Enhancing the MDTF software package with a process-oriented diagnostic framework to evaluate drought and fire hazard uncertainty in climate models competition: mapp process-oriented diagnostics for…

Enhancing the MDTF software package with a process-oriented diagnostic framework to evaluate drought and fire hazard uncertainty in climate models competition: mapp process-oriented diagnostics for NOAA climate model improvement and applications predictability of drought and cascading wildfire hazard are essential to inform proactive management and policy designed to reduce the negative impacts of these disasters. Despite the importance of predicting drought and fire hazard under a changing climate, climate model intercomparison project - general circulation models (CMIP-GCMS) have shown low skill in predicting land-atmosphere interactions during drought and substantial uncertainty in drought intensity and duration over many global regions. Drought projections from earth system models suffer from uncertainties partially due to a lack of knowledge and quantification of process-level deficiencies. For instance, the mapp model diagnostic task force (MDTF) software package is missing key diagnostics for drought and fire hazard indices. Our overarching goal is to develop and implement a globally-applicable and novel process-oriented diagnostics (POD) framework to quantify uncertainties and diagnose processlevel error sources for an integrated drought index which is tightly coupled with fire hazard, the standardized vapor pressure deficit drought index (SVDI). The proposed POD will enable the project team as well as developers and users of CMIP-GCMS to readily answer the following three science questions: (1) how skillful are climate model simulations of atmospheric and land surface processes related to drought and fire hazard? (2) how are these process-level variables and uncertainties interconnected in different climate models? (3) which uncertainty factors are the primary contributors to climate model uncertainty in SVDI? The proposed POD will: (1) calculate a suite of climate model error statistics at monthly to annual time scales relative to observational products; (2) quantify the transfer of uncertainty between SVDI and related process-level variables to provide a more complete picture of model uncertainty and associated key contributing processes through two measures from information theory: mutual information and transfer entropy; and (3) produce statistics and figures which synthesize results highlighting uncertainty estimates of SVDI and associated primary uncertainty contributors. The proposed analyses and software development will specifically focus on CMIP6- GCMS but the methodological framework will be designed to be applicable to a variety of climate models. The diagnosis and uncertainty analysis through our proposed POD package to be implemented in MDTF will provide valuable information and guidance for improving model processes that are key to modeling drought and fire hazard conditions. The proposed work directly responds to three priority areas of this mapp. (1) addressing model precipitation biases and how they drive key societal impacts such as droughts and wildfires; (2) advancing the simulation of climate and weather extremes; and (3) benchmarking model processes that undergird the successful prediction of extreme events. Ultimately the POD integrated within the MDTF package will support model development and optimization designed to improve the representation of drought and wildfire hazard in climate model simulations, which are communicated to the general public and policy makers. This work will contribute to noaas long-term goal of building a climate ready nation by promoting public understanding of climate science and climate-related events, providing climate-related support to the private sector and the nations economy, and serving people making climate-related decisions.