Global Forecast Model Research Articles

Relating Anomaly Correlation to Lead Time: Principal Component Analysis of NMME Forecasts of Summer Precipitation in China

AbstractThe skill of global climate model (GCM) forecasts is usually indicated by the anomaly correlation between ensemble mean and observation. For GCM forecasts, anomaly correlation does not steadily improve with decreasing lead time but oscillates instead. This paper aims to address the oscillation and illustrate the relationship between anomaly correlation and lead time. We formulate the anomaly correlation of forecasts at different initialization times as a vector and pool anomaly correlation vectors across grid cells in the analysis. We propose two patterns to characterize the spatial and temporal variation of anomaly correlation in the three‐dimensional space of latitude, longitude, and initialization time. The first pattern suggests that the anomaly correlation at different initialization times is at a similar level. The second pattern indicates that the anomaly correlation linearly increases with decreasing lead time. These two patterns are tested using the eigenvectors through principal component analysis. They are first illustrated using the GFDL‐CM2p1‐aer04 forecasts of summer precipitation in China. They are further verified by another nine sets of North‐American Multi‐Model Ensemble (NMME) forecasts. Overall, the first pattern explains more variation than the second pattern. In total, the two patterns explain 42% of the variation of anomaly correlation for CanCM3, 59% for CanCM4, 42% for COLA‐RSMAS‐CCSM3), 45% for COLA‐RSMAS‐CCSM4, 59% for GFDL‐CM2p1, 67% for GFDL‐CM2p1‐aer04, 65% for GFDL‐CM2p5‐FLOR‐A06, 57% for GFDL‐CM2p5‐FLOR‐B01, 48% for NCAR‐CESM1, and 60% for NCEP‐CFSv2. The percentage of explained variation demonstrates the effectiveness of the two patterns as exploratory tools to analyze the predictive performance of GCM forecasts.

Effects of Non-orographic Gravity Wave Drag on Seasonal and Medium-range Predictions in a Global Forecast Model

This study implements the parameterizations of convective and frontal gravity wave drag (GWD) with wide phase speed spectra into a global forecast model with a model top near 0.3 hPa. The new convective GWD scheme replaces the existing one that considers only a stationary convective GW, and the frontal GWD scheme is newly introduced. When the new GWD schemes are used, the Rayleigh friction, applied above 2 hPa to mimic the effects of missing GWD, is removed. The convective (frontal) GWs are generated mainly in the Intertropical Convergence Zone and winter extratropical storm track regions (extratropics where strong baroclinicity exists). The convective and frontal GWD derived from the new schemes are significant near the model top, with maxima of ~2-4 and ~26-58 m s−1 day−1, respectively. The differences in convective GWD between the stationary and non-stationary schemes appear mainly in the tropics and summer hemisphere, where stationary GWs cannot propagate upward. The new schemes improve the seasonal representation of stratospheric wind, through changes in both the GWD and the resolved wave forcing, which is modulated by the changed large-scale wind due to the GWD. The downward influence, in response to the changed GWD, is also positive in the tropospheric fields, such as subtropical jet and planetary-scale disturbances. For the medium-range forecasts, improved skill scores on wind speed are achieved globally with the new schemes. The improvements mostly appear only in the stratosphere during the early forecast period (~3 days) but expand to the troposphere as forecast time increases.

Medium-range reference evapotranspiration forecasts for the contiguous United States based on multi-model numerical weather predictions

Reference evapotranspiration (ET0) plays a fundamental role in agronomic, forestry, and water resources management. Estimating and forecasting ET0 have long been recognized as a major challenge for researchers and practitioners in these communities. This work explored the potential of multiple leading numerical weather predictions (NWPs) for estimating and forecasting summer ET0 at 101 U.S. Regional Climate Reference Network stations over nine climate regions across the contiguous United States (CONUS). Three leading global NWP model forecasts from THORPEX Interactive Grand Global Ensemble (TIGGE) dataset were used in this study, including the single model ensemble forecasts from the European Centre for Medium-Range Weather Forecasts (EC), the National Centers for Environmental Prediction Global Forecast System (NCEP), and the United Kingdom Meteorological Office forecasts (MO), as well as multi-model ensemble forecasts from the combinations of these NWP models. A regression calibration was employed to bias correct the ET0 forecasts. Impact of individual forecast variables on ET0 forecasts were also evaluated. The results showed that the EC forecasts provided the least error and highest skill and reliability, followed by the MO and NCEP forecasts. The multi-model ensembles constructed from the combination of EC and MO forecasts provided slightly better performance than the single model EC forecasts. The regression process greatly improved ET0 forecast performances, particularly for the regions involving stations near the coast, or with a complex orography. The performance of EC forecasts was only slightly influenced by the size of the ensemble members, particularly at short lead times. Even with less ensemble members, EC still performed better than the other two NWPs. Errors in the radiation forecasts, followed by those in the wind, had the most detrimental effects on the ET0 forecast performances.

