Improving Solar Radiation Improving Solar Radiation ForecastingForecasting

Betsy Weatherhead, U. ColoradoBetsy Weatherhead, U. Colorado

Craig S. Long, NOAA/NWS/NCEP/CPCCraig S. Long, NOAA/NWS/NCEP/CPC

Kathy Lantz, U. ColoradoKathy Lantz, U. Colorado

Patrick Disterhoft, U. ColoradoPatrick Disterhoft, U. Colorado

AMS Summer Meeting, August 12, 2009

Needs of Renewable Energy Needs of Renewable Energy CommunityCommunity

Solar radiation can reduce the Solar radiation can reduce the demand on the grid (personal demand on the grid (personal Photo-voltaic panels). Highly angle Photo-voltaic panels). Highly angle dependent. Total Solar Radiation.dependent. Total Solar Radiation.

PV plants depend on total solar PV plants depend on total solar radiation.radiation.• Fixed angleFixed angle• Single axisSingle axis• Full rotationFull rotation

Concentrated Solar Energy highly Concentrated Solar Energy highly dependent on direct beam radiationdependent on direct beam radiation• Various forms of energy storageVarious forms of energy storage

Time scales from seconds to Time scales from seconds to decades with a strong focus on decades with a strong focus on minutes to days.minutes to days.

Solar Radiation - BasicsSolar Radiation - Basics Solar radiation reaching the surface of the Earth Solar radiation reaching the surface of the Earth

is affected by many factors:is affected by many factors:• Sun angleSun angle• CloudsClouds• AerosolsAerosols• Surface reflectivitySurface reflectivity• Water vaporWater vapor• OzoneOzone

All of these factors can change dynamically on a All of these factors can change dynamically on a day-to-day basis.day-to-day basis.

Each part of the solar spectrum is affected Each part of the solar spectrum is affected differently by these factors.differently by these factors.

Technologies are Technologies are still developingstill developing

Wavelength Wavelength matters.matters.

Clouds, aerosols, Clouds, aerosols, ozone, water ozone, water vapor all matter.vapor all matter.

Multiple paths to making solar Multiple paths to making solar forecastsforecasts

ProprietaryProprietary Data extrapolationData extrapolation Statistical modelingStatistical modeling Physical modelingPhysical modeling Ensemble forecastsEnsemble forecasts Data assimilation, rapid refreshData assimilation, rapid refresh

Rapid Updatefor Aviation, Severe Weather

Air Quality

WRF NMMWRF ARW

WRF: NMM+ARW ETA, RSM GFS, Canadian Global Model

MostlySatellite+Radar

North AmericanMesoscaleWRF NMM

North American Ensemble Forecast System

Hurricane GFDLHWRF

GlobalForecastSystem

Dispersion

ARL’s HYSPLIT

Forecast

Severe Weather

ClimateCFS

~2B Obs/Day

Short-RangeEnsemble Forecast

NOAA’s NWS Model Production SuiteNOAA’s NWS Model Production Suite

MOM3

NOAH Land Surface Model

Coupled

Global DataAssimilation

OceansHYCOM

WaveWatch III

NAM+CMAQ

Courtesy Stan Benjamin

Forecasting Solar UV RadiationForecasting Solar UV Radiation

UV Forecasts have UV Forecasts have taken place for fifteen taken place for fifteen years in support of years in support of human health.human health.

NOAA, various weather NOAA, various weather services, Accu-Weatherservices, Accu-Weather

Private monitoring Private monitoring offers Nowcasts.offers Nowcasts.

NCEP 5-Day UV PredictionsFive Successive Days

UV Index Algorithm UV Index Algorithm Old (Old (virtually unchanged since 1994virtually unchanged since 1994))

• Weight UV relationship w/total ozone and SZA derived from Weight UV relationship w/total ozone and SZA derived from delta Eddingtondelta Eddington RT model. RT model.• Total ozone fcsts from GFS.Total ozone fcsts from GFS.• Surface UV albedo held constant at Surface UV albedo held constant at 5%,5%, (no allocation for snow or ice cover)(no allocation for snow or ice cover)..• Aerosols: AOD is Aerosols: AOD is constantconstant at 0.2 and SSA = 1.0. at 0.2 and SSA = 1.0.• Elevation increases UV flux by ~6% per km.Elevation increases UV flux by ~6% per km.• UVI clear sky field on 1x1 grid up to this point.UVI clear sky field on 1x1 grid up to this point.• Relationship between Relationship between NGM MOS cloud probNGM MOS cloud prob and cloud attenuation of UV flux. and cloud attenuation of UV flux.• Cloudy UV Index values at all (~800) MOS locations.Cloudy UV Index values at all (~800) MOS locations.• Bulletin, Eta grid 211 and 207 are created by interpolation from MOS cities.Bulletin, Eta grid 211 and 207 are created by interpolation from MOS cities.• Products are limited to next day solar noon over US.Products are limited to next day solar noon over US.

NewNew• Weight UV relationship with total ozone, and SZA derived from Weight UV relationship with total ozone, and SZA derived from NCAR-TUVNCAR-TUV RT model. RT model.• Total ozone fcsts from GFS.Total ozone fcsts from GFS.• Surface UV albedo held constant at Surface UV albedo held constant at 3%3% for non-snow conditions. for non-snow conditions.

