Land Surface Process Model Research Articles

Multi-year simulations and experimental seasonal predictions for rainy seasons in China by using a nested regional climate model (RegCM_NCC). Part I: Sensitivity study

A modified version of the NCAR/RegCM2 has been developed at the National Climate Center (NCC), China Meteorological Administration, through a series of sensitivity experiments and multi-year simulations and hindcasts, with a special emphasis on the adequate choice of physical parameterization schemes suitable for the East Asian monsoon climate. This regional climate model is nested with the NCC/IAP (Institute of Atmospheric Physics) T63 coupled GCM to make an experimental seasonal prediction for China and East Asia. The four-year (2001 to 2004) prediction results are encouraging. This paper is the first part of a two-part paper, and it mainly describes the sensitivity study of the physical process parameterization represented in the model. The systematic errors produced by the different physical parameterization schemes such as the land surface processes, convective precipitation, cloud-radiation transfer process, boundary layer process and large-scale terrain features have been identified based on multi-year and extreme flooding event simulations. A number of comparative experiments has shown that the mass flux scheme (MFS) and Betts-Miller scheme (BM) for convective precipitation, the LPMI (land surface process model I) and LPMII (land surface process model II) for the land surface process, the CCM3 radiation transfer scheme for cloud-radiation transfer processes, the TKE (turbulent kinetic energy) scheme for the boundary layer processes and the topography treatment schemes for the Tibetan Plateau are suitable for simulations and prediction of the East Asia monsoon climate in rainy seasons. Based on the above sensitivity study, a modified version of the RegCM2 (RegCM_NCC) has been set up for climate simulations and seasonal predictions.

The Hydrological Balance of the October 2000 Flood in Piedmont, Italy: Quantitative Analysis and Simulation

This paper aims to analyze in detail the flood event that occurred in October 2000 over the northwestern Italian regions and to test the capability of mesoscale model RAMS-LSPM in reproducing the main characteristics of the hydrological budget components during the event. On 13-16 October 2000, a deep, quasi-stationary trough hung over the western Mediterranean, causing extreme precipitation over northwestern Italy. Values exceeded 600 mm in four days, causing huge flooding and landslides with significant damage to houses and infrastructures. Also several lives were reported lost in the Italian regions of Piedmont and Aosta Valley. This disastrous flood was associated with heavy rainfall episodes over a relatively small area during the course of a four-day period. The intense rainfall over the Alps was enhanced by the confluence of the surface southeasterly flow and of the upper-level southerly current. This confluence was strongest over the zones where the maxima of rainfall and vertical wind velocity were observed. This event was analyzed in detail, using the observational network of the Piedmont Region, the GTS network, and model analyses. The interest was focused on the detailed distribution of precipitation in time and space, in order to understand the dominant role of complex orography in enhancing and modifying the precipitation distribution, and on the hydrological budget. The mesoscale Regional Atmospheric Modelling System (RAMS), driven by the European Centre for Medium-Range Weather Forecasts analyses, was used to simulate this event. The interest was focused on the hydrologic budget during the flood; thus the original soil water balance scheme of RAMS (a simplified version of BATS, the Biosphere Atmosphere Transfer Scheme) was replaced by the Land Surface Process Model (LSPM). This method, already applied in the analysis of a previous flood that occurred on the same region in 1994, allowed all components of the hydrologic budget to be inferred during the flood episode, particularly the surface runoff and soil moisture content. The results of the simulation demonstrated that the coupled model RAMS-LSPM was able to give a quite realistic representation of the flood. The coupling with LSPM improved RAMS not only because the coupled model could represent the upper- and surface-level fields (geopotential, potential temperature, and specific humidity), but also because it provided quantitative estimates of the hydrologic budget and other surface parameters.

A hybrid inversion method for mapping leaf area index from MODIS data: experiments and application to broadleaf and needleleaf canopies

Leaf area index (LAI) is an important variable needed by various land surface process models. It has been produced operationally from the Moderate Resolution Imaging Spectroradiometer (MODIS) data using a look-up table (LUT) method, but the inversion accuracy still needs significant improvements. We propose an alternative method in this study that integrates both the radiative transfer (RT) simulation and nonparametric regression methods. Two nonparametric regression methods (i.e., the neural network [NN] and the projection pursuit regression [PPR]) were examined. An integrated database was constructed from radiative transfer simulations tuned for two broad biome categories (broadleaf and needleleaf vegetations). A new soil reflectance index (SRI) and analytically simulated leaf optical properties were used in the parameterization process. This algorithm was tested in two sites, one at Maryland, USA, a middle latitude temperate agricultural area, and the other at Canada, a boreal forest site, and LAI was accurately estimated. The derived LAI maps were also compared with those from MODIS science team and ETM+ data. The MODIS standard LAI products were found consistent with our results for broadleaf crops, needleleaf forest, and other cover types, but overestimated broadleaf forest by 2.0–3.0 due to the complex biome types.

