the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 Jul 2026
Past and future projected radiative effects from irrigation
Abstract. Irrigation is the most dominant freshwater-use practice on Earth, yet its effect on the top of atmosphere (TOA) radiative budget remains highly uncertain because past estimates have relied on idealised experiments or single-model frameworks. Here we capitalise on the Irrigation Model Intercomparison Project (IRRMIP) to provide the first multi-model quantification of irrigation-induced effective radiative forcing (ERF), defined here as the change in net TOA radiative flux including rapid atmospheric adjustments under prescribed sea surface temperatures (SST), over the historical period (1902–2014). Across seven Earth system models, under fixed SST conditions, the global multi-model mean (MMM) irrigation-induced ERF is positive but small (~ 0.023 W m⁻²), reflecting a near-cancellation between the positive longwave contribution associated with irrigation-induced atmospheric moistening and the negative shortwave contribution associated with enhanced reflection, including cloud-related effects. While the global signal is small, it masks pronounced regional heterogeneity: South Asia and West Central Asia exhibit substantially higher positive ERF (~ 0.25 W m⁻²), with spatial maps revealing coherent forcing hotspots of up to +5 W m⁻² over the Indo-Gangetic Plain that are consistent across all models, while East Asia shows a persistent negative signal (-0.03 W m⁻²). The clear-sky ERF is greater than the all-sky ERF in most regions, indicating that cloud adjustments following irrigation exert a systematic negative contribution to the ERF. We additionally examine future changes in net TOA radiation imbalance (ΔRn,TOA) from a fully coupled CMIP-style model (SSP1-2.6 and SSP3-7.0, 2015–2070) to assess whether the historical ERF pattern persists under future irrigation. A more global positive ΔRn,TOA is found under the high-irrigation-demand SSP3-7.0 pathway relative to SSP1-2.6, consistent with projected differences in irrigation expansion between scenarios. Correlations between historical ERF variability and its potential drivers suggest that reduced outgoing longwave radiation is the dominant control at the global scale and in South Asia and West Central Asia, while near-surface temperature, precipitable water, albedo, and cloud cover contribute with substantial regional and inter-model variability. These results demonstrate that irrigation's radiative influence, while small in the global mean, is regionally substantial and scenario-dependent, with implications for the attribution of regional climate change and for the design of land-based mitigation trajectories.
- Preprint
(3407 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-4199', Anonymous Referee #1, 18 Aug 2026
-
AC1: 'Reply on RC1', Amen Al-Yaari, 11 Sep 2026
Response to reviewers’ comments on egusphere-2026-4199
We are very grateful for your constructive feedback, which has helped us clarify several important aspects of the manuscript and strengthen both the methodology and discussion. Below, you will find our responses to your comments:
Comments by Anonymous Referee #1
This manuscript investigates the impacts of irrigation on the top-of-atmosphere (TOA) radiative budget, a topic that has received sustained attention over the past decades. The primary novelty of this study lies in the use of seven models from the Irrigation Model Intercomparison Project (IRRMIP), which substantially reduces the model dependence. In addition, the authors compare clear-sky and all-sky radiation-related variables to indirectly assess the impact of cloud-related variables on the TOA radiative budget. Finally, they employ a fully coupled Earth system model (CESM2) to investigate the future evolution of irrigation-induced radiative forcing (Rn,TOA). Overall, this study focuses on a well-defined scientific question. Although the analytical framework is relatively straightforward, the results are presented in considerable detail. I have several comments that I hope will help further improve the manuscript.
We thank the reviewer for these very constructive comments. We have addressed the reviewer’s comments in this response and have made modifications to the manuscript.
- Model dependence in future projections. One of the major strengths of this study is the use of a multi-model ensemble to reduce model dependence in the historical analysis. However, this advantage is largely lost in the future projections, where only a single model (CESM2) is employed. As a result, the projected future changes may be strongly influenced by model-specific characteristics, weakening one of the manuscript's main innovations. The authors should either justify this choice more clearly or discuss the associated uncertainty more explicitly.
Response:
We thank the reviewer for highlighting this important limitation. CESM2 is currently the only model for which future irrigation experiments under the selected scenarios are available within the IRRMIP framework; no other participating groups have performed the corresponding future simulations. We therefore use CESM2 to provide an initial assessment of the potential future evolution of irrigation-induced radiative effects, while acknowledging that model-specific characteristics may influence these projections. We will add the following statement to the revised manuscript, noting that the magnitude and regional patterns of the projected responses should be interpreted with caution.
“Several areas for improvement remain and should be explored in future studies: ............ (iii) the future projections are currently available only from CESM2, as no other IRRMIP modelling groups have performed the corresponding future irrigation experiments. Consequently, the magnitude and regional patterns of the projected responses may be influenced by CESM2-specific characteristics and should be interpreted as preliminary rather than as a robust multi-model projection.”
- Uncertainty associated with cloud parameterization. Cloud-related processes form one of the primary mechanisms used to interpret the TOA radiative responses throughout the manuscript. However, cloud representation remains one of the largest sources of uncertainty in current climate models. Consequently, the inferred cloud contributions should be interpreted with caution. I recommend that the authors explicitly acknowledge this uncertainty, particularly in the Conclusions.
Response:
We thank the reviewer for highlighting this important point. We fully agree that uncertainties in the representation of clouds and cloud-radiation interactions are an important source of uncertainty when interpreting the irrigation-induced TOA radiative response. We will therefore revise the Conclusions to acknowledge this limitation.
