The American Meteorological Society continuously publishes research on climate, weather, and water in its 12 journals. Many of these articles are available for early online access — they are peer-reviewed, but not yet in their final published form. Below are some recent examples of online and early-online research.
JOURNAL ARTICLES
What follows are summaries that have not been peer-reviewed or vetted by the article authors; read the full article for peer-reviewed conclusions. Please note that no single study is ever definitive, and each must be taken in the context of the broader scientific literature.
Atmospheric Predictability Beyond 30 Days with Machine Learning
Artificial Intelligence for the Earth Systems
Identifying optimal starting conditions can reduce the weather forecast “butterfly effect.” It’s accepted wisdom, when running a weather forecast model, that small errors and uncertainties in the starting conditions run compound into larger errors when that forecast is extended, especially beyond two weeks (aka the “butterfly effect”). Using past forecast data covering 2020, this paper finds that the GraphCast ML model can quickly pinpoint the starting conditions for model runs that led to the most correct forecast for past events. When those conditions are used to create a new forecast, it reduces errors an average of 86% for a 10-day forecast and produces “skillful deterministic forecasts far beyond two weeks,” and even past 30 days. Figuring out how to identify these optimal conditions in real time could open ways to substantially increase forecast lead time.
Episodic Slowdown of Global Warming by a Multi-Year La Niña
Journal of Climate
Even a weak multiyear La Niña can have a strong global cooling effect. While a strong El Niño will usually increase global mean surface temperature (GMST), the cooling effect of a weak, multiyear La Niña event is less studied. Using observations and climate models, the authors find that multiple years of La Niña amplify each others’ cooling effects, even when each year’s La Niña is weaker than the last, causing a slowdown of global warming.
From Binary to Bilingual: How the National Weather Service is Using Artificial Intelligence to Develop a Comprehensive Translation Program
Artificial Intelligence for the Earth Systems
Using AI to develop warnings for all. The National Weather Service is working with LILT to train artificial intelligence to translate weather terminology and messaging into multiple languages, including Spanish, Simplified Chinese, and Vietnamese. A new website features experimental multilingual NWS products, including translated warnings, 7-day forecasts, and educational campaigns. The developers used GIS to identify communities in need and integrated ethical AI practices into translation oversight and development.
Vegetation Dynamics and Their Contribution to the Little Ice Age
Weather, Climate, and Society
Did plant growth and human deaths contribute to the Little Ice Age? There are many theories about what caused the Little Ice Age (LIA), an aberrantly cold period in the North Atlantic that lasted hundreds of years. This paper suggests the LIA may have been triggered by depopulation of the Americas due to disease after European incursions, with the abandonment of farmlands allowing plants to reclaim the land and lower global temperatures. It explores how plant growth can alter ocean currents, bringing cold deep-ocean water to the surface and cooling the planet.
Variability and Trends in Dates of the Last Spring Freeze and First Fall Freeze over the United States
Journal of Applied Meteorology and Climatology
“Freeze-free” season is lengthening across the United States. This study uses the MERRA-2 model to analyze the climatology, trends, and interannual variability of the first fall freeze and last spring freeze across the United States for the years 1980–2023. It finds that across most regions, the last spring freeze date has trended earlier, and the first fall freeze date has trended later, resulting in a longer freeze-free season that impacts agriculture, ecosystems, and water resources.
A Supplemental High Resolution Weather Radar Network for the Conterminous United States
Bulletin of the American Meteorological Society
Establishing better U.S. radar coverage. Many areas of the continental United States lack adequate radar coverage, and climate change increases the need for coverage in some underserved areas. This paper describes how U.S. startup Climavision is establishing a supplemental network of 200+ polarimetric X-band radars that will work with a planned upgrade of the NWS radar network deployed throughout the country. The supplemental radars can scan lower to the ground than distant federal radars, aiding in rainfall estimates, detection of severe weather features, and more.
Evidence of Surface Wave Reduction by Hurricane-Generated Ocean Currents
Journal of Physical Oceanography
Confirmation: Some hurricane-driven currents reduce ocean surface wave heights. Simulations in a recent study indicated that when tropical cyclone-induced currents align well with the direction of surface waves, these currents can increase the waves’ speed such that they spend less time in areas of high wind, ultimately leading to more-frequent, but shorter-height waves. This follow-on study validates that conclusion by comparing model predictions to observations from four recent hurricanes (Ian, Idalia, Helene, and Milton) and confirms that incorporating storm-driven currents improves a model’s wave forecast.
The Warming Edge: Increasing Risk of Extreme Temperatures across South Asia over the Last 45 Years
Journal of Climate
Pakistan, Nepal, Bangladesh are “sitting ducks” for unprecedented warm-season heat events. Combining data archives with the new UNSEEN ensemble forecasting model, the authors find that the risk of heat events considered extreme in 1981 has increased over the past several decades for most of Pakistan, Nepal, and Bangladesh, as well as parts of India. However, because the most extreme predictedevents have yet to materialize, nations may be unprepared for these now very likely events.
Projected Changes in Both Mean Climate and Climate Variability Drive Substantial Increases in Extreme Fire Weather in the Western United States
Journal of Climate
Extreme Western fire weather likely to spread more widely over connected regions. This study agrees with others that predict increasing western U.S. wildfire risk due to fire weather becoming more frequent, persistent, and widespread. However, the authors also find that extreme fire weather events are likely not only to increase in frequency but to spread over larger areas and become more spatially connected, a finding important for shaping adaptation strategies.
Characteristics of Temperature–Precipitation Compound Extreme Events in China and Relationship to ENSO and PDO
Journal of Hydrometeorology
Extremely hot and dry nights increasing rapidly in China. The frequency of both nighttime and daytime compound extreme events (e.g., extreme heat PLUS extreme aridity or precipitation) has increased between 1961 and 2021 across most areas of China, this study suggests. The authors find the greatest overall increase in compound extreme events in the Southwest River Basin. Hot and dry events, especially hot-dry nighttime extreme events, showed the largest increase over time.
Increasing Socioeconomic Exposure to Extreme Tropical Cyclone Precipitation over China
Journal of Hydrometeorology
Tropical cyclone precipitation intensity drives increased socioeconomic risk in China. This study finds that socioeconomic exposure to risk from extreme tropical cyclone-induced precipitation (TCP) has increased in southeastern and east-central China, while decreasing in parts of the Yangtze River Valley. The increase in exposure is due mostly to increased TCP intensity and duration, but is magnified by rapid urbanization and economic expansion.
Offering an Evidence-Based Recommendation for Improving the Storm Prediction Center’s Categorical Outlook
Weather, Climate, and Society
Storm Prediction Center risk categories may be better conveyed with a numerical scale. U.S. Storm Prediction Center (SPC) thunderstorm outlooks use a verbal scale to describe hazard likelihood from “Marginal” to “High.” Surveys of the public and focus groups of users like emergency managers suggest that a numerical scale (1-5) leads to more consistent outcomes and understanding, but that users still want descriptive words attached to the numbers.
You can view all research published in AMS Journals at journals.ametsoc.org.