Low latitude mesospheric clouds in a warmer climate

Observations show that mesospheric clouds (MCs) have been increasing in recent decades, presumably due to increased mesospheric water vapor which is mainly caused by greater methane (CH4) oxidation in the middle atmosphere. Past warm climates such as those of the early Cretaceous and Paleogene periods are thought to have had higher CH4 concentrations than present day, and future CH4 concentrations will also likely continue to rise. Here, idealized atmosphere chemistry‐climate model experiments forced with strong polar‐amplified sea‐surface temperatures and elevated carbon dioxide (CO2) and CH4 concentrations predict a substantial spreading of MCs to middle and low latitudes, well beyond regions where they are currently found. Sensitivity tests show that increased water vapor from CH4 oxidation and cooling from increased CO2 and CH4 concentrations create favorable conditions for cloud formation, producing MC fractions of 0.02 in the low latitudes and 0.1 in the mid‐latitudes in the Northern Hemisphere when CH4 concentration is 16× higher than pre‐industrial. Further increases in CH4 result in a monotonic increase in low‐ and mid‐latitude MCs. A uniform surface ocean warming, changes in polar amplification, or the solar constant do not significantly affect our results. While the appearance of these clouds is interesting, their ice and liquid water content is not sufficient to cause a significant radiative effect. On the other hand, dehydration of the mesosphere due to these low‐ and mid‐latitude MCs could potentially lead to a reduction in atomic hydrogen, thereby affecting mesospheric ozone concentration, although further study is required to confirm this.


| INTRODUCTION
Mesospheric clouds (MC) or "night-shining" clouds form in the summer near the mesopause ($85 km in summer) region and consist of ice particles (Hervig et al., 2001;Hesstvedt, 1962).They become visible when the sun illuminates the ice particles from below the horizon at a depression angle ranging between 8 and 12 (Fogle & Haurwitz, 1966) and are generally observed at middle to high latitudes ($55 to 65 ) (Fogle & Haurwitz, 1966;Olivero & Thomas, 1986).Formation and disappearance of MCs are abrupt and mainly depend on background temperature and water vapor concentration in the mesosphere.When the mesopause temperature reduces below the frost point ($150 K) and the air becomes supersaturated, ice particles start forming (Thomas, 1991).A reduction in mesospheric temperature is associated with an increase in density and concentration of ice in MC (Hervig et al., 2009).MCs redistribute water vapor downward in the mesosphere by dehydrating regions of ice particle growth and increasing water vapor during sublimation after falling into the warmer underlying air (Berger & Von Zahn, 2002;Witt, 1968).
The formation of MCs influences the chemical composition in the mesosphere by-(a) catalytic removal of oxygen atoms on the surface of ice (Murray & Plane, 2003;Olivero, 1974), (b) redistribution of water vapor concentration caused by dehydration when water vapor condenses onto ice (Siskind et al., 2018) and (c) photolysis of ice at the MC base that releases hydrogen atoms (Murray & Plane, 2005;Yabushita et al., 2004).All these processes directly or indirectly influence the concentration of ozone in the mesosphere (Murray & Plane, 2005;Siskind et al., 2018).
Observational studies suggest an increase in occurrence frequency of MC in recent years (Berger & Lübken, 2015;Dalin et al., 2023;Hervig et al., 2016;Lübken et al., 2018).There have also been recent observations of MCs at unusually low latitudes (Dalin et al., 2023;Nielsen et al., 2011;Suzuki et al., 2016).This is caused by a combination of different factors, such as an increase in carbon dioxide (CO 2 ), methane (CH 4 ) and water vapor concentration between approximately 79 and 89 km altitude (Dalin et al., 2023;Lübken et al., 2018).An increase in CO 2 and CH 4 is associated with a reduction in temperature due to infrared cooling and dynamical feedback in the mesosphere.Additionally, each mol of CH 4 oxidizes to produce about 2 mol water vapor in the upper stratosphere and mesosphere (Roble & Dickinson, 1989;Thomas, 1996;Thomas et al., 1989).However, other factors such as the cold-point temperature at the tropical tropopause (Mote et al., 1996;Randel & Park, 2019) and deep convection (Sherwood, 2002;Sherwood & Dessler, 2000) may also influence the amount of water vapor entering the tropical stratosphere and mesosphere.Therefore, changes in mesospheric temperature and water vapor concentration influence the plausibility of MC occurrences in a warmer than present-day climate.This sparked our curiosity about the conditions under which MCs could appear.
Proxy reconstructions and modeling studies for past warm climates, such as the early Eocene ($56 to 47.8 million years ago), suggest a much lower equator-to-pole temperature gradient than today.These climates are also associated with higher than present-day CO 2 (Rae et al., 2021) and CH 4 concentrations.While there are no paleoproxies for CH 4 during climates predating ice core records, previous modeling studies with variable complexity suggest a higher than pre-industrial CH 4 concentration (Beerling et al., 2011;Wilton et al., 2019).
In addition to surface ocean temperature and greenhouse gases, a change in solar activity may also change the water vapor concentration, thereby influencing MC formation.Since its formation, the intensity of the Sun has increased steadily from $70% of its current value in the last 4.7 billion years (Gough, 1981).Diminished solar activity in the past could have contributed to temperature reduction and increased water vapor concentration in the mesosphere.These changes might have led to increased ice content in MCs, resulting in a higher frequency of MC occurrences (DeLand et al., 2006;Garcia, 1989;Hervig et al., 2019).
In this study, we forced the Whole Atmosphere Community Climate Model (WACCM) with an idealized polar-amplified SST anomaly, rather than focusing on SSTs and greenhouse gas concentrations corresponding to any particular geological period.WACCM is a high-top chemistry climate model with a well-resolved middle atmosphere.Our approach enabled us to investigate potential variations in the middle and upper atmosphere across various climatic conditions.Additionally, we examined how the MCs respond to a reduced solar constant.We unexpectedly discovered that low-latitude (between 30 N and 30 S) MCs are possible under realizable past and future conditions.

