Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 14
Encontrada(s) 1 página(s)

e-Archivo. Repositorio Institucional de la Universidad Carlos III de Madrid
Publication Artículo científico (JournalArticle)

Opening the Black Box: A systematic review on explainable artificial intelligence in remote sensing

  • Höhl, Adrian
  • Obadic, Ivica
  • Fernández Torres, Miguel Ángel
  • Najjar, Hiba
  • Borges Oliveira, Darío Augusto
  • Akata, Zeynep
  • Dengel, Andreas
  • Zhu, Xiao Xiang
In recent years, black-box machine learning approaches have become a dominant modeling paradigm for knowledge extraction in remote sensing. Despite the potential benefits of uncovering the inner workings of these models with explainable AI, a comprehensive overview summarizing the explainable AI methods used and their objectives, findings, and challenges in remote sensing applications is still missing. In this paper, we address this gap by performing a systematic review to identify the key trends in the field and shed light on novel explainable AI approaches and emerging directions that tackle specific remote sensing challenges. We also reveal the common patterns of explanation interpretation, discuss the extracted scientific insights, and reflect on the approaches used for the evaluation of explainable AI methods. As such, our review provides a complete summary of the state-of-the-art of explainable AI in remote sensing. Further, we give a detailed outlook on the challenges and promising research directions, representing a basis for novel methodological development and a useful starting point for new researchers in the field., The work of A. Höhl was funded by the project ML4Earth by the German Federal Ministry for Economic Affairs and Climate Action under Grant 50EE2201C. The work of I. Obadic is funded by the Munich Center for Machine Learning. M.Á. Fernández-Torres acknowledges the support from the European Research Council (ERC) under the ERC Synergy Grant USMILE (Grant Agreement 855187) and the EU’s Horizon 2020 research and innovation program within the project “XAIDA: Extreme Events — Artificial Intelligence for Detection and Attribution,” under Grant Agreement 101003469. The work of M.Á. Fernández-Torres and X. Zhu is supported by the German Federal Ministry of Education and Research in the framework of the international future AI lab “AI4EO – Artificial Intelligence for Earth Observation: Reasoning, Uncertainties, Ethics and Beyond” (Grant 01DD20001). H. Najjar acknowledges support through a scholarship from the University of Kaiserslautern-Landau.
e-Archivo. Repositorio Institucional de la Universidad Carlos III de Madrid
Publication Artículo científico (JournalArticle)

Deep Learning with Noisy Labels for Spatio-Temporal Drought Detection

  • Cortés Andrés, Jordi
  • Fernández Torres, Miguel Ángel
  • Camps Valls, Gustau
Droughts pose significant challenges for accurate monitoring due to their complex spatiotemporal characteristics. Data-driven machine learning (ML) models have shown promise in detecting extreme events when enough well-annotated data is available. However, droughts do not have a unique and precise definition, which leads to noise in human-annotated events and presents an imperfect learning scenario for deep learning models. This article introduces a 3-D convolutional neural network (CNN) designed to address the complex task of drought detection, considering spatiotemporal dependencies and learning with noisy and inaccurate labels. Motivated by the shortcomings of traditional drought indices, we leverage supervised learning with labeled events from multiple sources, capturing the shared conceptual space among diverse definitions of drought. In addition, we employ several strategies to mitigate the negative effect of noisy labels (NLs) during training, including a novel label correction (LC) method that relies on model outputs, enhancing the robustness and performance of the detection model. Our model significantly outperforms state-of-the-art drought indices when detecting events in Europe between 2003 and 2015, achieving an AUROC of 72.28%, an AUPRC of 7.67%, and an ECE of 16.20%. When applying the proposed LC method, these performances improve by +5%, +15%, and +59%, respectively. Both the proposed model and the robust learning methodology aim to advance drought detection by providing a comprehensive solution to label noise and conceptual variability., This work was supported in part by the European Research Council (ERC) under the ERC Synergy Grant USMILE (grant agreement 855187), the European Union’s Horizon 2020 research and innovation program within the projects ‘XAIDA: Extreme Events - Artificial Intelligence for Detection and Attribution,’ (grant agreement 101003469), the ESA AI4Science project ‘Multi-Hazards, Compounds and Cascade events: DeepExtremes,’ 2022-2024, the computer resources provided by the Jülich Supercomputing Centre (JSC) (Project No. PRACE-DEV-2022D01-048), the computer resources provided by Artemisa (funded by the European Union ERDF and Comunitat Valenciana), as well as the technical support provided by the Instituto de Física Corpuscular, IFIC (CSIC-UV).
e-Archivo. Repositorio Institucional de la Universidad Carlos III de Madrid
Publication Artículo científico (JournalArticle)

Calibration and uncertainty quantification for deep learning-based drought detection