At the same time, we would like to emphasize that the multi-model framework of IRRMIP is a key strength in this context. The seven participating models employ different atmospheric and land-surface model configurations and different representations of irrigation, providing an opportunity to assess whether the diagnosed radiative responses are robust across structural model differences. Indeed, several of the main features identified in our analysis are consistent across models. For example, the strong positive ΔERF hotspot over the Indo-Gangetic Plain is reproduced across all models, and our spatial analysis identifies regions where at least 80% of the models agree on the sign of the response. These inter-model consistencies provide confidence that the main regional features are not attributable to a single model formulation. However, they do not eliminate structural uncertainty, particularly in the representation of clouds.
The following text will be added to the Conclusions:
“Clear-sky diagnostics indicate that cloud adjustments generally damp the all-sky radiative response to irrigation. This contribution should, however, be interpreted with caution, given structural uncertainties in cloud representation across models. At the same time, the multi-model consistency of several key spatial patterns and physical relationships provides confidence that the main features of the irrigation-induced radiative response are not specific to a single model formulation. Our all-sky versus clear-sky comparison provides a diagnostic estimate of the net cloud contribution rather than a process attribution, and its magnitude remains model-dependent.”.
- Mechanistic interpretation of regional heterogeneity. The manuscript provides detailed descriptions of the spatial heterogeneity of irrigation-induced radiative effects. However, the interpretation is largely qualitative, relying on spatial coincidence with potential controlling variables and previous literature. A more mechanistic explanation is needed to explain why different regions exhibit markedly different responses and why the dominant controls vary geographically. For example, in Line 373, the authors state that “EAS is distinct in showing a sign reversal under clear-sky conditions, with ΔRn,TOA transitioning from weakly positive to weakly negative after approximately 2030-2040 under both scenarios.”. This behavior is particularly interesting but is not adequately explained. Likewise, why do the clear-sky and all-sky responses remain so similar over East Asia? More detailed physical interpretation would substantially improve the manuscript.
Response:
We thank the reviewer for this important comment. The following text:
“EAS is distinct in showing a sign reversal under clear-sky conditions, with ΔRn,TOA transitioning from weakly positive to weakly negative after approximately 2030-2040 under both scenarios”
Will be replaced by:
“East Asia exhibits a distinct radiative regime in which the competing longwave and shortwave effects of irrigation are more closely balanced. Irrigation-induced moistening tends to increase atmospheric longwave trapping, producing a positive clear-sky contribution. Still, this effect is relatively weak compared with the strong positive response over South Asia and West Central Asia. At the same time, the historical analysis shows a consistent negative relationship between low- and medium-level cloud changes and ΔERF over EAS, indicating that cloud-related shortwave reflection can offset part of the longwave effect. Under the future CESM2 simulations, the clear-sky ΔRn,TOA over EAS changes from weakly positive to weakly negative around 2030-2040 in both scenarios, suggesting that the balance between the competing radiative components evolves as the background climate and irrigation-induced hydrological response change. In addition, the relatively small difference between the all-sky and clear-sky responses over EAS suggests that the net cloud contribution is insufficient to substantially modify the already weak clear-sky signal, indicating that non-cloud processes also contribute to the negative radiative response.”
- Line 91: The manuscript states that IRRMIP contains eight participating models. Why were only seven models included in the analysis.
Response:
We thank the reviewer for pointing this out. We confirm that eight models participated in the IRRMIP protocol, including IPSL-CM6. However, the atmospheric variables required for the present analysis were not available for IPSL-CM6. Consequently, IPSL-CM6 could not be included in the multi-model analysis, and the results presented here are based on the seven models for which the required atmospheric and radiative variables were available.
The following text will be included in Section 2.1:
“Although eight models participated in the IRRMIP protocol, the present analysis includes seven models (CESM2, CESM2_GW, CNRM-CM6-1, E3SMv2, MIROC-INTEG-ES, NASA-GISS, and NorESM2), as the atmospheric variables required for our analysis were not available for IPSL-CM6.”
- Figure 1A: Global cropland area has expanded substantially over the past century. Why, then, does the irrigation-induced ERF exhibit no trend? Is this because changes in shortwave and longwave radiation largely compensate each other? A brief discussion would be helpful.
Response:
We thank the reviewer for highlighting this point. The absence of a clear global trend, despite the AEI expansion since 1901, reflects the spatial cancellation shown in Fig. 3H and Table 1: the strengthening positive ΔERF trend in South Asia and West Central Asia is offset by a comparably negative signal over East Asia and mixed/near-zero signals elsewhere, so global-mean ΔERF stays flat even as the regional forcing intensifies.
The following text will be included in the revised manuscript:
“At the global scale, the irrigation-induced all-sky ΔERF remains small throughout the entire period (Fig. 1A), with the multi-model mean, reported in Table 1, being about +0.023 W m⁻². This small magnitude may reflect both the compensation between the longwave and shortwave components of the irrigation ΔERF (Boucher et al., 2004) and the spatial cancellation of opposing regional responses. Despite the expansion of irrigated areas since 1901, the strengthening positive ΔERF over South Asia and West Central Asia is offset by a negative response over East Asia, together with mixed or near-zero responses elsewhere (Fig. 3H and Table 1). Thus, the global-mean ΔERF remains relatively small and shows no clear long-term trend.”
- Lines 205–206: The statement “The rate of ΔERF change accelerates after approximately 1950, consistent with the rapid post-war expansion of irrigation infrastructure in Asia.” appears to relate Asian irrigation development to a globally averaged ERF. Please clarify whether this discussion refers to the global mean or only to the Asian contribution.
