Evaluating the carbon and nitrogen cycles of the QUINCY terrestrial biosphere model using remotely-sensed data

Miinalainen, Tuuli; Ojasalo, Amanda; Croft, Holly; Aurela, Mika; Peltoniemi, Mikko; Caldararu, Silvia; Zaehle, Sönke; Thum, Tea

doi:https://doi.org/10.5194/egusphere-2025-2987

Preprints

https://doi.org/10.5194/egusphere-2025-2987

Preprints

10 Jul 2025

| 10 Jul 2025

Evaluating the carbon and nitrogen cycles of the QUINCY terrestrial biosphere model using remotely-sensed data

Tuuli Miinalainen, Amanda Ojasalo, Holly Croft, Mika Aurela, Mikko Peltoniemi, Silvia Caldararu, Sönke Zaehle, and Tea Thum

Abstract. Accurate estimates of future land carbon sinks and thus the remaining carbon budget to achieve the Paris climate goals requires rigorous modelling of the carbon sequestration potential of the terrestrial biosphere. Estimating the terrestrial carbon budget requires an accurate understanding of the interlinkages between the land carbon and nitrogen cycles, yet coupled carbon-nitrogen cycle models exhibit large uncertainties. Leaf chlorophyll, chl_leaf, is an indicator of the leaf nitrogen content stored within photosynthetic nitrogen pools and is central to the exchange of carbon, water and energy between the biosphere and the atmosphere. In this work, we harness an advanced remote sensing (RS) chlleaf product to evaluate a terrestrial biosphere model, QUantifying Interactions between terrestrial Nutrient CYcles and the climate system (QUINCY), which explicitly models chl_leaf. We focus on comparing the spatial and seasonal patterns of modelled and observed chl_leaf, and then further assessing if modelled leaf area and productivity agree with a RS leaf area index product and in-situ eddy covariance-based gross primary production, respectively. In addition, we conduct additional simulations to test two alternative formulations of leaf-internal nitrogen allocation within QUINCY. Our analysis over a globally representative set of locations reveals that QUINCY chl_leaf magnitudes are mostly in line with the RS chlleaf values. However, QUINCY chl_leaf tends to show a narrower numerical range compared to RS for specific ecosystem types, such as grasslands. While the seasonal cycle of QUINCY chl_leaf mostly corresponds well to the observations, for many deciduous forests, the increase in QUINCY’s chl_leaf predictions in spring and the decrease in autumn were delayed compared to observations. Our results also show that compared to the original leaf nitrogen allocation scheme of QUINCY, the revised scheme produced a more reasonable sensitivity of gross primary production to increases in chl_leaf. Our study shows the value of RS products linked to N cycle that will be useful in both C and N modelling, and paves way for closer linking of RS and TBMs.

Received: 23 Jun 2025 – Discussion started: 10 Jul 2025

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Preprint (PDF, 908 KB)

Supplement (1556 KB)

Download & links

Tuuli Miinalainen, Amanda Ojasalo, Holly Croft, Mika Aurela, Mikko Peltoniemi, Silvia Caldararu, Sönke Zaehle, and Tea Thum

Status: final response (author comments only)

RC1: 'Comment on egusphere-2025-2987', Anonymous Referee #1, 04 Aug 2025

This study is centered on evaluating the spatio-temporal behaviour of the terrestrial biosphere model (TBM) QUINCY, using mainly a remote sensing (RS) product of leaf chlorophyll content (RSchlleaf). The interest in the use of such a product is motivated by the strong link between leaf chlorophyll and Nitrogen contents. Other datasets are exploited such as RS LAI, and GPP at FLUXNET sites. The study looks at the agreement between simulated and observed chlleaf, evaluating the magnitude and the seasonal cycles across and within plant functional types (PFTs). Alternative algorithms of nitrogen use are also evaluated, and a statistical model is used to look at the main drivers of observed and simulated chlleaf.
The topic is quite innovative. The paper is well structured and clearly written. The methodology is sound, and the results are accurately described. The analysis of the seasonal cycles would gain in clarity from actual figures of the seasonal cycles. To enrich the temporal analysis, the authors could present boxplots of RMSD distribution for each PFT. The Discussion section would benefit from a more clearly defined structure. Even though the model results are not yet fully satisfactory, this paper nonetheless offers a valuable contribution to the field, and paves the way for a larger use of RS chlfleaf products among TBMs, both for evaluation and data assimilation purposes.

