Accurate, rapid, and automated landslide detection is crucial for early warning, emergency management, and landslide mechanism analysis. Increasingly general-purpose detection models are being deployed for these complex and dynamic tasks involving features that are difficult to characterize. However, these models are computationally expensive and memory-hungry, while the accuracy and detection efficiency remain wanting. To address the above problems, this paper proposes an end-to-end model with high-precision and lightweight design for integrated landslide detection and segmentation. Here, we customized the backbone utilizing the advanced Efficient MOdel (EMO), and further used the linear cheap operation from GhostNet to reduce computational complexity. As a result, the total parameters of our models were reduced by up to 48.13%, compared to the baseline. Building on this, we employed a dynamic detection head with multiple attention mechanisms, and proposed a lightweight attention enhancement module for strengthened multi-scale feature extraction and fusion. The results demonstrate that our model outperforms the baseline on all metrics, achieving an outstanding F1 score of 96.75%.
{"title":"Using lightweight method to detect landslide from satellite imagery","authors":"Jinchi Dai, Xiaoai Dai, Renyuan Zhang, JiaXin Ma, Wenyu Li, Heng Lu, Weile Li, Shuneng Liang, Tangrui Dai, Yunfeng Shan, Donghui Zhang, Lei Zhao","doi":"10.1016/j.jag.2024.104303","DOIUrl":"https://doi.org/10.1016/j.jag.2024.104303","url":null,"abstract":"Accurate, rapid, and automated landslide detection is crucial for early warning, emergency management, and landslide mechanism analysis. Increasingly general-purpose detection models are being deployed for these complex and dynamic tasks involving features that are difficult to characterize. However, these models are computationally expensive and memory-hungry, while the accuracy and detection efficiency remain wanting. To address the above problems, this paper proposes an end-to-end model with high-precision and lightweight design for integrated landslide detection and segmentation. Here, we customized the backbone utilizing the advanced Efficient MOdel (EMO), and further used the linear cheap operation from GhostNet to reduce computational complexity. As a result, the total parameters of our models were reduced by up to 48.13%, compared to the baseline. Building on this, we employed a dynamic detection head with multiple attention mechanisms, and proposed a lightweight attention enhancement module for strengthened multi-scale feature extraction and fusion. The results demonstrate that our model outperforms the baseline on all metrics, achieving an outstanding F1 score of 96.75%.","PeriodicalId":50341,"journal":{"name":"International Journal of Applied Earth Observation and Geoinformation","volume":"63 1","pages":""},"PeriodicalIF":7.5,"publicationDate":"2024-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142825362","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Effectively leveraging landslide spatial location information is crucial for improving the accuracy of deep learning in predicting landslide susceptibility and exploring the impacts of predisposing factors. Current single deep learning models for landslide susceptibility assessment require enhancements in both prediction accuracy and robustness. Inclusion of non-interrelated positional information among samples leads to reduced prediction accuracy and challenges in quantifying landslide risk covariates. This study proposes a landslide susceptibility assessment method that integrates ensemble learning with geographically weighted concepts. Using a stacking method, a 1D convolutional neural network (1D-CNN), a recurrent neural network (RNN), and a long short-term memory (LSTM) network were combined to form the CRNN-LSTM ensemble model. Additionally, we constructed a deep learning geographically weighted regression (GW-DNN) model based on the deep learning principles and geographically weighted regression to quantify the impacts of landslide-predisposing factors.The experimental results show that the CRNN-LSTM model achieved AUC values of 0.977 and 0.961 on the training and validation sets, significantly outperforming the individual classifiers (AUC of 0.944 and 0.940 for the 1D-CNN model, 0.950 and 0.948 for the RNN model, and 0.956 and 0.952 for the LSTM model). Additionally, the GW-DNN model achieved R2 coefficients of 0.876 and 0.860 during the training and validation phases. These findings indicate that our proposed method not only highly accurately predicts landslide susceptibility but also provides a precise quantitative assessment of the impact of landslide-predisposing factors at specific spatial points (landslide units) in high-risk areas. These findings offer valuable technical support for landslide disaster prevention and mitigation.
