Improvement In Forecast Skill Research Articles

Regime‐dependent statistical post‐processing of ensemble forecasts

AbstractA number of realizations of one or more numerical weather prediction (NWP) models, initialised at a variety of initial conditions, compose an ensemble forecast. These forecasts exhibit systematic errors and biases that can be corrected by statistical post‐processing. Post‐processing yields calibrated forecasts by analysing the statistical relationship between historical forecasts and their corresponding observations. This article aims to extend post‐processing methodology to incorporate atmospheric circulation. The circulation, or flow, is largely responsible for the weather that we experience and it is hypothesized here that relationships between the NWP model and the atmosphere depend upon the prevailing flow. Numerous studies have focussed on the tendency of this flow to reduce to a set of recognisable arrangements, known as regimes, which recur and persist at fixed geographical locations. This dynamical phenomenon allows the circulation to be categorized into a small number of regime states. In a highly idealized model of the atmosphere, the Lorenz ‘96 system, ensemble forecasts are subjected to well‐known post‐processing techniques conditional on the system's underlying regime. Two different variables, one of the state variables and one related to the energy of the system, are forecasted and considerable improvements in forecast skill upon standard post‐processing are seen when the distribution of the predictand varies depending on the regime. Advantages of this approach and its inherent challenges are discussed, along with potential extensions for operational forecasters.

The future of forecasting for renewable energy

AbstractForecasting for wind and solar renewable energy is becoming more important as the amount of energy generated from these sources increases. Forecast skill is improving, but so too is the way forecasts are being used. In this paper, we present a brief overview of the state‐of‐the‐art of forecasting wind and solar energy. We describe approaches in statistical and physical modeling for time scales from minutes to days ahead, for both deterministic and probabilistic forecasting. Our focus changes then to consider the future of forecasting for renewable energy. We discuss recent advances which show potential for great improvement in forecast skill. Beyond the forecast itself, we consider new products which will be required to aid decision making subject to risk constraints. Future forecast products will need to include probabilistic information, but deliver it in a way tailored to the end user and their specific decision making problems. Businesses operating in this sector may see a change in business models as more people compete in this space, with different combinations of skills, data and modeling being required for different products. The transaction of data itself may change with the adoption of blockchain technology, which could allow providers and end users to interact in a trusted, yet decentralized way. Finally, we discuss new industry requirements and challenges for scenarios with high amounts of renewable energy. New forecasting products have the potential to model the impact of renewables on the power system, and aid dispatch tools in guaranteeing system security.This article is categorized under: Energy Infrastructure > Systems and Infrastructure Wind Power > Systems and Infrastructure Photovoltaics > Systems and Infrastructure

Tropical and Midlatitude Impact on Seasonal Polar Predictability in the Community Earth System Model

Abstract The impact on seasonal polar predictability from improved tropical and midlatitude forecasts is explored using a perfect-model experiment and applying a nudging approach in a GCM. We run three sets of 7-month long forecasts: a standard free-running forecast and two nudged forecasts in which atmospheric winds, temperature, and specific humidity (U, V, T, Q) are nudged toward one of the forecast runs from the free ensemble. The two nudged forecasts apply the nudging over different domains: the tropics (30°S–30°N) and the tropics and midlatitudes (55°S–55°N). We find that the tropics have modest impact on forecast skill in the Arctic or Antarctica both for sea ice and the atmosphere that is mainly confined to the North Pacific and Bellingshausen–Amundsen–Ross Seas, whereas the midlatitudes greatly improve Arctic winter and Antarctic year-round forecast skill. Arctic summer forecast skill from May initialization is not strongly improved in the nudged forecasts relative to the free forecast and is thus mostly a “local” problem. In the atmosphere, forecast skill improvement from midlatitude nudging tends to be largest in the polar stratospheres and decreases toward the surface.

