Ecological Informatics最新文献

{"title":"BirdRecon: A free open source tool for image based bird species recognition","authors":"Hari Kishan Kondaveeti ,&nbsp;Nabin Kumar Upadhaya ,&nbsp;Dheeraj Sai Tukkugudam ,&nbsp;Rahul Panigrahi ,&nbsp;Sirivella Madhan Chandra Mouli ,&nbsp;Valli Kumari Vatsavayi ,&nbsp;Nagendra Panini Challa","doi":"10.1016/j.ecoinf.2025.103193","DOIUrl":"10.1016/j.ecoinf.2025.103193","url":null,"abstract":"<div><div>Automated bird species recognition is a critical, yet challenging task, particularly for the systems aimed at supporting ornithologists, conservationists, and bird enthusiasts. This study introduces BirdRecon, an open-source bird species recognition system developed to enhance birdwatching, ornithological research, and biodiversity conservation. The system leverages a soft voting ensemble of four pretrained deep learning models—DenseNet201, EfficientNetB7, InceptionV3, and ResNet50V2—to improve classification accuracy and robustness. To address class imbalance problem and enhance generalization, data augmentation is applied and an early stopping optimization strategy is used to prevent overfitting during training. A benchmark dataset comprising 525 bird species with over 84,000 training images is used to evaluate the system. The experimental results demonstrate that the proposed ensemble model achieves a classification accuracy of 99.6%, precision of 99.7%, and recall of 99.6%, outperforming the existing state-of-the-art methods by a margin of 0.51%.</div><div>BirdRecon is implemented as both a web and mobile application, offering real-time bird species identification with multilingual support (English, Hindi, and Telugu) and additional features, such as species descriptions through Google Gemini and visual references from Wikimedia Commons. The open-source nature of the system, available on GitHub, promotes collaboration and further advancements. With its user-friendly design and practical deployment capability on resource-constrained devices, BirdRecon serves as a valuable tool for researchers, conservationists, and birdwatchers, contributing to biodiversity monitoring and conservation efforts.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"90 ","pages":"Article 103193"},"PeriodicalIF":5.8,"publicationDate":"2025-05-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144154956","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}

{"title":"Integrating species functional and architectural traits for improving crown width prediction in subtropical multispecies forests using nonlinear hierarchical models","authors":"Canming He ,&nbsp;Xianglin Tian ,&nbsp;Hongxiang Wang","doi":"10.1016/j.ecoinf.2025.103215","DOIUrl":"10.1016/j.ecoinf.2025.103215","url":null,"abstract":"<div><div>Crown width (CW) is a critical predictor of individual tree development and forest ecosystem function. Effective CW models are well established for plantations and structurally simpler forests, yet their applicability to diverse, species-rich natural forests remains inadequate. In this study, we analyzed CW data from 1802 individual trees of 23 species in a subtropical natural forest to assess the effects of tree-specific variables, interspecies variability, neighborhood effects, and topographical factors on CW predictions. Specifically, we aimed to elucidate how variations in CW models among individual species can be accounted for based on species architectural and functional traits. Among various candidate base models, the logistic model most effectively captured the relationship between CW and diameter at breast height (DBH). Model accuracy was improved by incorporating individual tree height, crown length, neighborhood competition index, and elevation as explanatory variables. A nonlinear mixed-effects model highlighted the significant role of tree species identity as a random effect in accounting for interspecific variability in CW predictions. We integrated species-level functional and architectural traits as covariates in the hierarchical model. The results revealed that 62 % of the interspecific variation originally captured by the random effects of species was explained by leaf thickness and mean crown diameter to DBH ratio. Our results highlight the importance of accounting for species variability in CW modeling and suggest that including species traits as covariates significantly improves model accuracy and generalizability, particularly for previously neglected architectural traits. The results obtained in this study expand our understanding of crown growth patterns and offer an improved basis for modeling tree crown sizes in species-rich forests.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"90 ","pages":"Article 103215"},"PeriodicalIF":5.8,"publicationDate":"2025-05-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144166881","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}