| Model description
We used the WACCM version 4 (WACCM4) (Marsh et al., 2013) of the Community Earth System Model (CESM) version 1.2.2 (Hurrell et al., 2013).This model includes all the physical parameterizations of Community Atmosphere Model version 4 (CAM4) but offers much more careful treatment of the middle and upper atmosphere including the chemistry.WACCM4 extends from the surface to the thermosphere ($140 km) and has a horizontal resolution of 1.9 latitude by 2.5 longitude.WACCM4 includes 66 vertical layers with a resolution of $3.5 km above 65 km, 1.75 km near the stratopause ($50 km) and 1.1-1.4km in the lower stratosphere (below 30 km) (Garcia et al., 2007).
The chemistry module of WACCM4 is based on the Model for Ozone and Related Chemical Tracers (MOZART), version 3 (Kinnison et al., 2007).In CAM4 clouds form based on simple relationships between temperature and water vapor concentration when relative humidity exceeds a predefined threshold (Neale et al., 2012).Bardeen et al. (2010) suggested that WACCM coupled to the Community Aerosol and Radiation Model for Atmospheres (CARMA) microphysical model provides better simulation of polar MC microphysical processes, such as ice particle growth, sublimation and sedimentation (Bardeen et al., 2010;Siskind et al., 2018).Using ensembles of simulations with the CARMA model, Merkel et al. (2009) developed a bulk parameterization scheme to simulate polar MCs (WACCM-PMC) in a cost-effective way compared to running WACCM with CARMA.WACCM-PMC calculates the saturation ratio by dividing the partial pressure of water vapor by the saturation pressure of water over ice.Upon air supersaturation, some of the available water vapor transforms to ice, initiating MC nucleation.Thereafter WACCM-PMC uses an empirical formula to calculate particle growth using local temperature and available ice content.The parameterized MC-induced water vapor redistribution is then fed back to WACCM chemistry module.Comparisons between WACCM-PMC and the Solar Occultation for Ice Experiment onboard the Aeronomy of Ice in the Mesosphere satellite suggest an improved representation of MCs and background atmospheric state, including temperature and water vapor (Merkel et al., 2009;Siskind et al., 2018).
Due to computational costs, most experiments were run with WACCM only.But we conducted a second experiment with WACCM-PMC for the polar amplified 128Â CH 4 experiment (Pol20_4C_128M; Table 1) as we found a substantial increase in summer MC fraction across the low-, mid-, and high latitudes for CH 4 concentrations of 128Â or higher.We compared the WACCMonly and WACCM-PMC results in Section 3.