Web-based decision support toolbox for Spanish reservoirs

Abstract. Under the S-ClimWaRe (Seasonal Climate prediction in support of Water Reservoirs management) initiative, a climate service to support decision-making process by water managers in Spanish reservoirs has been developed. It consists in a web-based toolbox jointly designed with stakeholders. The website is organized in two main areas. The first one allows the user to explore, for any water reservoir or grid point over continental Spain, the existing hydrological variability and risk linked to climate variability. This is performed through a set of indicators obtained from time series of hydrological and meteorological observations and North Atlantic Oscillation (NAO) index, identified as main climate driver in this geographical region. The second main area provides seasonal forecasts of NAO and both reservoir inflow and precipitation, complemented by information on probabilistic forecasts skill. Currently the NAO index is the only driver implemented for display, and forecasts come from a statistical forecasting system developed only for the extended winter NDJFM period. Through the MEDSCOPE (MEDiterranean Services Chain based On climate PrEdictions) project new sources of predictability and relationships with different climate drivers will be explored. Forecast skill improvement is expected after the combination and weighting of ensemble members of the Copernicus seasonal forecasting systems. These forecasts will feed more sophisticated hydrological models. The toolbox has been flexible designed with respect to sources of seasonal forecasts and extension to additional drivers, variables and seasons. In this way, user requirements and scientific progress will be easily incorporated to new versions of this climate service.

Theoretical Predictability Limits of Spatially Anisotropic Multifractal Processes: Implications for Weather Prediction

AbstractA correlation spectrum‐based approach is used to express the theoretical predictability limits of multifractal processes as an analytical function of their anisotropy parameters. This spatially anisotropic power law function is then used to investigate the general impact of anisotropy on the predictability of atmospheric fields in the weather regime. The investigation reveals that (i) vertical stratification of a field increases and decreases its super and subsphero‐scale predictability limits, respectively; (ii) trivial horizontal anisotropy slightly improves predictability at all scales; and (iii) horizontal anisotropy together with vertical stratification significantly enhances its predictability over almost the entire scale range. Applying these general results to the case of horizontal wind fields suggests that the interplay between spatial‐anisotropy and atmospheric predictability could account for improvements in forecast skill, commonly observed during the occurrence of rotating thunderstorms and breaks in the Indian summer monsoon.

Incorporating Antecedent Soil Moisture into Streamflow Forecasting

This study incorporates antecedent (preceding) soil moisture into forecasting streamflow volumes within the North Platte River Basin, Colorado/Wyoming (USA). The incorporation of antecedent soil moisture accounts for infiltration and can improve streamflow predictions. Current Natural Resource Conservation Service (NRCS) forecasting methods are replicated, and a comparison is drawn between current NRCS forecasts and proposed forecasting methods using antecedent soil moisture. Current predictors used by the NRCS in regression-based streamflow forecasting include precipitation, streamflow persistence (previous season streamflow volume) and snow water equivalent (SWE) from SNOTEL (snow telemetry) sites. Proposed methods utilize antecedent soil moisture as a predictor variable in addition to the predictors noted above. A decision system was used to segregate data based on antecedent soil moisture conditions (e.g., dry, wet or normal). Principal Components Analysis and Stepwise Linear Regression were applied to generate streamflow forecasts, and numerous statistics were determined to measure forecast skill. The results show that when incorporating antecedent soil moisture, the “poor” forecasts (i.e., years in which the NRCS forecast differed greatly from the observed value) were improved, while the overall forecast skill remains unchanged. The research presented shows the need to increase the monitoring and collection of soil moisture data in mountainous western U.S. watersheds, as this parameter results in improved forecast skill.

The Impact of Cloud Representation on the Sub-Seasonal Forecasts of Atmospheric Teleconnections and Preferred Circulation Regimes in the Northern Hemisphere

The impact of cloud representation on the simulation of mid-latitude recurrent large-scale flows and forecast skill of mid-latitude atmospheric teleconnections is evaluated using the Community Climate System Model, version 4 (CCSM4), and the super-parameterized CCSM4 (SP-CCSM4). Patterns of low-level atmospheric circulation anomalies and convection associated with the Madden–Julian oscillation (MJO) are affected by the method used for the representation of cloud processes. The configuration of the model using super-parameterization for the representation of cloud processes produces MJO-related patterns that agree better with observations than the configuration of the model using a conventional cloud parameterization scheme. The recurrent circulation regimes of the mid-latitudes are also sensitive to the representation of cloud processes. In the North Atlantic sector, the inability of CCSM4 to simulate the Scandinavian blocking regime is corrected in the super-parameterized version of the model. In the North Pacific sector, the strength of the clustering (measured by a variance ratio) is too large in CCSM4 compared with observations and SP-CCSM4. The SP-CCSM4 model has better forecast skill for the MJO amplitude and phase than the model with conventional representation of moist convective processes. In turn, the improved forecast skill of the super-parameterized model results in better forecast skill for mid-latitude teleconnections in 500 hPa geopotential height anomalies forced by the MJO convection.