Main comments
Abstract.
- L14-16: ‘Our results also show that compared to the original leaf nitrogen allocation scheme of QUINCY, the revised scheme produced a more reasonable sensitivity of gross primary production to increases in chlleaf ‘-> Where is this demonstrated in your results?
1 Introduction
- L54: Krinner et al., 2005 -> The N version of ORCHIDEE is described in Vuichard et al. (2019).
Vuichard, N., Messina, P., Luyssaert, S., Guenet, B., Zaehle, S., Ghattas, J., ... & Peylin, P. (2019). Accounting for carbon and nitrogen interactions in the global terrestrial ecosystem model ORCHIDEE (trunk version, rev 4999): Multi-scale evaluation of gross primary production. Geoscientific Model Development, 12(11), 4751-4779.
- L55-56: ‘evaluating and validating’ -> What’s the difference for you?
2 Materials and methods
2.4 Remote sensing data
2.4.3 Post-processing of the RS data
- L241-243: ‘In addition, the RS chlleaf for the needle-leaved sites was multiplied by π/2. This was done to account for the half-hemispherical needle geometry in the remote sensing retrieval (Stenberg et al., 1995).’ -> This is weird, why was this correction not applied in the native algorithm of Croft et al. (2020)? How can the users of chlorophyll products know whether this has been considered? For example, what about the OLCI product, do you know if such a correction is integrated in the processing chain?
2.5 In-situ observations
2.5.1 Eddy covariance flux observations
- L256: ‘Data from all years were used, and therefore, the GPP time series are not from the same time interval as RS chlleaf.’ -> Can you comment on this discrepancy? What are the possible implications regarding the accuracy and robustness of your findings?
2.6 Feature importance analysis
- In this section we need more information on your training/test/evaluation datasets and the scores of your statistical models to trust your further analyses.
2.7 Data-analysis
- L326-327: ‘We used the 90^th percentile of LAI instead of the mean values to reduce the effect of seasonal variation.’ -> This is not clear, how is the mean value more impacted by the seasonal variation? What would be the impact on your study if you considered the mean LAI?
- L330 : ‘We analyzed the seasonal cycle of chlleaf , LAI and GPP for one specific site, Hainich in Germany (DE-Hai, 51.08◦N, 10.45◦E)’ -> Why not show mean seasonal cycles per PFT? This would be justified as many parameters are PFT-dependent.
- L334-336: ‘In addition, we calculated the average values over April, May, October and November for the PLUMBER2 TeBS NH sites for the QUINCY results and observations, to study the differences in seasonal development’. -> Again, why not study the complete seasonal cycle?
3 Results
3.1 Evaluation of simulated chlleaf , LAI and GPP against observations
- These first paragraphs should be grouped under a first subsection ‘3.1.1 Mean values’ (as opposed to the following subsection dealing with seasonal cycles).
- L353: ‘For chlleaf in all cases apart from TrBE and TrH, there is a lack of variation in the QUINCY chlleaf’ -> Looking at Table S5, TeBS seems to be quite good with QUINCY’s statistics being close to those of the RS product.
3.1.1 Seasonal cycle
- The whole description of Figure 3 is too vague, please provide some quantification to support your assertions.
- L383-384: ‘The annual cycle of chlleaf at the Hainich site (Fig. 3) is very similar when comparing QUINCY and RS’ -> This seems somewhat optimistic, even accounting for the delay, as the shapes are different. Please provide some statistics.
- L384-385: ‘the simulated LAI increases approximately 20 days later in spring compared to the RS LAI. The delay is even more pronounced for chlleaf’ -> This is not what is seen on the subplots of Figure 3: RS chlfeaf and LAI seem to be starting around DOY 80, and QUINCY chlfeaf and LAI around 120, and we don’t see any delay regarding chlleaf as compared to LAI, the delay vs the RS products seem to be around 40 days for both variables. To settle this, explain how you determine the start of season for model and observations, provide the dates, and add the corresponding lines on the figure.
