Wahaj Habib, Rémi Cresson, Kevin McGuinness, John Connolly
Peatlands, constituting over half of terrestrial wetland ecosystems across the globe, hold critical ecological significance and are large stores of carbon (C). Irish oceanic raised bogs are a rare peatland ecosystem offering numerous ecosystem services, including C storage, biodiversity support and water regulation. However, they have been degraded over the centuries due to artificial drainage, followed by peat extraction, afforestation and agriculture. This has an overall negative impact on the functioning of peatlands, shifting them from a moderate C sink to a large C source. Recognizing the importance of these ecosystems, efforts are underway for conservation (rewetting and rehabilitation), while accurately accounting for C stock and greenhouse gas (GHG) emissions. However, the implementation of these efforts requires accurate identification and mapping of artificial drainage ditches. This study utilized very high‐resolution (25 cm) aerial imagery, and a deep learning (U‐Net) approach to map the visible artificial drainage (unobstructed by vegetation or infill) in raised bogs at a national scale. The results show that artificial drainage is widespread, with ~20 000 km of drains mapped. The overall accuracy of the model was 80% on an independent testing dataset. The data were also used to derive the Fracditch which was 0.03 (fraction of artificial drainage on industrial peat extraction sites). This is lower than IPCC Tier 1 Fracditch and can aid in IPCC Tier 2 reporting for Ireland. This is the first study to map drains with diverse sizes and patterns on Irish‐raised bogs using optical aerial imagery and deep learning methods. The map will serve as an important baseline dataset for evaluating the artificial drainage ditch conditions. It will prove useful for sustainable management, conservation and refined estimations of GHG emissions. The model's capacity for generalization implies its potential in mapping artificial drains in peatlands at a regional and global scale, thereby enhancing the comprehension of the global effects of artificial drainage ditches on peatlands.
{"title":"Mapping artificial drains in peatlands—A national‐scale assessment of Irish raised bogs using sub‐meter aerial imagery and deep learning methods","authors":"Wahaj Habib, Rémi Cresson, Kevin McGuinness, John Connolly","doi":"10.1002/rse2.387","DOIUrl":"https://doi.org/10.1002/rse2.387","url":null,"abstract":"Peatlands, constituting over half of terrestrial wetland ecosystems across the globe, hold critical ecological significance and are large stores of carbon (C). Irish oceanic raised bogs are a rare peatland ecosystem offering numerous ecosystem services, including C storage, biodiversity support and water regulation. However, they have been degraded over the centuries due to artificial drainage, followed by peat extraction, afforestation and agriculture. This has an overall negative impact on the functioning of peatlands, shifting them from a moderate C sink to a large C source. Recognizing the importance of these ecosystems, efforts are underway for conservation (rewetting and rehabilitation), while accurately accounting for C stock and greenhouse gas (GHG) emissions. However, the implementation of these efforts requires accurate identification and mapping of artificial drainage ditches. This study utilized very high‐resolution (25 cm) aerial imagery, and a deep learning (U‐Net) approach to map the visible artificial drainage (unobstructed by vegetation or infill) in raised bogs at a national scale. The results show that artificial drainage is widespread, with ~20 000 km of drains mapped. The overall accuracy of the model was 80% on an independent testing dataset. The data were also used to derive the Frac<jats:sub>ditch</jats:sub> which was 0.03 (fraction of artificial drainage on industrial peat extraction sites). This is lower than IPCC Tier 1 Frac<jats:sub>ditch</jats:sub> and can aid in IPCC Tier 2 reporting for Ireland. This is the first study to map drains with diverse sizes and patterns on Irish‐raised bogs using optical aerial imagery and deep learning methods. The map will serve as an important baseline dataset for evaluating the artificial drainage ditch conditions. It will prove useful for sustainable management, conservation and refined estimations of GHG emissions. The model's capacity for generalization implies its potential in mapping artificial drains in peatlands at a regional and global scale, thereby enhancing the comprehension of the global effects of artificial drainage ditches on peatlands.","PeriodicalId":21132,"journal":{"name":"Remote Sensing in Ecology and Conservation","volume":"217 1","pages":""},"PeriodicalIF":5.5,"publicationDate":"2024-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140640323","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}
Gustavo Perez, Wenlong Zhao, Zezhou Cheng, Maria Carolina T. D. Belotti, Yuting Deng, Victoria F. Simons, Elske Tielens, Jeffrey F. Kelly, Kyle G. Horton, Subhransu Maji, Daniel Sheldon
The exodus of flying animals from their roosting locations is often visible as expanding ring‐shaped patterns in weather radar data. The NEXRAD network, for example, archives more than 25 years of data across 143 contiguous US radar stations, providing opportunities to study roosting locations and times and the ecosystems of birds and bats. However, access to this information is limited by the cost of manually annotating millions of radar scans. We develop and deploy an AI‐assisted system to annotate roosts in radar data. We build datasets with roost annotations to support the training and evaluation of automated detection models. Roosts are detected, tracked, and incorporated into our developed web‐based interface for human screening to produce research‐grade annotations. We deploy the system to collect swallow and martin roost information from 12 radar stations around the Great Lakes spanning 21 years. After verifying the practical value of the system, we propose to improve the detector by incorporating both spatial and temporal channels from volumetric radar scans. The deployment on Great Lakes radar scans allows accelerated annotation of 15 628 roost signatures in 612 786 radar scans with 183.6 human screening hours, or 1.08 s per radar scan. We estimate that the deployed system reduces human annotation time by ~7×. The temporal detector model improves the average precision at intersection‐over‐union threshold 0.5 (APIoU = .50) by 8% over the previous model (48%→56%), further reducing human screening time by 2.3× in its pilot deployment. These data contain critical information about phenology and population trends of swallows and martins, aerial insectivore species experiencing acute declines, and have enabled novel research. We present error analyses, lay the groundwork for continent‐scale historical investigation about these species, and provide a starting point for automating the detection of other family‐specific phenomena in radar data, such as bat roosts and mayfly hatches.
