Journal of Atmospheric and Solar-Terrestrial Physics最新文献_第8页

Stacked hybridization of deep learning model with grey wolf optimization for accurate and explainable reference evapotranspiration 深度学习模型与灰狼优化的叠加杂交，以获得准确和可解释的参考蒸散发

IF 1.9 4区地球科学 Q3 GEOCHEMISTRY & GEOPHYSICS

Journal of Atmospheric and Solar-Terrestrial Physics

Pub Date : 2025-10-08 DOI: 10.1016/j.jastp.2025.106655

Truptimayee Suna , Bibhuti Bhusan Sahoo , Dipali Pawar , Nand Lal Kushwaha , Pradosh Kumar Paramaguru , P.S. Brahmanand , Himani Bisht

Accurate estimation of reference evapotranspiration (ET₀) is essential for effective irrigation scheduling and water resource management, particularly in data-scarce regions such as India, which lack advanced automatic meteorological stations. The present study developed a hybrid model (DNN-GWO) and conducted an in-depth evaluation against standalone data-driven models, including Random Forest (RF), Support Vector Machine (SVM), Deep Neural Network (DNN), Recurrent Neural Network (RNN), and Deep Belief Network (DBN) for forecasting monthly ET₀ in the Upper Ganga canal command region, Uttar Pradesh, India. Three input scenarios were evaluated for their correlation to ET₀ estimation. The results revealed that the DNN model showed the best performance in all three scenarios, achieving R² = 0.958, RMSE = 0.076 mm/day, NSE = 0.954, RMSLE = 0.024, MAE = 0.055, MBE = 0.012, MSRE = 0.032, and EVS = 0.987 with solar radiation (Rs), wind speed (U), maximum temperature (Tmax), minimum temperature (Tmin), and relative humidity (RH) as inputs. The developed hybrid DNN-GWO model further improved predictive accuracy, with R² = 0.992, RMSE = 0.0317 mm/day, NSE = 0.99, RMSLE = 0.023, MAE = 0.054, MBE = 0.018, and EVS = 0.992, reducing RMSE by nearly 60 % compared to the best-performing standalone DNN. SHapley Additive explanations (SHAP) analysis revealed that temperature and solar radiation were the most influential predictors of ET₀, while the model also provided stable predictions across different input scenarios, demonstrating robustness in data-limited conditions. The developed hybrid framework, by combining deep learning, swarm intelligence, and explainability, provides a robust, accurate, and interpretable solution for agricultural water management in data-constrained environments.

准确估算参考蒸散发（ET0）对于有效的灌溉调度和水资源管理至关重要，特别是在印度等缺乏先进自动气象站的数据稀缺地区。本研究开发了一个混合模型（DNN- gwo），并对独立的数据驱动模型进行了深入评估，包括随机森林（RF）、支持向量机（SVM）、深度神经网络（DNN）、循环神经网络（RNN）和深度信念网络（DBN），用于预测印度北方邦上恒河运河指挥地区的月度ET0。评估了三种输入情景与ET0估计的相关性。结果表明，以太阳辐射（Rs）、风速(U)、最高温度（Tmax）、最低温度（Tmin）和相对湿度（RH）为输入，DNN模型在3种情景下均表现最佳，R2 = 0.958, RMSE = 0.076 mm/day, NSE = 0.954, RMSLE = 0.024, MAE = 0.055, MBE = 0.012, MSRE = 0.032, EVS = 0.987。所开发的混合DNN- gwo模型进一步提高了预测精度，R2 = 0.992, RMSE = 0.0317 mm/day, NSE = 0.99, RMSLE = 0.023, MAE = 0.054, MBE = 0.018, EVS = 0.992，与表现最好的独立DNN相比，RMSE降低了近60%。SHapley加性解释（SHAP）分析表明，温度和太阳辐射是最具影响力的ET0预测因子，而该模型在不同输入情景下也能提供稳定的预测，在数据有限的条件下表现出鲁棒性。开发的混合框架结合了深度学习、群体智能和可解释性，为数据受限环境下的农业水资源管理提供了强大、准确和可解释的解决方案。

{"title":"Stacked hybridization of deep learning model with grey wolf optimization for accurate and explainable reference evapotranspiration","authors":"Truptimayee Suna , Bibhuti Bhusan Sahoo , Dipali Pawar , Nand Lal Kushwaha , Pradosh Kumar Paramaguru , P.S. Brahmanand , Himani Bisht","doi":"10.1016/j.jastp.2025.106655","DOIUrl":"10.1016/j.jastp.2025.106655","url":null,"abstract":"<div><div>Accurate estimation of reference evapotranspiration (ET<sub>0</sub>) is essential for effective irrigation scheduling and water resource management, particularly in data-scarce regions such as India, which lack advanced automatic meteorological stations. The present study developed a hybrid model (DNN-GWO) and conducted an in-depth evaluation against standalone data-driven models, including Random Forest (RF), Support Vector Machine (SVM), Deep Neural Network (DNN), Recurrent Neural Network (RNN), and Deep Belief Network (DBN) for forecasting monthly ET<sub>0</sub> in the Upper Ganga canal command region, Uttar Pradesh, India. Three input scenarios were evaluated for their correlation to ET<sub>0</sub> estimation. The results revealed that the DNN model showed the best performance in all three scenarios, achieving R<sup>2</sup> = 0.958, RMSE = 0.076 mm/day, NSE = 0.954, RMSLE = 0.024, MAE = 0.055, MBE = 0.012, MSRE = 0.032, and EVS = 0.987 with solar radiation (Rs), wind speed (U), maximum temperature (Tmax), minimum temperature (Tmin), and relative humidity (RH) as inputs. The developed hybrid DNN-GWO model further improved predictive accuracy, with R<sup>2</sup> = 0.992, RMSE = 0.0317 mm/day, NSE = 0.99, RMSLE = 0.023, MAE = 0.054, MBE = 0.018, and EVS = 0.992, reducing RMSE by nearly 60 % compared to the best-performing standalone DNN. SHapley Additive explanations (SHAP) analysis revealed that temperature and solar radiation were the most influential predictors of ET<sub>0</sub>, while the model also provided stable predictions across different input scenarios, demonstrating robustness in data-limited conditions. The developed hybrid framework, by combining deep learning, swarm intelligence, and explainability, provides a robust, accurate, and interpretable solution for agricultural water management in data-constrained environments.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"277 ","pages":"Article 106655"},"PeriodicalIF":1.9,"publicationDate":"2025-10-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145324924","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Intraseasonal and interannual variability of the Total Ozone Column (TCO) in Brazilian biomes: An analysis of the multibiome trend in Mato Grosso do Sul - Brazil 巴西生物群系中总臭氧柱（TCO）的季节内和年际变化：巴西南马托格罗索州多生物群系趋势分析