Response:
We fully agree that the sentence is ambiguous as written. The statement describes the multi-model mean ΔERF (Fig. 1B) over SAS (South Asia), with the post-1950 acceleration attributed to Asian irrigation expansion (see Fig. 1I).
To remove the ambiguity, we rephrase it as:
" The rate of ΔERF change over SAS accelerates after approximately 1950, consistent with the rapid post-war expansion of irrigation infrastructure in Asia.
- Line 310: The statement “However, reductions in albedo can also trigger rapid atmospheric adjustments.” requires supporting references. Relevant studies include, but are not limited to:
Laguë, M. M., Swann, A. L., & Boos, W. R. (2021). Journal of Climate, 34, 6651–6672. https://doi.org/10.1175/JCLI-D-20-0883.1IF: 4.3 Q2
Tang, S. et al. (2025). Geophysical Research Letters, 52, e2025GL117253. https://doi.org/10.1029/2025GL117253IF: 5.0 Q1
Response:
We thank the reviewer for the references. We have included them in the manuscript.
- Line 404: The manuscript states that “while over Central Asia and China, the shortwave reductions from enhanced cloud cover drive an overall decline.” However, Figure 6F shows that clear-sky ΔRn,TOA over East Asia is also negative. This suggests that factors other than cloud changes also contribute to the radiative response. The authors should discuss these additional mechanisms.
Response:
We thank the reviewer for pointing this out. As we addressed this issue earlier in the manuscript (see our response to Comment 3), we have omitted this sentence here to avoid repetition.
Citation: https://doi.org/10.5194/egusphere-2026-4199-AC1
-
AC1: 'Reply on RC1', Amen Al-Yaari, 11 Sep 2026
-
RC2: 'Comment on egusphere-2026-4199', Pierre Tiengou, 24 Aug 2026
This paper examines the impact of irrigation on the Earth’s radiative budget in Earth System Models (ESMs). It first uses land-atmosphere simulations of historical climate, conducted within the framework of IRRMIP, to quantify the Effective Radiative Forcing (ERF) induced by irrigation, at the global scale and in several various regions of interest for irrigation. An analysis of multiple physical processes possibly contributing to irrigation’s ERF is then conducted to identify the main drivers of the radiative response through correlations and formulate explanations relevant for each region. Finally, one ESM is used to run fully coupled simulations of future climate and study the evolution of the radiative impacts of irrigation identified previously, under two distinct SSP-RCP scenarios.
The paper relevantly addresses an important yet often overlooked impact of irrigation with multiple climate model runs that cover the entire historical period. It provides much more reliable and complete analysis than previously done thanks to ensemble runs and multi-model means and comparisons. Overall, it is well structured and the three sections make clear distinct points that complement each other. I have several questions and comments listed below to help improve the manuscript for the reader.
- l.62-71.
I am not fully familiar with radiative metrics and therefore have some questions. You mention RF (and RF at the TOA but I believe they are the same thing), and then ERF.. It is not fully clear to me why previous studies used RF ? Is it only that ERF was introduced later or that they had distinct objectives ? The paper also later uses TOA radiative imbalance with a clear justification. Assuming the orders of magnitude are directly comparable, it could be worth giving an idea of how do the values given for irrigation RF compare with estimated RF (and associated uncertainty) and/or radiative imbalance associated to major processes such as anthropic GHG emissions, aerosols, or clouds feedbacks under climate change. A mention of the uncertainty on estimates from the literature for irrigation (E)RF would also be useful in my opinion. - I assumed this was studied within IRRMIP, and think the introduction might be the good moment to justify the relevance of these ESMs to simulate impacts of irrigation compared to previous studies. Have they improved simulated irrigated volumes or been shown to perform better in irrigated regions maybe ?
- l.119. Why have you used 7 out of the 8 IRRMIP models mentioned in the introduction and not all ?
- l. 189 …whereas Gormley-Gallagher et al. represented irrigation interactively through the land model.
Unsure that I understand the point of this sentence correctly. I see how that is different from Sherwood et al. but it seems similar to what is done in IRRMIP models, right ? - Fig. 1 (and associated section)
To understand the trends in ERF, I would be curious to see a variable that shows simulated irrigation, unless it can be considered as behaving directly like irrigated areas (which I think would then be worth mentioning). - Fig. 1 (and associated section)
I also wondered if seeing a graph of the evolution of irrigation for each region might be more helpful than the share (which is quite easy to grasp from subfigure H) to compare it with trends in ERF shown for each regions. For example, there seem to be decreasing trends from the 1930s in North America and the 1960s in MED, is that associated with similar trends in irrigation ? - Fig. 1 (and associated section)
There is a white band in Fig1 H at the bottom which is not associated with any region - Fig. 1 (and associated section)
I am surprised by the lack of evolution of ERF at the global scale. Although this is partially commented on later by comparing with clear-sky ERF, I think it deserves more details. I am particularly curious of wether that there is a significant ERF in the beginning of the century where irrigation is supposedly quite small (at least compared to the rest of the period). More generally, I wonder about the uncertainty or statistical significance associated with the ERF values mentioned, since 0 often seems to be in the MM-spread. I am not fully familiar with the usual methods to assess significance in such a study but maybe additional details could be useful here. - l.262. 80% of the models
That means 6 out of 7 right ? It might be clearer to state it this way, unless a more complex calculation accounting for ensemble members is used ? - Fig 3
Hatches are barely visible, I would suggest exploring other visualisation, either only showing coloured data where the condition is verified, or maybe dots on grid cells that verify it rather than the hatches - Fig. 3
There are strong patches in oceans in CESM, NorESM, E3SM, what is the statistical significance of these changes ? How much do they contribute to the mean global ERF ? Have you checked the share of global ERF impact coming from land versus oceans ? - l. 288 Interestingly, this correlation reverses in CNA and WNA, suggesting that in these regions, other components of the TOA energy balance (likely SW changes) dominate the ΔERF response.
