Ecological Informatics最新文献_第5页

A novel statistical workflow using pollen records and regression kriging to reconstruct the spatially and temporally explicit demographic history of tree species 一种新的统计工作流程，利用花粉记录和回归克里格重建时空明确的树种人口统计历史

IF 7.3 2区环境科学与生态学 Q1 ECOLOGY

Ecological Informatics

Pub Date : 2025-12-19 DOI: 10.1016/j.ecoinf.2025.103561

Azzurra Pistone , Denis Allard , Christoph Schwörer , César Morales-Molino , Willy Tinner , Katalin Csilléry

Understanding the effects of past climate shifts on the demography of forest tree species is crucial to assessing their response to ongoing climate change. However, little effort has so far been made to quantify past demographic changes in a spatially and temporally explicit manner at a continental scale. We have developed a novel statistical workflow that integrates two regression kriging models to reconstruct the demographic history of tree species across Europe. Our workflow anticipates spatially the probability of species occurrence (PoO), and interpolates their relative abundances (RelAb) spatially and temporally. Climate variables can be included as covariates. Our approach can accommodate non-stationary species responses to climate, and incorporates the presence of source populations, colonization constraints, and population trends as factors influencing species RelAb. We applied this workflow to European fir species (Abies spp.) since the Last Glacial Maximum (LGM), using fossil pollen records from 241 sites, and simulated paleoclimate data on a 0.41-degree grid and 500-year time bins. Model performance, assessed with cross-validation, demonstrates that including climate as a covariate enhances the spatial heterogeneity. Climate has a positive effect on RelAb interpolation under millennial static spatial distribution structure conditions, while the presence of source populations plays a more important role during rapid demographic processes. Additionally, we applied our workflow to assess future regional changes in the RelAb of Abies spp. under the main future climate scenarios. Our workflow is particularly suited for temperate and boreal tree species and can be used in various downstream analyses.

了解过去气候变化对森林树种数量的影响对于评估它们对持续气候变化的反应至关重要。然而，迄今为止，在大陆范围内以明确的空间和时间方式量化过去的人口变化的努力很少。我们开发了一种新的统计工作流程，它集成了两种回归克里格模型来重建整个欧洲树种的人口统计历史。我们的工作流程在空间上预测物种发生的概率（PoO），并在空间和时间上插值它们的相对丰度（RelAb）。气候变量可以作为协变量包括在内。我们的方法可以适应物种对气候的非平稳响应，并将源种群的存在、定殖限制和种群趋势作为影响物种相关性的因素。我们将此工作流程应用于末次盛冰期（LGM）以来的欧洲冷杉物种（冷杉种），使用来自241个站点的化石花粉记录，并在0.41度网格和500年时间箱上模拟古气候数据。通过交叉验证评估的模型性能表明，将气候作为协变量增加了空间异质性。在千年静态空间分布结构条件下，气候对RelAb插值有积极影响，而在快速人口过程中，源种群的存在起着更重要的作用。此外，我们还应用我们的工作流程来评估冷杉属植物在未来主要气候情景下的区域变化。我们的工作流程特别适合温带和北方树种，可用于各种下游分析。

{"title":"A novel statistical workflow using pollen records and regression kriging to reconstruct the spatially and temporally explicit demographic history of tree species","authors":"Azzurra Pistone , Denis Allard , Christoph Schwörer , César Morales-Molino , Willy Tinner , Katalin Csilléry","doi":"10.1016/j.ecoinf.2025.103561","DOIUrl":"10.1016/j.ecoinf.2025.103561","url":null,"abstract":"<div><div>Understanding the effects of past climate shifts on the demography of forest tree species is crucial to assessing their response to ongoing climate change. However, little effort has so far been made to quantify past demographic changes in a spatially and temporally explicit manner at a continental scale. We have developed a novel statistical workflow that integrates two regression kriging models to reconstruct the demographic history of tree species across Europe. Our workflow anticipates spatially the probability of species occurrence (PoO), and interpolates their relative abundances (RelAb) spatially and temporally. Climate variables can be included as covariates. Our approach can accommodate non-stationary species responses to climate, and incorporates the presence of source populations, colonization constraints, and population trends as factors influencing species RelAb. We applied this workflow to European fir species (<em>Abies</em> spp.) since the Last Glacial Maximum (LGM), using fossil pollen records from 241 sites, and simulated paleoclimate data on a 0.41-degree grid and 500-year time bins. Model performance, assessed with cross-validation, demonstrates that including climate as a covariate enhances the spatial heterogeneity. Climate has a positive effect on RelAb interpolation under millennial static spatial distribution structure conditions, while the presence of source populations plays a more important role during rapid demographic processes. Additionally, we applied our workflow to assess future regional changes in the RelAb of <em>Abies</em> spp. under the main future climate scenarios. Our workflow is particularly suited for temperate and boreal tree species and can be used in various downstream analyses.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"93 ","pages":"Article 103561"},"PeriodicalIF":7.3,"publicationDate":"2025-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145925320","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Human Auditory Representation Learning for cross-dialect bird species recognition 跨方言鸟类物种识别的人类听觉表征学习

IF 7.3 2区环境科学与生态学 Q1 ECOLOGY

Ecological Informatics

Pub Date : 2025-12-18 DOI: 10.1016/j.ecoinf.2025.103554

Xingfeng Li , Ningfeng Luo , Feifei Yu , Junjie Li , Kai Li , Yongwei Li , Zhen Zhao , Yang Liu , Xiaohan Shi

Automated bird species recognition (BSR) is crucial for biodiversity monitoring, but its accuracy is often hampered by geographic variation in bird vocalizations, which is frequently described as dialectal at the population level. This work uses the term dialect region in an operational sense to refer to geographically separated recording regions defined in benchmark corpora that are dominated by different vocal variants across populations, rather than to fine-grained, species-specific dialect types. We aim to address the challenge of generalizing across such dialect-dominated regions by introducing a novel approach leveraging well-established auditory-inspired features known for their computational robustness. We propose Human Auditory Representation Learning (HARL), a framework that integrates Gammatone- and Mel-spectrogram features to capture frequency selectivity and invariance to acoustic variations through their spectral efficiency and empirical success in audio processing. These complementary auditory representations are processed by a dual-stream ResNet50 architecture, with a multi-head attention mechanism to emphasize discriminative spectral–temporal patterns. In cross-dialect evaluation on the D3BV benchmark and cross-site tests on two field datasets (S1 and S2), the approach outperformed strong baselines, raising F1-score by up to 24.31% on D3BV and 28.23% on S1 and S2. Performance remained stable across noise conditions from

-

10 to +10 dB signal-to-noise ratio, indicating robustness for real-world deployment. These findings showed that bridging HARL with deep learning delivered a scalable and accurate solution for biodiversity monitoring, enabling reliable species recognition across diverse geographies and acoustic conditions.

