the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 Apr 2026
TIPMIP-OCEAN experimental protocol phase 1: Tipping dynamics of the AMOC
Abstract. This paper describes the experimental protocol for a set of coordinated simulations involving oceanic surface freshwater flux perturbations, conducted as part of the international Tipping Points Modelling Intercomparison Project (TIPMIP). These simulations constitute the first phase of the TIPMIP-OCEAN domain. We propose this protocol for inclusion in the Coupled Model Intercomparison Project Phase 7 (CMIP7), although it can also be implemented within CMIP6+ or other types of coupled or ocean standalone models. This initial phase focuses primarily on the dynamics of the North Atlantic Ocean, particularly the Atlantic Meridional Overturning Circulation (AMOC). The different experiments are designed to (i) evaluate the impacts of a potential major AMOC weakening under a 2 °C global warming scenario, (ii) assess the sensitivity of the AMOC to combined global warming and freshwater forcing, (iii) investigate the potential recovery of the AMOC following the reversal of forcings, and (iv) compare past AMOC variations with available climate observations and reconstructions. Four categories of experiments are included. Experiment group A examines the effect of freshwater release around Greenland under ramp-up, stabilization, and ramp-down scenarios in both CO2 emissions and freshwater input. Experiment group B complements this idealized set by using historical climate simulations and projections for 1850–2100, incorporating realistic estimates of Greenland Ice Sheet melt based on observations for the historical period and ice-sheet model projections for the future. Experiment group C extends the existing North Atlantic Hosing Model Intercomparison Project (NAHosMIP) by applying large freshwater perturbations to both control and 2 °C-warming simulations to assess how global warming influences AMOC reversibility. Finally, experiment group D imposes freshwater inputs, consistent with those inferred for the 8.2 kyr before present event, under pre-industrial conditions, in order to constrain model sensitivity to freshwater forcing using paleoclimate reconstructions. Together, these coordinated experiments will allow systematic evaluation of how different climate models respond to identical freshwater perturbations—an essential step toward better understanding the wide inter-model spread in North Atlantic dynamics and projected future AMOC changes.
- Preprint
(1473 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2026-1698', Anonymous Referee #1, 01 May 2026
- AC1: 'Reply on RC1', Didier Swingedouw, 31 Jul 2026
-
RC2: 'Comment on egusphere-2026-1698', Anonymous Referee #2, 27 May 2026
This paper describes the experimental protocol for a set of coordinated simulations involving oceanic surface freshwater flux perturbations, conducted as part of the international Tipping Points Modelling Intercomparison Project (TIPMIP). As such, it is not a ‘normal research paper’, but instead provides details so that research group can carry out the needed experiments to contribute to the TIPMIP project. Given the numerical focus of TIPMIP, I do feel it is appropriate for GMD. In fact, I am happy to see such protocols published in the literature rather than just on a webpage or a hard to find report. Making the protocol easily accessible is of value. Additionally, given the amount of work needed to define the protocol, it is appropriate that the authors get the credit for a peer reviewed paper. After reviewing the draft, I would recommend minor revisions. Although generally clear and well presented, there are places where material could be improved. My specific comments are given below.
I prefer the term forced ocean models to ocean standalone models. Especially as it likely that no model is actually ocean only, instead including at least a sea-ice component.Line 79 – Likely newer references to the AMOC than just a 2016 paper. Especially given all the work on the topic.
Line 80 – Given the recent OSNAP results, this general statement of regions could be modernized and more clearly explained. And in general, more literature on the AMOC and its stability would be good. I’d also like to see more background on freshwater sources, and discussion of Greenland melt, versus sources of freshwater to the Arctic (including rivers) as well as the impact of fluxes through the gateway straits. For a freshwater hosing protocol, the key background concepts need to be presented.
Line 81 – Doesn’t a significant amount of the AMOC waters upwell in the Atlantic, before it leaves for the other basins?
