12.
Maso, Joan; Brobia, Alba; Voidrot, Marie-Francoise; Zabala, Alaitz; Serral, Ivette
In: Remote Sensing, vol. 15, iss. 6, no. 1589, 2023, ISSN: 2072-4292.
Abstract | Links | BibTeX | Tags:
@article{nokey,
title = {G-reqs, a New Model Proposal for Capturing and Managing In Situ Data Requirements: First Results in the Context of the Group on Earth Observations},
author = {Joan Maso and Alba Brobia and Marie-Francoise Voidrot and Alaitz Zabala and Ivette Serral},
url = {https://www.mdpi.com/2072-4292/15/6/1589},
doi = {10.3390/rs15061589},
issn = {2072-4292},
year = {2023},
date = {2023-03-15},
urldate = {2023-03-15},
journal = {Remote Sensing},
volume = {15},
number = {1589},
issue = {6},
abstract = {In the field of Earth observation, the importance of in situ data was recognized by the Group on Earth Observations (GEO) in the Canberra Declaration in 2019. The GEO community focuses on three global priority engagement areas: the United Nations 2030 Agenda for Sustainable Development, the Paris Agreement, and the Sendai Framework for Disaster Risk Reduction. While efforts have been made by GEO to open and disseminate in situ data, GEO did not have a general way to capture in situ data user requirements and drive the data provider efforts to meet the goals of its three global priorities. We present a requirements data model that first formalizes the collection of user requirements motivated by user-driven needs. Then, the user requirements can be grouped by essential variable and an analysis can derive product requirements and parameters for new or existing products. The work was inspired by thematic initiatives, such as OSCAR, from WMO, OSAAP (formerly COURL and NOSA) from NOAA, and the Copernicus In Situ Component Information System. The presented solution focuses on requirements for all applications of Earth observation in situ data. We present initial developments and testing of the data model and discuss the steps that GEO should take to implement a requirements database that is connected to actual data in the GEOSS platform and propose some recommendations on how to articulate it.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
In the field of Earth observation, the importance of in situ data was recognized by the Group on Earth Observations (GEO) in the Canberra Declaration in 2019. The GEO community focuses on three global priority engagement areas: the United Nations 2030 Agenda for Sustainable Development, the Paris Agreement, and the Sendai Framework for Disaster Risk Reduction. While efforts have been made by GEO to open and disseminate in situ data, GEO did not have a general way to capture in situ data user requirements and drive the data provider efforts to meet the goals of its three global priorities. We present a requirements data model that first formalizes the collection of user requirements motivated by user-driven needs. Then, the user requirements can be grouped by essential variable and an analysis can derive product requirements and parameters for new or existing products. The work was inspired by thematic initiatives, such as OSCAR, from WMO, OSAAP (formerly COURL and NOSA) from NOAA, and the Copernicus In Situ Component Information System. The presented solution focuses on requirements for all applications of Earth observation in situ data. We present initial developments and testing of the data model and discuss the steps that GEO should take to implement a requirements database that is connected to actual data in the GEOSS platform and propose some recommendations on how to articulate it.
11.
Vekuri, Henriikka; Tuovinen, Juha-Pekka; Kulmala, Liisa; Papale, Dario; Kolari, Pasi; Aurela, Mika; Laurila, Tuomas; Liski, Jari; Lohila, Annalea
A widely-used eddy covariance gap-filling method creates systematic bias in carbon balance estimates Journal Article
In: 2023.
Abstract | Links | BibTeX | Tags:
@article{nokeyg,
title = {A widely-used eddy covariance gap-filling method creates systematic bias in carbon balance estimates},
author = {Henriikka Vekuri and Juha-Pekka Tuovinen and Liisa Kulmala and Dario Papale and Pasi Kolari and Mika Aurela and Tuomas Laurila and Jari Liski and Annalea Lohila },
url = {https://www.nature.com/articles/s41598-023-28827-2#additional-information},
doi = {10.1038/s41598-023-28827-2},
year = {2023},
date = {2023-01-31},
abstract = {Climate change mitigation requires, besides reductions in greenhouse gas emissions, actions to increase carbon sinks in terrestrial ecosystems. A key measurement method for quantifying such sinks and calibrating models is the eddy covariance technique, but it requires imputation, or gap-filling, of missing data for determination of annual carbon balances of ecosystems. Previous comparisons of gap-filling methods have concluded that commonly used methods, such as marginal distribution sampling (MDS), do not have a significant impact on the carbon balance estimate. By analyzing an extensive, global data set, we show that MDS causes significant carbon balance errors for northern (latitude >60∘
) sites. MDS systematically overestimates the carbon dioxide (CO2
) emissions of carbon sources and underestimates the CO2
sequestration of carbon sinks. We also reveal reasons for these biases and show how a machine learning method called extreme gradient boosting or a modified implementation of MDS can be used to substantially reduce the northern site bias.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Climate change mitigation requires, besides reductions in greenhouse gas emissions, actions to increase carbon sinks in terrestrial ecosystems. A key measurement method for quantifying such sinks and calibrating models is the eddy covariance technique, but it requires imputation, or gap-filling, of missing data for determination of annual carbon balances of ecosystems. Previous comparisons of gap-filling methods have concluded that commonly used methods, such as marginal distribution sampling (MDS), do not have a significant impact on the carbon balance estimate. By analyzing an extensive, global data set, we show that MDS causes significant carbon balance errors for northern (latitude >60∘
) sites. MDS systematically overestimates the carbon dioxide (CO2
) emissions of carbon sources and underestimates the CO2
sequestration of carbon sinks. We also reveal reasons for these biases and show how a machine learning method called extreme gradient boosting or a modified implementation of MDS can be used to substantially reduce the northern site bias.