鸟类物种自动识别（BSR）对生物多样性监测至关重要，但其准确性往往受到鸟类发声的地理差异的影响，在种群水平上经常被描述为方言。这项工作在操作意义上使用了术语方言区域，指的是在基准语料库中定义的地理上分离的记录区域，这些区域由不同种群的不同声音变体主导，而不是细粒度的、物种特定的方言类型。我们的目标是通过引入一种利用成熟的听觉启发特征的新方法来解决跨越方言主导区域的挑战，这些特征以其计算稳健性而闻名。我们提出了人类听觉表征学习（Human Auditory Representation Learning， HARL），这是一个整合了伽玛酮和梅尔谱图特征的框架，通过其频谱效率和音频处理中的经验成功来捕获声音变化的频率选择性和不变性。这些互补的听觉表征由双流ResNet50架构处理，具有多头注意机制，以强调区分光谱-时间模式。在D3BV基准的跨方言评价和两个现场数据集（S1和S2）的跨站点测试中，该方法优于强基线，D3BV的f1得分最高提高24.31%，S1和S2的f1得分最高提高28.23%。从- 10到+10 dB信噪比的噪声条件下，性能保持稳定，表明在实际部署中的鲁棒性。这些发现表明，将HARL与深度学习相结合，为生物多样性监测提供了可扩展且准确的解决方案，能够在不同地理位置和声学条件下实现可靠的物种识别。

{"title":"Human Auditory Representation Learning for cross-dialect bird species recognition","authors":"Xingfeng Li , Ningfeng Luo , Feifei Yu , Junjie Li , Kai Li , Yongwei Li , Zhen Zhao , Yang Liu , Xiaohan Shi","doi":"10.1016/j.ecoinf.2025.103554","DOIUrl":"10.1016/j.ecoinf.2025.103554","url":null,"abstract":"<div><div>Automated bird species recognition (BSR) is crucial for biodiversity monitoring, but its accuracy is often hampered by geographic variation in bird vocalizations, which is frequently described as dialectal at the population level. This work uses the term dialect region in an operational sense to refer to geographically separated recording regions defined in benchmark corpora that are dominated by different vocal variants across populations, rather than to fine-grained, species-specific dialect types. We aim to address the challenge of generalizing across such dialect-dominated regions by introducing a novel approach leveraging well-established auditory-inspired features known for their computational robustness. We propose Human Auditory Representation Learning (HARL), a framework that integrates Gammatone- and Mel-spectrogram features to capture frequency selectivity and invariance to acoustic variations through their spectral efficiency and empirical success in audio processing. These complementary auditory representations are processed by a dual-stream ResNet50 architecture, with a multi-head attention mechanism to emphasize discriminative spectral–temporal patterns. In cross-dialect evaluation on the D3BV benchmark and cross-site tests on two field datasets (S1 and S2), the approach outperformed strong baselines, raising F1-score by up to 24.31% on D3BV and 28.23% on S1 and S2. Performance remained stable across noise conditions from <span><math><mo>−</mo></math></span>10 to +10 dB signal-to-noise ratio, indicating robustness for real-world deployment. These findings showed that bridging HARL with deep learning delivered a scalable and accurate solution for biodiversity monitoring, enabling reliable species recognition across diverse geographies and acoustic conditions.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"93 ","pages":"Article 103554"},"PeriodicalIF":7.3,"publicationDate":"2025-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145791013","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Effects of input-image size on performance of fish detection and species classification using deep learning 使用深度学习的输入图像大小对鱼类检测和物种分类性能的影响

IF 7.3 2区环境科学与生态学 Q1 ECOLOGY

Ecological Informatics

Pub Date : 2025-12-18 DOI: 10.1016/j.ecoinf.2025.103566

Yuka Iwahara , Yasutoki Shibata , Masahiro Manano , Tomoya Nishino , Ryosuke Kariya , Hiroki Yaemori

Deep learning has been extensively used in fisheries science, as it enables the acquisition of information regarding the body length and stock-abundance index of target fish from images, thereby facilitating stock assessment and management. However, generally, multiple species appear together in images obtained from fisheries, necessitating the classification of fish species before extracting relevant biological information. Improving the performance of fish detection and species classification is crucial as it affects the quality of biological information that could be inferred from images. Previous studies have reported that increasing the input-image size can affect the classification accuracy. Identification characteristics of fish are small in comparison with their body size, and increasing the image size can affect the classification accuracy; however, there are no reports on the effect of image size on fish species-classification accuracy. Herein, different input-image sizes were taken to evaluate the effect of input-image size on the performance of fish detection and species classification. Fish images (41,922 fish across 41 classes) were acquired from conveyor belts to sort set-net fish catches. Fish were detected and classified using a mask region-based convolutional neural network. The effect of input-image size on performance was examined using nine datasets in three image sizes of 1333 × 888, 2000 × 1333, and 2666 × 1777 pixels, obtaining an average mAP50–95 value of 0.586, 0.612, and 0.609, respectively. Larger image sizes offered improved performance compared with that of the smallest, averaging 0.026 and 0.023 improvements in mAP50–95 at two larger image sizes. However, when comparing the degree of improvement between image sizes of 2000 × 1333 pixels and 2666 × 1777 pixels under fine-tuning conditions, the former size resulted in higher performance. Performance was observed to improve for species with low performance at the smallest image size; therefore, we can say that increasing the input-image size is a simple and effective way for improving detection and classification performance for these species.