Skill of medium-range reforecast for summertime extraordinary Arctic Cyclones in 1986–2016

The forecast skill for 26 extraordinary Arctic cyclones (EACs) in summer (June–August) during 1986–2016 was assessed using a reforecast data by the National Oceanic and Atmospheric Administration (NOAA)'s second-generation Global Ensemble Forecast System (GEFS). On average, more than 90% of ensemble members predicted the existence of the EACs for lead times of ≤3.0 days. The average central position error for the EACs decreases to ≤433.1 km (half the mean radius of the mature 26 EACs) for a lead time of 3.0 days, with the average central pressure error of ∼6.9 hPa. The GEFS reforecast for the 26 EACs shows a similar forecast skill to the GEFS reforecast and operational medium-range ensemble forecasts for 10 EACs occurred in summer during 2008–2016. There is no trend of improvements in forecast skills of existence, central pressure and position of the EACs at all lead times. In the verification of the strike probability for EACs within a 400-km radius, the probability information is useful in 1.0- to 5.0-day forecasts, although the forecast probabilities were overconfident for lead times more than 1.0 days.

Progress in Forecast Skill at Three Leading Global Operational NWP Centers during 2015–17 as Seen in Summary Assessment Metrics (SAMs)

Abstract The summary assessment metric (SAM) method is applied to an array of primary assessment metrics (PAMs) for the deterministic forecasts of three leading numerical weather prediction (NWP) centers for the years 2015–17. The PAMs include anomaly correlation, RMSE, and absolute mean error (i.e., the absolute value of bias) for different forecast times, vertical levels, geographic domains, and variables. SAMs indicate that in terms of forecast skill ECMWF is better than NCEP, which is better than but approximately the same as UKMO. The use of SAMs allows a number of interesting features of the evolution of forecast skill to be observed. All three centers improve over the 3-yr period. NCEP short-term forecast skill substantially increases during the period. Quantitatively, the effect of the 11 May 2016 NCEP upgrade to the four-dimensional ensemble variational data assimilation (4DEnVar) system is a 7.37% increase in the probability of improved skill relative to a randomly chosen forecast metric from 2015 to 2017. This is the largest SAM impact during the study period. However, the observed impacts are within the context of slowly improving forecast skill for operational global NWP as compared to earlier years. Clearly, the systems lagging ECMWF can improve, and there is evidence from SAMs in addition to the 4DEnVar example that improvements in forecast and data assimilation systems are still leading to forecast skill improvements.

Spatial and Temporal Averaging Windows and Their Impact on Forecasting: Exactly Solvable Examples

Abstract In making weather and climate predictions, the goal is often not to predict the instantaneous, local value of temperature, wind speed, or rainfall; instead, the goal is often to predict these quantities after averaging in time and/or space-for example, over one day or one week. What is the impact of spatial and/or temporal averaging on forecasting skill?Here this question is investigated using simple stochastic models that can be solved exactly analytically. While the models are idealized, their exact solutions allow clear results that are not affected by errors from numerical simulations or from random sampling. As a model of time series of oscillatory weather fluctuations, the complex Ornstein-Uhlenbeck process is used. To furthermore investigate spatial averaging, the stochastic heat equation is used as an idealized spatiotemporal model for moisture and rainfall. Space averaging and time averaging are shown to have distinctly different impacts on prediction skill. Spatial averaging leads to improved forecast skill, in line with some forms of basic intuition. Time averaging, on the other hand, is more subtle: it may either increase or decrease forecast skill. The subtle effects of time averaging are seen to arise from the relative definitions of the time averaging window and the lead time. These results should help in understanding and comparing forecasts with different temporal and spatial averaging windows.

Relationships between potential, attainable, and actual skill in a decadal prediction experiment