| Experimental setup
We performed the control experiment with pre-industrial climatological SST and sea ice conditions, and with an annual cycle of latitudinally varying CO 2 and CH 4 concentrations of around 280 ppm and 791 ppb, respectively.Following the default pre-industrial setup, the solar parameters (solar irradiance, wavelength, etc.) in these experiments are based on climatology for the period between 1834 and 1867, and the 11-year solar cycle is not included (Marsh et al., 2013).These experiments were integrated for 60 years, and the last 50 years of model output is used in our analysis.We calculated the average of different variables from June to August and December to February when presenting results for the Northern (NH) and Southern (SH) hemisphere summers.Table 1 lists the experiments.
In the first set of sensitivity experiments (hereafter Pol10 and Pol20) we added polar amplified SST anomalies with varying magnitude to the pre-industrial SST and removed the sea ice cover following Dutta et al. (2021).These experiments were conducted to explore the potential impacts of polar amplified SST on the mesosphere.The SST anomalies added in the Pol10 and Pol20 experiments are as follows: T A B L E 1 List of experiments simulated with WACCM-only.We conducted another experiment with WACCM-PMC using the boundary conditions for the Pol20_4C_128M experiment to compare the results between WACCM-only and WACCM-PMC.

Experiment
We found a small decrease in upper-mesospheric clouds in the NH high latitudes in response to polar amplified SSTs.To test the sensitivity of MCs to uniform SST warming, we conducted another experiment by adding a globally uniform 4 K SST to Pol10 (Pol10_4K).
An increase in greenhouse gases leads to infrared cooling thereby favoring MC formation.Therefore another experiment was conducted by increasing CO 2 concentration by 4Â that of pre-industrial while keeping other boundary conditions the same as in Pol20.We similarly conducted F I G U R E 1 Boreal summer (June-August) mean zonal mean temperature (a-f), water vapor concentration (g-l) and cloud fraction (m-r) in the mesosphere (between 1 and 0.01 hPa) in different experiments.
CH 4 sensitivity experiments, increasing its concentration by 16Â, 64Â, 128Â and 256Â that of pre-industrial, with Pol20 SST and 4Â pre-industrial CO 2 .More details on the perturbation experiments can be found in Dutta et al. (2021).
Earlier studies suggest that a rise in solar influx results in an increase in temperature and photodissociation of water vapor in the mesosphere, consequently diminishing MC fraction (Garcia, 1989).This indicates that a smaller than present-day solar constant in past climates could potentially provide more favorable conditions for MC formation.We therefore examined MC sensitivity to a 10% reduction in solar constant, in the case of Pol20 SSTs, 4Â CO 2 and 128Â CH 4 concentration, denoting this experiment Pol20_128M_Solar.