IF 1.9 4区地球科学 Q3 GEOCHEMISTRY & GEOPHYSICS

Journal of Atmospheric and Solar-Terrestrial Physics

Pub Date : 2025-10-06 DOI: 10.1016/j.jastp.2025.106651

Mutambi Songa , Amaury de Souza , José Francisco de Oliveira-Júnior , Kelvy Rosalvo Alencar Cardoso , Sneha Gautam

The Total Ozone Column (TCO) is a key atmospheric indicator for shielding against ultraviolet (UV) radiation. Although global recovery of the ozone layer has been underway since the Montreal Protocol, uncertainties persist in tropical regions. This study analyzes the spatio-temporal variability and trends of TCO in three major Brazilian biomes—Cerrado, Pantanal, and Atlantic Forest—located in Mato Grosso do Sul (MS), during the period 2005–2020. Satellite data from the OMI/Aura Level-3 product (NASA) with 1° × 1° resolution were processed into monthly, seasonal, and annual averages, applying a 13-month centered moving average for trend visualization. Trend detection was carried out using linear regression and the Mann–Kendall test. Results showed that the Cerrado had the highest average TCO (346.66 DU) and the only statistically significant positive trend (+0.031 DU.year⁻¹, p-value = 0.012), while the Pantanal and Atlantic Forest exhibited lower averages (∼261 DU) and occasional drops below the 260 DU threshold established by the World Meteorological Organization (WMO). A marked seasonality was identified, with maxima between September and October (spring) and minima between January and March (summer). Interannual variability was strongly influenced by El Niño events and biomass burning. These findings highlight that TCO variability is shaped by the interaction of global atmospheric circulation and local anthropogenic pressures, emphasizing the need for biome-specific monitoring and public policies to reduce UV exposure and adapt to climate change, in alignment with the Montreal Protocol and the 2030 Agenda.

总臭氧柱（TCO）是一个关键的大气指标屏蔽紫外线（UV）辐射。尽管自《蒙特利尔议定书》以来，全球臭氧层的恢复一直在进行，但热带地区的不确定性仍然存在。本文分析了2005-2020年巴西南马托格罗索州塞拉多、潘塔纳尔和大西洋森林3个主要生物群落TCO的时空变化及其趋势。来自OMI/Aura Level-3产品（NASA）的1°× 1°分辨率的卫星数据被处理成月、季节和年平均值，应用13个月为中心的移动平均线进行趋势可视化。采用线性回归和Mann-Kendall检验进行趋势检测。结果塞拉多地区平均TCO最高（346.66 DU），唯一有统计学意义的上升趋势（+0.031 DU）。年−1，p值= 0.012)，而潘塔纳尔和大西洋森林表现出较低的平均值（~ 261 DU），偶尔低于世界气象组织（WMO）确定的260 DU阈值。具有明显的季节性，峰值在9月至10月（春季），最小值在1月至3月（夏季）。年际变率受El Niño事件和生物质燃烧的强烈影响。这些发现突出表明，总碳含量变异是由全球大气环流和当地人为压力的相互作用形成的，强调需要根据《蒙特利尔议定书》和《2030年可持续发展议程》开展针对生物群落的监测和公共政策，以减少紫外线照射并适应气候变化。

{"title":"Intraseasonal and interannual variability of the Total Ozone Column (TCO) in Brazilian biomes: An analysis of the multibiome trend in Mato Grosso do Sul - Brazil","authors":"Mutambi Songa , Amaury de Souza , José Francisco de Oliveira-Júnior , Kelvy Rosalvo Alencar Cardoso , Sneha Gautam","doi":"10.1016/j.jastp.2025.106651","DOIUrl":"10.1016/j.jastp.2025.106651","url":null,"abstract":"<div><div>The Total Ozone Column (TCO) is a key atmospheric indicator for shielding against ultraviolet (UV) radiation. Although global recovery of the ozone layer has been underway since the Montreal Protocol, uncertainties persist in tropical regions. This study analyzes the spatio-temporal variability and trends of TCO in three major Brazilian biomes—Cerrado, Pantanal, and Atlantic Forest—located in Mato Grosso do Sul (MS), during the period 2005–2020. Satellite data from the OMI/Aura Level-3 product (NASA) with 1° × 1° resolution were processed into monthly, seasonal, and annual averages, applying a 13-month centered moving average for trend visualization. Trend detection was carried out using linear regression and the Mann–Kendall test. Results showed that the Cerrado had the highest average TCO (346.66 DU) and the only statistically significant positive trend (+0.031 DU.year<sup>−1</sup>, p-value = 0.012), while the Pantanal and Atlantic Forest exhibited lower averages (∼261 DU) and occasional drops below the 260 DU threshold established by the World Meteorological Organization (WMO). A marked seasonality was identified, with maxima between September and October (spring) and minima between January and March (summer). Interannual variability was strongly influenced by El Niño events and biomass burning. These findings highlight that TCO variability is shaped by the interaction of global atmospheric circulation and local anthropogenic pressures, emphasizing the need for biome-specific monitoring and public policies to reduce UV exposure and adapt to climate change, in alignment with the Montreal Protocol and the 2030 Agenda.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"277 ","pages":"Article 106651"},"PeriodicalIF":1.9,"publicationDate":"2025-10-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145269012","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

The influence of ERA5 and FNL driven WRF on sensitivity experiments of small-scale underlying surface changes: a case study of poyang lake in China ERA5和FNL驱动WRF对小尺度下垫面变化敏感性试验的影响——以鄱阳湖为例

