the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Improved representation of phosphorus exchange on soil mineral surfaces reduces estimates of P limitation in temperate forest ecosystems
- 1Centre for Environmental and Climate Science, Lund University, Sölvegatan 37223 62 Lund, Sweden
- 2Max Planck Institute for Biogeochemistry, Hans-Knoell-Str. 10, Jena, 07745, Germany
- 3International Max Planck Research School (IMPRS) for Global Biogeochemical Cycles, Jena, 07745, Germany
- 4Agroscope, Reckenholzstrasse 191, 8046 Zürich, Switzerland
- 5Eco&Sols, Institut Agro, CIRAD, INRA, IRD, Place Viala 34060 Montpellier cedex 2, France
- 1Centre for Environmental and Climate Science, Lund University, Sölvegatan 37223 62 Lund, Sweden
- 2Max Planck Institute for Biogeochemistry, Hans-Knoell-Str. 10, Jena, 07745, Germany
- 3International Max Planck Research School (IMPRS) for Global Biogeochemical Cycles, Jena, 07745, Germany
- 4Agroscope, Reckenholzstrasse 191, 8046 Zürich, Switzerland
- 5Eco&Sols, Institut Agro, CIRAD, INRA, IRD, Place Viala 34060 Montpellier cedex 2, France
Abstract. Phosphorus (P) availability affects the response of terrestrial ecosystems to environmental and climate change (e.g. elevated CO2), yet the magnitude of this effect remains uncertain. This uncertainty arises mainly from a lack of quantitative understanding of the soil biological and geochemical P cycling processes, particularly the P exchange with soil mineral surfaces, which is often described by a Langmuir sorption isotherm. We first conducted a literature review on P sorption experiments and terrestrial biosphere models (TBMs) using Langmuir isotherm.
We then developed a new algorithm to describe the inorganic P exchange between soil solution and soil matrix based on the double-surface Langmuir isotherm and extracted empirical equations to calculate the sorption capacity and Langmuir coefficient. We finally tested the conventional and new models of P sorption at five beech forest sites in Germany along a soil P stock gradient using the QUINCY (QUantifying Interactions between terrestrial Nutrient CYcles and the climate system) TBM.
We found that the conventional (single-surface) Langmuir isotherm approach in most TBMs largely differed from P sorption experiments regarding the sorption capacities and Langmuir coefficients, and simulated a too low soil P buffering capacity. Conversely, the double-surface Langmuir isotherm approach adequately reproduced the observed patterns of soil inorganic P pools. The better representation of inorganic P cycling using the double Langmuir approach also improved simulated foliar N and P concentrations, and the patterns of gross primary production and vegetation carbon across the soil P gradient. The novel model generally reduces the estimates of P limitation compared to the conventional model, particularly at the low-P site, as the model constraint of slow inorganic P exchange on plant productivity is reduced.
- Preprint
(1835 KB) -
Supplement
(2160 KB) - BibTeX
- EndNote
Lin Yu et al.
Status: open (until 12 Jul 2022)
-
RC1: 'Comment on bg-2022-114', Hongxing He, 22 Jun 2022
reply
In this study, Lin et al presented a double surface Langmuir adsorption isotherm in the QUINCY model and compared it to the traditional/simplified (single surface) Langmuir isotherm, that is mostly used in current TBMs. After model calibration, the authors argue the double isotherm shows a better representation of the inorganic P cycling. The improved P (ad)sorption model also suggests the current assumed P limitation in temperature forests was likely overestimated. Overall, the paper is well written, and the results are also sound. Given our current understanding of P limitation is still very limited, thus before the P models can be applied to make sound predictions, the model structure needs to be evaluated and discussed. Study like this paper thus contributes to improving the process description of P exchange in forest ecosystems and advances in C-P coupling in TBMs. I would thus recommend publishing this work at Biogeosciences. However, I have the following comments for the authors to consider during their revisions.
