Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “model data”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,333 records · Page 74
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Clinical Natural Language Processing for Radiation Oncology: A Review and Practical Primer

Natural language processing (NLP), which aims to convert human language into expressions that can be analyzed by computers, is one of the most rapidly developing and widely used technologies in the field of artificial intelligence. Natural language processing algorithms convert unstructured free text data into structured data that can be extracted and analyzed at scale. In medicine, this unlocking of the rich, expressive data within clinical free text in electronic medical records will help untap the full potential of big data for research and clinical purposes. Recent major NLP algorithmic advances have significantly improved the performance of these algorithms, leading to a surge in academic and industry interest in developing tools to automate information extraction and phenotyping from clinical texts. Thus, these technologies are poised to transform medical research and alter clinical practices in the future. Radiation oncology stands to benefit from NLP algorithms if they are appropriately developed and deployed, as they may enable advances such as automated inclusion of radiation therapy details into cancer registries, discovery of novel insights about cancer care, and improved patient data curation and presentation at the point of care. However, challenges remain before the full value of NLP is realized, such as the plethora of jargon specific to radiation oncology, nonstandard nomenclature, a lack of publicly available labeled data for model development, and interoperability limitations between radiation oncology data silos. Successful development and implementation of high quality and high value NLP models for radiation oncology will require close collaboration between computer scientists and the radiation oncology community. Here, we present a primer on artificial intelligence algorithms in general and NLP algorithms in particular; provide guidance on how to assess the performance of such algorithms; review prior research on NLP algorithms for oncology; and describe future avenues for NLP in radiation oncology research and clinics.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2026

Quantifying mean, variability, and uncertainty in indoor radon exposure in Pennsylvania using random forest and quantile regression forest models

Radon is a naturally occurring radioactive gas that poses a serious health risk as the primary cause of lung cancer in non-smokers. Despite the well-known adverse association with health outcomes, current radon exposure assessments are limited to county-level or average-level estimates, which fail to capture regional variability. This study uses Machine Learning models, including Random Forest (RF) and Quantile Regression Forest (QRF), to estimate the indoor radon concentrations at the ZCTA (Zip code tabulation area)-level and characterize uncertainties in model estimates. Incorporating geological, meteorological, and building-specific data, the models aim to improve radon risk assessment by capturing mean exposure, variability, and extreme concentration levels. Processed radon test data (n = 718,111) were analyzed using average, variability, and quantile prediction methods. Models that estimate the average radon exposure at the ZCTA-level can yield promising model-fit results, but they do not capture the underlying variability of indoor radon exposure within a ZCTA. We utilize volatility analyses to identify characteristics indicative of high variability of indoor radon exposure. We also show that a QRF model can be used to estimate upper quantiles of residential radon exposure, thereby uncovering localized areas of elevated exposure that were not apparent in mean estimates. The results highlighted the need for a deep characterization of exposure risk and show that regions with moderate average exposure levels could still harbor extreme outliers with implications for evaluating health risks. Utilizing multiple radon exposure models allows for a deeper characterization of radon risk within a geographic area and can better identify high-risk areas. The results from this study provide a foundation for developing mitigation strategies and examining associations between radon exposure and health outcomes at fine scales. Future research should extend the geographic scope and incorporate additional environmental risk factors to establish a comprehensive framework for risk assessment.

Lee, Heechan [ORNL]↗
DOE OSTI2024

Statistical Complexity of Quantum Learning

Abstract Learning problems involve settings in which an algorithm has to make decisions based on data, and possibly side information such as expert knowledge. This study has two main goals. First, it reviews and generalizes different results on the data and model complexity of quantum learning, where the data and/or the algorithm can be quantum, focusing on information‐theoretic techniques. Second, it introduces the notion of copy complexity, which quantifies the number of copies of a quantum state required to achieve a target accuracy level. Copy complexity arises from the destructive nature of quantum measurements, which irreversibly alter the state to be processed, limiting the information that can be extracted about quantum data. As a result, empirical risk minimization is generally inapplicable. The paper presents novel results on the copy complexity for both training and testing. To make the paper self‐contained and approachable by different research communities, an extensive background material is provided on classical results from statistical learning theory, as well as on the distinguishability of quantum states. Throughout, the differences between quantum and classical learning are highlighted by addressing both supervised and unsupervised learning, and extensive pointers are provided to the literature.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Investigation of carbon incorporation in laser-assisted MOCVD of GaN

