Articles | Volume 16, issue 20
https://doi.org/10.5194/gmd-16-5895-2023
© Author(s) 2023. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/gmd-16-5895-2023
© Author(s) 2023. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Model evaluation paper
| 
20 Oct 2023
Model evaluation paper |
| 20 Oct 2023
| 20 Oct 2023
Key factors for quantitative precipitation nowcasting using ground weather radar data based on deep learning
Daehyeon Han
Department of Civil, Urban, Earth, and Environmental Engineering, Ulsan National Institute of Science and Technology, Ulsan, 44919, South Korea
Department of Civil, Urban, Earth, and Environmental Engineering, Ulsan National Institute of Science and Technology, Ulsan, 44919, South Korea
Yeji Shin
National Institute of Meteorological Sciences, Korea Meteorological Administration, Jeju-do, 63568, South Korea
Market Intelligence Team, CJ CheilJedang Corporation, Seoul, 04560, South Korea
Juhyun Lee
Department of Civil, Urban, Earth, and Environmental Engineering, Ulsan National Institute of Science and Technology, Ulsan, 44919, South Korea
Related authors
Young Jun Kim, Hyun-Cheol Kim, Daehyeon Han, Sanggyun Lee, and Jungho Im
The Cryosphere, 14, 1083–1104, https://doi.org/10.5194/tc-14-1083-2020, https://doi.org/10.5194/tc-14-1083-2020, 2020
Short summary
Short summary
In this study, we proposed a novel 1-month sea ice concentration (SIC) prediction model with eight predictors using a deep-learning approach, convolutional neural networks (CNNs). The proposed CNN model was evaluated and compared with the two baseline approaches, random-forest and simple-regression models, resulting in better performance. This study also examined SIC predictions for two extreme cases in 2007 and 2012 in detail and the influencing factors through a sensitivity analysis.
Young Jun Kim, Hyun-Cheol Kim, Daehyeon Han, Sanggyun Lee, and Jungho Im
The Cryosphere, 14, 1083–1104, https://doi.org/10.5194/tc-14-1083-2020, https://doi.org/10.5194/tc-14-1083-2020, 2020
Short summary
Short summary
In this study, we proposed a novel 1-month sea ice concentration (SIC) prediction model with eight predictors using a deep-learning approach, convolutional neural networks (CNNs). The proposed CNN model was evaluated and compared with the two baseline approaches, random-forest and simple-regression models, resulting in better performance. This study also examined SIC predictions for two extreme cases in 2007 and 2012 in detail and the influencing factors through a sensitivity analysis.
Seohui Park, Minso Shin, Jungho Im, Chang-Keun Song, Myungje Choi, Jhoon Kim, Seungun Lee, Rokjin Park, Jiyoung Kim, Dong-Won Lee, and Sang-Kyun Kim
Atmos. Chem. Phys., 19, 1097–1113, https://doi.org/10.5194/acp-19-1097-2019, https://doi.org/10.5194/acp-19-1097-2019, 2019
Short summary
Short summary
This study proposed machine-learning-based models to estimate ground-level particulate matter concentrations using satellite observations and numerical model-derived data. Aerosol optical depth was identified as the most significant for estimating ground-level PM concentrations, followed by wind speed and solar radiation. The results show that the proposed models produced better performance than the existing approaches, particularly in improving on the biases of the process-based models.
Sanggyun Lee, Hyun-cheol Kim, and Jungho Im
The Cryosphere, 12, 1665–1679, https://doi.org/10.5194/tc-12-1665-2018, https://doi.org/10.5194/tc-12-1665-2018, 2018
Short summary
Short summary
Arctic sea ice leads play a major role in exchanging heat and momentum between the Arctic atmosphere and ocean. In this study, we propose a novel lead
detection approach based on waveform mixture analysis. The performance of the proposed approach in detecting leads was promising when compared to the
existing methods. The robustness of the proposed approach also lies in the fact that it does not require the rescaling of parameters, as it directly uses L1B waveform data.
Sanggyun Lee, Hyangsun Han, Jungho Im, Eunna Jang, and Myong-In Lee
Atmos. Meas. Tech., 10, 1859–1874, https://doi.org/10.5194/amt-10-1859-2017, https://doi.org/10.5194/amt-10-1859-2017, 2017
Short summary
Short summary
Deterministic and probabilistic CI detection models based on decision trees (DT), random forest (RF), and logistic regression (LR) were developed using Himawari-8 AHI data obtained over the Korean Peninsula. We used a total of 12 interest fields including time trends to develop the models. We identified contributing variables for CI detection. DT showed a higher hit rate, while RF produced a higher critical success index. The mean lead times by the four models were in the range of 20–40 min.
Dongmin Kim, Myong-In Lee, Su-Jong Jeong, Jungho Im, Dong Hyun Cha, and Sanggyun Lee
Biogeosciences Discuss., https://doi.org/10.5194/bg-2016-536, https://doi.org/10.5194/bg-2016-536, 2016
Manuscript not accepted for further review
Short summary
Short summary
This study compares historical simulations of the terrestrial carbon cycle produced by 10 ESMs that participated in the CMIP5. The models show noticeable deficiencies compared to the MODIS data and large differences among the simulations, although the MME mean provides a realistic global mean value and spatial distributions. MME is reflected by the systematic biases of simulated biogeochemical processes which depends on temperature conditions strongly in every plant functional types.
H. M. Park, M. A. Kim, and J. Im
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLI-B7, 325–327, https://doi.org/10.5194/isprs-archives-XLI-B7-325-2016, https://doi.org/10.5194/isprs-archives-XLI-B7-325-2016, 2016
S. Park and J. Im
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLI-B7, 703–704, https://doi.org/10.5194/isprs-archives-XLI-B7-703-2016, https://doi.org/10.5194/isprs-archives-XLI-B7-703-2016, 2016
Related subject area
Atmospheric sciences
New submodel for emissions from Explosive Volcanic ERuptions (EVER v1.1) within the Modular Earth Submodel System (MESSy, version 2.55.1)
Quantifying the oscillatory evolution of simulated boundary-layer cloud fields using Gaussian process regression
Numerical investigations on the modelling of ultrafine particles in SSH-aerosol-v1.3a: size resolution and redistribution
The third Met Office Unified Model–JULES Regional Atmosphere and Land Configuration, RAL3
The sensitivity of aerosol data assimilation to vertical profiles: case study of dust storm assimilation with LOTOS-EUROS v2.2
Knowledge-inspired fusion strategies for the inference of PM2.5 values with a neural network
Tuning the ICON-A 2.6.4 climate model with machine-learning-based emulators and history matching
A novel method for quantifying the contribution of regional transport to PM2.5 in Beijing (2013–2020): combining machine learning with concentration-weighted trajectory analysis
Quantification of CO2 hotspot emissions from OCO-3 SAM CO2 satellite images using deep learning methods
Diagnosis of winter precipitation types using the spectral bin model (version 1DSBM-19M): comparison of five methods using ICE-POP 2018 field experiment data
Improving winter condition simulations in SURFEX-TEB v9.0 with a multi-layer snow model and ice
UA-ICON with the NWP physics package (version ua-icon-2.1): mean state and variability of the middle atmosphere
Integrated Methane Inversion (IMI) 2.0: an improved research and stakeholder tool for monitoring total methane emissions with high resolution worldwide using TROPOMI satellite observations
HTAP3 Fires: towards a multi-model, multi-pollutant study of fire impacts
Using a data-driven statistical model to better evaluate surface turbulent heat fluxes in weather and climate numerical models: a demonstration study
Pochva: a new hydro-thermal process model in soil, snow, and vegetation for application in atmosphere numerical models
ClimKern v1.2: a new Python package and kernel repository for calculating radiative feedbacks
Accounting for effects of coagulation and model uncertainties in particle number concentration estimates based on measurements from sampling lines – a Bayesian inversion approach with SLIC v1.0
Top-down CO emission estimates using TROPOMI CO data in the TM5-4DVAR (r1258) inverse modeling suit
The Multi-Compartment Hg Modeling and Analysis Project (MCHgMAP): mercury modeling to support international environmental policy
Similarity-based analysis of atmospheric organic compounds for machine learning applications
Porting the Meso-NH atmospheric model on different GPU architectures for the next generation of supercomputers (version MESONH-v55-OpenACC)
Estimation of aerosol and cloud radiative heating rate in the tropical stratosphere using a radiative kernel method
Evaluation of dust emission and land surface schemes in predicting a mega Asian dust storm over South Korea using WRF-Chem