The discussion of the 8.2 Kyr event focuses on discharge of the glacial lakes through the St. Lawrence into the Atlantic. Yet there is a literature that suggests that there is some debate on where and how the freshwater is released, such as into the McKenzie river into the Arctic Ocean.
Even if the authors want to use a Labrador Sea focused protocol, a proper discussion of the freshwater sources to the event should at least be referenced – and whether it leaves by Hudson Strait or the St. Lawrence River likely may mean different freshwater distributions. For example, might the whole sub-polar gyre be better than just the Labrador Sea, as the freshwater likely has to flow around the gyre to reach the West Greenland Current region. Or why not along the Labrador Coast – high resolution studies have suggested that leads to different freshwater fates for this event.
I didn’t notice any discussion about model resolution in the protocol, and wonder whether that is relevant. That may impact the distance criteria to input the freshwater and thus where the freshwater is input in models with such differences in resolution. At the very least, it may impact the locations of discharge around Greenland, especially if any fiords are at least partially resolved. How might that impact the input of freshwater?
For experiment C, there is no thought on considering the rivers that may provide the enhanced precipitation in a future climate?
Shouldn’t there be an optional for seasonally varying the freshwater input?
What about for models that include icebergs? How shall the discharge be partitioned between solid and liquid components?
For experiment B, Tier 1, what initial conditions are recommended for the 1920s? Wouldn’t it make sense to be consistent, as the choice of initial conditions may impact the model results.
Figure 6. Why use the same dark blue for both fluxes in the ramp up and ramp down scenarios? A slightly different shade might help show that there are two scenarios.
Citation: https://doi.org/10.5194/egusphere-2026-1698-RC2 - AC2: 'Reply on RC2', Didier Swingedouw, 31 Jul 2026
-
RC3: 'Comment on egusphere-2026-1698', Anonymous Referee #3, 04 Jun 2026
Review comments on “TIPMIP-OCEAN experimental protocol phase 1: Tipping dynamics of the AMOC” by D. Swingedouw et al.
The authors present a protocol for the ocean component of TIPMIP that focuses on the stability of the AMOC. The MIP proposes a new set of fresh water hosing experiments that partly complement the core TIPMIP experiments and partly extend earlier hosing experiments such as the NAHosMIP.
Many previous community-driven hosing MIPs have focused on adding fresh water to the ocean while keeping the background to the pre-industrial state. The novel aspect of the proposal here is that they add the hosing to climate change simulations or different background states. This is an important aspect as the background climate may influence the AMOC stability and it is also more relevant for the ongoing discussion on the effects of climate change on the AMOC and its consequences for the north Atlantic realm. The MIP proposes a relatively wide range of freshwater applications, from an addition of a “realistic” amount of meltwater from Greenland (exp. B) to very strong, idealized scenarios with (unrealistic) additions of up to 0.6Sv. Thus, the potential analyses can include uncertainty estimates and also address issues of structural differences in the models.
The protocol also includes an extension of the NAHosMIP with relatively simplified set-ups and an experiment with a paleo-backdrop, mimicking the 8.2 kyr event.
The manuscript reflects an extensive discussion in the community and is well balanced so that many modelling groups can participate, either in the full set or selected sub-experiments.
The manuscript is written in a clear and concise way and should be published in GMD.
I have only a few minor issues to consider:
Abstract:
Ln 60: In experiment C, the run at 2 degr. Warmer climate is only tier 2, so it cannot be the general purpose of this experiment to study the effect of background warming. I would formulate it more like Exp. C is extending the NAHosMIP to include the TIPMIP models with the potential to include the global warming aspect.
Ln 80ff: I would support Referee#2 here and include the “post OSNAP” discussion on the relative roles of deep water formation regions and that their representation is certainly model and resolution dependent.
Ln 155 constant background
Ln 207: is the precip bias issue really relevant here?
