| 16 Sep 2025
Soil moisture-induced changes in land carbon sink projections in CMIP6
Abstract. The terrestrial biosphere absorbs about one third of anthropogenic CO2 emissions, thereby significantly slowing human–induced climate change. Its capacity to act as a carbon sink strongly depends on climate conditions, particularly soil moisture (SM), which can constrain plant growth and amplify land–atmosphere feedbacks. Therefore, accurately capturing these effects in Earth System Models (ESMs) is critical.
Using dedicated experiments of the Land Feedback Intercomparison Project (LFMIP, an experiment within the Land Surface, Snow, and Soil Moisture Model Intercomparison Project, LS3MIP) from the latest generation of ESMs from the Coupled Model Intercomparison Project Phase 6 (CMIP6), we show that projected SM changes substantially reduce the land carbon sink by the end of the century (2070–2099). This reduction is mainly driven by SM variability, highlighting the importance of SM extremes, which are projected to become more frequent and intense under climate change. Our results confirm those of the previous model generation based on the Global Land-Atmosphere Climate Experiment–Coupled Model Intercomparison Project phase 5 (GLACE–CMIP5). The results show that the strong negative impact of SM changes on the land carbon sink shown for GLACE–CMIP5 is less severe in LFMIP. A more in–depth analysis reveals that this is due at least in part to the specific ESM sampling of the respective experiments, with participating ESMs from CMIP5 generally showing a stronger drying trend. Despite agreement on the negative impact of SM on the land carbon sink in most tropical and mid–latitude ecosystems in both sets of multi–model experiments, there are large intermodel differences in the projected magnitudes.
As SM can influence land carbon uptake both directly and indirectly via land–atmosphere coupling, we conduct a contribution analysis on the impact of direct and indirect SM effects on carbon uptake, which reveals that SM–atmosphere interaction dominate SM–induced changes globally. However, models show disagreement on the magnitude of these effects. Intermodel differences arise mainly from varying sensitivities of GPP to SM–related direct and indirect effects, suggesting that differences likely stem from varying representations of water–stress related processes across ESMs.
Our findings highlight SM–atmosphere coupling as a critical factor for future land carbon uptake. Improving the representation of water stress processes, plant hydraulics, and vegetation characteristics in ESMs is essential for reducing uncertainty in projections. Maintaining and possibly extending the experimental set up to a larger set of models in future CMIP generations will be key to advancing our understanding of SM–carbon interactions and consequently of the evolution of the land carbon sink under human–induced climate change.
Overarching comments:
This paper address water-carbon coupling in models leveraging soil-moisture sensitivity runs through the LFMIP experiments. The topic is important to the modeling community, and the work presented here is thorough and yields many good insights. I congratulate the authors on pulling together what is clearly a lot of work. However, this was a difficult read and left many points of confusion; I suggest there is room to further develop the presentation for clarity, primarily through 1) the inclusion of clear guiding hypotheses/questions that motivate and structure the analysis; 2) more consolidation of uncertainty information in the figures; and 3) clearer organization of information throughout.
Some specific comments are below.
L 30: Please define the sign of land source/sink in this terminology. For example., “The net carbon exchange on large spatial and temporal scales is referred to as Net Biome Production (NBP), where positive values indicate a net flux of carbon from atmosphere to land (sink) and negative values indicate a net flux of carbon from land to atmosphere (source)."
L 55: Although later in the paper (L 130) the focus on GPP over respiration is asserted, it is worth mentioning the important and variable impacts of SM on microbial respiration, CO2/CH4 partitioning, etc, as the context is set in the introduction. This is a major part of the hydrological feedback puzzle especially at high latitudes where, in addition, total soil C is highly uncertain.
L66-75: Please introduce some hypotheses or questions here that can be addressed. This will make the results section far easier to read and greatly improve the presentation.
