Location and Intensity Changes of the North Equatorial Countercurrent Tied to ITCZ Under Global Warming

Previous studies have suggested that global ocean circulation would be significantly changed under global warming, while the change of North Equatorial Countercurrent (NECC) and its mechanisms are still unclear. Here, we investigate the location and intensity changes of NECC under global warming based on CESM1 high‐resolution long‐term simulations from the perspective of the Inter‐Tropical Convergence Zone (ITCZ), considering the close connection between NECC and ITCZ and well‐established changes of ITCZ. It is found that the annual‐mean NECC shifts equatorward of 0.39° and weakens by 21.71% in the RCP8.5 scenario at the end of 21st century. The NECC change is seasonally dependent, with maximum shift and weakening during spring, consistent with the changes of ITCZ. Both the location and intensity changes of ITCZ are important in the NECC changes, especially for spring. The weakening and equatorward shift of ITCZ contributes almost equally to the strongest decrease of spring NECC (47.63%).

Previous studies have found that under global warming, global ocean circulation would be enhanced at the surface especially in the midlatitudes and subtropics (Hu et al., 2020;Peng et al., 2022;Wang et al., 2015).Peng et al. (2022) found the strengthened surface ocean circulation is mainly caused by the enhanced surface ocean warming by separating different drivers.However, this dominance is mainly over the subtropics and middle latitudes, while in the tropics, the wind forcing also becomes important (Hu et al., 2021;Qiu & Chen, 2012; see Figure 1 in Peng et al., 2022).As an important tropical atmospheric circulation system, ITCZ would contract equatorward due to the enhanced equatorial warming, with largest equatorward shift in boreal spring (Donohoe et al., 2019, Huang et al., 2013;Zhou et al., 2019Zhou et al., , 2020;;see Song et al., 2023 for review).This phenomenon has been identified in both precipitation and atmospheric circulation fields.However, how the intensity and location of NECC would change under global warming receives much less attention but deserves investigation considering their close connection (Hu et al., 2015;Ju et al., 2022;Luo & Rothstein, 2011).Luo and Rothstein (2011) found there is a weakening of NECC under global warming, which is suggested to be related to the weakening of tropical atmospheric circulation as constrained by the atmospheric energetics (Held & Soden, 2006;Vecchi & Soden, 2007).The weakening of NECC is also confirmed by Hu et al. (2015) based on CMIP5 and Ju et al. (2022) based on CMIP6 models.Although the weakening of NECC under global warming is consistent among different generations of climate models, whether the location and seasonality of NECC have changed is still largely unknown.Here we aim to explore the seasonal cycle of the intensity and location changes of NECC under global warming based on Community Earth System Model Version 1 (CESM1) high-resolution atmosphere-ocean coupled model long-term simulation, and explore the underlying mechanisms.In particular, the changes in location and intensity of ITCZ and NECC would be compared in the seasonal cycle.Section 2 shows the data and methods.Section 3 shows the main results and conclusions are shown in Section 4.

Observational Data
In this study, we use the following observational and reanalysis data sets: (a) The monthly zonal current from the Ocean Surface Current Analysis Real-time (OSCR) data, with a resolution of 1/3°× 1/3° (Bonjean & Lagerloef, 2002); (b) The monthly zonal and meridional wind at 1,000 hpa from the ERA5 reanalysis at a resolution of 0.25°× 0.25° (Hersbach et al., 2020); (c) Precipitation from the Global Precipitation Climatology Project (GPCP; Adler et al., 2003) with a resolution of 2.5°× 2.5°.The period of observational and reanalysis data is from 1993 to 2021.

CESM1 High-Resolution Simulations
To explore the future changes of ITCZ and NECC, the historical and RCP8.5 simulations during 1986-2099 from high-resolution CESM1 model (Chang et al., 2020;Kay and Coauthors, 2015) are used.The horizontal resolution is ∼25 km for the atmospheric component and ∼10 km for the oceanic component.Compared to the lowresolution CESM1 model, the high-resolution model simulations give more accurate descriptions of NECC and ITCZ, especially for the location changes.The following variables are used: precipitation and surface wind at resolution of 0.23°× 0.31°, and zonal current at resolution of 0.1°× 0.1°.