) sites. MDS systematically overestimates the carbon dioxide (CO2
) emissions of carbon sources and underestimates the CO2
sequestration of carbon sinks. We also reveal reasons for these biases and show how a machine learning method called extreme gradient boosting or a modified implementation of MDS can be used to substantially reduce the northern site bias.
10.
Nesha, Karimon; Herold, Martin; Sy, Veronique De; de Bruin, Sytze; Araza, Arnan; Málaga, Natalia; Gamarra, Javier G. P.; Hergoualc'h, Kristell; Pekkarinen, Anssi; Ramirez, Carla; Morales-Hidalgo, David; Tavani, Rebecca
Exploring characteristics of national forest inventories for integration with global space-based forest biomass data Journal Article
In: Science of The Total Environment, vol. 850, iss. 157788, 2022.
Abstract | Links | BibTeX | Tags: Aboveground biomass, CCI biomass, Creative Commons, Forest Resources Assessment, National Forest Inventory, NFI and space-based data integration, NFI plot design
@article{@article{NESHA2022157788,
title = {Exploring characteristics of national forest inventories for integration with global space-based forest biomass data},
author = {Karimon Nesha and Martin Herold and Veronique {De Sy} and Sytze {de Bruin} and Arnan Araza and Natalia Málaga and Javier G.P. Gamarra and Kristell Hergoualc'h and Anssi Pekkarinen and Carla Ramirez and David Morales-Hidalgo and Rebecca Tavani},
url = {https://www.sciencedirect.com/science/article/pii/S0048969722048872?via%3Dihub},
doi = {https://doi.org/10.1016/j.scitotenv.2022.157788},
year = {2022},
date = {2022-12-01},
journal = {Science of The Total Environment},
volume = {850},
issue = {157788},
abstract = {National forest inventories (NFIs) are a reliable source for national forest measurements. However, they are usually not developed for linking with remotely sensed (RS) biomass information. There are increasing needs and opportunities to facilitate this link towards better global and national biomass estimation. Thus, it is important to study and understand NFI characteristics relating to their integration with space-based products; in particular for the tropics where NFIs are quite recent, less frequent, and partially incomplete in several countries. Here, we (1) assessed NFIs in terms of their availability, temporal distribution, and extent in 236 countries from FAO's Global Forest Resources Assessment (FRA) 2020; (2) compared national forest biomass estimates in 2018 from FRA and global space-based Climate Change Initiative (CCI) product in 182 countries considering NFI availability and temporality; and (3) analyzed the latest NFI design characteristics in 46 tropical countries relating to their integration with space-based biomass datasets. We observed significant NFI availability globally and multiple NFIs were mostly found in temperate and boreal countries while most of the single NFI countries (94 %) were in the tropics. The latest NFIs were more recent in the tropics and many countries (35) implemented NFIs from 2016 onwards. The increasing availability and update of NFIs create new opportunities for integration with space-based data at the national level. This is supported by the agreement we found between country biomass estimates for 2018 from FRA and CCI product, with a significantly higher correlation in countries with recent NFIs. We observed that NFI designs varied greatly in tropical countries. For example, the size of the plots ranged from 0.01 to 1 ha and more than three-quarters of the countries had smaller plots of ≤0.25 ha. The existing NFI designs could pose specific challenges for statistical integration with RS data in the tropics. Future NFI and space-based efforts should aim towards a more integrated approach taking advantage of both data streams to improve national estimates and help future data harmonization efforts. Regular NFI efforts can be expanded with the inclusion of some super-site plots to enhance data integration with currently available space-based applications. Issues related to cost implications versus improvements in the accuracy, timeliness, and sustainability of national forest biomass estimation should be further explored.},
keywords = {Aboveground biomass, CCI biomass, Creative Commons, Forest Resources Assessment, National Forest Inventory, NFI and space-based data integration, NFI plot design},
pubstate = {published},
tppubtype = {article}
}
National forest inventories (NFIs) are a reliable source for national forest measurements. However, they are usually not developed for linking with remotely sensed (RS) biomass information. There are increasing needs and opportunities to facilitate this link towards better global and national biomass estimation. Thus, it is important to study and understand NFI characteristics relating to their integration with space-based products; in particular for the tropics where NFIs are quite recent, less frequent, and partially incomplete in several countries. Here, we (1) assessed NFIs in terms of their availability, temporal distribution, and extent in 236 countries from FAO's Global Forest Resources Assessment (FRA) 2020; (2) compared national forest biomass estimates in 2018 from FRA and global space-based Climate Change Initiative (CCI) product in 182 countries considering NFI availability and temporality; and (3) analyzed the latest NFI design characteristics in 46 tropical countries relating to their integration with space-based biomass datasets. We observed significant NFI availability globally and multiple NFIs were mostly found in temperate and boreal countries while most of the single NFI countries (94 %) were in the tropics. The latest NFIs were more recent in the tropics and many countries (35) implemented NFIs from 2016 onwards. The increasing availability and update of NFIs create new opportunities for integration with space-based data at the national level. This is supported by the agreement we found between country biomass estimates for 2018 from FRA and CCI product, with a significantly higher correlation in countries with recent NFIs. We observed that NFI designs varied greatly in tropical countries. For example, the size of the plots ranged from 0.01 to 1 ha and more than three-quarters of the countries had smaller plots of ≤0.25 ha. The existing NFI designs could pose specific challenges for statistical integration with RS data in the tropics. Future NFI and space-based efforts should aim towards a more integrated approach taking advantage of both data streams to improve national estimates and help future data harmonization efforts. Regular NFI efforts can be expanded with the inclusion of some super-site plots to enhance data integration with currently available space-based applications. Issues related to cost implications versus improvements in the accuracy, timeliness, and sustainability of national forest biomass estimation should be further explored.
