Volume 51, Issue 5 e2023GL107094
Research Letter
Open Access

Spring Irrigation Reduces the Frequency and Intensity of Summer Extreme Heat Events in the North China Plain

Guoshuai Liu

Guoshuai Liu

The National Key Laboratory of Water Disaster Prevention, Hohai University, Nanjing, China

College of Agricultural Science and Engineering, Hohai University, Nanjing, China

Contribution: Conceptualization, Methodology, Software, Validation, Data curation, Writing - original draft, Writing - review & editing, Visualization

Search for more papers by this author
Weiguang Wang

Corresponding Author

Weiguang Wang

The National Key Laboratory of Water Disaster Prevention, Hohai University, Nanjing, China

Yangtze Institute for Conservation and Development, Hohai University, Nanjing, China

College of Hydrology and Water Resources, Hohai University, Nanjing, China

Cooperative Innovation Center for Water Safety and Hydro Science, Hohai University, Nanjing, China

Correspondence to:

W. Wang,

[email protected]

Contribution: Conceptualization, Writing - original draft, Writing - review & editing, Supervision

Search for more papers by this author
Hui Xu

Hui Xu

The National Key Laboratory of Water Disaster Prevention, Hohai University, Nanjing, China

College of Agricultural Science and Engineering, Hohai University, Nanjing, China

Contribution: Conceptualization, Writing - review & editing, Supervision

Search for more papers by this author
First published: 02 March 2024

Abstract

Irrigation has distinct impacts on extreme temperatures. Due to the carryover effect of soil moisture into other seasons, temperature impacts of irrigation are not limited to irrigated seasons. Focusing on the North China Plain, where irrigation occurs in both spring (March-April-May) and summer (June-July-August), with a higher proportion of irrigation water applied during spring, we investigate the impact of spring irrigation on summer extreme heat events. Based on partial correlation analysis of data products, we find positive correlations between spring and summer soil moisture, suggesting that spring irrigation-induced water surplus persists into the following summer and affects regional climate by impacting surface energy partitioning. Regional climate simulations confirm cross-seasonal climatic effects and show that spring irrigation reduces the frequency and intensity of summer extreme heat events by approximately −2.5 days and −0.29°C, respectively. Our results highlight the importance of the cross-seasonal climatic effect of irrigation in mitigating climate extremes.

Key Points

  • Effect of multi-seasonal irrigation on summer extreme heat events is investigated

  • Spring irrigation is beneficial for reducing summer extreme heat events

  • Irrigation modulates the relationship between spring and summer soil moisture

Plain Language Summary

Irrigation exerts a stronger impact on extreme temperatures than on mean temperatures. The North China Plain (NCP) is a typical winter wheat-summer maize rotation planting area, where irrigation is necessary in both spring and summer, but with a higher proportion of irrigation water applied during spring. The climatic effects of spring and summer irrigation in the NCP are intertwined due to the carryover effects of soil moisture. Recently, the climatic effect of irrigation in the NCP has been extensively explored, whereas the cross-seasonal effects of irrigation on summer extreme heat events have never been quantified. In this study, we employ the Weather Research and Forecasting model coupled with a demand-driven irrigation algorithm to discern the effects of spring and/or summer irrigation on summer extreme heat events by means of idealized climate simulations. The results show that spring and summer irrigation significantly reduces the frequency and intensity of summer extreme heat events by approximately −6.5 days and −1.0°C, of which spring irrigation contributes about 38% and 30%, respectively. Our findings underline the importance of irrigation-induced climate impacts in mitigating extreme heat events and emphasize that climate change adaptation planning in terms of irrigation must account for cross-seasonal climatic effects.