Background carbon (C) impurity incorporation in metalorganic chemical vapor deposition (MOCVD) grown gallium nitride (GaN) represents one of the major issues in further improving GaN vertical power device performance. This work presents a laser-assisted MOCVD (LA-MOCVD) technique to address the high-C issue in MOCVD homoepitaxial GaN under different growth rate (Rg) regimes and studies the correlations between [C] and Rg. [C] in LA-MOCVD GaN is reduced by 50%–90% as compared to the conventional MOCVD GaN for a wide growth rate range between 1 and 16 μm/h. A mass-transport based model is developed to understand the C incorporation at different Rg regimes. The results obtained from the developed model are in good agreement with experimental data. The model further reveals that LA-MOCVD effectively suppresses C incorporation by reducing the active C species in the gas phase. Moreover, high step velocity in step flow growth mode can facilitate C incorporation at fast Rg, exhibiting steeper C increase. The theoretical model indicates that [C] can be suppressed below 1016 cm−3 with a fast growth rate (Rg) of 10 μm/h by utilizing higher power LA-MOCVD and freestanding GaN substrates with larger off-cut angles.

Physics↗
DOE OSTI2025

Automated RF Phase Adjustment for Beam Stabilization in the Fermilab Linac

The Fermilab Linac experiences longitudinal beam phase drift, leading to increased particle loss, conventionally corrected through labor-intensive manual RF adjustments. This project explores machine learning-based automation for drift correction, employing a prototype-based classification approach. Our model utilizes a 34-dimensional feature set (RF settings and BPM readings) and leverages a 7x27 response matrix for system modeling. To overcome limited real-world data, we generate synthetic data, enhancing model training and generalizability. Custom loss functions, including a surrogate energy-consistent loss and a temporal smoothness constraint, ensure physically plausible drift predictions. The goal is a robust system for autonomous phase adjustments, ensuring stable beam acceleration and reduced manual intervention.

Chichili, R. R. [Illinois U., Chicago]↗
DOE OSTI2022

Toward CO 2 Multimeasurement Geophysical Monitoring in the North Dakota CarbonSAFE Project

Paper presented at 16th International Conference on Greenhouse Gas Control Technologies (GHGT-16), October 23–27, 2022, Lyon, France. Overcoming data acquisition challenges associated with a lake in the study area, electrical infrastructure around the plant, noise from mine activities, complex near-surface conditions associated with reclaimed mine land, high wind speed, and extremely cold temperatures, data processing, modeling, and inversion of the baseline geophysical data demonstrated the importance of high-quality data for a CO 2 geophysical monitoring program. Implementing more advanced data acquisition and processing techniques in future monitoring surveys can further improve confidence in interpretations of geophysical survey data.

20 FOSSIL-FUELED POWER PLANTS↗
DOE OSTI2023

Water resource recovery modelling 2021 (WRRmod2021 conference)

Our society is transitioning fast into the digital age, spurred by development of cheap and new sensing technology, breakthroughs in computing, and development of efficient algorithms for optimization. This transition is also visible in the field of wastewater treatment and is driving new model developments, especially by exploiting the large data sets available with many utilities. Not surprisingly, WRRmod2021 had featured a strong session on ‘data-driven models and digitalization’ focused on this hot topic. At the same time, engineering practice calls for more robust models for performance evaluation and optimization of both conventional facilities and innovative processes. As a result, the WRRmod2021 program also exhibited sessions on modelling of new process units (e.g., aerobic granular sludge), modelling of the nitrogen cycle, and integrated/plant-wide modelling.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2020