9.
Câmara, Gilberto
Challenges to Achieving the Commitments of Brazil’s NDC in the Amazon Biome Journal Article
In: 2022.
Abstract | Links | BibTeX | Tags:
@article{nokeym,
title = {Challenges to Achieving the Commitments of Brazil’s NDC in the Amazon Biome},
author = {Gilberto Câmara},
url = {https://cebri.org/revista/en/artigo/63/challenges-to-achieving-the-commitments-of-brazils-ndc-in-the-amazon-biome},
year = {2022},
date = {2022-11-28},
abstract = {The Brazilian Nationally Determined Contribution (NDC) established ambitious greenhouse gas emissions reduction targets. Using the Brazilian Forestry Code as reference, we describe the challenges Brazil will face to achieve its deforestation and forest restoration NDC commitments in the Amazon biome. This paper proposes a new perspective for public policy focusing on the extent of illegal deforestation and on current land use in the region.
},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The Brazilian Nationally Determined Contribution (NDC) established ambitious greenhouse gas emissions reduction targets. Using the Brazilian Forestry Code as reference, we describe the challenges Brazil will face to achieve its deforestation and forest restoration NDC commitments in the Amazon biome. This paper proposes a new perspective for public policy focusing on the extent of illegal deforestation and on current land use in the region.
8.
Málaga, Natalia; de Bruin, Sytze; McRoberts, Ronald E.; Olivos, Alexs Arana; de la Cruz Paiva, Ricardo; Montesinos, Patricia Durán; Suarez, Daniela Requena; Herold, Martin
Precision of subnational forest AGB estimates within the Peruvian Amazonia using a global biomass map Journal Article
In: vol. 115, 2022.
Abstract | Links | BibTeX | Tags:
@article{nokeyh,
title = {Precision of subnational forest AGB estimates within the Peruvian Amazonia using a global biomass map},
author = {Natalia Málaga and Sytze de Bruin and Ronald E. McRoberts and Alexs Arana Olivos and Ricardo de la Cruz Paiva and Patricia Durán Montesinos and Daniela Requena Suarez and Martin Herold},
url = {https://www.sciencedirect.com/science/article/pii/S1569843222002904?via%3Dihub},
doi = {10.1016/j.jag.2022.103102},
year = {2022},
date = {2022-11-18},
volume = {115},
abstract = {National forest inventories (NFI) provide essential forest-related biomass and carbon information for country greenhouse gas (GHG) accounting systems. Several tropical countries struggle to execute their NFIs while the extent to which space-based global information on aboveground biomass (AGB) can support national GHG accounting is under investigation. We assess whether the use of a global AGB map as auxiliary information produces a gain in precision of subnational AGB estimates for the Peruvian Amazonia. We used model-assisted estimators with data from the country’s NFI and explored hybrid inferential techniques to account for the sources of uncertainty associated with the integration of remote sensing-based products and NFI plot data.
Our results show that the selected global biomass map tends to overestimate AGB values across the Peruvian Amazonia. For most strata, directly using the map in its published form did not reduce the precision of AGB estimates. However, after calibrating the map using the NFI data, the precision of our map-assisted AGB estimates increased by up to 50% at stratum level and 20% at Amazonia level. We further demonstrate how different sources of uncertainties can be incorporated in the map-NFI integrated estimates. With the hybrid inferential analysis, we found that the small spatial support of the NFI plots compared to the remote sensing-based sample units of aggregated pixels (within block variability) contributed the most to the total uncertainty associated with the AGB estimates from our map-NFI integration. Uncertainties caused by measurement variability and allometric model prediction uncertainty were the second largest contributors. When these uncertainties were incorporated, the increase in precision of our calibrated map-assisted AGB estimates was negligible, probably hindered by the great contribution of the within block variability to our map-plot assessment. We developed a reproducible method that countries can build upon and further improve while the global biomass products continue to evolve and better characterize the AGB distribution under large biomass conditions. We encourage further cross-country case studies that reflect a wider range of AGB distributions, especially within humid tropical forests, to further assess the contribution of global biomass maps to (sub)national AGB estimates and finally GHG monitoring and reporting.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
National forest inventories (NFI) provide essential forest-related biomass and carbon information for country greenhouse gas (GHG) accounting systems. Several tropical countries struggle to execute their NFIs while the extent to which space-based global information on aboveground biomass (AGB) can support national GHG accounting is under investigation. We assess whether the use of a global AGB map as auxiliary information produces a gain in precision of subnational AGB estimates for the Peruvian Amazonia. We used model-assisted estimators with data from the country’s NFI and explored hybrid inferential techniques to account for the sources of uncertainty associated with the integration of remote sensing-based products and NFI plot data.