1 Introduction

Irrigation is an important anthropogenic activity that accounts for approximately 70% of the total freshwater withdrawals globally (Wada et al., 2013) and generates distinct climate impacts (McDermid et al., 2023; Singh et al., 2018; Thiery et al., 2020). Numerous observational and modeling studies have explored the climatic effect of irrigation and yielded considerable achievements, showing that irrigation strongly affects the partitioning of surface turbulent heat fluxes, soil moisture, crop growth, and the coupling strength of land-atmosphere interactions and therefore climate impacts (e.g., Ambika & Mishra, 2022; Bonfils & Lobell, 2007; Chen & Dirmeyer, 2019; Hirsch et al., 2017; Jha et al., 2022; Kang & Eltahir, 2018; Liu & Wang, 2023; Liu et al., 2021a, 2023; Mishra et al., 2020; Ozdogan et al., 2010; Qian et al., 2013; Thiery et al., 2017, 2020; Wang et al., 2021; Yang et al., 2020). Of the climatic effects of irrigation, the cooling effect is the most famous since it counteracts greenhouse-gas-induced warming (Bonfils & Lobell, 2007). Previous studies have mainly focused on the irrigation cooling effect on mean temperatures, aiming to reveal the contribution of irrigation to climate warming mitigation on global or regional scales (e.g., Cook et al., 2015; Qian et al., 2013; Sacks et al., 2009; Wu et al., 2018; Yang et al., 2016, 2020). With extreme heat events frequently occurring, a growing body of literature has attempted to evaluate the irrigation cooling effect on extreme temperatures, and it has been concluded that the impact on extreme temperatures is much more pronounced than that on mean temperatures, highlighting the role of irrigation in modulating extreme temperatures (e.g., Ambika & Mishra, 2022; Hauser et al., 2019; Hirsch et al., 2017; Mishra et al., 2020; Mueller et al., 2016; Thiery et al., 2017).

However, the climatic effect of irrigation is not limited to irrigated seasons due to soil moisture memory (e.g., Liu et al., 2021a; Qian et al., 2013; Yang et al., 2016). Soil moisture memory is also referred to as the persistence of soil moisture anomalies. Atmospheric anomalies can produce a positive or negative anomaly in soil moisture, while the dissipation of soil moisture anomaly may take weeks to months (Koster & Suarez, 2001), resulting in an autocorrelated evolution of soil moisture over time (Li et al., 2021). Existing studies have demonstrated the crucial role of antecedent soil moisture conditions in the evolution of summer extreme and mean temperatures over Western Europe, Australia, and the Northern Hemisphere (Lian et al., 2020; Pascoa et al., 2022; Stegehuis et al., 2021). Irrigation, as the most intensive activity for water use, could exert significant influences on the autocorrelated evolution of soil moisture and further affect the interseasonal climate. However, the role of the cross-seasonal effects of irrigation in mitigating summer extreme heat events has not been adequately recognized and thoroughly investigated. Thus, comprehensively assessing the cross-seasonal climatic effects of irrigation is important for fully understanding the climate response to agricultural water management and thus more accurately simulating past and future climate extremes and improving attribution studies.

The North China Plain (NCP) with an area of about 400 thousand square kilometers is China's agricultural base providing about 37% of the national wheat and maize production (National Bureau of Statistics of China, 2020). Irrigation contributes an essential component of agriculture in this region due to the limited annual precipitation of approximately 475 mm relative to similar locations in South China (Kang & Eltahir, 2018; Koch et al., 2020). Moreover, owing to the influence of the sub-humid temperate continental monsoon climate, the seasonal variation in precipitation is highly uneven, with approximately 70% of the annual precipitation occurring from June-September (Figure S1 in Supporting Information S1). Therefore, the relatively high potential evapotranspiration and low precipitation from March-May render spring irrigation more important in ensuring crop growth, accounting for a high proportion of the annual irrigation water use (Koch et al., 2020; Liu et al., 2023). Furthermore, despite the considerably high precipitation from June-September, summer irrigation is still needed based on site-specific needs (Zhang et al., 2022). In this case, we argue that spring irrigation could affect the summer climate through soil moisture memory combined with summer irrigation in this region. To the best of our knowledge, while the climatic effects of multi-seasonal irrigation in the NCP have been extensively investigated in recent studies (e.g., Liu & Wang, 2023; Wu et al., 2018; Yang et al., 2016; Zhang et al., 2022), the cross-seasonal effects of irrigation on summer extreme heat events and the separate contributions of spring and summer irrigation have not been explored and quantified.

To bridge this knowledge gap, we first analyze data products from 1980 to 2018 to detect the potential impact of irrigation on the connections between spring and summer soil moisture. Subsequently, we employ the Weather Research and Forecasting (WRF) model coupled with a demand-driven irrigation algorithm within the Noah-Multiparameterization land-surface model (NoahMP LSM; Niu et al., 2011) to discern the effects of spring and/or summer irrigation on summer extreme heat events in the NCP over the 2004–2018 period by means of idealized climate simulations.