Ln 219: wasn’t that originally proposed by Gerdes et al (https://doi.org/10.1016/j.ocemod.2005.08.003)
Ln 240: it would be good to clarify here that some experiments are required in parallel to the core TIPMIP experiments, but for B and C, they can (at least for tier 1) be run in parallel to existing control or historical/scenario simulations.
Ln 373: is “pursued” the correct word here?
Ln 509: …testetd a range of freshwater range from 0 to 0.6 Sv ( no need to mention tier 2 here)
The experiment tables could be made a bit more consistent, e.g. include the initial conditions for all.
Citation: https://doi.org/10.5194/egusphere-2026-1698-RC3 - AC3: 'Reply on RC3', Didier Swingedouw, 31 Jul 2026
Status: closed
-
RC1: 'Comment on egusphere-2026-1698', Anonymous Referee #1, 01 May 2026
The manuscript by Swingedouw and co-authors describes an experimental protocol to systematically analyse AMOC sensitivity under external forcing from both freshwater input and climate change. The authors propose four interesting experiments, which will primarily be performed with Earth System Models (CMIP6+), and which are well suited to address the main scientific questions. The overarching aim is to obtain a better understanding of AMOC sensitivity and, in doing so, to assess AMOC-related impacts that are relevant for society.
I strongly recommend this manuscript for publication in Geoscientific Model Development (GMD). The experimental design is well motivated and clearly presented, with supporting figures and tables. This TIPMIP-OCEAN protocol will provide valuable information for the AMOC community and for assessing the risk of AMOC tipping under future climate change.
That being said, I have a few moderate and minor comments that should be addressed prior to publication.
Moderate comments and suggestions:
- General aim of Experiments A, and specifically the reversibility component (Lines 191–192): I find the proposed experiments useful for studying AMOC weakening due to the joint contributions of global warming and freshwater input. The imposed forcing leads to a “major AMOC weakening” (Line 548; could this be quantified?) during the first 150 model years. Thereafter, forcing conditions are ramped down over 100 years and continued for an additional 100 years under no external forcing.
Given the typical multi-centennial adjustment timescale of the AMOC (Bonan et al., 2022, already cited) and results from existing pulse-forcing experiments (Jackson et al., 2023, already cited; van Westen et al., 2025, https://doi.org/10.1029/2025JC022651), it is unlikely that a coherent picture of AMOC reversibility can be obtained from the outlined experimental design (Fig. 2). Addressing reversibility would require substantially longer simulations for experiment “Stabilized (Ad)”, which are not mandatory under Tier 1 (Table 1). The 250-year experiment “Stabilized (Ab)” is likely sufficiently long in duration, but because it includes a modified background forcing, it cannot be used to assess reversibility.
My main point is that some aspects of AMOC stability cannot be addressed using simulations that are shorter than the AMOC adjustment timescale. This also applies to Experiments C and D, where the minimum required “stabilization” periods are 50 and 100 years, respectively. Only if modelling groups are willing to extend the stabilization (i.e. no external forcing) phases beyond 100 years can these questions be meaningfully addressed.
I would therefore suggest using more careful language and tempering expectations regarding AMOC reversibility, which is one of the central scientific questions in the TIPMIP-OCEAN protocol (Line 551). - The proposed experiments involve relatively rapid changes in forcing conditions (typically 100 years), which are short compared to the multi-centennial adjustment timescale of the AMOC. As a result, the simulated AMOC responses are likely to be dominated by the imposed forcing rather than by intrinsic AMOC dynamics. While the forcing may still trigger destabilising feedbacks (e.g., salt-advection), the experimental design appears less suitable for investigating true “tipping dynamics of the AMOC”, as suggested by the manuscript’s title. I therefore suggest reconsidering the title, for example:“Tipping sensitivity of the AMOC.”
