Skip Navigation Links www.nws.noaa.gov 
NOAA logo - Click to go to the NOAA home page National Weather Service   NWS logo - Click to go to the NWS home page
Climate Prediction Center

 
HOME > Outreach > Meetings > 33rd Annual Climate Diagnostics & Prediction Workshop > Abstracts
 

Climate Prediction: ENSO, MJO and Teleconnections
Abstract

 

Abstract Author: Rongqian Yang, Kenneth Mitchell, and Jesse Meng

Abstract Title: Winter Season Forecast Experiments with the NCEP Coupled Forecast System (CFS) Using Different Land Models and Different Initial Land States

Abstract: It is well known that during the N.H. winter season, the influence on atmospheric circulation and global precipitation of ENSO-related tropical Pacific SST anomalies is relatively stronger than during the N.H. summer season, hence seasonal predictions by coupled global climate models manifests higher skill in the N.H. in winter than in summer, especially over land. Research over the past decade has demonstrated that proper treatment of land surface processes and land state initialization (especially anomalies of soil moisture and snow cover) is likely the next most important physical attribute (after SST influence) on successful seasonal predictions by coupled global prediction models.

In this study, we use a new experimental version of the NCEP global Coupled Forecast System (CFS) to examine to what extent the choice of land surface model and source of land surface initial conditions impacts the CFS N.H. winter season predictions. Hence this new work presents the winter season counterpart to the summer season results we reported on at an earlier conference. The CFS experiments consist of four CFS configurations (all at T126 resolution) comprising various combinations of two different land models and two different sources of initial land states. Additionally, we also study the impact of the number of years executed for establishing model climatology on evaluating model performance.

Three of the four CFS configurations consist of the modern Noah Land Surface Model (Noah LSM) initialized from three different sources of initial land states: 1) NCEP-DOE Global Reanalysis 2 (GR2), 2) Noah-based Global Land Data Assimilation System (GLDAS-Noah), and 3) climatology of GLDAS-Noah. The fourth CFS configuration uses the older OSU Land Surface Model (OSU LSM) with GR2 initial land states. The land model utilized in the GR2 is the OSU LSM. As a reference for performance, we also evaluate the N.H. winter forecasts of the current operational CFS, which uses the older OSU LSM with GR2 initial conditions, but executes at lower resolution (T62) and uses older physics in other components of the CFS. Winter forecast experiments have been carried out for 10 chosen years. For each year an ensemble of 10 CFS members is executed, whose initial starting dates are from November 29 to December 3 and from December 19 to December 23. The 10 years are comprised of El Nino years, La Nina years, and relatively ENSO-neutral years. We examine the impact in the CFS of using a different land surface model and different source of initial land states on N. H. winter seasonal predictions of SST, precipitation, 2-meter temperature, 200 mb and 500 mb heights, and on land surface characteristics including latent heat and sensible heat fluxes, among others. CFS N.H. winter forecasts were also executed for an additional 14 years (24 years total) for two of the four configurations (Noah LSM from GLDAS-Noah land states and OSU LSM from GR2-OSU land states) to determine the influence of the number of sampled years on the assessment of model performance.

Results from our experiments indicate that different sources of initial land states have a non-negligible impact on CFS winter season predictions. Specifically, our results thus far show that the best CFS performance is realized by means of execution of a companion global land data assimilation system (GLDAS) with the very same new land model as utilized in the land-component upgrade of the global climate model. Providing initial land states that are self-consistent with the chosen land surface model of the coupled global prediction model is important to seasonal predictions, while improper initialization of the land surface component of the global model can degrade the model performance. Thus it is naive to merely upgrade the land component of a global climate model for seasonal forecasting without simultaneously upgrading the land component of the companion global data assimilation system. Finally, assessment of global model seasonal forecasts over an insufficient number of sample years can lead to an incorrect impression of model performance, indicating that for objective assessment of seasonal forecast skill of a global coupled forecast model, a relatively large number of sampled years is needed (of order 25-30 years minimum).


NOAA/ National Weather Service
NOAA Center for Weather and Climate Prediction
Climate Prediction Center
5830 University Research Court
College Park, Maryland 20740
Page Author: Climate Prediction Center Internet Team
Page last modified: July 31, 2008
Disclaimer
Information Quality
Credits
Glossary
Privacy Policy
Freedom of Information Act (FOIA)
About Us
Career Opportunities