I would add that it’s also not clearly present for MED and EAS, and maybe relate it in more details to opposite behaviours of SW and near-surface T (possibly merging some of the next paragraph to examine the variables conjointly ?) - Fig 5, 6
I feel the need to visualise irrigation (or irrigated areas if it is sufficient as mentioned earlier) in both scenarios to understand the differences in the magnitude of the changes involved here - Fig 7
Although responses over the ocean are discussed here, the large values at the pole still raise the question of statistical significance of the impacts of irrigation since these regions are also very sensitive to climate change - (Generic comment)
I wondered about the relevance of considering mean values over the entire historical period since irrigation is evolving rather quickly in a century, have you considered looking into distinct 20- or 30-year periods (like for analysis under climate change) ? - (Generic comment)
Particularly because of the remote impacts observed on Fig 3, 7, I also am unsure about the physical meaning of local ERF (I admit do not know how common it is to look into this variable in that way). I understand the relevance of regional surface energy balance for instance since it is directly related to boundary layer conditions at the regional scale, but ERF is integrated over the entire atmospheric column which is supposedly very well mixed (I cannot think of studies linking irrigation with changes of cloud cover above the ABL and would be interested to know about any). Do you consider that the different regional behaviours could be an indicator of different regimes piloting interactions with general circulation patterns well above the ABL ? - (Generic comment)
To what extent have you considered the impact of modelling choices for parameterisations on the radiative response to irrigation ? I am particularly thinking of irrigation representation, clouds, deep convection, and radiative schemes, which all have rather large uncertainties. More specifically, is the diversity of modelling approaches a good way of assessing uncertainty in the responses in this study ? Could the presence of several models from the same family affect the relevance of a multi-model mean ? Have distinct models been found to consistently agree/disagree and could that be explained by differences in modelling choices ?
Although listing the modelling approaches for each in the Methods might be too much, I think this should be at least discussed in the final part of the paper.
Citation: https://doi.org/10.5194/egusphere-2026-4199-RC2 -
AC2: 'Reply on RC2', Amen Al-Yaari, 11 Sep 2026
Response to reviewers’ comments on egusphere-2026-4199
We thank Pierre Tiengou for the comments. We are very grateful for your constructive feedback, which has helped us clarify several important aspects of the manuscript and strengthen both the methodology and discussion. Below, we provide a point-by-point response to your comments.
Comments by Pierre Tiengou
This paper examines the impact of irrigation on the Earth’s radiative budget in Earth System Models (ESMs). It first uses land-atmosphere simulations of historical climate, conducted within the framework of IRRMIP, to quantify the Effective Radiative Forcing (ERF) induced by irrigation, at the global scale and in several various regions of interest for irrigation. An analysis of multiple physical processes possibly contributing to irrigation’s ERF is then conducted to identify the main drivers of the radiative response through correlations and formulate explanations relevant for each region. Finally, one ESM is used to run fully coupled simulations of future climate and study the evolution of the radiative impacts of irrigation identified previously, under two distinct SSP-RCP scenarios.
The paper relevantly addresses an important yet often overlooked impact of irrigation with multiple climate model runs that cover the entire historical period. It provides much more reliable and complete analysis than previously done thanks to ensemble runs and multi-model means and comparisons. Overall, it is well structured and the three sections make clear distinct points that complement each other. I have several questions and comments listed below to help improve the manuscript for the reader.
We thank the reviewer for these very constructive comments.
1. l.62-71.
I am not fully familiar with radiative metrics and therefore have some questions. You mention RF (and RF at the TOA but I believe they are the same thing), and then ERF.. It is not fully clear to me why previous studies used RF ? Is it only that ERF was introduced later or that they had distinct objectives ? The paper also later uses TOA radiative imbalance with a clear justification. Assuming the orders of magnitude are directly comparable, it could be worth giving an idea of how do the values given for irrigation RF compare with estimated RF (and associated uncertainty) and/or radiative imbalance associated to major processes such as anthropic GHG emissions, aerosols, or clouds feedbacks under climate change. A mention of the uncertainty on estimates from the literature for irrigation (E)RF would also be useful in my opinion.
Response:
We thank the reviewer for highlighting this point. The reviewer is correct that RF and RF at the TOA (same for ERF and ERF at the TOA) refer to the same thing. We have revised the manuscript to make this consistent throughout. RF at the TOA and RF will be replaced with RFTOA in the revised manuscript.
