ME
M.A. Eleveld
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
6 records found
1
Modeling gross primary productivity across different European ecosystem types
Evaluating the versatility of SARIMAX, XGBoost, and LSTM using ICOS FLUXNET and Sentinel-2 data
Journal article
(2026)
-
Anna Spinosa, Karisma Karisma, Marieke A. Eleveld, Mario Alberto Fuentes-Monjaraz, Valeria Mobilia, Ulf Mallast, Johannes Peterseil, Ghada El Serafy
Predicting Gross Primary Productivity (GPP) is key for understanding ecosystem health and quantifying the global carbon cycle. While data-driven models have shown strong performance in capturing GPP dynamics at specific sites, their ability to generalize across ecosystems without site-specific recalibration remains largely untested. This study addresses this gap by evaluating the applicability of XGBoost and LSTM models in estimating GPP across different European ecosystems. We developed a unified (cross-site) modeling framework that integrates in-situ eddy covariance observations and Sentinel-2–derived vegetation indices using incremental learning. Models’ performance was assessed via: (i) site-specific models, developed to capture individual site characteristics, and (ii) cross-site generalization, including evaluation on an independent dataset of unseen ecosystems. SARIMAX is included as a site-specific statistical benchmark for comparison. Our findings indicate that XGBoost consistently outperformed the other models, achieving site-specific R2 values above 0.90 in forest and grassland ecosystems and an average R2 of 0.72 across unseen sites (range 0.66–0.78). LSTM exhibited better accuracy in predicting GPP peaks at site-specific level, particularly in cropland and forest ecosystems. At site-level, SARIMAX showed comparable performance to XGBoost but struggled in capturing the rapid temporal variation of GPP. These findings demonstrate the feasibility of a data-driven framework for cross-site GPP monitoring within European flux-tower networks, making a first step toward transferable GPP prediction without site-specific recalibration.
...
Predicting Gross Primary Productivity (GPP) is key for understanding ecosystem health and quantifying the global carbon cycle. While data-driven models have shown strong performance in capturing GPP dynamics at specific sites, their ability to generalize across ecosystems without site-specific recalibration remains largely untested. This study addresses this gap by evaluating the applicability of XGBoost and LSTM models in estimating GPP across different European ecosystems. We developed a unified (cross-site) modeling framework that integrates in-situ eddy covariance observations and Sentinel-2–derived vegetation indices using incremental learning. Models’ performance was assessed via: (i) site-specific models, developed to capture individual site characteristics, and (ii) cross-site generalization, including evaluation on an independent dataset of unseen ecosystems. SARIMAX is included as a site-specific statistical benchmark for comparison. Our findings indicate that XGBoost consistently outperformed the other models, achieving site-specific R2 values above 0.90 in forest and grassland ecosystems and an average R2 of 0.72 across unseen sites (range 0.66–0.78). LSTM exhibited better accuracy in predicting GPP peaks at site-specific level, particularly in cropland and forest ecosystems. At site-level, SARIMAX showed comparable performance to XGBoost but struggled in capturing the rapid temporal variation of GPP. These findings demonstrate the feasibility of a data-driven framework for cross-site GPP monitoring within European flux-tower networks, making a first step toward transferable GPP prediction without site-specific recalibration.
Journal article
(2024)
-
A. de Fockert, M. A. Eleveld, W. Bakker, J. M. Felício, T. S. Costa, M. Vala, P. Marques, R. Calvert, T. S. van den Bremer, More authors...
Remote sensing technologies have the potential to support monitoring of floating plastic litter in aquatic environments. An experimental campaign was carried out in a large-scale hydrodynamic test facility to explore the detectability of floating plastics in ocean waves, comparing and contrasting different microwave and optical remote sensing technologies. The extensive experiments revealed that detection of plastics was feasible with microwave measurement techniques using X and Ku-bands with VV polarization at a plastic threshold concentration of 1 item/m2 or 1–10 g/m2. The optical measurements further revealed that spectral and polarization properties in the visible and infrared spectrum had diagnostic information unique to the floating plastics. This assessment presents a crucial step towards enabling the detection of aquatic plastics using advanced remote sensing technologies. We demonstrate that remote sensing has the potential for global targeting of plastic litter hotspots, which is needed for supporting effective clean-up efforts and scientific evidence-based policy making.