{"title":"Using spatiotemporal information in weather radar data to detect and track communal roosts","authors":"Gustavo Perez, Wenlong Zhao, Zezhou Cheng, Maria Carolina T. D. Belotti, Yuting Deng, Victoria F. Simons, Elske Tielens, Jeffrey F. Kelly, Kyle G. Horton, Subhransu Maji, Daniel Sheldon","doi":"10.1002/rse2.388","DOIUrl":"https://doi.org/10.1002/rse2.388","url":null,"abstract":"The exodus of flying animals from their roosting locations is often visible as expanding ring‐shaped patterns in weather radar data. The NEXRAD network, for example, archives more than 25 years of data across 143 contiguous US radar stations, providing opportunities to study roosting locations and times and the ecosystems of birds and bats. However, access to this information is limited by the cost of manually annotating millions of radar scans. We develop and deploy an AI‐assisted system to annotate roosts in radar data. We build datasets with roost annotations to support the training and evaluation of automated detection models. Roosts are detected, tracked, and incorporated into our developed web‐based interface for human screening to produce research‐grade annotations. We deploy the system to collect swallow and martin roost information from 12 radar stations around the Great Lakes spanning 21 years. After verifying the practical value of the system, we propose to improve the detector by incorporating both spatial and temporal channels from volumetric radar scans. The deployment on Great Lakes radar scans allows accelerated annotation of 15 628 roost signatures in 612 786 radar scans with 183.6 human screening hours, or 1.08 s per radar scan. We estimate that the deployed system reduces human annotation time by ~7×. The temporal detector model improves the average precision at intersection‐over‐union threshold 0.5 (AP<jats:sup>IoU = .50</jats:sup>) by 8% over the previous model (48%→56%), further reducing human screening time by 2.3× in its pilot deployment. These data contain critical information about phenology and population trends of swallows and martins, aerial insectivore species experiencing acute declines, and have enabled novel research. We present error analyses, lay the groundwork for continent‐scale historical investigation about these species, and provide a starting point for automating the detection of other family‐specific phenomena in radar data, such as bat roosts and mayfly hatches.","PeriodicalId":21132,"journal":{"name":"Remote Sensing in Ecology and Conservation","volume":"49 1","pages":""},"PeriodicalIF":5.5,"publicationDate":"2024-04-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140607842","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}
Elena Schall, Idil Ilgaz Kaya, Elisabeth Debusschere, Paul Devos, Clea Parcerisas
Passive acoustic monitoring (PAM) is commonly used to obtain year‐round continuous data on marine soundscapes harboring valuable information on species distributions or ecosystem dynamics. This continuously increasing amount of data requires highly efficient automated analysis techniques in order to exploit the full potential of the available data. Here, we propose a benchmark, which consists of a public dataset, a well‐defined task and evaluation procedure to develop and test automated analysis techniques. This benchmark focuses on the special case of detecting animal vocalizations in a real‐world dataset from the marine realm. We believe that such a benchmark is necessary to monitor the progress in the development of new detection algorithms in the field of marine bioacoustics. We ultimately use the proposed benchmark to test three detection approaches, namely ANIMAL‐SPOT, Koogu and a simple custom sequential convolutional neural network (CNN), and report performances. We report the performance of the three detection approaches in a blocked cross‐validation fashion with 11 site‐year blocks for a multi‐species detection scenario in a large marine passive acoustic dataset. Performance was measured with three simple metrics (i.e., true classification rate, noise misclassification rate and call misclassification rate) and one combined fitness metric, which allocates more weight to the minimization of false positives created by noise. Overall, ANIMAL‐SPOT performed the best with an average fitness metric of 0.6, followed by the custom CNN with an average fitness metric of 0.57 and finally Koogu with an average fitness metric of 0.42. The presented benchmark is an important step to advance in the automatic processing of the continuously growing amount of PAM data that are collected throughout the world's oceans. To ultimately achieve usability of developed algorithms, the focus of future work should be laid on the reduction of the false positives created by noise.
