Projected Changes in Mountain Precipitation Under CO2‐Induced Warmer Climate

Mountains play a vital role in shaping regional and global climate, altering atmospheric circulation and precipitation patterns. To this end, identifying projected changes in mountain precipitation is significantly challenging due to topographic complexity. This study explains how mountain precipitation could respond to rising greenhouse gases. Using a series of century‐long fully coupled high‐resolution simulations conducted with the Community Earth System Model, we aim to disentangle future changes in mountain precipitation in response to atmospheric carbon dioxide (CO2) perturbations. Our research findings indicate that the warming observed in global mountains is more pronounced when compared to the mean warming rates experienced globally and in the ocean under elevated CO2. We identify five low‐latitude mountain ranges with elevation‐dependent precipitation response, including New Guinea, East Africa, Eastern Himalayas, Central America, and Central Andes. Those mountains are expected to have a mixture of increasing and decreasing precipitation in response to CO2‐induced warming, especially over the summit and steep topography. To elucidate the mechanisms controlling future changes in mountain precipitation, we propose “Orographic moisture omega feedback” in which an increase in low‐level relative humidity enhances local precipitation by strengthening the upward motion through moist processes for the wetting response and vice versa for the drying response. The effects of Mountain precipitation changes can be extended to hydrology and could lead to significant consequences for human societies and ecosystems.

the mountain climate system (Hock et al., 2019).However, the IPCC report mainly focuses on the mountain cryosphere, combined with precipitation, snow, permafrost, glaciers, and ice in lakes and rivers.Recent regional changes and an increase in extremes imply a significant change in sediment loads and water quality provided by mountains (Immerzeel et al., 2020;Luna, 2021).Therefore, understanding local climate change in mountainous regions (Hock et al., 2019;UN, 2015) is crucial for policymakers and stakeholders.
Precipitation over the mountains is driven by the inflow of moisture-laden winds that are lifted as they move over the terrain and condense to form precipitation, warming the atmosphere by latent heat release (Smith, 2018;Wallace & Hobbs, 2006).However, the precipitation response to climate change depends on many other factors (Colle et al., 2013;Smith, 2018), such as the large-scale shifts in atmospheric circulation that can modify moisture transport, affecting regional precipitation (Shi & Durran, 2014, 2015;Siler & Roe, 2014).A study using numerical simulation (Siler & Roe, 2014) found that the increase in precipitation associated with orographic storms on the lee side slope is due to vertical shifts in condensation.A recent study (Tamarin-Brodsky & Hadas, 2019) demonstrated that increased precipitation extremes are triggered by enhanced atmospheric moisture content and upward vertical velocity.Furthermore, remarkable changes in precipitation over terrain during the last few decades have been observed in many regions globally (Napoli et al., 2019;N. C. Pepin et al., 2022;Roe & Baker, 2006;Smith, 2018), typically referred to as the orographic process.Thus far, most of the studies on mountain meteorology underscore the role of extreme events in the mid-latitude environment (Grose et al., 2019;Shi & Durran, 2015, 2016;Siler et al., 2013;Smith, 1979).However, understanding the precipitation and related processes over the mountainous region is restricted, partly due to reliability in precipitation data across mountainous regions (Hock et al., 2019;Zandler et al., 2019) that comes from less spatial coverages of observation stations, the influence of satellite algorithms, and data assimilation schemes (Sun et al., 2018).Therefore, the climate model is a valuable tool for discovering and understanding the physical processes underlying mountain precipitation change.
To account for the complexities above in mountainous precipitation, the scientific community primarily relies on regional climate models as the topographical features still need to be better resolved in the coarse-resolution Global Climate Models (GCMs) (Gutowski et al., 2021).However, the regional models have limitations in incorporating large-scale features (Di Luca et al., 2015;Giorgi, 2019) due to their sensitivity to domain size and lateral boundary conditions.To overcome this, we adopt an ultra-high-resolution global Earth system model configuration to answer an essential question on how the precipitation over mountainous regions will respond to projected greenhouse gas forcing.The exact impact of CO 2 -induced warming on mountain precipitation is complex and can vary depending on the specific conditions of the region.Here, we present a global assessment of future CO 2 -induced warming impacts on regional mountains and demonstrate the underlying feedback mechanism.

