Observations from geostationary satellites can provide spatially continuous coverage at continental scales with high spatial and temporal resolution. Because of this, they are commonly used to complement ground-based precipitation measurements, whose coverage is often more limited. We present Hydronn, a neural-network-based, near-real-time precipitation retrieval for Brazil based on visible and infrared (Vis–IR) observations from the Advanced Baseline Imager (ABI) on the Geostationary Operational Environmental Satellite 16 (GOES-16). The retrieval, which employs a convolutional neural network to perform Bayesian precipitation retrievals, was developed with the aims of (1) leveraging the full potential of latest-generation geostationary observations and (2) providing probabilistic precipitation estimates with well-calibrated uncertainties. The retrieval is trained using more than 3 years of collocations with combined radar and radiometer retrievals from the Global Precipitation Measurement (GPM) core observatory over South America. The accuracy of instantaneous precipitation estimates is assessed using a separate year of GPM combined retrievals and compared to retrievals from passive microwave (PMW) sensors and HYDRO, the Vis–IR retrieval that is currently in operational use at the Brazilian Institute for Space Research. Using all available channels of the ABI, Hydronn achieves accuracy close to that of state-of-the-art PMW precipitation retrievals in both precipitation estimation and detection despite the lower information content of the Vis–IR observations. Hourly, daily, and monthly precipitation accumulations are evaluated against gauge measurements for June and December 2020 and compared to HYDRO, the Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks (PERSIANN) Cloud Classification System (CCS), and the Integrated Multi-satellitE Retrievals for GPM (IMERG). Compared to HYDRO, Hydronn reduces the mean absolute error for hourly accumulations by 21 % (22 %) compared to HYDRO by 44 % (41 %) for the mean squared error (MSE) and increases the correlation by 138 % (312 %) for June (December) 2020. Compared to IMERG, the improvements correspond to 16 % (14 %), 12 % (12 %), and 20 % (56 %), respectively. Furthermore, we show that the probabilistic retrieval is well calibrated against gauge measurements when differences in the distributions of the training data and the gauge measurements are accounted for. Hydronn has the potential to significantly improve near-real-time precipitation retrievals over Brazil. Furthermore, our results show that precipitation retrievals based on convolutional neural networks (CNNs) that leverage the full range of available observations from latest-generation geostationary satellites can provide instantaneous precipitation estimates with accuracy close to that of state-of-the-art PMW retrievals. The high temporal resolution of the geostationary observation allows Hydronn to provide more accurate precipitation accumulations than any of the tested conventional precipitation retrievals. Hydronn thus clearly shows the potential of deep-learning-based precipitation retrievals to improve precipitation estimates from currently available satellite imagery.

Different algorithms for forecasting and tracking meteorological systems have been developed over the years. Many of them are used to study cloud propagation, precipitation and lightning for nowcasting. Therefore, it is necessary to define carefully the parameters (e.g., intensity thresholds and minimum size) that impact tracking of these variables. In order to represent the physical aspects of rain propagation over the Amazon region, several methods of correction and displacement detection were studied. Different parameters were used to validate the methods based on the extrapolated rain cell. A probability detection of 78.4% and 68.6% was achieved for 20 dBZ thresholds during the wet and dry season, respectively. However, the POD decreases for higher reflectivity thresholds. The results for corrections by Inner Nuclei showed that embedded convection can dictate the propagation of rain cells. Split and merge corrections performed well; however, they applied only to a few cases. Corrections performed better for precipitating systems with larger areas and longer duration. The correction methods showed similar skills for both seasons. Which shows that they are able to monitor rain cells throughout the year. The automated combination of different methods for the 20 dBZ threshold proved to be the best choice for tracking rainfall in the Amazon region.

This study aims to evaluate wind speed and significant wave height data from the Copernicus Marine Environment Monitoring Service (CMEMS) reanalysis using buoy measurements for offshore energy application off the east coast of Brazil. Such analysis has become important, since reanalysis datasets can be fundamental tools in identifying regions with wind energy potential that are suitable for the installation of offshore farms. Two sets of reanalysis were used: wind speed (with spatial resolution of 0.25° and temporal resolution of 6 h) and significant wave height (with spatial resolution of 0.2° and temporal resolution of 3 h). For validation, seven MetOcean buoys were selected. In the statistical validation, Pearson’s correlation, coefficient of determination (R²), slope of the straight line, root mean square error (RMSE), mean square error (MSE), probability density function (PDF), mean and standard deviation were calculated. In the evaluation of offshore wind energy resources, the calculation of energy density was performed. The results showed correlations above 0.70 for wind speed and above 0.91 for significant wave height, and additionally, the RMSE values showed maximums of 2.31 m/s for wind speed and 0.28 cm for significant wave height. In the PDF comparison of buoy data and reanalysis, similarities were observed, mainly in the PDF parameters. The energy density presented values consistent with other studies (352–461 W/m²). The results show that the reanalysis data can be applicable in studies focusing on offshore wind potential.