ERF was formally introduced into the IPCC assessment framework in AR5 and is now the preferred metric for comparing the climate effects of different forcing agents because it is more closely related to the subsequent temperature response. This was already mentioned in the revised manuscript:
“The traditional stratospherically-adjusted RFTOA metric is poorly suited for assessing near-surface water vapor perturbations in climate models because it requires the tropospheric state to remain fixed (IPCC, 2014). As a result, tropospheric radiative adjustments, which have been shown to play an important role in the response to irrigation forcing (Sherwood et al., 2018), are not accounted for (Sherwood et al., 2015). Effective RFTOA (ERFTOA) is a better predictor of long-term temperature change than RFTOA (Ramaswamy et al., 2018) and is the recommended method to assess the radiative impacts of climate perturbations (IPCC, 2023). Instantaneous radiative forcing refers to the immediate radiative effect of an imposed perturbation before the atmospheric and land-surface state adjusts.”. The earlier irrigation studies used different metrics partly because of differences in experimental design and partly because the ERFTOA framework became increasingly adopted over time. More importantly, the studies imposed different irrigation-related perturbations. For example, Boucher et al. (2004) considered idealized water-vapour perturbations, Cook et al. (2015) diagnosed the TOA radiative effect of irrigation in GISS ModelE2, whereas Sherwood et al. (2018) and Gormley-Gallagher et al. (2022) explicitly considered atmospheric adjustments associated with irrigation and therefore reported ERF.”
We will also add a comparison with major anthropogenic radiative forcings to provide context for the magnitude of the irrigation-induced signal. The IPCC AR6 assesses the total anthropogenic ERFTOA over 1750-2019 at 2.72 [1.96 to 3.48] W m⁻², the ERFTOA of greenhouse gases at 3.84 [3.46 to 4.22] W m⁻², and aerosol ERFTOA at –1.3 [–2.0 to –0.6] W m⁻². These values provide useful global-scale context, but we emphasize that they should not be interpreted as directly equivalent to our irrigation perturbation because they correspond to different reference periods, forcing agents, and experimental frameworks.
Our global irrigation ERFTOA of approximately +0.023 W m⁻² is therefore much smaller than the present-day anthropogenic ERF. However, the regional values are substantially larger than the global mean because irrigation is spatially concentrated. For example, the multi-model mean reaches approximately +0.22-0.24 W m⁻² in South Asia and West Central Asia, with local values exceeding 1 W m⁻² and reaching approximately 2-5 W m⁻² over some highly irrigated areas. We now emphasize that the small global mean should not be interpreted as a weak local radiative effect.
We also agree that uncertainty in previous irrigation estimates should be stated explicitly. Published studies have produced estimates ranging from negative to positive values, including approximately +0.03 to +0.1 W m⁻² in the idealized simulations of Boucher et al. (2004), approximately +0.02 W m⁻² in Cook et al. (2015), approximately −0.026 W m⁻² in Sherwood et al. (2018), and approximately +0.05 W m⁻² in Gormley-Gallagher et al. (2022). This spread reflects both structural model uncertainty and differences in the representation of irrigation and its associated atmospheric adjustments, rather than a formal uncertainty range from a single statistical distribution. This information will be included in the revised manuscript.
2. I assumed this was studied within IRRMIP, and think the introduction might be the good moment to justify the relevance of these ESMs to simulate impacts of irrigation compared to previous studies. Have they improved simulated irrigated volumes or been shown to perform better in irrigated regions maybe ?
Response:
We thank the reviewer for pointing this out. We agree that the motivation for using IRRMIP should be clarified in the Introduction. The purpose of IRRMIP is not to determine which model is better than another, but rather to provide a consistent multi-model framework in which the agreement and disagreement among models can be evaluated. Previous irrigation studies have often relied on a single model. In contrast, IRRMIP represents the transient historical expansion of irrigation and uses model-specific irrigation schemes to simulate water application. This provides a more systematic framework for assessing the robustness of irrigation-induced climate responses and the associated structural uncertainty.
3. l.119. Why have you used 7 out of the 8 IRRMIP models mentioned in the introduction and not all ?
Response:
We thank the reviewer for pointing this out. The eighth IRRMIP model, IPSL-CM6, was not included because the required atmospheric TOA radiative flux variables needed to calculate ΔERF were not available. We will state this explicitly in Section 2.1 rather than simply listing the seven models.
4. l. 189 …whereas Gormley-Gallagher et al. represented irrigation interactively through the land model.
Unsure that I understand the point of this sentence correctly. I see how that is different from Sherwood et al. but it seems similar to what is done in IRRMIP models, right ?
Response:
We agree with the reviewer that the irrigation representation in Gormley-Gallagher et al. (2022) is comparable to the interactive irrigation representations used in the IRRMIP models. In both cases, irrigation is represented within the land-surface model and can modify soil moisture, evapotranspiration, and the surface energy balance, rather than being prescribed solely as an external radiative or temperature perturbation. We will therefore revise the sentence to avoid implying that the approach of Gormley-Gallagher et al. is fundamentally different from that used in IRRMIP. The distinction we intended to make is instead between interactive irrigation representations and studies in which irrigation is represented more simply or prescribed externally. We will clarify this point in the revised manuscript.
5. Fig. 1 (and associated section)
To understand the trends in ERF, I would be curious to see a variable that shows simulated irrigation, unless it can be considered as behaving directly like irrigated areas (which I think would then be worth mentioning).
Response:
We thank the reviewer for this suggestion. The simulated irrigation is already presented and discussed in Figure 1 of Yao et al. (2025a), which shows the historical evolution and spatial distribution of irrigated areas and the associated irrigation water withdrawal across the IRRMIP models. To avoid duplicating a figure that has already been published, we have chosen not to reproduce this figure in the present manuscript. Instead, we will explicitly refer to Figure 1 of Yao et al. (2025a) in the relevant section and clarify that the simulated irrigation follows the historical expansion of irrigated areas, while inter-model differences remain in the magnitude of irrigation water withdrawal.
6. Fig. 1 (and associated section)
I also wondered if seeing a graph of the evolution of irrigation for each region might be more helpful than the share (which is quite easy to grasp from subfigure H) to compare it with trends in ERF shown for each regions. For example, there seem to be decreasing trends from the 1930s in North America and the 1960s in MED, is that associated with similar trends in irrigation ?
