Intraseasonal to Interannual Variability in the Intra-Americas Sea in Climate Models

This study assesses the ability of CMIP models to produce realistic intraseasonal to interannual variability (IAV to ISV) in the Atlantic warm pool (AWP) region and the implications for hurricanes, the ability of parameterization modifications in the GFDL AM3 to improve the simulation of AWP ISV, and how mean state biases in CMIP models develop and the implications for forecast biases in ISV and IAV. The following questions will be answered:

1. Can IAS-region intraseasonal variability in the GFDL AM3/CM3 be improved through modifications to the treatment of deep convection?

The sensitivity of IAS mean state and ISV to modifications in the Donner convection scheme in GFDL AM3/CM3 will be assessed, including different treatments of triggering, rain evaporation, and entrainment. The degree to which IASregion ISV is coherent with that in the Eastern Hemisphere will be assessed, which has consequences for prediction. Variables that impact hurricane genesis potential will be a focus.

2. How do model errors develop over the Atlantic warm pool?

In the AWP, the ensemble mean of CMIP3 models features SST errors of 2oC or larger in the annual mean, with considerable variability in rainfall errors among different atmosphere models forced by observed SSTs. A systematic investigation into errors of SST, rainfall, sea level pressure and wind in the AWP based on the CMIP3/5 database will be conducted, including the similarities and differences among models. Initial emphasis will be GFDL models, especially initialized seasonal forecasts, followed by a diagnosis in the broader suite of CMIP5 models.

3. How well can CMIP5 models simulate the ENSO-Atlantic hurricane teleconnection?

Substantial biases in the ENSO-Atlantic hurricane teleconnection occur in all CMIP3 models (Shaman and Maloney 2011). We will assess the ability of CMIP5 models to capture the ENSO teleconnection to the Atlantic and its manifestation in large-scale variables that affect tropical cyclone (TC) genesis, with specific focus on the GFDL CM3. We will also intercompare CMIP5 model ability to capture other modes of Atlantic IAV including the Atlantic meridional mode and Atlantic Multidecadal Oscillation, and the variables relevant for TCs.

4. How do IAS-region mean state biases affect forecasts of ISV and IAV and extreme events?

The climatology of a coupled prediction model drifts quickly, and the model errors approach the equilibrium in a matter of months. Biases in the mean state, such as those in SST, precipitation, and winds can have profound implications for Atlantic climate variability and how remote forcing from climate variability in other basins is manifest in the Atlantic. The effect of the climate drift on forecast results within the IAS will be assessed. The regional climate model from U. Hawaii will be used to examine how mean state biases affect biases in ISV and IAV and extreme events over the IAS region. ISV to IAV in boundary conditions will be retained while biasing the mean state to that of the GFDL CM and other CMIP models to determine how changes in the climate state and its statistics affect the simulation of extreme events like TCs.

This proposal directly addresses MAPP Priority Area Three. We intend a comprehensive study to document the ability of CMIP5 climate models including GFDL CM/AM to accurately simulate the ISV to IAV of the IAS region and associated TC activity. Parameterization modifications to the GFDL AM3 and their ability to improve ISV in the IAS region will be assessed. A close look into model biases in the IAS to identify their sources is an important step toward improving seasonal forecasts of extremes over the Americas. This proposal also supports NOAA’s NGSP by improving understanding of model biases that will allow more accurate predictions of future climate, allowing society to better anticipate and respond to the challenges of climate change. This proposal entails research that advances the nation’s core capabilities in understanding and modeling the climate system, which is a primary goal of the CPO.