IF 1.9 4区地球科学 Q3 GEOCHEMISTRY & GEOPHYSICS

Journal of Atmospheric and Solar-Terrestrial Physics

Pub Date : 2025-10-02 DOI: 10.1016/j.jastp.2025.106650

Shanshan Wu , Jing Zheng , Haibo Zou , Hai Wang , Lizhi Tao

ECMWF ERA5 and NCEP FNL are the two most commonly used datasets to drive WRF model, and the former has relatively higher spatiotemporal resolution than the latter. In order to explore the performances of these two datasets driving WRF model to conduct sensitivity experiments of small-scale underlying surface changes, four experiments are designed for the convective event near Poyang Lake (PL) in China on May 4, 2020. These experiments are ERA-Water (ERA5-driven, including the water body of PL), ERA-Cropland (ERA5-driven, the water body of PL is replaced by cropland), FNL-Water (FNL-driven, including the water body of PL) and FNL-Cropland (FNL-driven, the water body of PL is replaced by cropland), respectively. Simulation results show that ERA-Water and FNL-Water successfully reproduce this convection, while ERA-Cropland and FNL-Cropland don't simulate it, reflecting the role of PL in the convection development. However, the simulations of ERA-Cropland and FNL-Cropland differ greatly, and the former simulates a strong convection in the middle and east of the north shore of PL, while the latter did not simulate a convection. Diagnostic analysis indicates that the strong convection in ERA-Cropland is false, and is mainly caused by the mismatch between the high-resolution ERA5 and the changed underlying surface information. This suggests that when conducting sensitivity experiments of small-scale underlying surface changes, it may be more appropriate to use NCEP FNL than ECMWF ERA5.

ECMWF ERA5和NCEP FNL是驱动WRF模型最常用的两个数据集，前者具有相对较高的时空分辨率。为了探索这两个数据集驱动WRF模型进行小尺度下垫面变化敏感性试验的性能，设计了2020年5月4日中国鄱阳湖附近对流事件的4个试验。这些实验分别是ERA-Water （era5驱动，包括PL水体）、ERA-Cropland （era5驱动，PL水体被农田取代）、FNL-Water （fnl驱动，包括PL水体）和FNL-Cropland （fnl驱动，包括PL水体被农田取代）。模拟结果表明，ERA-Water和FNL-Water成功地模拟了这种对流，而ERA-Cropland和FNL-Cropland没有模拟这种对流，反映了PL在对流发展中的作用。而ERA-Cropland和FNL-Cropland的模拟结果差异较大，前者模拟了PL北岸中东部的强对流，后者没有模拟对流。诊断分析表明，ERA-Cropland区的强对流是假的，其主要原因是高分辨率ERA5与变化的下垫面信息不匹配。这表明，在进行小尺度下垫面变化敏感性实验时，使用NCEP FNL可能比使用ECMWF ERA5更合适。

{"title":"The influence of ERA5 and FNL driven WRF on sensitivity experiments of small-scale underlying surface changes: a case study of poyang lake in China","authors":"Shanshan Wu , Jing Zheng , Haibo Zou , Hai Wang , Lizhi Tao","doi":"10.1016/j.jastp.2025.106650","DOIUrl":"10.1016/j.jastp.2025.106650","url":null,"abstract":"<div><div>ECMWF ERA5 and NCEP FNL are the two most commonly used datasets to drive WRF model, and the former has relatively higher spatiotemporal resolution than the latter. In order to explore the performances of these two datasets driving WRF model to conduct sensitivity experiments of small-scale underlying surface changes, four experiments are designed for the convective event near Poyang Lake (PL) in China on May 4, 2020. These experiments are ERA-Water (ERA5-driven, including the water body of PL), ERA-Cropland (ERA5-driven, the water body of PL is replaced by cropland), FNL-Water (FNL-driven, including the water body of PL) and FNL-Cropland (FNL-driven, the water body of PL is replaced by cropland), respectively. Simulation results show that ERA-Water and FNL-Water successfully reproduce this convection, while ERA-Cropland and FNL-Cropland don't simulate it, reflecting the role of PL in the convection development. However, the simulations of ERA-Cropland and FNL-Cropland differ greatly, and the former simulates a strong convection in the middle and east of the north shore of PL, while the latter did not simulate a convection. Diagnostic analysis indicates that the strong convection in ERA-Cropland is false, and is mainly caused by the mismatch between the high-resolution ERA5 and the changed underlying surface information. This suggests that when conducting sensitivity experiments of small-scale underlying surface changes, it may be more appropriate to use NCEP FNL than ECMWF ERA5.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"277 ","pages":"Article 106650"},"PeriodicalIF":1.9,"publicationDate":"2025-10-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145269013","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Pre-seismic ionospheric disturbances (PIDs) associated With 2021 Mw 7.5 Northern Peru earthquake: GNSS and ground uplift observations 与2021 Mw 7.5秘鲁北部地震相关的震前电离层扰动（PIDs）： GNSS和地面隆起观测

IF 1.9 4区地球科学 Q3 GEOCHEMISTRY & GEOPHYSICS

Journal of Atmospheric and Solar-Terrestrial Physics

Pub Date : 2025-09-27 DOI: 10.1016/j.jastp.2025.106644

Oluwasegun M. Adebayo , Esfhan A. Kherani , Alexandre A. Pimenta , Babatunde Rabiu

Predicting natural disasters such as earthquakes remains a major challenge in geosciences, with critical implications for early warning systems and disaster risk reduction. Among various precursory signals, ionospheric anomalies have gained increasing attention as potential indicators of impending seismic events. In this study, we examine pre-seismic ionospheric disturbances (PIDs) associated with the Mw 7.5 Northern Peru earthquake on November 28, 2021, using Global Navigation Satellite System (GNSS) Total Electron Content (TEC) and ground vertical velocity data. Significant sequential negative TEC anomalies were observed up to two hours prior to the mainshock, with multiple disturbances recorded by several GNSS receivers. The amplitudes of these disturbances increased as the earthquake approached, suggesting a progressive ionospheric response to the buildup of tectonic stress. Spectral analysis using the Short-Time Fourier Transform revealed center frequencies between 3.63 mHz and 4.80 mHz — within the acoustic/infrasonic range — indicating that such waves, possibly generated by foreshocks, may be responsible for these PIDs. To rule out other sources of TEC anomalies, we examined the TEC data for the previous day along the same trajectories and found no similar disturbances. Furthermore, geomagnetic conditions were quiet during the period, as indicated by Kp and Dst indexes two days before and after the earthquake. These findings contribute to the understanding of seismo-ionospheric coupling and highlight the potential role of ionospheric monitoring as a complementary approach to conventional seismic methods in earthquake early warning systems.