  • Zhang, Mengxue
  • Fernández Torres, Miguel Ángel
  • Cohrs, Kai-Hendric
  • Camps Valls, Gustau
Droughts are hydrometeorological extreme events influenced by highly intricate land–atmosphere feedback mechanisms and climate variability. Deep learning models have recently succeeded in detecting extreme climate events and promise to uncover and understand droughts further. There are two main challenges of reliability and trustworthiness limiting their applications: miscalibration and inherent uncertainty. However, they remain rarely explored because deep learning models are overparameterized and seldom tractable. To address this shortcoming, we introduce methodologies for model calibration and entropy-based uncertainty quantification for deep learning-based drought detection. The calibration algorithm can deal with calibration errors by reducing distributional shifts and alleviating overconfident predictions. The uncertainty framework, in turn, decomposes and quantifies the total uncertainty according to several components: data uncertainty, procedural variability, parametric variability, and latent variability. Thus, our method identifies uncertain predictions and supports robust evaluations, benefiting the credibility of the decision-making process. Empirical evidence of performance in a wide range of European drought events is given, justifying the effectiveness of our approach. The calibration methodology yields the lowest expected calibration error (0.31%) and the precision of the uncertainty-based decision-making is improved from 72.27% to 74.06% and 76.59%, based on ensemble predictions and rejecting the predictions for the top 20% uncertain negative samples, respectively. In summary, our approach significantly enhances drought detection’s reliability and classification accuracy, constituting a key step toward more trustworthy and actionable climate decision-making., M Z appreciates the financial support from the China Scholarship Council (CSC) through the State Scholarship Fund for Overseas Study (No. 202106710031). All authors acknowledge the support from the European Research Council (ERC) under the ERC Synergy Grant USMILE (grant agreement 855187), the European Union‘s Horizon 2020 research and innovation program within the project ‘XAIDA: Extreme Events - Artificial Intelligence for Detection and Attribution,’ (GA 101003469), ‘AI4PEX: Artificial Intelligence and Machine Learning for Enhanced Representation of Processes and Extremes in Earth System Models’ (GA 101137682) and the computer resources provided by the Jülich Supercomputing Centre (JSC) (Project No. PRACE-DEV-2022D01-048), the computer resources provided by Artemisa (funded by the European Union ERDF and Comunitat Valenciana), as well as the technical support provided by the Instituto de Física Corpuscular, IFIC (CSIC-UV) .
e-Archivo. Repositorio Institucional de la Universidad Carlos III de Madrid
Publication Artículo científico (JournalArticle)

The AIDE Toolbox: Artificial intelligence for disentangling extreme events

  • Gonzalez Calabuig, Maria
  • Cortés Andrés, Jordi
  • Williams, Tristan Keith Ellis
  • Zhang, Mengxue
  • Pellicer Valero, Oscar José
  • Fernández Torres, Miguel Ángel
  • Camps-Valls, Gustavo
We introduce the Artificial Intelligence for Disentangling Extremes (AIDE) toolbox that allows for anomaly detection, extreme event analysis, and impact assessment in remote sensing and geoscience applications. AIDE integrates advanced machine learning (ML) models, ranging in complexity, assumptions, and sophistication, and can yield spatiotemporal explicit monitoring maps with probabilistic estimates. Supervised and unsupervised algorithms, deterministic and probabilistic, convolutional and recurrent neural networks (CNNs and RNNs), as well as methods based on density estimation, are covered by this framework. The open source toolbox is intended for scientists, engineers, and students with basic knowledge of remote sensing and geosciences working in anomaly detection, deep learning (DL), and explainable AI (XAI), and is available at https://github.com/IPL-UV/AIDE., This work has received funding from the European Union's Horizon 2020 Research and Innovation Projects "X AIDA: Extreme Events-Artificial Intelligence for Detection and Attribution" (Grant Agreement 101003469) and "DeepCube: Explainable AI Pipelines for Big Copernicus Data" (Grant Agreement 101004188). G.C.-V. was supported by the European Research Council (ERC) under the ERC Synergy Grant USMILE (Grant Agreement 855187). The authors thank the support of the European Space Agency through the project "DeepExtremes: Extreme Events, Multi-Hazards, and Com-pound Events," and the computer resources provided by Artemisa, funded by the European Union European Regional Development Fund and Comunitat Valenciana, as well as the technical support provided by the Instituto de Fisica Corpuscular. Maria Gonzalez-Calabuig and Jordi Cortes-Andres are co-first authors.
e-Archivo. Repositorio Institucional de la Universidad Carlos III de Madrid
Publication Artículo científico (JournalArticle)