| RESULTS
In the boreal summer (June-August) of the PI experiment, mesospheric temperatures below 165 K lead to the formation of MCs in the pressure levels between 0.02 and 0.01 hPa poleward of 60 N (Figure 1).Polar-amplified surface warming raises mesospheric temperatures above 0.1 hPa in high-and mid-latitudes (Figure 1a,b), reducing the MC fraction in Pol10 and Pol20 compared to PI (Figure 1).However, an increase in greenhouse gas concentrations combined with Pol20 SST causes an increase in MCs, particularly when the concentration of CH 4 is enhanced.Notably, we find that MCs appear at low latitudes when CH 4 concentrations are increased by 64Â or more.A continuous layer of MCs between 0.1 and 0.01 hPa is seen across all latitude bands in summer if CH 4 is increased by 128Â (Figure 1g-i) or more.However, an increase in CO 2 alone only slightly increases highlatitude MCs in Pol20_4C and further increases in CO 2 leads to a reduction in MCs (not shown).Our results indicate that as the concentration of CH 4 increases, the resulting mesospheric temperature reduction and water vapor increase could lead to a substantial increase in MCs.
Polar amplified surface warming in Pol10 and Pol20 results in a warming of approximately 3-4 K in the tropical cold-point temperature (Figure 2a,b).Consequently, there is an increase in water vapor entering the tropical stratosphere in Pol10 and Pol20 (Figure 2c,d).Water vapor at different latitude bands gradually increases with an increase in the tropical mixing ratio (Figure 2c,d).The increase in CO 2 alone in Pol20_4C has a very small impact on cold-point temperature and water vapor mixing ratio in the tropics.There is a significant increase in water vapor at 0.1 hPa compared to 70 hPa in the high CH 4 experiments for the same tropical water vapor mixing ratio.This mesospheric water vapor increase can be primarily attributed to CH 4 oxidation.Further details can be found in Dutta et al. (2022).
Modeling (Becker & Schmitz, 2003;Karlsson, McLandress, & Shepherd, 2009) and observational (Gumbel & Karlsson, 2011;Karlsson, Randall, et al., 2009) studies suggest that the winter stratosphere and troposphere may influence the summer mesosphere though gravity wave filtering.In this case, a strengthening of the NH polar stratospheric vortex during boreal winter reduces gravity waves reaching the mesosphere, which weakens the meridional overturning circulation.As a result, both downwelling in the NH mesosphere and upwelling in the SH mesosphere decrease, creating a negative temperature anomaly in the NH and a positive temperature anomaly in the SH.While Dutta et al. (2022) showed that polar amplified SSTs in Pol10 and Pol20 cause a stronger Arctic vortex in the boreal winter, we find an increase in polar mesospheric temperatures in both hemispheres and a reduction in low latitude mesospheric temperatures in these experiments (Figure 3a,b).This indicates involvement of processes other than a change in strength of the mesospheric meridional overturning circulation, such as radiative and chemical heating or a rearrangement of the meridional circulation, in changing mesospheric temperatures.Whatever its cause, this warming restricts MC formation (Figure 3g-i) despite a small increase in mesospheric water vapor (Figure 3d-f).
A four-fold increase in CO 2 in Pol20_4C results in a radiative cooling, reducing temperature by up to $20 K between 1 and 0.1 hPa compared to Pol20 (Figure 1d).However, this produces high-latitude MCs only above 0.02 hPa (Figure 1p).A further increase in CO 2 further cools the mesosphere at levels below 0.1 hPa but causes warming above (not shown).
An increase in CH 4 results in both a radiative cooling (Figure 3a-c) and an increase in water vapor (Figure 3d-f) in the mesosphere.With a 16-fold increase in CH 4 , the increase in MCs is limited above 0.05 hPa in the high and middle latitudes (Figure 3g,h).Further CH 4 increases result in additional temperature reductions in the low latitudes, leading to low latitude MCs.In the low and high latitudes, we find higher MCs in the NH than in the SH (Figure 3g-i).This can be mainly attributed to different temperature response between the hemispheres (Figure 3a-c) possibly caused by differences in large-scale dynamical processes.
Since we see a relatively large MC increase in the low, middle, and high latitudes in the polar amplified 128Â CH 4 experiment, we compared the free-running WACCM results with WACCM coupled with CARMA (using the in-build WACCM-PMC scheme) for this experiment.Below the 0.05 hPa level WACCM-PMC shows a slightly higher MC fraction compared to WACCM, while the opposite is true above 0.05 hPa (Figure 4).However, the MC ice content in the WACCM-PMC is lower than the free running WACCM in all latitude bands (Figure 4).These differences may be attributed to modification of water vapor during two-way interaction between WACCM and WACCM-PMC (Bardeen et al., 2010).During each model step, water vapor calculated by WACCM is input into WACCM-PMC, which adjusts the water vapor amount based on deposition or sublimation.The modified water vapor is then fed back into WACCM.
Despite the substantial increase in MC fraction (Figures 3g-i and 4a-c), these clouds do not provide a significant radiative effect in any of our experiments (not shown).Primarily composed of ice (Figure 4d-f), these MCs experience rapid sedimentation of cloud particles due to low air density in the mesosphere, resulting in exceedingly low water content (Figure 4g-i).Consequently, they have extremely low optical thickness.
Additionally, we tested the impacts of a 10% reduction in solar constant on MCs in the polar amplified 128Â CH 4 experiment.We find a small increase in MC fraction in both hemispheres, particularly in the low and middle latitudes (Figure 4a-c), but a reduced ice content (Figure 4d-f).The modest size of these changes suggests that the reduction in solar constant in deep time may not have played an important role in changing MCs.However, this may be attributed to abundance of mesospheric water vapor in the 128Â CH 4 experiment, thereby reducing the sensitivity of MCs to solar constant change.