Ultra-High-Resolution Simulation Strategy
The physical conditions in mesoscale processes, such as moisture advection, atmospheric circulation, and orographic lifting, can significantly affect mountain hydroclimate.Typically, these orographic features are not well resolved in coarse-resolution GCMs, and oftentimes their processes are parameterized (Small et al., 2014).Therefore, an adequate representation of these processes and their future change require a high-resolution climate model (advantages described in ref.Roberts et al., 2018).Hence, we employ a state-of-the-art, ultra-high-resolution, fully coupled climate model, the Community Earth System Model, version 1.2.2 (referred as CESM-UHR, see ref. Chu et al., 2020 for more details about CESM-UHR experiments and observed biases).The atmospheric component is the Community Atmosphere Model, version 5 (CAM5; Neale et al., 2012), with a horizontal resolution of about 25 km and 30 vertical layers, allowing realistic regional details such as topography and local processes responsible for orographic processes (Sandu et al., 2019;Small et al., 2014;Tao et al., 2020).The coupled ocean model in CESM-UHR is the Parallel Ocean Program version 2, with a horizontal resolution of about 10 km and 62 vertical levels.So, the sea surface temperature in CESM-UHR allows a more realistic representation of the ocean-atmosphere interactions and resolving of mesoscale oceanic features.We conducted a 140-year present-day (PD) simulation using an atmospheric CO 2 concentration of 367 ppm, initialized from a quasi-equilibrated climate state (Small et al., 2014).Two sensitivity experiments were carried out with CO 2 doubling (2xCO 2 , 734 ppm) and CO 2 quadrupling (4xCO 2 , 1468 ppm) (Chu et al., 2020).Other greenhouse gases have been kept at PD levels in each simulation.Each experiment was branched from the 71st year of the

Vertically Integrated Moisture Flux (kg m −1 day −1 )
The zonal and meridional vertically integrated moisture flux (Q) is the horizontal transport of atmospheric moisture by the penetrating winds, as:

𝑞𝑞𝑣𝑣𝑞𝑞𝑞𝑞
where Q u is zonal vertically integrated moisture flux Q v is meridional vertically integrated moisture flux u and v are horizontal components of wind components, q is specific humidity, g is the gravitational constant, and p is the atmospheric pressure.

Vertical Cross-Sections
Here, we studied the vertical structure to understand the associated processes rather than taking the vertical column average.Either latitude or longitude was chosen for cross-sections in detail based on the change in precipitation and associated vertically integrated moisture advection over the mountain.

Classification of Windward and Leeward Slopes
We identify windward and leeward slopes over cross-section based on the PD mean state.We consider the prevailing wind direction, specifically the dominant zonal or meridional wind component penetrating the mountain region.Subsequently, an additional main factor is the surface wind speed.

Local Saturated Condensation (g kg −1 day −1 )
Considering that environmental thermodynamics generally follows a moist adiabat, the local saturated condensation rate (Smith, 1979) caused by the adiabatic lifting can be approximated by: where ω a is the ascending vertical velocity, and q s is the saturated specific humidity.

Saturated Specific Humidity (g kg −1 )
Here, we define saturated specific humidity using an empirical method, where saturation vapor pressure (e*) is calculated using the Tetens equation (Murray, 1967).Static stability is stability of the atmosphere from hydrostatic equilibrium to vertical displacements.We investigated simple static stability using the following equation: where T is the absolute air temperature and θ is the potential temperature.

Total Diabatic Heating (K day −1 )
To understand the heating source in the atmosphere, the total diabatic heating rate (Nigam et al., 2000;Wang et al., 2019) is calculated using the potential temperature, which has conservation properties like those of dry static energy (DSE), approximated as: where ω is the vertical velocity, positive values of total diabatic heating rate represent heating, and negative values represent cooling.

Moist Static Energy (kJ kg −1 )
The moist static energy (MSE) (h) is used as a thermodynamic variable, which represents the addition of DSE (sum of dry air enthalpy and potential energy) and latent static energy (LSE, latent atmospheric heat), as: where C p is specific heat at constant pressure, z is the height above the surface, and L v is the latent heat of vapourization.

Moisture Budget Analysis (mm day −1 )
The vertically integrated moisture budget can be expressed as a linearized equation in terms of precipitation P, where E is evaporation, V is vertical moisture advection and H is horizontal moisture advection.We can neglect the moisture tendency term on the annual time scale, as it's small compared to the rest of terms in the moisture budget (Oueslati et al., 2019): Consequently, the change in mean precipitation can be expressed as follow: Here angle brackets indicate a vertical mass integral, and delta indicates the change in mean state response to CO 2 perturbation of the respective quantity.

Vertical Decomposition of Moist Static Energy Advection (kJ kg −1 day −1 )
Atmospheric deep convection is mainly constrained by the MSE budget (Neelin & Held, 1987) and can be explained using the vertical structure of MSE advection.We decomposed vertical MSE advection into its dynamic and thermodynamic components.
Here overbar indicates a mean state value and delta denotes the change in mean state response to CO 2 perturbation of the respective quantity.