The Landsat ETM+ spectral mixing space

Spectral mixture analysis (SMA) uses linear mixture models to provide physical representations of land surface reflectance as continuous fields of spectral endmember abundances. While most spectral mixture analyses are site-specific and use local image endmembers to model abundance of specific materials, the linear mixture model can also provide a physical basis for a more general representation of land surface reflectance as mixtures of generic endmembers. A spectral mixture analysis of a global composite of 30 spectrally diverse Landsat ETM+ subscenes indicates that a wide variety of ETM+ reflectance spectra can be accurately represented as linear combinations of soil and rock substrate, green vegetation, and dark surface reflectance spectra. The global analysis indicates that >98% of ETM+ image spectral variance can be represented within a three-dimensional spectral mixing space and that >90% of the variance can be described with a two-dimensional projection of the mixing space. A wide variety of individual subscene and composite mixing spaces consistently show similar triangular distributions of mixed reflectances bounded by linear mixtures of the substrate, vegetation, and dark surface endmembers. Distinct binary mixing spurs representing ice/snow and submarine/reef reflectance are also linear but separate from the primary triangular mixing space. While the three-endmember substrate, vegetation, dark surface (SVD) linear mixing model does not accommodate some significant characteristics of the three-dimensional mixing space, it is able to represent more than 95% of the 30 million observed ETM+ image spectra with misfits of less than 0.04 reflectance units. Linear binary mixing continua between the dark surface endmember and both the substrate and vegetation endmembers indicate the extent to which shadowing and nonreflective surfaces combine with illuminated substrate and vegetation at subpixel scales to modulate spectrally mixed ETM+ reflectances in a wide variety of land cover types. SVD fractions estimated with a standard set of generic endmember spectra could offer many of the benefits of standard metrics like vegetation indices (e.g. NDVI) and linear transformations (e.g. Kauth-Thomas Tasseled Cap), while retaining the unique benefits of physically based fraction estimates. These benefits include good agreement with ground based measurements and output units corresponding to physical properties of land cover that can be used directly as input to land surface process models.

Land surface influences on the West African summer monsoon: Implications for synoptic disturbances

The study examines regional atmospheric model (RM) simulations of the mean June–September (JJAS) climate and the implications of the mean state for the model representation of African wave disturbances (AWD). Two simulations are made with a version of the RM that computes soil moisture with multivariate functions that statistically relate it to rainfall, surface temperature, albedo, vegetation and terrain slope. These simulations differ according to the assignment of ground temperatures (Tg). While the control experiment is based on the fully interactive computation of soil moisture, the second experiment tests the response to fixed Tg whose seasonal means are more realistic than in the control within a swath along the Gulf of Guinea coast. A third simulation is made with the RM coupled to a sophisticated land surface process model (RM2). Results show a rather acute sensitivity of the mean circulation to land surface processes. The more realistic meridional temperature gradient created by fixing Tg in turn increased the vertical wind shear over West Africa and eliminated unrealistic westerly circulation at 700 mb. AWD composites were transformed from intense closed cyclonic circulations with copious rainfall to more realistic open waves that organized more moderate precipitation maxima. Lower vorticity variances in the specified change experiment imply that the open waves were characterized by more moderate vorticity extremes. Corresponding spectral amplitudes for 3–6 day periodicities of the 700 mb meridional wind were 40–80% of control values within the swath of maximum AWD activity. From among the three simulations, RM2 achieved seasonal mean precipitation, temperature, energy flux and circulation distributions that, despite some unrealistic features, were closest to observational evidence. RM2 AWD were much less intense and favored slightly longer periods. Results demonstrate that discrepancies in modeled ground temperatures caused by underestimating the cloudiness that intercepts short wave solar flux along a narrow swath of the West African coast have far-reaching consequences for the simulation of both the mean summer climate and individual synoptic disturbances. The study implies that this cooling along the Gulf of Guinea coast prevents AWD from developing into more intense storms with heavier precipitation.