{"title":"An approach for predicting landslide susceptibility and evaluating predisposing factors","authors":"Wanxin Guo, Jian Ye, Chengbing Liu, Yijie Lv, Qiuyu Zeng, Xin Huang","doi":"10.1016/j.jag.2024.104217","DOIUrl":"https://doi.org/10.1016/j.jag.2024.104217","url":null,"abstract":"Effectively leveraging landslide spatial location information is crucial for improving the accuracy of deep learning in predicting landslide susceptibility and exploring the impacts of predisposing factors. Current single deep learning models for landslide susceptibility assessment require enhancements in both prediction accuracy and robustness. Inclusion of non-interrelated positional information among samples leads to reduced prediction accuracy and challenges in quantifying landslide risk covariates. This study proposes a landslide susceptibility assessment method that integrates ensemble learning with geographically weighted concepts. Using a stacking method, a 1D convolutional neural network (1D-CNN), a recurrent neural network (RNN), and a long short-term memory (LSTM) network were combined to form the CRNN-LSTM ensemble model. Additionally, we constructed a deep learning geographically weighted regression (GW-DNN) model based on the deep learning principles and geographically weighted regression to quantify the impacts of landslide-predisposing factors.The experimental results show that the CRNN-LSTM model achieved AUC values of 0.977 and 0.961 on the training and validation sets, significantly outperforming the individual classifiers (AUC of 0.944 and 0.940 for the 1D-CNN model, 0.950 and 0.948 for the RNN model, and 0.956 and 0.952 for the LSTM model). Additionally, the GW-DNN model achieved R<ce:sup loc=\"post\">2</ce:sup> coefficients of 0.876 and 0.860 during the training and validation phases. These findings indicate that our proposed method not only highly accurately predicts landslide susceptibility but also provides a precise quantitative assessment of the impact of landslide-predisposing factors at specific spatial points (landslide units) in high-risk areas. These findings offer valuable technical support for landslide disaster prevention and mitigation.","PeriodicalId":50341,"journal":{"name":"International Journal of Applied Earth Observation and Geoinformation","volume":"120 1","pages":""},"PeriodicalIF":7.5,"publicationDate":"2024-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142825364","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-09DOI: 10.1016/j.jag.2024.104236
Zhiyong Zhou, Cheng Fu, Robert Weibel
Building generalization is an essential task in generating multi-scale topographic maps. The progress of deep learning offers a new paradigm to overcome the coordination challenges faced by conventional building generalization algorithms. Some studies have confirmed the feasibility of several original semantic segmentation networks, such as U-Net and its variants and the conditional generative adversarial network (cGAN), for building generalization in image maps. However, they suffer from critical deformation effects, especially for large and geometrically complex buildings. Since learning building generalization essentially means modeling the subtle transformation of building footprints across scales, we argue that the spatial awareness of a neural network, for instance, regarding building size and shape, is crucial to effective learning. Thus, we propose a spatially-aware generative adversarial network, SpaGAN. It takes a representative cGAN, pix2pix, as the backbone, and modifies two modules: In the U-Net-based generator, an atrous spatial pyramid pooling (ASPP) module replaces the conventional convolutional module to extract multi-scale features of buildings of varying sizes and shapes; in the PatchGAN-based discriminator, a signed distance map (SDM) module is used to capture the fine-grained shape difference for discrimination. The proposed network was comprehensively evaluated with a synthetic and a real-world dataset. The results demonstrate that SpaGAN outperforms existing baseline models (U-Net, ResU-Net, pix2pix) for building generalization, particularly in the real-world dataset. The new model can achieve more reasonable aggregation, simplification, and squaring generalization operators.
{"title":"SpaGAN: A spatially-aware generative adversarial network for building generalization in image maps","authors":"Zhiyong Zhou, Cheng Fu, Robert Weibel","doi":"10.1016/j.jag.2024.104236","DOIUrl":"https://doi.org/10.1016/j.jag.2024.104236","url":null,"abstract":"Building generalization is an essential task in generating multi-scale topographic maps. The progress of deep learning offers a new paradigm to overcome the coordination challenges faced by conventional building generalization algorithms. Some studies have confirmed the feasibility of several original semantic segmentation networks, such as U-Net and its variants and the conditional generative adversarial network (cGAN), for building generalization in image maps. However, they suffer from critical deformation effects, especially for large and geometrically complex buildings. Since learning building generalization essentially means modeling the subtle transformation of building footprints across scales, we argue that the spatial awareness of a neural network, for instance, regarding building size and shape, is crucial to effective learning. Thus, we propose a spatially-aware generative adversarial network, SpaGAN. It takes a representative cGAN, pix2pix, as the backbone, and modifies two modules: In the U-Net-based generator, an atrous spatial pyramid pooling (ASPP) module replaces the conventional convolutional module to extract multi-scale features of buildings of varying sizes and shapes; in the PatchGAN-based discriminator, a signed distance map (SDM) module is used to capture the fine-grained shape difference for discrimination. The proposed network was comprehensively evaluated with a synthetic and a real-world dataset. The results demonstrate that SpaGAN outperforms existing baseline models (U-Net, ResU-Net, pix2pix) for building generalization, particularly in the real-world dataset. The new model can achieve more reasonable aggregation, simplification, and squaring generalization operators.","PeriodicalId":50341,"journal":{"name":"International Journal of Applied Earth Observation and Geoinformation","volume":"11 1","pages":""},"PeriodicalIF":7.5,"publicationDate":"2024-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142825363","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-06DOI: 10.1016/j.jag.2024.104296