Sensitivity studies of a four-dimensional local ensemble transform Kalman filter coupled with WRF-Chem version 3.9.1 for improving particulate matter simulation accuracy
A Bayesian method for predicting background radiation at environmental monitoring stations in local-scale networks
Inclusion of the ECMWF ecRad radiation scheme (v1.5.0) in the MAR (v3.14), regional evaluation for Belgium, and assessment of surface shortwave spectral fluxes at Uccle
Development of a fast radiative transfer model for ground-based microwave radiometers (ARMS-gb v1.0): validation and comparison to RTTOV-gb
Indian Institute of Tropical Meteorology (IITM) High-Resolution Global Forecast Model version 1: an attempt to resolve monsoon prediction deadlock
Cell-tracking-based framework for assessing nowcasting model skill in reproducing growth and decay of convective rainfall
NeuralMie (v1.0): an aerosol optics emulator
A REtrieval Method for optical and physical Aerosol Properties in the stratosphere (REMAPv1)
Simulation performance of planetary boundary layer schemes in WRF v4.3.1 for near-surface wind over the western Sichuan Basin: a single-site assessment
FootNet v1.0: development of a machine learning emulator of atmospheric transport
Updates and evaluation of NOAA's online-coupled air quality model version 7 (AQMv7) within the Unified Forecast System
Quantifying the analysis uncertainty for nowcasting application
Improving the ensemble square root filter (EnSRF) in the Community Inversion Framework: a case study with ICON-ART 2024.01
The MESSy DWARF (based on MESSy v2.55.2)
Generalized local fractions – a method for the calculation of sensitivities to emissions from multiple sources for chemically active species, illustrated using the EMEP MSC-W model (rv5.5)
SanDyPALM v1.0: Static and Dynamic Drivers for the PALM-4U Model to Facilitate Realistic Urban Microclimate Simulations
An enhanced emission module for the PALM model system 23.10 with application for PM10 emission from urban domestic heating
Identifying lightning processes in ERA5 soundings with deep learning
Sensitivity of predicted ultrafine particle size distributions in Europe to different nucleation rate parameterizations using PMCAMx-UF v2.2
Explaining neural networks for detection of tropical cyclones and atmospheric rivers in gridded atmospheric simulation data
Implementation of a dry deposition module (DEPAC v3.11) in a large eddy simulation code (DALES v4.4)
Accurate and fast prediction of radioactive pollution by Kriging coupled with Auto-Associative Models
Mitigating Hail Overforecasting in the 2-Moment Milbrandt-Yau Microphysics Scheme (v2.25.2_beta_04) in WRF (v4.5.1) by Incorporating the Graupel Spongy Wet Growth Process (MY2_GSWG v1.0)
PALACE v1.0: Paranal Airglow Line And Continuum Emission model
Accurate space-based NOx emission estimates with the flux divergence approach require fine-scale model information on local oxidation chemistry and profile shapes
Exploring a high-level programming model for the NWP domain using ECMWF microphysics schemes
Matthias Kohl, Christoph Brühl, Jennifer Schallock, Holger Tost, Patrick Jöckel, Adrian Jost, Steffen Beirle, Michael Höpfner, and Andrea Pozzer
Geosci. Model Dev., 18, 3985–4007, https://doi.org/10.5194/gmd-18-3985-2025, https://doi.org/10.5194/gmd-18-3985-2025, 2025
Short summary
Short summary
SO2 from explosive volcanic eruptions reaching the stratosphere can oxidize and form sulfur aerosols, potentially persisting for several years. We developed a new submodel, Explosive Volcanic ERuptions (EVER), that seamlessly includes stratospheric volcanic SO2 emissions in global numerical simulations based on a novel standard historical model setup, successfully evaluated with satellite observations. Sensitivity studies on the Nabro eruption in 2011 evaluate different emission methods.
Gunho Loren Oh and Philip H. Austin
Geosci. Model Dev., 18, 3921–3940, https://doi.org/10.5194/gmd-18-3921-2025, https://doi.org/10.5194/gmd-18-3921-2025, 2025
Short summary
Short summary
It is difficult to study the behaviour of a cloud field due to internal fluctuations and observational noise. We perform a high-resolution simulation of the boundary-layer cloud field and introduce statistical and numerical techniques, including machine-learning models, to study the evolution of the cloud field, which shows a periodic behaviour. We aim to use the numerical techniques to identify the underlying behaviour within noisy observations.
Oscar Jacquot and Karine Sartelet
Geosci. Model Dev., 18, 3965–3984, https://doi.org/10.5194/gmd-18-3965-2025, https://doi.org/10.5194/gmd-18-3965-2025, 2025
Short summary
Short summary
Modelling the size distribution and the number concentration is important to represent ultrafine particles. A new analytic formulation is presented to compute coagulation partition coefficients, allowing us to lower the numerical diffusion associated with the resolution of aerosol dynamics. The significance of this effect is assessed in a 0D box model and over greater Paris with a chemistry transport model, using different size resolutions of the particle distribution.
Mike Bush, David L. A. Flack, Huw W. Lewis, Sylvia I. Bohnenstengel, Chris J. Short, Charmaine Franklin, Adrian P. Lock, Martin Best, Paul Field, Anne McCabe, Kwinten Van Weverberg, Segolene Berthou, Ian Boutle, Jennifer K. Brooke, Seb Cole, Shaun Cooper, Gareth Dow, John Edwards, Anke Finnenkoetter, Kalli Furtado, Kate Halladay, Kirsty Hanley, Margaret A. Hendry, Adrian Hill, Aravindakshan Jayakumar, Richard W. Jones, Humphrey Lean, Joshua C. K. Lee, Andy Malcolm, Marion Mittermaier, Saji Mohandas, Stuart Moore, Cyril Morcrette, Rachel North, Aurore Porson, Susan Rennie, Nigel Roberts, Belinda Roux, Claudio Sanchez, Chun-Hsu Su, Simon Tucker, Simon Vosper, David Walters, James Warner, Stuart Webster, Mark Weeks, Jonathan Wilkinson, Michael Whitall, Keith D. Williams, and Hugh Zhang
Geosci. Model Dev., 18, 3819–3855, https://doi.org/10.5194/gmd-18-3819-2025, https://doi.org/10.5194/gmd-18-3819-2025, 2025
Short summary
Short summary
RAL configurations define settings for the Unified Model atmosphere and Joint UK Land Environment Simulator. The third version of the Regional Atmosphere and Land (RAL3) science configuration for kilometre- and sub-kilometre-scale modelling represents a major advance compared to previous versions (RAL2) by delivering a common science definition for applications in tropical and mid-latitude regions. RAL3 has more realistic precipitation distributions and an improved representation of clouds and visibility.
Mijie Pang, Jianbing Jin, Ting Yang, Xi Chen, Arjo Segers, Batjargal Buyantogtokh, Yixuan Gu, Jiandong Li, Hai Xiang Lin, Hong Liao, and Wei Han
Geosci. Model Dev., 18, 3781–3798, https://doi.org/10.5194/gmd-18-3781-2025, https://doi.org/10.5194/gmd-18-3781-2025, 2025
Short summary
Short summary
Aerosol data assimilation has gained popularity as it combines the advantages of modelling and observation. However, few studies have addressed the challenges in the prior vertical structure. Different observations are assimilated to examine the sensitivity of assimilation to vertical structure. Results show that assimilation can optimize the dust field in general. However, if the prior introduces an incorrect structure, the assimilation can significantly deteriorate the integrity of the aerosol profile.
Matthieu Dabrowski, José Mennesson, Jérôme Riedi, Chaabane Djeraba, and Pierre Nabat
Geosci. Model Dev., 18, 3707–3733, https://doi.org/10.5194/gmd-18-3707-2025, https://doi.org/10.5194/gmd-18-3707-2025, 2025
Short summary
Short summary
This work focuses on the prediction of aerosol concentration values at the ground level, which are a strong indicator of air quality, using artificial neural networks. A study of different variables and their efficiency as inputs for these models is also proposed and reveals that the best results are obtained when using all of them. Comparison between network architectures and information fusion methods allows for the extraction of knowledge on the most efficient methods in the context of this study.
Pauline Bonnet, Lorenzo Pastori, Mierk Schwabe, Marco Giorgetta, Fernando Iglesias-Suarez, and Veronika Eyring
Geosci. Model Dev., 18, 3681–3706, https://doi.org/10.5194/gmd-18-3681-2025, https://doi.org/10.5194/gmd-18-3681-2025, 2025
Short summary
Short summary
Tuning a climate model means adjusting uncertain parameters in the model to best match observations like the global radiation balance and cloud cover. This is usually done by running many simulations of the model with different settings, which can be time-consuming and relies heavily on expert knowledge. To make this process faster and more objective, we developed a machine learning emulator to create a large ensemble and apply a method called history matching to find the best settings.