{"title":"Deep learning in marine bioacoustics: a benchmark for baleen whale detection","authors":"Elena Schall, Idil Ilgaz Kaya, Elisabeth Debusschere, Paul Devos, Clea Parcerisas","doi":"10.1002/rse2.392","DOIUrl":"https://doi.org/10.1002/rse2.392","url":null,"abstract":"Passive acoustic monitoring (PAM) is commonly used to obtain year‐round continuous data on marine soundscapes harboring valuable information on species distributions or ecosystem dynamics. This continuously increasing amount of data requires highly efficient automated analysis techniques in order to exploit the full potential of the available data. Here, we propose a benchmark, which consists of a public dataset, a well‐defined task and evaluation procedure to develop and test automated analysis techniques. This benchmark focuses on the special case of detecting animal vocalizations in a real‐world dataset from the marine realm. We believe that such a benchmark is necessary to monitor the progress in the development of new detection algorithms in the field of marine bioacoustics. We ultimately use the proposed benchmark to test three detection approaches, namely ANIMAL‐SPOT, Koogu and a simple custom sequential convolutional neural network (CNN), and report performances. We report the performance of the three detection approaches in a blocked cross‐validation fashion with 11 site‐year blocks for a multi‐species detection scenario in a large marine passive acoustic dataset. Performance was measured with three simple metrics (i.e., true classification rate, noise misclassification rate and call misclassification rate) and one combined fitness metric, which allocates more weight to the minimization of false positives created by noise. Overall, ANIMAL‐SPOT performed the best with an average fitness metric of 0.6, followed by the custom CNN with an average fitness metric of 0.57 and finally Koogu with an average fitness metric of 0.42. The presented benchmark is an important step to advance in the automatic processing of the continuously growing amount of PAM data that are collected throughout the world's oceans. To ultimately achieve usability of developed algorithms, the focus of future work should be laid on the reduction of the false positives created by noise.","PeriodicalId":21132,"journal":{"name":"Remote Sensing in Ecology and Conservation","volume":"447 1","pages":""},"PeriodicalIF":5.5,"publicationDate":"2024-04-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140607453","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}
Shulin Ren, Xiyan Xu, Gensuo Jia, Anqi Huang, Wei Ma
Fire events in South America are becoming more extensive and frequent as climate extremes and human pressures increase, and even repeatedly occurring in some areas within decades. However, the relationship between recurring fires and vegetation dynamics remains unclear. Here, we extracted the number of fire occurrences using burned area satellite product and analysed the relationship between recurring fires and vegetation dynamics with remote sensing land use and vegetation index datasets in South America. We show that approximately 1.39 × 106 km2 of burnt area has experienced recurring fires during 2001–2020. More than half of burnt area of recurring fires occurred in savannahs with remaining burnt area in grasslands, forests and croplands. Although forests tended to be less susceptible to recurring fires among all vegetation types, their coverage loss with recurring fires was the greatest. The greater proportion of forest conversion to croplands concurred with more recurring fires. Conversely, the coverage of croplands and grasslands gained the most with recurring fires. In the areas without vegetation conversion, more frequent recurring fires further suppressed canopy greenness and density, even in fire‐adapted savannahs and grasslands. Our results suggest that recurring fires and land use change are generally coincident, reflecting the intense pressure of human activities on natural vegetation in South America. Thus, coordinated efforts on vegetation conservation and sustainable management of human‐induced burning in the region are urgently needed.