COUPLING AN ADVANCED LAND SURFACE PROCESS MODEL WITH A NONHYDROSTATIC ATMOSPHERIC MODEL

In this study we have coupled land surface process model (SiBUC: Simple Biosphere Mode including Urban Canopy) into nonhydrostatic atmospheric model (ARPS: Advanced Regional Prediction System) developed by CAPS (Center for Analysis and Prediction of Storms). Preliminary test simulation was carried out by ARPS-SiBUC coupled model and ARPS within and surrounding area of the Lake Biwa Basin during the intensified observation period of the Lake Biwa Project 2001. We compare the simulation results of the two models paying attention to the accuracy of the land surface flux and the local circulaton within the Lake Biwa Basin.

Offline experiment of land surface processes in Qinghai-Xizang Plateau

The observational data of Beiluhe test of the Qinghai-Xizang railway in June 2002 are used as forcing field of the surface model to study the simulation performance in Qinghai-Xizang Plateau. The simulation result shows that the model (NCAR/LSM) can present the surface character tendency reasonably under observation forcing. In NCAR/LSM, the descriptions of the surface vegetation type in Qinghai-Xizang Plateau have much difference to what is actually happening. According to the actual characteristic of the surface, we define the grassland vegetation coverage rate is 0.6, and the vegetation canopy is 0.15 m, and the displacement height is 0.10 m. Compared the results with the old model’s parameter in experiment, we find that the simulation air (2 m) temperature and the ground temperature are close to those of the observation, and the simulation temperature of peak value is better, and the simulation flux of the ground is sure to be improved too. So, if the surface and vegetation parameters are revised in LSM model, the simulation capability of the land surface process model can be improved effec-tively in Qinghai-Xizang Plateau.

A process‐based model for quantifying the impact of climate change on permafrost thermal regimes

Air temperature at northern high latitudes has increased at a higher rate than the global mean, and most general circulation models project that this pattern will continue. Climate warming can increase summer thaw depth and induce permafrost degradation, which may alter the dynamics and functions of northern ecosystems and the lifestyles of northern residents. To address these issues, we developed a process‐based model to simulate permafrost thermal regimes by combining the strength of existing permafrost models and land surface process models. Soil temperature and active layer thickness were simulated by solving the heat conduction equation, with the upper boundary conditions being determined using the surface energy balance and the lower boundary conditions being defined as the geothermal flux. The model integrated the effects of climate, vegetation, ground features, and hydrological conditions on the basis of energy and water transfer in the soil‐vegetation‐atmosphere system. The model was validated against the measurements at four sites in Canada. The simulation results agreed with the measurements of energy fluxes, snow depth, soil temperature, and thaw depth. These results indicate that this physically based model captured the effects of climate, vegetation, and ground conditions on soil temperature and freezing/thawing dynamics, and the model is suitable to investigate the impacts of transient climate change on soil thermal regimes and permafrost degradation and their consequent effects on ecosystem dynamics.

Methods and examples for remote sensing data assimilation in land surface process modeling

Land surface process models describe the energy, water, carbon, and nutrient fluxes on a local to regional scale using a set of environmental land surface parameters and variables. They need time series of spatially distributed inputs to account for the large spatial and temporal variability of land surface processes. In principle many of these inputs can be derived through remote sensing using both optical and microwave sensors. New approaches in four-dimensional data-assimilation (4DDA) form the basis to combine remote sensing data and spatially explicit land surface process models more effectively. This paper describes basic techniques for 4DDA in land surface process modeling. Two case studies were carried out to demonstrate different successful approaches of remote sensing data assimilation into land surface process models. The assimilation of surface soil moisture estimates from European Remote Sensing (ERS) synthetic aperture radar data in a flood forecasting scheme is presented, as well as the combination of a land surface process model and a radiative transfer model to improve the accuracy of land surface parameter retrieval from optical data [Landsat Thematic Mapper (TM)].

