Article
Statistics & Probability
Michael Zamo, Liliane Bel, Olivier Mestre
Summary: This study focuses on combining forecasts from different numerical weather prediction models using expert advice theory to improve forecast performance. Additionally, it explores the use of two forecast performance criteria, CRPS and the Jolliffe-Primo test, to achieve reliable and skillful probabilistic forecasts.
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES C-APPLIED STATISTICS
(2021)
Article
Astronomy & Astrophysics
H. A. Elliott, C. N. Arge, C. J. Henney, M. A. Dayeh, G. Livadiotis, J. -M Jahn, C. E. DeForest
Summary: We analyze the residual errors in Wang-Sheeley-Arge (WSA) solar wind speed forecasts and find the systematic relationship between the residual speed errors and the photospheric magnetic field expansion factor and the minimum separation angle in the photosphere. By using these residual error maps, we apply corrections to the model speeds and test this correction approach using 3-day lead time speed forecasts. The improved accuracy of solar wind speed forecasts enables the prediction of multiday forecasts of various parameters, expanding the usefulness of the WSA forecasts for space weather clients.
SPACE WEATHER-THE INTERNATIONAL JOURNAL OF RESEARCH AND APPLICATIONS
(2022)
Article
Chemistry, Analytical
Rongnian Tang, Yuke Ning, Chuang Li, Wen Feng, Youlong Chen, Xiaofeng Xie
Summary: Achieving high-performance numerical weather prediction is crucial for livelihoods and socioeconomic development. A novel spatial LightGBM model is proposed to correct the numerical forecast results at each observation station, incorporating local spatial information and utilizing a specific strategy for high-performance correction.
Article
Chemistry, Physical
Ayse Tugba Dosdogru, Asli Boru Ipek
Summary: Energy sources are crucial for national economic growth, with wind energy playing a significant role in low-carbon energy technologies. This study focuses on improving wind speed prediction by utilizing a hybrid approach of XGBoost, AdaBoost, and ANN, aiming to provide more accurate results for wind speed forecasting.
INTERNATIONAL JOURNAL OF HYDROGEN ENERGY
(2022)
Article
Agronomy
Muhammad Tousif Bhatti, Arif A. Anwar
Summary: This paper presents a case study of developing a decision support service for farmers in Pakistan based on rainfall forecast data. The study analyzes the statistical verification of 16-day rainfall forecast data and explores the process of developing such a service using global weather forecast data. The findings show variations in the quality of the forecast across different stations.
AGRICULTURAL AND FOREST METEOROLOGY
(2022)
Article
Computer Science, Artificial Intelligence
Andrei Konstantinov, Lev Utkin
Summary: This method proposes a way to interpret black-box models locally and globally based on ensemble gradient boosting machines, using simple decision tree structures and the Lasso method for weight calculation and update. Compared to the neural additive model, it provides a more intuitive and easy-to-train approach.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Environmental Sciences
Xueliang Zhao, Qilong Sun, Wanru Tang, Shuang Yu, Boyu Wang
Summary: Wind speed forecasting is critical in various fields, but conventional methods lack accuracy and need post-processing. This paper applies deep learning algorithms for error correction in wind speed prediction, compared with time-series prediction methods. Experimental results show that deep learning methods can improve accuracy without modeling.
FRONTIERS IN ENVIRONMENTAL SCIENCE
(2022)
Article
Construction & Building Technology
Yiming Liu, Lin Liu, Liu Yang, Li Hao, Yi Bao
Summary: This paper introduces a method to measure distance using ultra-wideband radio technology, machine learning, and error mitigation to improve accuracy. In-situ measurement demonstrated sub-millimeter accuracy, revealing a tradeoff between measurement accuracy and frequency. This study is expected to enhance the capability of measuring distance in automation processes for construction and operation of engineering structures.