Yile He, Youping Xie, Junchen Liu, Zengyun Hu, Jun Liu, Yuhua Cheng, Lei Zhang, Zhihui Wang, Man Li
Under the background of climate change and global warming, extreme drought events in China are becoming increasingly frequent. Drought is one of the primary natural causes of damage to China’s agriculture, economy, and environment, making timely, accurate, and high-resolution drought monitoring particularly crucial. The global standardized precipitation − evapotranspiration index database (SPEIbase) is a widely accepted and used global-scale drought monitoring product. However, limited by its spatial resolution of 0.5 degrees, it is difficult to describe the local spatio-temporal structure of drought. How to improve its spatial resolution while maintaining spatio-temporal consistency is one of the current research hotspots. Based on the response of vegetation growth status to drought, this paper proposes a simple and feasible SPEI prediction method, which improves the resolution of SPEIbase from 0.5 degrees to 1 km. Sixteen remote sensing inversion indices, reflectance and elevation data related to drought were selected from Google Earth Engine (GEE) as features. After preprocessing such as gridding and sample balancing, a random forest regression model was constructed to achieve high spatial resolution prediction of SPEI. SPEI with time scales of 1, 3, 6, 9, 12 and 24 months in July 2020, August 2019 and August 2018 in China was selected for experiments. The accuracy of 1 km resolution SPEI was evaluated through metrics such as root mean square error (RMSE), Pearson correlation coefficient (PCC) and determination coefficient (R2). At the same time, it was compared with the existing 1 km resolution SPEI dataset and the site-scale SPEI values. The results show that the method in this paper can obtain accurate prediction results more stably. The PCC and R2 of different months and multiple time scales are all higher than 0.9 and 0.8, and the RMSE is lower than 0.4, showing a good application prospect. Despite the good consistency between the Proposed SPEI and SPEIbase with the site-scale SPEI values, there is still significant room for improvement.
{"title":"Generation of 1 km high resolution Standardized precipitation evapotranspiration Index for drought monitoring over China using Google Earth Engine","authors":"Yile He, Youping Xie, Junchen Liu, Zengyun Hu, Jun Liu, Yuhua Cheng, Lei Zhang, Zhihui Wang, Man Li","doi":"10.1016/j.jag.2024.104296","DOIUrl":"https://doi.org/10.1016/j.jag.2024.104296","url":null,"abstract":"Under the background of climate change and global warming, extreme drought events in China are becoming increasingly frequent. Drought is one of the primary natural causes of damage to China’s agriculture, economy, and environment, making timely, accurate, and high-resolution drought monitoring particularly crucial. The global standardized precipitation − evapotranspiration index database (SPEIbase) is a widely accepted and used global-scale drought monitoring product. However, limited by its spatial resolution of 0.5 degrees, it is difficult to describe the local spatio-temporal structure of drought. How to improve its spatial resolution while maintaining spatio-temporal consistency is one of the current research hotspots. Based on the response of vegetation growth status to drought, this paper proposes a simple and feasible SPEI prediction method, which improves the resolution of SPEIbase from 0.5 degrees to 1 km. Sixteen remote sensing inversion indices, reflectance and elevation data related to drought were selected from Google Earth Engine (GEE) as features. After preprocessing such as gridding and sample balancing, a random forest regression model was constructed to achieve high spatial resolution prediction of SPEI. SPEI with time scales of 1, 3, 6, 9, 12 and 24 months in July 2020, August 2019 and August 2018 in China was selected for experiments. The accuracy of 1 km resolution SPEI was evaluated through metrics such as root mean square error (RMSE), Pearson correlation coefficient (PCC) and determination coefficient (R<ce:sup loc=\"post\">2</ce:sup>). At the same time, it was compared with the existing 1 km resolution SPEI dataset and the site-scale SPEI values. The results show that the method in this paper can obtain accurate prediction results more stably. The PCC and R<ce:sup loc=\"post\">2</ce:sup> of different months and multiple time scales are all higher than 0.9 and 0.8, and the RMSE is lower than 0.4, showing a good application prospect. Despite the good consistency between the Proposed SPEI and SPEIbase with the site-scale SPEI values, there is still significant room for improvement.","PeriodicalId":50341,"journal":{"name":"International Journal of Applied Earth Observation and Geoinformation","volume":"10 1","pages":""},"PeriodicalIF":7.5,"publicationDate":"2024-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142793735","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-06DOI: 10.1016/j.jag.2024.104307
Zongze Li, Jinsong Chong, Yawei Zhao, Lijie Diao
Glacier velocity is one of the crucial parameters in the research of glacier dynamics. Synthetic aperture radar (SAR), as an active microwave sensor, represents a common method to monitor glacier velocity. However, the changes of glacier surface could cause the data missing of glacier velocity due to incoherence. To meet the demand for glacier velocity monitoring, this paper employs the SAR images of Sentinel-1 in long time series and optical images of Sentinel-2 to investigate the velocity of Petermann glacier in 2021. Firstly, the time series of glacier velocity in the whole year of 2021 is obtained by using SAR images. The glacier velocity extracted from the optical image pairs is used as the initial value of the large missing part of the glacier velocity field. Then the spatiotemporal glacier velocity matrix is constructed and empirical orthogonal function (EOF) analysis is carried out. Among them, the glacier velocity is reconstructed by the glacier velocity estimation method based on confidence, and the complete glacier velocity time series is obtained by iterating to minimize the error of the reconstructed glacier velocity. Finally, the obtained time series of Petermann Glacier velocity in 2021 were statistically analyzed. The statistical results quantified the seasonal differences of Petermann Glacier. In addition, the analysis results show that the temporal and spatial variations of Petermann Glacier velocity are affected by topography and temperature.