L 85: Please provide a) some brief acknowledgement that there’s no water budget closure when SM levels are prescribed as such and b) brief explanation on why the logic and feedback physics of these sensitivity experiments nonetheless works for the variables of interest.
L125: It will be clearer if the content of this sentence can be reformatted as equations:
“The experiments of LFMIP allow isolating the effect of SM trend and variability, where ∆NBPSM trend =
∆NBPrmLC−pdLC and ∆NBPSM var = ∆NBPCT L−rmLC , as well as the combined effects of SM expressed as ∆NBPSM all =
∆NBPSM trend + ∆NBPSM var = ∆NBPCT L−pdLC.”
L145: If focusing on total SM, the methods set up can be simplified for the aid of the reader by limiting description/equations to just the pdLC and CTR experiments.
L175: (related to comment on L145) Yet, if there is substantial discussion of SMtrend and SMvar, the authors should leave these definitions in. But then I would rephrase this to say that the other results are primarily in the supplement but discussed throughout.
L184-186: negative signs are redundant with phrasing: “reducing by…”
Figure 1: When suggesting a comparison to Green et al 2019 figure 1, please provide more direction on what comparison should be made. For instance, is Figure 1 panel b meant to be a reproduction (over a different time period) of data in Green et al 2019 figure 1? If so, this is confusing, as the black, blue and pink lines do not have the same decadal-scale dynamics—this paper’s lines reach a minimum in ~2020 while Green et al 2019 lines have an approximately monotonic increase. What accounts for the difference, how should we understand this comparison?
L 200-210/figure 3 (also applies to figure 2): This analysis would be enhanced by presentation of uncertainty in the figures as follows: a) in the model-specific plots, some indication on where the trends are insignificant with stippling or shading. Without this it is hard to assess in z-score space what is going in in the Sahara—some of the Sahel SM increase in CMCC-ESM2 may be “real” accompanying Sahel greening, but some of the other widespread changes over the desert (e.g. IPSL-CM6A-LR) may be noise. Similarly, b) some indication in the ensemble mean plots where there is sign disagreement within the 4 models would greatly enhance the interpretation of the visuals. The IPCC reports have good inspiration on how to make these multi-map plots more useful by conveying this type of uncertainty information.
L 256-265: This is not adequately interpreted. A) It would be helpful to state the hypotheses guiding this analysis. Related to that, B) to help us understand a statement such as “the results indicate that about 70–90 % of intermodel difference can be explained by either changes in the direct and indirect SM effects or the sensitivity of GPP to those effects” please list out what the possible contributors are, and how this relates to the guiding hypotheses. When I dig back to Eq4 and line 156, this statement appears to suggest that 70-90% of the effects are accounted for by “everything” which is a trivial conclusion. Please make this clearer for the reader. C) Please correct North South America D) 263-265 suggest to put this sentence higher in the paragraph, this is where the real conclusions are.
Overall discussion: This discussion has several good insights but more organization of information is needed to make this compelling and easy to digest. For example, I could see a header “Are the models fit for purpose?” and another one “Are the experiments fit for purpose?” to help frame the current discussion of model process richness and drought impact performance and the inclusion of SM extremes, respectively, alongside some indication in the opening discussion paragraph that the authors intend to explore these ideas? This is just an example to illustrate the suggestion of how to organize the information, the authors can of course approach this how they want.
L 276-277: “The latitudinal NBP of the northern mid–latitudes accounts for about 85 % of global NBP for the CMIP5 MMM.” Can this be rephrased? I don’t know what is meant here.
L 281-294 is informative but needs to be contextualized—currently the authors make a strong case these models are not fit to capture the key feedbacks. Including an objective model evaluation such as an ILAMB broadscore plot would help readers understand “who” these models are. The authors mention that the GFDL model has the best lagged responses, but does it outperform the other models on the benchmarks? This also pertains to the discussion on L323-333: The authors mention that CLM5 is the most process rich in the plant hydraulic space, but does it outperform other models in the benchmarks?