AUTOMATION IN CONSTRUCTION
(2021)
Article
Engineering, Multidisciplinary
Hongfei Ma, Wenqi Zhao, Yurong Zhao, Yu He
Summary: Accurate prediction of monthly oil and gas production is crucial for oil enterprises to plan production, avoid blind investment, and achieve sustainable development. This paper predicts initial single-layer production by utilizing the data-driven artificial intelligence algorithm GBDT, considering geological data, fluid PVT data, and well data. The results demonstrate that the GBDT algorithm has high accuracy, improves efficiency significantly, and has universal applicability. The trained GBDT method in this study can provide helpful predictions for well site optimization, perforation layer optimization, and engineering parameter optimization, thus guiding oilfield development.
CMES-COMPUTER MODELING IN ENGINEERING & SCIENCES
(2023)
Article
Mathematics
Luca Di Persio, Nicola Fraccarolo
Summary: In recent years, there has been a growing interest in developing accurate and efficient forecasting methods for energy production and consumption. Traditional linear approaches are insufficient in modeling the relationships between variables, especially when dealing with multiple features. This study proposes a Gradient-Boosting-Machine-based framework to forecast the demand of customers in different locations within the Italian electricity market. The main challenge is to provide precise one-day-ahead predictions using historical data that is two months old, which requires incorporating exogenous regressors and tailoring them to the specific case. Numerical simulations demonstrate that the Gradient Boosting method outperforms classical statistical models such as ARMA, particularly in capturing holidays.
Article
Geochemistry & Geophysics
Ali Danandeh Mehr
Summary: This study compared the classification and prediction capabilities of decision tree, genetic programming, and gradient boosting decision tree techniques for forecasting precipitation and evaporation indexes. GBT showed promising performance in Ankara, while GP was more accurate in central Antalya. The GP model with a scaled sigmoid function demonstrated ease in classifying and predicting different events in the case studies.
Article
Thermodynamics
Gabriele Casciaro, Francesco Ferrari, Daniele Lagomarsino-Oneto, Andrea Lira-Loarca, Andrea Mazzino
Summary: This study proposes a strategy to combine quasi-real time observed wind speed and weather model predictions using Ensemble Model Output Statistics, in order to fill the time gap between consecutive model runs and provide accurate predictions.
Article
Mathematical & Computational Biology
Yunyun Liang, Shengli Zhang, Huijuan Qiao, Yinan Cheng
Summary: In this study, a model named iEnhancer-MFGBDT was developed to identify enhancers and their strength by fusing multiple features and gradient boosting decision tree. The model achieved accuracies of 78.67% and 66.04% for identifying enhancers and their strength on the benchmark dataset, demonstrating its usefulness and effectiveness as an intelligent tool for enhancer identification.
MATHEMATICAL BIOSCIENCES AND ENGINEERING
(2021)
Article
Chemistry, Multidisciplinary
Xue Zhao, Xiaohui Li, Shuang Sun, Xu Jia
Summary: In recent years, federated GBDTs have replaced traditional GBDTs and become the focus of academic research for structured data mining. This paper proposes a secure and efficient FL algorithm for GBDTs (SeFB) based on horizontal federated learning, addressing the issues of information leakage, model accuracy, and communication cost. Experimental analysis shows that the algorithm protects data privacy and reduces communication cost effectively.
APPLIED SCIENCES-BASEL
(2023)
Article
Meteorology & Atmospheric Sciences
Sam Allen, Gavin R. Evans, Piers Buchanan, Frank Kwasniok
Summary: The study found that a regime-dependent mixture model based on the North Atlantic Oscillation can substantially improve wind speed forecasts compared to traditional post-processing methods, but the model's complexity may impact forecast performance. A measure of regime dependency was defined to differentiate situations when numerical model output benefits from regime-dependent post-processing, leading to further improvements in predictive performance and more accurate forecasts of extreme wind speeds when implemented based on a certain threshold value.
