Open AccessSubseasonal prediction performance for South American land–atmosphere coupling in extended austral summer
Author(s) -
Chevuturi Amulya,
Klingaman Nicholas P.,
Guo Liang,
Holloway Christopher E.,
Guimarães Bruno S.,
Coelho Caio A. S.,
Kubota Paulo Y.,
Young Matthew,
Black Emily,
Baker Jessica C.A.,
Vidale Pier Luigi
Publication year - 2022
Publication title -
climate resilience and sustainability
Language(s) - English
Resource type - Journals
ISSN - 2692-4587
DOI - 10.1002/cli2.28
Subject(s) - environmental science , predictability , evapotranspiration , precipitation , climatology , atmosphere (unit) , water content , moisture , subtropics , atmospheric sciences , meteorology , geography , geology , ecology , physics , geotechnical engineering , quantum mechanics , fishery , biology
Land–atmosphere feedbacks, through water and energy exchanges, provide subseasonal‐to‐seasonal predictability of the hydrological cycle. We analyse subseasonal land–atmosphere coupling over South America (SA) during extended austral summer for the soil moisture‐to‐precipitation and soil moisture‐to‐air temperature feedback pathways. We evaluate subseasonal hindcasts from global forecasting systems from the UK Met Office, the National Centers for Environmental Prediction (NCEP), the European Centre for Medium Range Weather Forecasts and the Center for Weather Forecast and Climate Studies (CPTEC), for the common period of 1999–2010, against two reanalyses. Biases in land–atmosphere states are established in the first week of hindcasts and increase with lead time. By Week 5, all the models only demonstrate good performance over northern, northeastern and southeastern SA for soil moisture and evapotranspiration and over tropical and subtropical SA for temperature. The hindcasts show stronger coupling at longer lead–lag between variables than reanalyses. Our results highlight possible deficiencies in feedbacks between soil moisture and precipitation in CPTEC and NCEP forecasts over the Amazon due to initial dry soil moisture biases, and in feedbacks between soil moisture and temperature for all four investigated models over southeastern SA due to erroneous representations of evapotranspiration.