Kang Hu, Hong Liao, Dantong Liu, Jianbing Jin, Lei Chen, Siyuan Li, Yangzhou Wu, Changhao Wu, Shitong Zhao, Xiaotong Jiang, Ping Tian, Kai Bi, Ye Wang, and Delong Zhao
Geosci. Model Dev., 18, 3623–3634, https://doi.org/10.5194/gmd-18-3623-2025, https://doi.org/10.5194/gmd-18-3623-2025, 2025
Short summary
Short summary
This study combines machine learning with concentration-weighted trajectory analysis to quantify regional transport PM2.5. From 2013–2020, local emissions dominated Beijing's pollution events. The Air Pollution Prevention and Control Action Plan reduced regional transport pollution, but the eastern region showed the smallest decrease. Beijing should prioritize local emission reduction while considering the east region's contributions in future strategies.
Joffrey Dumont Le Brazidec, Pierre Vanderbecken, Alban Farchi, Grégoire Broquet, Gerrit Kuhlmann, and Marc Bocquet
Geosci. Model Dev., 18, 3607–3622, https://doi.org/10.5194/gmd-18-3607-2025, https://doi.org/10.5194/gmd-18-3607-2025, 2025
Short summary
Short summary
We developed a deep learning method to estimate CO2 emissions from power plants using satellite images. Trained and validated on simulated data, our model accurately predicts emissions despite challenges like cloud cover. When applied to real OCO3 satellite images, the results closely match reported emissions. This study shows that neural networks trained on simulations can effectively analyse real satellite data, offering a new way to monitor CO2 emissions from space.
Wonbae Bang, Jacob T. Carlin, Kwonil Kim, Alexander V. Ryzhkov, Guosheng Liu, and GyuWon Lee
Geosci. Model Dev., 18, 3559–3581, https://doi.org/10.5194/gmd-18-3559-2025, https://doi.org/10.5194/gmd-18-3559-2025, 2025
Short summary
Short summary
Microphysics model-based diagnosis, such as the spectral bin model (SBM), has recently been attempted to diagnose winter precipitation types. In this study, the accuracy of SBM-based precipitation type diagnosis is compared with other traditional methods. SBM has a relatively higher accuracy for dry-snow and wet-snow events, whereas it has lower accuracy for rain events. When the microphysics scheme in the SBM was optimized for the corresponding region, the accuracy for rain events improved.
Gabriel Colas, Valéry Masson, François Bouttier, Ludovic Bouilloud, Laura Pavan, and Virve Karsisto
Geosci. Model Dev., 18, 3453–3472, https://doi.org/10.5194/gmd-18-3453-2025, https://doi.org/10.5194/gmd-18-3453-2025, 2025
Short summary
Short summary
In winter, snow- and ice-covered artificial surfaces are important aspects of the urban climate. They may influence the magnitude of the urban heat island effect, but this is still unclear. In this study, we improved the representation of the snow and ice cover in the Town Energy Balance (TEB) urban climate model. Evaluations have shown that the results are promising for using TEB to study the climate of cold cities.
Markus Kunze, Christoph Zülicke, Tarique A. Siddiqui, Claudia C. Stephan, Yosuke Yamazaki, Claudia Stolle, Sebastian Borchert, and Hauke Schmidt
Geosci. Model Dev., 18, 3359–3385, https://doi.org/10.5194/gmd-18-3359-2025, https://doi.org/10.5194/gmd-18-3359-2025, 2025
Short summary
Short summary
We present the Icosahedral Nonhydrostatic (ICON) general circulation model with an upper-atmospheric extension with the physics package for numerical weather prediction (UA-ICON(NWP)). We optimized the parameters for the gravity wave parameterizations and achieved realistic modeling of the thermal and dynamic states of the mesopause regions. UA-ICON(NWP) now shows a realistic frequency of major sudden stratospheric warmings and well-represented solar tides in temperature.
Lucas A. Estrada, Daniel J. Varon, Melissa Sulprizio, Hannah Nesser, Zichong Chen, Nicholas Balasus, Sarah E. Hancock, Megan He, James D. East, Todd A. Mooring, Alexander Oort Alonso, Joannes D. Maasakkers, Ilse Aben, Sabour Baray, Kevin W. Bowman, John R. Worden, Felipe J. Cardoso-Saldaña, Emily Reidy, and Daniel J. Jacob
Geosci. Model Dev., 18, 3311–3330, https://doi.org/10.5194/gmd-18-3311-2025, https://doi.org/10.5194/gmd-18-3311-2025, 2025
Short summary
Short summary
Reducing emissions of methane, a powerful greenhouse gas, is a top policy concern for mitigating anthropogenic climate change. The Integrated Methane Inversion (IMI) is an advanced, cloud-based software that translates satellite observations into actionable emissions data. Here we present IMI version 2.0 with vastly expanded capabilities. These updates enable a wider range of scientific and stakeholder applications from individual basin to global scales with continuous emissions monitoring.
Cynthia H. Whaley, Tim Butler, Jose A. Adame, Rupal Ambulkar, Steve R. Arnold, Rebecca R. Buchholz, Benjamin Gaubert, Douglas S. Hamilton, Min Huang, Hayley Hung, Johannes W. Kaiser, Jacek W. Kaminski, Christoph Knote, Gerbrand Koren, Jean-Luc Kouassi, Meiyun Lin, Tianjia Liu, Jianmin Ma, Kasemsan Manomaiphiboon, Elisa Bergas Masso, Jessica L. McCarty, Mariano Mertens, Mark Parrington, Helene Peiro, Pallavi Saxena, Saurabh Sonwani, Vanisa Surapipith, Damaris Y. T. Tan, Wenfu Tang, Veerachai Tanpipat, Kostas Tsigaridis, Christine Wiedinmyer, Oliver Wild, Yuanyu Xie, and Paquita Zuidema
Geosci. Model Dev., 18, 3265–3309, https://doi.org/10.5194/gmd-18-3265-2025, https://doi.org/10.5194/gmd-18-3265-2025, 2025
Short summary
Short summary
The multi-model experiment design of the HTAP3 Fires project takes a multi-pollutant approach to improving our understanding of transboundary transport of wildland fire and agricultural burning emissions and their impacts. The experiments are designed with the goal of answering science policy questions related to fires. The options for the multi-model approach, including inputs, outputs, and model setup, are discussed, and the official recommendations for the project are presented.
Maurin Zouzoua, Sophie Bastin, Fabienne Lohou, Marie Lothon, Marjolaine Chiriaco, Mathilde Jome, Cécile Mallet, Laurent Barthes, and Guylaine Canut
Geosci. Model Dev., 18, 3211–3239, https://doi.org/10.5194/gmd-18-3211-2025, https://doi.org/10.5194/gmd-18-3211-2025, 2025
Short summary
Short summary
This study proposes using a statistical model to freeze errors due to differences in environmental forcing when evaluating the surface turbulent heat fluxes from numerical simulations with observations. The statistical model is first built with observations and then applied to the simulated environment to generate possibly observed fluxes. This novel method provides insight into differently evaluating the numerical formulation of turbulent heat fluxes with a long period of observational data.
Oxana Drofa
Geosci. Model Dev., 18, 3175–3209, https://doi.org/10.5194/gmd-18-3175-2025, https://doi.org/10.5194/gmd-18-3175-2025, 2025
Short summary
Short summary
This paper presents the result of many years of effort of the author, who developed an original mathematical numerical model of heat and moisture exchange processes in soil, vegetation, and snow. The author relied on her 30 years of research experience in atmospheric numerical modelling. The presented model is the fruit of the author's research on physical processes at the surface–atmosphere interface and their numerical approximation and aims at improving numerical weather forecasting and climate simulations.
Tyler P. Janoski, Ivan Mitevski, Ryan J. Kramer, Michael Previdi, and Lorenzo M. Polvani
Geosci. Model Dev., 18, 3065–3079, https://doi.org/10.5194/gmd-18-3065-2025, https://doi.org/10.5194/gmd-18-3065-2025, 2025
Short summary
Short summary
We developed ClimKern, a Python package and radiative kernel repository, to simplify calculating radiative feedbacks and make climate sensitivity studies more reproducible. Testing of ClimKern with sample climate model data reveals that radiative kernel choice may be more important than previously thought, especially in polar regions. Our work highlights the need for kernel sensitivity analyses to be included in future studies.
Matti Niskanen, Aku Seppänen, Henri Oikarinen, Miska Olin, Panu Karjalainen, Santtu Mikkonen, and Kari Lehtinen
Geosci. Model Dev., 18, 2983–3001, https://doi.org/10.5194/gmd-18-2983-2025, https://doi.org/10.5194/gmd-18-2983-2025, 2025
Short summary
Short summary
Particle size is a key factor determining the properties of aerosol particles which have a major influence on the climate and on human health. When measuring the particle sizes, however, sometimes the sampling lines that transfer the aerosol to the measurement device distort the size distribution, making the measurement unreliable. We propose a method to correct for the distortions and estimate the true particle sizes, improving measurement accuracy.
