the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 24 Jan 2025
BOSSE v1.0: the Biodiversity Observing System Simulation Experiment
Abstract. 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.
- Preprint
(5819 KB) - Metadata XML
-
Supplement
(5476 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-318', Anonymous Referee #1, 24 Feb 2025
-
AC3: 'Reply on RC1', Javier Pacheco-Labrador, 13 May 2025
We thank the reviewer for the questions and comments that we would like to clarify here.
Regarding the multi-angular capabilities, so far, the model does not allow for off-nadir observation since the point spread function (PSF) is projected over the surface only from nadir. The emulators simulate the different spectroradiometric variables for multiple sun angles and nadir view. Recognizing the importance of off-nadir observation in modeling bidirectional reflectance, we seek to include off-nadir PSF projection in the next version of BOSSE. This will be stressed in the main text of the manuscript. In fact, this task is partly advanced, and we have already trained some of the multi-angular emulators, and the implementation of the remote sensing PSF needs to be developed.
Regarding the mixture of species, BOSSE species are individually simulated in the background and always correspond to a single pure pixel at the selected baseline. All the variables are thus handled at a relative resolution of 100 % (i.e., each pixel represents a single individual or a group of identical individuals). Then, this resolution can be degraded when the output (e.g., the simulated remote sensing imagery or plant trait maps) is provided to the user. In that case, the traits and spectral signals of the different nearby pixels are averaged in different ways: simple averaging for plant trait maps, simulating field sampling plots, or weighted averaging according to the PSF of the remote sensors. We have clarified this in sections 2.4.1 and 2.4.2, respectively. Additionally, any user could produce the maps at 100 % spatial resolution and aggregate the data using any approach of their choice.
We also thank the reviewer for the manuscript recommended. This is indeed a complementary approach in line with the aims of BOSSE, and we will make reference to it in the revised manuscript.
Citation: https://doi.org/10.5194/egusphere-2025-318-AC3
-
AC3: 'Reply on RC1', Javier Pacheco-Labrador, 13 May 2025
-
CEC1: 'Comment on egusphere-2025-318 - No compliance with the policy of the journal', Juan Antonio Añel, 05 Mar 2025
Dear authors,
Unfortunately, after checking your manuscript, it has come to our attention that it does not comply with our "Code and Data Policy".
https://www.geoscientific-model-development.net/policies/code_and_data_policy.html
You have archived your code on GitHub. However, GitHub is not a suitable repository for scientific publication. GitHub itself instructs authors to use other long-term archival and publishing alternatives, such as Zenodo. Therefore, the current situation with your manuscript is irregular. Please, publish your code in one of the appropriate repositories and reply to this comment with the relevant information (link and a permanent identifier for it (e.g. DOI)) as soon as possible, as we can not accept manuscripts in Discussions that do not comply with our policy.In this way, if you do not fix this problem, we will have to reject your manuscript for publication in our journal.
Also, you must include a modified 'Code Availability' section in a potentially reviewed manuscript, containing the link and DOI of the new repositories.
Juan A. Añel
Geosci. Model Dev. Executive EditorCitation: https://doi.org/10.5194/egusphere-2025-318-CEC1 -
AC1: 'Reply on CEC1', Javier Pacheco-Labrador, 05 Mar 2025
Dear Editor,
Thanks for the warning. I was not aware of this. A release of the code has been published in Zenodo under DOI: https://doi.org/10.5281/zenodo.14973471 . The DOI will be included in the manuscript during the review.
On behalf of the co-authors,
Javier Pacheco-Labrador
Citation: https://doi.org/10.5194/egusphere-2025-318-AC1
-
AC1: 'Reply on CEC1', Javier Pacheco-Labrador, 05 Mar 2025
-
RC2: 'Comment on egusphere-2025-318', Anonymous Referee #2, 22 Apr 2025
It is the understanding of this reviewer that the paper presents BOSSE (Biodiversity Observing System Simulation Experiment), a modelling framework designed to generate synthetic landscapes for the evaluation of remote sensing (RS) approaches to monitoring plant functional diversity (PFD) and biodiversity-ecosystem function (BEF) relationships. BOSSE simulates spatially explicit vegetation communities and generates RS images - such as hyperspectral reflectance, solar induced chlorophyll fluorescence (SIF) and land surface temperature (LST) - linked to ecosystem functions. The aim is to fill a critical gap in the field by providing a consistent, controlled environment for testing and benchmarking RS-based biodiversity metrics.
BOSSE appears to be a novel and timely contribution. It stands out as a pioneering observation system simulation experiment explicitly focused on plant diversity and BEF research, an area where dedicated tools have been lacking. The integration of different RS modalities (reflectance, fluorescence, LST) with ecological functional modelling makes BOSSE a flexible and forward-looking framework for hypothesis testing and method development. According to this reviewer, the use of radiative transfer models, emulators and phenological modelling adds scientific depth and enables realistic simulations that increase analytical confidence.
