhoon, and narrow convection, respectively). Qualitative evaluation of the model indicated better results in terms of displacement movement and reduced underestimation than other models under the high variability of precipitation patterns of the three cases. A comparison of model complexity among DL-QPN models was conducted, taking into consideration operational applications across various study areas and environments. The proposed approach is expected to provide a new baseline for DL-based QPN, and the improved prediction using the proposed model can lead to reduced socioeconomic damage incurred as a result of short-term intense precipitation.