{"title":"Open real-time, non-invasive fish detection and size estimation utilizing binocular camera system in a Portuguese river affected by hydropeaking","authors":"Jürgen Soom ,&nbsp;Isabel Boavida ,&nbsp;Renan Leite ,&nbsp;Maria João Costa ,&nbsp;Gert Toming ,&nbsp;Mairo Leier ,&nbsp;Jeffrey A. Tuhtan","doi":"10.1016/j.ecoinf.2025.103196","DOIUrl":"10.1016/j.ecoinf.2025.103196","url":null,"abstract":"<div><div>The need for efficient approaches to track and assess fish behavior in rivers impacted by hydropeaking is increasing. Nonetheless, employing an automated camera system for underwater monitoring requires that the algorithms function under highly variable environmental conditions, which affect the ability to detect and assess fish size. Additionally, there is a lack of openly accessible freshwater fish classification and size estimation datasets. To address these limitations, we propose a binocular underwater fish monitoring system capable of real-time fish detection and size estimation. The system was deployed and tested over one week in two Portuguese rivers affected by hydropeaking. The week-long analysis also provided new insights regarding wild fish behavior in rivers affected by hydropeaking. Results indicate that hydropeaking strongly influences how fish may use instream flow refuges during hydropeaking. Fish were less frequently detected in the flow refuge during peak flow events, suggesting that the flow conditions created habitat instability and difficulty accessing the flow refuge. In contrast, fish in the non-hydropeaking river consistently used refuge areas, reinforcing their importance as shelter during natural flow variations. This study demonstrates the potential of a computer vision-based pipeline for real-time, fully automated fish monitoring of hydropeaking’s impacts on riverine fish. Additionally, we provide PTFish, an open dataset with 18,523 manually annotated frames featuring infrared and color video frames. These findings emphasize that automated, camera-based solutions for hydropeaking monitoring can be used to develop evidence-based mitigation strategies to sustain fish populations in rivers impacted by hydropeaking.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"90 ","pages":"Article 103196"},"PeriodicalIF":5.8,"publicationDate":"2025-05-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144213035","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}

{"title":"Integrating multiple indices of habitat quality to inform habitat management for a sagebrush indicator species","authors":"Megan C. Milligan ,&nbsp;Peter S. Coates ,&nbsp;Shawn T. O'Neil ,&nbsp;Brianne E. Brussee ,&nbsp;Michael P. Chenaille ,&nbsp;Derek A. Friend ,&nbsp;Kathleen Steele ,&nbsp;Justin R. Small ,&nbsp;Timothy S. Bowden ,&nbsp;Arlene D. Kosic ,&nbsp;Katherine Miller ,&nbsp;Michael L. Casazza","doi":"10.1016/j.ecoinf.2025.103228","DOIUrl":"10.1016/j.ecoinf.2025.103228","url":null,"abstract":"<div><div>Robust science is needed to inform natural resource management and policy decisions. Predictive species habitat maps are frequently employed in conservation decision-making but are often based on a single metric representing habitat quality. We outlined a framework that combines multiple spatially explicit indices of potential habitat quality that could be used to identify and prioritize habitat management areas, using the greater sage-grouse (<em>Centrocercus urophasianus</em>; hereafter sage-grouse) as an example species. Due to large-scale changes in sagebrush ecosystems, sage-grouse have suffered significant population declines in recent decades and have become key to land management plans throughout the western United States, where comprehensive habitat maps are crucial to effective conservation efforts. We evaluated habitat selection and survival patterns of sage-grouse across six distinct annual life stages and seasons to generate predictive habitat map surfaces, mapped the distribution of current occupancy, and combined maps of potential selection and survival patterns with space use and occupancy indices to delineate example habitat management categories. Our approach facilitates identification of priority areas to target for habitat preservation and areas where anthropogenic impacts could occur with likely minimal impact to the species. Overall, by combining indices representing selection, survival, and current occupancy, we provide a framework to allow for a flexible and targeted management approach that could be adapted to a wide variety of species.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"90 ","pages":"Article 103228"},"PeriodicalIF":5.8,"publicationDate":"2025-05-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144221557","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}