This study evaluates the performance of two deep learning architectures, U-Net and convolutional long short-term memory (ConvLSTM), for predicting cloud-to-ground (CG) lightning probability in the central Amazon basin. Utilizing data from S-band weather radar and very low-frequency (VLF) lightning detection networks, we explored nowcasting capabilities at 12- and 6-min intervals. Various data configurations were tested, specifically examining the integration of volumetric radar reflectivity (dB Z ) and echo-top 35-dBZ (ET35). The evaluation encompassed objective skill metrics and spatial error analyses, accounting for sensor uncertainties. Results reveal a critical divergence in multimodal data assimilation. While the U-Net struggled with volumetric inputs, the ConvLSTM architecture successfully leveraged dB Z data, achieving the highest critical success index (CSI) of 0.546 and an effective spatial scale of 6 km, notably outperforming persistence baselines. Conversely, U-Net performed best using only quality-controlled lightning data or lightweight ET35 inputs. Furthermore, 6-min experiments demonstrated that extending the temporal lookback window notably enhances the skill of recurrent models. Spatial error analyses indicated well-collocated clusters with average centroid errors below 6 km at optimal probability thresholds. These findings suggest that while U-Net provides a robust baseline, ConvLSTM is superior for integrating high-dimensional radar variables in very short-term lightning nowcasting. Significance Statement: This exploratory study evaluates and compares emerging computational approaches for meteorological forecasting, specifically examining the effectiveness of U-Net and convolutional long short-term memory (ConvLSTM) architectures in predicting cloud-to-ground lightning probabilities in the central Amazon basin. By integrating data from weather radar and lightning detection networks, the research investigates methodological solutions for very short-term lightning prediction (6–12 min) in one of Earth’s most climatically complex regions. The comparative analysis reveals a critical distinction in data assimilation capabilities. While the ConvLSTM successfully leverages volumetric radar reflectivity to achieve effective spatial scales of 6 km, the U-Net performs optimally with simplified inputs like echo tops. This work identifies robust configurations that surpass traditional persistence baselines, laying the foundation for next-generation early warning systems capable of mitigating the socioeconomic impacts and risks of human fatalities caused by lightning.

The Amazon Basin is the largest rainforest in the world, and studying the rainfall in this region is crucial for understanding the functioning of the entire rainforest ecosystem and its role in regulating the regional and global climate. This work is part of the application of complex networks, which refer to a network modeled by graphs and are characterized by their high versatility, as well as the extraction of key information from the system under study. The main objective of this article is to examine the precipitation system in the Amazon basin during the austral summer. The networks are defined by nodes and connections, where each node represents a precipitation time series, while the connections can be represented by different similarity functions. For this study, three rainfall networks were created, which differ based on the correlation function used (Pearson, Spearman, and Kendall). By comparing these networks, we can identify the most effective method for analyzing the data and gain a better understanding of rainfall’s spatial structure, thereby enhancing our knowledge of its impact on different Amazon basin regions. The results reveal the presence of three important regions in the Amazon basin. Two areas were identified in the northeast and northwest, showing incursions of warm and humid winds from the oceans and favoring the occurrence of large mesoscale systems, such as squall lines. Additionally, the eastern part of the central Andes may indicate an outflow region from the basin with winds directed toward subtropical latitudes. The networks showed a high level of activity and participation in the center of the Amazon basin and east of the Andes. Regarding information transmission, the betweenness centrality identified the main pathways within a basin, and some of these are directly related to certain rivers, such as the Amazon, Purus, and Madeira. Indicating the relationship between rainfall and the presence of water bodies. Finally, it suggests that the Spearman and Kendall correlation produced the most promising results. Although they showed similar spatial patterns, the major difference was found in the identification of communities, this is due to the meridional differences in the network’s response. Overall, these findings highlight the importance of carefully selecting appropriate techniques and methods when analyzing complex networks.