Domain knowledge-driven variational recurrent networks for drought monitoring

  • Zhang, Mengxue
  • Fernández Torres, Miguel Ángel
  • Camps Valls, Gustau
In the context of climate change, droughts, increasingly frequent and severe, necessitate effective monitoring. Existing methods, such as drought indices and data-driven models, face important limitations. Drought indices are built on prior expert knowledge but lack calibration based on actual drought events, while data-driven models prioritize goodness of fit over real event identification, undermining their credibility and generalization, and also struggling to generalize from regional to large-scale contexts. To address these challenges, here we introduce a hybrid machine learning framework for time series that combines domain knowledge and observational data in a variational recurrent neural network. The network models the joint distribution of total precipitation, air temperature, and real drought events, providing accurate predictions and uncertainty estimates. Extensive experiments focusing on a wide range of European drought events from 2011 to 2018 consistently show that our hybrid model surpasses both drought indices and data-driven models in terms of accuracy in drought detection, underlining its effectiveness, robustness, and stability. Our model achieves the best ROC-AUC (%) results in Afghanistan (79.7 ± 0.5), Italy (84.3 ± 0.6), Russia (89.4 ± 0.2), Europe-0 (84.3 ± 0.1), and Europe-1 (82.8 ± 0.4), effectively capturing the starting and ending times of drought events with lower uncertainty, and also generalizing better for unseen locations., M Z appreciates the financial support from the China Scholarship Council (CSC) through the State Scholarship Fund for Overseas Study (No. 202106710031). All authors acknowledge the support from the European Research Council (ERC) under the ERC Synergy Grant USMILE (grant agreement 855187), the European Union's Horizon 2020 Research and Innovation Program within the project ‘XAIDA: Extreme Events - Artificial Intelligence for Detection and Attribution,’ under grant agreement No 101003469, the computer resources provided by the Jülich Supercomputing Centre (JSC) (Project No. PRACE-DEV-2022D01-048), the computer resources provided by Artemisa (funded by the European Union ERDF and Comunitat Valenciana), as well as the technical support provided by the Instituto de Física Corpuscular, IFIC (CSIC-UV).
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Within-season crop monitoring at continental scale utilizing new gap-filled Landsat temporal series

  • Rajadel-Lambistos, C.
  • Moreno-Martínez, A.
  • Maneta, M. P.
  • Beguería, Santiago
  • Kimball, J. S.
  • Clinton, N.
  • Atzberger, C.
  • Camps-Valls, G. S.W.
  • Running, S. W.
18 Pags.- 12 Figs.- 5 Tabls. Data availability: The JavaScript code and datasets (including training and test datasets) used in this work are available to the interested reader in the GEE repository: https://code.earthengine.google.com/?accept_repo=users/RemoteSensing_master/prompt_classification (see ‘Demo’ script) © 2024 The Author(s). This is an Open Access article distributed under the terms of the Creative Commons Attribution License., Timely and accurate crop acreage information is essential for food security and the informed decision-making by governmental bodies and stakeholders in the agro-economic system. Surveys and fieldwork are expensive and time consuming, and the information is usually only released after the cropping season. Remote sensing technology is inexpensive, scales well for large areas, and offers the opportunity to gather within-season information. However, environmental and technical constraints complicate model applications to larger scales. For example, clouds and snow introduce data gaps, and the weather-driven crop development creates highly variable spectral-temporal signatures. This study determines the earliest possible month that permits the accurate classification of the major crops in the CONUS using gap-filled data, transfer learning, and cloud computing technology. The resulting within-season, annual crop classifications are generated with precision rates of up to 80%, presented as annual crop maps and county-level area statistics. The crop classification products presented in this work are unique, in that they were created with shorter records for the first time, using data within a concise time-range (i.e. a few months) and at a 30-meter spatial resolution, while covering continental scale. This novel prospect allows for potential new applications in forecasting regional crop productions., A. Moreno-Martinez and G. Camps-Valls were supported by the European Research Council under the ERC-SyG-2019 USMILE project (grant agreement 855187)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Within-season crop monitoring at continental scale utilizing new gap-filled Landsat temporal series

  • Rajadel-Lambistos, C.
  • Izquierdo-Verdiguier, E.
  • Moreno-Martínez, A.
  • Maneta, M. P.
  • Beguería, Santiago
  • Kimball, J. S.
  • Clinton, N.
  • Atzberger, C.
  • Camps-Valls, G.
  • Running, S. W.
18 Pags.- 13 Figs.- 6 Tabls. The JavaScript code and datasets (including training and test datasets) used in this work are available to the interested reader in the GEE repository: https://code.earthengine.google.com/?accept_repo=users/RemoteSensing_master/prompt_classification (see ‘Demo’ script). © 2024 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution License., Timely and accurate crop acreage information is essential for food security and the informed decision-making by governmental bodies and stakeholders in the agro-economic system. Surveys and fieldwork are expensive and time consuming, and the information is usually only released after the cropping season. Remote sensing technology is inexpensive, scales well for large areas, and offers the opportunity to gather within-season information. However, environmental and technical constraints complicate model applications to larger scales. For example, clouds and snow introduce data gaps, and the weather-driven crop development creates highly variable spectral-temporal signatures. This study determines the earliest possible month that permits the accurate classification of the major crops in the CONUS using gap-filled data, transfer learning, and cloud computing technology. The resulting within-season, annual crop classifications are generated with precision rates of up to 80%, presented as annual crop maps and county-level area statistics. The crop classification products presented in this work are unique, in that they were created with shorter records for the first time, using data within a concise time-range (i.e. a few months) and at a 30-meter spatial resolution, while covering continental scale. This novel prospect allows for potential new applications in forecasting regional crop productions., A. Moreno-Martinez and G. Camps-Valls were supported by the European Research Council under the ERC-SyG-2019 USMILE project (grant agreement 855187)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Satellite remote sensing reveals the footprint of biodiversity on multiple ecosystem functions across the NEON eddy covariance network

