AbstractArtificial intelligence (AI) and machine learning (ML) pose a challenge for achieving science that is both reproducible and replicable. The challenge is compounded in supervised models that depend on manually labeled training data, as they introduce additional decision‐making and processes that require thorough documentation and reporting. We address these limitations by providing an approach to hand labeling training data for supervised ML that integrates quantitative content analysis (QCA)—a method from social science research. The QCA approach provides a rigorous and well‐documented hand labeling procedure to improve the replicability and reproducibility of supervised ML applications in Earth systems science (ESS), as well as the ability to evaluate them. Specifically, the approach requires (a) the articulation and documentation of the exact decision‐making process used for assigning hand labels in a “codebook” and (b) an empirical evaluation of the reliability” of the hand labelers. In this paper, we outline the contributions of QCA to the field, along with an overview of the general approach. We then provide a case study to further demonstrate how this framework has and can be applied when developing supervised ML models for applications in ESS. With this approach, we provide an actionable path forward for addressing ethical considerations and goals outlined by recent AGU work on ML ethics in ESS.

High‐resolution Taiwan Climate Change Projection Information and Adaptation Knowledge Platform (TCCIP) gridded precipitation data are used to characterize days in the Mei‐yu season with the most extreme precipitation (EP). These “EP days” are grouped into weather types based on the presence of features such as tropical cyclones (TCs) and atmospheric rivers (ARs), then analyzed from the perspective of weather type frequency and synoptic changes. During the 1979–2019 period, EP days associated with ARs were associated with significant increasing trends in season‐total precipitation. These AR‐related precipitation increases are due to four events in 2005, 2006, 2012, and 2017 which had long duration and unusually intense precipitation, and which were anomalous even within the longer 1960–2019 time period. Meanwhile, TC‐related EP days contribute less precipitation than they did in the 1980s due to decreased frequency of TCs on EP days and in the Mei‐yu season climatology. Over the 1979–2019 period, the AR‐related and TC‐related trends combine to produce EP increases in western Taiwan and decreases in eastern Taiwan. Mei‐yu season southwesterly integrated vapor transport (IVT), wind speed, and specific humidity have all increased in association with these extreme events. Low‐level winds appear to the primary factor influencing the IVT increase, with increased moisture also contributing. The wind trends are consistent with climatological pressure increases south of Taiwan and decreases over the East Asian landmass, which facilitate a strengthened circulation in a corridor extending from the southern China coastline over Taiwan during this season.

We introduce the National Science Foundation (NSF) AI Institute for Research on Trustworthy AI in Weather, Climate, and Coastal Oceanography (AI2ES). This AI institute was funded in 2020 as part of a new initiative from the NSF to advance foundational AI research across a wide variety of domains. To date AI2ES is the only NSF AI institute focusing on environmental science applications. Our institute focuses on developing trustworthy AI methods for weather, climate, and coastal hazards. The AI methods will revolutionize our understanding and prediction of high-impact atmospheric and ocean science phenomena and will be utilized by diverse, professional user groups to reduce risks to society. In addition, we are creating novel educational paths, including a new degree program at a community college serving underrepresented minorities, to improve workforce diversity for both AI and environmental science.

A review of the mean state over the tropical eastern Pacific (EPAC) and the Intra-Americas Sea (IAS) region is provided to assess the characteristics that impact the development and genesis of easterly waves (EWs). The EPAC-IAS region is characterized by complex topography, the Western Hemisphere warm pool, the ITCZ at 10 degrees N, and predominant deep convection over the Panama Bight around 9 degrees N, 78 degrees W. A prominent easterly jet at 600 hPa of about 5.5 m s(-1), is oriented approximately parallel to the Mexican coast. The jet is characterized by a strip of high potential vorticity (PV) on the cyclonic shear side and low PV on the anticyclonic side. This distribution of PV satisfies the necessary conditions for barotropic instability: the Charney-Stern condition, as well as the Fjortoft condition. Together these conditions suggest the potential for barotropic growth of EWs over the EPAC region. The mean high PV region over the EPAC is created in association with two different populations of cloud/convection systems: stratiform and shallow, with the former being key for the creation of positive PV anomalies at midlevels. Evidence is also provided that suggests that the low PV region arises in association with sources of negative PV anomalies over the Sierra Madre region likely resulting from frequent dry convection. This is a key and novel result that is basic for the setting up of a negative meridional PV gradient and fundamental for the Charney-Stern condition associated with barotropic instability and growth of EWs. Significance StatementThe tropical eastern Pacific is influenced by synoptic easterly waves that impact daily weather in the region and can trigger tropical cyclones. This research explores the nature of a midlevel jet that supports the development of easterly waves in this basin. The jet is established in association with moist convection over the ocean that leads to a midlevel potential vorticity maximum equatorward of the jet and, frequent dry convection over the Mexican Sierra Madre region that leads to a low-level potential vorticity minimum poleward of the jet. This finding highlights the need to better understand, and ultimately predict, these potential vorticity sources in order to better understand and predict the nature of the easterly wave developments in this region.