Our results show that the selected global biomass map tends to overestimate AGB values across the Peruvian Amazonia. For most strata, directly using the map in its published form did not reduce the precision of AGB estimates. However, after calibrating the map using the NFI data, the precision of our map-assisted AGB estimates increased by up to 50% at stratum level and 20% at Amazonia level. We further demonstrate how different sources of uncertainties can be incorporated in the map-NFI integrated estimates. With the hybrid inferential analysis, we found that the small spatial support of the NFI plots compared to the remote sensing-based sample units of aggregated pixels (within block variability) contributed the most to the total uncertainty associated with the AGB estimates from our map-NFI integration. Uncertainties caused by measurement variability and allometric model prediction uncertainty were the second largest contributors. When these uncertainties were incorporated, the increase in precision of our calibrated map-assisted AGB estimates was negligible, probably hindered by the great contribution of the within block variability to our map-plot assessment. We developed a reproducible method that countries can build upon and further improve while the global biomass products continue to evolve and better characterize the AGB distribution under large biomass conditions. We encourage further cross-country case studies that reflect a wider range of AGB distributions, especially within humid tropical forests, to further assess the contribution of global biomass maps to (sub)national AGB estimates and finally GHG monitoring and reporting.
Our results show that the selected global biomass map tends to overestimate AGB values across the Peruvian Amazonia. For most strata, directly using the map in its published form did not reduce the precision of AGB estimates. However, after calibrating the map using the NFI data, the precision of our map-assisted AGB estimates increased by up to 50% at stratum level and 20% at Amazonia level. We further demonstrate how different sources of uncertainties can be incorporated in the map-NFI integrated estimates. With the hybrid inferential analysis, we found that the small spatial support of the NFI plots compared to the remote sensing-based sample units of aggregated pixels (within block variability) contributed the most to the total uncertainty associated with the AGB estimates from our map-NFI integration. Uncertainties caused by measurement variability and allometric model prediction uncertainty were the second largest contributors. When these uncertainties were incorporated, the increase in precision of our calibrated map-assisted AGB estimates was negligible, probably hindered by the great contribution of the within block variability to our map-plot assessment. We developed a reproducible method that countries can build upon and further improve while the global biomass products continue to evolve and better characterize the AGB distribution under large biomass conditions. We encourage further cross-country case studies that reflect a wider range of AGB distributions, especially within humid tropical forests, to further assess the contribution of global biomass maps to (sub)national AGB estimates and finally GHG monitoring and reporting.
7.
Milenkovic, Milutin
Big EO Data Processing Approaches for Monitoring Tropical Forest Disturbance with Sentinel-1 Data Presentation
15.11.2022.
Abstract | BibTeX | Tags: Earth Observations, Presentation, Sentinel-1, tropical forests
@misc{nokey,
title = {Big EO Data Processing Approaches for Monitoring Tropical Forest Disturbance with Sentinel-1 Data},
author = {Milutin Milenkovic},
year = {2022},
date = {2022-11-15},
abstract = {An increase in the disturbance of tropical forests, mainly caused by human-driven deforestation, drouths, and forest fires, has been constantly reported over the last decades. Those stressors are pushing tropical ecosystems to their limits, and it is urgently needed to understand how far the ecosystems are from their point of no return. Novel global and free Earth Observation (EO) data, such as Sentinel-1 satellite radar images from the EU’s Copernicus programme, have high spatial (20 m) and temporal (6 or more days) resolution, providing an excellent possibility for monitoring tropical ecosystem change. However, those terabytes of data also require novel processing ways. In this talk, I will present three different big data approaches utilized to detect Sentinel-1 signal disturbances and recoveries over tropical forests in the Amazon and Congo basins. The presentation will give a user perspective and experience on the Sentinel-1 processing on local servers, the ESA openEO platform, and a national computing infrastructure used within the C-SCALE project. A novel EU-funded project on big EO data processing, the open Earth monitor cyberinfrastructure, will also be introduced.
Event: Enhancing Satellite Imagery and Geospatial Information Capabilities in Support of Nuclear Safeguards
Place: UN, IAEA Headquarters, Vienna, Austria
Date: 15–17 November 2022},
keywords = {Earth Observations, Presentation, Sentinel-1, tropical forests},
pubstate = {published},
tppubtype = {presentation}
}
An increase in the disturbance of tropical forests, mainly caused by human-driven deforestation, drouths, and forest fires, has been constantly reported over the last decades. Those stressors are pushing tropical ecosystems to their limits, and it is urgently needed to understand how far the ecosystems are from their point of no return. Novel global and free Earth Observation (EO) data, such as Sentinel-1 satellite radar images from the EU’s Copernicus programme, have high spatial (20 m) and temporal (6 or more days) resolution, providing an excellent possibility for monitoring tropical ecosystem change. However, those terabytes of data also require novel processing ways. In this talk, I will present three different big data approaches utilized to detect Sentinel-1 signal disturbances and recoveries over tropical forests in the Amazon and Congo basins. The presentation will give a user perspective and experience on the Sentinel-1 processing on local servers, the ESA openEO platform, and a national computing infrastructure used within the C-SCALE project. A novel EU-funded project on big EO data processing, the open Earth monitor cyberinfrastructure, will also be introduced.