深度学习在渔业科学中得到了广泛的应用，它可以从图像中获取目标鱼的体长和种群丰度指数等信息，从而促进种群评估和管理。然而，通常在渔业获得的图像中，多个物种会同时出现，因此在提取相关生物信息之前，需要对鱼类进行分类。提高鱼类检测和物种分类的性能至关重要，因为它会影响从图像中推断的生物信息的质量。已有研究报道，增加输入图像的大小会影响分类精度。鱼类的识别特征与体型相比较小，增大图像尺寸会影响分类精度；然而，目前还没有关于图像大小对鱼类分类精度影响的报道。本文采用不同的输入图像大小来评估输入图像大小对鱼类检测和物种分类性能的影响。从传送带获取鱼类图像（41个类别的41,922条鱼），用于对渔网渔获物进行分类。使用基于掩模区域的卷积神经网络对鱼进行检测和分类。使用1333 × 888、2000 × 1333和2666 × 1777像素的9个数据集考察了输入图像大小对性能的影响，得到mAP50-95的平均值分别为0.586、0.612和0.609。与最小的图像尺寸相比，较大的图像尺寸提供了更好的性能，在两个较大的图像尺寸下，mAP50-95的平均性能提高了0.026和0.023。然而，当比较2000 × 1333像素和2666 × 1777像素的图像尺寸在微调条件下的改善程度时，前者的性能更高。在最小的图像尺寸下，性能较低的物种的性能有所提高；因此，我们可以说，增加输入图像的大小是提高这些物种的检测和分类性能的一种简单而有效的方法。

{"title":"Effects of input-image size on performance of fish detection and species classification using deep learning","authors":"Yuka Iwahara , Yasutoki Shibata , Masahiro Manano , Tomoya Nishino , Ryosuke Kariya , Hiroki Yaemori","doi":"10.1016/j.ecoinf.2025.103566","DOIUrl":"10.1016/j.ecoinf.2025.103566","url":null,"abstract":"<div><div>Deep learning has been extensively used in fisheries science, as it enables the acquisition of information regarding the body length and stock-abundance index of target fish from images, thereby facilitating stock assessment and management. However, generally, multiple species appear together in images obtained from fisheries, necessitating the classification of fish species before extracting relevant biological information. Improving the performance of fish detection and species classification is crucial as it affects the quality of biological information that could be inferred from images. Previous studies have reported that increasing the input-image size can affect the classification accuracy. Identification characteristics of fish are small in comparison with their body size, and increasing the image size can affect the classification accuracy; however, there are no reports on the effect of image size on fish species-classification accuracy. Herein, different input-image sizes were taken to evaluate the effect of input-image size on the performance of fish detection and species classification. Fish images (41,922 fish across 41 classes) were acquired from conveyor belts to sort set-net fish catches. Fish were detected and classified using a mask region-based convolutional neural network. The effect of input-image size on performance was examined using nine datasets in three image sizes of 1333 × 888, 2000 × 1333, and 2666 × 1777 pixels, obtaining an average mAP50–95 value of 0.586, 0.612, and 0.609, respectively. Larger image sizes offered improved performance compared with that of the smallest, averaging 0.026 and 0.023 improvements in mAP50–95 at two larger image sizes. However, when comparing the degree of improvement between image sizes of 2000 × 1333 pixels and 2666 × 1777 pixels under fine-tuning conditions, the former size resulted in higher performance. Performance was observed to improve for species with low performance at the smallest image size; therefore, we can say that increasing the input-image size is a simple and effective way for improving detection and classification performance for these species.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"93 ","pages":"Article 103566"},"PeriodicalIF":7.3,"publicationDate":"2025-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145925483","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Innovative daily runoff prediction model integrating black-winged kite algorithm and Mamba2–Transformer architecture 集成黑翼风筝算法和Mamba2-Transformer架构的创新日径流预测模型

IF 7.3 2区环境科学与生态学 Q1 ECOLOGY

Ecological Informatics

Pub Date : 2025-12-17 DOI: 10.1016/j.ecoinf.2025.103565

Dong-mei Xu, Xiao-xue Hu, Wen-chuan Wang, Jun Wang, Can-can Shi, Hong-fei Zang

The accurate prediction of daily runoff is crucial for effective water resource management, flood prevention, and disaster mitigation. However, current runoff prediction models face dual challenges in extracting discriminative features from random and non-stationary data, as well as enhancing prediction performance. To overcome these challenges and improve prediction accuracy, this study proposes a hybrid feature optimization and variational prediction model (HFOVPM). The HFOVPM combines an optimized feature extraction framework with advanced techniques of temporal pattern recognition. First, the Black-winged Kite Algorithm was employed to optimize variational mode decomposition parameters for effective signal decomposition. To enhance feature representation, the resultant components were systematically organized into multivariate input vectors. Secondly, the Mamba2 architecture was used to model nonlinear dynamical interactions within the multivariate inputs. Subsequently, its learned representations were then integrated into Transformer layers to establish enhanced global temporal dependencies. Through error back-propagation optimization, the framework improved the accuracy of predictions. The model was validated using daily runoff data from three hydrological stations representing distinct eco-hydrological regimes: an alpine snowmelt-dominated basin, a subtropical rainfall-driven basin, and a mixed rain-snow basin. The key findings were as follows: (1) The HFOVPM consistently outperformed several benchmark models across all basins, achieving Kling–Gupta efficiency values of 0.94–0.99 and maintaining the mean absolute percentage error below 14.2 %. (2) The model exhibited particular strength in predicting extreme runoff events, capturing flood peaks more accurately than comparative models, which is critical for flood risk early warning. (3) Its robust performance across contrasting climates indicates a reliable generalization capability for diverse hydrological processes. In summary, the HFOVPM effectively addresses the feature extraction and temporal dependency challenges in runoff prediction, providing an accurate and transferable tool that can support hydrological forecasting and water-related decision-making in varied environmental settings.

日径流量的准确预测对于有效的水资源管理、防洪和减灾至关重要。然而，目前的径流预测模型在从随机和非平稳数据中提取判别特征以及提高预测性能方面面临着双重挑战。为了克服这些挑战，提高预测精度，本研究提出了一种混合特征优化和变分预测模型（HFOVPM）。HFOVPM结合了优化的特征提取框架和先进的时间模式识别技术。首先，采用黑翼风筝算法对变分模态分解参数进行优化，实现有效的信号分解；为了增强特征表示，结果组件被系统地组织成多变量输入向量。其次，利用Mamba2体系结构对多变量输入内的非线性动态相互作用进行建模。随后，将其学习到的表示集成到Transformer层中，以建立增强的全局时间依赖性。通过误差反向传播优化，提高了预测精度。该模型使用三个水文站的日径流数据进行了验证，这些水文站代表了不同的生态水文制度：高山融雪主导的盆地，亚热带降雨驱动的盆地和雨雪混合盆地。结果表明：(1)HFOVPM在所有流域均优于多个基准模型，克林-古普塔效率值为0.94 ~ 0.99，平均绝对百分比误差保持在14.2%以下。(2)该模型在预测极端径流事件方面表现出较强的能力，能较准确地捕捉洪峰，对洪水风险预警具有重要意义。(3)其在不同气候条件下的鲁棒性表明其对不同水文过程具有可靠的泛化能力。总之，HFOVPM有效地解决了径流预测中的特征提取和时间依赖性挑战，提供了一个准确和可转移的工具，可以支持各种环境下的水文预测和与水相关的决策。