预测地震等自然灾害仍然是地球科学的一项重大挑战，对早期预警系统和减少灾害风险具有重要意义。在各种前兆信号中，电离层异常作为即将发生的地震事件的潜在指标而受到越来越多的关注。在这项研究中，我们使用全球导航卫星系统（GNSS）总电子含量（TEC）和地面垂直速度数据，研究了与2021年11月28日秘鲁北部7.5 Mw地震相关的震前电离层扰动（ids）。在主震发生前两小时，观测到明显的连续负TEC异常，几个GNSS接收器记录了多次干扰。这些扰动的振幅随着地震的临近而增加，这表明电离层对构造应力的积累有一个渐进的反应。使用短时傅立叶变换的频谱分析显示，中心频率在3.63 mHz和4.80 mHz之间，在声学/次声范围内，表明这种可能由前震产生的波可能是这些PIDs的原因。为了排除TEC异常的其他来源，我们沿着相同的轨迹检查了前一天的TEC数据，没有发现类似的干扰。从地震前后两天的Kp和Dst指数可以看出，这一时期的地磁条件是安静的。这些发现有助于理解地震-电离层耦合，并突出了电离层监测作为地震预警系统中常规地震方法的补充方法的潜在作用。

{"title":"Pre-seismic ionospheric disturbances (PIDs) associated With 2021 Mw 7.5 Northern Peru earthquake: GNSS and ground uplift observations","authors":"Oluwasegun M. Adebayo , Esfhan A. Kherani , Alexandre A. Pimenta , Babatunde Rabiu","doi":"10.1016/j.jastp.2025.106644","DOIUrl":"10.1016/j.jastp.2025.106644","url":null,"abstract":"<div><div>Predicting natural disasters such as earthquakes remains a major challenge in geosciences, with critical implications for early warning systems and disaster risk reduction. Among various precursory signals, ionospheric anomalies have gained increasing attention as potential indicators of impending seismic events. In this study, we examine pre-seismic ionospheric disturbances (PIDs) associated with the Mw 7.5 Northern Peru earthquake on November 28, 2021, using Global Navigation Satellite System (GNSS) Total Electron Content (TEC) and ground vertical velocity data. Significant sequential negative TEC anomalies were observed up to two hours prior to the mainshock, with multiple disturbances recorded by several GNSS receivers. The amplitudes of these disturbances increased as the earthquake approached, suggesting a progressive ionospheric response to the buildup of tectonic stress. Spectral analysis using the Short-Time Fourier Transform revealed center frequencies between 3.63 mHz and 4.80 mHz — within the acoustic/infrasonic range — indicating that such waves, possibly generated by foreshocks, may be responsible for these PIDs. To rule out other sources of TEC anomalies, we examined the TEC data for the previous day along the same trajectories and found no similar disturbances. Furthermore, geomagnetic conditions were quiet during the period, as indicated by Kp and Dst indexes two days before and after the earthquake. These findings contribute to the understanding of seismo-ionospheric coupling and highlight the potential role of ionospheric monitoring as a complementary approach to conventional seismic methods in earthquake early warning systems.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"277 ","pages":"Article 106644"},"PeriodicalIF":1.9,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145221899","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Estimation of ERA5 tropospheric parameters using GNSS data over Tashkent 利用GNSS数据估算塔什干上空ERA5对流层参数

IF 1.9 4区地球科学 Q3 GEOCHEMISTRY & GEOPHYSICS

Journal of Atmospheric and Solar-Terrestrial Physics

Pub Date : 2025-09-27 DOI: 10.1016/j.jastp.2025.106648

H.E. Eshkuvatov , Sh.N. Mardonov , O.V. Xudoynazarov , Z.J. Ruziev , Sh.Sh. Numonjonov , J.R. Hoshimov , F.X. Asatullayev , I.M. Egamberdiev , M.A. Musurmonov

Remote sensing of atmospheric water vapor using Global Navigation Satellite System (GNSS) signals has become an important technique in meteorology, weather forecasting, and climate research. This study investigated regional atmospheric variability over Tashkent, Uzbekistan, by analyzing ten key atmospheric parameters from the ERA5 reanalysis and retrieving precipitable water vapor (PW) from GNSS-derived tropospheric delay data. The analysis covered the period from 12 to 22 February 2025 (day of year 43–53), using ground-based observations from the Tashkent (TASH) and Maidantal (MTAL) GNSS stations. The primary aim was to enhance the characterization of regional atmospheric dynamics and to evaluate the potential of GNSS-derived PW for improving precipitation forecasting when combined with reanalysis data. The results revealed a strong correlation between GNSS-derived and ERA5-derived PW values, indicating that GNSS tropospheric delay observations reliably capture short-term variations in atmospheric water vapor. These findings confirm the utility of integrating GNSS retrievals with reanalysis products for high-resolution monitoring of atmospheric processes in Central Asia.

利用全球导航卫星系统（GNSS）信号遥感大气水汽已成为气象学、天气预报和气候研究中的一项重要技术。利用ERA5再分析的10个关键大气参数和gnss对流层延迟数据反演的可降水量（PW），研究了乌兹别克斯坦塔什干地区的大气变率。分析涵盖了2025年2月12日至22日（43-53年的一天），使用了塔什干（TASH）和迈丹塔尔（MTAL）全球导航卫星系统站的地面观测数据。主要目的是增强区域大气动力学特征，并评估gnss衍生的PW与再分析数据相结合时改善降水预报的潜力。结果显示，GNSS反演的PW值与era5反演的PW值之间存在很强的相关性，表明GNSS对流层延迟观测可靠地捕获了大气水汽的短期变化。这些发现证实了将GNSS检索结果与再分析产品结合起来对中亚大气过程进行高分辨率监测的实用性。