Abstract A time‐series analysis of Easterly Wave (EW) activity, based on dynamical and convective variance measures, was carried out over the tropical northeastern Pacific (EPAC). A significant interdecadal change in EW‐activity was identified shifting from reduced activity in 1980–1997, to increased activity during 1998–2015. The changes in EW‐activity are modulated on interdecadal timescales by the Atlantic Multidecadal Oscillation (AMO) and the Pacific Decadal Oscillation (PDO). EW‐activity is increased when there is a positive AMO index and, to a lesser extent, a negative PDO index. The opposite occurred with a negative AMO index and positive PDO index. This relationship can be understood in terms of how the AMO and PDO impact sea surface temperatures (SSTs) in the EPAC. Warmer SSTs tend to be associated with more EW‐related convection and associated vorticity generation in the EW troughs and vice versa. These results suggest potential predictability of EW‐activity in the EPAC on interdecadal timescales.

This study uses Global Historical Climate Network (GHCN) data in each season to identify the days with the most extreme precipitation ("EP days") in the mid-Atlantic and northeast United States between 1979 and 2019. These days are sorted according to the fraction of extreme precipitation attributed to tropical cyclone (TC), atmospheric river (AR), and extreme integrated vapor transport (IVT) influences. In winter and spring, there have been increases in seasonal precipitation from the most extreme days, associated with a combination of frequency and intensity changes. These increasing trends come primarily from atmospheric rivers. In summer and fall, there have also been large increases in precipitation on extreme days, in this case due entirely to increased event frequency. These changes come from a combination of AR, TC, and extreme IVT influences. Synoptic characteristics of AR-related EP days in winter and spring have changed significantly. In winter, there has been an amplification of the Atlantic ridge and a deepening of the upstream trough over the upper Great Plains, as well as enhanced AR detection and IVT on these days. The composite low has shifted north and intensified. In spring, the trough has weakened and 1000-500-hPa thickness has increased broadly to the south. These changes are related to changes in the large-scale flow. In winter and spring, the North Atlantic subtropical high (NASH) has strengthened and shifted west, leading to increased southwesterly IVT over the mid-Atlantic and Northeast United States. In summer, southerly IVT along the east coast has increased, and 1000-500-hPa climatological thickness has increased broadly in all seasons. Significance StatementThis paper studies the days with the most extreme precipitation over the mid-Atlantic and Northeast United States in each season and finds that there have been large changes in the frequency, intensity, and characteristics of some of these days. Extreme days associated with atmospheric rivers (ARs) and tropical cyclones (TCs) have become more frequent. Precipitation on AR-related days has also become more intense in some areas. On winter extreme days associated with ARs, the pattern has become more amplified and moisture fluxes have become stronger; these changes are likely associated with strengthened high pressure over the Atlantic. Throughout all seasons, there is evidence of warming on extreme days. This is important because it adds to our understanding of how total and AR-related extreme precipitation is changing in a warming climate.

While considerable attention has been given to how convectively coupled Kelvin waves (CCKWs) influence the genesis of tropical cyclones (TCs) in the Atlantic Ocean, less attention has been given to their direct influence on African easterly waves (AEWs). This study builds a climatology of AEW and CCKW passages from 1981 to 2019 using an AEW-following framework. Vertical and horizontal composites of these passages are developed and divided into categories based on AEW position and CCKW strength. Many of the relationships that have previously been found for TC genesis also hold true for non-developing AEWs. This includes an increase in convective coverage surrounding the AEW center in phase with the convectively enhanced ("active") CCKW crest, as well as a buildup of relative vorticity from the lower to upper troposphere following this active crest. Additionally, a new finding is that CCKWs induce specific humidity anomalies around AEWs that are qualitatively similar to those of relative vorticity. These modifications to specific humidity are more pronounced when AEWs are at lower latitudes and interacting with stronger CCKWs. While the influence of CCKWs on AEWs is mostly transient and short lived, CCKWs do modify the AEW propagation speed and westward-filtered relative vorticity, indicating that they may have some longer-term influences on the AEW life cycle. Overall, this analysis provides a more comprehensive view of the AEW-CCKW relationship than has previously been established, and supports assertions by previous studies that CCKW-associated convection, specific humidity, and vorticity may modify the favorability of AEWs to TC genesis over the Atlantic.