{"title":"Coherence of recurring fires and land use change in South America","authors":"Shulin Ren, Xiyan Xu, Gensuo Jia, Anqi Huang, Wei Ma","doi":"10.1002/rse2.390","DOIUrl":"https://doi.org/10.1002/rse2.390","url":null,"abstract":"Fire events in South America are becoming more extensive and frequent as climate extremes and human pressures increase, and even repeatedly occurring in some areas within decades. However, the relationship between recurring fires and vegetation dynamics remains unclear. Here, we extracted the number of fire occurrences using burned area satellite product and analysed the relationship between recurring fires and vegetation dynamics with remote sensing land use and vegetation index datasets in South America. We show that approximately 1.39 × 10<jats:sup>6</jats:sup> km<jats:sup>2</jats:sup> of burnt area has experienced recurring fires during 2001–2020. More than half of burnt area of recurring fires occurred in savannahs with remaining burnt area in grasslands, forests and croplands. Although forests tended to be less susceptible to recurring fires among all vegetation types, their coverage loss with recurring fires was the greatest. The greater proportion of forest conversion to croplands concurred with more recurring fires. Conversely, the coverage of croplands and grasslands gained the most with recurring fires. In the areas without vegetation conversion, more frequent recurring fires further suppressed canopy greenness and density, even in fire‐adapted savannahs and grasslands. Our results suggest that recurring fires and land use change are generally coincident, reflecting the intense pressure of human activities on natural vegetation in South America. Thus, coordinated efforts on vegetation conservation and sustainable management of human‐induced burning in the region are urgently needed.","PeriodicalId":21132,"journal":{"name":"Remote Sensing in Ecology and Conservation","volume":"19 1","pages":""},"PeriodicalIF":5.5,"publicationDate":"2024-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140547553","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}
Patrick Kacic, Ursula Gessner, Stefanie Holzwarth, Frank Thonfeld, Claudia Kuenzer
Assessing the dynamics of forest structure complexity is a critical task in times of global warming, biodiversity loss and increasing disturbances in order to ensure the resilience of forests. Recent studies on forest biodiversity and forest structure emphasize the essential functions of deadwood accumulation and diversification of light conditions for the enhancement of structural complexity. The implementation of an experimental patch‐network in managed broad‐leaved forests within Germany enables the standardized analysis of various aggregated and distributed treatments characterized through diverse deadwood and light structures. To monitor the dynamics of enhanced forest structure complexity as seasonal and trend components, dense time‐series from high spatial resolution imagery of Sentinel‐1 (Synthetic‐Aperture Radar, SAR) and Sentinel‐2 (multispectral) are analyzed in time‐series decomposition models (BEAST, Bayesian Estimator of Abrupt change, Seasonal change and Trend). Based on several spatial statistics and a comprehensive catalog on spectral indices, metrics from Sentinel‐1 (n = 84) and Sentinel‐2 (n = 903) are calculated at patch‐level. Metrics best identifying the treatment implementation event are assessed by change point dates and probability scores. Heterogeneity metrics of Sentinel‐1 VH and Sentinel‐2 NMDI (Normalized Multi‐band Drought Index) capture the treatment implementation event most accurately, with clear advantages for the identification of aggregated treatments. In addition, aggregated structures of downed or no deadwood can be characterized, as well as more complex standing structures, such as snags or habitat trees. To conclude, dense time‐series of complementary high spatial resolution sensors have the potential to assess various aggregated forest structure complexities, thus supporting the continuous monitoring of forest habitats and functioning over time.
{"title":"Assessing experimental silvicultural treatments enhancing structural complexity in a central European forest – BEAST time‐series analysis based on Sentinel‐1 and Sentinel‐2","authors":"Patrick Kacic, Ursula Gessner, Stefanie Holzwarth, Frank Thonfeld, Claudia Kuenzer","doi":"10.1002/rse2.386","DOIUrl":"https://doi.org/10.1002/rse2.386","url":null,"abstract":"Assessing the dynamics of forest structure complexity is a critical task in times of global warming, biodiversity loss and increasing disturbances in order to ensure the resilience of forests. Recent studies on forest biodiversity and forest structure emphasize the essential functions of deadwood accumulation and diversification of light conditions for the enhancement of structural complexity. The implementation of an experimental patch‐network in managed broad‐leaved forests within Germany enables the standardized analysis of various aggregated and distributed treatments characterized through diverse deadwood and light structures. To monitor the dynamics of enhanced forest structure complexity as seasonal and trend components, dense time‐series from high spatial resolution imagery of Sentinel‐1 (Synthetic‐Aperture Radar, SAR) and Sentinel‐2 (multispectral) are analyzed in time‐series decomposition models (BEAST, Bayesian Estimator of Abrupt change, Seasonal change and Trend). Based on several spatial statistics and a comprehensive catalog on spectral indices, metrics from Sentinel‐1 (<jats:italic>n</jats:italic> = 84) and Sentinel‐2 (<jats:italic>n</jats:italic> = 903) are calculated at patch‐level. Metrics best identifying the treatment implementation event are assessed by change point dates and probability scores. Heterogeneity metrics of Sentinel‐1 VH and Sentinel‐2 NMDI (Normalized Multi‐band Drought Index) capture the treatment implementation event most accurately, with clear advantages for the identification of aggregated treatments. In addition, aggregated structures of downed or no deadwood can be characterized, as well as more complex standing structures, such as snags or habitat trees. To conclude, dense time‐series of complementary high spatial resolution sensors have the potential to assess various aggregated forest structure complexities, thus supporting the continuous monitoring of forest habitats and functioning over time.","PeriodicalId":21132,"journal":{"name":"Remote Sensing in Ecology and Conservation","volume":"53 1","pages":""},"PeriodicalIF":5.5,"publicationDate":"2024-04-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140346003","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}
Nicholas Pontone, Koreen Millard, Dan K. Thompson, Luc Guindon, André Beaudoin