{"title":"Estimation of ERA5 tropospheric parameters using GNSS data over Tashkent","authors":"H.E. Eshkuvatov , Sh.N. Mardonov , O.V. Xudoynazarov , Z.J. Ruziev , Sh.Sh. Numonjonov , J.R. Hoshimov , F.X. Asatullayev , I.M. Egamberdiev , M.A. Musurmonov","doi":"10.1016/j.jastp.2025.106648","DOIUrl":"10.1016/j.jastp.2025.106648","url":null,"abstract":"<div><div>Remote sensing of atmospheric water vapor using Global Navigation Satellite System (GNSS) signals has become an important technique in meteorology, weather forecasting, and climate research. This study investigated regional atmospheric variability over Tashkent, Uzbekistan, by analyzing ten key atmospheric parameters from the ERA5 reanalysis and retrieving precipitable water vapor (PW) from GNSS-derived tropospheric delay data. The analysis covered the period from 12 to 22 February 2025 (day of year 43–53), using ground-based observations from the Tashkent (TASH) and Maidantal (MTAL) GNSS stations. The primary aim was to enhance the characterization of regional atmospheric dynamics and to evaluate the potential of GNSS-derived PW for improving precipitation forecasting when combined with reanalysis data. The results revealed a strong correlation between GNSS-derived and ERA5-derived PW values, indicating that GNSS tropospheric delay observations reliably capture short-term variations in atmospheric water vapor. These findings confirm the utility of integrating GNSS retrievals with reanalysis products for high-resolution monitoring of atmospheric processes in Central Asia.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"277 ","pages":"Article 106648"},"PeriodicalIF":1.9,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145221900","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Identification and distribution of wet and dry season in the 70 most populated cities in India 印度70个人口最多的城市的湿季和旱季的识别和分布

IF 1.9 4区地球科学 Q3 GEOCHEMISTRY & GEOPHYSICS

Journal of Atmospheric and Solar-Terrestrial Physics

Pub Date : 2025-09-26 DOI: 10.1016/j.jastp.2025.106649

Sumanta Dandapath , Abhijit Patil , Dhanashri Suresh Shinde , Praveen Kumar Pathak

In this paper, for the first time, we have attempted to identify the length and distribution of the wet and dry season of the 70 most populated cities in India based on the relative intra-annual distribution of rainfall observed during the last three decades (1991–2020). The wet season in the cities under investigation is rather continuous and usually lasts for 4–8 months. The distribution of the wet months however indicates significant variation, particularly for the cities located in the southern and northern periphery of the country. The cities located in the remaining areas even though have been receiving rainfall primarily during the Indian Summer Monsoon (ISM), the length of the wet season however varies among them. The present study also highlights that the average length of the wet season in India is not always four months; instead, it is a little over 5 months long in general. The length and distribution of the dry season also indicate that most cities receive a very negligible amount of rainfall for more than five months. The average amount of rainfall received by the cities during their respective dry season (4.1 %) is about 22 times lower than the average amount of rainfall received during their respective wet season (90.2 %). Noticeable variation in the amount of annual rainfall, length of dry and wet season, and amount of rainfall received in their respective wet and dry season among cities within each of the four homogeneous regions is also a significant finding of the present work.

在本文中，我们首次尝试根据过去三十年（1991-2020年）观测到的降雨量的相对年内分布，确定印度70个人口最多的城市的干湿季节的长度和分布。被调查城市的雨季相当持续，通常持续4-8个月。然而，湿润月份的分布显示出显著的变化，特别是位于该国南部和北部边缘的城市。位于其余地区的城市虽然主要在印度夏季风（ISM）期间降雨，但雨季的长度在它们之间有所不同。目前的研究还强调，印度雨季的平均长度并不总是四个月；相反，它通常是5个多月的时间。旱季的长度和分布也表明，大多数城市在五个多月的时间里降雨量非常小。各城市在旱季的平均降雨量（4.1%）比其在雨季的平均降雨量（90.2%）低约22倍。在四个均匀区域内的城市之间，年降雨量、干湿季节长度以及各自干湿季节的降雨量的显著变化也是本工作的一个重要发现。

{"title":"Identification and distribution of wet and dry season in the 70 most populated cities in India","authors":"Sumanta Dandapath , Abhijit Patil , Dhanashri Suresh Shinde , Praveen Kumar Pathak","doi":"10.1016/j.jastp.2025.106649","DOIUrl":"10.1016/j.jastp.2025.106649","url":null,"abstract":"<div><div>In this paper, for the first time, we have attempted to identify the length and distribution of the wet and dry season of the 70 most populated cities in India based on the relative intra-annual distribution of rainfall observed during the last three decades (1991–2020). The wet season in the cities under investigation is rather continuous and usually lasts for 4–8 months. The distribution of the wet months however indicates significant variation, particularly for the cities located in the southern and northern periphery of the country. The cities located in the remaining areas even though have been receiving rainfall primarily during the Indian Summer Monsoon (ISM), the length of the wet season however varies among them. The present study also highlights that the average length of the wet season in India is not always four months; instead, it is a little over 5 months long in general. The length and distribution of the dry season also indicate that most cities receive a very negligible amount of rainfall for more than five months. The average amount of rainfall received by the cities during their respective dry season (4.1 %) is about 22 times lower than the average amount of rainfall received during their respective wet season (90.2 %). Noticeable variation in the amount of annual rainfall, length of dry and wet season, and amount of rainfall received in their respective wet and dry season among cities within each of the four homogeneous regions is also a significant finding of the present work.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"277 ","pages":"Article 106649"},"PeriodicalIF":1.9,"publicationDate":"2025-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145221941","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Dynamics of atmospheric temperature inversions in Dammam, Saudi Arabia: Long-term characterization and trends 达曼，沙特阿拉伯的大气温度逆温动力学：长期特征和趋势

IF 1.9 4区地球科学 Q3 GEOCHEMISTRY & GEOPHYSICS

Journal of Atmospheric and Solar-Terrestrial Physics

Pub Date : 2025-09-24 DOI: 10.1016/j.jastp.2025.106645

Abdullrahman Maghrabi, Abdulah Al-Dosari, Mohammed Altlasi, Abdulah Alsherhri, Maohammed Almutairi

This study investigates the temporal variability of temperature inversions below 5000 m in Dammam, Saudi Arabia(26.4°N, 50.1°E), over a 38-year period from 1985 to 2023, using radiosonde data to analyze six critical variables: base height (H_i), maximum height (H_f), temperature at the base (T_i), temperature at the maximum height (T_f), height difference (DH), and temperature difference (DT). Considering the temperature inversions occurred below 5000 m and with DT greater than 1 C, a total of 13744 temperature inversion events were recognized and investigated. Analysis revealed predominant low-level inversions (76.07 % below 1000 m) with strong nocturnal occurrence (∼70 %). High-level inversions showed seasonal variations in H_i (1900 ± 100 m in winter, 2500 ± 100 m in summer) and T_i (5.98 ± 2.0 °C in January, 22.30 ± 2.0 °C in June), with H_f consistently 200–300 m higher and Tf 2–3 °C warmer. DH ranged annually between 170 ± 50 m and 220 ± 50 m, with nocturnal low-level inversions showing the largest stable vertical extent (296 ± 23 m). Inversion frequencies increased over time, particularly at higher altitudes, with robust trends confirmed by Mann-Kendall and regression analyses (91.7 % concordance). These patterns, driven by radiative cooling, sea-land breezes, and urban heat island effects, suggest significant implications for air quality and urban planning in Dammam's coastal desert environment.