Taiwan regularly experiences precipitation extremes of hundreds of millimeters per day, especially between May and September. In this study, Taiwan's extreme rainfall (ER) is analyzed over a 56-yr time period in different seasons and geographic regions, using a recently released, high-resolution gridded rainfall dataset. ER is defined using a seasonally and geographically varying 99th-percentile threshold to better resolve the characteristics of the most intense rainfall seen in different locations and times of year. The resulting monthly ER rates are largest in typhoon season and smallest in fall, winter, and spring. ER is spatially homogeneous in the mei-yu and typhoon seasons and concentrated in northern Taiwan during the rest of the year. A trend analysis revealed a positive trend in island-mean ER for the winter, spring, and typhoon seasons. In winter and spring, these trends are most pronounced in the north. In the mei-yu season, ER has increased most over the southwestern mountain slopes; in typhoon season, ER has increased consistently over much of Taiwan. These changes often exceed 1% yr(-1). In many areas, typhoon season accounts for the largest fraction of the observed annual ER trend. TCs produce most of the observed typhoon season ER and ER trend, with nearly half of the typhoon season ER trend being associated with increases in TC frequency and duration around central and northern Taiwan. Certain regional changes in ER characteristics, particularly in areas with low sample size or complex seasonal contributions, merit further investigation in future work.

Our transformational science question is: can we revolutionize both the prediction and understanding of extreme events through trustworthy AI? Our use-cases include extreme weather such as tornadoes and hail as well as water-based events including extreme precipitation, compound flooding, harmful algal blooms, and sea turtle cold stunnings and nest inundations.

This paper explores a new mechanism for in situ genesis of easterly waves (EWs) over the tropical eastern Pacific Ocean (EPAC). Using an idealized primitive equation model, it is shown that EWs can be triggered by finite-amplitude transient heating close to the midlevel jet at about 15 degrees N over the EPAC and intra-Americas sea region. The atmospheric response to heating initiates EWs downstream, showing an EW structure within 4 days, with a wavelength and propagation speed of about 2000 km and 4.6 m s(-1), respectively, resembling EWs described in the literature. The most sensitive location for EW initiation from finite-amplitude transient heating is located over the northern part of South America and extends to the EPAC. The closer the heating is to the jet, the bigger the response is. A stratiform heating profile is the most efficient at triggering EPAC EWs. Comparisons of simulated EWs over the EPAC and West Africa reveal similar structures but with a shorter wavelength and much weaker amplitudes over the EPAC. EPAC EWs are dominated by horizontal tilts against the shear on the equatorial side of the jet, consistent with barotropic growth, with weaker low-level amplitudes relative to those seen over West Africa. These differences arise from differences in the mean state EPAC having a shorter and weaker midlevel jet with less baroclinicity.

Vertical profiles of atmospheric temperature, moisture, wind, and aerosols are essential information for weather monitoring and prediction. Their availability, however, is limited in space and time due to the significant resources required to observe them. To fill this gap, the New York State Mesonet (NYSM) Profiler Network has been deployed as a national testbed to facilitate the research, development and evaluation of ground-based profiling technologies and applications. The testbed comprises 17 profiler stations across the state, forming a long-term regional observational network. Each Profiler station is comprised of a ground-based Doppler lidar, a microwave radiometer (MWR) and an environmental Sky Imaging Radiometer (eSIR). Thermodynamic profiles (temperature and humidity) from the MWR; wind and aerosol profiles from the Doppler lidar; and solar radiance and optical depth parameters from the eSIR are collected, processed, disseminated, and archived every 10 minutes. This paper introduces the NYSM Profiler Network and reviews the network design and siting, instrumentation, network operations and maintenance, data and products, and some example applications highlighting the benefits of the network. Some sample applications include improved situational awareness and monitoring of the sea/land breeze, long-range wildfire smoke transport, air quality (PM 2.5 and AOD) and boundary layer height. Ground-based profiling systems promise a path forward for filling a critical gap in the nation’s observing system with the potential to improve analysis and prediction for many weather-sensitive sectors, such as aviation, ground transportation, health, and wind energy.

The New York State Mesonet (NYSM) is a network of 126 standard environmental monitoring stations deployed statewide with an average spacing of 27 km. The primary goal of the NYSM is to provide high-quality weather data at high spatial and temporal scales to improve atmospheric monitoring and prediction, especially for extreme weather events. As compared with other statewide networks, the NYSM faced considerable deployment obstacles with New York’s complex terrain, forests, and very rural and urban areas; its wide range of weather extremes; and its harsh winter conditions. To overcome these challenges, the NYSM adopted a number of innovations unique among statewide monitoring systems, including 1) strict adherence to international siting standards and metadata documentation; 2) a hardened system design to facilitate continued operations during extreme, high-impact weather; 3) a station design optimized to monitor winter weather conditions; and 4) a camera installed at every site to aid situational awareness. The network was completed in spring of 2018 and provides data and products to a variety of sectors including weather monitoring and forecasting, emergency management, agriculture, transportation, utilities, and education. This paper focuses on the standard network of the NYSM and reviews the network siting, site configuration, sensors, site communications and power, network operations and maintenance, data quality control, and dissemination. A few example analyses are shown that highlight the benefits of the NYSM.