Use of rainfall forecast from a high-resolution global NWP model in a hydrological stream flow model over Narmada river basin during monsoon

Use of NWP model rainfall forecast is useful for hydrological modelling for river basins for flood and water management. In this study, the SWAT model is applied to the Narmada river basin to estimate the surface runoff. The capability of the model is tested for the south west monsoon season (June–September) of 2016 and validated against the NCUM global NWP model forecast data up to Day-5. Two heavy rainfall events are used for verification of the global model forecast with the observed data. The two verification events occurring during 1 July to 15 July (mean of 15 days) and 30 Aug to 11 Sept (mean of 12 days) are considered. Using Arc GIS sub watershed boundaries, the drainage network, slope, land use map and soil map are generated for the whole Narmada river basin. The surface runoff is computed by forcing with observed rainfall data as well as the NCUM model rainfall forecasts data as input to the SWAT model. Model simulations are compared at five stations of Narmada namely—Balkheri, Sandia, Handia, Hoshangabad and Garudeshwar. Initially, the NCUM model rainfall forecast is compared with the actual observation to evaluate the model performances in terms of bias estimation. Then the surface runoff is computed using the observed rainfall data and NCUM forecast data up to Day-5 and made a comparison to evaluate the SWAT model performances. It is shown that the estimated surface runoff during the two active phases is simulated well with some overestimation during the season. The skill of hydrological simulations is examined during different length of forecasts. It is seen that the SWAT model output performance is related to the NWP rainfall forecast skills. NCUM global NWP model is overestimating rainfall amount (wet bias) by 50–70 mm/day over the basin during the two active phases. Further, in all the days’ forecasts, the model is showing wet bias over the upper part of the Narmada river basin and a dry bias over the lower part of the basin. The dry bias is seen to be around 5–35 mm/day.

Resolving Tropical Cyclone Intensity in Models

AbstractIn recent years, global weather forecast models and global climate models have begun to depict intense tropical cyclones, even up to category 5 on the Saffir‐Simpson scale. In light of the limitation of horizontal resolution in such models, the author performs calculations, using the extended Best Track data for Atlantic tropical cyclones, to estimate the ability of models with differing grid spacing to represent Atlantic tropical cyclone intensity statistically. Results indicate that, under optimistic assumptions, models with horizontal grid spacing of one fourth degree or coarser should not produce a realistic number of category 4 and 5 storms unless there are errors in spatial attributes of the wind field. Furthermore, the case of Irma (2017) is used to demonstrate the importance of a realistic depiction of angular momentum and to motivate the use of angular momentum in model evaluation.

Evaluation of Real‐Time Convection‐Permitting Precipitation Forecasts in China During the 2013–2014 Summer Season

AbstractForecasts at a 4 km convection‐permitting resolution over China during the summer season have been produced with the Weather Research and Forecasting model at Nanjing University since 2013. Precipitation forecasts from 2013 to 2014 are evaluated with dense rain gauge observations and compared with operational global model forecasts. Overall, the 4 km forecasts show very good agreement with observations over most parts of China, outperforming global forecasts in terms of spatial distribution, intensity, and diurnal variation. Quantitative evaluations with the Gilbert skill score further confirm the better performance of the 4 km forecasts over global forecasts for heavy precipitation, especially for the thresholds of 100 and 150 mm d−1. Besides bulk characteristics, the representations of some unique features of summer precipitation in China under the influence of the East Asian summer monsoon are further evaluated. These include the northward progression and southward retreat of the main rainband through the summer season, the diurnal variations of precipitation, and the meridional and zonal propagation of precipitation episodes associated with background synoptic flow and the embedded mesoscale convective systems. The 4 km forecast is able to faithfully reproduce most of the features while overprediction of afternoon convection near the southern China coast is found to be a main deficiency that requires further investigations.