Johann Rasmus Nüß, Nikos Daskalakis, Fabian Günther Piwowarczyk, Angelos Gkouvousis, Oliver Schneising, Michael Buchwitz, Maria Kanakidou, Maarten C. Krol, and Mihalis Vrekoussis
Geosci. Model Dev., 18, 2861–2890, https://doi.org/10.5194/gmd-18-2861-2025, https://doi.org/10.5194/gmd-18-2861-2025, 2025
Short summary
Short summary
We estimate carbon monoxide emissions through inverse modeling, an approach where measurements of tracers in the atmosphere are fed to a model to calculate backwards in time (inverse) where the tracers came from. We introduce measurements from a new satellite instrument and show that, in most places globally, these on their own sufficiently constrain the emissions. This alleviates the need for additional datasets, which could shorten the delay for future carbon monoxide source estimates.
Ashu Dastoor, Hélène Angot, Johannes Bieser, Flora Brocza, Brock Edwards, Aryeh Feinberg, Xinbin Feng, Benjamin Geyman, Charikleia Gournia, Yipeng He, Ian M. Hedgecock, Ilia Ilyin, Jane Kirk, Che-Jen Lin, Igor Lehnherr, Robert Mason, David McLagan, Marilena Muntean, Peter Rafaj, Eric M. Roy, Andrei Ryjkov, Noelle E. Selin, Francesco De Simone, Anne L. Soerensen, Frits Steenhuisen, Oleg Travnikov, Shuxiao Wang, Xun Wang, Simon Wilson, Rosa Wu, Qingru Wu, Yanxu Zhang, Jun Zhou, Wei Zhu, and Scott Zolkos
Geosci. Model Dev., 18, 2747–2860, https://doi.org/10.5194/gmd-18-2747-2025, https://doi.org/10.5194/gmd-18-2747-2025, 2025
Short summary
Short summary
This paper introduces the Multi-Compartment Mercury (Hg) Modeling and Analysis Project (MCHgMAP) aimed at informing the effectiveness evaluations of two multilateral environmental agreements: the Minamata Convention on Mercury and the Convention on Long-Range Transboundary Air Pollution. The experimental design exploits a variety of models (atmospheric, land, oceanic ,and multimedia mass balance models) to assess the short- and long-term influences of anthropogenic Hg releases into the environment.
Hilda Sandström and Patrick Rinke
Geosci. Model Dev., 18, 2701–2724, https://doi.org/10.5194/gmd-18-2701-2025, https://doi.org/10.5194/gmd-18-2701-2025, 2025
Short summary
Short summary
Machine learning has the potential to aid the identification of organic molecules involved in aerosol formation. Yet, progress is stalled by a lack of curated atmospheric molecular datasets. Here, we compared atmospheric compounds with large molecular datasets used in machine learning and found minimal overlap with similarity algorithms. Our result underlines the need for collaborative efforts to curate atmospheric molecular data to facilitate machine learning models in atmospheric sciences.
Juan Escobar, Philippe Wautelet, Joris Pianezze, Florian Pantillon, Thibaut Dauhut, Christelle Barthe, and Jean-Pierre Chaboureau
Geosci. Model Dev., 18, 2679–2700, https://doi.org/10.5194/gmd-18-2679-2025, https://doi.org/10.5194/gmd-18-2679-2025, 2025
Short summary
Short summary
The Meso-NH weather research code is adapted for GPUs using OpenACC, leading to significant performance and energy efficiency improvements. Called MESONH-v55-OpenACC, it includes enhanced memory management, communication optimizations and a new solver. On the AMD MI250X Adastra platform, it achieved up to 6× speedup and 2.3× energy efficiency gain compared to CPUs. Storm simulations at 100 m resolution show positive results, positioning the code for future use on exascale supercomputers.
Jie Gao, Yi Huang, Jonathon S. Wright, Ke Li, Tao Geng, and Qiurun Yu
Geosci. Model Dev., 18, 2569–2586, https://doi.org/10.5194/gmd-18-2569-2025, https://doi.org/10.5194/gmd-18-2569-2025, 2025
Short summary
Short summary
The aerosol in the upper troposphere and stratosphere is highly variable, and its radiative effect is poorly understood. To estimate this effect, the radiative kernel is constructed and applied. The results show that the kernels can reproduce aerosol radiative effects and are expected to simulate stratospheric aerosol radiative effects. This approach reduces computational expense, is consistent with radiative model calculations, and can be applied to atmospheric models with speed requirements.
Ji Won Yoon, Seungyeon Lee, Ebony Lee, and Seon Ki Park
Geosci. Model Dev., 18, 2303–2328, https://doi.org/10.5194/gmd-18-2303-2025, https://doi.org/10.5194/gmd-18-2303-2025, 2025
Short summary
Short summary
This study evaluates the Weather Research and Forecasting Model (WRF) coupled with Chemistry (WRF-Chem) to predict a mega Asian dust storm (ADS) over South Korea on 28–29 March 2021. We assessed combinations of five dust emission and four land surface schemes by analyzing meteorological and air quality variables. The best scheme combination reduced the root mean square error (RMSE) for particulate matter 10 (PM10) by up to 29.6 %, demonstrating the highest performance.
Jianyu Lin, Tie Dai, Lifang Sheng, Weihang Zhang, Shangfei Hai, and Yawen Kong
Geosci. Model Dev., 18, 2231–2248, https://doi.org/10.5194/gmd-18-2231-2025, https://doi.org/10.5194/gmd-18-2231-2025, 2025
Short summary
Short summary
The effectiveness of this assimilation system and its sensitivity to the ensemble member size and length of the assimilation window are investigated. This study advances our understanding of the selection of basic parameters in the four-dimensional local ensemble transform Kalman filter assimilation system and the performance of ensemble simulation in a particulate-matter-polluted environment.
Jens Peter Karolus Wenceslaus Frankemölle, Johan Camps, Pieter De Meutter, and Johan Meyers
Geosci. Model Dev., 18, 1989–2003, https://doi.org/10.5194/gmd-18-1989-2025, https://doi.org/10.5194/gmd-18-1989-2025, 2025
Short summary
Short summary
To detect anomalous radioactivity in the environment, it is paramount that we understand the natural background level. In this work, we propose a statistical model to describe the most likely background level and the associated uncertainty in a network of dose rate detectors. We train, verify, and validate the model using real environmental data. Using the model, we show that we can correctly predict the background level in a subset of the detector network during a known
anomalous event.
Jean-François Grailet, Robin J. Hogan, Nicolas Ghilain, David Bolsée, Xavier Fettweis, and Marilaure Grégoire
Geosci. Model Dev., 18, 1965–1988, https://doi.org/10.5194/gmd-18-1965-2025, https://doi.org/10.5194/gmd-18-1965-2025, 2025
Short summary
Short summary
The MAR (Modèle Régional Atmosphérique) is a regional climate model used for weather forecasting and studying the climate over various regions. This paper presents an update of MAR thanks to which it can precisely decompose solar radiation, in particular in the UV (ultraviolet) and photosynthesis ranges, both being critical to human health and ecosystems. As a first application of this new capability, this paper presents a method for predicting UV indices with MAR.
Yi-Ning Shi, Jun Yang, Wei Han, Lujie Han, Jiajia Mao, Wanlin Kan, and Fuzhong Weng
Geosci. Model Dev., 18, 1947–1964, https://doi.org/10.5194/gmd-18-1947-2025, https://doi.org/10.5194/gmd-18-1947-2025, 2025
Short summary
Short summary
Direct assimilation of observations from ground-based microwave radiometers (GMRs) holds significant potential for improving forecast accuracy. Radiative transfer models (RTMs) play a crucial role in direct data assimilation. In this study, we introduce a new RTM, the Advanced Radiative Transfer Modeling System – Ground-Based (ARMS-gb), designed to simulate brightness temperatures observed by GMRs along with their Jacobians. Several enhancements have been incorporated to achieve higher accuracy.
R. Phani Murali Krishna, Siddharth Kumar, A. Gopinathan Prajeesh, Peter Bechtold, Nils Wedi, Kumar Roy, Malay Ganai, B. Revanth Reddy, Snehlata Tirkey, Tanmoy Goswami, Radhika Kanase, Sahadat Sarkar, Medha Deshpande, and Parthasarathi Mukhopadhyay
Geosci. Model Dev., 18, 1879–1894, https://doi.org/10.5194/gmd-18-1879-2025, https://doi.org/10.5194/gmd-18-1879-2025, 2025
Short summary
Short summary
The High-Resolution Global Forecast Model (HGFM) is an advanced iteration of the operational Global Forecast System (GFS) model. HGFM can produce forecasts at a spatial scale of ~6 km in tropics. It demonstrates improved accuracy in short- to medium-range weather prediction over the Indian region, with notable success in predicting extreme events. Further, the model will be entrusted to operational forecasting agencies after validation and testing.