Response:
We thank the reviewer for this suggestion. We agree that the temporal evolution of irrigation could provide useful context for interpreting the regional ERF trends. However, the purpose of Figure 1 in our manuscript is to show the spatial contribution and relative importance of the different regions, rather than to reproduce the detailed temporal evolution of irrigation already presented in Yao et al. (2025a) in Figure 1(the historical evolution of irrigated areas and irrigation water withdrawal differs among regions). We will add a reference to this figure and clarify this point in the manuscript. The regional ERF reflects not only changes in irrigation extent or water application, but also the associated surface and atmospheric responses and their evolution over time. Therefore, a direct correspondence between the temporal trend in irrigation and the ERF trend should not necessarily be expected.7. Fig. 1 (and associated section)
There is a white band in Fig1 H at the bottom which is not associated with any region
Response:
We thank the reviewer for noticing this. The white band at the bottom of Fig. 1H corresponds to SEA (Southeast Asia). We will clarify this in the revised figure caption.
8. Fig. 1 (and associated section)
I am surprised by the lack of evolution of ERF at the global scale. Although this is partially commented on later by comparing with clear-sky ERF, I think it deserves more details. I am particularly curious of wether that there is a significant ERF in the beginning of the century where irrigation is supposedly quite small (at least compared to the rest of the period). More generally, I wonder about the uncertainty or statistical significance associated with the ERF values mentioned, since 0 often seems to be in the MM-spread. I am not fully familiar with the usual methods to assess significance in such a study but maybe additional details could be useful here.
Response:
We thank the reviewer for this suggestion. The absence of a pronounced global trend does not imply an absence of irrigation-induced radiative responses. Rather, the global mean is strongly affected by cancellation among spatially heterogeneous positive and negative responses and by the relatively small area occupied by irrigated land. The regional responses, particularly in South Asia and West Central Asia, become much larger as irrigation expands during the second half of the twentieth century. The confidence intervals reported in Table 1 are based on a t-distribution across the seven model-specific ensemble means, rather than treating individual ensemble members as independent models. We will add this information in the revised manuscript.
9. l.262. 80% of the models
That means 6 out of 7 right ? It might be clearer to state it this way, unless a more complex calculation accounting for ensemble members is used?
Response:
Yes. We will replace "≥80% of models" with "at least six of seven models" in the revised manuscript.
10. Fig 3
Hatches are barely visible, I would suggest exploring other visualisation, either only showing coloured data where the condition is verified, or maybe dots on grid cells that verify it rather than the hatches
Response:
We thank the reviewer for this suggestion. Instead of the current hatching, we will use dots.
11. Fig. 3
There are strong patches in oceans in CESM, NorESM, E3SM, what is the statistical significance of these changes ? How much do they contribute to the mean global ERF ? Have you checked the share of global ERF impact coming from land versus oceans ?
Response:
We thank the reviewer for raising this important point. In our analysis, ERF is calculated as a global mean, including both land and ocean areas, rather than being calculated separately for land and ocean. This approach is consistent with the conventional calculation of global-mean ERF and allows direct comparison with globally averaged radiative forcing estimates reported in the IPCC assessments. To further investigate the pronounced patches over the oceans in some models, we repeated the analysis using the models that provide the required clear-sky radiative variables (CESM2-GW, CNRM-CM6-1, and E3SM). The oceanic patches largely disappear under clear-sky conditions. This indicates that the spatial patterns over the oceans are strongly influenced by cloud-related processes and radiative adjustments, rather than representing a direct response to irrigation alone. We will add this analysis to the revised manuscript with more discussion.
12. l. 288 Interestingly, this correlation reverses in CNA and WNA, suggesting that in these regions, other components of the TOA energy balance (likely SW changes) dominate the ΔERF response.
I would add that it’s also not clearly present for MED and EAS, and maybe relate it in more details to opposite behaviours of SW and near-surface T (possibly merging some of the next paragraph to examine the variables conjointly ?)
Response:
We agree. We will revise the text to acknowledge that the relationship between ΔERF and outgoing longwave radiation is not uniform across all regions. In particular, MED and EAS do not show the same clear relationship as South Asia and West Central Asia. We will discuss the regional behavior in terms of the shortwave and longwave components of the TOA energy balance. In EAS, for example, the negative all-sky ERF despite a weakly positive or near-zero clear-sky response points to an important role for cloud-associated shortwave effects.
13. Fig 5, 6
I feel the need to visualise irrigation (or irrigated areas if it is sufficient as mentioned earlier) in both scenarios to understand the differences in the magnitude of the changes involved here.
Response:
We thank the reviewer for this suggestion. The irrigation area and annual mean irrigation water withdrawal (IWW) under SSP1-2.6 and SSP3-7.0 are already presented and discussed extensively in a dedicated paper by Yao et al. (2025). To avoid duplicating a figure that has already been published, we have chosen not to reproduce this figure in the present manuscript. Instead, we will explicitly refer to the figures of Yao et al. (2025) in the relevant section.
14. Fig 7
Although responses over the ocean are discussed here, the large values at the pole still raise the question of statistical significance of the impacts of irrigation since these regions are also very sensitive to climate change
Response:
Please see our reply to comment # 11.
15. (Generic comment)
I wondered about the relevance of considering mean values over the entire historical period since irrigation is evolving rather quickly in a century, have you considered looking into distinct 20- or 30-year periods (like for analysis under climate change) ?