The fact that both the data and the code (pyBOSSE) are openly shared is another major strength. Detailed documentation, including Jupyter notebooks, helps to ensure that the tool is accessible and reproducible, which should encourage wider adoption by the community. The potential of BOSSE to bridge ecology and RS is significant - it could support future research into biodiversity monitoring, ecosystem stability and even climate shift impacts.
However, there are some limitations and areas for development that authors are encouraged to consider. First, the reliance on 1D radiative transfer models, while common, limits realism - especially in structurally complex ecosystems such as forests. The authors acknowledge this and note plans to incorporate 3D models (e.g. DART, SCOPE 2.0), which would be a welcome improvement. Similarly, the current lack of LiDAR and radar simulations reduces the applicability of the framework to emerging biodiversity monitoring techniques. These planned extensions are promising, but their feasibility and implementation timeframe remain somewhat uncertain at this stage.
To make BOSSE even more accessible, the authors could consider adding a user-friendly GUI or web-based interface to complement the Python package. Providing pre-configured starter scenarios would also help lower the barrier to entry for new users.
In terms of presentation, the paper is generally well written, but some improvements would improve clarity. Some sections, particularly those describing the phenological and trait generation models, are quite technical and could benefit from simplified overviews or visual flowcharts. Similarly, figures such as Figure 2 are informative but could be made more intuitive with better annotation and more descriptive legends.
Authors are also encouraged to include a glossary of key terms (e.g. PFT, GSI), which would make the paper more accessible to readers from different disciplinary backgrounds. In addition, the methods section could be streamlined by moving some of the more technical implementation details to supplementary material.
While the synthetic nature of BOSSE is a strength, the paper would be further strengthened if the authors included a validation case study - perhaps comparing BOSSE results with real-world observations - or a sensitivity analysis to explore model robustness. A more thorough quantification of the emulator uncertainty would also improve the transparency and scientific rigour of the approach.
Finally, the paper does not currently discuss the computational feasibility of large-scale simulations in much detail. Including runtime metrics or performance benchmarks for larger scenes would give potential users a better sense of the scalability and practical implementation requirements of BOSSE.
In summary, this reviewer considers BOSSE to be a promising and well-executed contribution to geoscientific model development, with strong relevance to the RS and biodiversity communities. With some targeted improvements - particularly in the areas of accessibility, validation and computational performance - the paper could have an even wider impact.
Citation: https://doi.org/10.5194/egusphere-2025-318-RC2 -
AC2: 'Reply on RC2', Javier Pacheco-Labrador, 13 May 2025
We thank the reviewer for the detailed review and the suggestions regarding potential improvements and developments of the model. While it is not possible to tackle them all in the first version of this model, we will implement some of the recommendations in this review, and some others in the next version of the model. The open-source nature of pyBOSSE allows the community to contribute and include new features, foreseen or not, by the authors or the reviewers of this manuscript.
Regarding the improvement of radiative transfer models, we acknowledge the limitations and anticipate future model improvements, but we cannot provide accurate timelines for all of them. The inclusion of 3D modeling has already been foreseen, and we hope we can incorporate it in the next version of the model. LiDAR and radar radiative transfer modeling may not yet be so mature and might take longer. Consider this is the first version of the model, and it already includes a wide variety of spectral signals. BOSSE v1.0 is able, in its current state, to provide answers to “simple”, yet relevant and unanswered questions. Likely, future designs could be helped by the lessons learned during the efforts to answer these questions. Furthermore, the open-source code and modular design would allow researchers interested in questions BOSSE v1.0 could not answer to implement additional features, including those suggested by the reviewer. Finally, while there is a commitment to further developing BOSSE, implementing new functionalities will depend on available funding and the community’s interest in the model.
Regarding the inclusion of GUIs or pre-configured scenarios, we thank the reviewer for the suggestion. However, we had to balance modeling efforts and the development of interfaces. Still, this is a good idea that could be implemented once the model is more mature and we have received more feedback from potential users to understand their needs better. Regarding pre-configured scenarios, the current tutorial shows how to generate repeatable scenarios, which, in our opinion, is the most important aspect. Still, we could include a second tutorial showing how to store these scenarios in data formats accessible from GUI and non-programming platforms.
With respect to the improvements in clarity, we will improve the description of the processes mentioned by the reviewer and include a new flowchart focused on phenological and trait generation. We will also include colors to facilitate the reading of the current flowcharts, a legend in Fig. 2, and define the acronyms in the figures captions. We will also include a glossary of terms as in a new table or a box.