FORECAST STRUCTURE DEVELOPMENT FOR MODELLING THE SOUTHERN BRAZILIAN CONTINENTAL SHELF HYDRODYNAMIC

Global forecast models have been widely used worldwide; however, they lack ability to prescribe high quality local data. In this paper, an assembly of oceanographic (RTOFS) and atmospheric (GFS) global data base are used to set up a nested forecast system. The approach proposed relies on the sage of python and FORTRAN scripts to download and assemble global metocean data, in addition to interpolate it towards the study area numerical grid as boundary conditions. Therefore, the TELEMAC-3D hydrodynamic model is automatically initiated in order to provide the regional forecast for the Southern Brazilian Continental Shelf. Results shows that the methodology proposed represent properly the flow dynamics inside the Patos Lagoon and also in the Tramandaí beach, with low/high frequency dynamic effects and influences being reported.

The Representation of Tropospheric Water Vapor Over Low-Latitude Oceans in (Re-)analysis: Errors, Impacts, and the Ability to Exploit Current and Prospective Observations

This paper addresses the representation of lower tropospheric water vapor in the meteorological analyses—fully detailed estimates of atmospheric state—providing the wide temporal and spatial coverage used in many process studies. Analyses are produced in a cycle combining short forecasts from initial conditions with data assimilation that optimally estimates the state of the atmosphere from the previous forecasts and new observations, providing initial conditions for the next set of forecasts. Estimates of water vapor are among the less certain aspects of the state because the quantity poses special challenges for data assimilation while being particularly sensitive to the details of model parameterizations. Over remote tropical oceans observations of water vapor come from two sources: passive observations at microwave or infrared wavelengths that provide relatively strong constraints over large areas on column-integrated moisture but relatively coarse vertical resolution, and occultations of Global Positioning System provide much higher accuracy and vertical resolution but are relatively spatially coarse. Over low-latitude oceans, experiences with two systems suggest that current analyses reproduce much of the large-scale variability in integrated water vapor but have systematic errors in the representation of the boundary layer with compensating errors in the free troposphere; these errors introduce errors of order 10% in radiative heating rates through the free troposphere. New observations, such as might be obtained by future observing systems, improve the estimates of water vapor but this improvement is lost relatively quickly, suggesting that exploiting better observations will require targeted improvements to global forecast models.

Effects of Parameterized Orographic Drag on Weather Forecasting and Simulated Climatology Over East Asia During Boreal Summer

AbstractThe impact of subgrid orographic drag on weather forecasting and simulated climatology over East Asia in boreal summer is examined using two parameterization schemes in a global forecast model. The schemes consider gravity wave drag (GWD) with and without lower‐level wave breaking drag (LLWD) and flow‐blocking drag (FBD). Simulation results from sensitivity experiments verify that the scheme with LLWD and FBD improves the intensity of a summertime continental high over the northern part of the Korean Peninsula, which is exaggerated with GWD only. This is because the enhanced lower tropospheric drag due to the effects of lower‐level wave breaking and flow blocking slows down the wind flowing out of the high‐pressure system in the lower troposphere. It is found that the decreased lower‐level divergence induces a compensating weakening of middle‐ to upper‐level convergence aloft. Extended experiments for medium‐range forecasts for July 2013 and seasonal simulations for June to August of 2013–2015 are also conducted. Statistical skill scores for medium‐range forecasting are improved not only in low‐level winds but also in surface pressure when both LLWD and FBD are considered. A simulated climatology of summertime monsoon circulation in East Asia is also realistically reproduced.

Performance of the New NCEP Global Ensemble Forecast System in a Parallel Experiment

Abstract A new version of the Global Ensemble Forecast System (GEFS, v11) is tested and compared with the operational version (v10) in a 2-yr parallel run. The breeding-based scheme with ensemble transformation and rescaling (ETR) used in the operational GEFS is replaced by the ensemble Kalman filter (EnKF) to generate initial ensemble perturbations. The global medium-range forecast model and the Global Forecast System (GFS) analysis used as the initial conditions are upgraded to the GFS 2015 implementation version. The horizontal resolution of GEFS increases from Eulerian T254 (~52 km) for the first 8 days of the forecast and T190 (~70 km) for the second 8 days to semi-Lagrangian T574 (~34 km) and T382 (~52 km), respectively. The sigma pressure hybrid vertical layers increase from 42 to 64 levels. The verification of geopotential height, temperature, and wind fields at selected levels shows that the new GEFS significantly outperforms the operational GEFS up to days 8–10 except for an increased warm bias over land in the extratropics. It is also found that the parallel system has better reliability in the short-range probability forecasts of precipitation during warm seasons, but no clear improvement in cold seasons. There is a significant degradation of TC track forecasts at days 6–7 during the 2012–14 TC seasons over the Atlantic and eastern Pacific. This degradation is most likely a sampling issue from a low number of TCs during these three TC seasons. The results for an extended verification period (2011–14) and the recent two hurricane seasons (2015 and 2016) are generally positive. The new GEFS became operational at NCEP on 2 December 2015.