{"title":"Reconstruction of Petermann glacier velocity time series using multi-source remote sensing images","authors":"Zongze Li, Jinsong Chong, Yawei Zhao, Lijie Diao","doi":"10.1016/j.jag.2024.104307","DOIUrl":"https://doi.org/10.1016/j.jag.2024.104307","url":null,"abstract":"Glacier velocity is one of the crucial parameters in the research of glacier dynamics. Synthetic aperture radar (SAR), as an active microwave sensor, represents a common method to monitor glacier velocity. However, the changes of glacier surface could cause the data missing of glacier velocity due to incoherence. To meet the demand for glacier velocity monitoring, this paper employs the SAR images of Sentinel-1 in long time series and optical images of Sentinel-2 to investigate the velocity of Petermann glacier in 2021. Firstly, the time series of glacier velocity in the whole year of 2021 is obtained by using SAR images. The glacier velocity extracted from the optical image pairs is used as the initial value of the large missing part of the glacier velocity field. Then the spatiotemporal glacier velocity matrix is constructed and empirical orthogonal function (EOF) analysis is carried out. Among them, the glacier velocity is reconstructed by the glacier velocity estimation method based on confidence, and the complete glacier velocity time series is obtained by iterating to minimize the error of the reconstructed glacier velocity. Finally, the obtained time series of Petermann Glacier velocity in 2021 were statistically analyzed. The statistical results quantified the seasonal differences of Petermann Glacier. In addition, the analysis results show that the temporal and spatial variations of Petermann Glacier velocity are affected by topography and temperature.","PeriodicalId":50341,"journal":{"name":"International Journal of Applied Earth Observation and Geoinformation","volume":"20 1","pages":""},"PeriodicalIF":7.5,"publicationDate":"2024-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142793511","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-06DOI: 10.1016/j.jag.2024.104291
Yufang He, Mahdi Motagh, Xiaohang Wang, Xiaojie Liu, Hermann Kaufmann, Guochang Xu, Bo Chen
Recently Guangzhou and Foshan in China are experiencing significant urbanization and economic development. However, the accelerated urbanization process has contributed significantly to urban land subsidence, causing huge economic losses and endangering safety of infrastructure. This intricate activities on urban surfaces can also lead to pseudo danger in interpreting InSAR-based urban surface deformation, resulting in hazard misidentification in two cities. In order to more accurately identify the hazard of urban surface deformation, we innovatively present a combination of InSAR technology with multi-temporal optical remote sensing data. It can also analyze the specific causes of urban deformation at SAR pixel level in two cities. The SBAS-InSAR method was adopted to obtain an urban subsidence map from 2017 to 2020 based on 110 Sentinel-1 SAR image scenes. To obtain an urban surface change map with a high accuracy, an improved SwiT-UNet++ model was applied based on multi optical Google Earth imagery. By a combined analysis of SAR and optical images, we discovered multiple irregular funnels with subsidence at different scales in both cities, that are mostly relatable to urban surface constructions such as foundation compression, building demolition, and the construction of public facilities. Furthermore, to identify detailed hazard around surface changes, the buffer analysis based on InSAR surface deformation and urban surface change maps was conducted. It revealed the surface deformation signals around certain urban surface change areas are more obvious and pose certain hazard. Finally additional high-risk areas are found in the two cities. By subtracting the optical surface change detection map from the InSAR-based urban subsidence map, the “pseudo danger” caused by urban activities in the interpretation of InSAR-based urban surface deformation is eliminated, enabling precise identification of actual land subsidence hazards. It is realized through a risk assessment experiment in the research area by adding factors of urbanization processes. By combining multiple sources of data and using advanced analytical techniques, we could identify the determining factors contributing to urban subsidence and the detailed hazards and thus, provide valuable information for future urban developments.
