Atmospheric rivers (ARs) are crucial for the Earth's hydrological cycle, transporting vast amounts of moisture. Western North America, particularly coastal regions, is disproportionately impacted by ARs, receiving substantial precipitation from them. ARs' contribution ranges from beneficial drought relief to devastating floods. With climate warming, ARs are projected to intensify, potentially exacerbating extreme events. This study investigates the impact of stratospheric aerosol injection (SAI), a solar geoengineering technique, on western North American ARs. SAI aims to mitigate some of the worst effects of global warming by reflecting sunlight back into space, but its potential impacts on regional weather patterns remain largely uncertain. The scientific debate surrounding solar geoengineering has intensified, emphasizing the need to understand the potential consequences of such interventions. The National Academies of Sciences, Engineering, and Medicine has highlighted significant knowledge gaps and recommended further research to assess the benefits and risks of solar geoengineering. Earth System Models (ESMs) are valuable tools in this context. Previous GeoMIP (Geoengineering Model Intercomparison Project) experiments have often employed simplified approaches or equatorial injections, leading to unrealistic cooling patterns. However, recent work by Kravitz et al. (2017) has shown that using multiple injection locations with varying sulfur dioxide (SO2) amounts can mitigate this issue. This study utilizes simulations from the Geoengineering Large Ensemble (GLENS) project, which employs this approach, to investigate SAI's effects on western North American ARs, thus filling the knowledge gap regarding the impact of SAI on a crucial aspect of regional hydroclimate.

The literature extensively documents the importance of ARs in western North America's hydroclimate, highlighting their contribution to both beneficial and destructive precipitation events. Studies have projected increases in AR intensity and moisture content under climate change, potentially worsening floods and extreme events. The scientific literature on solar geoengineering, particularly SAI, is also substantial, encompassing numerous studies examining its potential benefits and unintended consequences. Previous GeoMIP and GLENS studies have explored broad-scale hydrological responses to geoengineering, but detailed investigations of western North American ARs have been limited. This research aims to address this gap by focusing specifically on the impacts of SAI on the dynamics and precipitation associated with ARs in this region.

This study utilizes simulations from the Geoengineering Large Ensemble (GLENS) project, which employs the Community Earth System Model version 1 with the Whole Atmosphere Community Climate Model (CESM1(WACCM)). CESM1(WACCM) is a high-resolution fully coupled Earth System Model with a high-top atmospheric component that extends to 140 km. It includes detailed representation of the stratospheric processes relevant to SAI. The GLENS simulations include a 20-member ensemble with time-varying SO2 injections at multiple latitudes to maintain 2020 temperature levels under a high forcing scenario (RCP8.5). A 20-member ensemble of RCP8.5 control simulations (2010-2030) and 3 members extended to 2097 were also used. The study analyzes changes in ARs based on the Shields/Kiehl atmospheric river detection method. This method focuses on identifying stronger moisture streams relative to their background environment, minimizing the influence of thermodynamic effects on AR counts. The method applies moisture and wind thresholds to identify ARs that make landfall in western North America, and it also restricts the analysis to "Pineapple Express" ARs, originating from subtropical moisture sources. It also places geometric requirements on length and width. The analysis focuses on cool-season precipitation, examining changes in AR frequency, precipitation intensity, and total precipitation. The impact of SAI is assessed by comparing end-of-century SAI simulations to both RCP8.5 end-of-century simulations and near-term (2010-2030) control simulations.

The study's key findings are as follows: 1. **SAI's Impact on AR Frequency:** Under SAI, AR frequency is projected to increase in southern California and decrease in the Pacific Northwest and coastal British Columbia by the end of the 21st century. These changes are primarily driven by shifts in low-level winds, which are consistent with those observed in other future and paleoclimate scenarios. The changes in AR frequency are more pronounced when comparing SAI to RCP8.5 end-of-century simulations, with less significant changes observed when compared to the near-term (2010-2030) control. 2. **SAI's Impact on AR Precipitation Intensity:** SAI results in a shift towards more moderate rainfall rates from higher intensity bands. This effect is most pronounced in southern California, where ARs are a dominant precipitation source. When comparing SAI to RCP8.5, significant decreases in extreme precipitation events are observed. However, the differences between SAI and the near-term climate are less significant. 3. **SAI's Impact on Total Precipitation:** Total cool-season precipitation shows a complex pattern under SAI. Southern California experiences increased precipitation, while the Pacific Northwest is drier. These regional differences highlight the complexities of SAI's impacts on hydroclimate beyond just ARs. 4. **Seasonal Variation:** Changes in AR frequency exhibit significant seasonal variability, with the strongest effects during the cool season. The water year accumulation diagnostics show a clear narrowing of precipitation amount by ARs for lower latitudes when comparing SAI to RCP8.5 end-of-century scenarios. Southern California shows increased AR precipitation throughout the year when compared to the near-term conditions. 5. **Dynamical Forcing:** The shift in low-level winds, consistent with other studies and driven by a change in jet stream location, is the primary mechanism driving the changes in AR frequency and associated precipitation.

The findings indicate that SAI can alter AR characteristics in western North America, with regional variations in both AR frequency and precipitation intensity. The shift towards more moderate rainfall events suggests a potential mitigation of extreme flood events in southern California, where ARs contribute significantly to precipitation. However, this benefit might be offset by reduced precipitation in other regions, such as the Pacific Northwest. The observed changes in AR frequency are consistent with changes in low-level winds, highlighting the role of dynamical forcing in shaping the response of ARs to climate interventions. While SAI may mitigate some aspects of AR-related precipitation, the study highlights that SAI's influence extends beyond ARs, impacting total precipitation patterns across the region. Therefore, a comprehensive assessment of SAI's regional hydroclimatic implications requires considering effects beyond ARs alone.

This study demonstrates that SAI, as simulated in GLENS, significantly alters the frequency and intensity of ARs impacting western North America, particularly in southern California and the Pacific Northwest. While SAI shows promise in moderating extreme rainfall events in southern California, a nuanced picture emerges when considering total precipitation patterns. This work underscores the regional complexities of geoengineering impacts, highlighting the need for further research to refine understanding and guide informed decision-making. Future research should focus on higher resolution simulations and expanded AR definitions beyond the "Pineapple Express" type to provide a comprehensive evaluation of the effectiveness and potential risks of SAI.

The study's limitations include the use of a single AR detection method, the limited ensemble size of the RCP8.5 end-of-century simulations, and the relatively coarse model resolution. The reliance on the "Pineapple Express" type of ARs also limits the generalizability of the findings, as other types of ARs may respond differently to SAI. Higher resolution datasets and an expanded ensemble size would improve the confidence in the study's conclusions. The use of multiple AR detection methods would allow for a better assessment of the uncertainties associated with AR identification. A regionally refined CESM framework with higher horizontal resolution is needed for more accurate characterization of landfalling ARs and orographic precipitation.