Quantifying the Precipitation Extremes
We used the following indices (Karl et al., 1999) to identify projected changes in extreme rainfall events.
The simple daily intensity index (SDII, mm day −1 ) describes the daily precipitation amount averaged over all wet days in a year.The wet days are when precipitation exceeds more or equals 1 mm day −1 .SDII is an absolute index used to assess the intensity of extreme precipitation.
The extreme flooding index (Rx5day, mm) describes the maximum precipitation amount in five consecutive days for each year.Rx5day is generally used to express the changes in likely flood risks, as heavy rain conditions can contribute to flooding conditions over consecutive days.

Projected Elevation-Dependent Precipitation Changes
Generally, the location of the mountain precipitation is determined by local factors (Boos & Pascale, 2021;Smith, 1979Smith, , 2018) such as mountain geometry, terrain steepness, surface wind, moisture source, etc. Considering the solidity of mountain geometry, which will not change in the future, changes in surface temperature due to global warming may lead to changes in other factors, such as surface wind and moisture sources.A global picture of projected temperature changes reveals that mountain systems are susceptible to greenhouse warming (Figure S1 in Supporting Information S1).Table 1 shows that the amplified warming rates observed in mountain regions are nearly comparable to those in the land and exhibit stronger warming than the global mean and ocean in response to CO 2 (Kad et al., 2022).The mountain warming rate is 3.44°C for 2xCO 2 and 7.35°C for 4xCO 2 (Table .1).These values represent the anticipated temperature increase in mountain regions under the respective CO 2 perturbation scenarios.The higher warming rate for 4xCO 2 than 2xCO 2 suggests a more temperature response in mountainous areas under more extreme CO 2 concentration levels.We examine the response of annual mean precipitation to surface local warming (Figure 1) response to 4xCO 2 .The response of annual mean precipitation seems more dominant over low-latitudinal mountains (30°S-30°N) as compared to the mountains in high-latitude in the 4xCO 2 experiment (Figure 1).Strong response in these low-latitude regions can also be linked to the more substantial enhancement of water vapor in the low-latitude than high-latitude and changes in large-scale circulation patterns such as ITCZ (Mamalakis et al., 2021), MJO (Maloney et al., 2019;Roxy et al., 2019), and ENSO (Latif & Keenlyside, 2009;Mamalakis et al., 2021) under greenhouse warming.However, this study emphasizes the regional scale process only because precipitation response is very high over limited areas, implying that changes in regional climate are the dominant factor in the context of mountain precipitation changes.This analysis identifies five mountain regions experiencing significant precipitation changes in response to CO 2 quadrupling.Based on the area that exceeds precipitation response to the local warming by a threshold ±0.1 mm/ day/°C, we selected the five most prominent regions among the global mountain range: New Guinea, East Africa, Eastern Himalaya, Central America, and Central Andes (Figures 1a-1e).Since the high-resolution simulation can decently capture topographic features, the mean precipitation pattern in the CESM-UHR simulation over mountain ranges seems reliable compared to the CESM-CMIP6 (100 km nominal resolution) (see Figure 2).The reliability and accuracy of Coupled Model Intercomparison Project Phase 6 (CMIP6) models are restricted due to their utilization of a coarser resolution.However, precipitation biases in CESM-UHR simulation over some regions analogized to satellite observations are very likely due to the model configuration.
Further, we aimed to thoroughly quantify precipitation changes by investigating precipitation sensitivity in mountain regions.To achieve this, we conducted an analysis of cross-sectional sensitivity across the selected mountain regions.By comparing the relative changes in precipitation (%) to the absolute temperature changes (°C) in response to CO 2 perturbation, we determined the sensitivity rates for each region based on the windward and leeward sides.In this context, East Africa and the Central Andes have been categorized as regions exhibiting windward and leeward features.For the remaining mountains, we have designated them as windward regions.
Our analysis revealed that Central America exhibits a negative sensitivity under CO 2 perturbation, indicating a potential dry response in this region.Furthermore, we found that precipitation sensitivity is more pronounced on the windward side of East Africa.Specifically, the windward side of East Africa and the Central Andes in our study exhibited higher precipitation sensitivity, suggesting a more robust response regarding precipitation changes.In contrast, the leeward side of these mountain regions demonstrated relatively lower sensitivity to changes in temperature.The differential sensitivity between windward and leeward sides has significant implications for regional climate dynamics and water resources in these two mountain regions.These results, presented in Table 2, underscore the importance of considering the precipitation response in mountainous areas.
Global precipitation changes in response to CO 2 perturbation (Figure S2 in Supporting Information S1) exhibit a higher magnitude of changes in tropical regions.The tropics are particularly sensitive to changes in CO 2 concentrations, leading to alterations in precipitation patterns (Bony et al., 2013;Mamalakis et al., 2021;Sohn et al., 2019).These changes can result in increased rainfall intensity, changed monsoon patterns, and shifts in wet and dry seasons (Intergovernmental Panel on Climate Change, 2023).The notable changes in precipitation are most likely at high elevations (Figures S3a-S3e in Supporting Information S1) under CO 2 -induced warming but still are highly uncertain (Hock et al., 2019).Based on the precipitation change over five mountain regions, we define wetting mountains (New Guinea, East Africa, Eastern Himalayas, windward side of Central Andes) where the precipitation increases and drying mountains (Central America, the leeward side of Central Andes) where the precipitation decreases in a CO 2 -enriched climate.Interestingly, the spatial pattern of anomalous precipi tation over an individual mountain is heterogeneous, predominantly evident at mount summits or steep terrain (Figure 1), and looks like elevation-dependent precipitation change.The most increase in precipitation intensity is exhibited at Puncak Jaya in New Guinea (Figure 1a and Figure S3a in Supporting Information S1).In contrast, there is an overall drying over Central America (Figure 1d).It should be noted that the highest mountain summit (e.g., Himalayan peaks) experiences inadequate precipitation due to a lack of moisture supply.In such cases (like Eastern Himalaya, Figure 1c, and Figure S3c in Supporting Information S1), it precipitates over steep topography before reaching the high summit.Central Andes exhibits both responses, wetting over the windward side of Central Andes and drying over the leeward side of Central Andes.

