You are here
Seattle Sunlight 2011
How much sun does Seattle get? The short answer is 3.3 on a linear scale where Yakima WA is 4.1, Los Angeles is 4.9 and Las Vegas is 5.7. Germany, the biggest solar market in the world, is generally less than 3. That puts Seattle at 67% of Southern California, better than you might have guessed, and better than Germany.
The long answer is that there is a lot of data available on the internet for folks who like to look things up for themselves. A good starting point is NREL, the National Renewable Energy Lab. They have been researching, collecting, archiving and displaying this kind of information for many years. The numbers above for example came from their archive of 30 year averages for 1961-1990. Other NREL resource links lead to a variety of sources, for example an interactive map called Solar Prospector where you can zoom in to your exact location and choose from a variety of data configurations.
When doing comparisons be sure to use the same configurations. For example the short answer numbers above are Global Horizontal Irradiance in kWh/m2/day. This is the most commonly measured value since it requires only a single sensor fix mounted horizontally to the ground, typically a pyranometer. Other ground measurements are possible such as Global at the Latitude Tilt Angle, Direct Normal and Global Diffuse Irradiance. Especially be aware that NREL's Prospector and PVWatts interface show Tilt at Latitude values by default. In recent years satellite data has been added and computer algorithms developed that allow accurate solar resource estimates anywhere on the globe for any array orientation including fixed tilt and tracking. Programs that go on to predict solar array performance, such as Retscreen, SAM and PVWatts use climate data files that include this solar resource information along with wind, temperature and component models. Its fairly easy to extract solar data from these programs with drop down location selections and they are all free downloads.
Because the weather is so variable, solar and climate resource data is typically averaged over a number of years. However, you can dig deeper and find raw sensor data in intervals as short as 1 minute at some locations, notably the national network's MIDC weather stations. In Washington State we do have an Agricultural Weather Network hosted by WSU with a surprising number of GHI pyranometers with 15 minute granularity that can be accessed after creating a login account. While AWN is more rural-centric there is 1 site in King County at 21 Acres, just north of Redmond.
Of special interest for Seattlites, there is a very well equipped NOAA ISIS station on the shores of Lake Washington at Sand Point, pictured left, that includes a sun tracking mount with a total solar pyranometer, a diffuse pyranometer, and a UV-B Biomete. These instruments have been collecting Global, Direct and Diffuse Solar Radiation data in 3 minute intervals since 2000. That's a lot of data. Fortunately the University of Oregon Solar Radiation Monitoring Lab has been collecting, managing and reducing this datastream and making the summary data available on their web site in various formats including for example Daily Totals and Monthly Averages.
Other University of Oregon data sites in Washington State include Cheney, Richland and White Bluff. They also have a bunch of sites in OR along with some others in ID, MT, UT and WY.
The reason I got interested in the actual amount of sun this year as opposed to average values was to confirm my suspicion that 2011 was the worst July for sunshine that I can ever remember. The chart below compares monthly average values for 2011 against the previous 11 years, from 2000 through 2010. Sure enough 2011 is the lowest July value in this 12 year dataset. I don't expect much sun in June but 2011 was the low point for this month too. In fact the entire 1st half of the year except for May has been on the low side of the envelope.
Actual values for sun are useful for correcting measured system performance to estimates that use average values. For example you might be concerned if your system underformed by 20% compared to what you expected from PVWatts or one of the other prediction programs. By looking at the nearest measured data you might find that the actual solar resource was in fact less than the average by that very amount. The next chart shows how 2011 compares to the prior 11 year average and to expected values from Retscreen, SAM and PVWatts for the same location, Sand Point WA. In this case June is 13% lower than the averages and July is 8.5% lower. Residential customers might not notice the shortfall at these relatively low values but commercial and utility sites would probably be interested to know that their month to month shortfalls were just due to less sun than expected rather than hardware problems.
Notice too in this chart how close the modeling program climate files are to the actual measured 11 year averages. I notice PVWatts seems a little lower in July. Retscreen actually has a drop down choice for Sand Point in their menu structure after choosing USA->WA. SAM has default TMY2 files for Olympia, Qullayute, Seattle, Spokane and Yakima, so their Seattle would the closest choice to Sand Point.
SAM also links to resources for additional file choices including TMY3 and U.S. sattelite-derived data (10 km grid cells in TMY2 Format). The TMY3 link goes to an alphabetical list by state and city that includes 29 cities in WA. The three nearest Seattle are Boeing Field, Sea-Tac airport and Renton Municipal airport. You can download these CSV files and add them to the SAM climate directory. There are some choices in the International EPW files for North America->USA->WA but these are the same TMY3 files. The satellite data link takes you to NREL's Solar Prospecter launcher. 1 click there takes you to an interactive map where you can move to any location by zooming and panning a map on the screen, including for example Magnuson Park, next to the site where the Sand Point NOAA site is on Lake Washington. You can also enter a location by street address or Lat/Long (decimal degrees). In our case this is 47.65°N, 122.25°W. The main menu includes a selection to Download which brings up a choice of files for various years or for a typical year based on averages. These are compressed as GZ files so you have to extract after downloading before putting it in SAM climate directory.
The various SAM files produce a little different results as shown in the next chart.
Besides revealing the expected solar resource the prediction programs go on to calculate the expected energy production in kWh and in some cases provide comprehensive resources for cost analysis. While this is a subject for another article the usual result for energy production in Western Washington winds up being about 950-1000 kWh/kW. That is if you install a 5kW PV system on your house you can expect to get up to ~5000kWh of annual energy production, enough to offset a lot of your power bill. Sunny Yakima would be 25% better.