Climate model projections for future seasonal rainfall cycle statistics in Northwest Costa Rica

Abstract

Increasing water scarcity due to rising demand and changes in climate and land use are expected to exert significant stress on water resources in many parts of the world. In many areas, distinctive patterns of seasonal precipitation play an important role in regional ecosystems, economies, and food and energy supplies. Given the inconsistency among climate model precipitation projections, this study develops a statistical approach to better characterize seasonal patterns of precipitation and to assess the potential impact of climate change on these patterns. The method is applied to evaluate the bimodal seasonal pattern of precipitation in northwest Costa Rica as a case study. A Gaussian mixture model is used to describe the bimodal pattern and quantify changes in the seasonal precipitation cycle projected by 19 Coupled General Circulation Models. The model simulations for the current period (1979–2005) are compared with observed monthly precipitation data based on four goodness‐of‐fit metrics to select the best performing models. The monthly bias‐corrected and spatially disaggregated (BCSD) climate projections from the selected models are used to investigate the projected change in the bimodal seasonal pattern, seasonal mean precipitation and interannual variability at the late twenty‐first century. Under RCP8.5, the degree of consistency associated with these BCSD climate projections is found to be higher for the projections of the midsummer drought than that for the early portion of the wet season (early season). For the late portion of the wet season (late season) projections, the degree of consistency is found to be the lowest. The proposed method provides (a) an overall characterization of inconsistencies in climate model projections for the region, (b) more information about the possible changes in the seasonal cycle of precipitation, and (c) the possibility of performing detailed comparison tests among climate models over a set of simulations and projections.