Definition of Location and Intensity of ITCZ and NECC
The changes of NECC and ITCZ may include changes in location and/or intensity, so it is necessary to separate these two components.The location refers to the latitude-weighted centroid of the variable considered, for example, precipitation, similar to Woelfle et al. (2019): In Equation 1, φ is latitude and P represents the variable examined, such as precipitation, wind convergence, vorticity and zonal current.Besides, a,b in Equation 1 are the latitude boundary (a is 0°and b is 15°N).The longitude range is over the Pacific (150°E 90°W).For precipitation, we only consider regions with precipitation greater than 5 mm/day but for the wind convergence, vorticity, zonal current and dvor dy (vor means vorticity, y is meridional distance), regions with values greater than 0 are considered.The intensity is defined as below: To isolate the intensity change, the area for calculating the intensity is not fixed, but changes along with the location shift.We use precipitation, wind convergence and vorticity to depict the location and intensity of ITCZ and zonal current to depict the characteristics of NECC.To examine whether the relationship is sensitive to the method to define the location and intensity of NECC, we also use the maximum value in each longitude and its latitude position to define the intensity and location, respectively.It is found that the results shown below are not sensitive to the definition of location and intensity.

Results
We first evaluate the climatology of ITCZ and NECC in CESM1 high-resolution simulation by comparing with observation.The ITCZ featuring strong precipitation and wind convergence and NECC featuring zonal current belts are generally well reproduced by the model, but with larger magnitude than the observation (Figure S1 in Supporting Information S1).The stronger ITCZ and NECC in the model are evident throughout the year (Figure S2 in Supporting Information S1).However, the locations of ITCZ and NECC, represented by precipitation/wind convergence and zonal current respectively, in the model are consistent with the observation.Hence, the CESM1 high-resolution model performs well in capturing the main characteristics of the ITCZ and NECC.
In order to explore the changes in the location/intensity and seasonality of NECC, we take a look at the differences of NECC between future state during 2080-2099 and historical state during 1986-2005 in Figure 1.It is evident that NECC shows a decreasing trend over the whole climatological NECC region and this weakening can extend up to 150 m below the water (Figures 1a and c).Meanwhile, the Equatorial Undercurrent would get enhanced and upward shift under global warming (Figure 1c).In addition, NECC shows a southward shift, which is most evident in spring and quite uniform in other seasons (Figure 1b).Furthermore, we take a quantitative look at the location and intensity change of NECC (Figures 1d and 1e).It shows that the southward shift of NECC is about 0.39°for annual mean and will weaken about 22% in the future.More importantly, the NECC changes are seasonally dependent, with the maximum southward shift and weakening in the boreal spring.The southward shift of NECC can reach 0.68°and the weakening of NECC can reach ∼48% in spring.These are also evident when taking long term trends during 1986-2099 (Figure S3 in Supporting Information S1).
In order to examine whether the location and intensity change of NECC and ITCZ are consistent with each other, we further examine the future changes of ITCZ represented by precipitation and wind convergence.It is evident that under global warming, precipitation and wind convergence will decrease in the main body of ITCZ area and increase near the south edge of ITCZ, indicating a southward shift of ITCZ (Figures 2a,2b,and 2e).Note that although the main body of ITCZ precipitation would decrease, the intensity of precipitation would increase (Figure 3f).In particular, the southward shift of precipitation and wind convergence is more evident in the eastern Pacific than the central Pacific, consistent with the NECC change (Figure S4 in Supporting Information S1).It is linked to the greater sea surface temperature change in the eastern Pacific, that is, the well-known El Nino-like warming pattern (e.g., Coats & Karnauskas, 2017;P. N. DiNezio, et al., 2009;Kociuba & Power, 2015;Plesca et al., 2018;Sun & Lu, 2021;Wang et al., 2024;Ying et al., 2016).In addition, the evident NECC changes in the western Pacific west of dateline (Figure 1a) in the absence of wind anomalies (Figure 2b) may be related to the westward-propagating Rossby waves as suggested in previous studies (e.g., Chen et al., 2016;Hsin & Qiu, 2012b).Here, although there are some zonally-varying features of NECC changes, we focus on the zonallyuniform changes of NECC, which are closely linked to ITCZ.The southward shift of ITCZ is also seasonally dependent and consistent in both precipitation and wind convergence (Figures 2c-2e).Noticeably, the southward shift of ITCZ is most evident in spring, with 1.2°in the ITCZ precipitation and ∼0.88°in the ITCZ convergence.However, the intensity changes of ITCZ represented by precipitation and wind convergence are opposite: the precipitation intensity is enhanced while the wind convergence intensity is reduced (Figures 2f and 2h).To understand this, a simplified moisture budget is utilized: .This means the enhanced surface specific humidity would dominate the increase of precipitation intensity (Figure S5c in Supporting Information S1), although the intensity of surface wind convergence would decrease in the future following the weakening of tropical circulation (Figure S5d in  2h).Quantitatively, the southward shift of ITCZ is over 1°when represented by precipitation and about 0.88°when represented by convergence in spring, much larger than the shifts in other seasons, while the southward shift of NECC is slightly weaker (∼0.68°).In contrast, the intensity of NECC exhibits a significant decrease, with a maximum decrease about 47.63% in spring, while the ITCZ represented by convergence decreases by 19.3% and the ITCZ represented by precipitation increases only 1.49% in boreal spring.Hence, there are two questions deserving further investigation: (a) In the annual mean, are the location and intensity changes of ITCZ and NECC related with each other?(b) During spring, is the more evident southward shift of ITCZ associated with more evident decrease of NECC?
In order to address these two questions, we analyze the relationship between the NECC and ITCZ in terms of both location and intensity changes.In what follows, the wind convergence is used to represent the ITCZ, as it is the dynamical component of ITCZ that interacts with the NECC change.Under global warming, the wind speed is weakened over the equatorial Pacific due to the weakened Walker circulation (Held & Soden, 2006;Vecchi & Soden, 2007).The weakened wind would lead to the reduced convergence as shown in Figures 2b-2d and 2f, 2h.In addition, it is well known that ITCZ would shift equatorward seasonally under global warming (e.g., Zhou et al., 2019).The NECC originates from the wind stress curl over the ITCZ (Masunaga & L'Ecuyer, 2011;Sun et al., 2019;Sverdrup, 1947).According to the mass transport equation of zonal current: where φ is latitude, , a is the radius of the Earth, τ x is zonal wind stress.By calculation, it is found the first term ( tan φ a × ∂τ x ∂y ) takes up a small proportion over the NECC region and the mass transport is mainly determined by the second term ( ∂ 2 τ x ∂y 2 ).Since the second term approximates to dvor dy (vor is the wind stress curl), the future change of NECC's mass transport is consistent with the change of dvor dy .As dvor dy is closely related to both the location and intensity of ITCZ (Figure S7 in Supporting Information S1), the corresponding relationships are investigated in Figure 3.The intensity of NECC is found to be related to the intensity of ITCZ convergence (Figure 3a, r = 0.85), which is consistent with previous findings that the weakening tropical circulation would lead to the weakened NECC (e.g., Luo & Rothstein, 2011).Meanwhile, the location of NECC also closely follows both the location (Figure 3b, r = 0.87) and intensity (Figure 3c, r = 0.65) of ITCZ.These relationships largely hold when detrending the variables.This indicates that under global warming, both the southward shift and weakening of ITCZ is in favor of the southward shift of NECC.
Seasonally, both the southward shift of ITCZ and weakening of NECC peak in spring as shown in Figures 1 and 2. Therefore, we further explore the relationship between ITCZ changes and NECC intensity changes in spring.It is found both the location and intensity of ITCZ convergence are closely related to the intensity of NECC in spring (Figures 4a and 4b), with correlation coefficients of 0.55 and 0.87, respectively.Previous studies only focusing on the annual-mean change suggest the role of ITCZ intensity change (Ju et al., 2022;Luo & Rothstein, 2011).Here, we found that the location changes are also important for the NECC intensity changes in spring.Similarly, this close relationship still holds when we use the ITCZ vorticity instead of the ITCZ convergence (Figures 4c and 4d).Here, we consider the vorticity as it is more directly linked to the NECC.By showing the future changes of vorticity in Figure 4e, it is found that the vorticity is decreased and increased to the north and south of the NECC respectively.Moreover, the axis of NECC is located in the region where dvor dy has a maximum weakening (Figure 4f).As the mass transport of NECC is proportional to dvor dy based on Equation 4, the weakening of dvor dy is closely related to the weakening NECC.To further quantify the contribution from the intensity and location of ITCZ to the NECC weakening, we reconstruct a linear regression model by considering both the intensity and location of ITCZ as predictors.The intensity of NECC based on this regression model has a good relationship with the simulated one (r = 0.75).Based on this regression model, it is found that the future weakening of ITCZ will cause the decrease of NECC's intensity by 4.01 cm ⋅ s 1 , while the southward shift of ITCZ will lead to the decreased intensity of NECC by 3.63 cm ⋅ s 1 .The actual weakening of spring NECC is about 4.71 cm ⋅ s 1 in the future high-resolution model, close to the sum of these two components.Therefore, the location and intensity change of ITCZ have a comparable contribution.In summary, the intensity of NECC is influenced by both the location and intensity of ITCZ in spring.