First, the authors did a literature survey to highlight uncertainties of the current parameterization of Langmuir isotherm in some TBMs. The parameter error in some previous TBMs was also noticed by myself when I develop a recent ecosystem CNP model. Thus I think it is important to highlight this for the community and a very good motivation for the current study. However, one would ask if this is just a parameterization issue or if it is a model structure issue (as the authors argued here)? Empirical data that fit different isotherm functions, including traditional Langmuir isotherm, generally show various but reasonably well-fit results (e.g. Brenner et al 2019, Lin et al. 2020, and much more). Thus, I would like to discuss this with the authors. First, which part of the improved model fits the measured data that could be attributed to the model structure, which part is from improved parameters? I have concerns about how the model comparison is made and how much conclusions can be drawn from such a comparison? In Line 138, the authors state separate calibrations were made for each site and each depth, what’s the influence of those separate calibrations for the comparison?
Second, as the authors argue the advantage of using double surface Langmuir, i.e. its higher buffer capacity. Then I would suggest a better separation of the influence of P release from other releases and uptakes? i.e. the feedback is of need. From the results, the main improvement is the ratio between Plab and exchangeable Pi (section 3.2). The P uptake across models seems rather similar (Fig. S2), i.e. for the P limited site LUE, the uptake PO4 for siLang, dbLang and 4pool model (Fig. 2Sf). The siLang shows a higher uptake in autumn but at an annual scale, the overall rates seem rather similar. The different approaches show a rather large influence on the C partitioning, (LAI, aboveground C, Fig. S2 bc). This is rather strange, what causes such large feedback on the aboveground plant properties, given the total P uptake seems rather similar? I also do not find evidence to support the statements of Line 203, i.e. the differing plant P uptake.
Third, the model performance of foliar P, Fig.4b shows a convergence of different models when P availability becomes smaller. In other words, in more P-limited conditions, the difference between the models becomes smaller, although all of them largely underestimated the measured P concentration. How come such large differences in the P-rich sites? Is this due to the calibration being mainly focused on the soil and thus less on the vegetation?
Some more specifics to consider:
Introduction
Line 36, missing references after “boreal forests are generally considered N limited”
Line 49-50, the argument is that organic P recycling is the major flux, while the geochemical P flux is small.
Line 50-55. In literature, several isotherms, or model functions, including double Langmuir, have been suggested to describe the phosphorus adsorption-desorption processes (i.e., McGechan and Lewis, 2002). I would also suggest not to use “a novel model concept, Line 54” as the authors propose in its current form. It has been in the literature for some time. I think the novelty is the implementation of the TBM models and evaluation of the implications? Besides, I am also lacking the field and experimental evidence to support the additional supplement of P from the adsorbed P pools. So, what is the, i.e. P isotopic data suggest, and do they support your hypothesis here? What are the mechanisms behind that? I would suggest adding those to the motivate current model development work.
Methods
Line 70, equ 1, the Langmuir isotherm, do the interaction with water considered? As the concentration also dependent on the water content at each time step?
Line 116 do you have leaf P/N concentration data over years? Or just sampled for one year?
Line 138 what is calibrated and what criteria were used for the calibration? Be specific here.
Results
Line 203 given the total P uptake by different approaches?
Line 246 the pool sizes differs also after the simulation, SOM top soil, the fluxes and the pools sizes. As also your sensitivity results show the importance of SOM pools for dbLang, Line 219, which indicates the potential feedback due to the biological mineralization. Also in your Line 243 on the plant and soil changes
So the different approaches show impacts on the fluxes and pool sizes. Why not show a complete P budget for each site with different fluxes simulated by various approaches? Also show the different pool sizes before the simulation and after the simulation, i.e. the pool size changes. This will give an overall picture of the ecosystems.
Some references mentioned:
McGechan and Lewis, 2002 presented an excellent review of the principles, equations, and models for the sorption of phosphorus. Biosystems Engineering 82 (1), 1-24.