  • Gomarasca, Ulisse
  • Duveiller, Gregory
  • Pacheco-Labrador, Javier
  • Ceccherini, Guido
  • Cescatti, Alessandro
  • Girardello, Marco
  • Nelson, Jacob A.
  • Reichstein, Markus
  • Wirth, Christian
  • Migliavacca, Mirco
Biodiversity relates to ecosystem functioning by modulating biogeochemical cycles of carbon, water, energy, and nutrients within and between multiple biotic and abiotic components of the ecosystems. However, large-scale, systematic measurements of plant biodiversity are still lacking, and the effects of biodiversity on measured biogeochemical processes are understudied. Here, we combined alpha (α) and beta (β) taxonomic measurements, spectral diversity from satellite observations, structural properties of the vegetation, and climatic drivers to assess the effect of biodiversity on ecosystem functional properties. Ecosystem functional properties were computed from eddy-covariance fluxes at 44 sites of the National Ecological Observatory Network. Based on the spectral variation hypothesis, we used the near-infrared reflectance of vegetation (NIRv) derived from Sentinel-2 satellite imagery to compute Rao's quadratic entropy (Rao Q), a distance metric related to spatial heterogeneity. Using an automatic model averaging technique, we found that biodiversity proxies hold substantial explanatory power when predicting several ecosystem functions related to carbon and water exchange. In particular, NIRv-based Rao Q (RaoQNIRv) reflected positive biodiversity effects on productivity, as expected from the literature. In contrast, traditional taxonomic α-diversity indices were generally not selected as relevant predictors of the ecosystem functional properties. Yet, β-diversity strongly contributed to the prediction of carbon use efficiency, surface conductance, and water use efficiency. We also found that the RaoQNIRv is less affected by issues of saturation and bare soil contribution compared to RaoQNDVI. We show that spectral heterogeneity based on remotely sensed NIRv holds the potential for globally characterizing the biodiversity-ecosystem functioning relationship (BEF). While systematic measurements of taxonomic diversity co-located at biogeochemical measurement stations could reduce the uncertainty surrounding the BEF relationship at whole-ecosystem scale, remotely- sensed metrics characterizing important functional and structural diversity aspects of the landscape will be crucial for continuous spatiotemporal monitoring of biodiversity with relevant implications for ecosystem services to humankind., JPL activity was carried out under the ESA Living Planet Fellowship IRS4BEF ‘Integrated Remote Sensing for Biodiversity-Ecosystem Function’ (C.N.4000140028/22/I-DT-lr), a program of and funded by the European Space Agency”. JRC activities were carried out under the exploratory project ‘Assessing Forest Biodiversity and Resilience from Space, ForBioRes’. MR and GD acknowledge funding by the European Research Council (ERC) Synergy Grant ‘Understanding and modelling the Earth System with Machine Learning (USMILE)’ under the Horizon 2020 research and innovation programme (Grant Agreement No. 855187), Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

BOSSE v1.0: the Biodiversity Observing System Simulation Experiment

  • Pacheco-Labrador, Javier
  • Gomarasca, Ulisse
  • Migliavacca, Mirco
  • Jung, Martin
  • Duveiller, Gregory
  • Pabon-Moreno, Daniel E.
  • Li, Wantong
As global and regional vegetation diversity loss threatens essential ecosystem services under climate change, monitoring biodiversity dynamics and its role in ecosystem services is crucial in predicting future states and providing insights into climate adaptation and mitigation. In this context, remote sensing (RS) offers a unique opportunity to assess long-term and large-scale biodiversity dynamics. However, the development of this capability suffers from the lack of consistent, global, and spatially matched ground diversity measurements that enable testing and validating generalizable methodologies. The Biodiversity Observing System Simulation Experiment (BOSSE) aims to alleviate the lack of this information by means of simulation. BOSSE simulates synthetic landscapes featuring communities of various vegetation species whose traits's seasonality and ecosystem functions (e.g., biospheric fluxes) respond to meteorology and environmental factors. Simultaneously, BOSSE can generate various types of remote sensing imagery linked to the traits and functions via radiative transfer theory. Specifically, it simulates hyperspectral reflectance factors (R), which can be convolved to the bands of specific RS missions, sun-induced chlorophyll fluorescence (SIF), and land surface temperature (LST). The resolution of the RS imagery can be degraded to test the robustness of different approaches to information loss and the capability of new methodologies to overcome this limitation. Therefore, BOSSE enables users to evaluate the capability of different methods to estimate plant functional diversity (PFD) from RS and link it to ecosystem functions. We expect BOSSE to support the benchmarking and improvement of old and novel methods dedicated to estimating plant diversity and exploring the biodiversity–ecosystem function (BEF) relationships, facilitating advances in this growing area of research and supporting the analysis and interpretation of real-world measurements. We also expect BOSSE to be extended and include new features that provide more realistic simulations that help answer more complex questions related to climate change and global warming., This research has been supported by the Living Planet Fellowship “Integrated Remote Sensing for Biodiversity-Ecosystems Function” IRS4BEF (ESA Contract no. 4000140028/22/I-DT-lr) of the European Space Agency, the “Integrated Observing Systems and Simulation Experiments to Analyze Biodiversity-Ecosystem Function Relationships in Savanna Ecosystems” research project (PID2023-151046NB-I00 funded by MCIU/AEI/10.13039/501100011033/FEDER, UE) funded by Ministerio de Ciencia, Innovación y Universidades, and the European Research Council, and the European Research Council (ERC) Synergy Grant “Understanding and modeling the Earth System with Machine Learning (USMILE)” under the EU Horizon 2020 research and innovation program (grant agreement no. 855187), Peer reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Evaluation of native Earth system model output with ESMValTool v2.6.0