GFS albedo is used when > 30% (where snow is possible).GFS albedo is used when > 30% (where snow is possible).• Aerosols: AOD and SSAAerosols: AOD and SSA vary vary, derived from , derived from seasonal global climatology grids.seasonal global climatology grids.• Elevation increases UV flux by ~ 6% per km.Elevation increases UV flux by ~ 6% per km.• Ratio of Ratio of GFS down-welling UVB flux w & w/o cloudsGFS down-welling UVB flux w & w/o clouds determines cloud attenuation. determines cloud attenuation.• Output remains on gaussian grid.Output remains on gaussian grid.• GFS inputs available at 3hr interval.GFS inputs available at 3hr interval.• UV Index products are outputted globally at 1 hr intervals out to 96 hoursUV Index products are outputted globally at 1 hr intervals out to 96 hours

UV Index : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

- f19

Satellite IR Clouds obs at 11:45 UTC Oct 12, 2003

- f19

Bondville

25 mWatts/sq m = 1 UV Index unit

__ Diurnal trace of observed UVx 1 hour mean about 18 UTC| St Dev of obs +/- 0.5 hr about 18 UTCo 18 hr FCST from 00Z run valid at 18 UTC

40.04N, 88.36W

Ery

them

al D

ose

Rat

e (m

W/s

q m

)

Improvements in NWS’ solar UV Improvements in NWS’ solar UV forecastforecast

Improvements in cloud and ozone Improvements in cloud and ozone forecasts result in a much improved forecasts result in a much improved solar forecast.solar forecast.

Verification depends strongly on time Verification depends strongly on time period of averaging when single period of averaging when single locations are considered.locations are considered.

Significant improvements in solar Significant improvements in solar radiation forecastsradiation forecasts

Cloud modeling has improved Cloud modeling has improved significantly over the past twenty significantly over the past twenty years.years.

Inclusion of “cloud optical depth” and Inclusion of “cloud optical depth” and “cloud height” are of considerable “cloud height” are of considerable importance to assessing importance to assessing transmission.transmission.

Aerosol forecasts are currently under Aerosol forecasts are currently under consideration in NWS.consideration in NWS.

Erythemal Dose Rate - Bondville, IL - Oct 2003

0

25

50

75

100

125

274 277 280 283 286 289 292 295 298 301 304

Day of Year

Ery

thm

al D

ose

Rat

e (m

W/s

q m

)

SURFRAD New UVI-1 Current UVI25 mWatts/sq m = 1 UV Index unit

1 Day Forecast Verification

Erythemal Dose Rate - Bondville, IL - October 13 & 14, 200340.04N, 88.36W

0

25

50

75

100

125

286.0 286.5 287.0 287.5 288.0

Day of Year

Ery

them

al D

ose

Rat

e (m

Wat

ts/s

q m

)

OBS UVI Clear UVI w/clouds

25 mWatts/sq m = 1 UV Index unit278 DU 309 DU

Fine temporal structure is extremely difficult.

Renewable Energy NeedsRenewable Energy Needs

Improved forecasts are necessary for optimal use of solar energy Improved forecasts are necessary for optimal use of solar energy integration.integration.

Multiple factors must be considered to make accurate forecasts: Multiple factors must be considered to make accurate forecasts:

cloudsclouds, aerosols, water vapor., aerosols, water vapor. NWS currently makes solar radiation forecasts for 57 locations for UV NWS currently makes solar radiation forecasts for 57 locations for UV

index. Significant improvements have been achieved using state of the index. Significant improvements have been achieved using state of the art cloud forecasting.art cloud forecasting.

Fundamentally, improvements require accurate measurements of Fundamentally, improvements require accurate measurements of both direct and diffuse radiation.both direct and diffuse radiation.

Actual hourly variation of solar energy potential- accurate forecasts are essential

Long-term estimates of solar radiation will need to take into multi-year factors including el nino, solar cycles, NAO, changes in climate, including in changes in the jet stream.

Predictions of long-term changes will require research into past long-term datasets as well as best available climate models.

Long-term Predictions

Foundation for Model Foundation for Model Improvement--MeasurementsImprovement--Measurements

Surface Radiation monitoringSurface Radiation monitoring Direct beam separate from total solar radiation.Direct beam separate from total solar radiation. DOE’s Atmospheric Radiation Measurement, DOE’s Atmospheric Radiation Measurement,

National Renewable Energy LaboratoryNational Renewable Energy Laboratory NOAA’s SurfRad, Central Calibration FacilityNOAA’s SurfRad, Central Calibration Facility USDA’s UV Monitoring USDA’s UV Monitoring

NetworkNetwork Direct beam measurements Direct beam measurements

are difficult, but not are difficult, but not impossible.impossible.

NOAA’s SurfRad Network

Satellite IR Clouds obs at 11:45 UTC Oct 12, 2003

SummarySummary

Data are the foundation for Data are the foundation for improving forecasts.improving forecasts.

Current solar monitoring is Current solar monitoring is inadequate, but can be easily inadequate, but can be easily adapted and expanded.adapted and expanded.

Multiple modeling approaches will be Multiple modeling approaches will be needed to address the multiple needed to address the multiple needs of the renewable energy needs of the renewable energy communities.communities.