- I note that the manuscript appears to miss a substantial portion of relevant AMOC literature, despite the authors being highly active in this research area and publishing influential work. While the AMOC literature is rapidly evolving and extensive, I have included a number of relevant (non-exhaustive) recent studies in the minor comments below to help complement the discussion. I do not expect the authors to include all of the listed references, but I encourage them to consider the most relevant ones.
Minor comments and suggestions:
- Lines 68 – 69: I also recommend including the more recent studies by Armstrong McKay et al. (2022) and Wunderling et al. (2024), which are already cited in the manuscript.
- Line 74: The term “exact thresholds” is somewhat vague and would benefit from clarification; please specify what is meant here. In addition, it may be helpful to explicitly identify the relevant ocean tipping elements in this context, particularly the two major ones: the AMOC and the North Atlantic subpolar gyre.
- Line 77: Please write out AMOC here; this is the first time the acronym is mentioned in the main text.
- Lines 77 – 78: Either quantify “warm and salty” and “cold and fresh” or add ‘relatively’ here.
- Line 80: after “Labrador, Irminger, and Nordic seas”. This should have a reference, e.g., Katsman et al. (2018, https://doi.org/10.1029/2017JC013329).
- Lines 81 – 82: “The Atlantic nortward upper limb of the AMOC”, the term “Atlantic” is redudant here and could be removed.
- Line 84: I would also add the study by Orihuela-Pinto et al. (2022, https://doi.org/10.1038/s41558-022-01380-y) here, which nicely provides an overview of the relevant global climatic responses (their Figure 6).
- Line 85: The term “climatological equator” is unclear in this context. Do you mean that the ITCZ and/or the maximum zonally-averaged temperature is located north of the geographic equator? Please clarify.
- Lines 85 – 89: These climate impacts require substantially more comprehensive and appropriate referencing, as a single citation is clearly insufficient. Moreover, the study by Ben-Yami et al. (2024) focuses primarily on ITCZ shifts under different AMOC strengths and does not cover the broader range of climate impacts discussed here. A number of recent modelling studies have addressed AMOC-related climate impacts more comprehensively, including the mentioned effects here.
For clarity, I provide below a (non-exhaustive) overview of relevant and recent studies:
General impacts:
- Orihuela-Pinto et al. (2022, https://doi.org/10.1038/s41558-022-01380-y)
- Bellomo et al. (2023, https://doi.org/10.1007/s00382-023-06754-2)
- van Westen et al. (2024, already cited)
- Bellomo & Mehling (2024, https://doi.org/10.1029/2023GL107624)
Temperature impacts and winter storms:
- Meccia et al. (2024, https://doi.org/10.1088/1748-9326/ad14b0)
- Meccia et al. (2025, https://doi.org/10.1088/1748-9326/ada3e7)
- van Westen & Baatsen (2025, already cited)
Hydroclimate:
- Saini et al. (2025, https://doi.org/10.1029/2024PA004967)
- van Westen et al. (2025, https://doi.org/10.5194/hess-29-6607-2025)
Sea level:
- Volkov et al. (2023, https://doi.org/10.1038/s41467-023-40848-z)
- Howard et al. (2024, https://doi.org/10.1088/2515-7620/ad3368)
- van Westen et al. (2026, https://doi.org/10.5194/os-22-1353-2026) - Line 91 – 92: A recent study by Nian et al. (2026, https://www.nature.com/articles/s43247-026-03427-w) might be worth including/discussing here.
- Line 98: Perhaps you could roughly quantify the “large freshwater outburst” for the general reader, for example by expressing it in Sverdrups and/or as a multiple of the present-day Greenland Ice Sheet (GrIS) melt rate. I do acknowledge that this is being discussed in greater detail in lines 505 – 518.
- Line 108: I can’t find the (Dijkstra, 2024) entry in the reference list, please include.