Peatlands in the Canadian boreal forest are being negatively impacted by anthropogenic climate change, the effects of which are expected to worsen. Peatland types and sub-classes vary in their ecohydrological characteristics and are expected to have different responses to climate change. Large-scale modelling frameworks such as the Canadian Model for Peatlands, the Canadian Fire Behaviour Prediction System and the Canadian Land Data Assimilation System require peatland maps including information on sub-types and vegetation as critical inputs. Additionally, peatland class and vegetation height are critical variables for wildlife habitat management and are related to the carbon cycle and wildfire fuel loading. This research aimed to create a map of peatland sub-classes (bog, poor fen, rich fen permafrost peat complex) for the Canadian boreal forest and create an inventory of peatland vegetation height characteristics using ICESat-2. A three-stage hierarchical classification framework was developed to map peatland sub-classes within the Canadian boreal forest circa 2020. Training and validation data consisted of peatland locations derived from various sources (field data, aerial photo interpretation, measurements documented in literature). A combination of multispectral data, L-band SAR backscatter and C-Band interferometric SAR coherence, forest structure and ancillary variables was used as model predictors. Ancillary data were used to mask agricultural areas and urban regions and account for regions that may exhibit permafrost. In the first stage of the classification, wetlands, uplands and water were classified with 86.5% accuracy. In the second stage, within the wetland areas only, peatland and mineral wetlands were differentiated with 93.3% accuracy. In the third stage, constrained to only the peatland areas, bogs, rich fens, poor fens and permafrost peat complexes were classified with 71.5% accuracy. Then, ICESat-2 ATL08 spaceborne lidar data were used to describe regional variations in peatland vegetation height characteristics and regional and class-wise variations based on a boreal forest wide sample. This research introduced a comprehensive large-scale peatland sub-class mapping framework for the Canadian boreal forest, presenting the first moderate resolution map of its kind.
{"title":"A hierarchical, multi-sensor framework for peatland sub-class and vegetation mapping throughout the Canadian boreal forest","authors":"Nicholas Pontone, Koreen Millard, Dan K. Thompson, Luc Guindon, André Beaudoin","doi":"10.1002/rse2.384","DOIUrl":"https://doi.org/10.1002/rse2.384","url":null,"abstract":"Peatlands in the Canadian boreal forest are being negatively impacted by anthropogenic climate change, the effects of which are expected to worsen. Peatland types and sub-classes vary in their ecohydrological characteristics and are expected to have different responses to climate change. Large-scale modelling frameworks such as the Canadian Model for Peatlands, the Canadian Fire Behaviour Prediction System and the Canadian Land Data Assimilation System require peatland maps including information on sub-types and vegetation as critical inputs. Additionally, peatland class and vegetation height are critical variables for wildlife habitat management and are related to the carbon cycle and wildfire fuel loading. This research aimed to create a map of peatland sub-classes (bog, poor fen, rich fen permafrost peat complex) for the Canadian boreal forest and create an inventory of peatland vegetation height characteristics using ICESat-2. A three-stage hierarchical classification framework was developed to map peatland sub-classes within the Canadian boreal forest circa 2020. Training and validation data consisted of peatland locations derived from various sources (field data, aerial photo interpretation, measurements documented in literature). A combination of multispectral data, L-band SAR backscatter and C-Band interferometric SAR coherence, forest structure and ancillary variables was used as model predictors. Ancillary data were used to mask agricultural areas and urban regions and account for regions that may exhibit permafrost. In the first stage of the classification, wetlands, uplands and water were classified with 86.5% accuracy. In the second stage, within the wetland areas only, peatland and mineral wetlands were differentiated with 93.3% accuracy. In the third stage, constrained to only the peatland areas, bogs, rich fens, poor fens and permafrost peat complexes were classified with 71.5% accuracy. Then, ICESat-2 ATL08 spaceborne lidar data were used to describe regional variations in peatland vegetation height characteristics and regional and class-wise variations based on a boreal forest wide sample. This research introduced a comprehensive large-scale peatland sub-class mapping framework for the Canadian boreal forest, presenting the first moderate resolution map of its kind.","PeriodicalId":21132,"journal":{"name":"Remote Sensing in Ecology and Conservation","volume":"8 1","pages":""},"PeriodicalIF":5.5,"publicationDate":"2024-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139957152","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}
David Singer, Jonas Hagge, Johannes Kamp, Hermann Hondong, Andreas Schuldt
Passive acoustic monitoring (PAM) has gained increasing popularity to study behaviour, habitat preferences, distribution and community assembly of birds and other animals. Automated species classification algorithms like ‘BirdNET’ are capable of detecting and classifying avian vocalizations within extensive audio data, covering entire species assemblages. PAM reveals substantial potential for biodiversity monitoring that informs evidence-based conservation. Nevertheless, fully realizing this potential remains challenging, especially due to the issue of false-positive species detections. Here, we introduce an optimized thresholding framework, which incorporates contextual information extracted from the time-series of automated species detections (i.e. covariates on quality and quantity of species' detections measured at varying time intervals) to improve the differentiation of true and false positives. We verified a sample of BirdNET detections per species and modelled species-specific thresholds using conditional inference trees. These thresholds were designed to minimize false-positive detections while maximizing the preservation of true positives in the dataset. We tested this framework for a large dataset of BirdNET detections (5760 h of audio data, 60 sites) recorded over an entire breeding season. Our results revealed considerable interspecific variability of precision (percentage of true positives) within raw BirdNET data. Our optimized thresholding approach achieved high precision (≥0.9) for 70% of the 61 detected species, while species-specific thresholds solely relying on the BirdNET confidence scores achieved high precision for only 31% of the species. Conservative universal thresholds (not species-specific) reached high precision for 48% of the species. Our thresholding approach outperformed previous thresholding approaches and enhanced interspecific comparability for bird community analyses. By incorporating contextual information from the time-series of species detections, the differentiation of true and false positives was substantially improved. Our approach may enhance a straightforward application of PAM in biodiversity research, landscape planning and evidence-based conservation.