Conclusions
Based on historical and RCP8.5 simulations from CESM1 high-resolution model, the future changes of location and intensity of the NECC and their relationships to the changes of ITCZ in the Pacific Ocean under global warming are examined.
In the future, the intensity of ITCZ rainfall would increase, which is dominated by the increased moisture under warming, but the intensity of wind convergence would decrease along with the weakening tropical circulation.Meanwhile, ITCZ shows a clear trend of equatorward contraction, especially in the boreal spring.Accordingly, NECC would move equatorward and weaken in the future, with a maximum weakening in the spring.However, the southward shift of NECC is a bit smaller than that of wind convergence in the boreal spring (NECC moves ∼0.68°southward, whereas wind convergence moves ∼0.88°southward), but the decrease is much stronger than the ITCZ wind (NECC decreases about 22% vs. ITCZ decrease about 14% in the annual mean and NECC decrease about 48% vs. ITCZ decreases about 19% in spring).For the annual-mean change, the weakening of NECC is mainly caused by the weakened ITCZ.However, for the springtime, the equatorward shift of ITCZ also has a comparable contribution to the weakening of NECC.The equatorward shift of NECC is contributed by both the weakening and southward shift of ITCZ.These relationships between the changes of NECC and ITCZ also hold in the low-resolution model.Given the narrow nature of NECC, it should be more precise in the high-resolution model.Considering the large uncertainty of NECC changes under global warming, it is important to examine the location and seasonality changes of NECC and their relationships with ITCZ in other CMIP6 high-resolution models.
surface specific humidity, c is the surface wind convergence, ∆ represents the future change, and the overbar represents the historical climatology.Equation 3 is modified from Huang et al. (2013) by replacing the vertical velocity with the surface wind convergence as the surface wind can better link the ITCZ with NECC.The future precipitation change ∆P can be well captured by the sum of ∆c ⋅ q + c ⋅ ∆q (Figures S5a and S5b in Supporting Information S1), suggesting it is reasonable to use Equation 3 to mimic the precipitation change and compare the relative roles of water vapor and surface wind convergence.It is found that the second term (c ⋅ ∆q) is much larger than the first term (∆c • q)