{"title":"Influence of geometrical shape on thermal heterogeneity in closed landfill sites","authors":"Arash Gitifar,&nbsp;Farzin Naghibalsadati,&nbsp;Nima Karimi,&nbsp;Anika Tahsin Abha,&nbsp;Rumpa Chowdhury,&nbsp;Kelvin Tsun Wai Ng","doi":"10.1016/j.ecoinf.2025.103219","DOIUrl":"10.1016/j.ecoinf.2025.103219","url":null,"abstract":"<div><div>Thermal heterogeneity assessment in landfill sites is essential for identification of potential hazards. The relationship between landfill geometrical shape and land surface thermal heterogeneity is not well understood. This study examines the association between landfills' shape configuration and thermal heterogeneity by using two mathematical shape factors on thirty-eight closed landfills. Three different multiple linear regression models were developed for landfill sites of various sizes. Geometrical shape analysis of the sites shows that all landfills surpass the 0.5 threshold, suggesting a tendency toward regular shapes and a systematic approach in their design and operation, with a mean elongation and compactness shape factor of 0.819, and 0.724, respectively. This pattern likely accommodates land use constraints and proximity to neighboring properties, with boundaries confined by the surrounding road network. In larger landfill sites, the elongation shape factor exhibits a higher coefficient (−0.46) than the compactness shape factor (−0.35), indicating its stronger association on thermal heterogeneity of the site. This finding helps to develop strategies for better thermal management and environmental safety of large landfill sites. The negative coefficients for all the site groups (small, medium, and large) suggest that a more compact and regular shape may promote thermal homogeneity in closed landfills. The proposed method improves monitoring of closed landfills and contributes to the development of evidence-based landfill design guidelines and regulations.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"90 ","pages":"Article 103219"},"PeriodicalIF":5.8,"publicationDate":"2025-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144196411","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}

{"title":"Patterns and drivers of plant C:N:P stoichiometry across a 3000 km aridity gradient","authors":"Jianzhao Wu ,&nbsp;Lingbo Dong ,&nbsp;Jiwei Li ,&nbsp;Yang Liao ,&nbsp;Zhouping Shangguan ,&nbsp;Bing Wang ,&nbsp;Lei Deng","doi":"10.1016/j.ecoinf.2025.103221","DOIUrl":"10.1016/j.ecoinf.2025.103221","url":null,"abstract":"<div><div>Leaf element stoichiometry is crucial for understanding nutrient dynamics and carbon (C) cycling in terrestrial ecosystems. However, the biogeographical patterns of leaf C, nitrogen (N), and phosphorus (P) content and their stoichiometric relationships along aridity gradients remain poorly understood, particularly regarding their driving factors. This study examined leaf C:N stoichiometry across a 3000 km aridity gradient in China, encompassing 36 sampling sites representing forest, grassland, and desert ecosystems. We further investigated the relationships between leaf biochemical traits and environmental drivers. Results revealed that the mean leaf contents of C, N, and P at 588.29 ± 6.9, 19.11 ± 0.3, and 1.33 ± 0.03 g kg⁻¹, respectively. The C:N, C:P, and N:P ratios were obtained as 32.43 ± 0.64, 480.65 ± 11.36, and 15.71 ± 0.4, respectively. The leaf C:N:P stoichiometry exhibited a pervasive nonlinear pattern, and a threshold of approximately 0.7 on an aridity index (AI). Below this threshold (AI &lt;0.7), the leaf C:P and N:P ratios decreased as AI increased, and N limitation became more evident. Conversely, these ratios increased above this threshold (AI &gt;0.7), indicating that P availability increasingly constrained plant growth. Furthermore, plants in arid regions (AI &lt;0.7) demonstrated strong stoichiometric homeostasis, suggesting effective physiological adaptation to environmental fluctuations. This homeostatic capacity substantially weakened in humid regions (AI &gt;0.7), where plants showed greater stoichiometric plasticity. These findings advance our understanding of spatial patterns in leaf nutrient stoichiometry and provide critical insights for modeling ecosystem nutrient cycling under global climate change scenarios.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"90 ","pages":"Article 103221"},"PeriodicalIF":5.8,"publicationDate":"2025-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144154957","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}