{"title":"Aggregated time-series features boost species-specific differentiation of true and false positives in passive acoustic monitoring of bird assemblages","authors":"David Singer, Jonas Hagge, Johannes Kamp, Hermann Hondong, Andreas Schuldt","doi":"10.1002/rse2.385","DOIUrl":"https://doi.org/10.1002/rse2.385","url":null,"abstract":"Passive acoustic monitoring (PAM) has gained increasing popularity to study behaviour, habitat preferences, distribution and community assembly of birds and other animals. Automated species classification algorithms like ‘BirdNET’ are capable of detecting and classifying avian vocalizations within extensive audio data, covering entire species assemblages. PAM reveals substantial potential for biodiversity monitoring that informs evidence-based conservation. Nevertheless, fully realizing this potential remains challenging, especially due to the issue of false-positive species detections. Here, we introduce an optimized thresholding framework, which incorporates contextual information extracted from the time-series of automated species detections (i.e. covariates on quality and quantity of species' detections measured at varying time intervals) to improve the differentiation of true and false positives. We verified a sample of BirdNET detections per species and modelled species-specific thresholds using conditional inference trees. These thresholds were designed to minimize false-positive detections while maximizing the preservation of true positives in the dataset. We tested this framework for a large dataset of BirdNET detections (5760 h of audio data, 60 sites) recorded over an entire breeding season. Our results revealed considerable interspecific variability of precision (percentage of true positives) within raw BirdNET data. Our optimized thresholding approach achieved high precision (≥0.9) for 70% of the 61 detected species, while species-specific thresholds solely relying on the BirdNET confidence scores achieved high precision for only 31% of the species. Conservative universal thresholds (not species-specific) reached high precision for 48% of the species. Our thresholding approach outperformed previous thresholding approaches and enhanced interspecific comparability for bird community analyses. By incorporating contextual information from the time-series of species detections, the differentiation of true and false positives was substantially improved. Our approach may enhance a straightforward application of PAM in biodiversity research, landscape planning and evidence-based conservation.","PeriodicalId":21132,"journal":{"name":"Remote Sensing in Ecology and Conservation","volume":"3 1","pages":""},"PeriodicalIF":5.5,"publicationDate":"2024-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139957154","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}
Xiang Liu, Julian Frey, Catalina Munteanu, Martin Denter, Barbara Koch
Increasingly available spaceborne sensors provide unprecedented opportunities for large-scale, timely and continuous tree species diversity (TSD) monitoring. However, given differences in spectral and spatial resolutions, the choice of sensor is not always straightforward. In this work, we investigated the effects of spatial and spectral resolutions for four spaceborne sensors (RapidEye, Landsat-8, Sentinel-2 and PlanetScope) on TSD mapping in an area of approximately 4000 km2 within the Black Forest, Germany. We employed a random forest (RF) regression model to predict Shannon–Wiener diversity based on seven types of spectral heterogeneity metrics (texture, coefficient of variation, Rao's Q, convex hull volume, spectral angle mapper, convex hull area and spectral species diversity) and a full survey dataset from 135 one-ha sample plots. We compared the RF model's performance across sensors and spatial resolutions. Our results demonstrated that the Sentinel-2-based TSD model achieved the highest accuracy (mean R2: 0.477, mean root-mean-square error (RMSE): 0.274). The RapidEye-based TSD model produced lower accuracy (mean R2: 0.346, mean RMSE: 0.303), but it was better than the PlanetScope- and Landsat-based TSD models. The 10 m (for Sentinel-2 and RapidEye) and 15 m (for PlanetScope) were the best spatial resolutions for predicting TSD. The NIR band was the most favourable spectral band for predicting TSD. Texture metrics and Rao's Q outperformed the other spectral heterogeneity metrics. Our results highlighted that spaceborne optical imagery (especially Sentinel-2) can be successfully used for large-scale TSD mapping but that the choice of sensors can significantly affect the resulting mapping accuracy in temperate montane forests.