{"title":"Innovative daily runoff prediction model integrating black-winged kite algorithm and Mamba2–Transformer architecture","authors":"Dong-mei Xu, Xiao-xue Hu, Wen-chuan Wang, Jun Wang, Can-can Shi, Hong-fei Zang","doi":"10.1016/j.ecoinf.2025.103565","DOIUrl":"10.1016/j.ecoinf.2025.103565","url":null,"abstract":"<div><div>The accurate prediction of daily runoff is crucial for effective water resource management, flood prevention, and disaster mitigation. However, current runoff prediction models face dual challenges in extracting discriminative features from random and non-stationary data, as well as enhancing prediction performance. To overcome these challenges and improve prediction accuracy, this study proposes a hybrid feature optimization and variational prediction model (HFOVPM). The HFOVPM combines an optimized feature extraction framework with advanced techniques of temporal pattern recognition. First, the Black-winged Kite Algorithm was employed to optimize variational mode decomposition parameters for effective signal decomposition. To enhance feature representation, the resultant components were systematically organized into multivariate input vectors. Secondly, the Mamba2 architecture was used to model nonlinear dynamical interactions within the multivariate inputs. Subsequently, its learned representations were then integrated into Transformer layers to establish enhanced global temporal dependencies. Through error back-propagation optimization, the framework improved the accuracy of predictions. The model was validated using daily runoff data from three hydrological stations representing distinct eco-hydrological regimes: an alpine snowmelt-dominated basin, a subtropical rainfall-driven basin, and a mixed rain-snow basin. The key findings were as follows: (1) The HFOVPM consistently outperformed several benchmark models across all basins, achieving Kling–Gupta efficiency values of 0.94–0.99 and maintaining the mean absolute percentage error below 14.2 %. (2) The model exhibited particular strength in predicting extreme runoff events, capturing flood peaks more accurately than comparative models, which is critical for flood risk early warning. (3) Its robust performance across contrasting climates indicates a reliable generalization capability for diverse hydrological processes. In summary, the HFOVPM effectively addresses the feature extraction and temporal dependency challenges in runoff prediction, providing an accurate and transferable tool that can support hydrological forecasting and water-related decision-making in varied environmental settings.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"93 ","pages":"Article 103565"},"PeriodicalIF":7.3,"publicationDate":"2025-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145791014","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Spatiotemporal analysis of long-term vegetation dynamics using KNDVI and machine learning-based multifactor analysis 基于KNDVI和基于机器学习的多因素分析的植被长期动态时空分析

IF 7.3 2区环境科学与生态学 Q1 ECOLOGY

Ecological Informatics

Pub Date : 2025-12-14 DOI: 10.1016/j.ecoinf.2025.103559

Ying Liu , Mou Leong Tan , Zaher Mundher Yaseen , Fei Zhang

In the context of accelerated climate change and intensified human activities, exploring vegetation responses to diverse environmental stressors is essential for understanding ecosystem resilience and long-term sustainability. This study focuses on the Jialing River Basin in China to evaluate vegetation changes and their driving mechanisms during the growing season from 2000 to 2024. The Kernel Normalized Difference Vegetation Index (KNDVI), computed using the Google Earth Engine platform, was used to characterize vegetation dynamics. A Random Forest model, coupled with SHapley Additive exPlanations (SHAP), was then employed to quantify the effects of topography, climate, soil texture, air pollutants, and land-use attributes on vegetation dynamics. The Random Forest model demonstrated robust performance (training R² = 0.774–0.871; testing R² = 0.690–0.826; RMSE = 0.038), with minimal overfitting (4.45–8.43 %), indicating strong generalization in capturing KNDVI–environment relationships. Results indicated that: (1) the mean annual KNDVI was 0.263, exhibiting a north–south gradient with higher values in the north. Vegetation degradation occurred in 52.53 % of the basin (85,371.56 km²), mainly in urban and agricultural areas; (2) the Hurst exponent suggested unstable future trends, with both improvement and degradation; and (3) SHAP analysis revealed that vegetation dynamics in the Jialing River Basin were mainly governed by elevation and land use, while climatic factors, particularly temperature, solar radiation, and precipitation, have increasingly influenced vegetation patterns under accelerated warming. The impacts of these factors were characterized by significant nonlinear responses, interactive effects, and threshold behaviors. These findings deepen our understanding of vegetation dynamics in topographically complex basins and highlight the intertwined roles of climate, topography, and anthropogenic factors in shaping ecosystem changes.

在气候变化加速和人类活动加剧的背景下，探索植被对各种环境胁迫的响应对于理解生态系统的恢复力和长期可持续性至关重要。以嘉陵江流域为研究对象，对2000 - 2024年生长季植被变化及其驱动机制进行了评价。利用谷歌Earth Engine平台计算的核归一化植被指数（KNDVI）来表征植被动态。然后采用随机森林模型，结合SHapley加性解释（SHAP），量化了地形、气候、土壤质地、空气污染物和土地利用属性对植被动态的影响。随机森林模型表现出鲁棒性（训练R2 = 0.774-0.871；检验R2 = 0.690-0.826; RMSE = 0.038），过拟合最小（4.45 - 8.43%），表明在捕获kndvi -环境关系方面具有很强的泛化能力。结果表明：(1)年平均KNDVI为0.263，呈南北梯度，北部较高；流域植被退化面积达52.53% (85,371.56 km2)，主要集中在城市和农业区；(2) Hurst指数表明未来趋势不稳定，既有改善趋势，也有退化趋势；(3) SHAP分析表明，嘉陵江流域植被动态主要受高程和土地利用的支配，而气候因子，尤其是温度、太阳辐射和降水对加速增温条件下植被格局的影响越来越大。这些因素的影响表现为显著的非线性响应、交互效应和阈值行为。这些发现加深了我们对地形复杂盆地植被动态的理解，并突出了气候、地形和人为因素在塑造生态系统变化中的相互交织的作用。