Atmospheric Conditions
The moisture advection over the mountains strongly depends on its terrain pattern, which plays an important role in shaping the vertical profile of moisture content and regional atmospheric conditions.The overall atmospheric relative humidity is assumed to have increased (Tamarin-Brodsky & Hadas, 2019) as global temperatures rise due to the increased greenhouse gases concentrations, where the Clausius-Clapeyron relation governs the increase in saturation-specific humidity (O'Gorman & Muller, 2010).It is observed that the change in precipitation (Figure 1) is linked to the vertical structure of the relative humidity (Figures 2k-2o).The major contributor to the precipitation changes is due to changes in atmospheric humidity and surface wind, which influence upslope lifting over the mountain terrain (Smith, 1979(Smith, , 2018;;Tao et al., 2020).The enhancement and reduction of vertical motion are more evident on the mountain summit and slope (Figure 5).These enhancements are consistently replicated in the 2xCO 2 and 4xCO 2 experiments.Also, updrafts are widespread in the upper level over projected wet (Figures 4k-4o) and downdrafts over projected drying (Figures 4n and 4o) mountains.These results show that moisture influences the vertical motion over mountainous terrain in the projected CO 2 -induced warming scenario.The projected precipitation changes in mountain regions are described in terms of the changes in local saturated condensation (Figure 5), which is a function of saturated humidity and upward vertical velocity.
Figure 6 displays the vertical distribution of atmospheric static stability in shading.The stability increase at upper levels reflects a higher resistance of the atmosphere to highlight the atmospheric stabilization condition.Atmos- pheric stabilization refers to the resistance of the atmosphere to vertical motion, which can inhibit the development of deep convection, as it makes it harder for air to rise and form deep convection.The diabatic heating influences atmospheric stability (black contour in Figure 6) at the upper level.
The specific humidity (red contour represented in Figure 6) indicates water vapor levels at various altitudes.By comparing the profiles in each panel, the study aims to comprehend the changes in these atmospheric variables under the PD mean state, 2xCO 2 , and 4xCO 2 perturbation.These vertical profiles yield valuable insights into atmospheric processes and changes in the examined regions under distinct conditions, thus enhancing our understanding of the implications of increased greenhouse gas concentrations on the climate system.
Changes in vertical velocity (Figure 5) show good agreement with strengthening MSE (Figure 7) through latent heat release.A unique core of least moist static energetics can be observed in the vicinity of mountain summits        ing scenario, followed by lower-level LSE.Additionally, precipitation causes additional local moistening and subsequently enhances the LSE with a vertical extension of MSE (Figure 7).In drying mountain regions, raised lower-level DSE (Figures 7n and 7o) can be seen where restricted diabatic surface heating and upper-level cooling (Figures 6n and 6o).In addition, we marked an anomalous dry static environment under CO 2 perturbation in these unfavorable regions for upward motion.

