Publications
Publications by categories in reversed chronological order.
Journal Articles
- Toward robust parameterizations in ecosystem-level photosynthesis modelsJournal of Advances in Modeling Earth Systems, 2023
Abstract In a model simulating dynamics of a system, parameters can represent system sensitivities and unresolved processes, therefore affecting model accuracy and uncertainty. Taking a light use efficiency (LUE) model as an example, which is a typical approach for estimating gross primary productivity (GPP), we propose a Simultaneous Parameter Inversion and Extrapolation approach (SPIE) to overcome issues stemming from plant-functional-type (PFT)-dependent parameterizations. SPIE refers to predicting model parameters using an artificial neural network based on collected variables, including PFT, climate types, bioclimatic variables, vegetation features, atmospheric nitrogen and phosphorus deposition, and soil properties. The neural network was optimized to minimize GPP errors and constrain LUE model sensitivity functions. We compared SPIE with 11 typical parameter extrapolating methods, including PFT- and climate-specific parameterizations, global and PFT-based parameter optimization, site-similarity, and regression approaches. All methods were assessed using Nash-Sutcliffe model efficiency (NSE), determination coefficient and normalized root mean squared error, and contrasted with site-specific calibrations. Ten-fold cross-validated results showed that SPIE had the best performance across sites, various temporal scales and assessing metrics. Taking site-level calibrations as a benchmark (NSE = 0.95), SPIE performed with an NSE of 0.68, while all the other investigated approaches showed lower NSE. The Shapley value, layer-wise relevance and partial dependence showed that vegetation features, bioclimatic variables, soil properties and some PFTs determine parameters. SPIE overcomes strong limitations observed in many standard parameterization methods. We argue that expanding SPIE to other models overcomes current limits and serves as an entry point to investigate the robustness and generalization of different models.
@article{Bao2023, author = {Bao, S. and Alonso, L. and Wang, S. and Gensheimer, J. and De, R. and Carvalhais, N.}, title = {Toward robust parameterizations in ecosystem-level photosynthesis models}, journal = {Journal of Advances in Modeling Earth Systems}, volume = {15}, number = {8}, pages = {e2022MS003464}, keywords = {parameter extrapolation, parameterization, ecosystem properties, hybrid model, feature importance, vegetation productivity}, doi = {10.1029/2022MS003464}, url = {https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2022MS003464}, year = {2023}, }
Pre-prints/ In Preparation
- Addressing Challenges in Simulating Inter–annual Variability of Gross Primary ProductionJan 2025
A long-standing challenge in studying the global carbon cycle has been understanding the factors controlling inter–annual variation (IAV) of carbon fluxes, and improving their representations in existing biogeochemical models. Here, we compared an optimality-based model and a semi-empirical light use efficiency model to understand how current models can be improved to simulate IAV of gross primary production (GPP). Both models simulated hourly GPP and were parameterized for (1) each site–year, (2) each site with an additional constraint on IAV (CostIAV), (3) each site, (4) each plant–functional type, and (5) globally. This was followed by forward runs using calibrated parameters, and model evaluations using Nash–Sutcliffe efficiency (NSE) as a model-fitness measure at different temporal scales across 198 eddy-covariance sites representing diverse climate–vegetation types. Both models simulated hourly GPP better (median normalized NSE: 0.83 and 0.85) than annual GPP (median normalized NSE: 0.54 and 0.63) for most sites. Specifically, the optimality-based model substantially improved from NSE of -1.39 to 0.92 when drought stress was explicitly included. Most of the variability in model performances was due to model types and parameterization strategies. The semi-empirical model produced statistically better hourly simulations than the optimality-based model, and site–year parameterization yielded better annual model performance. Annual model performance did not improve even when parameterized using CostIAV. Furthermore, both models underestimated the peaks of diurnal GPP, suggesting that improving predictions of peaks could produce better annual model performance. Our findings reveal current modelling deficiencies in representing IAV of carbon fluxes and guide improvements in further model development.
