“Problem statement: The North American Multi-Model Ensemble (NMME) includes global ensemble forecasts at lead times of weeks to months since 2011 from several leading climate models. Approximately 30 years of hindcasts from each model are available for calibrating the model forecast skill. It is hypothesized that suitably bias-corrected and calibrated probabilistic forecasts based on NMME can help predict the risk of temperature extremes at monthly to seasonal lead times. Temperature extremes impact many sectors, including agriculture, water resources, energy, transportation, and health.
Work plan: We will apply and extend methods that we and others have developed for quantifying and improving forecast skill, including trend extrapolation, quantification of forecast information gain and over- or under-confidence, and different approaches to probability distribution estimation and model weighting, to provide NOAA with usable probability forecasts and with diagnostic outputs for model improvement. The research will be oriented toward developing and testing general algorithms that can be applied to operational forecasts of any desired climate properties with limited manual tuning. The PI will solicit feedback from the NOAA Climate Prediction Center (CPC) operational prediction branch, which has expressed interest in the problem, to develop pilot products of interest to CPC operations and stakeholders.
Relevance to competition: This proposal targets exploration of new applications within the NMME System Evaluation and Application competition. The proposed work will seek to explore a novel application of NMME output for providing warnings of heightened risk of temperature extremes, at time scales and lead times of 1-3 months. The results from the proposed work will extend and apply previous findings across NMME models and forecast products, resulting in pilot probabilistic seasonal temperature forecasts that could find a wide variety of applications.
Relevance to NOAA goals: NOAA’s Next-Generation Strategic Plan goals of im-
proved scientific understanding of the climate system and of providing climate services to complement currently available weather analysis and forecast services are challenged by the current gap in prediction tools for the intraseasonal to interannual timescale. This project will help NOAA move toward filling this gap by developing scientific and computational tools to utilize the predictive abilities in current and future generations of numerical models to produce well-calibrated probabilistic forecasts of quantities of practical interest.”