Role of Moist Dynamics
Moisture budget analysis (see Section 3 and Figure 8) shows vertical moisture advection has a close relationship with precipitation in the PD and CO 2 experiments (this close relationship can only be found over the wetter area), consistent with previous studies (Oueslati et al., 2019;Yang et al., 2014).Sufficient moisture in the atmospheric column and strong vertical motions resulted in wetting over mountain regions.This framework has been generally utilized to compare local changes in precipitation (Bony et al., 2013;Chou et al., 2012;Huang et al., 2013;Oueslati et al., 2019;Wang et al., 2019).Wetting mountains consistently increase precipitation and vertical moisture advection, whereas drying mountains do not.However, horizontal moisture advection is small compared to vertical moisture advection (Figure 9).An anomalous increase in moisture advection can be found in the atmosphere over mountains, consistent with precipitation increase, whereas an anomalous decrease in moisture advection with precipitation decrease (Figures 8 and 9).The moisture advection response (see Section 3 and Figure S4 in Supporting Information S1) is analyzed using vertical moisture advection (in terms of dynamic and thermodynamic components) and horizontal moisture advection.Our results suggest that vertical advection plays an important role (Figure 7), which may be benefited from latent heating, strengthening the upward motion through the thermodynamic factors (Figure S4 in Supporting Information S1) at lower-level closer to the steep mountain terrain.However, it is important to acknowledge that budget analyses alone cannot establish definitive causality relationships between the components.As such, we cannot conclusively attribute the observed effects solely to either thermodynamic or dynamic factors.For instance, Eastern Himalayas have a thermodynamic component that appears to be triggered at a lower level from foothills to steep terrain, which may further influence the upper-level dynamic contribution (see Figures S4c and S4h in Supporting Information S1).Even though horizontal advection has slight changes over most of the mountains, it plays a crucial role in the windward side of the Central Andes (peripheral to the Pacific Ocean; Figure 9o and Figure S4o in Supporting Information S1).
Using the moist advection approach to understand the seasonal cycle (Figure 9) and its complexity to explain the observed pattern seems complicated.We endeavored to improve our understanding of the intricate patterns observed in both Central America and the Central Andes using deep convection.Our objective was to gain insights into the multiple factors contributing to the complex patterns observed, including moisture distribution and atmospheric circulation.To achieve this, we studied the role of deep convection and its impact on these patterns.

Atmospheric Deep Convection
Atmospheric deep convection occurs in the tropics (Houze et al., 2015;Johnston et al., 2018) and is mainly associated with vertical motion, causing diabatic heating and MSE export (Bui et al., 2016;Neelin & Held, 1987;Yan et al., 2020;Zhang & Fueglistaler, 2020).A recent study by Imamovic et al. (2019) also highlighted that the orographic height and width large width aspect control deep convection over Mountains.We observed evidence of deep convection in the vertical profiles of atmospheric variables over the mountain regions, as illustrated in previous sections.Deep convection refers to the vertical transport of heat and moisture, leading to the formation of towering cumulonimbus clouds that can extend deep into the upper atmosphere (Figure S5 in Supporting Information S1).Deep convection is indeed a complex phenomenon.For the model analysis, cloud fraction at 100-200 hPa has been utilized as an indirect measurement of deep convection (Peña-Ortiz et al., 2019).These findings align with previous observational studies, providing evidence of deep convection in New Guinea (Johnston et al., 2018;Permana et al., 2016), East Africa (Finney et al., 2019;Hart et al., 2019), Eastern Himalayas (Hunt et al., 2022), Central America (Houze et al., 2015), and Central Andes (Rasmussen & Houze, 2011).These observational studies support the significance of deep convection in influencing precipitation patterns over mountainous regions.Moist processes driving precipitation in mountain regions can vary depending on the specific geography, climate conditions, and seasonality.Also, as vertical moisture advection plays a dominant role in mountain hydroclimate response to CO 2 quadrupling, we split vertical MSE advection into dynamic and thermodynamic components (see Section 3).Positive (negative) vertical MSE advection involves the transport of higher (lower) energy in the vertical direction, resulting in an increase (decrease) of available energy in the atmosphere.This can help us explain the respective contribution of energy import or export and its possible linkage to vertical motion.A symmetric pattern (Figure 10) is observed over mountains with a low-level energy import.Interestingly, this symmetric pattern of moist energetic response shows consistent results with precipitation changes.In wet regions such as East Africa, the Eastern Himalayas, Central America, and the windward side of Central Andes, there is a positive thermodynamic component in the lower atmosphere (Figures 10i, 10m, and 10o) and a negative dynamic component at the upper level ( Figures 10g, 10h, and 10e).In New Guinea, both 10.1029/2023EF003886 14 of 19 thermodynamic and dynamic responses are similar in the lower atmosphere (Figures 8f and 8k), which leads to enhanced deep convection and intense precipitation.However, drying mountains have energy imports at lower and upper levels (Figures 10d and 10e).At the upper level in Central America and the leeward side of Central Andes, positive dynamic responses (as shown in Figures 10i and 10j) impede convection and reduce precipitation.We demonstrated that warming in mountainous areas amplifies the thermodynamic effect (Moustakis et al., 2020) in the lower atmosphere.Based on our analysis, we can infer that the cause of deep convection in wet regions is attributed to the increase in thermodynamic components at a low level and the decrease in dynamic components at the upper level.However, shallow-hinder convection is observed in dry regions due to the increase in the thermodynamic component at a low level and the dynamic component at an upper level.