A sequence of ensemble forecasts permits the estimation of the “predictable” and “unpredictable” components of the forecasts and hence the predictability of the system. Predictability is a feature of a physical and/or mathematical system which characterizes its “ability to be predicted” as measured, for instance, by the rate at which initially close states separate. Modern climate models may be used to predict climate evolution on seasonal to decadal timescales and they may also be used to estimate predictability. The predictability in this case measures the model’s ability to predict its own evolution and, for a well behaved climate model, this may be a useful measure of the predictability of the climate system itself. Forecast skill, where forecasts are compared to observations, indicates a forecast system’s “ability to predict” the evolution of the climate system. Common measures of skill include correlation, mean square error and mean square skill score. When the model is used to predict its own evolution, rather than the evolution of the actual climate system, the same measures of skill may be used but are then predictability measures and termed measures of “potential skill”. The expectation is that the model’s skill in predicting its own evolution will be greater than its ability to predict the actual evolution of the system in the face of observational uncertainty and model deficiencies. The further expectation is that the differences between potential and actual skill can give an indication of where and to what extent there is the potential for improvement in forecast skill. Predictability studies may be undertaken with available models but it may not be immediately obvious which statistical conditions are necessary if potential skill is to be a reasonable estimate of attainable skill. The purpose of this paper is to investigate and illustrate the spatial and temporal behaviour of the predictable and unpredictable components of surface air temperature in a decadal prediction experiment and to discuss some of the implications for predictability studies and for analyzing model behaviour. Formal relationships between potential, actual, and attainable skill are developed which, however, provide only modest constraints on model results if they are to be used to infer predictability. The results of various scalings of the forecasts on potential and actual skill and their relationship are also considered. Differences in the actual and potential forecast skill between models and/or model versions can be partitioned into components associated with known statistics as a guide to forecast improvement.

A Flexible and Efficient Radiation Scheme for the ECMWF Model

AbstractThis paper describes a new radiation scheme ecRad for use both in the model of the European Centre for Medium‐Range Weather Forecasts (ECMWF), and off‐line for noncommercial research. Its modular structure allows the spectral resolution, the description of cloud and aerosol optical properties, and the solver, to be changed independently. The available solvers include the Monte Carlo Independent Column Approximation (McICA), Tripleclouds, and the Speedy Algorithm for Radiative Transfer through Cloud Sides (SPARTACUS), the latter which makes ECMWF the first global model capable of representing the 3‐D radiative effects of clouds. The new implementation of the operational McICA solver produces less noise in atmospheric heating rates, and is 41% faster, which can yield indirect forecast skill improvements via calling the radiation scheme more frequently. We demonstrate how longwave scattering may be implemented for clouds but not aerosols, which is only 4% more computationally costly overall than neglecting longwave scattering and yields further modest forecast improvements. It is also shown how a sequence of radiation changes in the last few years has led to a substantial reduction in stratospheric temperature biases.

Superensemble forecast of respiratory syncytial virus outbreaks at national, regional, and state levels in the United States

Respiratory syncytial virus (RSV) infections peak during the winter months in the United States, yet the timing, intensity, and onset of these outbreaks vary each year. An RSV vaccine is on the cusp of being released; precise models and accurate forecasts of RSV epidemics may prove vital for planning where and when the vaccine should be deployed. Accurate forecasts with sufficient spatial and temporal resolution could also be used to support the prevention or treatment of RSV infections. Previously, we developed and validated an RSV forecast system at the regional scale in the United States. This model-inference system had considerable forecast skill, relative to the historical expectance, for outbreak peak intensity, total outbreak size, and onset, but only marginal skill for predicting the timing of the outbreak peak. Here, we use a superensemble approach to combine three forecasting methods for RSV prediction in the US at three different spatial resolutions: national, regional, and state. At the regional and state levels, we find a substantial improvement of forecast skill, relative to historical expectance, for peak intensity, timing, and onset outbreak up to two months in advance of the predicted outbreak peak. Moreover, due to the greater variability of RSV outbreaks at finer spatial scales, we find that improvement of forecast skill at the state level exceeds that at the regional and national levels. Such finer scale superensemble forecasts may be more relevant for effecting local-scale interventions, particularly in communities with a high burden of RSV infection.

A THREE-DIMENSIONAL AEROSOL VARIATIONAL DATA ASSIMILATION SYSTEM FOR AIRCRAFT AND SURFACE OBSERVATIONS

Abstract. A three-dimensional variational data assimilation system is implemented within the Weather Research and Forecasting/Chemistry model, and the control variables consist of eight species of the Model for Simulation Aerosol Interactions and Chemistry scheme. In the experiments, the three-dimensional profiles of aircraft speciated observations and surface concentration observations acquired during the California Research at the Nexus of Air Quality and Climate Change field campaign are assimilated. The data assimilation experiments are performed at 02:00 local time 2 June 2010, assimilating surface observations at 02:00 and aircraft observations from 01:30 to 02:30 local time. The results show that the assimilation of both aircraft and surface observations improves the subsequent forecasts. The improved forecast skill resulting from the assimilation of the aircraft profiles persists a time longer than the assimilation of the surface observations, which suggests the necessity of vertical profile observations for extending aerosol forecasting time.