Response:
Thank you for this suggestion. In addition to showing the temporal evolution of irrigation and its effects in Figures 1, 2, 5, and 7, we also analyze distinct periods. Specifically, Figure 3 focuses on the 1981-2014 period (30 years), while Figure 7 examines the 2050-2070 period (20 years). These analyses complement the historical-period mean values and allow us to assess how irrigation and its associated effects evolve over time.
16. (Generic comment)
Particularly because of the remote impacts observed on Fig 3, 7, I also am unsure about the physical meaning of local ERF (I admit do not know how common it is to look into this variable in that way). I understand the relevance of regional surface energy balance for instance since it is directly related to boundary layer conditions at the regional scale, but ERF is integrated over the entire atmospheric column, which is supposedly very well mixed (I cannot think of studies linking irrigation with changes of cloud cover above the ABL and would be interested to know about any). Do you consider that the different regional behaviours could be an indicator of different regimes piloting interactions with general circulation patterns well above the ABL ?
Response:
We appreciate this important point. Atmospheric circulation dynamically connects different regions, so the TOA radiative response at a particular location does not necessarily originate from processes over the local surface. We agree with the reviewer that the contrasting regional responses, particularly the remote signals in Figs. 3 and 7, may reflect interactions between irrigation-induced surface perturbations and different background atmospheric circulation regions. Irrigation initially modifies the local surface energy and water budgets through changes in evapotranspiration, surface temperature, and atmospheric humidity. These perturbations may subsequently influence moisture transport, cloud and radiative properties, latent heating, and atmospheric circulation via initial land-atmosphere interactions, thereby producing radiative responses beyond the irrigated regions (local effect) via atmospheric wave propagation and/or circulation changes and advection (non-local effects). Therefore, the spatial pattern of ERF may contain contributions from both local thermodynamic adjustments and induced circulation (advection) remote responses, so that the cloud cover will be altered. However, we didn't diagnose the vertical cloud response or the detailed circulation pathways connecting irrigation to the remote ERF signals, so we present these mechanisms as plausible interpretations. This information will be included in the revised manuscript.
17. (Generic comment)
To what extent have you considered the impact of modelling choices for parameterisations on the radiative response to irrigation ? I am particularly thinking of irrigation representation, clouds, deep convection, and radiative schemes, which all have rather large uncertainties. More specifically, is the diversity of modelling approaches a good way of assessing uncertainty in the responses in this study ? Could the presence of several models from the same family affect the relevance of a multi-model mean ? Have distinct models been found to consistently agree/disagree and could that be explained by differences in modelling choices ?Although listing the modelling approaches for each in the Methods might be too much, I think this should be at least discussed in the final part of the paper.
Response:
We thank the reviewer for this important point, and we agree that this is an important limitation. Several elements relevant to this question are already present in the manuscript, though not yet drawn together into a dedicated paragraph. Section 2.1 describes the two categories of irrigation representation across the seven IRRMIP models (external forcing vs. soil-moisture-triggered schemes). The irrigation schemes themselves differ substantially: CNRM and NASA-GISS use external irrigation forcing, whereas the other models use soil-moisture-based irrigation schemes with different triggers, timing, and amounts. In addition, CESM2 and CESM2-GW share the same atmospheric model (meaning the same parameterizations for the cloud and PBL, for example); the only difference is the irrigation scheme and the groundwater component in CESM2-GW. E3SM also shares many aspects of the land model used in CESM2 and CESM2-GW, although they differ in atmospheric models and irrigation implementation. However, with only seven models in total, three of which share a land component and four of which are each structurally distinct, the current ensemble is too small to meaningfully partition into "shared" and "independent" subsets and compare their behavior statistically. We will add a new paragraph in the conclusion Section to highlight that inter-model disagreement in ΔERF may reflect differences in irrigation, cloud, and convection parameterizations. This flags parameterization diversity as future work and recommends that future, larger multi-model irrigation ensembles be designed with explicit attention to sampling independent land-atmosphere-irrigation structures.