2 Materials and Methods

2.1 Data Sets

To investigate the relationship between spring and summer soil moisture in the NCP, we use monthly root-zone soil moisture data derived from the Global Land Evaporation Amsterdam Model (GLEAM), which exhibits a spatial resolution of 0.25° covering 1980–2018 (Martens, et al., 2017). The fifth version of the global map of irrigation areas at a 5-arc min spatial resolution is used to depict where irrigation occurs (Siebert et al., 2013). The new ERA5 Reanalysis data (Copernicus Climate Change Service, 2017), with a spatial resolution of 0.25° and a temporal resolution of 1 hr, are obtained from the European Center for Medium-Range Weather Forecasting (ECMWF), which are used as lateral boundary forcing in the WRF simulations and for evaluating the WRF outputs of climatic variables. Leaf area index (LAI), the fraction of vegetation cover (FVC), and surface albedo data accessed from the Global Land Surface Satellite (GLASS) data sets (Liang et al., 2021) are used to prescribe the land surface vegetation characteristics in the WRF simulations. The GLASS data sets exhibit a spatial resolution of 0.05° and a temporal resolution of 8 days and are considered one of the best vegetation data sets in China (Li et al., 2018a, 2018b).

2.2 Model and Experimental Design

The advanced research WRF model version 4.1 (Skamarock et al., 2019), which has been widely used for regional climate modeling due to its flexible resolution and parameterization, is applied to simulate the regional climate response to irrigation in the NCP. The selected physics parameterization schemes are based on our previous study (Liu & Wang, 2023; see Table S1 in Supporting Information S1). To mimic irrigation processes, a demand-driven irrigation algorithm is added into the NoahMP LSM. This irrigation algorithm determines the occurrence of irrigation by checking if the WRF-simulated root zone soil moisture availability falls below a certain threshold during the irrigated seasons, which has been commonly used in regional irrigation-related studies (e.g., Liu & Wang, 2023; Liu et al., 2023; Nie et al., 2021; Ozdogan et al., 2010; Qian et al., 2013; Yang et al., 2019). Given that the winter wheat-summer maize rotation is the dominant crop system in the NCP (Jeong et al., 2014), the irrigation algorithm is activated from March 1 to May 31 and from June 20 to September 15, corresponding to the main crop growth period of winter wheat and summer maize. Moreover, to accurately depict crop growth, we turn off the dynamic vegetation scheme in NoahMP LSM and use the remotely sensed vegetation data sets to prescribe the land surface vegetation characteristics. We calibrated the soil moisture threshold values based on the Governmental agricultural census-based irrigation water use (IWU) data to make the simulated IWU more in line with observed IWU (Jha et al., 2022). Note that the NoahMP LSM is a one-dimensional land surface model and does not account for horizontal flow of water. Thus, we neglect the impact of water resources management on spatial redistribution of water and assume that water availability for irrigation is unlimited (Qian et al., 2020). For more details on the irrigation algorithm, refer to Text S1 and Figure S2 in Supporting Information S1.

The WRF model is configured with one domain with a center of 36°N, 116°E and a horizontal grid spacing of 25 km (containing 180 and 170 grid points along the north-south and east-west directions, respectively, as shown in Figure S3 in Supporting Information S1). There are 37 vertical layers from the surface to the 50-hPa level. In total, we conduct three sets of WRF experiments. The first experiment does not contain any representation of irrigation (hereafter referred to as NOIR). The second experiment is designed to have the irrigation algorithm activated in both spring and summer (hereafter referred to as IRS1). In the third experiment, the irrigation algorithm is only activated in spring (hereafter referred to as IRS2). With this numerical experiment design, the fingerprint of spring irrigation on summer climate variables can be calculated as the difference between the IRS2 and NOIR experiments, while the combined effect of spring and summer irrigation can be calculated as the difference between the IRS1 and NOIR experiments. All three experiments are performed from 2004 to 2018 year by year (the simulation of each year extends from January 1 to December 31) using lateral and lower boundary conditions from ERA5 Reanalysis data.

