Adjoint-Derived Sensitivity Diagnosis of Mid-Latitude Weather Systems

University Of Wisconsin-Madison, Madison WI

Investigators

Abstract

In the context of an numerical weather prediction model, adjoint model output provides the sensitivity of a given function of forecast model output (the response function) to changes in the model state as represented by the model forecast trajectory. Despite their potential value, adjoint models remained an under-utilized tool in synoptic case studies. This research is aimed at developing a set of diagnostic tools that would help with interpretation of adjoint model sensitivities. These adjoint tools will then be applied to a number of sensitivity studies related to the structure, intensity, and predictability of idealized and observed midlatitude weather systems. The results from the research will be useful in better designing of traditional forecast and impact studies. Intellectual merit: The PI will address the following scientific questions. 1. How does the sensitivity of a particular response function change with respect to time during the evolution of idealized and observed cyclogenesis events? 2. What are the principal regions of short range (i.e., 24 to 60 h) forecast error sensitivity over the eastern United States and how do those regions relate to kinematic and dynamic characteristics of the flow field? Broader Impacts: The research will be integrated within graduate education and continued training of students in the application of adjoint tools and provision of the diagnostic tools to a wider research community. The research will lead to development of community resources, and furtherance of predictability and data assimilation research.