Figure 1 .
Figure 1.Future changes (shaded, 2080-2099 minus 1986-2005) and historical climatology (contour) of NECC in (a) spatial (m ⋅ s 1 , with an interval of 0.2 m ⋅ s 1 ), (b) seasonal (m ⋅ s 1 , with an interval of 0.2 m ⋅ s 1 ), (c) vertical (m ⋅ s 1 , with an interval of 0.1 m ⋅ s 1 ), (d) southward shift, (e) intensity percentage change.The thick contours represent 0, solid contours represent positive values and dashed contours represent negative values.The magenta line in (b) represents future centroid, the green line represents historical centroid.The magenta and green dots in (c) represent the maximum values of Equatorial Under Current in future and historical climatology, respectively.The longitude range in (b-e) is 150°E ∼ 90°W.

Figure 2 .
Figure 2. Future changes (shaded, 2080-2099 minus 1986-2014) and historical climatology (contour) of Inter-Tropical Convergence Zone represented by precipitation and wind convergence in (a): precipitation spatial change (mm ⋅ day 1 , with an interval of 4 mm ⋅ day 1 ), (b): convergence spatial change (×10 5 s 1 , with an interval of 0.5 × 10 5 s 1 ), (c): precipitation seasonal change (mm ⋅ day 1 , with an interval of 6 mm ⋅ day 1 ); (d): convergence seasonal change (×10 5 s 1 , with an interval of 0.4 × 10 5 s 1 ).(e): Solid line is the location change of precipitation and wind convergence from 1986 to 2099, and the dashed line is their linear fit.(f): Solid line is the intensity change of precipitation (mm ⋅ day 1 ) and wind convergence (×10 5 s 1 ) in 1986-2099, and the dashed line is their linear fit.The changes in (e, f) are significant at 95% confidence level.(g): Southward shift of precipitation and convergence.(h): Intensity change of precipitation and convergence.The thick contours represent 5 mm ⋅ day 1 in (a), 0 in (b).The solid contours represent positive values and dash lines represent negative values in (b, d).The magenta line in (c, d) represents future centroid, the green line represents historical centroid.The longitude range in (c-h) is 150°E ∼ 90°W.

Figure 3 .
Figure 3.The 5-year running mean relationship between NECC and Inter-Tropical Convergence Zone (ITCZ) during 1986-2099.The blue line represents (a) the intensity of NECC (m ⋅ s 1 ), (b, c) the location of NECC.The red line in (a, c) is the intensity of ITCZ represented by convergence (×10 5 s 1 ), but it is the location of ITCZ represented by convergence in (b).The correlation coefficients and p-values are noted in the upper right corner of each panel.

Figure 4 .
Figure 4.The 5-year running mean relationship between NECC and Inter-Tropical Convergence Zone during 1986-2099 in spring.The blue line represents (a-d) the intensity of NECC (m ⋅ s 1 ).The red line is: the location of (a) convergence and (c) vorticity, the intensity of (b) convergence (×10 5 s 1 ) and (d) vorticity (×10 5 s 1 ).The correlation coefficients and p-values are noted in the upper right corner of each panel.Future changes (shaded) and historical climatology (contour) of (e) vorticity (×10 5 s 1 ) and dvor dy (×10 10 m 1 s 1 ) in spring.The thick contours represent 0, the solid contours represent positive values and dashed contours are negative values.The magenta line is the historical location of NECC in spring during 1986-2005.