LGM lake records from China and an analysis of climate dynamics using a modelling approach

Lake-geological studies in China have reported that there were much higher lake levels and much fresher water than today at the last glacial maximum (LGM) in western China. A compilation of lake data in this study showed LGM conditions much drier than today in eastern China but somewhat wetter in western China. These E–W differential patterns of climate conditions were completely different from the modern dry-wet conditions with a N–S differential distribution. In this study palaeoclimate simulations by an AGCM coupled with land surface process model were used to explore the possible mechanisms of LGM climate in China. The results confirmed that the dry conditions in eastern China resulted from less summer precipitation due to the Pacific Subtropical High occupying eastern China and the decline in the summer monsoon. The wet conditions in western China were produced by a decrease in evaporation due to a low temperature on land surface at the LGM and increase in precipitation. Two experiments of the palaeoclimate simulations with different land surface of modern and palaeo-vegetations have been designed to test the discrepancies of simulated LGM climate with in precipitation and P–E fields. The results suggested that the feedback from the Asian land surface within the climate system would amplify and modify external forcing, leading to marked climate changes in China.

Automated Parameterization of Land Surface Process Models Using Fuzzy Logic

All land surface process models require parameters that are proxies for spatial processes that are impractical or impossible to measure. Recent developments in model parameter estimation theory suggest that information obtained from calibrating such models is inherently uncertain in nature. As a consequence, identification of optimum parameter values is often highly non–specific. A calibration framework using fuzzy logic is presented to deal with such uncertain information. An application of this technique to calibrate the sub–canopy controls on transpiration in a land surface process model demonstrates that objective estimates of parameter values and expected ranges of predictions can be obtained with suitable choices for objective functions. An iterative refinement in parameter estimates was possible with conditional sampling techniques. The automated approach was able to correctly identify parameter tradeoffs such that two strongly different sets of parameters could

Remote sensing data assimilation using coupled radiative transfer models

This paper discusses data assimilation of biophysical parameters retrieved from optical remote sensing images in land surface process models by means of image simulation and model inversion. Two different approaches are presented. The first is based on model inversion of atmospherically corrected Landsat TM surface reflectance images and assimilation of the retrieved parameters in a crop growth model. In the second approach top-of-atmosphere (TOA) hyperspectral radiance images have been simulated for the future ESA mission SPECTRA. In this case only the simulation of the images has been executed in order to demonstrate the feasibility of this task with existing software running on a PC. The radiative transfer models that have been used are PROSPECT (leaf level), GeoSAIL (canopy level) and MODTRAN4 (atmosphere). Coupling of this chain of models to land use information of the area can be used to generate TOA radiance images. Comparison of simulated images with actual remote sensing data can be applied to retrieve biophysical parameters and in turn these can be employed to update process models of crop growth.

Assimilating remote sensing data into a land-surface process model

A PROcess-oriented Modular Environment and Vegetation model (PROMET-V) was developed to calculate plant growth, water and nitrogen fluxes, their spatial patterns and interactions. A central goal of PROMET-V was to develop interfaces to use remote sensing data for spatially distributed modelling. Remote sensing measurements are used by PROMET-V to initialize the model, to adjust and update model parameters, and to validate model results. Model results show that management related parameters, such as the cutting date of meadows or the sowing density of crops, are largely responsible for spatial heterogeneities. These impacts cannot be modelled accurately. They must be observed instead. The same is true for pests, diseases or catastrophic events, e.g. flood, fire, storm damage. Thematic Mapper (TM) images were used to update model parameters and to estimate management related parameters. The use of remote sensing data not only leads to improved model results, it also allows the detection of spatial heterogeneities which otherwise cannot be accounted for. Remote sensing data were also used to validate model results. The calculated surface soil moisture compared favourably to soil moisture patterns derived from European Remote Sensing Synthetic Aperture Radar (ERS SAR) data.

Thermal features of the Mackenzie basin from NOAA AVHRR observations for summer 1994

A series of mid‐afternoon Advanced Very High Resolution Radiometer (AVHRR) thermal radiance scenes were assembled in order to develop a better understanding of the complex energy and water processes leading to variations in surface temperature. An in‐depth knowledge of the temperature variability is of interest to land surface process modelling and its application to the Mackenzie Global Energy and Water Cycle Experiment (GEWEX) Study (MAGS). Clear‐sky land surface temperatures are estimated by applying a split window technique to remove atmospheric effects. A maximum land surface temperature map of the Mackenzie basin at 1‐km scale for summer 1994 is produced. The patterns are related to land surface features and elevation. The basin's maximum land surface temperature patterns can be subdivided into three land zones (≥ 35°C, 33–34°C and 27–32°C) and a water dominated zone (20.5°C on average). The highest maximum temperature zone (≥35°C) corresponds to a combination of minimal vegetation, drier soils and low terrain. This zone is not in the southern part of the basin as might be speculated in the absence of these data, but in a wide low elevation corridor from west of Great Bear Lake along the Mackenzie River down to 50°N, 120°W. The maximum land surface temperatures tend to decrease with increasing vegetation density and surface moisture; they also decrease with elevation at a rate of –4.5°C km–1. This is confirmed by weather station data. The AVHRR data extend this relationship to the 1200 – 2200 m altitude ranges, where there are no station data. The data suggest that elevation and land cover should be taken into account in the objective analysis (spatial interpolation) of station data.