Jenna Ritvanen, Seppo Pulkkinen, Dmitri Moisseev, and Daniele Nerini
Geosci. Model Dev., 18, 1851–1878, https://doi.org/10.5194/gmd-18-1851-2025, https://doi.org/10.5194/gmd-18-1851-2025, 2025
Short summary
Short summary
Nowcasting models struggle with the rapid evolution of heavy rain, and common verification methods are unable to describe how accurately the models predict the growth and decay of heavy rain. We propose a framework to assess model performance. In the framework, convective cells are identified and tracked in the forecasts and observations, and the model skill is then evaluated by comparing differences between forecast and observed cells. We demonstrate the framework with four open-source models.
Andrew Geiss and Po-Lun Ma
Geosci. Model Dev., 18, 1809–1827, https://doi.org/10.5194/gmd-18-1809-2025, https://doi.org/10.5194/gmd-18-1809-2025, 2025
Short summary
Short summary
Particles in the Earth's atmosphere strongly impact the planet's energy budget, and atmosphere simulations require accurate representation of their interaction with light. This work introduces two approaches to represent light scattering by small particles. The first is a scattering simulator based on Mie theory implemented in Python. The second is a neural network emulator that is more accurate than existing methods and is fast enough to be used in climate and weather simulations.
Andrin Jörimann, Timofei Sukhodolov, Beiping Luo, Gabriel Chiodo, Graham Mann, and Thomas Peter
EGUsphere, https://doi.org/10.5194/egusphere-2025-145, https://doi.org/10.5194/egusphere-2025-145, 2025
Short summary
Short summary
Aerosol particles in the stratosphere affect our climate. Climate models therefore need an accurate description of their properties and evolution. Satellites measure how strongly aerosol particles extinguish light passing through the stratosphere. We describe a method to use such aerosol extinction data to retrieve the number and sizes of the aerosol particles and calculate their optical effects. The resulting data sets for models are validated against ground-based and balloon observations.
Qin Wang, Bo Zeng, Gong Chen, and Yaoting Li
Geosci. Model Dev., 18, 1769–1784, https://doi.org/10.5194/gmd-18-1769-2025, https://doi.org/10.5194/gmd-18-1769-2025, 2025
Short summary
Short summary
This study evaluates the performance of four planetary boundary layer (PBL) schemes in near-surface wind fields over the Sichuan Basin, China. Using 112 sensitivity experiments with the Weather Research and Forecasting (WRF) model and focusing on 28 wind events, it is found that wind direction was less sensitive to the PBL schemes. The quasi-normal scale elimination (QNSE) scheme captured temporal variations best, while the Mellor–Yamada–Janjić (MYJ) scheme had the least error in wind speed.
Tai-Long He, Nikhil Dadheech, Tammy M. Thompson, and Alexander J. Turner
Geosci. Model Dev., 18, 1661–1671, https://doi.org/10.5194/gmd-18-1661-2025, https://doi.org/10.5194/gmd-18-1661-2025, 2025
Short summary
Short summary
It is computationally expensive to infer greenhouse gas (GHG) emissions using atmospheric observations. This is partly due to the detailed model used to represent atmospheric transport. We demonstrate how a machine learning (ML) model can be used to simulate high-resolution atmospheric transport. This type of ML model will help estimate GHG emissions using dense observations, which are becoming increasingly common with the proliferation of urban monitoring networks and geostationary satellites.
Wei Li, Beiming Tang, Patrick C. Campbell, Youhua Tang, Barry Baker, Zachary Moon, Daniel Tong, Jianping Huang, Kai Wang, Ivanka Stajner, and Raffaele Montuoro
Geosci. Model Dev., 18, 1635–1660, https://doi.org/10.5194/gmd-18-1635-2025, https://doi.org/10.5194/gmd-18-1635-2025, 2025
Short summary
Short summary
The study describes the updates of NOAA's current UFS-AQMv7 air quality forecast model by incorporating the latest scientific and structural changes in CMAQv5.4. An evaluation during the summer of 2023 shows that the updated model overall improves the simulation of MDA8 O3 by reducing the bias by 8%–12% in the contiguous US. PM2.5 predictions have mixed results due to wildfire, highlighting the need for future refinements.
Yanwei Zhu, Aitor Atencia, Markus Dabernig, and Yong Wang
Geosci. Model Dev., 18, 1545–1559, https://doi.org/10.5194/gmd-18-1545-2025, https://doi.org/10.5194/gmd-18-1545-2025, 2025
Short summary
Short summary
Most works have delved into convective weather nowcasting, and only a few works have discussed the nowcasting uncertainty for variables at the surface level. Hence, we proposed a method to estimate uncertainty. Generating appropriate noises associated with the characteristic of the error in analysis can simulate the uncertainty of nowcasting. This method can contribute to the estimation of near–surface analysis uncertainty in both nowcasting applications and ensemble nowcasting development.
Joël Thanwerdas, Antoine Berchet, Lionel Constantin, Aki Tsuruta, Michael Steiner, Friedemann Reum, Stephan Henne, and Dominik Brunner
Geosci. Model Dev., 18, 1505–1544, https://doi.org/10.5194/gmd-18-1505-2025, https://doi.org/10.5194/gmd-18-1505-2025, 2025
Short summary
Short summary
The Community Inversion Framework (CIF) brings together methods for estimating greenhouse gas fluxes from atmospheric observations. The initial ensemble method implemented in CIF was found to be incomplete and could hardly be compared to other ensemble methods employed in the inversion community. In this paper, we present and evaluate a new implementation of the ensemble mode, building upon the initial developments.
Astrid Kerkweg, Timo Kirfel, Duong H. Do, Sabine Griessbach, Patrick Jöckel, and Domenico Taraborrelli
Geosci. Model Dev., 18, 1265–1286, https://doi.org/10.5194/gmd-18-1265-2025, https://doi.org/10.5194/gmd-18-1265-2025, 2025
Short summary
Short summary
Normally, the Modular Earth Submodel System (MESSy) is linked to complete dynamic models to create chemical climate models. However, the modular concept of MESSy and the newly developed DWARF component presented here make it possible to create simplified models that contain only one or a few process descriptions. This is very useful for technical optimisation, such as porting to GPUs, and can be used to create less complex models, such as a chemical box model.
Peter Wind and Willem van Caspel
EGUsphere, https://doi.org/10.5194/egusphere-2024-3571, https://doi.org/10.5194/egusphere-2024-3571, 2025
Short summary
Short summary
This paper presents a numerical method to assess the origin of air pollution. Combined with a numerical air pollution transport and chemistry model, it can follow the contributions from a large number of emission sources. The result is a series of maps that give the relative contributions from for example all European countries at each point.
Julian Vogel, Sebastian Stadler, Ganesh Chockalingam, Afshin Afshari, Johanna Henning, and Matthias Winkler
EGUsphere, https://doi.org/10.5194/egusphere-2025-144, https://doi.org/10.5194/egusphere-2025-144, 2025
Short summary
Short summary
This study presents a toolkit to simplify input data creation for the urban microclimate model PALM-4U. It introduces novel methods to automate the use of open data sources. Our analysis of four test cases created from different geographic data sources shows variations in temperature, humidity, and wind speed, influenced by data quality. Validation indicates that the automated methods yield results comparable to expert-driven approaches, facilitating user-friendly urban climate modeling.
Edward C. Chan, Ilona J. Jäkel, Basit Khan, Martijn Schaap, Timothy M. Butler, Renate Forkel, and Sabine Banzhaf
Geosci. Model Dev., 18, 1119–1139, https://doi.org/10.5194/gmd-18-1119-2025, https://doi.org/10.5194/gmd-18-1119-2025, 2025
Short summary
Short summary
An enhanced emission module has been developed for the PALM model system, improving flexibility and scalability of emission source representation across different sectors. A model for parametrized domestic emissions has also been included, for which an idealized model run is conducted for particulate matter (PM10). The results show that, in addition to individual sources and diurnal variations in energy consumption, vertical transport and urban topology play a role in concentration distribution.
Gregor Ehrensperger, Thorsten Simon, Georg J. Mayr, and Tobias Hell
Geosci. Model Dev., 18, 1141–1153, https://doi.org/10.5194/gmd-18-1141-2025, https://doi.org/10.5194/gmd-18-1141-2025, 2025
Short summary
Short summary
As lightning is a brief and localized event, it is not explicitly resolved in atmospheric models. Instead, expert-based auxiliary descriptions are used to assess it. This study explores how AI can improve our understanding of lightning without relying on traditional expert knowledge. We reveal that AI independently identified the key factors known to experts as essential for lightning in the Alps region. This shows how knowledge discovery could be sped up in areas with limited expert knowledge.