Brenner, Julia, et al. 2019, Phosphorus sorption on tropical soils with relevance to Earth system model needs, Soil Research, 57, 17-27.
Lin Yang, et al 2020, Anoxic conditions maintained high phosphorus sorption in humid tropical forest soils, Biogeosciences, 17, 89-101.
Hongxing He
McGill, Montreal.
-
RC2: 'Comment on bg-2022-114', Anonymous Referee #2, 27 Jun 2022
reply
This paper by Yu et al. describes a new algorithm to better represent soil phosphorus sorption dynamics in a terrestrial biosphere model – QUINCY. The authors proposed the use of a double-surface Langmuir isotherm to better capture the non-linear relationships between solution P and labile P pools in the soil. They performed a review on both published data and model assumptions on P sorption. They then compared their simulation against data at a range of P availability sites, performed sensitivity on parameters and compared simulations for CO2 and P enrichment scenarios. They argued that the double-surface Langmuir isotherm is a better modeling scheme because it simulated observed pattern of soil organic pools well, it maintained a relatively stable solution P pool to act as a buffer against instability, which then led to less P limitation at the P-poor site, and it led to improved simulation of folia N and P concentration.
Overall, this is a clearly-written manuscript. The rationale and objectives are crystally clear. The discussion is also well written. My comments mostly focus on two aspects of the results that I want to discuss with the authors and receive their clarifications:
- Dd it indeed lead to improved estimate relative to the conventional single surface approach? All models performed well for reproducing the measured SOC etc. as reported in figure 3. The novelty of the double-surface scheme, as the authors argued, is that it better reproduced the ratio between Plab and exchangeable Pi (L190-191; Table 3). But looking at Table 3, the statistical significance is relatively weak ( p = 0.014 for lab-to-exchangeable P ratio, and 0.044 for SIP). At the same time, I wonder if the new scheme actually increase model complexity or not. May be the authors should make a paragraph discussing whether the gained benefits in terms of improved simulation accuracy is worth the added complexity, if there’s indeed additional complexity associated with the new scheme. In particular, does it require additional parameters relative to the conventional approach? And, if we want to constrain the parameters in the new model scheme, what data collection should we make? If it doesn’t involve additional complexity, I think it’s very useful to highlight.
- What does it mean for the land C sink estimates under future rising CO2 if the model simulated a less P limitation at the P-limited site. As the authors introduced, there has been a lot of model development to add P-cycle into models. The relative magnitude of the P limitation is obviously different, but one of the crucial argument for the inclusion of P-cycle in models is that they would impose additional processes to constrain ecosystem productivity for P-poor regions of the world. The new scheme seemed to alleviate the extent of P limitation, and therefore I wonder how does it compare to a simulation without the P-cycle turned on. Do you obtain similar CO2 responses for the P-limited site? Obviously the CN-only simulation does not have the capacity to accurate reflect the processes limiting CO2 responses at the P-poor site, but it would be interesting to see if there’s indeed difference between the two approaches.
Lastly, one question I have, which isn’t a criticism per se, is that why the authors didn’t use the Jena Soil Model (that they developed) to investigate the effects of different soil P sorption functions. It appears to be a soil question and therefore I wonder is there any particular reason that deemed JSM unsuitable for this work?
Specific comments:
Title: replace P with phosphorus.
Figures: Figure presentation needs to improve quite significantly. Figures are currently in low quality resolution. Units and variable names in the figure legend/axis need to be properly labeled.
For Figure 3, the authors may need to think of a better way to show the results, as it’s not very clear to see differences in panel a and b (but maybe because they are similar and therefore it’s not important to show the differences?). But still, the 4pool and control color are too hard to differentiate from each other.
Table 3. One could argue the statistical significance is very weak.
Lin Yu et al.
Lin Yu et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 183 | 46 | 8 | 237 | 10 | 1 | 1 |
- HTML: 183
- PDF: 46
- XML: 8
- Total: 237
- Supplement: 10
- BibTeX: 1
- EndNote: 1
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1