Search

Atmospheric rivers (ARs) are long, narrow synoptic scale weather features important for Earth's hydrological cycle typically transporting water vapor poleward, delivering precipitation important for local climates. Understanding ARs in a warming climate is problematic because the AR response to climate change is tied to how the feature is defined. The Atmospheric River Tracking Method Intercomparison Project (ARTMIP) provides insights into this problem by comparing 16 atmospheric river detection tools (ARDTs) to a common data set consisting of high resolution climate change simulations from a global atmospheric general circulation model. ARDTs mostly show increases in frequency and intensity, but the scale of the response is largely dependent on algorithmic criteria. Across ARDTs, bulk characteristics suggest intensity and spatial footprint are inversely correlated, and most focus regions experience increases in precipitation volume coming from extreme ARs. The spread of the AR precipitation response under climate change is large and dependent on ARDT selection. Plain Language Summary: Atmospheric rivers (ARs) are long and narrow weather features often referred to as "rivers in the sky." They often transport water from lower latitudes to higher latitudes typically across climate zones and produce precipitation necessary for local climates. Understanding ARs in a warming climate is challenging because of the variety of ways an AR can be defined on gridded data sets. Unlike weather features such as tropical cyclones where identification methodologies are similar, algorithms that determine the characteristics of ARs vary depending on the science question posed. Because there is no real consensus on AR identification methodology, we aim to quantify the algorithmic uncertainty in AR metrics and precipitation. We compare 16 different ways of defining an AR on gridded data sets and present the range of possibilities in which an AR could change under global warming. Generally, ARs are projected to increase but the amount of that increase is a function of the algorithm. Across all algorithms and focus regions, AR precipitation is projected to become more extreme. Key Points: High‐resolution historical and future simulations are used to evaluate atmospheric river detection tools (ARDT) uncertaintyARDTs mostly show increases in frequency and intensity of future atmospheric rivers (ARs) but the scale of response is dependent on algorithmic restrictivenessMost regions experience an increase in precipitation volume coming from extreme ARs [ABSTRACT FROM AUTHOR]

Atmospheric rivers (ARs)—narrow corridors of high atmospheric water vapor transport—occur globally and are associated with flooding and maintenance of the water supply. Therefore, it is important to improve forecasts of AR occurrence and characteristics. Although prior work has examined the skill of numerical weather prediction (NWP) models in forecasting atmospheric rivers, these studies only cover several years of reforecasts from a handful of models. Here, we expand this previous work and assess the performance of 10–30 years of wintertime (November–February) AR landfall reforecasts from the control runs of nine operational weather models, obtained from the International Subseasonal to Seasonal (S2S) Project database. Model errors along the west coast of North America at leads of 1–14 days are examined in terms of AR occurrence, intensity, and landfall location. Occurrence-based skill approaches that of climatology at 14 days, while models are, on average, more skillful at shorter leads in California, Oregon, and Washington compared to British Columbia and Alaska. We also find that the average magnitude of landfall integrated water vapor transport (IVT) error stays fairly constant across lead times, although overprediction of IVT is common at later lead times. Finally, we show that northward landfall location errors are favored in California, Oregon, and Washington, although southward errors occur more often than expected from climatology. These results highlight the need for model improvements, while helping to identify factors that cause model errors. [ABSTRACT FROM AUTHOR]