Hyemi Kim

Under multiple anthropogenic global warming scenarios considered by the Coupled Model Intercomparison Project Phase 6 (CMIP6), Arctic sea ice is projected to disappear seasonally as early as 2035. Stratospheric Aerosol Injection (SAI) is a climate intervention strategy that has been proposed to mitigate some of the impacts of global warming. In this study, we evaluate the effectiveness of SAI in preserving Arctic sea ice, focusing on its sensitivity to the injection latitude of the aerosols. Using the 2nd version of the Community Earth System Model (CESM2) coupled with the Whole Atmosphere Community Climate Model (WACCM6), we analyze experiments with aerosol injection latitudes ranging from 45°S to 45°N. The results reveal that as the injection latitude shifts closer to the North Pole, Arctic sea ice rapidly recovers in both its extent and volume. This recovery is driven by coordinated shifts in clear-sky and cloud-related radiation, along with changes in surface reflectivity, that collectively reshape the surface energy balance in favor of ice growth. Importantly, we also find that, under fixed SAI injection rates, Arctic sea ice recovery varies substantially with injection latitude and does not scale directly with global mean surface temperature.

By analyzing five Earth Science I textbooks developed based on the 2015 Revised National Curriculum, this study examines content related to the “El Niño-Southern Oscillation (ENSO),” a core concept in Earth Science. The analysis focuses on four key areas: the description of ENSO mechanisms, potential misconceptions caused by ENSO schematic diagrams, the depiction of ENSO’s global impacts, and inquiry activities related to ENSO. To address factors contributing to misconceptions in existing textbooks, an improved ENSO diagram is proposed, offering a comprehensive view of ENSO as a dynamically coupled ocean-atmosphere interactive system.

The Madden‐Julian oscillation (MJO) has remarkable impacts on global weather and climate systems. This study examines the future changes in MJO projected by 23 Coupled Model Intercomparison Project Phase 6 (CMIP6) models that produce a realistic MJO propagation in their historical runs. Results from the multi‐model mean show a ∼17% increase in MJO precipitation amplitude, a ∼11%–14% increase in MJO circulation amplitude, a ∼9% increase in propagation speed, a ∼2‐day decrease in MJO period, and a ∼5° eastward extension. Analysis of the lower tropospheric moisture budget suggests the dominant role of an increased meridional advection of mean moisture caused by the steepening of mean moisture gradient over the Indo‐Pacific warm pool in a warming climate in the majority of models. The enhanced zonal moisture advection plays a secondary or comparable role in some models. The stronger anticyclonic gyres to the east of the MJO convection center along with the enhanced moisture gradient favor an enhanced export of moisture away from the Equator and a local moistening at the flanks of the MJO convection center. This leads to a stronger positive moisture tendency to the east of MJO convection and hence an enhanced eastward MJO propagation with strengthened amplitude and faster speed. As the MJO changes in the future, its relationships with the precipitation and temperature extremes over the western United States (US) are found to change accordingly. The MJO impacts on precipitation tend to be stronger with the stronger MJO amplitude, whereas its impacts on temperature are generally weaker.

A process-level understanding of the volcanically forced climate response is an urgent challenge due to its similarities to the potential effects of geoengineering techniques. Although the influence of volcanic forcing on El Niño events has been studied extensively, the mechanisms driving the volcanically-induced immediate onset of El Niño remain uncertain, with many climate models producing a delayed El Niño response compared to observations. In this study, using large ensemble simulations that allow us to isolate the impacts of volcanic forcing on the El Niño response, we demonstrate a mechanism that highlights the central triggering role of the Madden–Julian oscillation (MJO), which has been overlooked in existing literature. Because the land areas surrounding the Indo-Pacific warm pool dry more quickly after a volcanic eruption, the background moisture distribution becomes more favorable for the MJO to propagate eastward from the Indian Ocean into the Pacific. This increases the likelihood of ensemble members having stronger MJO activity in the western Pacific by about 35% compared to non-volcanic years, which subsequently increases the frequency of westerly wind bursts by about 76%, ultimately enhancing the probability of the onset of an El Niño by about 98% following major volcanic eruptions.

Previous studies have highlighted the significant impacts of El Niño–Southern Oscillation (ENSO) and the Madden–Julian Oscillation (MJO) on wintertime precipitation over the contiguous United States (CONUS). Here, we demonstrate skillful statistical prediction of subseasonal precipitation over the CONUS using the information of ENSO and the MJO. Simple statistical tools, such as multiple linear regression, exhibit significant improvement in prediction when trained with large ensemble climate simulations, surpassing those trained solely on observational data. Despite the biases in ENSO and MJO teleconnections in the climate simulations, the abundance of data, exceeding observational records by 100 times, allows more robust statistical relationships to be established, leading to such improvement. The utilization of machine learning tools yields additional gains in prediction skill beyond multiple linear regression. ENSO emerges as a dominant contributor to prediction skill, surpassing the influence of the MJO, whose impact diminishes with increasing forecast lead time.