The flood of November 1994 in Piedmont, Italy: a quantitative analysis and simulation

AbstractThe disastrous flood that occurred in the southern area of Piedmont (north‐western Italy) at the beginning of November 1994 was associated with heavy rainfall episodes over a relatively small area in a course of events spanning the period 3–6 November 1994. The most intense rainfall (for which values over 200 mm were reported) occurred in two regions located in the upper reach of the Tanaro and Bormida river basins (Maritime Alps) and over the north‐west sector of the Alps. The distribution of rainfall intensities was different in the two areas where the precipitation peaks were observed. In the former, almost all rainfall was concentrated between the 4 and 5 November; in the second, rainfall was intense but spread over the entire period. The intense rainfall over these two mountain areas was enhanced by the confluence of the surface south‐easterly flow and of the upper‐level southerly current. This confluence was strongest over the two zones in which the maxima of rainfall and vertical wind velocity were observed. The cyclonic area was stationary over Western Europe for three consecutive days (4–6 November), due to the blocking situation over Eastern Europe. The mesoscale Regional Atmospheric Modelling System (RAMS), driven by the European Centre for Medium‐Range Weather Forecasts analyses, was used to simulate this event. We focused our attention on the target area (Tanaro and Bormida valleys). As we were interested in the hydrologic budget during the flood, we replaced the original soil scheme of RAMS (a version of Biosphere Atmosphere Transfer Scheme (BATS), too simplified to allow very detailed pattern resolution) with the Land Surface Process Model (LSPM). This method allowed all components of the hydrologic budget to be inferred during the flood episode, particularly the runoff and soil moisture content. The results of the simulation demonstrated that the coupled model RAMS–LSPM was able to give a quite realistic representation of a mesoscale event such as the flood event in this study. The coupling with LSPM improved RAMS not only because the coupled model can represent the upper‐ and surface‐level fields (geopotential, potential temperature and specific humidity), but also because it provides quantitative estimates of the hydrologic budget and other surface parameters. Copyright © 2002 John Wiley & Sons, Ltd.

Impact of soil surface moisture initialization on rainfall in a limited area model: a case study of the 1995 South Ticino flash flood

AbstractThe influence of the soil moisture initialization on predicting rainfall in a limited‐area model (LAM) was analysed. The model selected for the simulation was the LAM of Bologna (LAMBO). The case study selected in this paper was the flash flood episode that occurred during September 1995 in the South Ticino area (Switzerland). Two different experiments were conducted. In the first experiment, all initial fields including the initial soil moisture were provided by European Centre for Medium‐range Weather Forecast (ECMWF) analyses. In the second one, the initial soil moisture fields were calculated by running a soil–vegetation–atmosphere transfer scheme (the Land Surface Process Model—LSPM) in stand‐alone mode. The LSPM was driven by the synoptic observations carried out on a mesoscale area that includes Northern Italy. In order to overcome the well‐known problem of the initial soil moisture content, LSPM was run for a long period (5 months) preceding the flood, in order to reach a stability state. The soil moisture field calculated by LSPM is more structured than the ECMWF one, and the numerical values of soil moisture are generally lower than those of ECMWF. The results seemed to indicate that the precipitation field obtained for the flood episode in the second experiment is closer to the observations than in the control run, even if there were still some underestimations. The precipitation peak modelled on South Ticino was closer to the observations, and this result seems to confirm the importance of the correct initialization of the soil moisture field even in cases of short‐range forecasting. Copyright © 2002 John Wiley & Sons, Ltd.