本文研究了1985年至2023年38年间沙特阿拉伯达曼（26.4°N, 50.1°E） 5000 m以下气温的时间变率，利用探空数据分析了6个关键变量：基线高度（Hi）、最高高度（Hf）、基线温度（Ti）、最高高度温度（Tf）、高差（DH）和温差（DT）。考虑到发生在5000 m以下且DT大于1 C的逆温事件，共识别和研究了13744个逆温事件。分析显示主要的低空逆温（76.07%在1000米以下）和强烈的夜间发生（约70%）。高空逆温呈现出Hi（冬季1900±100 m，夏季2500±100 m）和Ti（1月5.98±2.0°C， 6月22.30±2.0°C）的季节变化，Hf持续升高200-300 m， Tf持续升高2-3°C。年DH在170±50 ~ 220±50 m之间，夜间低空逆温显示最大的稳定垂直范围（296±23 m）。逆温频率随着时间的推移而增加，特别是在高海拔地区，Mann-Kendall和回归分析证实了这种强劲的趋势（91.7%的一致性）。这些模式由辐射冷却、海陆风和城市热岛效应驱动，对达曼沿海沙漠环境的空气质量和城市规划具有重要意义。

{"title":"Dynamics of atmospheric temperature inversions in Dammam, Saudi Arabia: Long-term characterization and trends","authors":"Abdullrahman Maghrabi, Abdulah Al-Dosari, Mohammed Altlasi, Abdulah Alsherhri, Maohammed Almutairi","doi":"10.1016/j.jastp.2025.106645","DOIUrl":"10.1016/j.jastp.2025.106645","url":null,"abstract":"<div><div>This study investigates the temporal variability of temperature inversions below 5000 m in Dammam, Saudi Arabia(26.4°N, 50.1°E), over a 38-year period from 1985 to 2023, using radiosonde data to analyze six critical variables: base height (H<sub>i</sub>), maximum height (H<sub>f</sub>), temperature at the base (T<sub>i</sub>), temperature at the maximum height (T<sub>f</sub>), height difference (DH), and temperature difference (DT). Considering the temperature inversions occurred below 5000 m and with DT greater than 1 C, a total of 13744 temperature inversion events were recognized and investigated. Analysis revealed predominant low-level inversions (76.07 % below 1000 m) with strong nocturnal occurrence (∼70 %). High-level inversions showed seasonal variations in H<sub>i</sub> (1900 ± 100 m in winter, 2500 ± 100 m in summer) and T<sub>i</sub> (5.98 ± 2.0 °C in January, 22.30 ± 2.0 °C in June), with H<sub>f</sub> consistently 200–300 m higher and Tf 2–3 °C warmer. DH ranged annually between 170 ± 50 m and 220 ± 50 m, with nocturnal low-level inversions showing the largest stable vertical extent (296 ± 23 m). Inversion frequencies increased over time, particularly at higher altitudes, with robust trends confirmed by Mann-Kendall and regression analyses (91.7 % concordance). These patterns, driven by radiative cooling, sea-land breezes, and urban heat island effects, suggest significant implications for air quality and urban planning in Dammam's coastal desert environment.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"277 ","pages":"Article 106645"},"PeriodicalIF":1.9,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145156582","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Drought and flood evolution characteristics during winter rapeseed season based on the Z-index in Hubei Province, China 基于z指数的湖北省冬季油菜籽季旱涝演变特征

IF 1.9 4区地球科学 Q3 GEOCHEMISTRY & GEOPHYSICS

Journal of Atmospheric and Solar-Terrestrial Physics

Pub Date : 2025-09-22 DOI: 10.1016/j.jastp.2025.106646

Yaotian Tian, Enhao Zhang, Yongyuan Huang, Ming Huang, Haoran Shi, Hui Chen

Analyzing drought and flood characteristics is of great significance to ensure production security, disaster prevention, and mitigation. Hubei Province, as a major rapeseed-producing area in China, is crucial for the nation's edible oil supply. However, there were no studies on drought and flood characteristics during rapeseed season in Hubei Province. To study the drought and flood evolution, daily meteorological data from 28 surface weather stations in Hubei Province during the rapeseed season from 1960 to 2019 was adopted to calculate the Z-index. The Mann-Kendall trend test and wavelet analysis were employed. The results indicated that precipitation at the seedling, flowering, ripening, and whole growth stages of rapeseed generally decreased by 4.79, 2.14, 2.58, and 7.38 mm (10 yr)⁻¹, respectively. Precipitation at the budding stage of rapeseed increased by 1.74 mm (10 yr)⁻¹. Based on the Mann-Kendall trend test, the mutation of precipitation at the seedling, budding, flowering, ripening, and whole growth stages of rapeseed began in 1987, 1983, 1961, 1960, and 1968, respectively. Precipitation at the seedling, ripening, and whole growth stages decreased from south to north across Hubei Province. From southeast to northwest of Hubei Province, precipitation at the budding and flowering stages decreased. A shift from drought to flood was observed at the budding stage, whereas a shift from flood to drought occurred at the other stages. Severe drought (15.0 %) and mild drought (20 %) occurred most frequently at the flowering stage. Severe flood occurred most often at the budding stage (16.7 %), moderate flood at the ripening stage (13.3 %), and mild flood at the seedling stage (13.3 %). According to the wavelet analysis, the cycles of drought and flood disasters at the seedling, budding, flowering, ripening, and whole growth stages were 28, 23, 14, 18, and 7 years, respectively. This study provides a scientific basis for predicting drought and flood disasters and guiding rapeseed production in Hubei Province.

