Assessment of the benefits of climate model weights for ensemble analysis in three urban precipitation frequency studies

摘要

In hydrology, projected climate change impact assessment studies typically rely onensembles of downscaled climate model outputs. Due to large modeling uncertainties,the ensembles are often averaged to provide a basis for studying the effects ofclimate change. A key issue when analyzing averages of a climate model ensemble iswhether to weight all models in the ensemble equally, often referred to as the equal-weightsor unweighted approach, or to use a weighted approach, where, in general,each model would have a different weight. Many studies have advocated for the latter,based on the assumption that models that are better at simulating the past, that is, themodels with higher hindcast accuracy, will give more accurate forecasts for the futureand thus should receive higher weights. To examine this issue, observed and modeleddaily precipitation frequency (PF) estimates for three urban areas in the UnitedStates, namely Boston, Massachusetts; Houston, Texas; and Chicago, Illinois, wereanalyzed. The comparison used the raw output of 24 Coupled Model IntercomparisonProject Phase 5 (CMIP5) models. The PFs from these models were compared withthe observed PFs for a specific historical training period to determine model weightsfor each area. The unweighted and weighted averaged model PFs from a more recenttesting period were then compared with their corresponding observed PFs to determineif weights improved the estimates. These comparisons indeed showed that theweighted averages were closer to the observed values than the unweighted averagesin nearly all cases. The study also demonstrated how weights can help reduce modelspread in future climate projections by comparing the unweighted and weighted ensemblestandard deviations in these projections. In all studied scenarios, the weightsactually reduced the standard deviations compared to the equal-weightsapproach.Finally, an analysis of the results' sensitivity to the areal reduction factor used to allowcomparisons between point station measurements and grid-boxaverages is provided.