{"title":"Spatiotemporal analysis of long-term vegetation dynamics using KNDVI and machine learning-based multifactor analysis","authors":"Ying Liu , Mou Leong Tan , Zaher Mundher Yaseen , Fei Zhang","doi":"10.1016/j.ecoinf.2025.103559","DOIUrl":"10.1016/j.ecoinf.2025.103559","url":null,"abstract":"<div><div>In the context of accelerated climate change and intensified human activities, exploring vegetation responses to diverse environmental stressors is essential for understanding ecosystem resilience and long-term sustainability. This study focuses on the Jialing River Basin in China to evaluate vegetation changes and their driving mechanisms during the growing season from 2000 to 2024. The Kernel Normalized Difference Vegetation Index (KNDVI), computed using the Google Earth Engine platform, was used to characterize vegetation dynamics. A Random Forest model, coupled with SHapley Additive exPlanations (SHAP), was then employed to quantify the effects of topography, climate, soil texture, air pollutants, and land-use attributes on vegetation dynamics. The Random Forest model demonstrated robust performance (training R<sup>2</sup> = 0.774–0.871; testing R<sup>2</sup> = 0.690–0.826; RMSE = 0.038), with minimal overfitting (4.45–8.43 %), indicating strong generalization in capturing KNDVI–environment relationships. Results indicated that: (1) the mean annual KNDVI was 0.263, exhibiting a north–south gradient with higher values in the north. Vegetation degradation occurred in 52.53 % of the basin (85,371.56 km<sup>2</sup>), mainly in urban and agricultural areas; (2) the Hurst exponent suggested unstable future trends, with both improvement and degradation; and (3) SHAP analysis revealed that vegetation dynamics in the Jialing River Basin were mainly governed by elevation and land use, while climatic factors, particularly temperature, solar radiation, and precipitation, have increasingly influenced vegetation patterns under accelerated warming. The impacts of these factors were characterized by significant nonlinear responses, interactive effects, and threshold behaviors. These findings deepen our understanding of vegetation dynamics in topographically complex basins and highlight the intertwined roles of climate, topography, and anthropogenic factors in shaping ecosystem changes.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"93 ","pages":"Article 103559"},"PeriodicalIF":7.3,"publicationDate":"2025-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145925484","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Machine learning and species distribution models for crops: A review 机器学习与作物物种分布模型研究进展

IF 7.3 2区环境科学与生态学 Q1 ECOLOGY

Ecological Informatics

Pub Date : 2025-12-13 DOI: 10.1016/j.ecoinf.2025.103563

Emanuele Serra , Marta Debolini , Helder Fraga , Antonio Trabucco , Valentina Mereu , Donatella Spano

The agricultural sector is facing enormous challenges to sustain food security, with rising pressures from population growth, climate change, water scarcity and land degradation. In this regard, Agricultural Land Suitability Analysis (ALSA) serves as a fundamental tool in identifying the most appropriate areas for crop cultivation and optimizing land use planning and spatial crop allocation. Recent advances in Artificial Intelligence (AI), particularly Machine Learning (ML) and Species Distribution Models (SDMs), have significantly enhanced ALSA by improving prediction robustness and accuracy. In this regard, this scoping review aims to provide an overview of current trends of the most applied applications in ALSA concerning the most relevant crops for human subsistence and prosperity, in order to identify possible future development and research needs. This research focuses on feature space-based machine learning application to infer cash crop suitability. Following PRISMA guidelines and the Population Concept Context (PCC) framework, a qualitative-quantitative analysis of 113 peer-reviewed articles was performed to provide a comprehensive overview of the theme, showing a significant increase in interest in this area since 2021. Among 55 studied crops, rice, coffee, and wheat were the most frequently analyzed. The most commonly used model was Maximum Entropy (MaxEnt), followed by Random Forest (RF), with limited applications of modelling ensemble approaches. Environmental variables considered in studies are mostly bioclimatic, followed by topographic and pedological factors, with limited socio-economic thematic consideration. Climate change scenarios were incorporated in 63.7% of studies, with Representative Concentration Pathways (RCP) and Shared Socioeconomic Pathways (SSP) scenarios being considered in 50.6% and 43.3%, respectively. Findings highlight the growing interest for Artificial Intelligence in ALSA, and emphasize the need to integrate a larger spectrum of variables, especially socioeconomic ones, and multi-model ensemble approaches to enhance model robustness and ensure more reliable assessments for a wider range of crops.

由于人口增长、气候变化、水资源短缺和土地退化带来的压力越来越大，农业部门在维持粮食安全方面面临巨大挑战。在这方面，农业用地适宜性分析（ALSA）是确定最适宜作物种植区域、优化土地利用规划和空间作物配置的基本工具。人工智能（AI）的最新进展，特别是机器学习（ML）和物种分布模型（SDMs），通过提高预测的鲁棒性和准确性，显着增强了als。因此，本文旨在综述与人类生存和繁荣最相关的作物在als中最广泛应用的当前趋势，以确定未来可能的发展和研究需求。本研究的重点是基于特征空间的机器学习应用来推断经济作物的适宜性。根据PRISMA指南和人口概念背景（PCC）框架，对113篇同行评议文章进行了定性-定量分析，以全面概述该主题，显示自2021年以来对该领域的兴趣显着增加。在55种被研究的作物中，水稻、咖啡和小麦是最常被分析的。最常用的模型是最大熵（MaxEnt），其次是随机森林（RF），建模集成方法的应用有限。研究中考虑的环境变量主要是生物气候，其次是地形和土壤因素，有限的社会经济主题考虑。63.7%的研究纳入了气候变化情景，分别有50.6%和43.3%的研究考虑了代表性浓度路径（RCP）和共享社会经济路径（SSP）情景。研究结果强调了人工智能在als研究中的日益增长的兴趣，并强调需要整合更大范围的变量，特别是社会经济变量，以及多模型集成方法，以提高模型的鲁棒性，并确保对更广泛的作物进行更可靠的评估。

{"title":"Machine learning and species distribution models for crops: A review","authors":"Emanuele Serra , Marta Debolini , Helder Fraga , Antonio Trabucco , Valentina Mereu , Donatella Spano","doi":"10.1016/j.ecoinf.2025.103563","DOIUrl":"10.1016/j.ecoinf.2025.103563","url":null,"abstract":"<div><div>The agricultural sector is facing enormous challenges to sustain food security, with rising pressures from population growth, climate change, water scarcity and land degradation. In this regard, Agricultural Land Suitability Analysis (ALSA) serves as a fundamental tool in identifying the most appropriate areas for crop cultivation and optimizing land use planning and spatial crop allocation. Recent advances in Artificial Intelligence (AI), particularly Machine Learning (ML) and Species Distribution Models (SDMs), have significantly enhanced ALSA by improving prediction robustness and accuracy. In this regard, this scoping review aims to provide an overview of current trends of the most applied applications in ALSA concerning the most relevant crops for human subsistence and prosperity, in order to identify possible future development and research needs. This research focuses on feature space-based machine learning application to infer cash crop suitability. Following PRISMA guidelines and the Population Concept Context (PCC) framework, a qualitative-quantitative analysis of 113 peer-reviewed articles was performed to provide a comprehensive overview of the theme, showing a significant increase in interest in this area since 2021. Among 55 studied crops, rice, coffee, and wheat were the most frequently analyzed. The most commonly used model was Maximum Entropy (MaxEnt), followed by Random Forest (RF), with limited applications of modelling ensemble approaches. Environmental variables considered in studies are mostly bioclimatic, followed by topographic and pedological factors, with limited socio-economic thematic consideration. Climate change scenarios were incorporated in 63.7% of studies, with Representative Concentration Pathways (RCP) and Shared Socioeconomic Pathways (SSP) scenarios being considered in 50.6% and 43.3%, respectively. Findings highlight the growing interest for Artificial Intelligence in ALSA, and emphasize the need to integrate a larger spectrum of variables, especially socioeconomic ones, and multi-model ensemble approaches to enhance model robustness and ensure more reliable assessments for a wider range of crops.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"93 ","pages":"Article 103563"},"PeriodicalIF":7.3,"publicationDate":"2025-12-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145791098","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Deep learning-based canopy gap detection using a cross-technological approach with airborne laser scanning and aerial imagery data 基于深度学习的基于机载激光扫描和航空图像数据的跨技术方法的冠层间隙检测