QUARTERLY JOURNAL OF THE ROYAL METEOROLOGICAL SOCIETY
(2021)
Article
Environmental Sciences
Wenqing Xu, Like Ning, Yong Luo
Article
Meteorology & Atmospheric Sciences
Zhizhen Wang, Chesheng Zhan, Like Ning, Hai Guo
Summary: This study evaluated the global terrestrial ET of CMIP6 models and found that no single model could perform optimally in all aspects of the comparison. The performance of the CMIP6 ensemble was better than that of individual models, but most models and the ensemble overestimated ET. The uncertainty of the CMIP6 ensemble was generally low, and the estimation reliability varied by geographical region.
THEORETICAL AND APPLIED CLIMATOLOGY
(2021)
Article
Astronomy & Astrophysics
Lei Huang, Tammo S. Steenhuis, Yong Luo, Qiuhong Tang, Ronglin Tang, Junqing Zheng, Wen Shi, Chen Qiao
Summary: This study compared various algorithms for estimating evapotranspiration and found that the original method had higher accuracy but some errors due to compensating inaccuracies. Six of the proposed algorithms underestimated daily ET, mainly because of inaccurate estimation of daily net radiation. One algorithm, using observed flux tower data for daily net radiation instead of estimated values, successfully calculated EF and ET with low relative errors.
EARTH AND SPACE SCIENCE
(2021)
Article
Engineering, Civil
Haoyue Zhang, Chesheng Zhan, Jun Xia, Pat J-F Yeh, Like Ning, Shi Hu, Xu-Sheng Wang
Summary: Groundwater affects water and carbon cycles by providing moisture to plants in semi-arid and arid regions. This study found that in semi-arid areas, WUE increases with decreasing GD due to improved water availability, mainly regulated by biological processes, while in arid zones, WUE is insensitive to GD changes, primarily controlled by physical processes. During drought, groundwater-independent vegetation in arid zones can also access groundwater.
JOURNAL OF HYDROLOGY
(2022)
Article
Geosciences, Multidisciplinary
Shaofei Jin, Ziyan Zheng, Like Ning
Summary: The study reveals that climate change and human activities have significant impacts on runoff in Beijing, with human activities contributing more to the decline in runoff compared to climate change. Economic factors, changes in arable land area, and urbanization rate are identified as major contributing factors to the variation in runoff caused by human activities.
PHYSICS AND CHEMISTRY OF THE EARTH
(2021)
Article
Water Resources
Zhonghe Li, Chesheng Zhan, Shi Hu, Like Ning, Lanfang Wu, Hai Guo
Summary: Multimodel ensembles provide a powerful tool for evaluating agricultural production. This study evaluates the performance of nine global gridded crop models in simulating the yields of four major crops in China. The results show that the models better simulate maize yields compared to other crops in several regions. Wheat has the largest yield gap, mainly due to nutritional stress in the northwest region. Optimizing nitrogen management in wheat production can effectively mitigate the negative impact of climate change on crop production in northwest China.
HYDROLOGY RESEARCH
(2022)
Article
Geosciences, Multidisciplinary
Ziyan Zheng, Like Ning, Danqiong Dai, Liang Chen, Yongli Wang, Zhuguo Ma, Zong-Liang Yang, Chesheng Zhan
Summary: Water resources in the Haihe River Basin in North China have been a major concern. Recent studies have shown that while precipitation has increased since 2000, evapotranspiration has also risen due to warming temperatures. This has resulted in significant stress on water resources in the region, particularly in the plain area. Human water withdrawals have led to severe depletion of groundwater storage, and water resource vulnerability is high in subbasins with large populations. Projections indicate that water storage depletion will continue in the future, highlighting the need for a water crisis response system and sustainable water development.