@unpublished{De_IAV_GPP_2025, author = {De, R. and Bao, S. and Koirala, S. and Brenning, A. and Reichstein, M. and Tagesson, T. and Liddell, M. and Ibrom, A. and Wolf, S. and Šigut, L. and Hörtnagl, L. and Woodgate, W. and Korkiakoski, M. and Merbold, L. and Black, T. A. and Roland, M. and Klosterhalfen, A. and Blanken, P. D. and Knox, S. and Sabbatini, S. and Gielen, B. and Montagnani, L. and Fensholt, R. and Wohlfahrt, G. and Desai, A. R. and Paul-Limoges, E. and Galvagno, M. and Hammerle, A. and Jocher, G. and Ruiz Reverter, B. and Holl, D. and Chen, J. and Vitale, L. and Arain, M. A. and Carvalhais, N.}, title = {{Addressing Challenges in Simulating Inter--annual Variability of Gross Primary Production}}, journal = {{ESS Open Archive}}, volume = {2025}, month = jan, year = {2025}, pages = {1--75}, url = {https://essopenarchive.org/doi/full/10.22541/essoar.172656939.93739740/v2}, doi = {10.22541/essoar.172656939.93739740/v2}, }
Conference Presentations
- Addressing challenges in representing inter-annual variability of gross primary productivity fluxes using robust empirical and theory-based modelsIn Book of Abstracts, ICOS Science Conference 2024, Versailles, France, Sep 2024
Terrestrial vegetation is crucial in the carbon cycle, primarily for mediating atmospheric CO2 fluxes through photosynthesis. The inter-annual variation (IAV) of gross primary production (GPP) is highly correlated with peak GPP frequency in diurnal cycles, yet biogeochemical models often fail to simulate these peaks. Here, we compare a light use efficiency (LUE) model and an eco-evolutionary optimality-based P-model to simulate the IAV of GPP. Both models estimate GPP at sub-daily scale, and are optimized for each site-year, site, plant functional type, and globally, followed by an evaluation at different temporal resolutions across 198 eddy-covariance sites of diverse climate-vegetation types. Hourly model (site-year optimization) evaluations show significantly higher performance (Nash-Sutcliffe efficiency, viz. NSE ≥ 0.6 for 93% and 86% of sites for LUE and P-model respectively) compared to yearly aggregated evaluations (NSE ≥ 0.6 for 44% and 29% of sites for LUE and P-model respectively). The LUE model performs better than P-model as we optimize parameters in temperature, vapor-pressure deficit functions and consider the cloudiness effects. Further investigation of the role of soil moisture stress reveals that it substantially improves the P-model’s (which already considers acclimation of parameters) performance (from an NSE of -0.94 to 0.93) in simulating annual average GPP per site-year. For both models, optimized parameter values vary more across the sites than in the site-years. Moreover, we analyze the role of data versus model epistemic uncertainty in estimating the IAV of GPP. Our model-data-integration experiments reveal current models’ limitations in capturing the IAV of carbon flux and guide their improvements.
@inproceedings{DeR_ICOS_Sc_Con2024, title = {{Addressing challenges in representing inter-annual variability of gross primary productivity fluxes using robust empirical and theory-based models}}, author = {De, R. and Bao, S. and Koirala, S. and Brenning, A. and Reichstein, M. and Carvalhais, N.}, year = {2024}, month = sep, booktitle = {Book of Abstracts, ICOS Science Conference 2024, Versailles, France}, pages = {37}, url = {https://www.icos-cp.eu/news-and-events/science-conference/icos2024sc/all-abstracts}, } - Evaluating a generic parameterization approach for modelling photosynthesis across eddy-covariance sitesIn EGU General Assembly 2023, Vienna, Austria, Apr 2023
@inproceedings{DeR_EGU2023, title = {{Evaluating a generic parameterization approach for modelling photosynthesis across eddy-covariance sites}}, author = {De, R. and Bao, S. and Weber, U. and Koirala, S. and Yang, H. and Carvalhais, N.}, year = {2023}, month = apr, booktitle = {{EGU General Assembly 2023, Vienna, Austria}}, doi = {10.5194/egusphere-egu23-13032}, url = {https://meetingorganizer.copernicus.org/EGU23/EGU23-13032.html}, } - Towards Robust Parameterizations in Ecosystem-level Photosynthesis ModelsIn EGU General Assembly 2023, Vienna, Austria, Apr 2023