IF 7.3 2区环境科学与生态学 Q1 ECOLOGY

Ecological Informatics

Pub Date : 2025-12-12 DOI: 10.1016/j.ecoinf.2025.103558

Florian Franz , Dominik Seidel , Philip Beckschäfer

Canopy gaps are crucial structural elements of forests, supporting biodiversity and influencing forest dynamics and ecosystem health. Airborne laser scanning (ALS) is commonly used for forest gap analysis and typically outperforms digital aerial photogrammetry (DAP), especially in detecting smaller gaps. However, ALS data availability remains limited compared to DAP. Given the broader availability and cost-effectiveness of DAP, this study aimed to overcome its technical drawbacks in canopy gap detection by applying a cross-technological approach with multiple data sources. This involves ALS-derived reference data fused with spectral and height information from DAP. We developed a deep learning-based method, employing a convolutional neural network (CNN), specifically the U-Net architecture, for detecting canopy gaps. The U-Net was trained using gap polygons automatically generated from ALS-derived canopy height models (CHMs), combined with true digital orthophotos (TDOPs) and DAP-based CHMs. Adding spectral information from TDOPs was intended to help detect shadows typically associated with smaller canopy gaps, which are often missed in DAP-based CHMs. The model was tested in the Solling, a forest area in a low mountain range in Central Germany. Performance was evaluated in independent test areas representing a gradient of structural heterogeneity. Overall, our model achieved moderate to high segmentation performance (IoU: 0.67–0.77; F1-score: 0.56–0.74). Once trained, it can be applied to image-derived inputs, improving canopy gap detection F1-score by on average 0.08 compared to using DAP-based CHMs alone. Our results demonstrate a novel approach for detecting canopy gaps without ALS data, suggesting applications across broader spatial and temporal scales.

林冠间隙是森林的重要结构要素，支持生物多样性，影响森林动态和生态系统健康。机载激光扫描（ALS）通常用于森林林隙分析，通常优于数字航空摄影测量（DAP），特别是在探测较小的林隙方面。然而，与DAP相比，ALS的数据可用性仍然有限。考虑到DAP的广泛可用性和成本效益，本研究旨在通过应用多数据源的跨技术方法来克服其在冠层间隙检测中的技术缺陷。这涉及als衍生的参考数据与DAP的光谱和高度信息融合。我们开发了一种基于深度学习的方法，使用卷积神经网络（CNN），特别是U-Net架构来检测树冠间隙。U-Net使用由als衍生的冠层高度模型（CHMs）自动生成的间隙多边形，结合真实数字正射影像（TDOPs）和基于dap的CHMs进行训练。添加来自TDOPs的光谱信息旨在帮助检测通常与较小冠层间隙相关的阴影，这些阴影在基于dap的CHMs中经常被遗漏。该模型在德国中部低山脉的索林森林地区进行了测试。性能在独立的测试区域进行评估，代表结构异质性的梯度。总体而言，我们的模型实现了中等到较高的分割性能（IoU: 0.67-0.77; F1-score: 0.56-0.74）。经过训练后，它可以应用于图像衍生的输入，与单独使用基于dap的CHMs相比，它可以将冠层间隙检测的f1分数平均提高0.08分。我们的研究结果展示了一种新的方法来检测树冠间隙没有ALS数据，建议应用在更广泛的空间和时间尺度。

{"title":"Deep learning-based canopy gap detection using a cross-technological approach with airborne laser scanning and aerial imagery data","authors":"Florian Franz , Dominik Seidel , Philip Beckschäfer","doi":"10.1016/j.ecoinf.2025.103558","DOIUrl":"10.1016/j.ecoinf.2025.103558","url":null,"abstract":"<div><div>Canopy gaps are crucial structural elements of forests, supporting biodiversity and influencing forest dynamics and ecosystem health. Airborne laser scanning (ALS) is commonly used for forest gap analysis and typically outperforms digital aerial photogrammetry (DAP), especially in detecting smaller gaps. However, ALS data availability remains limited compared to DAP. Given the broader availability and cost-effectiveness of DAP, this study aimed to overcome its technical drawbacks in canopy gap detection by applying a cross-technological approach with multiple data sources. This involves ALS-derived reference data fused with spectral and height information from DAP. We developed a deep learning-based method, employing a convolutional neural network (CNN), specifically the U-Net architecture, for detecting canopy gaps. The U-Net was trained using gap polygons automatically generated from ALS-derived canopy height models (CHMs), combined with true digital orthophotos (TDOPs) and DAP-based CHMs. Adding spectral information from TDOPs was intended to help detect shadows typically associated with smaller canopy gaps, which are often missed in DAP-based CHMs. The model was tested in the Solling, a forest area in a low mountain range in Central Germany. Performance was evaluated in independent test areas representing a gradient of structural heterogeneity. Overall, our model achieved moderate to high segmentation performance (IoU: 0.67–0.77; F1-score: 0.56–0.74). Once trained, it can be applied to image-derived inputs, improving canopy gap detection F1-score by on average 0.08 compared to using DAP-based CHMs alone. Our results demonstrate a novel approach for detecting canopy gaps without ALS data, suggesting applications across broader spatial and temporal scales.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"93 ","pages":"Article 103558"},"PeriodicalIF":7.3,"publicationDate":"2025-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145791011","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Drivers of forest fires under fire management policies: A comparison of machine learning models based on multi-temporal MODIS fire data 火灾管理政策下的森林火灾驱动因素：基于多时相MODIS火灾数据的机器学习模型比较