2.3 Bias Correction of the Regional Climate Simulations

In the WRF simulations, systematic biases remain in the simulation outputs due to the uncertainties persisting in the physical parametrizations (Yang et al., 2015). These systematic biases could cause overestimation or underestimation of the simulated extreme heat events. Hence, to capture the characteristics of summer extreme heat events in the regional climate simulations, we apply the bias correction procedure developed by Pal and Eltahir (2016) to reduce the systematic bias of the daily maximum surface air temperature at 2 m height (Tmax). Note that because the design of the IRS1 experiment is closer to reality, we use the IRS1 experiment output and ERA5 Reanalysis data to detect the systematic bias.

The bias correction procedure consists of five steps. (a) Tmax for each day is computed for both the IRS1 experiment hourly output and the ERA5 Reanalysis hourly 0.25° × 0.25° data. (b) The ERA5 Tmax data are interpolated from the 0.25° × 0.25° horizontal grid to the 25-km WRF grid. (c) Consistent IRS1 experiment and ERA5 2004–2018 mean values of Tmax are computed for each day of the year on the 25-km WRF grid. (d) The magnitude of the bias for each day of the year is estimated as the difference between the 30-day running means of the two types of data from 2004 to 2018 means. (e) The daily bias is finally applied to the WRF daily values of Tmax for all three experiments.

2.4 Indices of Extreme Heat Events

Here, we use two indices to evaluate the effect of irrigation on summer extreme heat events, according to a previous study of extreme temperature detection (Perkins & Alexander, 2013). One index, namely, extreme hot days (EHD), denotes the frequency of extreme heat events, expressed as the number of summer days, with daily values of Tmax exceeding the 90% percentile of the baseline period. Percentiles are calculated for each day over a 15-day moving window using the ERA5 Reanalysis data for the 1980–2018 period. The other one index, namely, extreme hot temperatures (EHT), denotes the intensity of extreme heat events, expressed as the 90% percentile of summer Tmax. Based on the bias-corrected WRF daily values of Tmax, these two indices are computed over the 2004–2018 period for all three experiments.

3 Results

3.1 Connections Between Spring and Summer Soil Moisture Contents

The NCP is the most intensively irrigated area in China, wherein the irrigation fraction of most grid cells exceeds 50% (Figure 1a). Additional water delivered to the soil by irrigation could affect the connections between spring and summer soil moisture in this area. Apart from irrigation, the climate variables also have influences on the evolution of soil moisture. To statistically control the covarying effects of climatic variables on the relationship between spring and summer soil moisture, we apply the partial correlation analysis, which is a method used to describe the relationship between two variables whilst taking away the effects of other variables on this relationship. The partial correlation coefficient is a measure of the linear dependence after eliminating the effects of other variables (Encyclopedia of Mathematics, https://www.encyclopediaofmath.org/wiki/Partial_correlation_coefficient). The climatic variables, including summer temperature, precipitation, and solar radiation, are obtained from the ERA5 Reanalysis. As shown in Figure 1b, the partial correlations between spring and summer soil moisture exhibit a widespread positive pattern across the NCP. The widespread positive pattern infers that when the effects of interannual variations of summer precipitation are factored out, the increase in spring soil moisture due to irrigation will result in local soil wetting in subsequent summers. Moreover, we find that the partial correlation coefficient generally decreases with increasing irrigation fraction (Figure 1c). This may be because although spring irrigation increase summer soil moisture, the increase in summer soil moisture is still mainly caused by summer irrigation. Overall, the concurrence of spring and summer irrigation modulates the relationship between spring and summer soil moisture in the NCP.

Details are in the caption following the image

Partial correlations between spring and summer soil moisture during 1980-2018. Spatial distributions of (a) irrigation fractions (%) and (b) partial correlation coefficients in the North China Plain. Black stipples in (b) indicate regions with a statistically significant correlation (P < 0.05). (c) Cropland partial correlation coefficients between spring and summer soil moisture, averaged over grid cells with the irrigation fraction ranging from 0% to 100% and binned into 5% intervals.

3.2 Irrigation Effects on Surface Water and Energy Balance

Given the combined effects of spring and summer irrigation on summer soil moisture in the NCP, distinguishing the individual influences of spring and summer irrigation on summer climate is more complex and elusive, especially from the perspective of data product analysis. Hence, we turn to the WRF regional climate model to conduct model sensitivity tests, aiming to discern the effects of spring and summer irrigation on summer climate.