- L388-390: ‘However, despite the fact that QUINCY chlleaf and LAI remain higher, their winter level is reached almost at the same time as in the Hainich observations, because the senescence occurs more rapidly in QUINCY than in the observations’ -> That might be overstated, explain how you determine the winter levels for model and observations, provide them and the dates when they are reached, add all the corresponding horizontal and vertical lines on the figure.
- L390: ‘Therefore, the overestimation in GPP is not as pronounced.’ -> Why “therefore”? I don’t see the link. Your later sentence on L395-396: ‘In addition, although the simulated LAI remains at the summer level until DOY ∼280, the simulated GPP decreases due to the environmental conditions’ is more convincing.
- L397-398: ‘For the TrH sites (Fig. 2i), the lowest PFT mean for QUINCY is in April, suggesting that the phenological cycle for these sites needs further tuning in QUINCY’ -> It’s not clear from the scatter plot what the seasonal cycles look like. It would be much easier to see the mean seasonal cycles for both model and observations than trying to imagine them from the scatter plot. Please add them, at least in supplementary, otherwise it’s very difficult to follow this paragraph. Also, you could add the corresponding RMSD and r.
- L402-403: ‘could be due to a drought’ -> Do you mean a modeled water stress? Please check whether this is the case in your simulations. Plus, wouldn’t that be the standard phenology for a tropical deciduous PFT, to be limited by water availability rather than temperature?
- L408: ‘The mean IAV’ -> Why do you present the mean IAV in the “Seasonal cycle” section, and not in the former one which presents the mean annual values?
3.2 Nitrogen limitations in QUINCY
- L455: ‘indicating that chlleaf is more influenced by other factors than leaf N levels, compared to BNE and TeBS’ -> Provide a few examples of these other potential factors.
- L478:479: ‘The US-MMS QUINCY leaf C:N is close to the lower leaf C:N threshold ‘-> You could provide the leaf C:N low and high thresholds per PFT in the supplementary material.
4 Discussion
4.1 QUINCY’s ability to reproduce chlleaf magnitude
- This section spans over two pages and lacks clear organization. Please add subsection titles to better guide readers through the logic of your argument.
- L569-570: ‘Overestimation of LAI can lead to too strong shading, which can result in too small GPP in lower canopy layers’ -> I’m not sure about that, it will depend on the radiative transfer model.
- L570-571: ‘In addition, the radiative transfer model might play a role in the underestimated GPP’ -> Why ‘in addition’, as this has a direct link to your former sentence?
- L596-598: ‘This indicates that the alternative N allocation scheme produces more in line with our current ecophysiological understanding of plant dynamics: increasing leaf N in chlleaf does not decrease other photosynthetic fractions , but more structural part (fN,struct)’ -> I’m not sure what ‘this’ refers to. While you argue that your new model is more realistic, it is unclear how this is supported by your results.
- L600-601: ‘QUINCY chlleaf for evergreen sites was driven by N deposition, with other environmental variables contributing less. The same was true for the RS chlleaf for BNE and TrBR, but not for TeNE’ ‑> Why is TrBR mentioned here, it’s not an evergreen PFT. Do you mean TrBE?
Code and data availability
- L774: ‘RS chlleaf by Croft et al. (2020) will be available by request from the authors.’ -> Why not make it available on a public repository? This would ensure broader use.