Closed-Form Analytic Solution of Cloud Dissipation for a Mixed-Layer Model

Abstract Stratocumulus clouds play an important role in climate cooling and are hard to predict using global climate and weather forecast models. Thus, previous studies in the literature use observations and numerical simulation tools, such as large-eddy simulation (LES), to solve the governing equations for the evolution of stratocumulus clouds. In contrast to the previous works, this work provides an analytic closed-form solution to the cloud thickness evolution of stratocumulus clouds in a mixed-layer model framework. With a focus on application over coastal lands, the diurnal cycle of cloud thickness and whether or not clouds dissipate are of particular interest. An analytic solution enables the sensitivity analysis of implicitly interdependent variables and extrema analysis of cloud variables that are hard to achieve using numerical solutions. In this work, the sensitivity of inversion height, cloud-base height, and cloud thickness with respect to initial and boundary conditions, such as Bowen ratio, subsidence, surface temperature, and initial inversion height, are studied. A critical initial cloud thickness value that can be dissipated pre- and postsunrise is provided. Furthermore, an extrema analysis is provided to obtain the minima and maxima of the inversion height and cloud thickness within 24 h. The proposed solution is validated against LES results under the same initial and boundary conditions.

Direct climate change impacts on growth and drought risk in blue gum (Eucalyptus globulus) plantations in Australia

ABSTRACTAustralia’s climate is changing and Australia’s forests have been identified as vulnerable to climate change impacts. The process-based model CABALA, an ensemble of global circulation models, and a range of scenarios on plant response to elevated CO2 (eCO2) and site conditions, were used to predict the direct effects of climate change on the productivity and mortality risks for Australia’s blue gum (Eucalyptus globulus) plantation estate. The modelling showed considerable uncertainty about future outcomes across large parts of the estate, with best-case and worst-case scenarios varying from decreased to increased production. In some areas we can be confident of future outcomes. Nevertheless, it is clear that, across the whole estate, appropriate management can reduce risk and ensure that we are able to capitalise on potential beneficial aspects of climate change. Under most future scenarios, without adaptation, the drier parts of the plantation estate are at risk. However, in all but the eastern areas of Western Australia, and at the driest margins of the estate in South Australia and Victoria, adaptation options can reduce risk and ensure productivity. Other areas currently at the cold margins of the estate, notably the highlands of Victoria and across Tasmania, appear to increase in productivity under most future scenarios. Even in those areas where productivity may potentially increase, experience suggests that pests and diseases may pose a risk. If we are to benefit from any aspect of climate change, then pursuit of best-practice sustainable forest management is critical. The greatest source of uncertainty in forecasts is associated with how photosynthesis in field-grown trees will respond to elevated atmospheric carbon dioxide levels (eCO2). Regional differences in global circulation model forecasts are important, particularly in Western Australia, where the extent of future rainfall decline significantly impacts predictions. Local site factors, exemplified in this study by soil depth, will play an important role in modulating climate change impacts, particularly through their role in constraining responses to eCO2.

A Simplified Extended Kalman Filter assimilation of soil moisture for the SL-AV global medium-range forecast model

The implementation of the Simplified Extended Kalman Filter (SEKF) for the deep soil moisture initialization in the SL-AV global atmosphere model is described. Special attention is paid to the calculation of the observation operator and analysis increment. SL-AV screen-level parameters forecasts are estimated with SEKF and optimal interpolation initialization methods. It is demonstrated that the implementation of the assimilation algorithm improves the model forecast quality for screen-level temperature and relative humidity.