  • Schlund, Manuel
  • Hassler, Birgit
  • Lauer, Axel
  • Andela, Bouwe
  • Jöckel, Patrick
  • Loosveldt Tomas, Saskia
  • Vegas Regidor, Javier
Earth system models (ESMs) are state-of-the-art climate models that allow numerical simulations of the past, present-day, and future climate. To extend our understanding of the Earth system and improve climate change projections, the complexity of ESMs heavily increased over the last decades. As a consequence, the amount and volume of data provided by ESMs has increased considerably. Innovative tools for a comprehensive model evaluation and analysis are required to assess the performance of these increasingly complex ESMs against observations or reanalyses. One of these tools is the Earth System Model Evaluation Tool (ESMValTool), a community diagnostic and performance metrics tool for the evaluation of ESMs. Input data for ESMValTool needs to be formatted according to the CMOR (Climate Model Output Rewriter) standard, a process that is usually referred to as “CMORization”. While this is a quasi-standard for large model intercomparison projects like the Coupled Model Intercomparison Project (CMIP), this complicates the application of ESMValTool to non-CMOR-compliant climate model output. In this paper, we describe an extension of ESMValTool introduced in v2.6.0 that allows seamless reading and processing of “native” climate model output, i.e., operational output produced by running the climate model through the standard workflow of the corresponding modeling institute. This is achieved by an extension of ESMValTool's preprocessing pipeline that performs a CMOR-like reformatting of the native model output during runtime. Thus, the rich collection of diagnostics provided by ESMValTool is now fully available for these models. For models that use unstructured grids, a further preprocessing step required to apply many common diagnostics is regridding to a regular latitude–longitude grid. Extensions to ESMValTool's regridding functions described here allow for more flexible interpolation schemes that can be used on unstructured grids. Currently, ESMValTool supports nearest-neighbor, bilinear, and first-order conservative regridding from unstructured grids to regular grids. Example applications of this new native model support are the evaluation of new model setups against predecessor versions, assessing of the performance of different simulations against observations, CMORization of native model data for contributions to model intercomparison projects, and monitoring of running climate model simulations. For the latter, new general-purpose diagnostics have been added to ESMValTool that are able to plot a wide range of variable types. Currently, five climate models are supported: CESM2 (experimental; at the moment, only surface variables are available), EC-Earth3, EMAC, ICON, and IPSL-CM6. As the framework for the CMOR-like reformatting of native model output described here is implemented in a general way, support for other climate models can be easily added., The development of ESMValTool is supported by the projects “Climate-Carbon Interactions in the Current Century” (4C; grant agreement no. 821003), “Earth System Models for the Future” (ESM2025; grant agreement no. 101003536), “Infrastructure for the European Network for Earth System Modelling - Phase 3” (IS-ENES3; grant agreement no. 824084), and the European Research Council (ERC) Synergy Grant “Understanding and Modelling the Earth System with Machine Learning” (USMILE; grant agreement no. 855187) under the European Union's Horizon 2020 Research and Innovation Programme. This study is a contribution to the project S1 of the Collaborative Research Centre TRR 181 “Energy Transfers in Atmosphere and Ocean” funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) (project no. 274762653). Brian Medeiros acknowledges support by the U.S. Department of Energy (award no. DE-SC0022070), the National Science Foundation (NSF) (IA 1947282), the National Center for Atmospheric Research, which is a major facility sponsored by the NSF (cooperative agreement no. 1852977), and the National Oceanic and Atmospheric Administration (award no. NA20OAR4310392). Tobias Stacke acknowledges funding support from the European Research Council (ERC) under the European Union's Horizon 2020 Programme (grant agreement no. 951288). This work used resources of the Deutsches Klimarechenzentrum (DKRZ) granted by its Scientific Steering Committee (WLA) under the project IDs bd0854, bd1179, and id0853. We acknowledge the World Climate Research Programme (WCRP), which, through its Working Group on Coupled Modeling, coordinated and promoted CMIP6. We thank the climate modeling groups for producing and making their model output available, the Earth System Grid Federation (ESGF) for archiving the data and providing access, and the multiple funding agencies who support CMIP and ESGF. We would like to thank Mattia Righi (DLR) for providing helpful comments about the manuscript., Peer Reviewed, "Article signat per 15 autors/es:Manuel Schlund, Birgit Hassler, Axel Lauer, Bouwe Andela, Patrick Jöckel, Rémi Kazeroni, Saskia Loosveldt Tomas, Brian Medeiros, Valeriu Predoi, Stéphane Sénési, Jérôme Servonnat, Tobias Stacke, Javier Vegas-Regidor, Klaus Zimmermann, and Veronika Eyring"
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Evaluation of native Earth system model output with ESMValTool v2.6.0

