Agniv Sengupta

The value of improving longer-lead precipitation forecasting in the water-stressed, semi-arid western United States cannot be overstated, especially considering the severity and frequency of droughts that have plagued the region for much of the 21st century. Seasonal prediction skill of current operational forecast systems, however, remain insufficient for decision-making purposes across a variety of applications. To address this capability gap, we develop a seasonal forecasting system that leverages the long-term memory of leading global and basin-scale modes of sea surface temperature variability. This approach focuses on characterizing and capitalizing on the spatiotemporal evolution of predictor modes over multiple antecedent seasons, instead of the customary use of predictive information from just the current season. Another distinctive methodological feature is the incorporation of sources of predictability spanning multiple timescales, from interannual to decadal-multidecadal. An evaluation of the forecast system’s performance from cross-validation analyses demonstrates skill over core winter precipitation regions—California, Pacific Northwest, and the Upper Colorado River basin. The developed model exhibits superior skill compared to dynamical and statistical benchmarks in predicting winter precipitation. Experimental seasonal precipitation forecasts from the model have the potential to provide critical situational awareness guidance to stakeholders in the water resources, agriculture, and disaster preparedness communities.

Accurate precipitation forecasting often relies on high-resolution numerical weather prediction (NWP) models, which are essential for capturing fine-scale and nonlinear atmospheric dynamics. However, the computational demands of these models can be substantial. Leveraging recent advancements in artificial intelligence (AI), we present a stretched-grid AI-driven weather model with 6-km horizontal grid increments over the Western United States and  31 km in other regions globally. The model employs an autoregressive framework to generate forecasts in minutes and is evaluated against global and regional NWP systems, as well as a lower-resolution AI model. Our results show that the regional AI model reduces 24-h accumulated precipitation errors, performs competitively with the regional NWP model, and effectively captures extreme precipitation events, particularly those linked to atmospheric rivers, which global coarser models often underestimate. This work underscores the potential of regional, high-resolution AI models for precipitation forecasting at km-scales, and discusses some of the challenges for future development.

Artificial Intelligence (AI) weather models are explored for initial condition sensitivity studies to analyze the physicality of the relationships learned. Gradients (or sensitivities) of the target metric of interest are computed with respect to the variable fields at initial time by means of the backpropagation algorithm, which does not assume linear perturbation growth. Here, sensitivities from an AI model at 36-h lead time were compared to those produced by an adjoint of a dynamical model for an extreme weather event, cyclone Xynthia, presenting very similar structures and with the evolved perturbations leading to similar impacts. This demonstrates the ability of the AI weather model to learn physically meaningful spatio-temporal links between atmospheric processes. These findings should enable researchers to conduct initial condition studies in minutes, potentially at lead times into the non-linear regime (typically >5 days), with important applications in observing network design and the study of atmospheric dynamics.