...
Remote sensing technologies have the potential to support monitoring of floating plastic litter in aquatic environments. An experimental campaign was carried out in a large-scale hydrodynamic test facility to explore the detectability of floating plastics in ocean waves, comparing and contrasting different microwave and optical remote sensing technologies. The extensive experiments revealed that detection of plastics was feasible with microwave measurement techniques using X and Ku-bands with VV polarization at a plastic threshold concentration of 1 item/m2 or 1–10 g/m2. The optical measurements further revealed that spectral and polarization properties in the visible and infrared spectrum had diagnostic information unique to the floating plastics. This assessment presents a crucial step towards enabling the detection of aquatic plastics using advanced remote sensing technologies. We demonstrate that remote sensing has the potential for global targeting of plastic litter hotspots, which is needed for supporting effective clean-up efforts and scientific evidence-based policy making.
Digital Elevation Models (DEMs) are a necessity for modelling many large-scale environmental processes. In this study, we investigate the potential of data from two spaceborne lidar altimetry missions, ICESat-2 and GEDI—with respect to their vertical accuracies and planimetric data collection patterns—as sources for rasterisation towards creating global DEMs. We validate the terrain measurements of both missions against airborne lidar datasets over three areas in the Netherlands, Switzerland, and New Zealand and differentiate them using land-cover classes. For our experiments, we use five years of ICESat-2 ATL03 data and four years of GEDI L2A data for a total of 252 million measurements. The datasets are filtered using parameter flags provided by the higher-level products ICESat-2 ATL08 and GEDI L3A. For all areas and land-cover classes combined, ICESat-2 achieves a bias of −0.11 m, an MAE of 0.43 m, and an RMSE of 0.93 m. From our experiments, we find that GEDI is less accurate, with a bias of 0.09 m, an MAE of 0.98 m, and an RMSE of 2.96 m. Measurements in open land-cover classes, such as “Cropland” and “Grassland”, result in the best accuracy for both missions. We also find that the slope of the terrain has a major influence on vertical accuracy, more so for GEDI than ICESat-2 because of its larger horizontal geolocation error. In contrast, we find little effect of either beam power or background solar radiation, nor do we find noticeable seasonal effects on accuracy. Furthermore, we investigate the spatial coverage of ICESat-2 and GEDI by deriving a DEM at different horizontal resolutions and latitudes. GEDI has higher spatial coverage than ICESat-2 at lower latitudes due to its beam pattern and lower inclination angle, and a derived DEM can achieve a resolution of 500 m. ICESat-2 only reaches a DEM resolution of 700 m at the equator, but it increases to almost 200 m at higher latitudes. When combined, a 500 m resolution lidar-based DEM can be achieved globally. Our results indicate that both ICESat-2 and GEDI enable accurate terrain measurements anywhere in the world. Especially in data-poor areas—such as the tropics—this has potential for new applications and insights.
...
Digital Elevation Models (DEMs) are a necessity for modelling many large-scale environmental processes. In this study, we investigate the potential of data from two spaceborne lidar altimetry missions, ICESat-2 and GEDI—with respect to their vertical accuracies and planimetric data collection patterns—as sources for rasterisation towards creating global DEMs. We validate the terrain measurements of both missions against airborne lidar datasets over three areas in the Netherlands, Switzerland, and New Zealand and differentiate them using land-cover classes. For our experiments, we use five years of ICESat-2 ATL03 data and four years of GEDI L2A data for a total of 252 million measurements. The datasets are filtered using parameter flags provided by the higher-level products ICESat-2 ATL08 and GEDI L3A. For all areas and land-cover classes combined, ICESat-2 achieves a bias of −0.11 m, an MAE of 0.43 m, and an RMSE of 0.93 m. From our experiments, we find that GEDI is less accurate, with a bias of 0.09 m, an MAE of 0.98 m, and an RMSE of 2.96 m. Measurements in open land-cover classes, such as “Cropland” and “Grassland”, result in the best accuracy for both missions. We also find that the slope of the terrain has a major influence on vertical accuracy, more so for GEDI than ICESat-2 because of its larger horizontal geolocation error. In contrast, we find little effect of either beam power or background solar radiation, nor do we find noticeable seasonal effects on accuracy. Furthermore, we investigate the spatial coverage of ICESat-2 and GEDI by deriving a DEM at different horizontal resolutions and latitudes. GEDI has higher spatial coverage than ICESat-2 at lower latitudes due to its beam pattern and lower inclination angle, and a derived DEM can achieve a resolution of 500 m. ICESat-2 only reaches a DEM resolution of 700 m at the equator, but it increases to almost 200 m at higher latitudes. When combined, a 500 m resolution lidar-based DEM can be achieved globally. Our results indicate that both ICESat-2 and GEDI enable accurate terrain measurements anywhere in the world. Especially in data-poor areas—such as the tropics—this has potential for new applications and insights.