Observation and study of land surface parameters over Gobi in typical arid region

According to the need of popular land surface process models, characteristics and rules of some key land surface process and soil parameters over Gobi in typical arid region of Northwest China are analyzed by using the data observed during the intensive observation period of the Dunhuang Land–Surface Process Field Experiment (DLSPFE) (May–June 2000). Using the relative reflection as weighting factor, the weighted mean of the surface albedo over Dunhuang Gobi in typical arid region is calculated and its values are 0.255 ± 0.021. After canceling the interference of the buildings, the mean values of the roughness length averaged with logarithm is 0.0019 ± 0.00071 m. After removing the influence of the oasis, the soil wetness factor computed with data under condition of no precipitation is 0.0045. After removing the influence of the precipitation , the mean values of the soil heat capacity over Dunhuang Gobi in typical arid region is 1.12 × 106 J m−3K−1, a bit smaller than the values observed in HEIFE. But the soil heat diffusivity and conductivity are about one of those observed in HEIFE. The soil water content over Dunhuang Gobi in typical synoptic condition is very little and does not exceed 1% basically.

Energy and Water Balance at Soil-Air Interface in a Sahelian Region

The aim of this work is an improvement of the parameterization of the soil moisture in the scheme of the Land Surface Process Model (LSPM) for applications over desert areas. In fact, in very dry conditions, the water vapour flux plays an important role in the evaporation processes and influences the underground profiles of humidity and temperature. The improved version of soil moisture parameterization in the LSPM scheme has been checked by using the data taken from the database of the field experiment HAPEX-Sahel (Hydrology-Atmosphere Pilot Experiment in the Sahel. 1990-1992). Model simulations refer to three different stations located in Niger (Fallow, Millet and Tiger sites) where input data for LSPM and observations were simultaneously available. The results of simulations, taking into account the water vapour flux in the soil model LSPM, seem to compare better with the observed behaviour of soil moisture and turbulent heat fluxes than those overlooking the water vapour flux, confirming the great importance of the water vapour in such dry conditions.

Simulations of a Boreal Grassland Hydrology at Valdai, Russia: PILPS Phase 2(d)

The Project for the Intercomparison of Land-Surface Parameterization Schemes (PILPS) aims to improve understanding and modeling of land surface processes. PILPS phase 2(d) uses a set of meteorological and hydrological data spanning 18 yr (1966‐83) from a grassland catchment at the Valdai water-balance research site in Russia. A suite of stand-alone simulations is performed by 21 land surface schemes (LSSs) to explore the LSSs’ sensitivity to downward longwave radiative forcing, timescales of simulated hydrologic variability, and biases resulting from single-year simulations that use recursive spinup. These simulations are the first in PILPS to investigate the performance of LSSs at a site with a well-defined seasonal snow cover and frozen soil. Considerable model scatter for the control simulations exists. However, nearly all the LSS scatter in simulated root-zone soil moisture is contained within the spatial variability observed inside the catchment. In addition, all models show a considerable sensitivity to longwave forcing for the simulation of the snowpack, which during the spring melt affects runoff, meltwater infiltration, and subsequent evapotranspiration. A greater sensitivity of the ablation, compared to the accumulation, of the winter snowpack to the choice of snow parameterization is found. Sensitivity simulations starting at prescribed conditions with no spinup demonstrate that the treatment of frozen soil (moisture) processes can affect the long-term variability of the models. The single-year recursive runs show large biases, compared to the corresponding year of the control run, that can persist through the entire year and underscore the importance of performing multiyear simulations.

The interannual variability and predictability in a global climate model

The interannual variabilities of the climatological simulation (V1) and the AMIP (Atmospheric Model Intercomparison Project) simulation (V2) by the IAP 9-Level Atmospheric General Circulation Model are studied and discussed in this paper. Based on the analysis of ratio of variability (R) of above two simulations the predictability of the model on the interannual climate variation are studied as well. Results show that V2 is bigger than V1 generally and V2 is more comparable to the real variability of the atmosphere, the major difference of V1 and V2 is in the tropics, for temperature and geopotential height the predictability is higher in the tropics while in the extra-tropics there is almost no predictability and the predictability is bigger in higher level than in lower level. The predictability for precipitation is generally low in the globe, and generally the predictability is high in the tropical eastern Pacific for the lower level. This study suggests that the possible way of increasing the model predictability is the improvement of land surface process modelling and the inclusion of the interannual variations of the land surface conditions (snow cover, albedo, soil moisture, etc.) as the forcing factor for climate modelling and prediction.