- Lines 117–119: There is indeed a significant weakening trend across all SSP scenarios (Weijer et al., 2020, already cited; Dijkstra & van Westen, 2026;https://doi.org/10.1146/annurev-marine-040324-024822). However, more recent studies provide additional nuance. For instance, Bonan et al. (2025, https://doi.org/10.1038/s41561-025-01709-0) suggest that the decline is more coherent across CMIP6 models, while Portmann et al. (2026, https://www.science.org/doi/10.1126/sciadv.adx4298) report even larger declines when accounting for model biases. It may be worthwhile to also discuss these more recent findings here.
- Lines 119 – 120: The 21st century climate change-induced AMOC weakening is primarily driven by changing surface heat fluxes, with a limited contribution by surface freshwater fluxes. See the surface buoyancy flux analyses in van Westen et al. (2025, https://doi.org/10.1029/2025JC022651) and Drijfhout et al. (2025, already cited).
- Line 124: The ‘substantial biases in CMIP-class model’ needs some references:
- van Westen & Dijkstra (2024, already cited)
- Dijkstra & van Westen (2024, https://doi.org/10.16993/tellusa.3246)
- Portmann et al. (2026, https://www.science.org/doi/10.1126/sciadv.adx4298) - Line 126 – 127: ‘with indications that a tipping point may have been crossed’, this has been suggested in the analysis by van Westen et al. (2025, https://doi.org/10.1029/2025JC022651).
- Line 128 – 130: A reference to Vanderborght et al. (2025, https://doi.org/10.1175/JCLI-D-24-0570.1) is also appropriate here, as they quantify the ‘key feedbacks governing its stability and tipping behaviour’.
- Line 131: entrance -> southern entrance
- Line 132: Atlantic -> North Atlantic; as the salt-advection feedback convergences freshwater anomalies over the northern North Atlantic
- Line 134: ‘presence of an AMOC tipping point’, the review study by Dijkstra et al. (2026, https://doi.org/10.1002/wcc.70049) points to the strong evidence that the present-day AMOC is indeed in a multi-stable regime (i.e., the presence of an AMOC tipping point).
- Line 137: The statement “may not behave” is somewhat vague. This likely reflects that (quasi-)equilibrium conditions are not satisfied. Only under such conditions is an appropriate stability indicator, as it can then be meaningfully linked to the strength of the salt-advection feedback (Rahmstorf, 1996, already cited; Vanderborght et al., 2025, https://doi.org/10.1175/JCLI-D-24-0570.1). Please clarify this point.
- Lines 141 – 142: A reference to Hankel (2025, https://doi.org/10.1073/pnas.2411357121) is also appropriate here.
- Lines 145 – 147: What about the results from Weijer et al. (2012, https://doi.org/10.1029/2012GL051611), where a comparison is made between a high-resolution and low-resolution stand-alone ocean model simulation.
- Lines 150–153: I do not fully agree with this statement. The pulse-forcing experiments in Jackson et al. (2023) (+0.3 Sv for 100 years) are relatively short in duration (mostly 200 years in total), which may be insufficient to assess AMOC stability. For example, van Westen et al. (2025,https://doi.org/10.1029/2025JC022651) show in their pulse-forcing experiments (their Fig. 6a, blue curve) that a 200-year integration is too short to robustly assess AMOC stability based on AMOC strength alone. This limitation should be acknowledged or discussed more carefully.
- Lines 157–161: The statement regarding an overshoot does not appear to be supported by all relevant studies. For instance, the CO₂ hysteresis experiment in Willeit & Ganopolski (2024,https://doi.org/10.5194/esd-15-1417-2024) does not exhibit an overshoot, nor do the results presented by Gérard & Crucifix (2024, https://doi.org/10.5194/esd-15-293-2024).
- Lines 233 & 469: The compensation indeed prevents model drift; however, more importantly, it is required to obtain equilibrium solutions (i.e., to ensure an autonomous system).
- Line 235: Maybe you can already state here that you are proposing a volume compensation.
- Line 361: ‘Impact the GMSAT’, maybe good to add a reference here? For example, Drijfhout (2015, https://www.nature.com/articles/srep14877) or van Westen et al. (2024, already cited).