Event: Enhancing Satellite Imagery and Geospatial Information Capabilities in Support of Nuclear Safeguards
Place: UN, IAEA Headquarters, Vienna, Austria
Date: 15–17 November 2022
Event: Enhancing Satellite Imagery and Geospatial Information Capabilities in Support of Nuclear Safeguards
Place: UN, IAEA Headquarters, Vienna, Austria
Date: 15–17 November 2022
6.
Mîrț, Andrei; Reiche, Johannes; Verbesselt, Jan; Herold, Martin
A Downsampling Method Addressing the Modifiable Areal Unit Problem in Remote Sensing Journal Article
In: Remote Sensing, vol. 14, iss. 22, no. 5538, 2022.
Abstract | Links | BibTeX | Tags: big data, downsampling, modifiable areal unit problem, multscale, remote sensing
@article{nokey,
title = {A Downsampling Method Addressing the Modifiable Areal Unit Problem in Remote Sensing},
author = {Andrei Mîrț and Johannes Reiche and Jan Verbesselt and Martin Herold},
url = {https://www.mdpi.com/2072-4292/14/21/5538},
doi = {https://doi.org/10.3390/rs14215538},
year = {2022},
date = {2022-11-03},
journal = {Remote Sensing},
volume = {14},
number = {5538},
issue = {22},
abstract = {Handling multiple scales efficiently is one avenue for processing big remote sensing imagery data. Unfortunately, imagery is also affected by the infamous modifiable areal unit problem, which creates unpredictable errors at different scales. We developed a downsampling method that attempts to keep the data distribution in a downsampled image constant, reducing the modifiable areal unit problem. We tested our method against classic downsampling methods (mean, central pixel selection, random) under a range of typical remote sensing scenarios. Under our experimental conditions, our downsampling method consistently outperformed the classical downsampling methods within a 95% confidence level. The downsampling method can be used in most typical situations where downsampling is needed, but it is likely to shine when used as a pyramid building policy in geocomputing platforms, such as Google Earth Engine.},
keywords = {big data, downsampling, modifiable areal unit problem, multscale, remote sensing},
pubstate = {published},
tppubtype = {article}
}
Handling multiple scales efficiently is one avenue for processing big remote sensing imagery data. Unfortunately, imagery is also affected by the infamous modifiable areal unit problem, which creates unpredictable errors at different scales. We developed a downsampling method that attempts to keep the data distribution in a downsampled image constant, reducing the modifiable areal unit problem. We tested our method against classic downsampling methods (mean, central pixel selection, random) under a range of typical remote sensing scenarios. Under our experimental conditions, our downsampling method consistently outperformed the classical downsampling methods within a 95% confidence level. The downsampling method can be used in most typical situations where downsampling is needed, but it is likely to shine when used as a pyramid building policy in geocomputing platforms, such as Google Earth Engine.
5.
Labrière, Nicolas; Davies, Stuart J.; Disney, Mathias I.; Duncanson, Laura I.; Herold, Martin; Lewis, Simon L.; Phillips, Oliver L.; Quegan, Shaun; Saatchi, Sassan S.; Schepaschenko, Dmitry G.; Scipal, Klaus; Sist, Plinio; Chave, Jérôme
Toward a forest biomass reference measurement system for remote sensing applications Journal Article
In: vol. 29, iss. 3, pp. 827-840, 2022.
Abstract | Links | BibTeX | Tags:
@article{nokeye,
title = {Toward a forest biomass reference measurement system for remote sensing applications},
author = {Nicolas Labrière and Stuart J. Davies and Mathias I. Disney and Laura I. Duncanson and Martin Herold and Simon L. Lewis and Oliver L. Phillips and Shaun Quegan and Sassan S. Saatchi and Dmitry G. Schepaschenko and Klaus Scipal and Plinio Sist and Jérôme Chave},
url = {https://onlinelibrary.wiley.com/doi/10.1111/gcb.16497},
doi = {10.1111/gcb.16497},
year = {2022},
date = {2022-10-21},
volume = {29},
issue = {3},
pages = {827-840},
abstract = {Forests contribute to climate change mitigation through carbon storage and uptake, but the extent to which this carbon pool varies in space and time is still poorly known. Several Earth Observation missions have been specifically designed to address this issue, for example, NASA's GEDI, NASA-ISRO's NISAR and ESA's BIOMASS. Yet, all these missions' products require independent and consistent validation. A permanent, global, in situ, site-based forest biomass reference measurement system relying on ground data of the highest possible quality is therefore needed. Here, we have assembled a list of almost 200 high-quality sites through an in-depth review of the literature and expert knowledge. In this study, we explore how representative these sites are in terms of their coverage of environmental conditions, geographical space and biomass-related forest structure, compared to those experienced by forests worldwide. This work also aims at identifying which sites are the most representative, and where to invest to improve the representativeness of the proposed system. We show that the environmental coverage of the system does not seem to improve after at least the 175 most representative sites are included, but geographical and structural coverages continue to improve as more sites are added. We highlight the areas of poor environmental, geographical, or structural coverage, including, but not limited to, Canada, the western half of the USA, Mexico, Patagonia, Angola, Zambia, eastern Russia, and tropical and subtropical highlands (e.g. in Colombia, the Himalayas, Borneo, Papua). For the proposed system to succeed, we stress that (1) data must be collected and processed applying the same standards across all countries and continents; (2) system establishment and management must be inclusive and equitable, with careful consideration of working conditions; and (3) training and site partner involvement in downstream activities should be mandatory.