The simulated irrigation rates reach two peaks in May and August, corresponding to the winter wheat-summer maize rotation system (Figure 2a). Due to the uneven seasonal pattern of precipitation in the NCP, the simulated IWU of about 136 mm in spring is higher than the value of 97 mm in summer. Along with the additional water input, soil moisture increases accordingly (Figure 2b). In the IRS1 experiment, the changes in soil moisture reach two peaks in May and August similar to the IWU. It should be noted that although the summer IWU is lower than the spring IWU, the two-peak values of the changes in soil moisture in spring and summer are comparable, suggesting that spring irrigation also contributes to the increase in summer soil moisture. In spring, the increase in soil moisture in the top 1-m soil layer is greater than that in the bottom 1-m soil layer. After deactivating summer irrigation, the spring irrigation-induced increase in soil moisture progressively declines to about zero from June to December (Figure 2b). In addition, the increase in summer soil moisture in the top 1-m soil layer is less than that in the bottom 1-m soil layer (Figures S4b and S4d in Supporting Information S1) because the added irrigation water in the top 1-m soil layer first evaporates into the atmosphere. The additional water input alters the surface energy budget, resulting in more of the available energy being partitioned into the latent heat flux and less into the sensible heat flux. In summer, spring and summer irrigation increases the latent heat flux by 9.8 W m−2 and decreases sensible heat flux by −9.0 W m−2 (Figures 2c and 2d, and Figures S5a and S5c in Supporting Information S1), while the spring irrigation-induced changes in latent and sensible heat fluxes are 4.1 and −3.8 W m−2, respectively (Figures S5b and S5d in Supporting Information S1). The changes in latent and sensible heat fluxes result in the evaporative cooling effect, reducing the near-surface air temperature (i.e., surface air temperature at 2 m height). The spring irrigation-induced cooling diminishes over time after May as the change in near-surface air temperature decline to about zero (Figure 2e). The spatial pattern of the irrigation-induced changes in near-surface air temperature is consistent with that of the changes in sensible heat flux (Figures S5 and S6 in Supporting Information S1).

Details are in the caption following the image

Effects of irrigation on surface water and energy balance in the North China Plain. (a) Temporal variations in irrigation rates (mm day−1) in IRS1 and IRS2 experiments. Irrigation-induced monthly changes (i.e., IRS1 minus NOIR and IRS2 minus NOIR) in (b) soil moisture (m3 m−3), (c) latent heat flux (W m−2), (d) sensible heat flux (W m−2), and (e) near-surface air temperature (°C) during 2004–2018, which are averaged in the NCP's irrigated grid cells (Irrigation fraction >20%).

3.3 Irrigation Effects on Summer Extreme Heat Events

Before evaluating the effects of irrigation on the frequency and intensity of summer extreme heat events in the NCP, it is imperative that the WRF model can realistically represent the near-surface air temperature. As shown in Figure S7 in Supporting Information S1, the spatial pattern of Tmax in summer is represented well, but with an overestimation of bias of 2.1°C in the irrigated areas in the NOIR experiment. Although the irrigation cooling effect reduces this seasonal bias by 0.7°C, there still remains a positive bias. We therefore use the bias correction procedure to minimize the systematic bias. After systematic bias corrections, the positive biases of the simulated EHT and EHD are reduced from 3.7 to 1.1°C and from 21.1 to 4.3 days, respectively (Figures S8–S10 in Supporting Information S1). Overall, the bias-corrected Tmax ensures that the simulated summer extreme heat events are closer to reality.

Based on the bias-corrected Tmax of the three experiments, we assess the effect of irrigation on summer extreme heat events in the NCP. In the bias-corrected NOIR experiment, the simulated EHT and EHD in some grid cells even reach 39°C and 30 days, and the regional means of EHT and EHD are 35.8°C and 21.7 days, respectively (Figures 3a and 3d). When irrigation activates in both spring and summer, irrigation-induced cooling notably reduces the regional means of EHT and EHD by −1.0°C and −6.5 days, with the largest declines exceeding −1.6°C and −10.0 days, respectively (Figures 3b, 3e, 3g, and 3h). Although the effects of spring irrigation on EHT and EHD are less than the combined effect of spring and summer irrigation, the larger decreases in EHT and EHD in some grid cells reach −0.6°C and −6.0 days, respectively (Figures 3c, 3f, 3g, and 3h), indicating that the effect of spring irrigation cannot be overlooked. Spring irrigation-induced cooling reduces regional means of EHT and EHD by −0.29°C and −2.5 days, contributing about 30% and 38% of the combined effects, respectively.