- Line 371: The more recent study by Pontes & Menviel (2024, https://doi.org/10.1038/s41561-024-01568-1) might be worth adding here.
- Line 447: This claim needs some references, for example Dijkstra (2024, https://doi.org/10.1016/j.physd.2023.133984).
- Line 451 (and elsewhere): It may be helpful to refer to this forcing protocol as a “pulse forcing” experiment, following the terminology used in Dijkstra & van Westen (2026,https://doi.org/10.1146/annurev-marine-040324-024822). Suggestion: “… constant freshwater flux for a fixed time, a so-called pulse forcing (Dijkstra & van Westen, 2026).”
- Line 454: “there is an indication that a threshold has been crossed”, this is quite vague. Please clarify which threshold.
- Line 482: There is a consensus on the sign (i.e., weakening) in AMOC projections (Weijer et al., 2020, already cited; Dijkstra & van Westen, 2026, https://doi.org/10.1146/annurev-marine-040324-024822), but there is much more uncertainty on the magnitude of AMOC weakening (Bonan et al., 2025, https://doi.org/10.1038/s41561-025-01709-0; Portmann et al., 2026, https://www.science.org/doi/10.1126/sciadv.adx4298). Maybe rewrite this sentence.
Citation: https://doi.org/10.5194/egusphere-2026-1698-RC1 - AC1: 'Reply on RC1', Didier Swingedouw, 31 Jul 2026
- General aim of Experiments A, and specifically the reversibility component (Lines 191–192): I find the proposed experiments useful for studying AMOC weakening due to the joint contributions of global warming and freshwater input. The imposed forcing leads to a “major AMOC weakening” (Line 548; could this be quantified?) during the first 150 model years. Thereafter, forcing conditions are ramped down over 100 years and continued for an additional 100 years under no external forcing.
-
RC2: 'Comment on egusphere-2026-1698', Anonymous Referee #2, 27 May 2026
This paper describes the experimental protocol for a set of coordinated simulations involving oceanic surface freshwater flux perturbations, conducted as part of the international Tipping Points Modelling Intercomparison Project (TIPMIP). As such, it is not a ‘normal research paper’, but instead provides details so that research group can carry out the needed experiments to contribute to the TIPMIP project. Given the numerical focus of TIPMIP, I do feel it is appropriate for GMD. In fact, I am happy to see such protocols published in the literature rather than just on a webpage or a hard to find report. Making the protocol easily accessible is of value. Additionally, given the amount of work needed to define the protocol, it is appropriate that the authors get the credit for a peer reviewed paper. After reviewing the draft, I would recommend minor revisions. Although generally clear and well presented, there are places where material could be improved. My specific comments are given below.
I prefer the term forced ocean models to ocean standalone models. Especially as it likely that no model is actually ocean only, instead including at least a sea-ice component.Line 79 – Likely newer references to the AMOC than just a 2016 paper. Especially given all the work on the topic.
Line 80 – Given the recent OSNAP results, this general statement of regions could be modernized and more clearly explained. And in general, more literature on the AMOC and its stability would be good. I’d also like to see more background on freshwater sources, and discussion of Greenland melt, versus sources of freshwater to the Arctic (including rivers) as well as the impact of fluxes through the gateway straits. For a freshwater hosing protocol, the key background concepts need to be presented.
Line 81 – Doesn’t a significant amount of the AMOC waters upwell in the Atlantic, before it leaves for the other basins?
The discussion of the 8.2 Kyr event focuses on discharge of the glacial lakes through the St. Lawrence into the Atlantic. Yet there is a literature that suggests that there is some debate on where and how the freshwater is released, such as into the McKenzie river into the Arctic Ocean.