David Patoulias, Kalliopi Florou, and Spyros N. Pandis
Geosci. Model Dev., 18, 1103–1118, https://doi.org/10.5194/gmd-18-1103-2025, https://doi.org/10.5194/gmd-18-1103-2025, 2025
Short summary
Short summary
The effect of the assumed atmospheric nucleation mechanism on particle number concentrations and size distribution was investigated. Two quite different mechanisms involving sulfuric acid and ammonia or a biogenic organic vapor gave quite similar results which were consistent with measurements at 26 measurement stations across Europe. The number of larger particles that serve as cloud condensation nuclei showed little sensitivity to the assumed nucleation mechanism.
Tim Radke, Susanne Fuchs, Christian Wilms, Iuliia Polkova, and Marc Rautenhaus
Geosci. Model Dev., 18, 1017–1039, https://doi.org/10.5194/gmd-18-1017-2025, https://doi.org/10.5194/gmd-18-1017-2025, 2025
Short summary
Short summary
In our study, we built upon previous work to investigate the patterns artificial intelligence (AI) learns to detect atmospheric features like tropical cyclones (TCs) and atmospheric rivers (ARs). As primary objective, we adopt a method to explain the AI used and investigate the plausibility of learned patterns. We find that plausible patterns are learned for both TCs and ARs. Hence, the chosen method is very useful for gaining confidence in the AI-based detection of atmospheric features.
Leon Geers, Ruud Janssen, Gudrun Thorkelsdottir, Jordi Vilà-Guerau de Arellano, and Martijn Schaap
EGUsphere, https://doi.org/10.5194/egusphere-2025-426, https://doi.org/10.5194/egusphere-2025-426, 2025
Short summary
Short summary
High-resolution data on reactive nitrogen deposition are needed to inform cost-effective policies. Here, we describe the implementation of a dry deposition module into a large eddy simulation code. With this model, we are able to represent the turbulent exchange of tracers at the hectometer resolution. The model calculates the dispersion and deposition of NOx and NH3 in great spatial detail, clearly showing the influence of local land use patterns.
Raphaël Périllat, Sylvain Girard, and Irène Korsakissok
EGUsphere, https://doi.org/10.5194/egusphere-2024-3838, https://doi.org/10.5194/egusphere-2024-3838, 2025
Short summary
Short summary
We developed a method to improve decision-making during nuclear crises by predicting the spread of radiation more efficiently. Existing approaches are often too slow, especially when analyzing complex data like radiation maps. Our method combines techniques to simplify these maps and predict them quickly using statistical tools. This approach could help authorities respond faster and more accurately in emergencies, reducing risks to the population and the environment.
Shaofeng Hua, Gang Chen, Baojun Chen, Mingshan Li, and Xin Xu
EGUsphere, https://doi.org/10.5194/egusphere-2024-3834, https://doi.org/10.5194/egusphere-2024-3834, 2025
Short summary
Short summary
Hail forecasting using numerical models remains a challenge. In this study, we found that the commonly used graupel-to-hail conversion parameterization method led to hail overforecasting in heavy rainfall cases where no hail was observed. By incorporating the spongy wet growth process, we successfully mitigated hail overforecasting. The modified scheme also produced hail in real hail events. This research contributes to a better understanding of hail formation.
Stefan Noll, Carsten Schmidt, Patrick Hannawald, Wolfgang Kausch, and Stefan Kimeswenger
EGUsphere, https://doi.org/10.5194/egusphere-2024-3512, https://doi.org/10.5194/egusphere-2024-3512, 2025
Short summary
Short summary
Non-thermal emission from chemical reactions in the Earth's middle und upper atmosphere strongly contributes to the brightness of the night sky below about 2.3 µm. The new Paranal Airglow Line and Continuum Emission model calculates the emission spectrum and its variability with an unprecedented accuracy. Relying on a large spectroscopic data set from astronomical spectrographs and theoretical molecular/atomic data, it is valuable for airglow research and astronomical observatories.
Felipe Cifuentes, Henk Eskes, Enrico Dammers, Charlotte Bryan, and Folkert Boersma
Geosci. Model Dev., 18, 621–649, https://doi.org/10.5194/gmd-18-621-2025, https://doi.org/10.5194/gmd-18-621-2025, 2025
Short summary
Short summary
We tested the capability of the flux divergence approach (FDA) to reproduce known NOx emissions using synthetic NO2 satellite column retrievals from high-resolution model simulations. The FDA accurately reproduced NOx emissions when column observations were limited to the boundary layer and when the variability of the NO2 lifetime, the NOx : NO2 ratio, and NO2 profile shapes were correctly modeled. This introduces strong model dependency, reducing the simplicity of the original FDA formulation.
Stefano Ubbiali, Christian Kühnlein, Christoph Schär, Linda Schlemmer, Thomas C. Schulthess, Michael Staneker, and Heini Wernli
Geosci. Model Dev., 18, 529–546, https://doi.org/10.5194/gmd-18-529-2025, https://doi.org/10.5194/gmd-18-529-2025, 2025
Short summary
Short summary
We explore a high-level programming model for porting numerical weather prediction (NWP) model codes to graphics processing units (GPUs). We present a Python rewrite with the domain-specific library GT4Py (GridTools for Python) of two renowned cloud microphysics schemes and the associated tangent-linear and adjoint algorithms. We find excellent portability, competitive GPU performance, robust execution on diverse computing architectures, and enhanced code maintainability and user productivity.
Cited articles
Adewoyin, R. A., Dueben, P., Watson, P., He, Y., and Dutta, R.: TRU-NET: a deep learning approach to high resolution prediction of rainfall, Mach. Learn., 110, 2035–2062, https://doi.org/10.1007/s10994-021-06022-6, 2021.
Agrawal, S., Barrington, L., Bromberg, C., Burge, J., Gazen, C., and Hickey, J.: Machine learning for precipitation nowcasting from radar images, arXiv [preprint], arXiv:1912.12132, 2019.
Albu, A.-I., Czibula, G., Mihai, A., Czibula, I. G., Burcea, S., and Mezghani, A.: NeXtNow: A Convolutional Deep Learning Model for the Prediction of Weather Radar Data for Nowcasting Purposes, Remote Sens.-Basel, 14, 3890, https://doi.org/10.3390/rs14163890, 2022.
Aswin, S., Geetha, P., and Vinayakumar, R.: Deep learning models for the prediction of rainfall, 2018 International Conference on Communication and Signal Processing (ICCSP), 0657–0661, 2018.
Ayzel, G.: RainNet: a convolutional neural network for radar-based precipitation nowcasting, GitHub [code], https://github.com/hydrogo/rainnet (last access: 18 September 2023), 2020.
Ayzel, G., Scheffer, T., and Heistermann, M.: RainNet v1.0: a convolutional neural network for radar-based precipitation nowcasting, Geosci. Model Dev., 13, 2631–2644, https://doi.org/10.5194/gmd-13-2631-2020, 2020.
Bai, S., Kolter, J. Z., and Koltun, V.: An empirical evaluation of generic convolutional and recurrent networks for sequence modeling, arXiv [preprint], https://doi.org/10.48550/arXiv.1803.01271, 2018.
Barth, A., Alvera-Azcárate, A., Troupin, C., and Beckers, J.-M.: DINCAE 2.0: multivariate convolutional neural network with error estimates to reconstruct sea surface temperature satellite and altimetry observations, Geosci. Model Dev., 15, 2183–2196, https://doi.org/10.5194/gmd-15-2183-2022, 2022.
Bonnet, S. M., Evsukoff, A., and Morales Rodriguez, C. A.: Precipitation nowcasting with weather radar images and deep learning in são paulo, brasil, Atmosphere, 11, 1157, https://doi.org/10.3390/atmos11111157, 2020.
Bouget, V., Béréziat, D., Brajard, J., Charantonis, A., and Filoche, A.: Fusion of rain radar images and wind forecasts in a deep learning model applied to rain nowcasting, Remote Sens.-Basel, 13, 246, https://doi.org/10.3390/rs13020246, 2021.
Bowler, N. E., Pierce, C. E., and Seed, A. W.: STEPS: A probabilistic precipitation forecasting scheme which merges an extrapolation nowcast with downscaled NWP, Q. J. Roy. Meteor. Soc., 132, 2127–2155, https://doi.org/10.1256/qj.04.100, 2006.
Chattopadhyay, A., Mustafa, M., Hassanzadeh, P., Bach, E., and Kashinath, K.: Towards physics-inspired data-driven weather forecasting: integrating data assimilation with a deep spatial-transformer-based U-NET in a case study with ERA5, Geosci. Model Dev., 15, 2221–2237, https://doi.org/10.5194/gmd-15-2221-2022, 2022.
Chen, G. and Wang, W. C.: Short-term precipitation prediction for contiguous United States using deep learning, Geophys. Res. Lett., 49, e2022GL097904, https://doi.org/10.1029/2022GL097904, 2022.
Chen, L., Cao, Y., Ma, L., and Zhang, J.: A deep learning-based methodology for precipitation nowcasting with radar, Earth Space Sci., 7, e2019EA000812, https://doi.org/10.1029/2019EA000812, 2020.