  • Schlund, Manuel
  • Hassler, Birgit
  • Lauer, Axel
  • Andela, Bouwe
  • Jöckel, Patrick
  • Loosveldt Tomas, Saskia
  • Vegas Regidor, Javier
Earth system models (ESMs) are state-of-the-art climate models that allow numerical simulations of the past, present-day, and future climate. To extend our understanding of the Earth system and improve climate change projections, the complexity of ESMs heavily increased over the last decades. As a consequence, the amount and volume of data provided by ESMs has increased considerably. Innovative tools for a comprehensive model evaluation and analysis are required to assess the performance of these increasingly complex ESMs against observations or reanalyses. One of these tools is the Earth System Model Evaluation Tool (ESMValTool), a community diagnostic and performance metrics tool for the evaluation of ESMs. Input data for ESMValTool needs to be formatted according to the CMOR (Climate Model Output Rewriter) standard, a process that is usually referred to as “CMORization”. While this is a quasi-standard for large model intercomparison projects like the Coupled Model Intercomparison Project (CMIP), this complicates the application of ESMValTool to non-CMOR-compliant climate model output. In this paper, we describe an extension of ESMValTool introduced in v2.6.0 that allows seamless reading and processing of “native” climate model output, i.e., operational output produced by running the climate model through the standard workflow of the corresponding modeling institute. This is achieved by an extension of ESMValTool's preprocessing pipeline that performs a CMOR-like reformatting of the native model output during runtime. Thus, the rich collection of diagnostics provided by ESMValTool is now fully available for these models. For models that use unstructured grids, a further preprocessing step required to apply many common diagnostics is regridding to a regular latitude–longitude grid. Extensions to ESMValTool's regridding functions described here allow for more flexible interpolation schemes that can be used on unstructured grids. Currently, ESMValTool supports nearest-neighbor, bilinear, and first-order conservative regridding from unstructured grids to regular grids. Example applications of this new native model support are the evaluation of new model setups against predecessor versions, assessing of the performance of different simulations against observations, CMORization of native model data for contributions to model intercomparison projects, and monitoring of running climate model simulations. For the latter, new general-purpose diagnostics have been added to ESMValTool that are able to plot a wide range of variable types. Currently, five climate models are supported: CESM2 (experimental; at the moment, only surface variables are available), EC-Earth3, EMAC, ICON, and IPSL-CM6. As the framework for the CMOR-like reformatting of native model output described here is implemented in a general way, support for other climate models can be easily added., The development of ESMValTool is supported by the projects “Climate-Carbon Interactions in the Current Century” (4C; grant agreement no. 821003), “Earth System Models for the Future” (ESM2025; grant agreement no. 101003536), “Infrastructure for the European Network for Earth System Modelling - Phase 3” (IS-ENES3; grant agreement no. 824084), and the European Research Council (ERC) Synergy Grant “Understanding and Modelling the Earth System with Machine Learning” (USMILE; grant agreement no. 855187) under the European Union's Horizon 2020 Research and Innovation Programme. This study is a contribution to the project S1 of the Collaborative Research Centre TRR 181 “Energy Transfers in Atmosphere and Ocean” funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) (project no. 274762653). Brian Medeiros acknowledges support by the U.S. Department of Energy (award no. DE-SC0022070), the National Science Foundation (NSF) (IA 1947282), the National Center for Atmospheric Research, which is a major facility sponsored by the NSF (cooperative agreement no. 1852977), and the National Oceanic and Atmospheric Administration (award no. NA20OAR4310392). Tobias Stacke acknowledges funding support from the European Research Council (ERC) under the European Union's Horizon 2020 Programme (grant agreement no. 951288). This work used resources of the Deutsches Klimarechenzentrum (DKRZ) granted by its Scientific Steering Committee (WLA) under the project IDs bd0854, bd1179, and id0853. We acknowledge the World Climate Research Programme (WCRP), which, through its Working Group on Coupled Modeling, coordinated and promoted CMIP6. We thank the climate modeling groups for producing and making their model output available, the Earth System Grid Federation (ESGF) for archiving the data and providing access, and the multiple funding agencies who support CMIP and ESGF. We would like to thank Mattia Righi (DLR) for providing helpful comments about the manuscript., Peer Reviewed, "Article signat per 15 autors/es:Manuel Schlund, Birgit Hassler, Axel Lauer, Bouwe Andela, Patrick Jöckel, Rémi Kazeroni, Saskia Loosveldt Tomas, Brian Medeiros, Valeriu Predoi, Stéphane Sénési, Jérôme Servonnat, Tobias Stacke, Javier Vegas-Regidor, Klaus Zimmermann, and Veronika Eyring", Postprint (published version)
Dipòsit Digital de Documents de la UAB
Publication Artículo preliminar (WorkingPaper)

Intercomparison of global foliar trait maps reveals fundamental differences and limitations of upscaling approaches