{"title":"Detailed hazard identification of urban subsidence in Guangzhou and Foshan by combining InSAR and optical imagery","authors":"Yufang He, Mahdi Motagh, Xiaohang Wang, Xiaojie Liu, Hermann Kaufmann, Guochang Xu, Bo Chen","doi":"10.1016/j.jag.2024.104291","DOIUrl":"https://doi.org/10.1016/j.jag.2024.104291","url":null,"abstract":"Recently Guangzhou and Foshan in China are experiencing significant urbanization and economic development. However, the accelerated urbanization process has contributed significantly to urban land subsidence, causing huge economic losses and endangering safety of infrastructure. This intricate activities on urban surfaces can also lead to pseudo danger in interpreting InSAR-based urban surface deformation, resulting in hazard misidentification in two cities. In order to more accurately identify the hazard of urban surface deformation, we innovatively present a combination of InSAR technology with multi-temporal optical remote sensing data. It can also analyze the specific causes of urban deformation at SAR pixel level in two cities. The SBAS-InSAR method was adopted to obtain an urban subsidence map from 2017 to 2020 based on 110 Sentinel-1 SAR image scenes. To obtain an urban surface change map with a high accuracy, an improved SwiT-UNet++ model was applied based on multi optical Google Earth imagery. By a combined analysis of SAR and optical images, we discovered multiple irregular funnels with subsidence at different scales in both cities, that are mostly relatable to urban surface constructions such as foundation compression, building demolition, and the construction of public facilities. Furthermore, to identify detailed hazard around surface changes, the buffer analysis based on InSAR surface deformation and urban surface change maps was conducted. It revealed the surface deformation signals around certain urban surface change areas are more obvious and pose certain hazard. Finally additional high-risk areas are found in the two cities. By subtracting the optical surface change detection map from the InSAR-based urban subsidence map, the “pseudo danger” caused by urban activities in the interpretation of InSAR-based urban surface deformation is eliminated, enabling precise identification of actual land subsidence hazards. It is realized through a risk assessment experiment in the research area by adding factors of urbanization processes. By combining multiple sources of data and using advanced analytical techniques, we could identify the determining factors contributing to urban subsidence and the detailed hazards and thus, provide valuable information for future urban developments.","PeriodicalId":50341,"journal":{"name":"International Journal of Applied Earth Observation and Geoinformation","volume":"9 1","pages":""},"PeriodicalIF":7.5,"publicationDate":"2024-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142793565","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-05DOI: 10.1016/j.jag.2024.104297
Meng Luo, Shengwei Zhang, Ruishen Li, Xi Lin, Shuai Wang, Lin Yang, Kedi Fang
Vegetation carbon sequestration is a fundamental process that supports ecosystem biodiversity and ecological services. It is a key factor in shaping ecosystem state and energy flow. Global climate change has intensified in recent years. Frequent drought events affect the stabilization of carbon cycle. In this study, we used correlation analysis method to explore the relationship between standardized precipitation evapotranspiration index (SPEI) and gross primary productivity (GPP). Our study found that the global drought degree is decreasing, and drought sensitivity of global surface vegetation decreased. The drought index value increased 91.3% and the sensitivity decreased 35.71% during the 2010–2020 period (P2) compared to the 2000–2010 period (P1). Our study also found that the global area of drought decreased by 4.03% in P2, but the global area with high drought frequency increased by 0.21%. The drought response time scale shortened by 5.19%. GPP showed an increasing trend, with the largest increase in agricultural land. By studying the interaction between drought and different vegetation types, we can better understand the mechanisms by which vegetation responds, adapts and regulates to climate change. It is necessary for understanding the sustainable development of global ecosystems and climate change response.