@inproceedings{Bao_EGU2023, title = {{Towards Robust Parameterizations in Ecosystem-level Photosynthesis Models}}, author = {Bao, S. and Carvalhais, N. and Alonso, L. and Wang, S. and Gensheimer, J. and De, R. and Shi, J.}, year = {2023}, month = apr, booktitle = {{EGU General Assembly 2023, Vienna, Austria}}, doi = {10.5194/egusphere-egu23-5431}, url = {https://meetingorganizer.copernicus.org/EGU23/EGU23-5431.html}, }
Theses
- Crop Yield Estimation from Simulated Carbon Flux using SCOPE: A Case Study of Samrakalwana Village In IndiaRanit DeUniversity of Twente, Jul 2021
The world has been facing a huge population boom for the last few decades. In this context, food security is a big challenge for many developing countries. Crop yield estimates play a crucial role in formulating food-related policies. They are generally produced using statistical data. But integrating actual remote sensing observations and deeper understanding of the dynamics of crops can help us to provide more accurate crop yield estimates. Moreover, we can have a sense of how crops respond to changing climatic conditions. Several parameters (such as chlorophyll content, leaf area index [LAI], water content in leaf etc.) defining crop dynamics are essential inputs for modelling ecosystem carbon and water fluxes. It is possible to retrieve values of these parameters from remote sensing observations. This study focuses on two wheat-growing seasons (2018-19 and 2019-20) at Samrakalwana village, located in the northern part of India. The main objective of this study is to simulate ecosystem fluxes using Soil Canopy Observation of Photosynthesis and Energy fluxes (SCOPE) with vegetation parameters retrieved from Sentinel-3 (S3) and Sentinel-2 (S2) data. Then, the simulated carbon flux was used to provide crop yield estimate. Two radiative transfer models (i.e., Optical Radiative Transfer Routine [RTMo] of SCOPE and Soil-Plant-Atmosphere Radiative Transfer [SPART]) were inverted to retrieve mainly crop parameters from S3 Ocean and Land Color Imager (OLCI) and S2 Multispectral Instrument (MSI) observations. The RTMo in SCOPE only represents a soil-vegetation system, whereas the SPART includes atmosphere also and it is possible to retrieve parameters defining atmospheric conditions (Aerosol Optical Thickness [AOT], columnar water and ozone content). The data from S3 OLCI was used as it has observations from 21 different bands with a higher temporal resolution of 1.1 days. In contrast, the advantage of S2 MSI is its higher spatial resolutions (4, 6 and 3 bands with 10m, 20m and 60m resolution respectively). The retrieved parameters and meteorological data fromECMWFERA5 datasetwere then used to model ecosystem fluxes (Gross Primary Production [GPP] and Evapotranspiration [ET]) using SCOPE. The SCOPE simulated GPP and ET were compared against MODIS and ECOSTRESS bases GPP and ET products. Then simulated GPP fluxes were used to provide crop yield estimate. Supplementary information, such as Water Use Efficiency (WUE), Light Use Efficiency (LUE) and Evaporative fraction (EF), were also calculated. The retrieved parameters, in general, are affected by spikes due to noisy input data. In some cases, the expected pattern of crop dynamics can be observed and retrieved LAI agrees with field measured LAI. The SCOPE simulated GPP flux was in a range of 0 to 12 μmol m2s-1. The simulated ET was in a range of 0 to 11 mm day-1. It was found that the values of simulated fluxes are mostly higher than the MODIS based estimate. Crop yield estimates from simulated carbon fluxes were also bit higher than the actual field measurements.
@phdthesis{de_2021_12951938, author = {De, Ranit}, title = {{Crop Yield Estimation from Simulated Carbon Flux using SCOPE: A Case Study of Samrakalwana Village In India}}, school = {{University of Twente}}, address = {{Enschede, Netherlands}}, year = {2021}, month = jul, doi = {10.5281/zenodo.12951938}, url = {https://essay.utwente.nl/88913/}, }