Orographic Moisture Omega Feedback
An increase in static stability is unfavorable for vertical motion as more upper-level warming will create a more stable troposphere.Several studies (Li & O'Gorman, 2020;Maloney et al., 2019;Sharmila & Walsh, 2018;Shi & Durran, 2015) have shown that global warming leads to increased atmospheric stability, which in turn can cause the atmosphere to become more stratified.Hypothetically, the mountains are supposed to be wetter under a warming scenario, as they can accumulate additional moisture content owing to temperature rise.Similar to conclusions from previous studies (see refs. Li & O'Gorman, 2020;Shi & Durran, 2016, 2015, 2014;Zhao et al., 2020) on the midlatitude mountains, raising moisture in vertical ascending motion results in anomalous diabatic heating through enhanced condensation and triggers precipitation extremes.This anomalous heating is thermodynamically compensated with ascending vertical motion, which sucks moisture from the lower atmosphere (Lau et al., 2020;Tamarin-Brodsky & Hadas, 2019).What determines the precipitation changes over mountains under the future warming climate?We attempted to answer this question by using feedback mechanisms.Herein, we introduce the concept of "Orographic moisture omega" feedback, explaining the loop mechanism in which vertical motion is reshaped by moisture over a mountain in a warming climate that can further amplify or dampen through feedback and vice versa (schematically illustrated in Figure 11).In response to CO 2 forcing, the wetting mountain regions are found to be associated with wetting response.Wind speed should reduce due to the weaker zonal temperature gradient in warm tropical climates but more moisture gradient (Maloney et al., 2019;Sohn et al., 2019;Vecchi & Soden, 2007).This abundant moisture content elongates the ascending motion by diabatic heating.Precipitation increases are attributed to the wetting response, further amplified by enhanced ascending motion under CO 2 perturbation (Figure 11a).A substantial increase in diabatic heating leads to deeper vertical heating at upper levels of the atmosphere, supporting deep convection with vertically rising MSE from a low level is also favorable for ascending motion.But it coexists with an overall increase in atmospheric stability in the background, which shows this increase in stability would increase the lapse rate (Figures S3u-S3y in Supporting Information S1).Deep moist convection in the mountains can counteract the regional lapse rate within this framework.We confirm that the resultant Orographic moisture omega feedback appears positive after compensating with processes.On the other hand, the drying mountains are manifested by the drying response of Orographic moisture omega feedback (Figure 11b).The drying regions experience a decrease in saturated moisture content, which leads to a reduction in upward air movement.This is caused by a diabatic cooling anomaly at a lower level, resulting in shallow convection.As a result, the initial humidity is further decreased.Atmospheric stabilization is a counterpart in both responses to maintain equilibrium within the feedback loop.Low-level relative humidity response shows why the initial trigger in some mountainous areas differs from others, as the response increased nearby wetting mountains and decreased nearby drying mountains (Figure S6 in Supporting Information S1).These results indicate that the Orographic moisture omega feedback that modifies regional precipitation is evident.Differences in the strength of the feedback on a regional scale can result in varying precipitation responses.

Extreme Event Reverberation
Extreme events are strongly associated with a change in the mean climate state.Regional atmospheric conditions and local topography strongly affect these extreme events (Zhang & Liang, 2020).Geographical features concerning topography significantly modulate the spatial changes in extreme events (Herold et al., 2016;Shi & Durran, 2015).Besides, change in the background temperature due to CO 2 -induced greenhouse warming also contributes to the variability of extreme precipitation.To further understand the change in extreme events in response to CO 2 quadrupling, we examine extremes (Karl et al., 1999) using absolute precipitation criteria (see Section 3 and Figure 12), such as the precipitation intensity (SDII) and extreme flooding events (Rx5day).These events (Figures 12a-12e and Figures S5a-S5e in Supporting Information S1) coincide with their mean precipitation changes (Figures 1a-1e and Figures S7a-S7e in Supporting Information S1) under CO 2 -induced greenhouse gas forcing, increasing the events over New Guinea, East Africa, Eastern Himalaya, mountains part of Central Andes, and decreasing over Central America and leeward side of Central Andes.However, certain extreme events are not related to changes in precipitation mean states, like the highest precipitation in 1-day and 5-day over the Eastern Himalayan region (Figures S7h and S7m in Supporting Information S1) and heavy rainy days over the Central Andes (Figure S7t in Supporting Information S1).These amplified events exceed precipitation anomalies, which further exclusively need to be investigated.Time-dependent indices like heavy rainy days R10, R20 (see Table S1 in Supporting Information S1, Figures S7p-S7t and S7u-S7y in Supporting Information S1) concurrently pick through vertical moisture advection (consistent with previous attributional studies (O'Gorman et al., 2021;Oueslati et al., 2019;Zhao et al., 2020)).The consequences of such excessive extremes are responsible for excessive surface runoff (Figures S3k-S3o in Supporting Information S1), further exacerbating hazards such as river floods, mountain landslides, and debris flow in the mountains and their surrounding regions (Moustakis et al., 2020).It is possible that the CESM-UHR model may have limitations when simulating extreme precipitation in mountainous regions, as these events are often influenced by complex and localized processes that can be challenging to represent accurately in models.