Evaluation of performance of seasonal precipitation prediction at regional scale over India

The seasonal scale precipitation amount is an important ingredient in planning most of the agricultural practices (such as a type of crops, and showing and harvesting schedules). India being an agroeconomic country, the seasonal scale prediction of precipitation is directly linked to the socioeconomic growth of the nation. At present, seasonal precipitation prediction at regional scale is a challenging task for the scientific community. In the present study, an attempt is made to develop multi-model dynamical-statistical approach for seasonal precipitation prediction at the regional scale (meteorological subdivisions) over India for four prominent seasons which are winter (from December to February; DJF), pre-monsoon (from March to May; MAM), summer monsoon (from June to September; JJAS), and post-monsoon (from October to December; OND). The present prediction approach is referred as extended range forecast system (ERFS). For this purpose, precipitation predictions from ten general circulation models (GCMs) are used along with the India Meteorological Department (IMD) rainfall analysis data from 1982 to 2008 for evaluation of the performance of the GCMs, bias correction of the model results, and development of the ERFS. An extensive evaluation of the performance of the ERFS is carried out with dependent data (1982–2008) as well as independent predictions for the period 2009–2014. In general, the skill of the ERFS is reasonably better and consistent for all the seasons and different regions over India as compared to the GCMs and their simple mean. The GCM products failed to explain the extreme precipitation years, whereas the bias-corrected GCM mean and the ERFS improved the prediction and well represented the extremes in the hindcast period. The peak intensity, as well as regions of maximum precipitation, is better represented by the ERFS than the individual GCMs. The study highlights the improvement of forecast skill of the ERFS over 34 meteorological subdivisions as well as India as a whole during all the four seasons.

Assimilation of Himawari‐8 All‐Sky Radiances Every 10 Minutes: Impact on Precipitation and Flood Risk Prediction

AbstractJapan's new geostationary satellite “Himawari‐8” enables updating precipitation and flood predictions as frequently as every 10 min and to capture as early as possible the flood risk associated with rapidly changing severe weather. This study focuses on the advantage of the frequent update of precipitation and flood predictions by assimilating all‐sky Himawari‐8 infrared radiances in the case of September 2015 Kanto‐Tohoku heavy rainfall, a major flooding event in Japan. The analyzed tropical cyclone representation and moisture transport are improved by the Himawari‐8 data assimilation. Deterministic runs from the analyses with the Himawari‐8 data provide much improved forecasts of a strong precipitation band. Moreover, every 10 min updates of river discharge forecasts driven by these improved precipitation forecasts show a large improvement in forecast skill with longer lead times. The results demonstrate that the frequent refresh aids to capture a signature of flooding at an earlier stage.

Validation and deployment of the first Lidar based weather observation network in New York State: The NYS MesoNet Project

The number and quality of atmospheric observations used by meteorologists and operational forecasters are increasing year after year, and yet, consistent improvements in forecast skill remains a challenge. While contributing factors involving these challenges have been identified, including the difficulty in accurately establishing initial conditions, improving the observations at regional and local scales is necessary for accurate depiction of the atmospheric boundary layer (below 2km), particularly the wind profile, in high resolution numerical models. Above the uncertainty of weather forecasts, the goal is also to improve the detection of severe and extreme weather events (severe thunderstorms, tornadoes and other mesoscale phenomena) that can adversely affect life, property and commerce, primarily in densely populated urban centers. This paper will describe the New York State Mesonet that is being deployed in the state of New York, USA. It is composed of 126 stations including 17 profiler sites. These sites will acquire continuous upper air observations through the combination of WINDCUBE Lidars and microwave radiometers. These stations will provide temperature, relative humidity & “3D” wind profile measurements through and above the planetary boundary layer (PBL) and will retrieve derived atmospheric quantities such as the PBL height, cloud base, momentum fluxes, and aerosol & cloud optical properties. The different modes and configurations that will be used for the Lidars are discussed. The performances in terms of data availability and wind accuracy and precision are evaluated. Several profiles with specific wind and aerosol features are presented to illustrate the benefits of the use of Coherent Doppler Lidars to monitor accurately the PBL.

