Inter-comparison of multiple downscaled climate datasets for the Pacific Northwest

Abstract:

Climate change is already impacting forest ecosystems in the Pacific Northwest region. To assess these impacts and forest responses, gridded data from global climate models (GCM) outputs are used to drive ecosystem model simulations. However, the spatial resolutions of outputs from currently available GCMs are too coarse to be used in many impact assessments at regional or landscape scales. Therefore, climatologists have downscaled GCM projections based on different methods (e.g., statistical and dynamical models). These downscaled climate data are being used as inputs to impact models, such as forest dynamics, crop growth, and hydrological models. However, without comparing differences and quantifying the uncertainties among the downscaled driving datasets, it is difficult for ecosystem modelers and resources managers to select the most appropriate dataset for their specific interests. Our study is aimed at evaluating the performance of different downscaled datasets in representing local and regional climates. We are also investigating the uncertainties among different data sources, especially for projected future periods. We analyzed and contrasted five recently developed downscaled dataset: NEX-DCP30, WorldClim, MACAv2-METDATA, MACAv2-LIVNEH, and ClimateNA. We evaluate the biases in each downscaled dataset by comparing it with a gridded observation-based dataset, such as PRISM (Parameter-elevation Relationships on Independent Slopes Model) for the historical period. In this work, we aim to provide recommendations for scientists and natural resource managers to enhance science-based climate change adaptive management across the Pacific Northwest.