National Solar Radiation Data Base User's Manual (1961-1990)

Table of Contents

3.0 Reading and Understanding Data Base Products


Part 2: How the Data Base was Produced

Familiarity with the history of solar radiation measurements in the United States and understanding the source data and the processes used to produce the NSRDB are not necessary for using the hourly data and statistics from the data base. Nevertheless, this information can lead to enhanced use of the data and will reduce the likelihood of misusing the data. You are encouraged, therefore, to become familiar with the subject matter of Part 2 and to refer to it as appropriate for your applications.

4.0 Brief History of Solar Radiation Measurements in the United States

4.1   NOAA'S Solar Radiation Network

From 1951 through 1975 there were about 60 stations in the National Weather Service (NWS) SOLRAD network.3 Each of the stations measured global horizontal solar radiation. Some of these stations made continuous recordings on strip charts; other stations only recorded the daily total energy (insolation) received. Much of these data were digitized at the station, using instructions that changed over the years, and were forwarded to regional centers (later centralized at the National Climatic Data Center [NCDC]). There they were subjected to varying degrees of quality control. These historical solar radiation data contain errors resulting from a host of calibration and instrument problems. When these data were evaluated, only 26 stations, identified with a black dot on Figure 4-1, were deemed suitable for inclusion in a national data base.

During 1976, the network was essentially shut down while new equipment was purchased and installed. From 1977 through 1980, the NWS collected data at the 39 sites identified with a triangle on Figure 4-1. Almost all of these stations recorded both global horizontal and direct normal radiation, and nine of the stations also recorded diffuse horizontal radiation. None of the stations collected data for the entire 48 months, and data for some elements at some stations were collected for less than a year.

During the upgrade of the equipment at the NWS stations, the National Oceanic and Atmospheric Administration (NOAA) established a solar radiation facility at Boulder, Colorado, where all of the pyranometers and pyrheliometers were periodically recalibrated. During the four years from 1977 through 1980, the operation of the network was jointly funded by the Department of Energy (DOE) and NOAA. The quantity and quality of the data collected during these four years was impacted favorably by adequate funding, improved instrumentation, and improved instrument calibrations.

Beginning in 1981, funding was inadequate to fully support the operation of the network. From 1981 through October of 1985 the network gradually decayed, and none of the data were quality controlled and processed for archival and distribution. The data that had been recorded at the stations on cassette tapes were transferred to nine-track magnetic tapes, and copies were shipped to the Solar Energy Research Institute (SERI; now known as the National Renewable Energy Laboratory or NREL). During these years, neither NCDC nor SERI had the resources to process these data. By October 1985, equipment failures had reduced the number of stations to fewer than 10, and NOAA made the decision to close the network to permit another upgrade of the equipment.

It should also be noted that from the spring of 1982 through 1985 no diffuse horizontal data were collected. The diffuse instruments, at the nine stations where they had been in use, were replaced with the Eppley Model 50 pyranometers used prior to 1976. This provided important comparative data, which clearly established the need for additional work to upgrade the pre-1976 global horizontal data. More information on these comparisons is given in Volume 2 - Final Technical Report of the NSRDB documentation.

Essentially no solar radiation data were collected by the National Weather Service from October 1985 through December 1987. During this period, an improved solar tracker was developed, and improved data recording equipment was installed at each station. Beginning in January 1988, the network resumed the collection of global horizontal and direct normal data at most of the 29 sites selected for the upgraded network. Because of a shortage of operational funds, two of the proposed stations for this reduced network, Sterling, Virginia, and Los Angeles, California, never achieved operational status, and some of the operating stations collected less than 50% of all possible hours of data.

As shown in Figure 4-1, 44 stations have participated in the NWS-SOLRAD Network from 1951 through 1990. However, only 16 of these stations have been active during the entire time that the network was in an operational status. Those 16 stations are identified by the circle enclosing the triangle that encloses the black dot. It is worth noting that during the entire period from 1977 through 1990, the solar radiation facility at Boulder, Colorado, continued to maintain calibrations of the pyranometers and pyrheliometers that were in use. Therefore, although the quantity of data varied greatly during these 14 years, the quality of the data that were collected remained relatively good.

Because of the increased interest in solar energy during the mid 1970s, a number of other organizations began collecting solar radiation data. These include data collected by universities and data collected by utilities. Some of these data have been included in the NSRDB.

4.2 The SOLMET/ERSATZ Data Base

The SOLMET/ERSATZ data base (SOLMET Vol. 1 1978 and Vol. 2 1979) for the United States was produced by NOAA and DOE during the latter half of the 1970s. SOLMET refers to the combination of solar radiation and meteorological data. This term is also used to identify the 26 stations that collected global horizontal data from 1951 through 1975. The term ERSATZ refers to the synthetic or modeled solar radiation data that were generated for 222 NWS stations not part of the SOLRAD network. The period of record for most stations in the SOLMET/ERSATZ data base is 23-1/2 years (July 1, 1952 through December 31, 1975).