Vadose Zone Model for B-3A/B Pond for Composite Analysis

The objectives of the vadose modeling for the updated Hanford Site composite analysis (CA) are to simulate the flow and transport of water and radionuclide releases from the surface to the water table and to provide radionuclide transfer rates for the plateau to river (P2R) model, version 8.3 (CP-57037, Model Package Report: Plateau to River Groundwater Model, Version 8.3). Water additions include natural recharge and water discharged to the ground as a result of industrial processes associated with Hanford Site operations. Contaminant sources include radionuclides in water discharged to the ground during operations and radionuclides disposed “dry” in solid waste burial grounds or other means. The following 16 radionuclides were selected for this modeling effort; carbon-14 (C-14), chlorine-36 (Cl-36), tritium (H-3), iodine-129 (I-129), neptunium-237 (Np-237), rhenium-187 (Re-187), strontium-90 (Sr-90), technetium-99 (Tc-99), uranium-232 (U-232), uranium-233 (U-233), uranium-234 (U-234), uranium-235 (U-235), uranium-236 (U-236), uranium-238 (U-238), radium-226 (Ra-226), and thorium-230 (Th-230). The simulation time starts in 1943 and ends at 12070, which is 10,000 years after assumed Hanford Site closure in 2070. The parallel version of the Subsurface Transport Over Multiple Phases (STOMP) simulator officially named the exascale Subsurface Transport Over Multiple Phases (eSTOMP), is used to simulate flow and transport for the vadose models. The documentation for the STOMP code is comprehensive. The theoretical and numerical approaches applied in the STOMP code are documented in a published theory guide (PNNL-12030, STOMP Subsurface Transport Over Multiple Phases Version 2.0 Theory Guide). The code has undergone a rigorous verification procedure against analytical solutions, laboratory-scale experiments, and field-scale demonstrations. The application guide (PNNL-11216, STOMP Subsurface Transport Over Multiple Phases Application Guide) provides instructive examples in the application of the code to classical groundwater problems. The user’s guide (PNNL-15782, STOMP: Subsurface Transport Over Multiple Phases Version 4.0: User’s Guide) describes the general use, input file formatting, compilation, and execution of the code. The primary output of the vadose zone modeling is radionuclide transfer rates to the groundwater for input into the P2R model. The rates will be summed over the 100 by 100 m P2R grid cells that fall within the vadose zone model source domain. The Hanford Site Central Plateau was subdivided into 26 individual vadose zone models, with 13 in the 200 East Area and 13 in the 200 West Area. Waste sites that have a completed performance assessment (PA) or past-leak analysis were not included as sources of radionuclides. Instead the vadose zone to groundwater transfer rates of the Environmental Restoration Disposal Facility, Integrated Disposal Facility, US Ecology, and Waste Management Area C (WMA C) PAs and the past-leak analysis for WMA C were used as direct input to the P2R model. Each of the vadose zone models is documented in separate environmental calculation files (ECFs). This ECF describes the B-3A/B Ponds model. The scope of this ECF is to document the development and results of the B-3 A/B Ponds vadose zone model. CP-63515, Model Package Report: Central Plateau Vadose Zone Models, describes the approach, assumptions, process of determining the number of models required and domain of each model, input data, and processing common to all the models.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2020