Technical comments
- L77-78: Evans and Clarke, 2018 -> Evans and Clarke, 2019. To be corrected also L79, 155, 862.
- L155: ‘Site description -> Description of the sites’
- L355: ‘for TeC and TeH sites, which gives’ -> ‘for TeC and TeH sites, which give’
- L359: ‘PFTs.Whilst’ -> ‘PFTs. Whilst’
- L378-379: ‘for the boreal and temperate evergreen sites (Fig. 2a,b,c,d)’ -> That would rather be ‘Fig2. a, c, f’.
- Figure 2e: ‘TrBR (1 sites)’ -> ‘TrBR (1 site)’
- L395: ‘between years the 2003–2011’ -> ‘between years 2003–2011’
- L401: ‘compared rest of the year’ -> ‘compared to the rest of the year’
- L405: ‘The April and May chlleaf values are mostly underestimated by QUINCY for chlleaf , LAI and GPP’ -> ‘The April and May values are mostly underestimated by QUINCY for chlleaf, LAI and GPP’
- L435: ‘Fig. S7 and S6’ -> ‘Fig. S7 and Table S6’
- L454: ‘p< 1×10⁻⁴⁰’ -> ‘p < 1×10⁻⁴⁰’
- L458: ‘For the TeBS site’-> ‘For the TeBS sites’
- Legend Figure5: ‘Teh’ -> ‘TeH'
- L486: ‘3.3 Alternative leaf N allocation scheme’ -> ‘3.3 Alternative leaf N allocation schemes’
- L496: ‘PFTS’ -> ‘PFTs’
- L567: ‘the QUINCY mean chlleaf is underestimated at majority of the the TeBS sites’ -> ‘the QUINCY mean chlleaf is underestimated at the majority of the TeBS sites’
- L579-580: ‘Another missing processes in are fertilization and management of croplands’ ->’ Other missing processes in QUINCY are fertilization and management of croplands’
- L596-598: ‘This indicates that the alternative N allocation scheme produces more in line with our current ecophysiological understanding of plant dynamics: increasing leaf N in chlleaf does not decrease other photosynthetic fractions , but more structural part (fN,struct)’ -> ‘This indicates that what the alternative N allocation scheme produces is more in line with our current ecophysiological understanding of plant dynamics: increasing leaf N in chlleaf does not decrease other photosynthetic fractions, but raher the structural part (fN,struct)’
- L599: ‘Our machine learning based analysis’ -> ‘Our machine learning-based analysis’
- L626-627: ‘The Sentinel-3 chlleaf shows the strongest seasonal cycle for the US-NR1 compared to other products’ -> ‘The Sentinel-3 chlleaf shows the strongest seasonal cycle at the US-NR1 site compared to other products’
- L633: ‘one of the state-of-the art TBMs that includes’ -> ‘one of the state-of-the art TBMs that include’
- L653: ‘overestimated at the certain sites’ -> ‘overestimated at certain sites’
- L679-680: ‘RS observations from the Sentinel-3 satellite could be included as it was tested for two sites’ -> ‘RS observations from the Sentinel-3 satellite could be included as they were tested for two sites’
- L692-693: ‘the low elevation angles of the sun, which limits the reliability of the measurements throughout the winter months and, in mid-winter, results in polar night’ -> ‘the low elevation angles of the sun, which limit the reliability of the measurements throughout the winter months and, in mid-winter, result in polar night’
- L708: ‘Similarly, RS inversion algorithm does not consider’ -> ‘Similarly, RS inversion algorithms do not consider’
- L710-711: ‘In addition, PFT can be a very broad category’ -> ‘In addition, a PFT can be a very broad category’
- L711-712: ‘different tree species may have different characteristics, which is taken into account in our PFT-based modeling scheme and parameterization’ -> ‘different tree species may have different characteristics, which is not taken into account in our PFT-based modeling scheme and parameterization
- L738-739: ‘However, this would be possible only if other TBMs to provide’ -> ‘However, this would be possible only if other TBMs were to provide’
- L743: ‘near-time’ -> Do you mean ‘near-real-time’?
- L768: ‘las access’ -> ‘last access’
Supplementary material
- ‘Table S2. Lanc cover’ -> ‘Table S2. Land cover’
- Table S5. chlleaf (μg cm⁻² -> chlleaf (μg cm⁻²)
- Table S7: ("Qdef."-> ("Qdef.")
- Legend Figure S2: ‘Subplot (h) has a different scale on the x- and y-axis than the other subplots’ -> Why h and not b or e, where values also are not higher than 60 mg cm^-2?