{"title":"Tree species diversity mapping from spaceborne optical images: The effects of spectral and spatial resolution","authors":"Xiang Liu, Julian Frey, Catalina Munteanu, Martin Denter, Barbara Koch","doi":"10.1002/rse2.383","DOIUrl":"https://doi.org/10.1002/rse2.383","url":null,"abstract":"Increasingly available spaceborne sensors provide unprecedented opportunities for large-scale, timely and continuous tree species diversity (TSD) monitoring. However, given differences in spectral and spatial resolutions, the choice of sensor is not always straightforward. In this work, we investigated the effects of spatial and spectral resolutions for four spaceborne sensors (RapidEye, Landsat-8, Sentinel-2 and PlanetScope) on TSD mapping in an area of approximately 4000 km<sup>2</sup> within the Black Forest, Germany. We employed a random forest (RF) regression model to predict Shannon–Wiener diversity based on seven types of spectral heterogeneity metrics (texture, coefficient of variation, Rao's Q, convex hull volume, spectral angle mapper, convex hull area and spectral species diversity) and a full survey dataset from 135 one-ha sample plots. We compared the RF model's performance across sensors and spatial resolutions. Our results demonstrated that the Sentinel-2-based TSD model achieved the highest accuracy (mean <i>R</i><sup>2</sup>: 0.477, mean root-mean-square error (RMSE): 0.274). The RapidEye-based TSD model produced lower accuracy (mean <i>R</i><sup>2</sup>: 0.346, mean RMSE: 0.303), but it was better than the PlanetScope- and Landsat-based TSD models. The 10 m (for Sentinel-2 and RapidEye) and 15 m (for PlanetScope) were the best spatial resolutions for predicting TSD. The NIR band was the most favourable spectral band for predicting TSD. Texture metrics and Rao's Q outperformed the other spectral heterogeneity metrics. Our results highlighted that spaceborne optical imagery (especially Sentinel-2) can be successfully used for large-scale TSD mapping but that the choice of sensors can significantly affect the resulting mapping accuracy in temperate montane forests.","PeriodicalId":21132,"journal":{"name":"Remote Sensing in Ecology and Conservation","volume":"29 1","pages":""},"PeriodicalIF":5.5,"publicationDate":"2024-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139911333","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}
Mountain meadows are an essential part of the alpine–subalpine ecosystem; they provide ecosystem services like pollination and are home to diverse plant communities. Changes in climate affect meadow ecology on multiple levels, for example, by altering growing season dynamics. Tracking the effects of climate change on meadow diversity through the impacts on individual species and overall growing season dynamics is critical to conservation efforts. Here, we explore how to combine crowd-sourced camera images with machine learning to quantify flowering species richness across a range of elevations in alpine meadows located in Mt. Rainier National Park, Washington, USA. We employed three machine-learning techniques (Mask R-CNN, RetinaNet and YOLOv5) to detect wildflower species in images taken during two flowering seasons. We demonstrate that deep learning techniques can detect multiple species, providing information on flowering richness in photographed meadows. The results indicate higher richness just above the tree line for most of the species, which is comparable with patterns found using field studies. We found that the two-stage detector Mask R-CNN was more accurate than single-stage detectors like RetinaNet and YOLO, with the Mask R-CNN network performing best overall with mean average precision (mAP) of 0.67 followed by RetinaNet (0.5) and YOLO (0.4). We found that across the methods using anchor box variations in multiples of 16 led to enhanced accuracy. We also show that detection is possible even when pictures are interspersed with complex backgrounds and are not in focus. We found differential detection rates depending on species abundance, with additional challenges related to similarity in flower characteristics, labeling errors and occlusion issues. Despite these potential biases and limitations in capturing flowering abundance and location-specific quantification, accuracy was notable considering the complexity of flower types and picture angles in this dataset. We, therefore, expect that this approach can be used to address many ecological questions that benefit from automated flower detection, including studies of flowering phenology and floral resources, and that this approach can, therefore, complement a wide range of ecological approaches (e.g., field observations, experiments, community science, etc.). In all, our study suggests that ecological metrics like floral richness can be efficiently monitored by combining machine learning with easily accessible publicly curated datasets (e.g., Flickr, iNaturalist).
