From the Watertown Daily News, and the “law of unintended consequences” department comes an inconvenient truth from the National Weather Service, that upon further investigation appears to be a nationwide problem for the WSR-88D doppler weather radar network used to predict, track, and analyze severe weather. According to NOAA’s Radar operations center, forecasters are faced with “little or no workaround”.
Part of the reason is that the WSR-88D national deployment in the early to mid 1990’s preceded the mass deployment of wind turbines to provide “green energy”. They had no way of knowing then that their field of view would be polluted by an army of rotating blades.
h/t to John Droz for the Watertown Daily News article below.
Document from the National Weather Service lists possible radar interference impact from wind turbines
WATERTOWN — A new document from the National Weather Service expands on potential interference with the weather radar in Montague, used by personnel at Fort Drum.
In addition to the Buffalo and Burlington, Vt. weather stations which cover Jefferson, Lewis and St. Lawrence counties, the document lists impacts to the Albany and Binghamton National Weather Service stations, which also use the Montague KTYX radar.
Among the possible concerns listed in the document for the Binghamton station is that beam blockage from wind turbines could hamper tracking of thunderstorms in Oneida and Madison counties, delaying tornado warnings. It could also make it difficult to track lake effect snow and rainfall, which in turn could delay travel advice and flash flood warnings.
For the Albany station, the document said that clutter from turbines could create false storm identification and tracking over Lewis and northern Herkimer counties, as well as possibly masking lake effect snow.
The document was sent out by Jessica A. Schultz with the National Oceanic and Atmospheric Administration and lists the possible impact of wind turbines for the four NWS weather stations that use the weather radar. According to documents published online by NOAA, Ms. Schultz works at the NOAA NWS Radar Operations Center in Oklahoma. Although the document itself is unsigned, the document’s properties list the author as being a JSchultz.
There are 4 National Weather Service (NWS) offices that use the Fort Drum KTYX radarto accomplish their mission of protection of life and property in the nearby counties.These offices are: NWS Albany, NWS Buffalo, NWS Binghamton, NWS Burlington.
NWS Albany Impacts:
The turbines can cause beam blockage and under-sampling of the radar echoesdownstream (25-30 nautical miles) into northern Herkimer County.
Precipitation underestimate is likely (warm and cool season) in lake effect andwidespread precipitation events.
Turbines can also cause partial beam blockage impacting dual polarization data,and display large amounts of erroneous data.
Downstream turbine clutter can impact precipitation data by over/underestimation, incorrect wind speed data, and false storm identification andtracking over Lewis County before moving into northern Herkimer County.
In the winter, lake effect snow features could be masked or underestimated,negatively impacting warnings and advisories.
During severe weather, erroneous data (especially wind velocities) can impactearly detection and warnings of high winds, hail, and tornadoes
NWS Buffalo Impacts:
The height of existing turbine towers and turbines’ spinning blades are causing beam blockage and under-sampling of the radar echoes downstream for Jefferson,Lewis, and Oswego counties.
Resulting precipitation estimates in the vicinity of turbines are not useable, while precipitation estimates downstream have been degraded.
Turbines are causing partial beam blockage impacting dual polarization products.This results in large amounts of erroneous data.
Additional turbine installations will nearly surround the radar, further exacerbating these issues and will make radar interpretation and the detection ofsevere weather increasingly difficult.
Wind turbine clutter has a negative impact on several radar capabilities:
o Precipitation estimation algorithms produce false estimates.
o Velocity products are often not useable near the turbines, particularly during severe weather.
o This can cause false and/or missed detection of tornadoes by radar algorithms and forecasters.
Thunderstorm or winter storm characteristics will be further masked or misinterpreted, reducing warning effectiveness in the vicinity of and downrangeof existing and future wind turbines.
False signatures contaminating Doppler velocity data will further reduce forecasters’ situational awareness, especially during hazardous weather events.
Potential radar relocation, particularly east or northeast from the current location will further reduce radar coverage south of Lake Ontario from Monroe, Wayne and Cayuga counties, with completely unseen lake effect events by radar. In any move, beam blockage will continue to be an issue near and over the Tug Hill Plateau.
NWS Binghamton Impacts:
The beam blockage will hamper our abilities to detect thunderstorm circulations in Oneida/Madison Counties and hence tornado warnings could be delayed. It is important that we have good radar coverage in Oneida/Madison Counties because there is a local maximum in tornadoes in these areas since the Mohawk Valley will often skew winds to the southeast leading to increased atmospheric rotation.
Thunderstorm or winter storm characteristics will be further masked or misinterpreted, reducing warning effectiveness in the vicinity of and down range of existing and future wind turbines.
False signatures contaminating Doppler velocity data will further reduce forecasters’ situational awareness, especially during hazardous weather events.