Citation: https://doi.org/10.5194/egusphere-2025-2987-RC1
RC2: 'Comment on egusphere-2025-2987', Anonymous Referee #2, 17 Aug 2025

Review report for "Evaluating the carbon and nitrogen cycles of the QUINCY terrestrial biosphere model using remotely-sensed data"
This manuscript (MS) addresses an interesting and relevant topic by combining remote sensing and terrestrial modeling to explore the leaf chlorophyll and its role in carbon and nitrogen cycles. The study makes use of a very comprehensive dataset for model evaluation and remote sensing analysis, covering both spatial and temporal perspectives, and presents a number of interesting results and discussion points.
However, the overall presentation requires improvement to enhance clarity and readability. In particular, the M&M section would benefit from a more structured and concise description of the datasets that the model uses as input or for comparison, as the current presentation feels somewhat disorganized. Similarly, some points in the Discussion appear scattered and not well-connected, which makes it difficult to follow the logical flow of the arguments.

Detailed comments:
Introduction:
L. 33: It appears that the sentence “TBMs use different modeling approaches to represent N limitation of photosynthesis” repeats the previous one.
L. 35: Why does increasing model complexity introduce additional uncertainties? In the previous sentence, you only mention different approaches that may lead to different predictions, but do not explain why greater complexity itself results in more uncertainty.
L. 76: Is this a modeled response? I recommend clarifying this point here, as the previous paragraph mentioned models and remote sensing, which may confuse readers.
M&M
L. 97-119: The description of the model feels somewhat fragmented. I suggest establishing a clearer logical link between N cycling, its relationship with leaf chlorophyll, photosynthesis, and photosynthesis-related parameters. In contrast, the detailed descriptions of processes such as N uptake functions, maintenance respiration functions, and soil pools/layers seem less relevant to the main focus on leaf chlorophyll, carbon, and nitrogen, from my point of view.
This leads to information spreading out, and the description of chlorophyll itself—just one sentence—seems insufficient.
L. 120-130: Here are considerable details on the start and end of the growing season. It would be helpful if the authors could clarify whether and how this influences the LAI simulation, which is the main output of focus. In lines 120 and 122, leaf biomass development and plant growth are mentioned—are these the same as, or directly related to, LAI development in the model? Overall, I suggest that the model description could be streamlined with a clearer emphasis on the outputs that are central to the study.
L. 165: What are “other issues”? I suggest avoiding such vague descriptions, and it would be better to specify what these issues are.
L167: The RS chlleaf data is available here, but you still retrieve the RS chlleaf data later for these sites? Which one do you use for analysis?
L177: Input is in which period?
L 183: In Section 2.2, you note that meteorological data are available for PLUMBER2, but there is no similar information provided for GLOBAL, making here a bit abrupt. It may help readers if you clearly specify, for each site set, which input datasets are available. Additionally, summarizing all available datasets used for either model input or comparison—including remote sensing and in-situ data—in a table could make the information clearer and improve the overall clarity, as you have various and large amounts of datasets.
L190&192: How do you deal with the N and P deposition input in spin-up and simulations?
L 192: I cannot understand the sentence: this was continued ...; what are the respective years?
L259: Why do you specifically focus on evergreen needle-leaved forests? Does it mean that model simulations were also performed for these two sites? If so, this step is not described in the Model Simulation section.
Overall, to improve clarity, I would suggest first describing all the datasets available at each site—for model input and model comparison—before explaining the model simulations. This would make the overall workflow easier for the reader to follow.
Sect. 2.7: It seems you did not describe your data analysis for the two sites with in-situ chlleaf data? And the mention of additionally analyzing data for the Hainich site is a bit abrupt. Why do you want to investigate this site?