DeltaDTM
A global coastal digital terrain model
Journal article
(2024)
-
Maarten Pronk, Aljosja Hooijer, Dirk Eilander, Arjen Haag, Tjalling de Jong, Michalis Vousdoukas, Ronald Vernimmen, Hugo Ledoux, Marieke Eleveld
Coastal elevation data are essential for a wide variety of applications, such as coastal management, flood modelling, and adaptation planning. Low-lying coastal areas (found below 10 m +Mean Sea Level (MSL)) are at risk of future extreme water levels, subsidence and changing extreme weather patterns. However, current freely available elevation datasets are not sufficiently accurate to model these risks. We present DeltaDTM, a global coastal Digital Terrain Model (DTM) available in the public domain, with a horizontal spatial resolution of 1 arcsecond (∼30 m) and a vertical mean absolute error (MAE) of 0.45 m overall. DeltaDTM corrects CopernicusDEM with spaceborne lidar from the ICESat-2 and GEDI missions. Specifically, we correct the elevation bias in CopernicusDEM, apply filters to remove non-terrain cells, and fill the gaps using interpolation. Notably, our classification approach produces more accurate results than regression methods recently used by others to correct DEMs, that achieve an overall MAE of 0.72 m at best. We conclude that DeltaDTM will be a valuable resource for coastal flood impact modelling and other applications.
...
Coastal elevation data are essential for a wide variety of applications, such as coastal management, flood modelling, and adaptation planning. Low-lying coastal areas (found below 10 m +Mean Sea Level (MSL)) are at risk of future extreme water levels, subsidence and changing extreme weather patterns. However, current freely available elevation datasets are not sufficiently accurate to model these risks. We present DeltaDTM, a global coastal Digital Terrain Model (DTM) available in the public domain, with a horizontal spatial resolution of 1 arcsecond (∼30 m) and a vertical mean absolute error (MAE) of 0.45 m overall. DeltaDTM corrects CopernicusDEM with spaceborne lidar from the ICESat-2 and GEDI missions. Specifically, we correct the elevation bias in CopernicusDEM, apply filters to remove non-terrain cells, and fill the gaps using interpolation. Notably, our classification approach produces more accurate results than regression methods recently used by others to correct DEMs, that achieve an overall MAE of 0.72 m at best. We conclude that DeltaDTM will be a valuable resource for coastal flood impact modelling and other applications.
Conference paper
(2024)
-
Anna Spinosa, Marieke Eleveld, Ulf Mallast, Johannes Peterseil, Valeria Mobilia, Karisma Karisma, Mario Alberto Fuentes-Monjaraz, Ghada El Serafy
This study addresses the challenges posed by climatic changes and biodiversity loss to ecosystem stability, by quantifying gross primary production (GPP) changes. An improved earth observation product is obtained by integrating in-situ and remote sensing data via data-driven models. Employing a user-centered strategy, our methodology builds on users' engagement, ensuring both the identification of user needs and practical product demonstrations. With GEOSS as central and integrated stakeholder, we strive for a broad interoperability and accessibility of generated outcomes. The project outcomes include a curated dataset with FAIR metadata, openly available code, and reports for reproducibility, contributing to the broader Earth Intelligence supply chain.
