NOAA Technology Transfer Award Archive

Melinda Marquis, Curtis Alexander, Stan Benjamin, John Brown, James Wilczak, Robert Banta, and Allison McComiskey

Melinda Marquis, Curtis Alexander, Stan Benjamin, John Brown, James Wilczak, Robert Banta, and Allison McComiskey

Oceanic and Atmospheric Research

Author: Derek Parks/Tuesday, December 12, 2017/Categories: 2017 Awards

During the last five years, NOAA researchers across all ESRL divisions worked with the Department of Energy (DOE) and the private sector to develop foundational improvements in wind and solar forecasts in the NOAA hourly-updated 13-km Rapid Refresh (RAP) and 3-km High-Resolution Rapid Refresh (HRRR) models to support the renewable energy industry and the larger energy industry. These model changes allow improved decision-making for power generation from different sources.  They also benefit all users of the RAP and HRRR models, including those in aviation, surface transportation, severe weather, and fire weather. Feedback from private-sector stakeholders indicates the value of these improvements. 

In the Wind Forecast Improvement Project (WFIP-1), the NOAA ESRL ASRE program partnered with private forecasting companies, WindLogics and AWS Truepower, and DOE to develop more accurate methods for wind forecasts. NOAA temporarily (2011-2012) installed instruments in the Great Plains and Texas to collect data. Additionally, NOAA established new relationships with energy companies, who voluntarily began sharing proprietary observations collected on wind farms with NOAA under non-disclosure agreements. ASRE model developers improved model physics in the RAP and HRRR, particularly with regard to modeling the boundary layer (the lowest 1-2 km of the atmosphere) more accurately. Results of these model improvements and additional meteorological observations led to a rough 15% reduction in error in forecasts of power production in the first four hours after the forecast is made. 

        In the Solar Forecast Improvement Project (SFIP), the ASRE program partnered with DOE, NCAR and IBM to develop more accurate methods for producing solar forecasts. The program modified the RAP and HRRR models so that, for the first time, they began providing forecasts of the three relevant components of irradiance: direct normal irradiance (DNI); diffuse horizontal irradiance (DHI); and global horizontal (GHI), which are all useful for the power sector. Further relationships were developed in this project, which led to additional sharing of private-sector observations with NOAA. 

In the Wind Forecast Improvement Project 2 (WFIP-2), ESRL ASRE scientists are currently working with DOE and Vaisala to improve forecasts of turbine-height winds in areas of complex terrain, specifically in the Columbia River Gorge and Basin. The scientific goals of WFIP-2 include improving understanding of physical processes, such as atmospheric stability, turbulence, and the low-level jet. An emphasis is on improving the RAP and HRRR model physics and parameterizations of processes that affect low-level winds in complex terrain. Areas of focus include the mass flux and turbulent kinetic energy schemes, shallow cumulus cloud scheme, and a wind farm parameterization (that models the effects of wind turbines on the atmosphere). Still more new relationships were developed in this project, which led to additional sharing of private-sector observations with NOAA. “NOAA’s continual advancement of forecast skill improves the stability of our electric grid, and directly reduces the cost of integrating more renewable energy into our resource mix,” said Dr. Jim McCaa, Manager, Vaisala Energy Advanced Applications.

Better forecasts enable plant operators to better predict how much electricity their wind turbines or solar collectors will generate, leading to more wind and solar power being used and reduced overall energy costs. “Our work (WFIP-2) will examine a wide range of meteorological phenomena that impact the operation of wind farms in mountainous terrain worldwide and improve the short-term predictability of wind energy, ultimately reducing the challenges of wind energy integration not just in North America, but also in emerging markets around the globe,” said McCaa.

“Improvements to weather-prediction models that produce higher-resolution forecasts with greater frequency are helping Xcel Energy and the electric power industry cost effectively and reliably integrate higher levels of renewable generation. We can also use more accurate weather forecasts to reduce costs and improve reliability by proactively sending crews to respond to storm damage before it occurs,” said Drake Bartlett, Xcel Energy.

“Improvements that have been made within the HRRR model over the years have been important factors to assist with real-time solar and wind forecasting accuracy.  One example, was in relation to HRRR v2 changes were made to the downward shortwave flux at the surface assisting with the model bias that existed previously of having an average excess of ~80 to 100 W/m2 of incoming shortwave radiation.  More changes have been made to the model since then, making this model an important tool to use for more real time forecasting for the energy industry.” - Amber L. Motley, Manager, Short Term Forecasting, California Independent System Operator (ISO).

“UVIG provides an annual forum to bring together weather modelers and forecasters with the utility organizations that use their products.  NOAA has become a permanent fixture in that forum due to the indispensable work they are doing for the industry on the RAP and the HRRR for wind forecasting improvement.  UVIG has recognized NOAA for its efforts on this work with its annual Achievement Award in 2013 and 2015.” - Charlie Smith, Utility Variable generation Integration Group (UVIG) Executive Director.

Model improvements, including enhancements to physics, parameterizations, and data assimilation, resulting from the WFIP1 and SFIP were transferred to operations at NCEP on August 23, 2016 in the RAPv3 and HRRRv2. These forecast model improvements significantly reduced the warm dry bias seen at the surface. The data assimilation improvements included canopy water cycling, innovations using pseudo observations for temperature in the boundary layer, and more consistent use of surface temperature and dewpoint data. There was reduced (excessive) shortwave radiation in the microphysics scheme, improved mixing length parameterization, and coupling of boundary layer clouds to the radiation radiative transfer model, and an improved boundary layer scheme. Finally, the land surface scheme included a reduced wilting point for more accurate modeling of transpiration.  


Number of views (4167)/Comments (0)



1315 East-West Highway
Silver Spring, MD 20910


Our mission is to foster preeminent science and technological innovation through federal investments in research and development (R&D), partnerships or licensing opportunities at NOAA.