Details are in the caption following the image

The role of irrigation on summer extreme heat events. Spatial distributions of summer (a) extreme hot temperatures (EHT, °C) and (d) extreme hot days (EHD, days) in BC-NOIR experiment. Irrigation-induced changes (i.e., BC-IRS1 minus BC-NOIR and BC-IRS2 minus BC-NOIR) in summer (b and c) EHT and (e and f) EHD. Probability density functions (PDF) of irrigation-induced changes in summer (g) EHT and (h) EHD in the NCP's irrigated grid cells (Irrigation fraction >20%). The vertical lines in (g) and (h) represent the regional means of the changes in EHT and EHD.

The spring irrigation-induced soil moisture anomaly persists later into the following summer due to the carryover effects of soil moisture, thereby impacting the summer climate and extreme heat events. We thus analyze the relationship between the spring irrigation-induced changes in soil moisture and summer extreme heat events and find a negative correlation (Figure 4). Especially when the spring irrigation-induced increase in spring soil moisture exceeds 0.027 m3 m−3, EHT and EHD could decrease by about −0.4°C and −4 days, respectively. In general, our finding that spring irrigation preconditions summer soil to be wetter thus suggests a shift in the distribution of the summer near-surface air temperature toward fewer hot extremes, reducing the summer extreme heat events.

Details are in the caption following the image

WRF-simulated changes in summer extreme heat event indices related to changes in spring soil moisture induced by irrigation in the North China Plain. The changes in summer extreme hot temperatures (EHT, °C) and extreme hot days (EHD, days) due to spring irrigation for different spring soil moisture (mean soil moisture in the top 2 m soil layer) changes binned into 0.003 (m3 m−3) intervals. The insets are the kernel density functions. The summer EHT and EHD values are obtained as the difference between IRS2 and NOIR experiments.

4 Summary and Discussion

Irrigation activities have been demonstrated to cool near-surface temperatures in both climate modeling and satellite-based studies (Chen & Dirmeyer, 2019; Cook et al., 2015; Thiery et al., 2020; Yang et al., 2020), and further shift the distribution of near-surface temperature toward fewer extreme heat events (Thiery et al., 2017). Understanding the impacts of irrigation on climate extremes rather than climate averages is more crucial for the sustainability of social and natural systems. In this study, we use the WRF model coupled with a demand-driven irrigation algorithm to conduct ideal experiments to evaluate the role of irrigation in mitigating summer heat events and discern the separate contributions of spring and summer irrigation to these mitigations in the NCP.

From the analysis of data products, there is a positive pattern between spring and summer soil moisture in this region (Figure 1b), suggesting that the spring irrigation-induced water surplus could persist into the following summer and thus influence the interseasonal climate. From the regional climate simulations, the increase in soil moisture induced by spring irrigation indeed lasts several months and affects the summer surface energy balance combined with summer irrigation (Figure 2). This lag effect of irrigation on soil moisture has been reported in several studies (e.g., Ding et al., 2020; Qian et al., 2013; Yang et al., 2016). For example, Qian et al. (2013) found that in the southern Great Plains, the irrigation-induced increases in soil moisture in late spring and summer persist into the early fall. Yang et al. (2016) noted that in the Huang-Huai-Hai Plain, spring irrigation can result in a wetting of the summer soil moisture. Ding et al. (2020) indicated that in the middle and lower reaches of the Yangtze River, excessive irrigation during the rice growing season could result in waterlogging in the field during the following winter wheat growth period. The aforementioned studies corroborate the cross-seasonal effect of irrigation and reflect the importance of this topic.