Numerical diagnosis of situations causing heavy rainfall over the Western Himalayas

Heavy rainfall events frequently occur over the western and central Himalayas during the monsoon season causing losses of lives and damages to properties over the fragile mountain environment. A heavy rainfall event that occurred over Uttarkashi (30.73°N, 78.45°E) in the Western Himalayas on 3rd August 2012 is investigated. The formation of the storm, its evolution and the initial physical mechanisms responsible for this event are analyzed by using a double nested Weather Research and Forecasting (WRF) model. The model skill is evaluated against available observations and analysis fields. The impact of assimilation of Global Telecommunication System (GTS) data in the model initial condition is also studied. The model simulated rainfall is compared with daily rainfall data from satellite-gauge merged rainfall and 3 hourly Tropical Rainfall Measuring Mission (TRMM) estimated rainfall. The double nested configuration of the model successfully simulates this heavy rainfall event. 3DVAR assimilation of GTS data in the model initial condition improves prediction of precipitation amount and location of heavy rainfall. The results of double nested WRF model simulation with GTS data assimilation are compared with the results of global model forecasts of National Centre for Medium Range Weather Forecasting (NCMRWF), India. Forecast skill of the model with and without data assimilation is computed with respect to TRMM estimated daily rainfall.

Tropical precipitation extremes: Response to SST‐induced warming in aquaplanet simulations

AbstractScaling of tropical precipitation extremes in response to warming is studied in aquaplanet experiments using the global Weather Research and Forecasting (WRF) model. We show how the scaling of precipitation extremes is highly sensitive to spatial and temporal averaging: while instantaneous grid point extreme precipitation scales more strongly than the percentage increase (∼7% K−1) predicted by the Clausius‐Clapeyron (CC) relationship, extremes for zonally and temporally averaged precipitation follow a slight sub‐CC scaling, in agreement with results from Climate Model Intercomparison Project (CMIP) models. The scaling depends crucially on the employed convection parameterization. This is particularly true when grid point instantaneous extremes are considered. These results highlight how understanding the response of precipitation extremes to warming requires consideration of dynamic changes in addition to the thermodynamic response. Changes in grid‐scale precipitation, unlike those in convective‐scale precipitation, scale linearly with the resolved flow. Hence, dynamic changes include changes in both large‐scale and convective‐scale motions.

Forecast Advisory for a Cold-Season Heavy Rainfall/Flood Event That Developed from Multiple Interactions of the Cold-Surge Vortex with Cold-Surge Flows in the South China Sea

Abstract The peak intensity occurrence frequency over the life cycles of parent cold-surge vortices (CSVs) for heavy rainfall/flood (HRF) events is classified into two types depending on their life cycles having two or three peak intensities, denoted as HRF2 or HRF3, respectively. The formation of an HRF2 event from its parent CSV(HRF2) formation is ≤5 days, while the formation of an HRF3 event is ≥6 days. The latter group contributes ~57% of the total number of HRF events. As a result of some model constraints, the formation and development of HRF3 events are not well forecasted by the Global Forecast System (GFS) and regional forecast models. The life cycle and second peak intensity for CSV(HRF3) allow for the introduction of a forecast advisory for HRF3 events. Identification of CSVs and two sufficient requirements for the formation and occurrence of HRF events were developed by previous studies. Nevertheless, two new necessary steps are now included in the proposed forecast advisory. The population ratio for CSV(HRF3) and the regular CSV is only about 15%. The occurrence optimum time to for the CSV(HRF3) second peak intensity from this vortex formation is about 3 days 6 h. The GFS forecast over to is utilized to identify CSV(HRF3). Then, the relay of the GFS forecast from the occurrence time of the CSV(HRF3) second peak is used to predict the formation/occurrence of HRF3 events. Six HRF3 events during cold seasons for 2013–16 are used to test the feasibility of this forecast advisory. Results clearly demonstrate this advisory is a success for the forecast of HRF3 events over the entire life cycles of their parent CSV(HRF3)s.

A Global Statistical–Dynamical Tropical Cyclone Wind Radii Forecast Scheme

Abstract Forecasts of tropical cyclone (TC) surface wind structure have recently begun to show some skill, but the number of reliable forecast tools, mostly regional hurricane and select global models, remains limited. To provide additional wind structure guidance, this work presents the development of a statistical–dynamical method to predict tropical cyclone wind structure in terms of wind radii, which are defined as the maximum extent of the 34-, 50-, and 64-kt (1 kt = 0.514 m s−1) winds in geographical quadrants about the center of the storm. The basis for TC size variations is developed from an infrared satellite-based record of TC size, which is homogenously calculated from a global sample. The change in TC size is predicted using a statistical–dynamical approach where predictors are based on environmental diagnostics derived from global model forecasts and observed storm conditions. Once the TC size has been predicted, the forecast intensity and track are used along with a parametric wind model to estimate the resulting wind radii. To provide additional guidance for applications and users that require forecasts of central pressure, a wind–pressure relationship that is a function of TC motion, intensity, wind radii (i.e., size), and latitude is then applied to these forecasts. This forecast method compares well with similar wind structure forecasts made by global forecast and regional hurricane models and when these forecasts are used as a member of a simple consensus; its inclusion improves the forecast performance of the consensus.