| DISCUSSION AND CONCLUSION
We investigate the response of mesospheric clouds (MCs) to idealized polar amplified surface warmings with higher than present-day CO 2 and CH 4 concentrations.This setup is reminiscent of warmer past climates such as the early Cretaceous and Paleogene periods.We find that an increase in CH 4 concentration is necessary to support an increase in MC fraction: increased CH 4 leads to an increase in mesospheric water vapor concentration and a reduction in mesospheric temperature, thereby promoting MC formation.For CH 4 concentrations of 64Â and more we find an increase in MC fraction not only at high latitudes, but also at middle and low latitudes.This is a surprising result given MCs have historically only been reported at mid-to-high latitudes.Our results support the new findings of Yu et al. (2023), who demonstrated the expansion of MCs to mid-latitudes over the 21st century across different socioeconomic pathways with WACCM version 6.They attributed these MC changes to an increase in mesospheric water vapor due to CH 4 oxidation and tropospheric water vapor transport.
Although the increased MC fractions in some experiments are striking, they are not accompanied by an increase in optical thickness of MCs, which remains negligibly small due to minuscule water content.Therefore, we do not find any significant surface radiative effect due to these new clouds.The degree of polar amplification, globally uniform SST warming, and total solar irradiance do not substantially alter the cloud response, which is mainly due to the local heating and moistening effects of greenhouse gases.It is worth noting that WACCM4 underestimates gravity wave fluxes, particularly in the SH (Garcia et al., 2017;Marsh et al., 2013).This could potentially influence mesospheric temperatures and winds, subsequently affecting MC fraction in the model.
Earlier studies suggest that an increase in MC may play an important role in changing the mesospheric chemistry through its impact on ozone (Murray & Plane, 2005;Siskind et al., 2018).Dehydration of the mesosphere due to MCs causes a reduction of atomic hydrogen, subsequently leading to an ozone increase in the mesosphere.Using observations and WACCM-PMC, Siskind et al. (2018) demonstrated that such MC-induced ozone changes can extend to lower latitudes beyond the MC formation region, possibly through the mean meridional circulation.An ozone increase, in turn, reduces the mesospheric temperatures, thereby influencing mesospheric circulation and MCs.However, our study does not examine the chemical effects of MCs.Future studies could focus on the changes in the chemical composition of the mesosphere in the presence of abundant mid-and low-latitude MCs.Finally, our findings support the idea that MCs can be considered as an indicator of climate change.Moreover, with substantially large CH 4 emissions in the future, MCs might become more prevalent, even in the tropics.

F
I G U R E 2 Zonal mean cold-point temperature averaged in (a) boreal (June-August) and (b) austral (December-February) summer, with different experiments represented by various colors.Scatter diagrams depict the (c) boreal and (d) austral summer mean water vapor mixing ratios in different latitude bands versus the tropical (10 N-10 S area weighted mean) cold-point mixing ratio.Circular, square and triangular markers indicate the area weighted mean water vapor mixing ratios between ±60 -90 , ±30 -60 and ±0 -30 latitudes, respectively.Filled and empty markers distinguish water vapor at 70 hPa and 0.1 hPa, respectively.

F
I G U R E 3 Boreal (June-August) and austral (December-February) summer-mean mesospheric temperature, water vapor concentration and cloud fraction in the Northern (solid lines) and Southern (dotted lines) Hemisphere high (top panel), middle (middle panel), and low (bottom panel) latitudes in different experiments in a-c, d-f and g-i, respectively.The regions between 60 to 90 , 30 to 60 , and 0 to 30 latitudes are considered as the high, middle, and low latitudes and area weighted averages in these regions are shown.

F
I G U R E 4 Boreal (June-August) and austral (December-February) summer-mean mesospheric cloud fraction, ice, and water content in Pol20_128M (orange), Pol20_128M_CARMA (green) and Pol20_128M_Solar (red) in a-c, d-f and g-i, respectively.