Impact of Sea Surface Temperature Forcing on Weeks 3 and 4 Forecast Skill in the NCEP Global Ensemble Forecasting System

Abstract The Global Ensemble Forecasting System (GEFS) is being extended from 16 to 35 days to cover the subseasonal period, bridging weather and seasonal forecasts. In this study, the impact of SST forcing on the extended-range land-only global 2-m temperature, continental United States (CONUS) accumulated precipitation, and MJO skill are explored with version 11 of the GEFS (GEFSv11) under various SST forcing configurations. The configurations consist of 1) the operational GEFS 90-day e-folding time of the observed real-time global SST (RTG-SST) anomaly relaxed to climatology, 2) an optimal AMIP configuration using the observed daily RTG-SST analysis, 3) a two-tier approach using the CFSv2-predicted daily SST, and 4) a two-tier approach using bias-corrected CFSv2-predicted SST, updated every 24 h. The experimental period covers the fall of 2013 and the winter of 2013/14. The results indicate that there are small differences in the ranked probability skill scores (RPSSs) between the various SST forcing experiments. The improvements in forecast skill of the Northern Hemisphere 2-m temperature and precipitation for weeks 3 and 4 are marginal, especially for North America. The bias-corrected CFSv2-predicted SST experiment generally delivers superior performance with statistically significant improvement in spatially and temporally aggregated 2-m temperature RPSSs over North America. Improved representation of the SST forcing (AMIP) increased the forecast skill for MJO indices up through week 2, but there is no significant improvement of the MJO forecast skill for weeks 3 and 4. These results are obtained over a short period with weak MJO activity and are also subject to internal model weaknesses in representing the MJO. Additional studies covering longer periods with upgraded model physics are warranted.

Data Assimilation in the Solar Wind: Challenges and First Results.

Data assimilation (DA) is used extensively in numerical weather prediction (NWP) to improve forecast skill. Indeed, improvements in forecast skill in NWP models over the past 30 years have directly coincided with improvements in DA schemes. At present, due to data availability and technical challenges, DA is underused in space weather applications, particularly for solar wind prediction. This paper investigates the potential of advanced DA methods currently used in operational NWP centers to improve solar wind prediction. To develop the technical capability, as well as quantify the potential benefit, twin experiments are conducted to assess the performance of the Local Ensemble Transform Kalman Filter (LETKF) in the solar wind model ENLIL. Boundary conditions are provided by the Wang‐Sheeley‐Arge coronal model and synthetic observations of density, temperature, and momentum generated every 4.5 h at 0.6 AU. While in situ spacecraft observations are unlikely to be routinely available at 0.6 AU, these techniques can be applied to remote sensing of the solar wind, such as with Heliospheric Imagers or interplanetary scintillation. The LETKF can be seen to improve the state at the observation location and advect that improvement toward the Earth, leading to an improvement in forecast skill in near‐Earth space for both the observed and unobserved variables. However, sharp gradients caused by the analysis of a single observation in space resulted in artificial wavelike structures being advected toward Earth. This paper is the first attempt to apply DA to solar wind prediction and provides the first in‐depth analysis of the challenges and potential solutions.

Assessment of Radio Occultation Observations from the COSMIC-2 Mission with a Simplified Observing System Simulation Experiment Configuration

AbstractThe mainstay of the global radio occultation (RO) system, the COSMIC constellation of six satellites launched in April 2006, is already past the end of its nominal lifetime and the number of soundings is rapidly declining because the constellation is degrading. For about the last decade, COSMIC profiles have been collected and their retrievals assimilated in numerical weather prediction systems to improve operational weather forecasts. The success of RO in increasing forecast skill and COSMIC’s aging constellation have motivated planning for the COSMIC-2 mission, a 12-satellite constellation to be deployed in two launches. The first six satellites (COSMIC-2A) are expected to be deployed in December 2017 in a low-inclination orbit for dense equatorial coverage, while the second six (COSMIC-2B) are expected to be launched later in a high-inclination orbit for global coverage. To evaluate the potential benefits from COSMIC-2, an earlier version of the NCEP’s operational forecast model and data assimilation system is used to conduct a series of observing system simulation experiments with simulated soundings from the COSMIC-2 mission. In agreement with earlier studies using real RO observations, the benefits from assimilating COSMIC-2 observations are found to be most significant in the Southern Hemisphere. No or very little gain in forecast skill is found by adding COSMIC-2A to COSMIC-2B, making the launch of COSMIC-2B more important for terrestrial global weather forecasting than that of COSMIC-2A. Furthermore, results suggest that further improvement in forecast skill might better be obtained with the addition of more RO observations with global coverage and other types of observations.