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Forests contribute to climate change mitigation through carbon storage and uptake, but the extent to which this carbon pool varies in space and time is still poorly known. Several Earth Observation missions have been specifically designed to address this issue, for example, NASA's GEDI, NASA-ISRO's NISAR and ESA's BIOMASS. Yet, all these missions' products require independent and consistent validation. A permanent, global, in situ, site-based forest biomass reference measurement system relying on ground data of the highest possible quality is therefore needed. Here, we have assembled a list of almost 200 high-quality sites through an in-depth review of the literature and expert knowledge. In this study, we explore how representative these sites are in terms of their coverage of environmental conditions, geographical space and biomass-related forest structure, compared to those experienced by forests worldwide. This work also aims at identifying which sites are the most representative, and where to invest to improve the representativeness of the proposed system. We show that the environmental coverage of the system does not seem to improve after at least the 175 most representative sites are included, but geographical and structural coverages continue to improve as more sites are added. We highlight the areas of poor environmental, geographical, or structural coverage, including, but not limited to, Canada, the western half of the USA, Mexico, Patagonia, Angola, Zambia, eastern Russia, and tropical and subtropical highlands (e.g. in Colombia, the Himalayas, Borneo, Papua). For the proposed system to succeed, we stress that (1) data must be collected and processed applying the same standards across all countries and continents; (2) system establishment and management must be inclusive and equitable, with careful consideration of working conditions; and (3) training and site partner involvement in downstream activities should be mandatory.
4.
Masolele, Robert N.; Sy, Veronique De; Marcos, Diego; Verbesselt, Jan; Gieseke, Fabian; Mulatu, Kalkidan Ayele; Moges, Yitebitu; Sebrala, Heiru; Martius, Christopher; Herold, Martin
Using high-resolution imagery and deep learning to classify land-use following deforestation: a case study in Ethiopia Journal Article
In: GIScience & Remote Sensing, vol. 59, no. 1, pp. 1446-1472, 2022.
Abstract | Links | BibTeX | Tags: Attention U-Net, deep learning, deforestation drivers, Planet-NFCI, remote sensing
@article{@article{doi:10.1080/15481603.2022.2115619,
title = {Using high-resolution imagery and deep learning to classify land-use following deforestation: a case study in Ethiopia},
author = {Robert N. Masolele and Veronique De Sy and Diego Marcos and Jan Verbesselt and Fabian Gieseke and Kalkidan Ayele Mulatu and Yitebitu Moges and Heiru Sebrala and Christopher Martius and Martin Herold},
url = {https://doi.org/10.1080/15481603.2022.2115619
},
doi = {https://doi.org/10.1080/15481603.2022.2115619},
year = {2022},
date = {2022-09-07},
urldate = {2022-09-07},
journal = {GIScience & Remote Sensing},
volume = {59},
number = {1},
pages = {1446-1472},
abstract = {National-scale assessments of post-deforestation land-use are crucial for decreasing deforestation and forest degradation-related emissions. In this research, we assess the potential of different satellite data modalities (single-date, multi-date, multi-resolution, and an ensemble of multi-sensor images) for classifying land-use following deforestation in Ethiopia using the U-Net deep neural network architecture enhanced with attention. We performed the analysis on satellite image data retrieved across Ethiopia from freely available Landsat-8, Sentinel-2 and Planet-NICFI satellite data. The experiments aimed at an analysis of (a) single-date images from individual sensors to account for the differences in spatial resolution between image sensors in detecting land-uses, (b) ensembles of multiple images from different sensors (Planet-NICFI/Sentinel-2/Landsat-8) with different spatial resolutions, (c) the use of multi-date data to account for the contribution of temporal information in detecting land-uses, and, finally, (d) the identification of regional differences in terms of land-use following deforestation in Ethiopia. We hypothesize that choosing the right satellite imagery (sensor) type is crucial for the task. Based on a comprehensive visually interpreted reference dataset of 11 types of post-deforestation land-uses, we find that either detailed spatial patterns (single-date Planet-NICFI) or detailed temporal patterns (multi-date Sentinel-2, Landsat-8) are required for identifying land-use following deforestation, while medium-resolution single-date imagery is not sufficient to achieve high classification accuracy. We also find that adding soft-attention to the standard U-Net improved the classification accuracy, especially for small-scale land-uses. The models and products presented in this work can be used as a powerful data resource for governmental and forest monitoring agencies to design and monitor deforestation mitigation measures and data-driven land-use policy.},
keywords = {Attention U-Net, deep learning, deforestation drivers, Planet-NFCI, remote sensing},
pubstate = {published},
tppubtype = {article}
}
National-scale assessments of post-deforestation land-use are crucial for decreasing deforestation and forest degradation-related emissions. In this research, we assess the potential of different satellite data modalities (single-date, multi-date, multi-resolution, and an ensemble of multi-sensor images) for classifying land-use following deforestation in Ethiopia using the U-Net deep neural network architecture enhanced with attention. We performed the analysis on satellite image data retrieved across Ethiopia from freely available Landsat-8, Sentinel-2 and Planet-NICFI satellite data. The experiments aimed at an analysis of (a) single-date images from individual sensors to account for the differences in spatial resolution between image sensors in detecting land-uses, (b) ensembles of multiple images from different sensors (Planet-NICFI/Sentinel-2/Landsat-8) with different spatial resolutions, (c) the use of multi-date data to account for the contribution of temporal information in detecting land-uses, and, finally, (d) the identification of regional differences in terms of land-use following deforestation in Ethiopia. We hypothesize that choosing the right satellite imagery (sensor) type is crucial for the task. Based on a comprehensive visually interpreted reference dataset of 11 types of post-deforestation land-uses, we find that either detailed spatial patterns (single-date Planet-NICFI) or detailed temporal patterns (multi-date Sentinel-2, Landsat-8) are required for identifying land-use following deforestation, while medium-resolution single-date imagery is not sufficient to achieve high classification accuracy. We also find that adding soft-attention to the standard U-Net improved the classification accuracy, especially for small-scale land-uses. The models and products presented in this work can be used as a powerful data resource for governmental and forest monitoring agencies to design and monitor deforestation mitigation measures and data-driven land-use policy.