Conclusions
The CESM-UHR provides a valuable opportunity for studying regional climate change in a specific region.
Our study reveals that mountain regions experience amplified warming in response to CO 2 perturbation.These amplified warming rates are stronger warming than the global mean and ocean.In the present study, we employed CESM-UHR to address future changes in precipitation patterns in mountain regions due to increasing CO 2 concentrations in the earth's atmosphere.The five most sensitive mountains in the low latitude, including New Guinea, East Africa, Eastern Himalayas, Central America, and Central Andes, are identified based on precipitation response.A comprehensive feedback framework describes the change in precipitation response to adjustments in the mean state of vertical structure and its related processes.Our study proposes "Orographic moisture omega feedback" for precipitation attributions which appear as positive feedback in the climate system, amplifying an initial state.This feedback consists of two primary net responses, wet and drying responses.In the case of wetting response, the warming-induced moisture addition in the mountain terrain favors a strengthening of ascending motion and anomalous diabatic heating through enhanced precipitation, which further enhances the local build-up of humidity.But, in the case of drying response, the moisture shortage restricts the ascending motion, reducing local precipitation and further reducing the moisture.The Orographic moisture omega feedback response to CO 2 perturbations could explain the projected wetting in New Guinea, East Africa, the Eastern Himalayas, the windward side of Central Andes, and projected drying in Central America and the leeward side of Central Andes.Also, higher precipitation sensitivity is observed on the windward side of both East Africa and the Central Andes mountains.
Even though our study concentrates on the mean state changes, the proposed feedback mechanism can potentially improve our comprehension of future changes in mountain precipitation variability from diurnal to interannual timescales.The precipitation changes over the mountains can cause other significant threats, such as mountain-ice melting (Chen et al., 2013), loss and degradation of soil (Borrelli et al., 2020), and biodiversity reduction (Peters et al., 2019;Viviroli et al., 2007), which pose severe consequences to humans and the entire our ecosystem (Conway et al., 2019;Elsen et al., 2020).Thus, climate change-induced regional hydroclimatic changes pose formidable challenges to decision-makers in ensuring mountain water management and resilience.
The scientific community can apply this framework to investigate the potential threats to water resource management and related biodiversity.Policymakers need to adopt strategic planning risk mitigation related to the mountains and regions highly dependent on mountain resources.

Figure 1 .
Figure 1.Projected elevation-dependent precipitation changes over the global mountain in response to 4xCO 2 .The center panel indicates precipitation response (absolute change in precipitation rate) to local warming associated with future change.The sub-panels indicate precipitation changes in 4xCO 2 compared to presentday over (a) New Guinea, (b) East Africa, (c) Eastern Himalaya, (d) Central America, and (f) Central Andes.The blue vectors represent change in vertically integrated horizontal moisture flux in response 4xCO 2 and the white contour shows an elevation orography of 0.5 km interval.The magenta line indicates either the longitude or latitude of the cross-section for further analysis.

Figure 2 .
Figure 2. Annual mean precipitation in observation and model simulation.Annual mean precipitation climatology over chosen mountains from (a-e) satellite product, (f-j) CESM2 simulation having about 100 km nominal resolution, and (k-o) present-day (PD) simulation having about 25 km resolution from UHR-CESM.The satellite data set was obtained from the Tropical Rainfall Measurement Mission (TRMM) (Huffman et al., 2010) 3B43 (this data merges the TRMM 3B42 product adjusted with the GPCC (Global Precipitation Climatology Center) rain gauge) during 2000-2019 and the CESM2 model data from Coupled Model Intercomparison Project Phase 6 (CMIP6) (Danabasoglu et al., 2020) during 1996-2015.The white contour shows an elevation orography of 1 km interval from the UHR-CESM for consistency and comparison.