Even if the authors want to use a Labrador Sea focused protocol, a proper discussion of the freshwater sources to the event should at least be referenced – and whether it leaves by Hudson Strait or the St. Lawrence River likely may mean different freshwater distributions. For example, might the whole sub-polar gyre be better than just the Labrador Sea, as the freshwater likely has to flow around the gyre to reach the West Greenland Current region. Or why not along the Labrador Coast – high resolution studies have suggested that leads to different freshwater fates for this event.
I didn’t notice any discussion about model resolution in the protocol, and wonder whether that is relevant. That may impact the distance criteria to input the freshwater and thus where the freshwater is input in models with such differences in resolution. At the very least, it may impact the locations of discharge around Greenland, especially if any fiords are at least partially resolved. How might that impact the input of freshwater?
For experiment C, there is no thought on considering the rivers that may provide the enhanced precipitation in a future climate?
Shouldn’t there be an optional for seasonally varying the freshwater input?
What about for models that include icebergs? How shall the discharge be partitioned between solid and liquid components?
For experiment B, Tier 1, what initial conditions are recommended for the 1920s? Wouldn’t it make sense to be consistent, as the choice of initial conditions may impact the model results.
Figure 6. Why use the same dark blue for both fluxes in the ramp up and ramp down scenarios? A slightly different shade might help show that there are two scenarios.
Citation: https://doi.org/10.5194/egusphere-2026-1698-RC2 - AC2: 'Reply on RC2', Didier Swingedouw, 31 Jul 2026
-
RC3: 'Comment on egusphere-2026-1698', Anonymous Referee #3, 04 Jun 2026
Review comments on “TIPMIP-OCEAN experimental protocol phase 1: Tipping dynamics of the AMOC” by D. Swingedouw et al.
The authors present a protocol for the ocean component of TIPMIP that focuses on the stability of the AMOC. The MIP proposes a new set of fresh water hosing experiments that partly complement the core TIPMIP experiments and partly extend earlier hosing experiments such as the NAHosMIP.
Many previous community-driven hosing MIPs have focused on adding fresh water to the ocean while keeping the background to the pre-industrial state. The novel aspect of the proposal here is that they add the hosing to climate change simulations or different background states. This is an important aspect as the background climate may influence the AMOC stability and it is also more relevant for the ongoing discussion on the effects of climate change on the AMOC and its consequences for the north Atlantic realm. The MIP proposes a relatively wide range of freshwater applications, from an addition of a “realistic” amount of meltwater from Greenland (exp. B) to very strong, idealized scenarios with (unrealistic) additions of up to 0.6Sv. Thus, the potential analyses can include uncertainty estimates and also address issues of structural differences in the models.
The protocol also includes an extension of the NAHosMIP with relatively simplified set-ups and an experiment with a paleo-backdrop, mimicking the 8.2 kyr event.
The manuscript reflects an extensive discussion in the community and is well balanced so that many modelling groups can participate, either in the full set or selected sub-experiments.
The manuscript is written in a clear and concise way and should be published in GMD.
I have only a few minor issues to consider:
Abstract:
Ln 60: In experiment C, the run at 2 degr. Warmer climate is only tier 2, so it cannot be the general purpose of this experiment to study the effect of background warming. I would formulate it more like Exp. C is extending the NAHosMIP to include the TIPMIP models with the potential to include the global warming aspect.
Ln 80ff: I would support Referee#2 here and include the “post OSNAP” discussion on the relative roles of deep water formation regions and that their representation is certainly model and resolution dependent.
Ln 155 constant background
Ln 207: is the precip bias issue really relevant here?
Ln 219: wasn’t that originally proposed by Gerdes et al (https://doi.org/10.1016/j.ocemod.2005.08.003)
Ln 240: it would be good to clarify here that some experiments are required in parallel to the core TIPMIP experiments, but for B and C, they can (at least for tier 1) be run in parallel to existing control or historical/scenario simulations.
Ln 373: is “pursued” the correct word here?
Ln 509: …testetd a range of freshwater range from 0 to 0.6 Sv ( no need to mention tier 2 here)
The experiment tables could be made a bit more consistent, e.g. include the initial conditions for all.