References
Gormley-Gallagher, A.M., Sterl, S., Hirsch, A.L., Seneviratne, S.I., Davin, E.L., Thiery, W., 2022. Agricultural management effects on mean and extreme temperature trends. Earth Syst. Dyn. 13, 419–438. https://doi.org/10.5194/esd-13-419-2022
Intergovernmental Panel on Climate Change (IPCC) (Ed.), 2023. The Earth’s Energy Budget, Climate Feedbacks and Climate Sensitivity, in: Climate Change 2021 – The Physical Science Basis: Working Group I Contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, pp. 923–1054. https://doi.org/10.1017/9781009157896.009
Intergovernmental Panel on Climate Change (IPCC) (Ed.), 2014. Anthropogenic and Natural Radiative Forcing, in: Climate Change 2013 – The Physical Science Basis: Working Group I Contribution to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, pp. 659–740. https://doi.org/10.1017/CBO9781107415324.018
Ramaswamy, V., Collins, W., Haywood, J., Lean, J., Mahowald, N., Myhre, G., Naik, V., Shine, K.P., Soden, B., Stenchikov, G., Storelvmo, T., 2018. Radiative Forcing of Climate: The Historical Evolution of the Radiative Forcing Concept, the Forcing Agents and their Quantification, and Applications. Meteorol. Monogr. 59, 14.1-14.101. https://doi.org/10.1175/AMSMONOGRAPHS-D-19-0001.1
Sherwood, S.C., Bony, S., Boucher, O., Bretherton, C., Forster, P.M., Gregory, J.M., Stevens, B., 2015. Adjustments in the Forcing-Feedback Framework for Understanding Climate Change. Bull. Am. Meteorol. Soc. 96, 217–228. https://doi.org/10.1175/BAMS-D-13-00167.1
Sherwood, S.C., Dixit, V., Salomez, C., 2018. The global warming potential of near-surface emitted water vapour. Environ. Res. Lett. 13, 104006. https://doi.org/10.1088/1748-9326/aae018
Yao, Y., Ducharne, A., Cook, B.I., De Hertog, S.J., Aas, K.S., Arboleda-Obando, P.F., Buzan, J., Colin, J., Costantini, M., Decharme, B., Lawrence, D.M., Lawrence, P., Leung, L.R., Lo, M.-H., Devaraju, N., Wieder, W.R., Wu, R.-J., Zhou, T., Jägermeyr, J., McDermid, S., Pokhrel, Y., Elling, M., Hanasaki, N., Muñoz, P., Nazarenko, L.S., Otta, K., Satoh, Y., Yokohata, T., Jin, L., Wang, X., Mishra, V., Ghosh, S., Thiery, W., 2025a. Impacts of irrigation expansion on moist-heat stress based on IRRMIP results. Nat. Commun. 16, 1045. https://doi.org/10.1038/s41467-025-56356-1
Yao, Y., Satoh, Y., van Maanen, N., Taranu, S., Keune, J., De Hertog, S.J., Lampe, S., Lawrence, D.M., Sacks, W.J., Wada, Y., Ducharne, A., Cook, B.I., Seneviratne, S.I., Liu, L., Buzan, J.R., Jägermeyr, J., Thiery, W., 2025b. Compounding future escalation of emissions- and irrigation-induced increases in humid-heat stress. Nat. Commun. 16, 9326. https://doi.org/10.1038/s41467-025-64375-1
Citation: https://doi.org/10.5194/egusphere-2026-4199-AC2
- l.62-71.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 204 | 73 | 19 | 296 | 19 | 22 |
- HTML: 204
- PDF: 73
- XML: 19
- Total: 296
- BibTeX: 19
- EndNote: 22
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript investigates the impacts of irrigation on the top-of-atmosphere (TOA) radiative budget, a topic that has received sustained attention over the past decades. The primary novelty of this study lies in the use of seven models from the Irrigation Model Intercomparison Project (IRRMIP), which substantially reduces the model dependence. In addition, the authors compare clear-sky and all-sky radiation-related variables to indirectly assess the impact of cloud-related variables on the TOA radiative budget. Finally, they employ a fully coupled Earth system model (CESM2) to investigate the future evolution of irrigation-induced radiative forcing (Rn,TOA).
Overall, this study focuses on a well-defined scientific question. Although the analytical framework is relatively straightforward, the results are presented in considerable detail. I have several comments that I hope will help further improve the manuscript.
1. Model dependence in future projections. One of the major strengths of this study is the use of a multi-model ensemble to reduce model dependence in the historical analysis. However, this advantage is largely lost in the future projections, where only a single model (CESM2) is employed. As a result, the projected future changes may be strongly influenced by model-specific characteristics, weakening one of the manuscript's main innovations. The authors should either justify this choice more clearly or discuss the associated uncertainty more explicitly.
2. Uncertainty associated with cloud parameterization. Cloud-related processes form one of the primary mechanisms used to interpret the TOA radiative responses throughout the manuscript. However, cloud representation remains one of the largest sources of uncertainty in current climate models. Consequently, the inferred cloud contributions should be interpreted with caution. I recommend that the authors explicitly acknowledge this uncertainty, particularly in the Conclusions.
3. Mechanistic interpretation of regional heterogeneity. The manuscript provides detailed descriptions of the spatial heterogeneity of irrigation-induced radiative effects. However, the interpretation is largely qualitative, relying on spatial coincidence with potential controlling variables and previous literature. A more mechanistic explanation is needed to explain why different regions exhibit markedly different responses and why the dominant controls vary geographically. For example, in Line 373, the authors state that “EAS is distinct in showing a sign reversal under clear-sky conditions, with ΔRn,TOA transitioning from weakly positive to weakly negative after approximately 2030–2040 under both scenarios.”. This behavior is particularly interesting but is not adequately explained. Likewise, why do the clear-sky and all-sky responses remain so similar over East Asia? More detailed physical interpretation would substantially improve the manuscript.
4. Line 91: The manuscript states that IRRMIP contains eight participating models. Why were only seven models included in the analysis.
5. Figure 1A: Global cropland area has expanded substantially over the past century. Why, then, does the irrigation-induced ERF exhibit no trend? Is this because changes in shortwave and longwave radiation largely compensate each other? A brief discussion would be helpful.
6. Lines 205–206: The statement “The rate of ΔERF change accelerates after approximately 1950, consistent with the rapid post-war expansion of irrigation infrastructure in Asia.” appears to relate Asian irrigation development to a globally averaged ERF. Please clarify whether this discussion refers to the global mean or only to the Asian contribution.
7. Line 310: The statement “However, reductions in albedo can also trigger rapid atmospheric adjustments.” requires supporting references. Relevant studies include, but are not limited to:
8. Line 404: The manuscript states that “while over Central Asia and China, the shortwave reductions from enhanced cloud cover drive an overall decline.” However, Figure 6F shows that clear-sky ΔRn,TOA over East Asia is also negative. This suggests that factors other than cloud changes also contribute to the radiative response. The authors should discuss these additional mechanisms.