The uncertainties surrounding the measurement of global horizontal radiation prior to 1976 required a major effort to upgrade the data. In addition to the problems already mentioned, the response characteristics of the Eppley Model 50 pyranometer, used for most stations and most years from 1951 through 1975, had to be addressed. These pyranometers were used without a temperature correction circuit, although their sensitivity was known to change on the average by 0.08% per degree Fahrenheit. Therefore, the sensitivity of these pyranometers could change by as much as 10% from summer to winter at mid-continent northern locations. A universal temperature correction was applied to these data, although it was known that the temperature sensitivity could vary considerably from instrument to instrument.

Furthermore, the response of these pyranometers was known to change as a function of exposure to solar radiation, by as much as 10% to 15% over a period of four to six years (Flowers and Starke 1966). In addition, the response of these instruments was sensitive to changes in the angle and direction of incidence of the solar radiation.

Nontemperature-related errors in the pre-1976 global horizontal data used in the SOLMET/ERSATZ data base were corrected using a technique known as the SYI/CSN procedure. Data for clear solar noons (CSN) were compared with standard year irradiance (SYI) values obtained from model calculations (SOLMET, Vol. 2 1979). In effect, a model was used to calculate global horizontal radiation values under clear sky conditions at solar noon for each of the 26 locations with measured data. These calculations made use of long-term monthly mean precipitable water and turbidity data. Every time a cloudless sky was observed at solar noon, the measured solar radiation was compared with the modeled standard year radiation value. The difference between the measured and modeled values as used to establish a synthetic calibration (correction) factor for the pyranometers. Linear interpolations were used to obtain correction factors for times between the occurrence of CSNs.

The corrected global horizontal data were used to develop clear sky and cloud regression equations for estimating global horizontal data from sunshine, opaque cloud, sky condition, and precipitation data. The coefficients for the regression equations were unique to each of the 26 SOLMET sites. These regression equations were used to fill-in missing data for the 26 SOLMET sites; they were also used to create global horizontal data for 222 ERSATZ sites. The climate at Central Park in New York City was considered to be unique to that location and was not used to estimate global horizontal radiation at any other location. Each of the other 25 SOLMET sites was used to create global horizontal data for a group of ERSATZ stations having similar climate conditions.

Unfortunately, the climates of some of the ERSATZ stations could not be well matched with any of the 25 SOLMET sites. For example, the regression equations for Dodge City, Kansas, were used to estimate solar radiation for Denver and Colorado Springs, Colorado. Dodge City is in the middle of the Great Plains whereas Denver and Colorado Springs are at the foot of the Rocky Mountains. Other questionable matches include Great Falls, Montana, at a latitude of 47.5o, as the reference station for Bettles, Alaska, at 67o north latitude. Also, Seattle is the reference station for both Barrow, Alaska, on the Arctic Ocean, and Redmond, Oregon, at an inland location at an elevation of 940 meters. Nevertheless, there were many good matches between the SOLMET reference stations and the ERSATZ stations.

Direct normal data for all stations were estimated using regression equations. Global horizontal and direct normal data for five stations (Albuquerque, New Mexico; Fort Hood, Texas; Livermore, California; Maynard, Massachusetts; and Raleigh, North Carolina) were used to develop regression equations to calculate direct normal values from global horizontal values (Randall and Whitson 1977). These few direct normal data were collected from 1974 to 1975, with the exception of Albuquerque (1961 to 1964). The regression equations were used to calculate all of the direct normal data for the 26 SOLMET stations for the entire period of record (16-1/2 to 24 years). Similar regression equations were used to calculate direct normal data for Typical Meteorological Year (TMY) data sets for the 222 ERSATZ stations (NCDC 1981).

This brief historical summary of solar radiation measurements and data base developments for the United States reveals shortcomings and limitations that must be considered when using data from the NSRDB. Although the NSRDB has benefitted from improved data and improved models, the uncertainties attached to much of the data are still unacceptably high. The user may want to use the source and uncertainty flags to screen data to be used for critical computations and decisions.


3Although the NWS network has carried different names during the past 40 years, for the sake of clarity and consistency we will refer to it as the National Weather Service Solar Radiation (SOLRAD) Network or NWS-SOLRAD Network, for all time periods.


5.0 Sources of Solar Radiation and Meteorological Data

Table of Contents


Return to RReDC Homepage ( http://www.nrel.gov/rredc )