{"title":"Global vegetation productivity has become less sensitive to drought in the first two decades of the 21st century","authors":"Meng Luo, Shengwei Zhang, Ruishen Li, Xi Lin, Shuai Wang, Lin Yang, Kedi Fang","doi":"10.1016/j.jag.2024.104297","DOIUrl":"https://doi.org/10.1016/j.jag.2024.104297","url":null,"abstract":"Vegetation carbon sequestration is a fundamental process that supports ecosystem biodiversity and ecological services. It is a key factor in shaping ecosystem state and energy flow. Global climate change has intensified in recent years. Frequent drought events affect the stabilization of carbon cycle. In this study, we used correlation analysis method to explore the relationship between standardized precipitation evapotranspiration index (SPEI) and gross primary productivity (GPP). Our study found that the global drought degree is decreasing, and drought sensitivity of global surface vegetation decreased. The drought index value increased 91.3% and the sensitivity decreased 35.71% during the 2010–2020 period (P2) compared to the 2000–2010 period (P1). Our study also found that the global area of drought decreased by 4.03% in P2, but the global area with high drought frequency increased by 0.21%. The drought response time scale shortened by 5.19%. GPP showed an increasing trend, with the largest increase in agricultural land. By studying the interaction between drought and different vegetation types, we can better understand the mechanisms by which vegetation responds, adapts and regulates to climate change. It is necessary for understanding the sustainable development of global ecosystems and climate change response.","PeriodicalId":50341,"journal":{"name":"International Journal of Applied Earth Observation and Geoinformation","volume":"28 1","pages":""},"PeriodicalIF":7.5,"publicationDate":"2024-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142793570","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-05DOI: 10.1016/j.jag.2024.104305
Ziwei Liu, Mingchang Wang, Xingnan Liu, Xiaoyue Lyu, Minshui Wang, Fengyan Wang, Xue Ji, Xiaoyan Li
Long-term climate change significantly affects the spatiotemporal dynamics of soil erosion. To explore this, remote sensing technology, future climate scenarios, and deep learning are combined to model the historical and future variations in soil erosion, investigating its spatiotemporal dynamics influenced by climate change. This paper uses the Revised Universal Soil Loss Equation (RUSLE) to assess the historical changes in erosion in northeast China from 1980 to 2020. A soil erosion simulation (SES) model was developed, incorporating deep learning models, to forecast future trends in soil erosion under various climate scenarios. The SES model achieves an R-squared (R2) value of 0.7513. The SES model can simulate the Spatiotemporal dynamics of soil erosion influenced by long-term climate change. Soil erosion from 2001 to 2020 is lower than that from 1980 to 2000, indicating a decrease in soil erosion under natural variability conditions. Unlike historical trends, future soil erosion demonstrates significant variation across three scenarios: SSP1-RCP1.9 (SSP119), SSP2-RCP4.5 (SSP245), and SSP5-RCP8.5 (SSP585). The simulation results show that the SSP119 climate scenario has a minor impact on soil erosion, whereas the SSP245 scenario leads to a gradual increase in soil erosion. The SSP585 scenario, characterized by high social vulnerability and substantial radiative forcing, exacerbates the risk of soil erosion. The study provides valuable references for maintaining soil stability and managing surface runoff.
{"title":"Spatiotemporal simulation and projection of soil erosion as affected by climate change in Northeast China","authors":"Ziwei Liu, Mingchang Wang, Xingnan Liu, Xiaoyue Lyu, Minshui Wang, Fengyan Wang, Xue Ji, Xiaoyan Li","doi":"10.1016/j.jag.2024.104305","DOIUrl":"https://doi.org/10.1016/j.jag.2024.104305","url":null,"abstract":"Long-term climate change significantly affects the spatiotemporal dynamics of soil erosion. To explore this, remote sensing technology, future climate scenarios, and deep learning are combined to model the historical and future variations in soil erosion, investigating its spatiotemporal dynamics influenced by climate change. This paper uses the Revised Universal Soil Loss Equation (RUSLE) to assess the historical changes in erosion in northeast China from 1980 to 2020. A soil erosion simulation (SES) model was developed, incorporating deep learning models, to forecast future trends in soil erosion under various climate scenarios. The SES model achieves an R-squared (R<ce:sup loc=\"post\">2</ce:sup>) value of 0.7513. The SES model can simulate the Spatiotemporal dynamics of soil erosion influenced by long-term climate change. Soil erosion from 2001 to 2020 is lower than that from 1980 to 2000, indicating a decrease in soil erosion under natural variability conditions. Unlike historical trends, future soil erosion demonstrates significant variation across three scenarios: SSP1-RCP1.9 (SSP119), SSP2-RCP4.5 (SSP245), and SSP5-RCP8.5 (SSP585). The simulation results show that the SSP119 climate scenario has a minor impact on soil erosion, whereas the SSP245 scenario leads to a gradual increase in soil erosion. The SSP585 scenario, characterized by high social vulnerability and substantial radiative forcing, exacerbates the risk of soil erosion. The study provides valuable references for maintaining soil stability and managing surface runoff.","PeriodicalId":50341,"journal":{"name":"International Journal of Applied Earth Observation and Geoinformation","volume":"69 1","pages":""},"PeriodicalIF":7.5,"publicationDate":"2024-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142793566","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Active landslides pose significant global risks, underscoring precise displacement monitoring for effective geohazard management and early warning. The Three Gorges Reservoir Area (TGRA) in China, a pivotal section of the world’s largest water conservancy project, has developed thousands of landslides due to unique hydrogeological conditions and reservoir operations. Many of these landslides are oriented north–south and covered by seasonal vegetation, which