Evaluation of CMIP6 precipitation simulations during the last century along the Subtropical Central Andes

JUAN ANTONIO RIVERA; ELIZABETH NARANJO TAMAYO; JULIÁN VILLAMAYOR; MYRIAM KHODRI; GABRIEL ARNOULD; MAXIMILIANO VIALE; VALERIE DAUX

San Francisco

Conferencia; AGU Fall Meeting 2019; 2019

American Geophysical Union

The Subtropical Central Andes (SCA) correspond to a complex orographic region located between 25° and 45°S. The Andes cordillera largely impacts the local climate, especially in winter, when frontal systems and moisture from the Pacific Ocean interact with the topography. An increasingly number of studies have reported a significant decrease of precipitations in the SCA, particularly marked along Central Chile and Northern Patagonia, in relation with the southward displacement of the descending branch of the Hadley cell and other interannual factors such as El Niño-Southern Oscillation. The SCA is a high populated area where economy, mainly based on agriculture and hydropower's exploitation, depends on water resources originated along the Andes. Nowadays, a mega-drought event is taking place over the SCA; thus, future projections of precipitation are mandatory for the development of adaptation strategies and the improvement of water management.This study aims to evaluate the performance of the available CMIP6 Global Climate Models (GCM) precipitation outputs over the SCA during the last century against observations from the Global Precipitation Climatology Centre and local rain gauges. This will allow to select the models that best represent current precipitation characteristics, and can be suitable for future climate investigation over the region. Spatio-temporal patterns of precipitation for the period 1895-2014 are adequately captured by the majority of the GCMs, considering both winter and summer patterns, as verified through Taylor diagrams and climatology maps. Observations show that the long-term decreasing trends are located mainly south of 36°S along the Chilean territory, while some GCMs extend this area towards the north, with trend values large as -3 mm/year. The distribution of precipitation totals was assessed through a binned score based on several precipitation categories, showing the existence of both wet and dry biases that are model dependent. Based on these results, a subset of models that best represent the precipitation characteristics will be proposed for the assessment of future precipitation trends.