Chi, J. and Kim, H.-C.: Retrieval of daily sea ice thickness from AMSR2 passive microwave data using ensemble convolutional neural networks, GISci. Remote Sens., 58, 812–830, 2021.
Chinita, M. J., Witte, M., Kurowski, M. J., Teixeira, J., Suselj, K., Matheou, G., and Bogenschutz, P.: Improving the representation of shallow cumulus convection with the simplified-higher-order-closure–mass-flux (SHOC+MF v1.0) approach, Geosci. Model Dev., 16, 1909–1924, https://doi.org/10.5194/gmd-16-1909-2023, 2023.
Cho, D., Yoo, C., Im, J., Lee, Y., and Lee, J.: Improvement of spatial interpolation accuracy of daily maximum air temperature in urban areas using a stacking ensemble technique, GISci. Remote Sens., 57, 633–649, 2020.
Cho, K., Van Merriënboer, B., Gulcehre, C., Bahdanau, D., Bougares, F., Schwenk, H., and Bengio, Y.: Learning phrase representations using RNN encoder-decoder for statistical machine translation, arXiv [preprint], arXiv:1406.1078, 2014.
Choi, S. and Kim, Y.: Rad-cGAN v1.0: Radar-based precipitation nowcasting model with conditional generative adversarial networks for multiple dam domains, Geosci. Model Dev., 15, 5967–5985, https://doi.org/10.5194/gmd-15-5967-2022, 2022.
Chung, K.-S. and Yao, I.-A.: Improving radar echo Lagrangian extrapolation nowcasting by blending numerical model wind information: Statistical performance of 16 typhoon cases, Mon. Weather Rev., 148, 1099–1120, https://doi.org/10.1175/MWR-D-19-0193.1, 2020.
Cuomo, J. and Chandrasekar, V.: Use of Deep Learning for Weather Radar Nowcasting, J. Atmos. Ocean. Tech., 38, 1641–1656, https://doi.org/10.1175/JTECH-D-21-0012.1, 2021.
Dixon, M. and Wiener, G.: TITAN: Thunderstorm identification, tracking, analysis, and nowcasting – A radar-based methodology, J. Atmos. Ocean. Tech., 10, 785–797, https://doi.org/10.1175/1520-0426(1993)010<0785:TTITAA>2.0.CO;2, 1993.
Franch, G., Nerini, D., Pendesini, M., Coviello, L., Jurman, G., and Furlanello, C.: Precipitation nowcasting with orographic enhanced stacked generalization: Improving deep learning predictions on extreme events, Atmosphere, 11, 267, https://doi.org/10.3390/atmos11030267, 2020.
Franch, G., Tomasi, E., Poli, V., Cardinali, C., Cristoforetti, M., and Alberoni, P. P.: Ensemble precipitation nowcasting by combination of generative and transformer deep learning models, Copernicus Meetings, https://doi.org/10.5194/egusphere-egu23-15153, 2023.
Gardoll, S. and Boucher, O.: Classification of tropical cyclone containing images using a convolutional neural network: performance and sensitivity to the learning dataset, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2022-147, 2022.
Gehring, J., Auli, M., Grangier, D., Yarats, D., and Dauphin, Y. N.: Convolutional Sequence to Sequence Learning, in: Proceedings of the 34th International Conference on Machine Learning – Volume 70, ICML'17, 6–11 August 2017, Sydney, Australia, 1243–1252, 2017.
Geiss, A., Silva, S. J., and Hardin, J. C.: Downscaling atmospheric chemistry simulations with physically consistent deep learning, Geosci. Model Dev., 15, 6677–6694, https://doi.org/10.5194/gmd-15-6677-2022, 2022.
Germann, U. and Zawadzki, I.: Scale-dependence of the predictability of precipitation from continental radar images. Part I: Description of the methodology, Mon. Weather Rev., 130, 2859–2873, https://doi.org/10.1175/1520-0493(2002)130<2859:SDOTPO>2.0.CO;2, 2002.
Germann, U. and Zawadzki, I.: Scale dependence of the predictability of precipitation from continental radar images. Part II: Probability forecasts, J. Appl. Meteorol., 43, 74–89, https://doi.org/10.1175/1520-0450(2004)043<0074:SDOTPO>2.0.CO;2, 2004.
Haiden, T., Kann, A., Wittmann, C., Pistotnik, G., Bica, B., and Gruber, C.: The Integrated Nowcasting through Comprehensive Analysis (INCA) system and its validation over the Eastern Alpine region, Weather Forecast., 26, 166–183, https://doi.org/10.1175/2010WAF2222451.1, 2011.
Han, D.: Supplementary code and data: Key factors for quantitative precipitation nowcasting using ground weather radar data based on deep learning, Zenodo [code and data set], https://doi.org/10.5281/zenodo.8353423, 2023.
Han, D., Lee, J., Im, J., Sim, S., Lee, S., and Han, H.: A novel framework of detecting convective initiation combining automated sampling, machine learning, and repeated model tuning from geostationary satellite data, Remote Sens.-Basel, 11, 1454, https://doi.org/10.3390/rs11121454, 2019.
Han, D., Choo, M., Im, J., Shin, Y., Lee, J., and Jung, S.: Precipitation nowcasting using ground radar data and simpler yet better video prediction deep learning, GISci. Remote Sens., 60, 2203363, https://doi.org/10.1080/15481603.2023.2203363, 2023.
Han, L., Zhang, J., Chen, H., Zhang, W., and Yao, S.: Toward the Predictability of a Radar-Based Nowcasting System for Different Precipitation Systems, IEEE Geosci. Remote S., 19, 1–5, 2022.
Hochreiter, S. and Schmidhuber, J.: Long short-term memory, Neural Comput., 9, 1735–1780, https://doi.org/10.1162/neco.1997.9.8.1735, 1997.
Jeong, C. H., Kim, W., Joo, W., Jang, D., and Yi, M. Y.: Enhancing the encoding-forecasting model for precipitation nowcasting by putting high emphasis on the latest data of the time step, Atmosphere, 12, 261, https://doi.org/10.3390/atmos12020261, 2021.
Kang, J., Wang, H., Yuan, F., Wang, Z., Huang, J., and Qiu, T.: Prediction of precipitation based on recurrent neural networks in Jingdezhen, Jiangxi Province, China, Atmosphere, 11, 246, https://doi.org/10.3390/atmos11030246, 2020.
Kim, D.-K., Suezawa, T., Mega, T., Kikuchi, H., Yoshikawa, E., Baron, P., and Ushio, T.: Improving precipitation nowcasting using a three-dimensional convolutional neural network model from Multi Parameter Phased Array Weather Radar observations, Atmos. Res., 262, 105774, https://doi.org/10.1016/j.atmosres.2021.105774, 2021.
Kim, M., Lee, J., and Im, J.: Deep learning-based monitoring of overshooting cloud tops from geostationary satellite data, GISci. Remote Sens., 55, 763–792, 2018.
Kim, Y. and Hong, S.: Very Short-Term Rainfall Prediction Using Ground Radar Observations and Conditional Generative Adversarial Networks, IEEE T. Geosci. Remote, 60, 4104308, https://doi.org/10.1109/TGRS.2021.3108812, 2021.
Kim, Y. J., Kim, H.-C., Han, D., Lee, S., and Im, J.: Prediction of monthly Arctic sea ice concentrations using satellite and reanalysis data based on convolutional neural networks, The Cryosphere, 14, 1083–1104, https://doi.org/10.5194/tc-14-1083-2020, 2020.
Kim, Y. J., Han, D., Jang, E., Im, J., and Sung, T.: Remote sensing of sea surface salinity: challenges and research directions, GISci. Remote Sens., 60, 2166377, 2023.
Kim, Y.-T., Park, M., and Kwon, H.-H.: Spatio-temporal summer rainfall pattern in 2020 from a rainfall frequency perspective, Journal of Korean Society of Disaster and Security, 13, 93–104, https://doi.org/10.21729/ksds.2020.13.4.93, 2020.
Kingma, D. P. and Ba, J.: Adam: A method for stochastic optimization, arXiv [preprint], arXiv:1412.6980, 2014.
Ko, J., Lee, K., Hwang, H., Oh, S.-G., Son, S.-W., and Shin, K.: Effective training strategies for deep-learning-based precipitation nowcasting and estimation, Comput. Geosci., 161, 105072, https://doi.org/10.1016/j.cageo.2022.105072, 2022.
Korea Meteorological Administration: Ground weather radar over South Korea, Korea Meteorological Administration [data set], https://data.kma.go.kr/data/rmt/rmtList.do?code=11&pgmNo=62&tabNo=2, last access: 18 September 2023.
