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: open (until 02 Sep 2026)
- RC1: 'Comment on egusphere-2026-4199', Anonymous Referee #1, 18 Aug 2026 reply
-
RC2: 'Comment on egusphere-2026-4199', Pierre Tiengou, 24 Aug 2026
reply
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 - l.62-71.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 139 | 40 | 11 | 190 | 14 | 17 |
- HTML: 139
- PDF: 40
- XML: 11
- Total: 190
- BibTeX: 14
- EndNote: 17
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.