3.
Hengl, Tomislav
Ways forward for Machine Learning to make useful global environmental datasets from legacy observations and measurements Journal Article
In: Nature Communications, vol. 13, no. 5178, 2022.
Abstract | Links | BibTeX | Tags:
@article{nokey,
title = {Ways forward for Machine Learning to make useful global environmental datasets from legacy observations and measurements},
author = {Tomislav Hengl },
doi = {https://doi.org/10.1038/s41467-022-32693-3},
year = {2022},
date = {2022-09-07},
urldate = {2022-09-07},
journal = {Nature Communications},
volume = {13},
number = {5178},
abstract = {Advances in geospatial and Machine Learning techniques for large datasets of georeferenced observations have made it possible to produce model-based global maps of ecological and environmental variables. However, the implementation of existing scientific methods (especially Machine Learning models) to produce accurate global maps is often complex. Tomislav Hengl (co-founder of OpenGeoHub foundation), Johan van den Hoogen (researcher at ETH Zürich), and Devin Routh (Science IT Consultant at the University of Zürich) shared with Nature Communications their perspectives for creators and users of these maps, focusing on the key challenges in producing global environmental geospatial datasets to achieve significant impacts.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Advances in geospatial and Machine Learning techniques for large datasets of georeferenced observations have made it possible to produce model-based global maps of ecological and environmental variables. However, the implementation of existing scientific methods (especially Machine Learning models) to produce accurate global maps is often complex. Tomislav Hengl (co-founder of OpenGeoHub foundation), Johan van den Hoogen (researcher at ETH Zürich), and Devin Routh (Science IT Consultant at the University of Zürich) shared with Nature Communications their perspectives for creators and users of these maps, focusing on the key challenges in producing global environmental geospatial datasets to achieve significant impacts.
2.
Masiliūnas, Dainius; Marcos, Diego; Tsendbazar, Nandin-Erdene; Herold, Martin; Verbesselt, Jan
High frequency land cover classification method for supporting global monitoring Journal Article
In: 2022.
@article{nokey,
title = {High frequency land cover classification method for supporting global monitoring},
author = {Dainius Masiliūnas and Diego Marcos and Nandin-Erdene Tsendbazar and Martin Herold and Jan Verbesselt},
url = {https://zenodo.org/records/8310036},
doi = {10.5281/zenodo.8310036},
year = {2022},
date = {2022-06-14},
urldate = {2022-06-14},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
1.
Ouyang, Zutao; Jackson, Robert B.; McNicol, Gavin; Fluet-Chouinard, Etienne; Runkle, Benjamin R. K.; Papale, Dario; Knox, Sara H.; Cooley, Sarah; Delwiche, Kyle B.; Feron, Sarah; Irvin, Jeremy Andrew; Malhotra, Avni; Muddasir, Muhammad; Sabbatini, Simone; Alberto, Ma. Carmelita R.; Cescatti, Alessandro; Chen, Chi-Ling; Dong, Jinwei; Fong, Bryant N.; Guo, Haiqiang; Hao, Lu; Iwata, Hiroki; Jia, Qingyu; Ju, Weimin; Kang, Minseok; Li, Hong; Kim, Joon; Reba, Michele L.; Nayak, Amaresh Kumar; Roberti, Debora Regina; Ryu, Youngryel; Swain, Chinmaya Kumar; Tsuang, Benjei; Xiao, Xiangming; Yuan, Wenping; Zhang, Geli; Zhang, Yongguang
Paddy rice methane emissions across Monsoon Asia Journal Article
In: Remote Sensing of Environment, vol. 284, pp. 113335, 0000, ISSN: 0034-4257.