{"title":"Using photographs and deep neural networks to understand flowering phenology and diversity in mountain meadows","authors":"Aji John, Elli J. Theobald, Nicoleta Cristea, Amanda Tan, Janneke Hille Ris Lambers","doi":"10.1002/rse2.382","DOIUrl":"https://doi.org/10.1002/rse2.382","url":null,"abstract":"Mountain meadows are an essential part of the alpine–subalpine ecosystem; they provide ecosystem services like pollination and are home to diverse plant communities. Changes in climate affect meadow ecology on multiple levels, for example, by altering growing season dynamics. Tracking the effects of climate change on meadow diversity through the impacts on individual species and overall growing season dynamics is critical to conservation efforts. Here, we explore how to combine crowd-sourced camera images with machine learning to quantify flowering species richness across a range of elevations in alpine meadows located in Mt. Rainier National Park, Washington, USA. We employed three machine-learning techniques (Mask R-CNN, RetinaNet and YOLOv5) to detect wildflower species in images taken during two flowering seasons. We demonstrate that deep learning techniques can detect multiple species, providing information on flowering richness in photographed meadows. The results indicate higher richness just above the tree line for most of the species, which is comparable with patterns found using field studies. We found that the two-stage detector Mask R-CNN was more accurate than single-stage detectors like RetinaNet and YOLO, with the Mask R-CNN network performing best overall with mean average precision (mAP) of 0.67 followed by RetinaNet (0.5) and YOLO (0.4). We found that across the methods using anchor box variations in multiples of 16 led to enhanced accuracy. We also show that detection is possible even when pictures are interspersed with complex backgrounds and are not in focus. We found differential detection rates depending on species abundance, with additional challenges related to similarity in flower characteristics, labeling errors and occlusion issues. Despite these potential biases and limitations in capturing flowering abundance and location-specific quantification, accuracy was notable considering the complexity of flower types and picture angles in this dataset. We, therefore, expect that this approach can be used to address many ecological questions that benefit from automated flower detection, including studies of flowering phenology and floral resources, and that this approach can, therefore, complement a wide range of ecological approaches (e.g., field observations, experiments, community science, etc.). In all, our study suggests that ecological metrics like floral richness can be efficiently monitored by combining machine learning with easily accessible publicly curated datasets (e.g., Flickr, iNaturalist).","PeriodicalId":21132,"journal":{"name":"Remote Sensing in Ecology and Conservation","volume":"6 1","pages":""},"PeriodicalIF":5.5,"publicationDate":"2024-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139911341","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}
Colin R. Swider, Daniela Hedwig, Peter H. Wrege, Susan E. Parks
Passive acoustic monitoring (PAM) is an effective remote sensing approach for sampling acoustically active animal species and is particularly useful for elusive, visually cryptic species inhabiting remote or inaccessible habitats. Key advantages of PAM are large spatial coverage and continuous, long-term monitoring. In most cases, a signal detection algorithm is utilized to locate sounds of interest within long sequences of audio data. It is important to understand the demographic/contextual usage of call types when choosing a particular signal to use for detection. Sampling biases may result if sampling is restricted to subsets of the population, for example, when detectable vocalizations are produced only by a certain demographic class. Using the African forest elephant repertoire as a case study, we test for differences in call type usage among different age-sex classes. We identified disproportionate usage by age-sex class of four call types—roars, trumpets, rumbles, and combination calls. This differential usage of signals by demographic class has implications for the use of particular call types in PAM for this species. Our results highlight that forest elephant PAM studies that have used rumbles as target signals may have under-sampled adult males. The addition of other call types to PAM frameworks may be useful to leverage additional population demographic information from these surveys. Our research exemplifies how an examination of a species' acoustic behavior can be used to better contextualize the data and results from PAM and to strengthen the resulting inference.
{"title":"Implications of target signal choice in passive acoustic monitoring: an example of age- and sex-dependent vocal repertoire use in African forest elephants (Loxodonta cyclotis)","authors":"Colin R. Swider, Daniela Hedwig, Peter H. Wrege, Susan E. Parks","doi":"10.1002/rse2.380","DOIUrl":"https://doi.org/10.1002/rse2.380","url":null,"abstract":"Passive acoustic monitoring (PAM) is an effective remote sensing approach for sampling acoustically active animal species and is particularly useful for elusive, visually cryptic species inhabiting remote or inaccessible habitats. Key advantages of PAM are large spatial coverage and continuous, long-term monitoring. In most cases, a signal detection algorithm is utilized to locate sounds of interest within long sequences of audio data. It is important to understand the demographic/contextual usage of call types when choosing a particular signal to use for detection. Sampling biases may result if sampling is restricted to subsets of the population, for example, when detectable vocalizations are produced only by a certain demographic class. Using the African forest elephant repertoire as a case study, we test for differences in call type usage among different age-sex classes. We identified disproportionate usage by age-sex class of four call types—roars, trumpets, rumbles, and combination calls. This differential usage of signals by demographic class has implications for the use of particular call types in PAM for this species. Our results highlight that forest elephant PAM studies that have used rumbles as target signals may have under-sampled adult males. The addition of other call types to PAM frameworks may be useful to leverage additional population demographic information from these surveys. Our research exemplifies how an examination of a species' acoustic behavior can be used to better contextualize the data and results from PAM and to strengthen the resulting inference.","PeriodicalId":21132,"journal":{"name":"Remote Sensing in Ecology and Conservation","volume":"56 1","pages":""},"PeriodicalIF":5.5,"publicationDate":"2024-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139396180","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}
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：