{"title":"Automated note annotation after bioacoustic classification: Unsupervised clustering of extracted acoustic features improves detection of a cryptic owl","authors":"Callan Alexander ,&nbsp;Robert Clemens ,&nbsp;Paul Roe ,&nbsp;Susan Fuller","doi":"10.1016/j.ecoinf.2025.103222","DOIUrl":"10.1016/j.ecoinf.2025.103222","url":null,"abstract":"<div><div>Passive acoustic monitoring and machine learning are increasingly being used to survey threatened species. When automated detection models are applied to large novel datasets, false-positive detections are likely even for high-performing models, and arbitrary thresholds may result in missed detections. Manual validation of outputs is time consuming, and additional fine-scale annotation of individual notes is impractical for large datasets and difficult to automate when using passive field recordings. This research presents an acoustic monitoring pipeline which employs a multi-stage hybrid approach: initial detection using a convolutional neural network classifier, followed by segmentation and iterative unsupervised clustering of extracted acoustic features using UMAP and HDBSCAN to remove label noise. We applied the pipeline to a large acoustic dataset comprised of 2764 h of environmental recordings and test the utility of the approach on territorial calls of Australia's largest owl: the threatened Powerful Owl (<em>Ninox strenua</em>). The pipeline reduced the large acoustic dataset into 10,116 annotations, of which 9399 (93 %) were correctly annotated individual notes of the target species. The clustering process also eliminated 88 % of false positive detections while retaining 95 % true positives (F1 = 0.94). The approach is highly scalable, can be applied to very large acoustic datasets, and can rapidly collect note-level annotations from noisy field recordings. The acoustic features derived from this methodology identified population differences in our test dataset and enable further exploration of song structure, geographic variation, and vocal individuality. The clustering process also facilitates a semi-supervised learning approach, allowing rapid selection of uncertain examples for model improvement. The pipeline helps to address two key challenges in bioacoustic monitoring: the need for manual validation of automated detections and the difficulty of obtaining accurate note-level annotations in noisy field recordings. Adaptation of these methods to other species and vocalisations may facilitate improved detection and investigation of vocal characteristics across different populations or regions.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"90 ","pages":"Article 103222"},"PeriodicalIF":5.8,"publicationDate":"2025-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144185326","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}

{"title":"Adaptation of deep learning models to distinguish species with strongly overlapping characteristics: the example of spores of the genus Riccia","authors":"Stefan Herdy ,&nbsp;Martina Pöltl ,&nbsp;Christian Berg ,&nbsp;Bettina Weber","doi":"10.1016/j.ecoinf.2025.103226","DOIUrl":"10.1016/j.ecoinf.2025.103226","url":null,"abstract":"<div><div>The liverwort genus <em>Riccia</em> is fairly characteristic in itself, but differentiation between some of the species herein proved as complex. Morphological thallus characteristics show a wide overlap and variation, also influenced by environmental conditions, whereas spore characteristics might be more consistent within individual species. Thus, here we investigated if morphological spore characteristics can be analyzed by means of generative as well as discriminative deep learning models, allowing a differentiation between very similar species. We applied a modified Generative Adversarial Network on spore images of the genus <em>Riccia</em> to generate images, that have a better expression of class specific morphological features. We also created one single spore image for every species that comprises the representative species characteristics. This approach allowed us to also distinguish morphologically very similar taxa and to quantify the morphological similarity between them. Our results show that generative modeling can improve morphological species classification in biology and provides a framework to quantify qualitative species features and similarities. These new tools facilitate more accurate and efficient species identification, thus greatly advancing methodologies in taxonomic research and environmental monitoring.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"90 ","pages":"Article 103226"},"PeriodicalIF":5.8,"publicationDate":"2025-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144166879","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}