IF 7.3 2区环境科学与生态学 Q1 ECOLOGY

Ecological Informatics

Pub Date : 2025-12-11 DOI: 10.1016/j.ecoinf.2025.103555

Lingling Tian , Yunlin Zhang , Jibin Ning , Guang Yang

Forest fires across China are influenced by various factors due to the different forest fire management policies. This study utilized MODIS fire point data (2001−2020) from Guizhou Province. The datasets were divided into two periods based on the implementation of the new Forest Fire Prevention Regulations (2001–2008 and 2009–2020), and the spatiotemporal dynamics of forest fire points were analyzed. By integrating meteorological, topographic, vegetation, anthropogenic, and socioeconomic driving factors, the differences in the main drivers across periods were analyzed. Multiple machine learning models were built to determine the optimal model to generate regional fire risk maps. Our analysis reveals that over 80 % of forest fires occurred from January to April from 2001 to 2008, and fire counts trended upwards. Conversely, from 2009 to 2020, approximately 85 % of fires remained concentrated during these months, but the frequency of fires showed a steady decline. Meteorological, socioeconomic, and vegetation factors were the main drivers of fire occurrence in both periods. The Random Forest model achieved optimal performance in both periods, with an accuracy over 88.30 % and Area Under Curve value ≥0.953, significantly outperforming eXtreme Gradient Boosting, Support Vector Machines, and Artificial Neural Networks. Across the two periods, the probability of forest fire occurrence was highest in spring and lowest in winter. This study revealed the main drivers of forest fires in Guizhou across different periods and built an optimal prediction model, thereby providing a scientific basis for forest fire management departments to conduct forest fire prevention, control, zoning, and other management work. These findings are vital for protecting forest resources, maintaining ecological and environmental security, and safeguarding human lives.

由于不同的森林火灾管理政策，中国各地的森林火灾受到多种因素的影响。本研究利用贵州省MODIS火点数据（2001 - 2020）。基于新《森林防火条例》的实施，将数据集分为2001-2008年和2009-2020年两个时期，分析了森林火点的时空动态。综合考虑气象、地形、植被、人为和社会经济驱动因素，分析了不同时期主要驱动因素的差异。建立了多个机器学习模型，以确定生成区域火灾风险图的最优模型。分析表明，2001 - 2008年，80%以上的森林火灾发生在1 - 4月，火灾数量呈上升趋势。相反，从2009年到2020年，大约85%的火灾集中在这几个月，但火灾频率呈稳步下降趋势。气象、社会经济和植被因素是两个时期火灾发生的主要驱动因素。随机森林模型在两个时间段都取得了最优的性能，准确率超过88.30%，曲线下面积值≥0.953，显著优于极端梯度增强、支持向量机和人工神经网络。两个时期森林火灾发生概率春季最高，冬季最低。本研究揭示了贵州不同时期森林火灾的主要驱动因素，并建立了最优预测模型，为森林消防管理部门开展森林防火、控制、分区等管理工作提供科学依据。这些发现对保护森林资源、维护生态环境安全、保障人类生命安全具有重要意义。

{"title":"Drivers of forest fires under fire management policies: A comparison of machine learning models based on multi-temporal MODIS fire data","authors":"Lingling Tian , Yunlin Zhang , Jibin Ning , Guang Yang","doi":"10.1016/j.ecoinf.2025.103555","DOIUrl":"10.1016/j.ecoinf.2025.103555","url":null,"abstract":"<div><div>Forest fires across China are influenced by various factors due to the different forest fire management policies. This study utilized MODIS fire point data (2001−2020) from Guizhou Province. The datasets were divided into two periods based on the implementation of the new Forest Fire Prevention Regulations (2001–2008 and 2009–2020), and the spatiotemporal dynamics of forest fire points were analyzed. By integrating meteorological, topographic, vegetation, anthropogenic, and socioeconomic driving factors, the differences in the main drivers across periods were analyzed. Multiple machine learning models were built to determine the optimal model to generate regional fire risk maps. Our analysis reveals that over 80 % of forest fires occurred from January to April from 2001 to 2008, and fire counts trended upwards. Conversely, from 2009 to 2020, approximately 85 % of fires remained concentrated during these months, but the frequency of fires showed a steady decline. Meteorological, socioeconomic, and vegetation factors were the main drivers of fire occurrence in both periods. The Random Forest model achieved optimal performance in both periods, with an accuracy over 88.30 % and Area Under Curve value ≥0.953, significantly outperforming eXtreme Gradient Boosting, Support Vector Machines, and Artificial Neural Networks. Across the two periods, the probability of forest fire occurrence was highest in spring and lowest in winter. This study revealed the main drivers of forest fires in Guizhou across different periods and built an optimal prediction model, thereby providing a scientific basis for forest fire management departments to conduct forest fire prevention, control, zoning, and other management work. These findings are vital for protecting forest resources, maintaining ecological and environmental security, and safeguarding human lives.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"93 ","pages":"Article 103555"},"PeriodicalIF":7.3,"publicationDate":"2025-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145791015","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A 3D data processing pipeline to automatically estimate tree dendrometric parameters from a single mobile phone video 一个三维数据处理流水线，自动估计树木的树形参数，从一个单一的手机视频

IF 7.3 2区环境科学与生态学 Q1 ECOLOGY

Ecological Informatics

Pub Date : 2025-12-11 DOI: 10.1016/j.ecoinf.2025.103552

Joris Ravaglia , Franck Hétroy-Wheeler , Pierre-Alexis Herrault , Philip M. Wheeler

Measurement of tree dendrometric parameters, such as height or trunk diameter at breast height (DBH), has been facilitated in the last decade using handheld devices such as smartphones or tablet computers. However, the solutions so far often require manual interaction, specific expertise or advanced technology. We present a simple and fully automatic method to compute the height, DBH and crown volume of individual urban trees using a single RGB video taken with any kind of device. It uses Structure-from-Motion to build a 3D point cloud from the video, scale it, isolate the tree within the cloud and fit geometric models to the trunk and the crown. In testing with a variety of tree taxa and sizes, the method accurately measures DBH and height and is robust to most environmental and video-recording parameters. This makes it suitable for use by expert and non-expert surveyors and a wide range of applications, including in citizen or community science.