Figure 3 .
Figure 3. Mountain precipitation sensitivity in response to CO 2 perturbation.Precipitation sensitivity over mountain cross-sections (a) New Guinea, (b) East Africa, (c) Eastern Himalaya, (d) Central America, and (e) Central Andes.The blue and red lines represent the precipitation sensitivity to 2xCO 2 and 4xCO 2 , respectively.The green line depicts the wind speed from the present-day mean state, while the vector indicates the dominant surface wind component (zonal or meridional, depending on the cross-section).The gray shading is the elevation of each mountain region.Additionally, the black color box indicates the windward side, while the cyan color box represents the leeward side of the mountain.

Figure 4 .
Figure 4. Vertical structure of relative humidity.Vertical profile of relative humidity (shading) and its anomalies (contour) over New Guinea, East Africa, Eastern Himalaya, Central America, and Central Andes.(a)-(e) For the present-day mean state, (f)-(j) for 2xCO 2 mean state, and (k)-(o) for 4xCO 2 mean state respectively.Contours represent relative humidity anomalies due to CO 2 perturbation, where positive values indicate boosts and negative values indicate reduction.

Figure 5 .
Figure 5. Vertical structure of vertical velocities with local saturated condensation rate.Vertical profile of mean vertical velocity (shading) and local saturated condensation rate (contour) over New Guinea, East Africa, Eastern Himalaya, Central America, and Central Andes.(a)-(e) present-day, (f)-(j) 2xCO 2 experiment, and (k)-(o) 4xCO 2 perturbation experiment respectively.Here, we consider upward motion using the daily scale as precipitation events are associated with upward motions, which offers a clearer idea about ascending motion.

Figure 7 .
Figure 7. Vertical structure of moist static energy (MSE).MSE (shading) over New Guinea, East Africa, Eastern Himalaya, Central America, and Central Andes for (a)-(e) present-day mean state, (f)-(j) 2xCO 2 anomalies, and (k)-(o) 4xCO 2 anomalies.The green color contour represents changes in latent static energy, and the blue color contour represents changes in dry static energy.

Figure 8 .
Figure8.Moisture budget analysis over selected mountain regions.Cross-section of moisture budget terms (a-e) for present-day mean state, (f-j) for 2xCO 2 anomalies, (k-o) 4xCO 2 anomalies respectively.The colored lines indicate the mean state of moisture budget terms; red for precipitation (P), green for evaporation (E), blue for vertical moisture advection (V), and cyan for horizontal moisture advection (H).The shaded gray color represents associated elevation orography.

Figure 9 .
Figure 9. Seasonal cycle evaluation of dominant terms from moisture budget analysis.Cross-section of moisture budget terms for change in precipitation (P), change in vertical moisture advection (V), and horizontal moisture advection (H), respectively.The shaded gray color in the bottom panels represents associated elevation orography.

Figure 10 .
Figure 10.Vertical structure of anomalous vertical moist static energy (MSE) advection in response to 4xCO 2 .Shading and contour denote projected changes in vertical MSE advection over New Guinea, East Africa, Eastern Himalayas, Central America, and Central Andes.(a)-(e) Change in net vertical MSE advection, (f)-(j) change in the dynamical component of vertical MSE advection, and (k)-(o) change in a thermodynamical component of vertical MSE advection, respectively.Positive contours represent energy import and negative energy export.

Figure 11 .
Figure 11.Schematic representation of "Orographic moisture omega feedback" mechanism over low-latitude mountains.(a) Wetting response: Atmospheric relative humidity in the mountain terrain will increase due to greenhouse warming and favors a drop strengthen ascending motion in a humid climate through total diabatic heating.Simultaneously, vertical moist static energy (MSE) export will feed ascending motion by deep convection.It boosts local precipitation and further contributes to the initial humidity.(b) Drying response: Atmospheric relative humidity in the mountain will reduce, and this limited humidity weakens ascending motion through total diabatic cooling.Dry static energy contributes to low-level warming and dry conditions.Vertical MSE import will feed descending motion.Local precipitation decreases and initial humidity is further reduced as a result.

Figure 12 .
Figure 12.Projected change in precipitation extremes over global mountain system in response to 4xCO 2 .The center panel indicates the extreme intensity index (SDII), and sub-panels (a)-(e) show the extreme flooding index (Rx5day).The white contour shows an elevation orography of 0.5 km interval.

Table 1
The global mean warming rates across global, ocean, land, and mountain regions are assessed using 2 m air temperature in this table.These rates are determined by calculating the absolute changes in the CO 2 experiment relative to the PD simulation.This analysis defines global mountain regions as areas with elevations higher than 1 km, excluding Antarctica.Amplified Warming Rates in Response to CO 2 Perturbation

Table 2
Mountain Precipitation Sensitivity Across Mountain Regions