...
This study addresses the challenges posed by climatic changes and biodiversity loss to ecosystem stability, by quantifying gross primary production (GPP) changes. An improved earth observation product is obtained by integrating in-situ and remote sensing data via data-driven models. Employing a user-centered strategy, our methodology builds on users' engagement, ensuring both the identification of user needs and practical product demonstrations. With GEOSS as central and integrated stakeholder, we strive for a broad interoperability and accessibility of generated outcomes. The project outcomes include a curated dataset with FAIR metadata, openly available code, and reports for reproducibility, contributing to the broader Earth Intelligence supply chain.
River ice is a major contributor to flood risk in cold regions due to the physical impediment of flow caused by ice jamming. Although a variety of classifiers have been developed to distinguish ice types using HH or VV intensity of SAR data, mostly based on data from RADARSAT-1 and -2, these classifiers still experience problems with breakup classification, because meltwater development causes overlap in co-polarization backscatter intensities of open water and sheet ice pixels. In this study, we develop a Random Forest classifier based on multiple features of Sentinel-1 data for three main classes generally present during breakup: rubble ice, sheet ice and open water, in a case study over the Athabasca River in Canada. For each ice stage, intensity of the VV and VH backscatter, pseudo-polarimetric decomposition parameters and Grey Level Co-occurrence Matrix texture features were computed for 70 verified sample areas. Several classifiers were developed, based on i) solely intensity features or on ii) a combination of intensity, pseudo-polarimetric and texture features and each classifier was evaluated based on Recursive Feature Elimination with Cross-Validation and pair-wise correlation of the studied features. Results show improved classifier performance when including GLCM mean of VV intensity, and VH intensity features instead of the conventional classifier based solely on intensity. This highlights the complementary nature of texture and intensity for the classification of breaking river ice. GLCM mean incorporates spatial patterns of the co-polarized intensity and sensitivity to context, while VH intensity introduces cross-polarized surface and volume scattering signals and is less sensitive to wind than the commonly used co-polarized intensity. We conclude that the proposed method based on the combination of texture and intensity features is suitable for and performs well in physically complex situations such as breakup, which are hard to classify otherwise. This method has a high potential for classifying river ice operationally, also for data from other SAR missions. Since it is a generic approach, it also has potential to classify river ice along other rivers globally.
...
River ice is a major contributor to flood risk in cold regions due to the physical impediment of flow caused by ice jamming. Although a variety of classifiers have been developed to distinguish ice types using HH or VV intensity of SAR data, mostly based on data from RADARSAT-1 and -2, these classifiers still experience problems with breakup classification, because meltwater development causes overlap in co-polarization backscatter intensities of open water and sheet ice pixels. In this study, we develop a Random Forest classifier based on multiple features of Sentinel-1 data for three main classes generally present during breakup: rubble ice, sheet ice and open water, in a case study over the Athabasca River in Canada. For each ice stage, intensity of the VV and VH backscatter, pseudo-polarimetric decomposition parameters and Grey Level Co-occurrence Matrix texture features were computed for 70 verified sample areas. Several classifiers were developed, based on i) solely intensity features or on ii) a combination of intensity, pseudo-polarimetric and texture features and each classifier was evaluated based on Recursive Feature Elimination with Cross-Validation and pair-wise correlation of the studied features. Results show improved classifier performance when including GLCM mean of VV intensity, and VH intensity features instead of the conventional classifier based solely on intensity. This highlights the complementary nature of texture and intensity for the classification of breaking river ice. GLCM mean incorporates spatial patterns of the co-polarized intensity and sensitivity to context, while VH intensity introduces cross-polarized surface and volume scattering signals and is less sensitive to wind than the commonly used co-polarized intensity. We conclude that the proposed method based on the combination of texture and intensity features is suitable for and performs well in physically complex situations such as breakup, which are hard to classify otherwise. This method has a high potential for classifying river ice operationally, also for data from other SAR missions. Since it is a generic approach, it also has potential to classify river ice along other rivers globally.