This study continues the evaluation of the prediction of MJO teleconnections in two versions of the NOAA Unified Forecast System (UFS): prototype 5 (UFS5) and prototype 6 (UFS6). The key difference between the two prototypes is in the number of vertical layers (127 in UFS6 vs. 64 in UFS5) and model top (80 km in UFS6 vs. 54 km in UFS5), and the role of this increased resolution for the stratospheric pathway of MJO teleconnections is explored. The higher resolution prototype (UFS6) displays larger biases in its representation of tropospheric stationary waves, though both simulate a reasonable flux of wave activity entering the stratosphere. The stratospheric polar vortex in both prototypes is overly-sensitive to incoming wave activity, while both simulate a reasonable downward propagation of vortex anomalies within the stratosphere and down to the troposphere. The net effect is that both simulate the stratospheric pathway of the MJO better than has been documented for any forecasting system, though the hindcast period available for UFS differs from that used in previous studies. This allows them to predict, e.g., near-surface temperature over Eurasia with some skill on subseasonal timescales (week 3 to 5). However, this success could be the result of “two wrongs making a right”, and as individual biases are fixed, skill could temporarily be reduced. Overall, the two prototypes show similar performance in simulating the stratospheric route, suggesting that in the UFS, the increase in model vertical resolution has a limited impact on the prediction of MJO teleconnections via the stratosphere.

This study evaluates the prediction skill of the North Atlantic Oscillation (NAO) pattern and its associated energy budget as simulated by the European Center for Medium‐Range Weather Forecasts ensemble forecast system. By classifying NAO events into high‐ and low‐skill cases, we analyzed the stationarity of NAO patterns and the role of baroclinic energy conversion in NAO prediction. In both positive and negative NAO phases, high‐skill cases exhibited more stationary NAO patterns than low‐skill cases. The analysis of processes indicates that high‐skill NAO cases are due to stronger baroclinic maintenance of NAO, with its initial position at the climatological thermal trough, whereas low‐skill NAO cases result from forecast biases in wave propagation from the North Pacific. Specifically, biases in baroclinic energy conversion in the meridional direction from week 2 lead to weak advection of the eddy available potential energy (EAPE), resulting in lower prediction skill.

To better understand and improve the prediction of the seasonal surface temperature (TS) across the United States, we employed convolutional neural network (CNN) models trained on the Community Earth System Model Version 2 Large Ensemble (CESM2 LENS). We used lagged sea surface temperatures (SST) over the tropical Pacific region, containing the information of the El Niño Southern Oscillation (ENSO), as input for the CNN models. ENSO is the principal driver of variability in seasonal US surface temperatures (TS_US) and employing CNN models allows for spatiotemporal aspects of ENSO to be analyzed to make seasonal TS_US predictions. For predicting TS_US, the CNN models exhibited significantly improved skill over standard statistical multilinear regression (MLR) models and dynamical forecasts across most regions in the US, for lead times ranging from 1 to 6 months. Furthermore, we employed the CNN models to predict seasonal TS_US during extreme ENSO events. For these events, the CNN models outperformed the MLR models in predicting the effects on seasonal TS_US, suggesting that the CNN models are able to capture the ENSO‐TS_US teleconnection more effectively. Results from a heatmap analysis demonstrate that the CNN models utilize spatial features of ENSO rather than solely the magnitude of the ENSO events, indicating that the improved skill of seasonal TS_US is due to analyzing spatial variation in ENSO events. The proposed CNN model demonstrates a promising improvement in prediction skill compared to existing methods, suggesting a potential path forward for enhancing TS_US forecast skill from subseasonal to seasonal timescales.

The characteristics of the Madden–Julian oscillation (MJO) have changed and are projected to continue changing with the expansion of the Indo–Pacific warm pool, which is the Earth's largest region of warm sea surface temperatures (SSTs). However, the likelihood of a change in MJO predictability following warm pool expansion remains unaddressed. Therefore, this study investigated the effect of warm pool expansion on MJO variability and predictability using the highly idealized aquaplanet configuration of Community Earth System Model 2 (CESM2). By expanding the warm pool in the Indo–Pacific, MJO‐like waves become more regionally confined, short‐lived convective events with weaker magnitude and less robust eastward propagating signals, possibly due to stronger zonal SST gradients and wider meridional widths of the warm pool. Perfect‐model ensemble forecast experiments revealed that the MJO predictability decreased by approximately 5 days, the forecast error proliferated, and the signal rapidly reduced following warm pool expansion.

Studies have indicated exaggerated Maritime Continent (MC) barrier effect in simulations of the Madden–Julian oscillation (MJO), a dominant source of subseasonal predictability in the tropics. This issue has plagued the modeling and operational forecasting communities for decades, while the sensitivity of MC barrier on MJO predictability has not been addressed quantitatively. In this study, perfect-model ensemble forecasts are conducted with an aquaplanet configuration of the Community Earth System Model version 2 (CESM2) in which both basic state and tropical modes of variability are reasonably simulated with a warm pool–like SST distribution. When water-covered terrain mimicking MC landmasses is added to the warm pool–like SST framework, the eastward propagation of the MJO is disturbed by the prescribed MC aqua-mountain. The MJO predictability estimate with the perfect-model experiment is about 6 weeks but reduces to about 4 weeks when the MJO is impeded by the MC aqua-mountain. Given that the recent operational forecasts show an average of 3–4 weeks of MJO prediction skill, we can conclude that improving the MJO propagation crossing the MC could improve the MJO skill to 5–6 weeks, close to the potential predictability found in this study (6 weeks). Therefore, more effort toward understanding and improving the MJO propagation is needed to enhance the MJO and MJO-related forecasts to improve the subseasonal-to-seasonal prediction.