The beam blockage could also significantly hamper our ability to forecast and detect lake effect snow. Oneida County (especially northern Oneida County) sees more than 200″ of snow per year on average and is one of the snowiest places east of the Rockies. The beam blockage could affect our ability to detect lake effect snow along the NY State Thruway between Syracuse and Utica of which is a major travel corridor. Our office provides almost daily briefings to the NYS Thruway Authority when a lake effect snow pattern is present. Significant beam blockage could erode our ability to time and track heavy lake effect snow bands that severely impact travel which would lead to less accurate decision support to the Thruway Authority.
Oneida and Madison Counties have a history of severe local flash flooding and beam blockage will hamper our ability to accurately estimate rainfall in thesecounties which would negatively impact the timeliness of flash flood warnings.
NWS Burlington Impacts:
Wind turbines close to the radar close to the radar cause some uncertainty/confusion about actual storm characteristics while monitoring storms that are moving north or northeast. This can delay warnings, resulting in a lower lead time prior to the storm reaching St. Lawrence or Franklin Counties.
The wind turbines “look” like precipitation, even on a clear day. This can cause confusion to users of the data, including the media, pilots, and general public.
If the radar is forced to be relocated because of wind turbines, concerns would be magnified. Any move to the east or south of the current location would result reduced radar coverage over St. Lawrence and Franklin Counties. This would mean poorer detection of lake effect snow and low level severe weather features,such as tornadoes, high winds, and hail.
It seems the issue isn’t limited to New York, the NWS has also been investigating wind turbine impacts on weather radars in the midwest. From the NOAA WSR-88D Operations center:
HOW ROTATING WIND TURBINE BLADES IMPACT THE NEXRAD DOPPLER WEATHER RADAR
Rotating wind turbine blades can impact the radar in several ways. Wind turbines can impact the NEXRAD radar base data, algorithms, and derived products when the turbine blades are moving and in the radar’s line of sight (RLOS); and, if turbines are sited very near to the radar their large nacelles and blades can also physically block the radar beam or reflect enough energy back to the radar to damage the radar’s receiver hardware.
Radar Receiver: The NEXRAD radar has a very sensitive receiver. The radar’s Receiver Protector prevents damage from strong reflected signals; however its upper limit is 53 dBm. Large objects sited very near the radar (< 4 km), such as turbine nacelles, have the potential to return signals that exceed the limit of receiver protector and render the radar inoperable.
Beam Blockage: If sited within a few kilometers of the radar, wind turbines can partially or fully block the radar beam. This beam blockage attenuates the strength of the beam and impacts data beyond the wind farm, causing shadows or spikes in the data through the entire range of the radar (460 km for reflectivity data, and up to 300 km for velocity and spectrum width data).
Radar Base Data: Turbines in RLOS can reflect energy back to the radar and visually contaminate the reflectivity, velocity, and spectrum width data. Forecasters look for certain “signatures” in the data that indicate the severity of the storms. The wind farm clutter can sometimes look just like showers and thunderstorms, or can alter the appearance of a storm (e.g. hook echoes). This visually corrupted data adds uncertainty to the analysis and could cause forecasters to delay/miss a severe weather warning or to warn unnecessarily.
Algorithms and Derived Products: The base reflectivity, velocity, and spectrum-width data are also used by many algorithms in the radar processor to detect certain storm characteristics, such as mesocyclones, relative storm motion, hail, turbulence, etc. Corrupted base data can cause the radar algorithms to generate false alerts or to miss alerts. The radar also generates many additional products using this base data, such as wind profiles and rainfall estimates. Wind turbine clutter can impact the accuracy of these derived products.
The graph below depicts the relative impact of wind turbines (or wind farms) on NEXRAD radars and forecasters as a function of distance (on level terrain) if wind turbines are in the RLOS.
Impacts increase greatly as wind turbines are sited closer to the radar, especially within 18 km (assuming level terrain), as radar operator workarounds become more difficult. Turbines sited at least 18 km from the radar generally only impact the lowest radar scan at 0.5 degrees elevation, and clutter is confined to the wind farm area. Within 18 km wind turbines cause additional impacts including: clutter on multiple elevation scans above 0.5 degrees, multipath clutter down range of the wind turbines, and greater impacts to radar algorithms. Multipath scattering from wind turbines can extend the contaminated data up to 40 km beyond the wind farm. Turbines sited within 4 km of the radar may also cause significant (>10%) attenuation/blockage of the radar beam impacting data throughout the entire range (460 km-reflectivity, 300 km-velocity) of the radar. When turbines are sited within 200 m, construction or maintenance personnel may be exposed to microwave energy exceeding OSHA (Occupational Safety and Health Administration) thresholds. The above distances assume a level terrain and a Standard Atmosphere Index of Refraction profile. Therefore, actual impacts may occur closer or further away from the radar than this chart indicates depending on the terrain and current atmospheric refraction. Accurate determination of the RLOS and impact distances requires a detailed site-by-site analysis.