分析旱涝特征对保障生产安全、防灾减灾具有重要意义。湖北省作为中国主要的油菜籽产区，对全国的食用油供应至关重要。但目前尚无对湖北省油菜季旱涝特征的研究。利用1960 - 2019年湖北省28个地面气象站的油菜籽季逐日气象资料，计算z指数，研究旱涝演变。采用Mann-Kendall趋势检验和小波分析。结果表明：油菜籽苗期、开花期、成熟期和全生育期降水总体上分别减少4.79、2.14、2.58和7.38 mm (10 yr)−1；油菜出芽期降水量增加1.74 mm (10 yr)−1。根据Mann-Kendall趋势检验，油菜苗期、出芽期、开花期、成熟期和全生育期降水分别在1987年、1983年、1961年、1960年和1968年发生突变。苗期、成熟期和全生育期降水量由南向北递减。湖北从东南向西北，出芽期和开花期降水呈减少趋势。出芽期由干旱向洪水转变，其他阶段由洪水向干旱转变。重度干旱（15.0%）和轻度干旱（20%）发生在花期。出芽期多发生重度洪水（16.7%），成熟期多发生中度洪水（13.3%），苗期多发生轻度洪水（13.3%）。根据小波分析，苗期、出芽期、花期、成熟期和全生育期的旱涝灾害周期分别为28年、23年、14年、18年和7年。该研究为预测湖北省旱涝灾害和指导油菜籽生产提供了科学依据。

{"title":"Drought and flood evolution characteristics during winter rapeseed season based on the Z-index in Hubei Province, China","authors":"Yaotian Tian, Enhao Zhang, Yongyuan Huang, Ming Huang, Haoran Shi, Hui Chen","doi":"10.1016/j.jastp.2025.106646","DOIUrl":"10.1016/j.jastp.2025.106646","url":null,"abstract":"<div><div>Analyzing drought and flood characteristics is of great significance to ensure production security, disaster prevention, and mitigation. Hubei Province, as a major rapeseed-producing area in China, is crucial for the nation's edible oil supply. However, there were no studies on drought and flood characteristics during rapeseed season in Hubei Province. To study the drought and flood evolution, daily meteorological data from 28 surface weather stations in Hubei Province during the rapeseed season from 1960 to 2019 was adopted to calculate the Z-index. The Mann-Kendall trend test and wavelet analysis were employed. The results indicated that precipitation at the seedling, flowering, ripening, and whole growth stages of rapeseed generally decreased by 4.79, 2.14, 2.58, and 7.38 mm (10 yr)<sup>−1</sup>, respectively. Precipitation at the budding stage of rapeseed increased by 1.74 mm (10 yr)<sup>−1</sup>. Based on the Mann-Kendall trend test, the mutation of precipitation at the seedling, budding, flowering, ripening, and whole growth stages of rapeseed began in 1987, 1983, 1961, 1960, and 1968, respectively. Precipitation at the seedling, ripening, and whole growth stages decreased from south to north across Hubei Province. From southeast to northwest of Hubei Province, precipitation at the budding and flowering stages decreased. A shift from drought to flood was observed at the budding stage, whereas a shift from flood to drought occurred at the other stages. Severe drought (15.0 %) and mild drought (20 %) occurred most frequently at the flowering stage. Severe flood occurred most often at the budding stage (16.7 %), moderate flood at the ripening stage (13.3 %), and mild flood at the seedling stage (13.3 %). According to the wavelet analysis, the cycles of drought and flood disasters at the seedling, budding, flowering, ripening, and whole growth stages were 28, 23, 14, 18, and 7 years, respectively. This study provides a scientific basis for predicting drought and flood disasters and guiding rapeseed production in Hubei Province.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"277 ","pages":"Article 106646"},"PeriodicalIF":1.9,"publicationDate":"2025-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145156447","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Time series analysis of the impact of global warming on Türkiye 全球变暖对土壤影响的时间序列分析

IF 1.9 4区地球科学 Q3 GEOCHEMISTRY & GEOPHYSICS

Journal of Atmospheric and Solar-Terrestrial Physics

Pub Date : 2025-09-22 DOI: 10.1016/j.jastp.2025.106647

Arif Ozbek, Mehmet Bilgili

According to the assessments of the Intergovernmental Panel on Climate Change (IPCC), Türkiye, located within the Mediterranean basin, is among the regions most susceptible to the adverse impacts of climate change. This heightened vulnerability is largely attributed to its geographic location, climatic characteristics, and socio-economic structure, which together amplify the risks associated with rising temperatures and increasing climate variability. In the present study, monthly mean air temperature data for Türkiye, recorded by the Turkish State Meteorological Service between 1970 and 2022 (TSMS dataset), were analyzed in combination with reanalysis-based satellite observations obtained from the ERA5 (ERA5 dataset). These historical records formed the foundation for developing temperature projections extending to the year 2050. To achieve this, two complementary time-series forecasting approaches were applied: the Long Short-Term Memory (LSTM) deep-learning model, known for its ability to capture nonlinear dependencies and long-range temporal patterns, and the Seasonal Auto-Regressive Integrated Moving Average (SARIMA) model, a classical statistical method suitable for handling seasonality and trend components in climate data. The projection results revealed that Türkiye's mean temperature anomaly relative to the 1970–1980 baseline period is expected to rise by approximately 2.52 °C when based on in-situ observational data, and by about 3.48 °C when derived from ERA5 reanalysis estimates. These findings consistently indicate a significant warming trajectory, regardless of the dataset or modeling approach applied.

根据政府间气候变化专门委员会（IPCC）的评估，位于地中海盆地内的t rkiye是最容易受到气候变化不利影响的地区之一。这种脆弱性的增加主要归因于其地理位置、气候特征和社会经济结构，这些因素共同放大了与气温上升和气候变率增加相关的风险。在本研究中，结合ERA5 （ERA5数据集）获得的基于再分析的卫星观测数据，分析了土耳其国家气象局1970 - 2022年记录的 rkiye月平均气温数据（TSMS数据集）。这些历史记录为发展到2050年的温度预测奠定了基础。为了实现这一目标，采用了两种互补的时间序列预测方法：长短期记忆（LSTM）深度学习模型，以其捕获非线性依赖关系和长期时间模式的能力而闻名，以及季节自回归综合移动平均（SARIMA）模型，这是一种经典的统计方法，适用于处理气候数据中的季节性和趋势成分。预估结果显示，与1970-1980年基线期相比，基于原位观测资料的 rkiye平均温度距平预计将上升约2.52℃，而基于ERA5再分析估计的距平预计将上升约3.48℃。无论采用何种数据集或建模方法，这些发现一致表明一个显著的变暖轨迹。