complicates the conventional remote sensing-based displacement monitoring, particularly in estimating the three-dimensional (3D) deformation and long-term time series displacement. To address these challenges, we propose an approach that integrates interferometric synthetic aperture radar (InSAR), pixel offset tracking (POT), stacking, and priori kinematic models to fully utilize the phase and amplitude information of multi-platform, multi-band SAR images (i.e., L-band ALOS-1, C-band Sentinel-1, and X-band TerraSAR-X). This approach is employed to scrutinize the long-term spatiotemporal deformation and evolution mechanism of two slow-moving, north-facing reservoir landslides in the TGRA. The results reveal for the first time the 15-year-long displacement evolution of these landslides before and after reservoir impoundment, highlighting the spatiotemporal heterogeneity of landslide deformation induced by hydrologic triggers. The impoundment in September 2008 induced transient acceleration in both landslides, followed by a relatively stable, step-like deformation pattern subject to rainfall and reservoir water level (RWL) fluctuations. Rainfall, with a lag of approximately 20 days, predominantly affects both landslides, while RWL fluctuations mainly influence the deformation at landslide toes. Notably, as the distance from the reservoir increases, the influence of RWL diminishes, with lag times increasing from 8 to about 40 days. This quantitative characterization of landslide responses to triggers represents a crucial step towards improved hazard mitigation capabilities.
{"title":"Unravelling long-term spatiotemporal deformation and hydrological triggers of slow-moving reservoir landslides with multi-platform SAR data","authors":"Fengnian Chang, Shaochun Dong, Hongwei Yin, Xiao Ye, Zhenyun Wu, Wei Zhang, Honghu Zhu","doi":"10.1016/j.jag.2024.104301","DOIUrl":"https://doi.org/10.1016/j.jag.2024.104301","url":null,"abstract":"Active landslides pose significant global risks, underscoring precise displacement monitoring for effective geohazard management and early warning. The Three Gorges Reservoir Area (TGRA) in China, a pivotal section of the world’s largest water conservancy project, has developed thousands of landslides due to unique hydrogeological conditions and reservoir operations. Many of these landslides are oriented north–south and covered by seasonal vegetation, which complicates the conventional remote sensing-based displacement monitoring, particularly in estimating the three-dimensional (3D) deformation and long-term time series displacement. To address these challenges, we propose an approach that integrates interferometric synthetic aperture radar (InSAR), pixel offset tracking (POT), stacking, and priori kinematic models to fully utilize the phase and amplitude information of multi-platform, multi-band SAR images (i.e., L-band ALOS-1, C-band Sentinel-1, and X-band TerraSAR-X). This approach is employed to scrutinize the long-term spatiotemporal deformation and evolution mechanism of two slow-moving, north-facing reservoir landslides in the TGRA. The results reveal for the first time the 15-year-long displacement evolution of these landslides before and after reservoir impoundment, highlighting the spatiotemporal heterogeneity of landslide deformation induced by hydrologic triggers. The impoundment in September 2008 induced transient acceleration in both landslides, followed by a relatively stable, step-like deformation pattern subject to rainfall and reservoir water level (RWL) fluctuations. Rainfall, with a lag of approximately 20 days, predominantly affects both landslides, while RWL fluctuations mainly influence the deformation at landslide toes. Notably, as the distance from the reservoir increases, the influence of RWL diminishes, with lag times increasing from 8 to about 40 days. This quantitative characterization of landslide responses to triggers represents a crucial step towards improved hazard mitigation capabilities.","PeriodicalId":50341,"journal":{"name":"International Journal of Applied Earth Observation and Geoinformation","volume":"210 1","pages":""},"PeriodicalIF":7.5,"publicationDate":"2024-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142793568","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-05DOI: 10.1016/j.jag.2024.104298
Jia Tian, Qingjiu Tian, Suju Li, Qianjing Li, Sen Zhang, Shuang He
Sunlit and shaded components are commonly present in both airborne and satellite remote sensing images. In vegetated ecosystems, shaded component often result from sunlight being obstructed by topographic relief or canopy structures, and shaded component may impact plant growth, leaf photosynthesis, and ultimately carbon sequestration. To accurately estimate the fractional cover of the shaded and sunlit components, including both green and non-green vegetation within vegetated ecosystems, a novel method called the quasi-Hue-Saturation-Lightness (quasi-HSL) method is proposed in this study. Inspired by the RGB to HSL conversion, this method utilizes near-infrared, green, and red bands to compute hue (and normalized hue), saturation, and lightness. Subsequently, two indices, namely Hue-Lightness Index (HLI) and Saturation-Lightness Index (SLI), are introduced to construct a triangular space for estimating the fractional cover of the three components. Through unmanned aerial vehicle field experiments conducted in two forested areas, the accuracy of fractional cover estimation for three components reaches an R2 value of 0.50–0.67. Furthermore, this fractional cover estimation approach can be extended to a four-component estimation, including sunlit green vegetation, sunlit non-green vegetation, shaded green vegetation, and shaded non-green vegetation. With this detailed fractional cover estimation in vegetated area, the fractional vegetation coverage can be retrieved. Cross-validated with the fractional vegetation coverage retrieved by NDVI, the accuracy reaches R2 = 0.92. The advantages of the proposed method are (1) estimating fractional cover of shaded component without blue band, which is easily impacted by atmospheric conditions and sensor performance, and (2) differentiating the sunlit green and non-green vegetation components in the vegetated ecosystem.