  • Dechant, Benjamin|||0000-0001-5171-2364
  • Kattge, Jens|||0000-0002-1022-8469
  • Pavlick, Ryan
  • Schneider, Fabian D.|||0000-0003-1791-2041
  • Sabatini, Francesco Maria|||0000-0002-7202-7697
  • Moreno-Martinez, Alvaro
  • Butler, Ethan E.|||0000-0003-3482-1950
  • van Bodegom, Peter|||0000-0003-0771-4500
  • Vallicrosa Pou, Helena|||0000-0002-5860-3096
  • Kattenborn, Teja|||0000-0001-7381-3828
  • Boonman, Coline C. F.|||0000-0003-2417-1579
  • Madani, Nima
  • Wright, Ian J.|||0000-0001-8338-9143
  • Dong, Ning|||0000-0003-0793-8854
  • Feilhauer, Hannes
  • Peñuelas, Josep|||0000-0002-7215-0150
  • Sardans i Galobart, Jordi|||0000-0003-2478-0219
  • Aguirre-Gutiérrez, Jesús|||0000-0001-9190-3229
  • Reich, Peter|||0000-0003-4424-662X
  • Leitão, Pedro J.
  • Cavender-Bares, Jeannine|||0000-0003-3375-9630
  • Myers-Smith, I. H.|||0000-0002-8417-6112
  • Duran, Sandra M.|||0000-0003-2044-8139
  • Croft, Holly
  • Prentice, Iain Colin|||0000-0002-1296-6764
  • Huth, Andreas
  • Rebe, Karin
  • Zaehle, Sönke|||0000-0001-5602-7956
  • Simova, Irena
  • Díaz, Sandra|||0000-0003-0012-4612
  • Reichstein, Markus|||0000-0001-5736-1112
  • Schiller, Christopher
  • Bruelheide, Helge|||0000-0003-3135-0356
  • Mahecha, Miguel
  • Wirth, Christian|||0000-0003-2604-8056
  • Malhi, Yadvinder|||0000-0002-3503-4783
  • Townsend, Philip A.
Foliar traits such as specific leaf area (SLA), leaf nitrogen (N) and phosphorus (P) concentrations play an important role in plant economic strategies and ecosystem functioning. Various global maps of these foliar traits have been generated using statistical upscaling approaches based on in-situ trait observations. Here, we intercompare such global upscaled foliar trait maps at 0.5° spatial resolution (six maps for SLA, five for N, three for P), categorize the upscaling approaches used to generate them, and evaluate the maps with trait estimates from a global database of vegetation plots (sPlotOpen). We disentangled the contributions from different plant functional types (PFTs) to the upscaled maps and characterized the differences between two trait metrics: community weighted mean (CWM) and top-of-canopy weighted mean (TWM). We found that the global foliar trait maps of SLA and N differ drastically and fall into two groups that are almost uncorrelated (for P only maps from one group were available). The primary factor explaining the differences between these groups is the exclusive use of PFT information combined with remote sensing-derived land cover products in one group while the other group mostly relied on environmental predictors. The impact of using TWM or CWM on spatial patterns was considerably smaller than that of including PFT and land cover information. The maps that used PFT and land cover information exhibit considerable similarities in spatial patterns that are strongly driven by land cover. The maps not using PFTs show a lower level of similarity and tend to be strongly driven by individual environmental variables. Overall, the maps using PFT and land cover information better reproduce the between-PFT trait differences and trait distributions of the plot-level sPlotOpen data, while the two groups performed similarly in capturing within-PFT trait variation. Upscaled maps of both groups were moderately correlated to grid-cell-level sPlotOpen data (R = 0.2-0.6), with considerable differences between upscaling approaches and overall higher correlations for SLA and N. Our findings highlight the importance of explicitly accounting for within-grid-cell trait variation, which has important implications for applications using existing maps and future upscaling efforts.
Dipòsit Digital de Documents de la UAB
Publication Artículo científico (JournalArticle)

Leaf-level coordination principles propagate to the ecosystem scale

  • Gomarasca, Ulisse|||0000-0001-7389-7793
  • Migliavacca, Mirco|||0000-0003-3546-8407
  • Kattge, Jens|||0000-0002-1022-8469
  • Nelson, Jacob|||0000-0002-4663-2420
  • Niinemets, Ülo|||0000-0002-3078-2192
  • Wirth, Christian|||0000-0003-2604-8056
  • Cescatti, Alessandro|||0000-0002-2769-2591
  • Bahn, Michael|||0000-0001-7482-9776
  • Nair, Richard|||0000-0002-6293-3610
  • Acosta, Alicia T. R.|||0000-0001-6572-3187
  • Arain, M. Altaf|||0000-0002-1433-5173
  • Beloiu, Mirela|||0000-0002-3592-8170
  • Black, T. Andrew|||0000-0002-7494-9767
  • Bruun, Hans Henrik|||0000-0003-0674-2577
  • Bucher, Solveig Franziska|||0000-0002-2303-4583
  • Buchmann, Nina|||0000-0003-0826-2980
  • Byun, Chaeho|||0000-0003-3209-3275
  • Carrara, Arnaud|||0000-0002-9095-8807
  • Conte, Adriano|||0000-0002-1309-2923
  • da Silva, Ana Carolina|||0000-0002-1285-640X
  • Duveiller, Gregory|||0000-0002-6471-8404
  • Fares, Silvano|||0000-0002-1990-0928
  • Ibrom, Andreas|||0000-0002-1341-921X
  • Knohl, Alexander|||0000-0002-7615-8870
  • Komac, Benjamin
  • Limousin, Jean-Marc|||0000-0002-2734-2495
  • Lusk, Christopher H.|||0000-0002-9037-7957
  • Mahecha, Miguel D.|||0000-0003-3031-613X
  • Martini, David|||0000-0003-2180-5126
  • Minden, Vanessa|||0000-0002-4933-5931
  • Montagnani, Leonardo|||0000-0003-4412-4243
  • Mori, Akira S.|||0000-0002-8422-1198
  • Onoda, Yusuke|||0000-0001-6245-2342
  • Peñuelas, Josep|||0000-0002-7215-0150
  • Perez-Priego, Oscar|||0000-0002-3138-3177
  • Poschlod, Peter
  • Powell, Thomas L.
  • Reich, Peter|||0000-0003-4424-662X
  • Šigut, Ladislav|||0000-0003-1951-4100
  • van Bodegom, Peter|||0000-0003-0771-4500
  • Walther, Sophia|||0000-0003-1681-9304
  • Wohlfahrt, Georg|||0000-0003-3080-6702
  • Wright, Ian J.|||0000-0001-8338-9143
  • Reichstein, Markus|||0000-0001-5736-1112
Fundamental axes of variation in plant traits result from trade-offs between costs and benefits of resource-use strategies at the leaf scale. However, it is unclear whether similar trade-offs propagate to the ecosystem level. Here, we test whether trait correlation patterns predicted by three well-known leaf- and plant-level coordination theories - the leaf economics spectrum, the global spectrum of plant form and function, and the least-cost hypothesis - are also observed between community mean traits and ecosystem processes. We combined ecosystem functional properties from FLUXNET sites, vegetation properties, and community mean plant traits into three corresponding principal component analyses. We find that the leaf economics spectrum (90 sites), the global spectrum of plant form and function (89 sites), and the least-cost hypothesis (82 sites) all propagate at the ecosystem level. However, we also find evidence of additional scale-emergent properties. Evaluating the coordination of ecosystem functional properties may aid the development of more realistic global dynamic vegetation models with critical empirical data, reducing the uncertainty of climate change projections.
Dipòsit Digital de Documents de la UAB
Publication Artículo científico (JournalArticle)