在过去十年中，使用手持设备（如智能手机或平板电脑）方便地测量树木的树形参数，如高度或胸径（DBH）。然而，迄今为止的解决方案通常需要手动交互、特定的专业知识或先进的技术。我们提出了一种简单且全自动的方法来计算单个城市树木的高度、胸径和树冠体积，使用任何类型的设备拍摄的单个RGB视频。它使用Structure-from-Motion从视频中构建一个3D点云，对其进行缩放，在云中隔离树，并将几何模型拟合到树干和树冠上。在对各种树木分类群和大小的测试中，该方法准确地测量了胸径和高度，并且对大多数环境和视频记录参数具有鲁棒性。这使得它适用于专家和非专家测量人员以及广泛的应用，包括公民或社区科学。

引用次数: 0

GIS-based assessment of sika deer (Cervus nippon) habitat suitability and human conflict risk using integrated analytical hierarchy process-multi-criteria decision analysis 基于gis的梅花鹿生境适宜性与人类冲突风险综合评价——层次分析法-多准则决策分析

IF 7.3 2区环境科学与生态学 Q1 ECOLOGY

Ecological Informatics

Pub Date : 2025-12-11 DOI: 10.1016/j.ecoinf.2025.103562

Santa Pandit , Kazuo Oki , Timothy Dube , Satoko Morofuji , Salem Ibrahim Salem

This study employed geospatial techniques to determine habitat suitability for sika deer (Cervus nippon) and evaluated the risk of human-deer conflict (HDC) in the Hatase district of Taki Town, Japan. The assessment of deer habitat appropriateness and HDC risk involved a structured four-step approach: criterion selection, decision hierarchy development, expert ranking collection, and Analytical Hierarchy Process (AHP) for variable weight assignment. Habitat suitability was categorized into land factors (elevation, slope, land use, and agricultural land area) and environmental factors (proximity to rivers, total wetness index (TWI), land surface temperature (LST), normalized difference vegetation index (NDVI), and precipitation). The risk of HDC was assessed using proximity (distance to deer occurrence, forests, fences, roads, and croplands) and exposure metrics (habitat suitability, population density, and settlement density). Insights were derived from ten experts in diverse fields. The results revealed that environmental factors played a more significant role than land-based parameters in determining habitat suitability for the species. Among these, NDVI was the most influential factor, followed by precipitation and land use. The risk assessment identified habitat suitability as the primary driver of HDC potential, with settlement and population density as notable secondary contributors. In the risk matrix, the exposure metrics substantially outweighed the proximity metrics. Spatial analysis of habitat suitability revealed a heterogeneous distribution, with moderate suitability covering approximately 45 % of the study area. The HDC risk assessment further highlighted conflict hotspots in the northern and eastern regions, where optimal deer habitats overlap with human settlements. The integration of these analyses provides a robust spatial framework for targeted conservation planning and the development of conflict mitigation strategies. These findings enhance our understanding of sika deer ecology within human-modified landscapes and provide practical guidance for wildlife management, particularly in regions experiencing intensified human–wildlife interactions due to evolving socioecological conditions.

本研究采用地理空间技术确定了梅花鹿（Cervus nippon）栖息地的适宜性，并评估了日本泷镇Hatase地区人鹿冲突（HDC）的风险。鹿群生境适宜性和HDC风险的评估采用了标准选择、决策层次建立、专家排名收集和层次分析法（AHP）四步评价方法。生境适宜性分为土地因子（高程、坡度、土地利用和农用地面积）和环境因子（靠近河流、总湿度指数（TWI）、地表温度（LST）、归一化植被指数（NDVI）和降水）。利用邻近性（与鹿发生的距离、森林、围栏、道路和农田）和暴露指标（生境适宜性、人口密度和定居密度）评估HDC风险。见解来自不同领域的10位专家。结果表明，环境因子对生境适宜性的影响大于陆地因子。其中，NDVI的影响最大，其次是降水和土地利用。生境适宜性是HDC潜力的主要驱动因素，聚落和人口密度是次要驱动因素。在风险矩阵中，暴露度量大大超过了接近度量。生境适宜性空间分布呈异质性，中等适宜性约占研究区面积的45%。HDC风险评估进一步强调了北部和东部地区的冲突热点，在这些地区，鹿的最佳栖息地与人类住区重叠。这些分析的整合为有针对性的保护规划和制定冲突缓解战略提供了一个强有力的空间框架。这些发现增强了我们对人类改造景观中梅花鹿生态的理解，并为野生动物管理提供了实用指导，特别是在社会生态条件不断变化导致人类与野生动物互动加剧的地区。

{"title":"GIS-based assessment of sika deer (Cervus nippon) habitat suitability and human conflict risk using integrated analytical hierarchy process-multi-criteria decision analysis","authors":"Santa Pandit , Kazuo Oki , Timothy Dube , Satoko Morofuji , Salem Ibrahim Salem","doi":"10.1016/j.ecoinf.2025.103562","DOIUrl":"10.1016/j.ecoinf.2025.103562","url":null,"abstract":"<div><div>This study employed geospatial techniques to determine habitat suitability for sika deer (<em>Cervus nippon</em>) and evaluated the risk of human-deer conflict (HDC) in the Hatase district of Taki Town, Japan. The assessment of deer habitat appropriateness and HDC risk involved a structured four-step approach: criterion selection, decision hierarchy development, expert ranking collection, and Analytical Hierarchy Process (AHP) for variable weight assignment. Habitat suitability was categorized into land factors (elevation, slope, land use, and agricultural land area) and environmental factors (proximity to rivers, total wetness index (TWI), land surface temperature (LST), normalized difference vegetation index (NDVI), and precipitation). The risk of HDC was assessed using proximity (distance to deer occurrence, forests, fences, roads, and croplands) and exposure metrics (habitat suitability, population density, and settlement density). Insights were derived from ten experts in diverse fields. The results revealed that environmental factors played a more significant role than land-based parameters in determining habitat suitability for the species. Among these, NDVI was the most influential factor, followed by precipitation and land use. The risk assessment identified habitat suitability as the primary driver of HDC potential, with settlement and population density as notable secondary contributors. In the risk matrix, the exposure metrics substantially outweighed the proximity metrics. Spatial analysis of habitat suitability revealed a heterogeneous distribution, with moderate suitability covering approximately 45 % of the study area. The HDC risk assessment further highlighted conflict hotspots in the northern and eastern regions, where optimal deer habitats overlap with human settlements. The integration of these analyses provides a robust spatial framework for targeted conservation planning and the development of conflict mitigation strategies. These findings enhance our understanding of sika deer ecology within human-modified landscapes and provide practical guidance for wildlife management, particularly in regions experiencing intensified human–wildlife interactions due to evolving socioecological conditions.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"93 ","pages":"Article 103562"},"PeriodicalIF":7.3,"publicationDate":"2025-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145976636","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0