Vadose Zone Model for SALDS for Composite Analysis

The objectives of the vadose modeling for the updated Hanford Site composite analysis (CA) are to simulate the flow and transport of water and radionuclide releases from the surface to the water table and to provide radionuclide transfer rates for the plateau to river (P2R) model, version 8.3 (CP-57037, Model Package Report: Plateau to River Groundwater Model, Version 8.3). Water additions include natural recharge and water discharged to the ground as a result of industrial processes associated with Hanford Site operations. Contaminant sources include radionuclides in water discharged to the ground during operations and radionuclides disposed “dry” in solid waste burial grounds or other means. The following 16 radionuclides were selected for this modeling effort: carbon-14 (C-14), chlorine-36 (Cl-36), tritium (H-3), iodine-129 (I-129), neptunium-237 (Np-237), rhenium-187 (Re-187), strontium-90 (Sr-90), technetium-99 (Tc-99), uranium-232 (U-232), uranium-233 (U-233), uranium-234 (U-234), uranium-235 (U-235), uranium-236 (U-236), uranium-238 (U-238), radium-226 (Ra-226), and thorium-230 (Th-230). The simulation time starts in 1943 and ends at 12070, which is 10,000 years after assumed Hanford Site closure in 2070. The parallel version of the Subsurface Transport Over Multiple Phases (STOMP1) simulator, officially named the exascale Subsurface Transport Over Multiple Phases (eSTOMP) is used to simulate flow and transport for the vadose models. The documentation for the STOMP code is comprehensive. The theoretical and numerical approaches applied in the STOMP code are documented in a published theory guide (PNNL-12030, STOMP Subsurface Transport Over Multiple Phases Version 2.0 Theory Guide). The code has undergone a rigorous verification procedure against analytical solutions, laboratory-scale experiments, and field-scale demonstrations. The application guide (PNNL-11216, STOMP Subsurface Transport Over Multiple Phases Application Guide) provides instructive examples in the application of the code to classical groundwater problems. The user’s guide (PNNL-15782, STOMP: Subsurface Transport Over Multiple Phases Version 4.0: User’s Guide) describes the general use, input file formatting, compilation, and execution of the code. The primary output of the vadose zone modeling is radionuclide transfer rates to the groundwater for input into the P2R model. The rates will be summed over the 100 by 100 m P2R grid cells that fall within the vadose zone model source domain. The Hanford Site Central Plateau was subdivided into 26 individual vadose zone models, with 13 in the 200 East Area and 13 in the 200 West Area. Waste sites that have a completed performance assessment (PA) or past-leak analysis were not included as sources of radionuclides. Instead the vadose zone to groundwater transfer rates of the Environmental Restoration Disposal Facility, Integrated Disposal Facility, US Ecology, and Waste Management Area C (WMA C) PAs and the past-leak analysis for WMA C were used as direct input to the P2R model. Each of the vadose zone models is documented in separate environmental calculation files (ECFs). This ECF describes the State-Approved Land Disposal Site (SALDS) model. The scope of this ECF is to document the development and results of the SALDS vadose zone model. CP-63515, Model Package Report: Central Plateau Vadose Zone Models, describes the approach, assumptions, process of determining the number of models required and domain of each model, input data, and processing common to all the models.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2020

Data Repository for Power system Open models With Evolving Resources (DR POWER) (Final Scientific/Technical Report)

Pacific Northwest National Laboratory (PNNL) and the National Rural Electric Cooperative Association (NRECA) created a Data Repository for Power system Open models With Evolving Resources (DR POWER) to establish, curate, and evolve open-access power grid models and scenarios, and measurements and observations, collectively referred to as datasets. Existing open-access datasets are inadequate (too small and not representative of the complexity of the modern grid, and scattered all over the world wide web) for the investigation and benchmarking of the advanced optimization methods needed for next-generation grids. Open-access datasets are static and inadequate in other aspects: they do not provide format transformation tools; there is missing information needed for advanced applications; hard to cite; no model evolution history; difficult to update current collections; and little to no community involvement and feedback. Overall, the current landscape is hindering model evolution and benchmarking standards needed for transformational approaches to be developed, tested, and accepted by the power modeling community.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

Automated RF Phase Adjustment for Beam Stabilization in the Fermilab Linac

The Fermilab Linac experiences longitudinal beam phase drift, leading to increased particle loss, conventionally cor- rected through labor-intensive manual RF adjustments. This project explores machine learning-based automation for drift correction, employing a prototype-based classification approach. Our model utilizes a 34-dimensional feature set (RF settings and BPM readings) and leverages a 7x27 response matrix for system modeling. To overcome limited real-world data, we generate synthetic data, enhancing model training and generalizability. Custom loss functions, including a sur- rogate energy-consistent loss and a temporal smoothness constraint, ensure physically plausible drift predictions. The goal is a robust system for autonomous phase adjustments, ensuring stable beam acceleration and reduced manual intervention.