Intercomparison of global foliar trait maps reveals fundamental differences and limitations of upscaling approaches

  • Dechant, Benjamin|||0000-0001-5171-2364
  • Kattge, Jens|||0000-0002-1022-8469
  • Pavlick, Ryan
  • Schneider, Fabian D.|||0000-0003-1791-2041
  • Sabatini, Francesco M.
  • Álvaro Moreno-Martínez, Álvaro
  • Butler, Ethan E.|||0000-0003-3482-1950
  • van Bodegom, Peter|||0000-0003-0771-4500
  • Vallicrosa Pou, Helena|||0000-0002-5860-3096
  • Kattenborn, Teja|||0000-0001-7381-3828
  • Boonman, Coline C. F.|||0000-0003-2417-1579
  • Madani, Nima
  • Wright, Ian J.|||0000-0001-8338-9143
  • Dong, Ning|||0000-0003-0793-8854
  • Feilhauer, Hannes
  • Peñuelas, Josep|||0000-0002-7215-0150
  • Sardans i Galobart, Jordi|||0000-0003-2478-0219
  • Aguirre-Gutiérrez, Jesús|||0000-0001-9190-3229
  • Reich, Peter|||0000-0003-4424-662X
  • J. Leitão, Pedro J.
  • Townsend, Philip A.
Foliar traits such as specific leaf area (SLA), leaf nitrogen (N), and phosphorus (P) concentrations play important roles in plant economic strategies and ecosystem functioning. Various global maps of these foliar traits have been generated using statistical upscaling approaches based on in-situ trait observations. Here, we intercompare such global upscaled foliar trait maps at 0.5° spatial resolution (six maps for SLA, five for N, three for P), categorize the upscaling approaches used to generate them, and evaluate the maps with trait estimates from a global database of vegetation plots (sPlotOpen). We disentangled the contributions from different plant functional types (PFTs) to the upscaled maps and quantified the impacts of using different plot-level trait metrics on the evaluation with sPlotOpen: community weighted mean (CWM) and top-of-canopy weighted mean (TWM). We found that the global foliar trait maps of SLA and N differ drastically and fall into two groups that are almost uncorrelated (for P only maps from one group were available). The primary factor explaining the differences between these groups is the use of PFT information combined with remote sensing-derived land cover products in one group while the other group mostly relied on environmental predictors alone. The maps that used PFT and corresponding land cover information exhibit considerable similarities in spatial patterns that are strongly driven by land cover. The maps not using PFTs show a lower level of similarity and tend to be strongly driven by individual environmental variables. Upscaled maps of both groups were moderately correlated to sPlotOpen data aggregated to the grid-cell level (R = 0.2-0.6) when processing sPlotOpen in a way that is consistent with the respective trait upscaling approaches, including the plot-level trait metric (CWM or TWM) and the scaling to the grid cells with or without accounting for fractional land cover. The impact of using TWM or CWM was relevant, but considerably smaller than that of the PFT and land cover information. The maps using PFT and land cover information better reproduce the between-PFT trait differences of sPlotOpen data, while the two groups performed similarly in capturing within-PFT trait variation. Our findings highlight the importance of explicitly accounting for within-grid-cell trait variation, which has important implications for applications using existing maps and future upscaling efforts. Remote sensing information has great potential to reduce uncertainties related to scaling from in-situ observations to grid cells and the regression-based mapping steps involved in the upscaling.