Results
L420: Since the years of remote sensing data and model simulations do not fully match, whether this might affect the comparison in terms of magnitude and pattern? Additionally, averaging over several years may reduce the apparent seasonal variation. Have you looked at the seasonal patterns for each year individually?

Discussion
The study analyzes a large amount of data from many perspectives and therefore raises a number of valuable discussion points. However, the discussion as a whole feels somewhat scattered (except 4.2).
For example, in Sect. 4.1, you started with chlleaf-GPP-LAI, but in line 554 you mix the discussion of GPP and LAI, and then in the following paragraph, you return to GPP simulations, which were already discussed in the second paragraph. While I understand that these variables are closely related, this back-and-forth may confuse readers and make it difficult to grasp the main points.
Secondly, the discussion of the relation between chlleaf and GPP residuals appears in L568 and L654.
And in 4.4, you mentioned the limitation from RS to the model, and again back to RS and the flux tower in L714.
I suggest considering summarizing and reorganizing the discussion points to make the overall discussion clearer and more structured.
L560: In addition, the sudden focus on a specific site (Hainich forest) feels abrupt. I always have the question about the rationale for highlighting this site? For example, is it because the simulations or RS data for this PFT show a distinct pattern compared to others?
L627: Which assumptions?
L655: This argument appears somewhat abrupt and does not have a clear logical connection with the preceding or following sentences.
Technical corrections:
L496: PFTs
L505: due to

Citation: https://doi.org/10.5194/egusphere-2025-2987-RC2

Tuuli Miinalainen, Amanda Ojasalo, Holly Croft, Mika Aurela, Mikko Peltoniemi, Silvia Caldararu, Sönke Zaehle, and Tea Thum

Supplement

https://doi.org/10.5194/egusphere-2025-2987-supplement

Data sets

QUINCY simulation data Tuuli Miinalainen and Tea Thum https://fmi.b2share.csc.fi/records/6a3849a7694b4f4a9efba39abde734af

Tuuli Miinalainen, Amanda Ojasalo, Holly Croft, Mika Aurela, Mikko Peltoniemi, Silvia Caldararu, Sönke Zaehle, and Tea Thum

Viewed

Total article views: 367 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
304	52	11	367	22	7	11

HTML: 304
PDF: 52
XML: 11
Total: 367
Supplement: 22
BibTeX: 7
EndNote: 11

Views and downloads (calculated since 10 Jul 2025)

Month	HTML	PDF	XML	Total
Jul 2025	96	29	6	131
Aug 2025	149	23	4	176
Sep 2025	59	1	60

Cumulative views and downloads (calculated since 10 Jul 2025)

Month	HTML	PDF	XML	Total
Jul 2025	96	29	6	131
Aug 2025	149	23	4	176
Sep 2025	59	1	60

Viewed (geographical distribution)

Total article views: 357 (including HTML, PDF, and XML) Thereof 357 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 09 Sep 2025

Short summary

Estimating the future carbon budget requires an accurate understanding of the interlinkages between the land carbon and nitrogen cycles. We use a remote sensing leaf chlorophyll product to evaluate a terrestrial biosphere model, QUINCY (QUantifying Interactions between terrestrial Nutrient CYcles and the climate system). Our study showcases how the latest advancements in remote sensing-based vegetation monitoring can be harnessed for improving and evaluating process-based vegetation models.


Total:	0
HTML:	0
PDF:	0
XML:	0