{"title":"Time series analysis of the impact of global warming on Türkiye","authors":"Arif Ozbek, Mehmet Bilgili","doi":"10.1016/j.jastp.2025.106647","DOIUrl":"10.1016/j.jastp.2025.106647","url":null,"abstract":"<div><div>According to the assessments of the Intergovernmental Panel on Climate Change (IPCC), Türkiye, located within the Mediterranean basin, is among the regions most susceptible to the adverse impacts of climate change. This heightened vulnerability is largely attributed to its geographic location, climatic characteristics, and socio-economic structure, which together amplify the risks associated with rising temperatures and increasing climate variability. In the present study, monthly mean air temperature data for Türkiye, recorded by the Turkish State Meteorological Service between 1970 and 2022 (TSMS dataset), were analyzed in combination with reanalysis-based satellite observations obtained from the ERA5 (ERA5 dataset). These historical records formed the foundation for developing temperature projections extending to the year 2050. To achieve this, two complementary time-series forecasting approaches were applied: the Long Short-Term Memory (LSTM) deep-learning model, known for its ability to capture nonlinear dependencies and long-range temporal patterns, and the Seasonal Auto-Regressive Integrated Moving Average (SARIMA) model, a classical statistical method suitable for handling seasonality and trend components in climate data. The projection results revealed that Türkiye's mean temperature anomaly relative to the 1970–1980 baseline period is expected to rise by approximately 2.52 °C when based on in-situ observational data, and by about 3.48 °C when derived from ERA5 reanalysis estimates. These findings consistently indicate a significant warming trajectory, regardless of the dataset or modeling approach applied.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"277 ","pages":"Article 106647"},"PeriodicalIF":1.9,"publicationDate":"2025-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145156583","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Long-term projections of global, northern hemisphere, and arctic sea ice concentration using statistical and deep learning approaches 使用统计和深度学习方法的全球、北半球和北极海冰浓度的长期预测

IF 1.9 4区地球科学 Q3 GEOCHEMISTRY & GEOPHYSICS

Journal of Atmospheric and Solar-Terrestrial Physics

Pub Date : 2025-09-20 DOI: 10.1016/j.jastp.2025.106634

Mehmet Bilgili , Engin Pinar , Md. Najmul Mowla , Tahir Durhasan , Muhammed M. Aksoy

The accelerating decline in sea ice concentration (SIC) poses significant challenges for global climate regulation, maritime navigation, and arctic ecosystem stability. This study develops and evaluates two advanced time-series forecasting models, seasonal autoregressive integrated moving average (SARIMA) and long short-term memory (LSTM) networks, to project SIC trends through 2050 across three spatial domains: the globe, the northern hemisphere, and the arctic. Utilizing the ERA5 reanalysis dataset (1970–2024) from the European center for medium-range weather forecasts (ECMWF), the models capture seasonal cycles and complex temporal dependencies to enable robust long-term projections. Comparative analysis demonstrates that SARIMA effectively models periodic fluctuations, while LSTM excels at learning nonlinear dependencies inherent in SIC dynamics. Performance metrics, including mean absolute percentage error (MAPE), root mean square error (RMSE), and correlation coefficient (R), confirm the high accuracy of both models, with SARIMA showing superior capability in representing structured seasonal patterns. Projections indicate a persistent decline in SIC, with arctic concentrations decreasing from 55.60% in 2023 to approximately 46.84% by 2050, underscoring the pronounced effects of arctic amplification. These results provide valuable insights for climate modeling, arctic policy formulation, and the development of adaptive navigation strategies in a rapidly changing polar environment.

海冰浓度的加速下降对全球气候调节、海上航行和北极生态系统的稳定提出了重大挑战。本研究开发并评估了两个先进的时间序列预测模型，即季节自回归综合移动平均（SARIMA）和长短期记忆（LSTM）网络，以预测到2050年全球、北半球和北极三个空间域的SIC趋势。利用欧洲中期天气预报中心（ECMWF）的ERA5再分析数据集（1970-2024），这些模式捕捉季节周期和复杂的时间依赖性，从而实现可靠的长期预测。对比分析表明，SARIMA有效地模拟了周期波动，而LSTM在学习SIC动力学中固有的非线性依赖方面表现出色。包括平均绝对百分比误差（MAPE）、均方根误差（RMSE）和相关系数(R)在内的性能指标证实了两种模型的高准确性，其中SARIMA在表示结构化季节模式方面表现出卓越的能力。预估表明SIC持续下降，北极浓度从2023年的55.60%下降到2050年的约46.84%，强调了北极放大的显著影响。这些结果为气候建模、北极政策制定以及在快速变化的极地环境中自适应导航策略的发展提供了有价值的见解。

{"title":"Long-term projections of global, northern hemisphere, and arctic sea ice concentration using statistical and deep learning approaches","authors":"Mehmet Bilgili , Engin Pinar , Md. Najmul Mowla , Tahir Durhasan , Muhammed M. Aksoy","doi":"10.1016/j.jastp.2025.106634","DOIUrl":"10.1016/j.jastp.2025.106634","url":null,"abstract":"<div><div>The accelerating decline in sea ice concentration (SIC) poses significant challenges for global climate regulation, maritime navigation, and arctic ecosystem stability. This study develops and evaluates two advanced time-series forecasting models, seasonal autoregressive integrated moving average (SARIMA) and long short-term memory (LSTM) networks, to project SIC trends through 2050 across three spatial domains: the globe, the northern hemisphere, and the arctic. Utilizing the ERA5 reanalysis dataset (1970–2024) from the European center for medium-range weather forecasts (ECMWF), the models capture seasonal cycles and complex temporal dependencies to enable robust long-term projections. Comparative analysis demonstrates that SARIMA effectively models periodic fluctuations, while LSTM excels at learning nonlinear dependencies inherent in SIC dynamics. Performance metrics, including mean absolute percentage error (MAPE), root mean square error (RMSE), and correlation coefficient (R), confirm the high accuracy of both models, with SARIMA showing superior capability in representing structured seasonal patterns. Projections indicate a persistent decline in SIC, with arctic concentrations decreasing from 55.60% in 2023 to approximately 46.84% by 2050, underscoring the pronounced effects of arctic amplification. These results provide valuable insights for climate modeling, arctic policy formulation, and the development of adaptive navigation strategies in a rapidly changing polar environment.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"277 ","pages":"Article 106634"},"PeriodicalIF":1.9,"publicationDate":"2025-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145119565","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0