Assimilation of High-Resolution Satellite-Derived Atmospheric Motion Vectors: Impact on HWRF Forecasts of Tropical Cyclone Track and Intensity

It is well known that global numerical model analyses and forecasts benefit from the routine assimilation of atmospheric motion vectors (AMVs) derived from meteorological satellites. Recent studies have also shown that the assimilation of enhanced (spatial and temporal) AMVs can benefit research-mode regional model forecasts of tropical cyclone track and intensity. In this study, the impact of direct assimilation of enhanced (higher resolution) AMV datasets in the NCEP operational Hurricane Weather Research and Forecasting Model (HWRF) system is investigated. Forecasts of Atlantic tropical cyclone track and intensity are examined for impact by inclusion of enhanced AMVs via direct data assimilation. Experiments are conducted for AMVs derived using two methodologies (“HERITAGE” and “GOES-R”), and also for varying levels of quality control in order to assess and inform the optimization of the AMV assimilation process. Results are presented for three selected Atlantic tropical cyclone events and compared to Control forecasts without the enhanced AMVs as well as the corresponding operational HWRF forecasts. The findings indicate that the direct assimilation of high-resolution AMVs has an overall modest positive impact on HWRF forecasts, but the impact magnitudes are dependent on the 1) availability of rapid scan imagery used to produce the AMVs, 2) AMV derivation approach, 3) level of quality control employed in the assimilation, and 4) vortex initialization procedure (including the degree to which unbalanced states are allowed to enter the model analyses).

Framework for improvement by vertical enhancement: A simple approach to improve representation of low and high‐level clouds in large‐scale models

AbstractLow and high clouds of shallow extent, especially stratocumulus and even more so for high‐level cirrus clouds that reside where vertical resolution is particularly coarse, are poorly represented in large‐scale models such as global climate models and weather forecasting models. This adversely affects, among others, estimation of cloud feedbacks for climate prediction and weather forecasts. Here we address vertical resolution as a reason for the failure of these models to adequately represent shallow clouds. We introduce a new methodology, the Framework for Improvement by Vertical Enhancement (FIVE). FIVE computes selected processes on a one‐dimensional vertical grid with local high resolution in the boundary layer and near the tropopause. In addition to the host model, variables on the locally high‐resolution grid are predicted in parallel so that high‐resolution information is retained. By exchanging tendencies with one another, the host model and high‐resolution field are always synchronized. The methodology is demonstrated for drizzling stratocumulus capped by a sharp inversion. First, FIVE is applied to a single‐column model to identify the cause of biases associated with computing an assigned process at low resolution. Second, a two‐dimensional regional model coupled with FIVE is shown to produce results comparable to those performed with high vertical resolution. FIVE is thus expected to represent low clouds more realistically and hence reduce the low‐cloud bias in large‐scale models. Finally, we propose a number of methods that will be developed and tested to further optimize FIVE.

The Advantage of Using International Multimodel Ensemble for Seasonal Precipitation Forecast over Israel

This study analyzes the results of monthly and seasonal precipitation forecasting from seven different global climate forecast models for major basins in Israel within October–April 1982–2010. The six National Multimodel Ensemble (NMME) models and the ECMWF seasonal model were used to calculate an International Multimodel Ensemble (IMME). The study presents the performance of both monthly and seasonal predictions of precipitation accumulated over three months, with respect to different lead times for the ensemble mean values, one per individual model. Additionally, we analyzed the performance of different combinations of models. We present verification of seasonal forecasting using real forecasts, focusing on a small domain characterized by complex terrain, high annual precipitation variability, and a sharp precipitation gradient from west to east as well as from south to north. The results in this study show that, in general, the monthly analysis does not provide very accurate results, even when using the IMME for one-month lead time. We found that the IMME outperformed any single model prediction. Our analysis indicates that the optimal combinations with the high correlation values contain at least three models. Moreover, prediction with larger number of models in the ensemble produces more robust predictions. The results obtained in this study highlight the advantages of using an ensemble of global models over single models for small domain.