{"title":"Phytoplankton group classification by integrating trait information and observed environmental thresholds","authors":"Hoang Vuong Dang ,&nbsp;Kermode Stephanie ,&nbsp;Peisheng Huang ,&nbsp;Cayelan C. Carey ,&nbsp;Matthew R. Hipsey","doi":"10.1016/j.ecoinf.2025.103212","DOIUrl":"10.1016/j.ecoinf.2025.103212","url":null,"abstract":"<div><div>Assigning phytoplankton taxa into functional groups is a common requirement for process-based models of aquatic ecology, yet it can be challenging in systems with large taxonomic diversity and remains a largely subjective task. In the absence of a clear and transferrable framework, modellers often default to the delineation of phytoplankton groups at the phyla or class level (e.g., diatoms, greens, etc.). However, this approach aggregates the substantial functional and trait diversity that occurs within these groups, creating challenges for model parameterization and assessment. To address this issue, we developed a data-driven approach to define phytoplankton functional groups considering species trait information and occurrence data. The framework calculates the observed environmental thresholds for species monitored in a 12-year dataset from the Hawkesbury-Nepean River (Sydney, Australia), combined with a priori species-level trait information (e.g., organism structure, biovolume, movement types, and nutrient acquisition strategies). We minimized subjectivity in phytoplankton group classification by first applying multiple correlation analysis and principal component analysis to identify the most important environmental factors for threshold analysis. Second, we used Threshold Indicator Taxa Analysis (TITAN) to detect the ecological threshold ranges summarizing species occurrence along environmental gradients of total phosphorus, total nitrogen, the ratio of total nitrogen to total phosphorus, and temperature. Third, we applied K-prototype clustering for group classification based on the identified thresholds and associated traits. Our approach identified five discrete phytoplankton groups with statistically distinct features of environmental preference and morphological and physiological characteristics. The advantage of the method is that the identified groups better reflect the ecological characteristics of the phytoplankton community considering the local environmental requirements, which better aligns with the process parameterizations used in numerical phytoplankton models. This framework can be applied in other aquatic systems as a robust and repeatable way to integrate long-term phytoplankton taxonomic and environmental datasets for water quality analyses.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"90 ","pages":"Article 103212"},"PeriodicalIF":5.8,"publicationDate":"2025-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144185327","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}

{"title":"Selecting of global phenological field observations for validating coarse AVHRR-derived forest phenology products based on spatial heterogeneity and temporal consistency","authors":"Qi Shao ,&nbsp;Chao Huang ,&nbsp;Yuanjun Xiao ,&nbsp;Li Liu ,&nbsp;Weiwei Liu ,&nbsp;Ran Huang ,&nbsp;Chang Zhou ,&nbsp;Wei Weng ,&nbsp;Jingfeng Huang","doi":"10.1016/j.ecoinf.2025.103216","DOIUrl":"10.1016/j.ecoinf.2025.103216","url":null,"abstract":"<div><div>Global phenological field observations play a crucial role in validating remote sensing products and algorithms. However, due to the spatial mismatch and scale effect between the field observations and the pixels of remote sensing phenology products, a direct comparison often leads to scale errors and increased uncertainty. Therefore, evaluating the spatial representativeness of field observations for remote sensing product validation is essential. This study developed a novel “bottom-up” evaluation framework named MSPT (<strong>M</strong>ain land cover type, <strong>S</strong>patial heterogeneity, <strong>P</strong>oint-area consistency and <strong>T</strong>emporal consistency), which comprehensively assesses the spatial representativeness of forest phenological field observations within the coarse spatial scale. Based on MSPT method, the capability of global forest phenological field observations to support coarse-scale remote sensing validation was evaluated. Compared with the general method, MSPT significantly improved validation performance. For the start of the growing season (SOS), the root mean square error (RMSE) decreased from 49.70 to 33.75 days, and the percent bias (PBIAS) changed from −0.14 to 0.03. For the end of the growing season (EOS), the RMSE was reduced from 83.42 to 42.53 days, and the PBIAS decreased from 0.15 to 0.08. These findings demonstrate that MSPT enhances the reliability of validation datasets and effectively reducing uncertainty in the evaluation of coarse AVHRR-derived forest phenology products. The framework offers new insights into resolving the scale mismatch between field observations and the pixels of remote sensing products.</div></div>","PeriodicalId":51024,"journal":{"name":"Ecological Informatics","volume":"90 ","pages":"Article 103216"},"PeriodicalIF":5.8,"publicationDate":"2025-05-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144138398","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}