You may wonder why we can’t filter out this clutter since we know where the wind farms are located. The NEXRAD has a sophisticated clutter removal scheme. Since weather is usually in motion, the scheme was designed to filter returns that have essentially no or very low motion. This is effective for removing the returned signals from terrain, buildings, and other non-moving structures. However, the radar sees rotating wind turbine blades as targets having motion, hence processes these returns as weather. At this time there is no filtering scheme available to identify and remove wind turbine clutter while preserving real weather returns.
Wind turbine clutter has not had a major negative impact on forecast or warning operations, yet. However, with more and larger wind turbines coming on line, radars in some parts of the country will have multiple wind farms in their line of sight. Cumulative negative impacts should be anticipated – which, at some point, may become sufficient to compromise the ability of radar data users to perform their missions.
Examples of Wind Turbine Clutter
Zoomed-in Display of WTC-contaminated data from Fort Drum NEXRAD
The image above is a zoomed 0.5 degree elevation Reflectivity product from the Ft Drum, NY NEXRAD. There is a large wind farm nearby with turbines oriented from due north through southeast of the radar. The turbines are close enough (within 18 km) to cause spurious multipath scattering that extends well beyond the wind farm and contaminates data at multiple scanning elevation angles.
Display of WTC-contaminated data from the Dyess AFB, TX NEXRAD
Sequence (left to right) of 0.5 deg reflectivity images showing thunderstorms developing over a wind farm (purple rectangle) 18–30 km (10-16 nm) west of Dyess AFB, TX WSR-88D. Left: thunderstorms have not yet developed, high reflectivity values due to wind turbines alone. Middle and Right: storm has developed to where in right image a distinct notch structure, indicative of severe weather, formed – note: turbine and weather echoes indistinguishable.
This radar-estimated Storm Total Precipitation accumulation product from the Dodge City, Kansas NEXRAD on April 22, 2010 at 1403 GMT depicts how wind farms can impact radar-derived products. Erroneous 4+ inch radar-estimated Storm Total Precipitation accumulations (indicated by the arrows) in the image on the left are due to wind farms northeast and southwest of the NEXRAD. The anomalous accumulations make estimates of rainfall over an area/river basin more difficult to determine. However, radar operators can apply exclusion zones to mitigate these anomalous accumulations, as seen on the right. (Radar precipitation algorithms do not use the returns from the exclusion zone to accumulate precipitation.)
Dodge City, KS NEXRAD (KDDC) reflectivity (upper right) and mean radial velocity (lower right) imagery for 0150 UTC on 23 Feb 2007 showing two wind farms within the radar’s line of sight. The yellow area in the upper left image depicts areas where the radar line of sight is within 130 m of the ground. The reflectivity and velocity values are anomalous and can confuse users. The lower left panel shows the effects of the wind farm to the southwest whose influence has resulted in a false tornado alert generated by the NEXRAD algorithms. The Weather Forecast Office did not issue a warning because, in this case, other data indicated that there was no severe weather in the wind farm area.
In 2016, the NWS produced a plan to prevent bad siting of future weather radar installtions in proximity with wind farms:
Based on the wind farm proposal the ROC receives, the ROC provides a case-by-case analysis of potential wind farm impacts on WSR-88D data and forecast/warning operations. The ROC uses a geographic information system (GIS) database that utilizes data from the Space Shuttle Radar Topography Mission to create a RLOS map with delineated areas corresponding to a turbine height of 160 m AGL. Multiple radar elevation angles are considered for projects close to the radar.
The ROC then performs a meteorological and engineering analysis using: distance from radar to turbines; maximum height of turbine blade tips; the number of wind turbines; radar azimuths impacted; elevation of the nearby WSR-88D antenna; an average 1.0 degree beam width spread; and terrain (GIS database). From this data the ROC determines if the main radar beam will intersect any tower or turbine blade based on the Standard Atmosphere’s Refractive Index profile.
Finally, the ROC estimates operational impacts based on amount of turbine blade intrusion into RLOS, number of radar elevation tilts impacted by turbines, location and size of the wind farm, number of turbines, orientation of the wind farm with respect to the radar (radial vs azimuthal alignment), severe weather climatology, and operational experience. The ROC also compares the wind farm to other operational wind farms to estimate impacts.
The problem is being addressed through a NWS training course, which is open to the public here:
Here they talk about the problem, when blades are turning, algorithms can’t remove the false signal. But when blades are stationary, they can. The problem is that Doppler radar is designed to detect motion, or more specifically, storm motion.
And their worst nightmare:
And there’s more. In 2007, a presentation was made about the Weatherford wind farm in Oklahoma:
Weatherford Wind Farm Blue Canyon Wind Farm Appearance of OK Wind Farms Varies with Time and Radar Beam Propagation.
Note the two stationary echoes in the left and lower left parts of the animation. These are the Weatherford (along the Interstate) and Blue Canyon wind farms
While it may be news to Watertown, NWS offices around the country have been dealing with this problem for awhile. It should be noted that there’s not one document citing radar interference from coal, nuclear, or hydroelectric power plants.