Lee, J., Kim, M., Im, J., Han, H., and Han, D.: Pre-trained feature aggregated deep learning-based monitoring of overshooting tops using multi-spectral channels of GeoKompsat-2A advanced meteorological imagery, GISci. Remote Sens., 58, 1052–1071, https://doi.org/10.1080/15481603.2021.1960075, 2021.
Leinonen, J., Hamann, U., Nerini, D., Germann, U., and Franch, G.: Latent diffusion models for generative precipitation nowcasting with accurate uncertainty quantification, arXiv [preprint], arXiv:2304.12891, 2023.
Lepetit, P., Ly, C., Barthès, L., Mallet, C., Viltard, N., Lemaitre, Y., and Rottner, L.: Using deep learning for restoration of precipitation echoes in radar data, IEEE T. Geosci. Remote, 60, 1–14, 2021.
Liu, J., Xu, L., and Chen, N.: A spatiotemporal deep learning model ST-LSTM-SA for hourly rainfall forecasting using radar echo images, J. Hydrol., 609, 127748, https://doi.org/10.1016/j.jhydrol.2022.127748, 2022.
Lu, Y., James, T., Schillaci, C., and Lipani, A.: Snow detection in alpine regions with Convolutional Neural Networks: discriminating snow from cold clouds and water body, GISci. Remote Sens., 59, 1321–1343, https://doi.org/10.1080/15481603.2022.2112391, 2022.
Marshall, J. S. and Palmer, W. K. M.: The distribution of raindrops with size, J. Atmos. Sci., 5, 165–166, https://doi.org/10.1175/1520-0469(1948)005<0165:TDORWS>2.0.CO;2, 1948.
Mecklenburg, S., Joss, J., and Schmid, W.: Improving the nowcasting of precipitation in an Alpine region with an enhanced radar echo tracking algorithm, J. Hydrol., 239, 46–68, https://doi.org/10.1016/S0022-1694(00)00352-8, 2000.
Ni, L., Wang, D., Singh, V. P., Wu, J., Wang, Y., Tao, Y., and Zhang, J.: Streamflow and rainfall forecasting by two long short-term memory-based models, J. Hydrol., 583, 124296, https://doi.org/10.1016/j.jhydrol.2019.124296, 2020.
Nielsen, A. H.: Video-Prediction-using-PyTorch, GitHub [code], https://github.com/holmdk/Video-Prediction-using-PyTorch/tree/master (last access: 18 September 2023), 2019.
Prudden, R., Adams, S., Kangin, D., Robinson, N., Ravuri, S., Mohamed, S., and Arribas, A.: A review of radar-based nowcasting of precipitation and applicable machine learning techniques, arXiv [preprint], arXiv:2005.04988, 2020.
Pulkkinen, S., Nerini, D., Pérez Hortal, A. A., Velasco-Forero, C., Seed, A., Germann, U., and Foresti, L.: Pysteps: an open-source Python library for probabilistic precipitation nowcasting (v1.0), Geosci. Model Dev., 12, 4185–4219, https://doi.org/10.5194/gmd-12-4185-2019, 2019.
Qichi, Y., Lihui, W., Jinliang, H., Linzhi, L., Xiaodong, L., Fei, X., Yun, D., Xue, Y., and Feng, L.: A novel alpine land cover classification strategy based on a deep convolutional neural network and multi-source remote sensing data in Google Earth Engine, GISci. Remote Sens., 60, 2233756, https://doi.org/10.1080/15481603.2023.2233756, 2023.
Ravuri, S., Lenc, K., Willson, M., Kangin, D., Lam, R., Mirowski, P., Fitzsimons, M., Athanassiadou, M., Kashem, S., and Madge, S.: Skilful precipitation nowcasting using deep generative models of radar, Nature, 597, 672–677, https://doi.org/10.1038/s41586-021-03854-z, 2021.
Ronneberger, O., Fischer, P., and Brox, T.: U-net: Convolutional networks for biomedical image segmentation, International Conference on Medical image computing and computer-assisted intervention, 234–241, https://doi.org/10.1007/978-3-319-24574-4_28, 2015. .
Samsi, S., Mattioli, C. J., and Veillette, M. S.: Distributed deep learning for precipitation nowcasting, 2019 IEEE High Performance Extreme Computing Conference (HPEC), 1–7, https://doi.org/10.1109/HPEC.2019.8916416, 2019.
Sato, H. and Ise, T.: Predicting global terrestrial biomes with the LeNet convolutional neural network, Geosci. Model Dev., 15, 3121–3132, https://doi.org/10.5194/gmd-15-3121-2022, 2022.
Shi, X., Chen, Z., Wang, H., Yeung, D.-Y., Wong, W.-K., and Woo, W.-C.: Convolutional LSTM Network: a machine learning approach for precipitation nowcasting, in: Proceedings of the 29th International Conference on Neural Information Processing Systems, Montreal, Canada, 7–12 December 2015, NeurIPS, 802–810, https://proceedings.neurips.cc/paper_files/paper/2015/file/07563a3fe3bbe7e3ba84431ad9d055af-Paper.pdf (last access: 14 October 2023), 2015.
Shi, X., Gao, Z., Lausen, L., Wang, H., Yeung, D.-Y., Wong, W.-K., and Woo, W.-C.: Deep Learning for Precipitation Nowcasting: A Benchmark and A New Model, in: Advances in Neural Information Processing Systems, 30, 5617–5627, https://papers.nips.cc/paper_files/paper/2017/file/a6db4ed04f1621a119799fd3d7545d3d-Paper.pdf (last access: 14 October 2023), 2017.
Trebing, K., Staǹczyk, T., and Mehrkanoon, S.: SmaAt-UNet: Precipitation nowcasting using a small attention-UNet architecture, Pattern Recogn. Lett., 145, 178–186, https://doi.org/10.1016/j.patrec.2021.01.036, 2021.
Turner, B., Zawadzki, I., and Germann, U.: Predictability of precipitation from continental radar images. Part III: Operational nowcasting implementation (MAPLE), J. Appl. Meteorol. Clim., 43, 231–248, https://doi.org/10.1175/1520-0450(2004)043<0231:POPFCR>2.0.CO;2, 2004.
Wang, P., Smeaton, A., Lao, S., O'Connor, E., Ling, Y., and O'Connor, N.: Short-term rainfall nowcasting: Using rainfall radar imaging, Eurographics Ireland 2009: The 9th Irish Workshop on Computer Graphics, Dublin, Ireland, 2009.
Wu, X., Liu, X., Zhang, D., Zhang, J., He, J., and Xu, X.: Simulating mixed land-use change under multi-label concept by integrating a convolutional neural network and cellular automata: A case study of Huizhou, China, GISci. Remote Sens., 59, 609–632, https://doi.org/10.1080/15481603.2022.2049493, 2022.
Xiong, T., He, J., Wang, H., Tang, X., Shi, Z., and Zeng, Q.: Contextual sa-attention convolutional LSTM for precipitation nowcasting: A spatiotemporal sequence forecasting view, IEEE J. Sel. Top. Appl., 14, 12479–12491, https://doi.org/10.1109/JSTARS.2021.3128522, 2021.
Yano, J.-I., Ziemiañski, M. Z., Cullen, M., Termonia, P., Onvlee, J., Bengtsson, L., Carrassi, A., Davy, R., Deluca, A., and Gray, S. L.: Scientific challenges of convective-scale numerical weather prediction, B. Am. Meteorol. Soc., 99, 699–710, https://doi.org/10.1175/BAMS-D-17-0125.1, 2018.
Zhang, F., Wang, X., Guan, J., Wu, M., and Guo, L.: RN-Net: A deep learning approach to 0–2 h rainfall nowcasting based on radar and automatic weather station data, Sensors, 21, 1981, https://doi.org/10.3390/s21061981, 2021.
Zhang, Y., Bi, S., Liu, L., Chen, H., Zhang, Y., Shen, P., Yang, F., Wang, Y., Zhang, Y., and Yao, S.: Deep Learning for Polarimetric Radar Quantitative Precipitation Estimation during Landfalling Typhoons in South China, Remote Sens.-Basel, 13, 3157, https://doi.org/10.3390/rs13163157, 2021.
Zhang, Y., Long, M., Chen, K., Xing, L., Jin, R., Jordan, M. I., and Wang, J.: Skilful nowcasting of extreme precipitation with NowcastNet, Nature, 619, 526–532, https://doi.org/10.1038/s41586-023-06184-4, 2023.
Short summary
To identify the key factors affecting quantitative precipitation nowcasting (QPN) using deep learning (DL), we carried out a comprehensive evaluation and analysis. We compared four key factors: DL model, length of the input sequence, loss function, and ensemble approach. Generally, U-Net outperformed ConvLSTM. Loss function and ensemble showed potential for improving performance when they synergized well. The length of the input sequence did not significantly affect the results.
To identify the key factors affecting quantitative precipitation nowcasting (QPN) using deep...