Chichili, R. R. [U. Illinois, Chicago]↗
DOE OSTI2022

Current and Projected Future Distribution of Plant Community Types for the Southern Seward Peninsula, Alaska

Landscape scale maps of plant community distribution at 5 m resolution were developed for southern Seward Peninsula. A Random Forest based environmental niche model was developed using topographic and climate data sets. Models were trained for the 2010-2019 period and applied to develop plant community distributions for contemporary (2010-2019) period. Trained model was applied to project the plant community distributions under future climate. Using downscaled projections for RCP8.5 scenario from five climate models (CCSM4, GFDL-CM3, GISS-E2-R, IPSL-CM5A-LR, MRI-CGCM3) and multi-model mean, a six member ensemble of plant community distribution were developed for current decade (2010-2019) and for future decades (2020-2029, 2030-2039, 2040-2049, 2050-2059). Contains *.tif and *.sh files. The Next-Generation Ecosystem Experiments: Arctic (NGEE Arctic), was a research effort to reduce uncertainty in Earth System Models by developing a predictive understanding of carbon-rich Arctic ecosystems and feedbacks to climate. NGEE Arctic was supported by the Department of Energy's Office of Biological and Environmental Research. The NGEE Arctic project had two field research sites: 1) located within the Arctic polygonal tundra coastal region on the Barrow Environmental Observatory (BEO) and the North Slope near Utqiagvik (Barrow), Alaska and 2) multiple areas on the discontinuous permafrost region of the Seward Peninsula north of Nome, Alaska. Through observations, experiments, and synthesis with existing datasets, NGEE Arctic provided an enhanced knowledge base for multi-scale modeling and contributed to improved process representation at global pan-Arctic scales within the Department of Energy's Earth system Model (the Energy Exascale Earth System Model, or E3SM), and specifically within the E3SM Land Model component (ELM).

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Data and scripts associated with “Sequential Precipitation Input Tagging (SPIT) to Estimate Water Transit Times and Hydrologic Tracer Dynamics within Water-Tagging Enabled Hydrologic Models” (v3)

This data package is associated with the publication “Sequential Precipitation Input Tagging (SPIT) to Estimate Water Transit Times and Hydrologic Tracer Dynamics within Water-Tagging Enabled Hydrologic Models” submitted to Journal of Advances in Modeling Earth Systems (Butler et al. 2025). This study developed the Sequential Precipitation Input Tagging (SPIT) framework to tag input precipitation and estimate water transit times and hydrologic tracers. SPIT tags all precipitation events at regular intervals over an extended period (monthly tags over seven years) in a hydrologic model from 2016-2022. SPIT is applied at six National Ecological Observatory Network (NEON) sites across the continental United States to calculate transit time distributions (TTD) and derive from these mean transit times (MTT), fractions of young water (Fyw), and hydrologic tracer concentrations in stream water (δ18O) within a water-tagging enabled version of the Weather Research and Forecast (WT-WRF-Hydro) model with national water model (NWM) configurations. We go on to validate WT-WRF-Hydro estimates against Butler et al. (2023), who analyzed the same NEON sites using stable water isotope data to estimate water transit times. This new tracking method provides a detailed picture of water movement and helps improve predictions about water availability in the future. This data package was originally published in January 2025. It was updated May 2025 (v2; new and modified files) and October 2025 (v3; new and modified files). File and folder names were not revised to indicate changes. See the change history section in the readme for more details. This data package contains the data and scripts used to develop the SPIT framework WT-WRF-Hydro (Water Tagging Weather Research and Forecasting Hydrologic) model and is associated with the following GitHub repository: https://github.com/zbutler33/SPIT-Framework. This data package contains five parent folders: (1) “Manipulated_outputs”, (2) “Metadata”, (3) “Observed”, (4) “Outputs”, and (5) “Scripts”. Each of these parent folders contains additional subfolders and files. Please see the FLMD (“v*_Butler_2024_WT_WRF_Hydro_flmd.csv”) for a list of all the files contained in this data package and descriptions for each. See the data dictionary (“v*_Butler_2024_WT_WRF_Hydro_dd.csv”) for definitions and units of all of the tabular (files ending in “.csv” and ".tsv") column headers.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Learning Implicit Models of Complex Dynamical Systems From Partial Observations [Slides]

Conclusions: certain data-driven models do implicitly what physics simulations do explicitly; theoretical underpinnings in Koopman theory, delay-coordinate embeddings, and Mori-Zwanzig formalism; best approach for given system with finite data an open question; physics-informed machine learning (PIML) emerging framework for combining explicit and implicit modeling.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