在航空和卫星遥感图像中，阳光照射和阴影部分通常都存在。在植被生态系统中，遮荫成分通常是由于地形起伏或冠层结构阻挡阳光的结果，遮荫成分可能影响植物生长、叶片光合作用和最终的碳固存。为了准确估计植被生态系统中遮阳和日照组分（包括绿色和非绿色植被）的覆盖度，本文提出了一种新的方法——准色度-饱和度-亮度（quasi- saturation - lightness, hsl）方法。受RGB到HSL转换的启发，该方法利用近红外，绿色和红色波段来计算色调（和归一化色调），饱和度和亮度。随后，引入Hue-Lightness Index （HLI）和Saturation-Lightness Index （SLI）两个指标，构建一个三角空间来估计这三个分量的分数覆盖度。通过在2个林区进行的无人机野外试验，3个分量的覆盖度估算精度R2值为0.50-0.67。此外，这种分数覆盖度估计方法可以扩展为四分量估计，包括阳光照射下的绿色植被、阳光照射下的非绿色植被、阴影下的绿色植被和阴影下的非绿色植被。利用这种详细的植被覆盖度估算方法，可以反演植被覆盖度。与NDVI反演植被覆盖度交叉验证，精度达到R2 = 0.92。该方法的优点是：(1)估算无蓝带遮挡成分的覆盖度分数，蓝带容易受到大气条件和传感器性能的影响；(2)区分植被生态系统中阳光照射下的绿色和非绿色植被成分。
{"title":"Quasi-HSL color space and its application: Sunlit and shaded component fractional cover estimation in vegetated ecosystem","authors":"Jia Tian, Qingjiu Tian, Suju Li, Qianjing Li, Sen Zhang, Shuang He","doi":"10.1016/j.jag.2024.104298","DOIUrl":"https://doi.org/10.1016/j.jag.2024.104298","url":null,"abstract":"Sunlit and shaded components are commonly present in both airborne and satellite remote sensing images. In vegetated ecosystems, shaded component often result from sunlight being obstructed by topographic relief or canopy structures, and shaded component may impact plant growth, leaf photosynthesis, and ultimately carbon sequestration. To accurately estimate the fractional cover of the shaded and sunlit components, including both green and non-green vegetation within vegetated ecosystems, a novel method called the quasi-Hue-Saturation-Lightness (quasi-HSL) method is proposed in this study. Inspired by the RGB to HSL conversion, this method utilizes near-infrared, green, and red bands to compute hue (and normalized hue), saturation, and lightness. Subsequently, two indices, namely Hue-Lightness Index (HLI) and Saturation-Lightness Index (SLI), are introduced to construct a triangular space for estimating the fractional cover of the three components. Through unmanned aerial vehicle field experiments conducted in two forested areas, the accuracy of fractional cover estimation for three components reaches an R<ce:sup loc=\"post\">2</ce:sup> value of 0.50–0.67. Furthermore, this fractional cover estimation approach can be extended to a four-component estimation, including sunlit green vegetation, sunlit non-green vegetation, shaded green vegetation, and shaded non-green vegetation. With this detailed fractional cover estimation in vegetated area, the fractional vegetation coverage can be retrieved. Cross-validated with the fractional vegetation coverage retrieved by NDVI, the accuracy reaches R<ce:sup loc=\"post\">2</ce:sup> = 0.92. The advantages of the proposed method are (1) estimating fractional cover of shaded component without blue band, which is easily impacted by atmospheric conditions and sensor performance, and (2) differentiating the sunlit green and non-green vegetation components in the vegetated ecosystem.","PeriodicalId":50341,"journal":{"name":"International Journal of Applied Earth Observation and Geoinformation","volume":"92 1","pages":""},"PeriodicalIF":7.5,"publicationDate":"2024-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142793574","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：