Citation: https://doi.org/10.5194/egusphere-2026-1698-RC3 - AC3: 'Reply on RC3', Didier Swingedouw, 31 Jul 2026
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 1,345 | 596 | 80 | 2,021 | 79 | 86 |
- HTML: 1,345
- PDF: 596
- XML: 80
- Total: 2,021
- BibTeX: 79
- EndNote: 86
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The manuscript by Swingedouw and co-authors describes an experimental protocol to systematically analyse AMOC sensitivity under external forcing from both freshwater input and climate change. The authors propose four interesting experiments, which will primarily be performed with Earth System Models (CMIP6+), and which are well suited to address the main scientific questions. The overarching aim is to obtain a better understanding of AMOC sensitivity and, in doing so, to assess AMOC-related impacts that are relevant for society.
I strongly recommend this manuscript for publication in Geoscientific Model Development (GMD). The experimental design is well motivated and clearly presented, with supporting figures and tables. This TIPMIP-OCEAN protocol will provide valuable information for the AMOC community and for assessing the risk of AMOC tipping under future climate change.
That being said, I have a few moderate and minor comments that should be addressed prior to publication.
Moderate comments and suggestions:
Given the typical multi-centennial adjustment timescale of the AMOC (Bonan et al., 2022, already cited) and results from existing pulse-forcing experiments (Jackson et al., 2023, already cited; van Westen et al., 2025, https://doi.org/10.1029/2025JC022651), it is unlikely that a coherent picture of AMOC reversibility can be obtained from the outlined experimental design (Fig. 2). Addressing reversibility would require substantially longer simulations for experiment “Stabilized (Ad)”, which are not mandatory under Tier 1 (Table 1). The 250-year experiment “Stabilized (Ab)” is likely sufficiently long in duration, but because it includes a modified background forcing, it cannot be used to assess reversibility.
My main point is that some aspects of AMOC stability cannot be addressed using simulations that are shorter than the AMOC adjustment timescale. This also applies to Experiments C and D, where the minimum required “stabilization” periods are 50 and 100 years, respectively. Only if modelling groups are willing to extend the stabilization (i.e. no external forcing) phases beyond 100 years can these questions be meaningfully addressed.
I would therefore suggest using more careful language and tempering expectations regarding AMOC reversibility, which is one of the central scientific questions in the TIPMIP-OCEAN protocol (Line 551).
Minor comments and suggestions:
For clarity, I provide below a (non-exhaustive) overview of relevant and recent studies:
General impacts:
- Orihuela-Pinto et al. (2022, https://doi.org/10.1038/s41558-022-01380-y)
- Bellomo et al. (2023, https://doi.org/10.1007/s00382-023-06754-2)
- van Westen et al. (2024, already cited)
- Bellomo & Mehling (2024, https://doi.org/10.1029/2023GL107624)
Temperature impacts and winter storms:
- Meccia et al. (2024, https://doi.org/10.1088/1748-9326/ad14b0)
- Meccia et al. (2025, https://doi.org/10.1088/1748-9326/ada3e7)
- van Westen & Baatsen (2025, already cited)
Hydroclimate:
- Saini et al. (2025, https://doi.org/10.1029/2024PA004967)
- van Westen et al. (2025, https://doi.org/10.5194/hess-29-6607-2025)
Sea level:
- Volkov et al. (2023, https://doi.org/10.1038/s41467-023-40848-z)
- Howard et al. (2024, https://doi.org/10.1088/2515-7620/ad3368)
- van Westen et al. (2026, https://doi.org/10.5194/os-22-1353-2026)
- van Westen & Dijkstra (2024, already cited)
- Dijkstra & van Westen (2024, https://doi.org/10.16993/tellusa.3246)
- Portmann et al. (2026, https://www.science.org/doi/10.1126/sciadv.adx4298)