3-D Radiological Data Acquisition, Visualization and Modeling - 20211

The U.S. Army Corps of Engineers (USACE) was tasked to investigate and remediate low activity radiological contamination from research and production of the nation's first nuclear weapons at the former DuPont Chambers Works Formerly Utilized Sites Remedial Action Program (FUSRAP) site (DuPont). The DuPont site had several buildings used for the Manhattan project that were demolished in the 1940's and 1950's apparently using heavy earthmoving equipment. Some of the contaminated rubble from the demolition appears to have been spread out by this equipment resulting in somewhat random scattering of radiologically contaminated soil and debris along with aqueous spills. Traditional investigative methods such as soil borings, test pits and 2-dimensional gamma walkovers were only partially successful in delineating the radiological contamination at the site. It was feared that even 'chasing' the contamination during remediation would miss contamination if the demolition resulted in discontinuous trails of radiologically contaminated soils. In evaluating the data generated over the interceding decades, the USACE determined that a better method to collect and process the remedial action radiological data was needed to enable the project team to optimize predictive planning and meet documentation expectations. The purpose of this paper is to provide an overview of the effort and progress to combine and organize radiological survey methods into a highly flexible sampling, modeling, and decision analysis approach that emphasizes the quality of decision-making during remediation. This innovative system blends multiple tools to develop a methodology that can extend MARRSIM [1] into the subsurface and provide tools that can be applied to other sites. (authors)

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

A machine-learning-driven data labeling pipeline for scientific analysis in MLExchange

This study introduces a novel labeling pipeline to accelerate the labeling process of scientific data sets by using artificial intelligence (AI)-guided tagging techniques. This pipeline includes a set of interconnected web-based graphical user interfaces (GUIs), where Data Clinic and MLCoach enable the preparation of machine learning (ML) models for data reduction and classification, respectively, while Label Maker is used for label assignment. Throughout this pipeline, data can be accessed through a direct connection to a file system or through Tiled for access through Hypertext Transfer Protocol (HTTP). Our experimental results present three use cases where this labeling pipeline has been instrumental for the study of large X-ray scattering data sets in the area of pattern recognition, the remote analysis of resonant soft X-ray scattering data and the fine-tuning process of foundation models. These use cases highlight the labeling capabilities of this pipeline, including the ability to label large data sets in a short period of time, to perform remote data analysis while minimizing data movement and to enhance the fine-tuning process of complex ML models with human involvement.

Chavez, Tanny (ORCID:0000000193172896)↗
DOE OSTI2023

Soil Respiration Responses to Throughfall Exclusion Are Decoupled From Changes in Soil Moisture for Four Tropical Forests, Suggesting Processes for Ecosystem Models

Abstract Climatic drying is predicted for many tropical forests yet models remain poorly parameterized for these ecosystems, hampering predictions of forest‐climate interactions. We applied an integrated model–experiment approach, parameterizing an ecosystem model with tropical forest observational data and comparing model predictions to a field drying manipulation. We hypothesized that drying suppresses soil CO 2 fluxes (i.e., respiration) in relatively dry tropical forests but increases CO 2 fluxes in wetter tropical forests by alleviating anaerobiosis. We measured soil CO 2 fluxes during wet‐dry cycles from 2015 to 2022 in four Panamanian forests that vary in rainfall and soil fertility. Measured soil CO 2 fluxes declined in the dry season and peaked in the early wet season ahead of peak soil moisture, resulting in lower soil moisture optima for respiration than previously modeled. We then parameterized the model using field data and the new moisture‐respiration response functions. The updated model predicted increased soil CO 2 fluxes with drying in wetter and fertile forests and suppressed fluxes in drier, infertile forests. In contrast to model predictions, a chronic throughfall exclusion experiment initially suppressed soil respiration across forests, with sustained suppression for four years in the wettest forest only (−28% ± 4% during the dry season). In the fertile forest, drying eventually elevated CO 2 fluxes over this period (+75% ± 28% during the late wet season). The unexpected negative drying effect in the wettest, infertile forest could have resulted from reduced vertical flushing of nutrients into soils. Including hydro‐nutrient interactions in ecosystem models could improve predictions of tropical forest‐climate feedbacks.

54 ENVIRONMENTAL SCIENCES↗