We find that irrigation in the NCP significantly reduces summer extreme heat events and the effect of irrigation on extreme temperatures is clearly stronger than that on the mean temperatures, which is consistent with previous studies indicating that irrigation exerts a notable impact on extreme temperatures (Bonfils & Lobell, 2007; Hauser et al., 2019; Hirsch et al., 2017; Mueller et al., 2016). In our recent study (Liu et al., 2023), we found that the irrigation cooling effect in the NCP decreased during past decades due to the decrease in the IWU intensity (Huang et al., 2017; Zhou et al., 2020). The annual IWU intensity in this region declined from about 430 mm in the 1980s to about 270 mm in the 2000s, and the decreasing trend of IWU in spring is higher than that in summer (Liu et al., 2023). The large reduction in spring IWU could weaken the beneficial effects of spring irrigation on mitigating summer extreme heat events. However, within this context, the fingerprint of spring irrigation on summer extreme heat events is still evident and nonnegligible from 2004 to 2018, further underlining the critical role of the cross-seasonal effect of irrigation in mitigating extreme heat events. Moreover, the reduction in summer extreme heat events due to spring irrigation can promote crop growth and reduce the irrigation water uses in summer. In future climate change where both water and heat stresses are expected to become more severe (Kala et al., 2022; Pfahl et al., 2017), the benefit of the cross-seasonal climatic effect of irrigation would be more important and cannot be overlooked in climate change adaptation planning. In addition, it should be emphasized that this study focuses on how spring and summer irrigation jointly affect summer extreme dry heat events, which has profound implications for agriculture. Recent studies have shown that notwithstanding the reduction in dry heat due to irrigation, irrigation-induced increases in specific humidity enhance the moist heat stress, thereby threatening human health (Kang & Eltahir, 2018; Mishra et al., 2020). Therefore, more indices should be considered in the evaluation of irrigation effects on extreme heat events in future studies.

There are several uncertainties and limitations in this study that need to be pointed out. In terms of data products, we select the ERA5 Reanalysis and GLEAM product. Recently, the performance of ERA5 Reanalysis has been evaluated in China based on meteorological station observational data (Jiao et al., 2021; Xu et al., 2022). Although the ERA5 Reanalysis can well capture the annual and seasonal patterns of observed climate variables in China, there remain uncertainties in this data set resulting from the quality of observational inputs, limitations in numerical weather prediction models, and uncertainties in boundary conditions (Hersbach et al., 2020). Moreover, the GLEAM product is the most widely used soil moisture data set, extensively adopted for examining land-atmosphere interactions and evaluating the evolution of soil moisture (Li et al., 2018a, 2018b; Wouters et al., 2022). However, some uncertainties exist in the GLEAM product because the root-zone soil moisture cannot be directly measurable at large spatial scales (Lian et al., 2020). In terms of regional climate simulations, uncertainties remain regarding the robustness of our single-model results, despite calibrating the irrigation algorithm based on the Governmental agricultural census-based IWU data and conducting simulations for 15 years instead of a few years to minimize the uncertainties and provide robust results. Yuan et al. (2023) recently assessed the potential impacts of different irrigation methods on regional climate, highlighting differences in climate feedback among various irrigation methods. In this study, irrigation water is directly added to the soil surface, because surface irrigation is the main method in China (Frenken, 2012; Liu et al., 2021b). Overall, for future studies, we encourage other researchers to use more regional or global climate models as well as different irrigation methods to investigate the cross-seasonal effects of irrigation on extreme heat events.

Acknowledgments

This work was jointly supported by the National Science Foundation of China (U2240218, 42301040), the Fundamental Research Funds for the Central Universities (B230201032), the Jiangsu Funding Program for Excellent Postdoctoral Talent (2023ZB194), the QingLan Project of Jiangsu Province, the National “Ten Thousand Program” Youth Talent, the “333 project” of Jiangsu Province, and the Six Talent Peaks Project in Jiangsu Province. Cordial thanks are extended to the editor and anonymous reviewers for their critical and constructive comments, which highly improve the quality of the manuscript.

    Conflict of Interest

    The authors declare no conflicts of interest relevant to this study.

    Data Availability Statement

    The fifth version of the global map of irrigation areas (Siebert et al., 2013) is available from (http://mygeohub.org/publications/8/2). The ERA5 global reanalysis data (Copernicus Climate Change Service, 2017) are from European Centre for Medium-Range Weather Forecasts (http://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-single-levels; https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-pressure-levels). The Global Land Surface Satellite (GLASS) data sets (Liang et al., 2021) are available from (http://www.glass.umd.edu/Download.html). The modified code can be accessed on GitHub (https://github.com/hhu607/modified-NoahMP).