This study investigates precipitation variability over the Arabian Peninsula (AP) during its wet season. The wet season is split into winter (November – February) and spring (March and April) seasons, and early (1950–1986) and late (1986–2021) periods to understand sub-seasonal characteristics of precipitation variability and long-term changes in global teleconnections. The first three Empirical Orthogonal Functions explain ~70% of the interannual wet season precipitation variance, which shows an increase (decrease) in the late period winter (spring). Linear regression of the sea surface temperatures and geopotential height onto associated principal components reveals many oceanic and atmospheric variability patterns, which exhibit significant differences between winter and spring and early and late periods. Further, linear regressions of AP precipitation onto 14 natural modes of climate variability reveal a complex network of global teleconnections. El Niño-Southern Oscillation (ENSO) is one of the key contributors to precipitation variability but considering ENSO diversity is crucial to fully understand its influence. While the direct ENSO influence only becomes robust after the 1980s, its indirect effect persists through projection onto atmospheric modes, such as East Atlantic West Russia Pattern and East Atlantic Mode, or inter-basin interaction (e.g., via the Indian Ocean). The Northern Hemisphere atmospheric modes also mediate influences of other natural modes in tropical Indian and Atlantic oceans and extra-tropical regions over the AP. Several precipitation teleconnections exhibit a shift in the 1980s. Some may be related to the introduction of satellite data, but further investigations are warranted to understand the causes of these shifts.

A key consideration for evaluating climate projections is uncertainty in radiative forcing scenarios. Although it is straightforward to monitor greenhouse gas concentrations and compare those observations with specified climate scenarios, it remains less obvious on how to connect regional climate patterns with these scenarios in real time. Here we introduce a machine learning approach for linking patterns of climate change with radiative forcing scenarios and use an attribution method to understand how these linkages are made. We train a neural network using output from the SPEAR Large Ensemble to classify whether temperature or precipitation maps are most likely to originate from one of several potential radiative forcing scenarios. The neural network learns to identify “fingerprint” patterns that associate signals of climate change with the scenarios. We illustrate this using output from additional mitigation experiments and highlight regions that are critical for associating the new simulations with likely radiative forcing scenarios.

Seasonal snowpack in the Western United States (WUS) is vital for meeting summer hydrological demands, reducing the intensity and frequency of wildfires, and supporting snow-tourism economies. While the frequency and severity of snow droughts (SD) are expected to increase under continued global warming, the uncertainty from internal climate variability remains challenging to quantify. Using a 30-member large ensemble from a state-of-the-art global climate model, the Seamless System for Prediction and EArth System Research (SPEAR), and an observations-based dataset, we find WUS SD changes are already significant. By 2100, SPEAR projects SDs to be nearly 9 times more frequent under shared socioeconomic pathway 5-8.5 (SSP5-8.5) and 5 times more frequent under SSP2-4.5. By investigating the influence of the two primary drivers of SD, temperature and precipitation amount, we find the average WUS SD will become warmer and wetter. To assess how these changes affect future summer water availability, we track April 15th snowpack across WUS watersheds, finding differences in the onset time of a “no-snow” threshold between regions and large internal variability within the ensemble that are both on the order of decades. For example, under SSP5-8.5, SPEAR projects California could experience no-snow anywhere between 2058 and 2096, while in the Pacific Northwest, the earliest transition happens in 2091. We attribute the inter-regional uncertainty to differences in the regions’ mean winter temperature and the intra-regional uncertainty to irreducible internal climate variability. This analysis indicates that internal climate variability will remain a significant source of uncertainty for WUS hydrology through 2100.

To better understand the regional changes in summertime temperatures across the conterminous United States (CONUS), we adopt a recently developed machine learning framework that can be used to reveal the timing of emergence of forced climate signals from the noise of internal climate variability. Specifically, we train an artificial neural network (ANN) on seasonally-averaged temperatures across the CONUS and then task the ANN to output the year associated with an individual map. In order to correctly identify the year, the ANN must therefore learn time-evolving patterns of climate change amidst the noise of internal climate variability. The ANNs are first trained and tested on data from large ensembles and then evaluated using observations from a station-based dataset. To understand how the ANN is making its predictions, we leverage a collection of ad hoc feature attribution methods from explainable artificial intelligence (XAI). We find that anthropogenic signals in seasonal mean minimum temperature have emerged by the early 2000s for the CONUS, which occurred earliest in the Eastern United States. While our observational timing of emergence estimates are not as sensitive to the spatial resolution of the training data, we find a notable improvement in ANN skill using a higher resolution climate model, especially for its early 20th century predictions. Composites of XAI maps reveal that this improvement is linked to temperatures around higher topography. We find that increases in spatial resolution of the ANN training data may yield benefits for machine learning applications in climate science.

The monsoons in Pakistan have been exceptionally harsh in recent decades, resulting in extraordinary drought conditions and record flooding events. The changing frequency of extreme events is widely attributed to climate change. However, given this region's long history of floods and droughts, the role of natural climate variability cannot be rejected without a careful diagnosis. Here, we examine how oceanic and atmospheric variability has contributed to unusual precipitation distributions in West South Asia. Variations in sea surface temperatures in the tropical Pacific and northern Arabian Sea, and internal atmospheric variability related to the circumglobal teleconnection pattern and the subtropical westerly jet stream, explain more than 70% of monthly summer precipitation variability in the 21st century. Several of these forcings have co-occurred with record strength during episodes of extreme monsoons, which have exacerbated the overall effect. Climate change may have contributed to increased variability and the in-phase co-occurrences of the identified mechanisms, but further research is required to confirm any such connection.