PHYSICS AND CHEMISTRY OF THE EARTH
(2022)
Article
Meteorology & Atmospheric Sciences
Wei Cheng, Douglas G. MacMartin, Ben Kravitz, Daniele Visioni, Ewa M. Bednarz, Yangyang Xu, Yong Luo, Lei Huang, Yongyun Hu, Paul W. Staten, Peter Hitchcock, John C. Moore, Anboyu Guo, Xiangzheng Deng
Summary: This study examines the effects of stratospheric aerosol geoengineering on the Hadley circulation intensity and the intertropical convergence zone. The results show that the geoengineering has a minor impact on the HC intensity in the northern hemisphere winter, but significantly weakens it in the southern hemisphere winter. Additionally, the geoengineering leads to a southward migration of the ITCZ. Further research can explore strategies to reduce the residual changes caused by geoengineering on HC intensity and ITCZ shifts.
NPJ CLIMATE AND ATMOSPHERIC SCIENCE
(2022)
Article
Meteorology & Atmospheric Sciences
Hai Guo, Chesheng Zhan, Like Ning, Zhonghe Li, Shi Hu
Summary: This study evaluates and compares the performance of CMIP6 and CMIP5 models in simulating the global and basin-scale runoff. Results show that CMIP6 models have less uncertainty on the global scale compared to CMIP5 models, but there is not significant progress on the basin scale. The overestimation of runoff is more prominent in arid and semi-arid areas. Furthermore, the CMIP6 multi-model ensemble mean performs better than individual models, and there are differences in trends and PBIAS between the reference datasets.
THEORETICAL AND APPLIED CLIMATOLOGY
(2022)
Article
Geosciences, Multidisciplinary
Danyi Sun, Wenyu Huang, Yong Luo, Jingjia Luo, Jonathon S. Wright, Haohuan Fu, Bin Wang
Summary: This study combines the numerical wave model and deep learning with the BU-Net model to accurately predict the significant wave height (SWH) of the Northwest Pacific Ocean. The results show promising improvements in forecasting accuracy and performance, especially during typhoon passages.
GEOPHYSICAL RESEARCH LETTERS
(2022)
Article
Environmental Sciences
Xin Zhang, Lihua Zhou, Xingying Zhang, Yong Luo, Lei Sun
Summary: The East Asian summer monsoon had a significant impact on surface ozone in China in 2018. The strong monsoon led to a decrease in surface ozone in the northeast and eastern coast of China, while most other regions experienced an increase in ozone levels.
Article
Geosciences, Multidisciplinary
Danyi Sun, Wenyu Huang, Zifan Yang, Yong Luo, Jingjia Luo, Jonathon S. Wright, Haohuan Fu, Bin Wang
Summary: Wintertime precipitation, especially snowstorms, has a significant impact on people's lives, but the current forecast skill for wintertime precipitation is still low. By combining data augmentation (DA) and deep learning, a DABU-Net model is proposed to improve the forecast accuracy of wintertime precipitation in southeastern China. Three independent models are built for forecast lead times of 24, 48, and 72 hours, and the DABU-Net model reduces the root mean squared errors (RMSEs) of wintertime precipitation at these lead times by 19.08%, 25.00%, and 22.37% respectively. The threat scores (TS) are also significantly increased at various thresholds for the three lead times. Moreover, during heavy precipitation days, the RMSEs are decreased by 14% and TS is increased by 7% within 48 hours lead time. This highlights the great prospects of combining DA and deep learning in precipitation forecasting.
GEOPHYSICAL RESEARCH LETTERS
(2023)
Article
Astronomy & Astrophysics
Lei Huang, Yong Luo, Tammo Steenhuis, Qiuhong Tang, Wei Cheng, Wen Shi, Xin Xia, Dingchi Zhao, Zhouyi Liao
Summary: Evapotranspiration (ET) is an important process that regulates heat and water transfer between land and the atmosphere. The VISEA(2023) model improves the accuracy of net radiation and ET by incorporating an atmosphere emissivity model and correcting the calculation of downward long-wave radiation on cloudy days. It demonstrates good agreement with measurements at seven ChinaFlux sites and outperforms other ET models in terms of accuracy and robustness.
EARTH AND SPACE SCIENCE
(2023)