Abstract | Links | BibTeX | Tags: climate change, machine learning, remote sensing
@article{nokey,
title = {Paddy rice methane emissions across Monsoon Asia},
author = {Zutao Ouyang and Robert B. Jackson and Gavin McNicol and Etienne Fluet-Chouinard and Benjamin R.K. Runkle and Dario Papale and Sara H. Knox and Sarah Cooley and Kyle B. Delwiche and Sarah Feron and Jeremy Andrew Irvin and Avni Malhotra and Muhammad Muddasir and Simone Sabbatini and Ma. Carmelita R. Alberto and Alessandro Cescatti and Chi-Ling Chen and Jinwei Dong and Bryant N. Fong and Haiqiang Guo and Lu Hao and Hiroki Iwata and Qingyu Jia and Weimin Ju and Minseok Kang and Hong Li and Joon Kim and Michele L. Reba and Amaresh Kumar Nayak and Debora Regina Roberti and Youngryel Ryu and Chinmaya Kumar Swain and Benjei Tsuang and Xiangming Xiao and Wenping Yuan and Geli Zhang and Yongguang Zhang},
url = {https://www.sciencedirect.com/science/article/pii/S0034425722004412},
doi = {https://doi.org/10.1016/j.rse.2022.113335},
issn = {0034-4257},
journal = {Remote Sensing of Environment},
volume = {284},
pages = {113335},
abstract = {Although rice cultivation is one of the most important agricultural sources of methane (CH4) and contributes ∼8% of total global anthropogenic emissions, large discrepancies remain among estimates of global CH4 emissions from rice cultivation (ranging from 18 to 115 Tg CH4 yr−1) due to a lack of observational constraints. The spatial distribution of paddy-rice emissions has been assessed at regional-to-global scales by bottom-up inventories and land surface models over coarse spatial resolution (e.g., > 0.5°) or spatial units (e.g., agro-ecological zones). However, high-resolution CH4 flux estimates capable of capturing the effects of local climate and management practices on emissions, as well as replicating in situ data, remain challenging to produce because of the scarcity of high-resolution maps of paddy-rice and insufficient understanding of CH4 predictors. Here, we combine paddy-rice methane-flux data from 23 global eddy covariance sites and MODIS remote sensing data with machine learning to 1) evaluate data-driven model performance and variable importance for predicting rice CH4 fluxes; and 2) produce gridded up-scaling estimates of rice CH4 emissions at 5000-m resolution across Monsoon Asia, where ∼87% of global rice area is cultivated and ∼ 90% of global rice production occurs. Our random-forest model achieved Nash-Sutcliffe Efficiency values of 0.59 and 0.69 for 8-day CH4 fluxes and site mean CH4 fluxes respectively, with land surface temperature, biomass and water-availability-related indices as the most important predictors. We estimate the average annual (winter fallow season excluded) paddy rice CH4 emissions throughout Monsoon Asia to be 20.6 ± 1.1 Tg yr−1 for 2001–2015, which is at the lower range of previous inventory-based estimates (20–32 CH4 Tg yr−1). Our estimates also suggest that CH4 emissions from paddy rice in this region have been declining from 2007 through 2015 following declines in both paddy-rice growing area and emission rates per unit area, suggesting that CH4 emissions from paddy rice in Monsoon Asia have likely not contributed to the renewed growth of atmospheric CH4 in recent years.},
keywords = {climate change, machine learning, remote sensing},
pubstate = {published},
tppubtype = {article}
}
Although rice cultivation is one of the most important agricultural sources of methane (CH4) and contributes ∼8% of total global anthropogenic emissions, large discrepancies remain among estimates of global CH4 emissions from rice cultivation (ranging from 18 to 115 Tg CH4 yr−1) due to a lack of observational constraints. The spatial distribution of paddy-rice emissions has been assessed at regional-to-global scales by bottom-up inventories and land surface models over coarse spatial resolution (e.g., > 0.5°) or spatial units (e.g., agro-ecological zones). However, high-resolution CH4 flux estimates capable of capturing the effects of local climate and management practices on emissions, as well as replicating in situ data, remain challenging to produce because of the scarcity of high-resolution maps of paddy-rice and insufficient understanding of CH4 predictors. Here, we combine paddy-rice methane-flux data from 23 global eddy covariance sites and MODIS remote sensing data with machine learning to 1) evaluate data-driven model performance and variable importance for predicting rice CH4 fluxes; and 2) produce gridded up-scaling estimates of rice CH4 emissions at 5000-m resolution across Monsoon Asia, where ∼87% of global rice area is cultivated and ∼ 90% of global rice production occurs. Our random-forest model achieved Nash-Sutcliffe Efficiency values of 0.59 and 0.69 for 8-day CH4 fluxes and site mean CH4 fluxes respectively, with land surface temperature, biomass and water-availability-related indices as the most important predictors. We estimate the average annual (winter fallow season excluded) paddy rice CH4 emissions throughout Monsoon Asia to be 20.6 ± 1.1 Tg yr−1 for 2001–2015, which is at the lower range of previous inventory-based estimates (20–32 CH4 Tg yr−1). Our estimates also suggest that CH4 emissions from paddy rice in this region have been declining from 2007 through 2015 following declines in both paddy-rice growing area and emission rates per unit area, suggesting that CH4 emissions from paddy rice in Monsoon Asia have likely not contributed to the renewed growth of atmospheric CH4 in recent years.