Seasonal Prediction with CCSM3.0: Impact of Atmosphere and Land Surface Initialization.
Jim Kinter1Dan PaolinoDavid Straus1Ben Kirtman2Dughong Min2Center for Ocean-Land-Atmosphere Studies1 also George Mason University thanks to NCAR CISL for2 University of Miami computing resources
Shukla and Mintz, 1982
Influence of Land-Surface Evapotranspiration on the Earth's Climate. Science, 215, 1498-1501.
GLACE showed that coupling between land and atmosphere is strongest in transitional zones between humid and arid regions (and lots of inter-model variance!).
Koster et al., 2004
Koster, R. D., P. A. Dirmeyer, Z. Guo, G. Bonan, E. Chan, P. Cox, H. Davies, T. Gordon, S. Kanae, E. Kowalczyk, D. Lawrence, P. Liu, S. Lu, S. Malyshev, B. McAvaney, K. Mitchell, T. Oki, K. Oleson, A. Pitman, Y. Sud, C. Taylor, D. Verseghy, R. Vasic, Y. Xue, and T. Yamada, 2004: Regions of strong coupling between soil moisture and precipitation. Science, 305, 1138-1140.
* Evaporation variability times a land-atmospheric flux function based on the tightness of the dependence of surface fluxes on soil moisture
Soil Moisture Memory Land Surface Initialization
GSWP-2 results from multiple models provide quantitative information about the effect of soil moisture on predictability. The season (color) and duration (intensity) of the maximum soil moisture memory is shown.
Series of papers by Guo and Dirmeyer; Guo et al.; Seneviratne et al.
GLACE2 - Forecast Correlation Land Surface Initialization
100 initial times: 10 years (1986-1995) X 5 months (Apr.-Aug.) X 2 days (the 1st and 15th)
10-member, 2-month COLA AGCM runs with observed SST
Correlations for CONUS region average: 70-125W, 22-50N
______ realistic land ICs runs ______ random land ICs runs ---------- 95% significance level
Courtesy of Zhichang Guo
Model Land Surface Initialization: CCSM3.0 is a coupled ice-ocean-atmosphere-land climate model with state-of-the-art formulations of dynamics and subgrid-scale physical parameterizations. The atmosphere is CAM3 (Eulerian dynamical core) at T85 (~150 km) horizontal resolution with 26 vertical levels. The ocean is POP with 1 degree resolution, stretched to 1/3 degree near the equator. Re-forecast Experiments: Retrospective forecasts cover the period 1982–1998 for the July initial state experiments, and 1981-2000 for the January initial state experiments. Ensembles of 6 (10) hindcasts were run in the OCN-only (ATM-OCN-LND) experiments (see below). Ocean Initialization: The ocean initialization uses the GFDL ocean data assimilation system, based on the MOM3 global ocean model using a variational optimal interpolation scheme. The GFDL ocean initial states were interpolated (horizontally and vertically) to the POP grid using a bi-linear interpolation scheme. (Climatological data from long simulations of CCSM3 were used poleward of 65°N and 75°S.) The ocean initial state is identical for each ensemble member.
EXPERIMENTSOne-year re-forecast ensembles with CCSM3.0 Initial states: 1 January and 1 July for 1981-2000Two sets of re-forecasts
The atmospheric and land surface initial states were taken from an extended atmosphere/land-only (CAM3) simulation with observed, prescribed SST. The atmospheric ensemble members were obtained by resetting the model calendar back one week and integrating the model forward one week with prescribed observed SST.
In this way, it is possible to generate initial conditions that are synoptically independent (separated by one week) but have the same initial date.
Thus all ensemble members were initialized at the same model clock time (1 Jan or 1 July) with independent atmospheric initial conditions.
Land/Atmosphere Initial Conditions in Two Sets of ExperimentsATM-OCN-LND ExperimentLand and atmosphere were initialized for each of the 10 days preceding the date of each ocean initial state * 22-31 December for the 1 January ocean states * 22-30 June for the 1 July ocean datesAtmosphere initialized by interpolating from daily Reanalysis. Land surface initialized from daily GSWP (1986-1995) and daily ERA40 (1982-1985 and 1996-1998). Observed anomalies superimposed on Common Land Model (CLM) climatology. Snow cover initialized from ERA40. Sea-ice initialized to climatological monthly condition based on a long simulation of CCSM3.0.
CCSM Performance - Predicting ENSO Experiments
Time-longitude cross-sections of equatorial Pacific SST anomaly
CCSM Re-Forecasts with Land ICs Experiments
(Top) Soil Moisture Prediction Skill Experiments
top 9 cm
top 7 cm
top 9 cm
top 7 cm
July (1-month lead) Soil Moisture (top level) Prediction Skill
(Mid) Soil Moisture Prediction Skill Experiments
July (1-month lead) Soil Moisture (mid-level) Prediction Skill
Model Drydown Experiments
7-month lead (Jan ICs) vs. 1-month lead (Jul ICs) - percent difference
Global Surface Air Temperature Forecasts Experiments
JAN: 1-month lead
FEB: 2-month lead
January initial conditions
Global Precipitation Forecasts for July Experiments
Simultaneous correlation for July initial conditions, 1-month lead forecasts and CMAP
Indian Monsoon Rainfall Experiments
The JAS mean precipitation in south Asia.
The simulated interannual variability of JAS rain over land (not shown) is much smaller than observed.
Nordeste Brazil Forecast Experiments
1 January ICs
Leading EOF of JAS mean 850 hPa rotational winds, 1982-1998.