Regarding validation, we thank the reviewer for the suggestion. However, while validation is important, please notice that BOSSE generates the spatial distribution of species and their traits in a fully randomized way, constrained by the meteorological data and the plant functional types to which each species is assigned. This feature allows for generating multiple scenarios but also makes comparing the simulations with observational datasets difficult. Validation would require having spatialized information of species distributions and a large set of their traits and biophysical and phenological properties so that the dynamics of other observables (e.g., remote sensing imagery or eddy covariance fluxes) could be compared with the simulations. However, this is not trivial, and it is precisely the lack of such information at wide scales that motivated the development of BOSSE. On the one hand, a partial lack of this information could compromise the comparison, making it impossible to understand whether the differences between simulations and observations are due to uncertainties in the unknown input parameters. On the other hand, these variables could be estimated by data assimilation, but then the validation would be circular, since it would depend on the fit and equifinality of the constrained solutions of the inverse problem. BOSSE relies on validated or at least tested models (e.g., SCOPE). Partial validations (e.g., validating only one scene) would be just validations of these models, which is not the aim of this work. Another key element is the distributions and ranges of the biophysical parameters simulated. While literature and observational data inform these, we will include in the Jupyter notebook tutorial a section dedicated to how to modify these thresholds for a scene, so that the simulations can be adjusted to more local or better-known conditions. We still acknowledge that randomness could lead to implausible simulations. For this reason, BOSSE is able to generate a large number of plots describing the temporal and spatial distribution of species traits and the sensitivity of the simulated species to meteorological variables for inspection, allowing the users to determine the plausibility or similarity of a particular scene with the behaviour of a specific study area or ecosystem type. Future developments of the model might allow for some tuning of the scene parameters to observations, but addressing this problem to provide an “improved” solution requires the BOSSE to mature and some understanding of the requirements from the community.
With respect to the emulation uncertainty, we have tested not only the uncertainty in the variables predicted by the emulators in fully independent “test” datasets, but we have also propagated these uncertainties to functional diversity metrics such as Rao Q or the variance-based partitioning of diversity for all of them. These results are presented in Table S6.1 of the supplementary material. We will stress this point in the main text of the manuscript. Notice that this is a step further with respect to former similar simulations in the context of plant functional diversity (e.g., Pacheco-Labrador et al., 2022, Ludwig et al., 2024), or other ecological applications (e.g., Katteborn et al. 2017).
Finally, computational feasibility is indeed an important point that BOSSE takes care of by replacing expensive models with emulators. Following the reviewer’s recommendation, we will provide an additional script to benchmark the computational cost on any machine prior to the large-scale use of the model to improve the user’s understanding of computational resources and time
References:
* Kattenborn, T., Fassnacht, F.E., Pierce, S., Lopatin, J., Grime, J.P., & Schmidtlein, S. (2017). Linking plant strategies and plant traits derived by radiative transfer modelling. Journal of Vegetation Science, 28, 717-727
* Ludwig, A., Doktor, D., & Feilhauer, H. (2024). Is spectral pixel-to-pixel variation a reliable indicator of grassland biodiversity? A systematic assessment of the spectral variation hypothesis using spatial simulation experiments. Remote Sensing of Environment, 302, 113988
* Pacheco-Labrador, J., Migliavacca, M., Ma, X., Mahecha, M.D., Carvalhais, N., Weber, U., Benavides, R., Bouriaud, O., Barnoaiea, I., Coomes, D.A., Bohn, F.J., Kraemer, G., Heiden, U., Huth, A., & Wirth, C. (2022). Challenging the link between functional and spectral diversity with radiative transfer modeling and data. Remote Sensing of Environment, 280, 113170
Citation: https://doi.org/10.5194/egusphere-2025-318-AC2
-
AC2: 'Reply on RC2', Javier Pacheco-Labrador, 13 May 2025
Data sets
The Biodiversity Observing System Simulation Experiment (BOSSE v1.0) ERA5Land Meteorological time series Javier Pacheco-Labrador, Ulirch Webber, Ulisse Gomarasca, Daniel E. Pabon-Moreno, Wantong Li, Mirco Migliavacca, Martin Jung, and Gregory Duveiller https://doi.org/10.5281/zenodo.14717038
Model code and software
pyBOSSE Javier Pacheco-Labrador https://github.com/JavierPachecoLabrador/pyBOSSE
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 647 | 207 | 27 | 881 | 61 | 24 | 38 |
- HTML: 647
- PDF: 207
- XML: 27
- Total: 881
- Supplement: 61
- BibTeX: 24
- EndNote: 38
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The manuscript provides a dynamic implementation of surface reflectances using a modeling framework to apply to biodiversity questions. The approach is described well, including with references to documented code environments. The description is fairly technical, following more as an operators guide for the model, but this is balanced toward the end with descriptions for model performance.
Two minor comments. First, there is no mention of bidirectional reflectance calculations other than a quick mention of solar angles. How would this modeling approach handle view angles to more appropriately integrate with remote sensing observations, and how would BRDF be estimated or what limitations are there. Second, it wasn't clear to me whether the modeling approach for estimating reflectances/emissivities can handle mixed species/PFTs and how the mixing is carried out.
Also, please see similar approach in development (no need to cite) -> https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2022JG006935