Jameco Electronics Puzzler Solution: Scientist Are Wrong

By Forrest M. Mims III

How did amateur scientis Joe Novice prove a satellite expert wrong?


The satellite expert who said that the aerosols from a major volcano eruption had dissipated a year or so earlier than expected should have followed Joe Novice's example by simply observing and timing twilight glows. He could have also double checked his satellite instrument with a homemade circuit he could have assembled in half an hour using several components. Here are the details:

1. Brilliant, Colorful Twilights.

Major volcanoes propel sulfur dioxide high into the stratosphere, where it soon forms a blanket of sulfate aerosols that can cover most of the earth. A volcanic aerosol blanket causes twilight glows to be much more brilliant that usual, as in this photo of a stunning twilight glow caused by the 1991 eruption of Mount Pinatubo.

Pinatubo Twilight Glow

If the satellite expert had watched the twilight glows before sunrise and after sunset, he might have realized that his satellite instrument was in trouble. The glow in the photo above is brighter than what Joe Novice was seeing two years after the eruption, but colorful glows were still occurring. They were much like the glows seen around the world for several years after the eruptions of Krakatau in 1883, El Chichon in 1982 and Pinatubo in 1991.

2. Lengthy Twilight Glows.

Volcanoes that inject aerosols into the stratosphere cause twilights to last much longer than normal. Expert sky watchers Aden and Marjorie described how the time between sunset and the setting of the twilight glow indicates the height of volcanic aerosols. Details are in their book Sunsets, twilights, and evening skies (1983, Cambridge University Press, pp. 66-70). If volcanic aerosols are floating through the stratosphere at an elevation of around 20 kilometers, the twilight glow will last around 45 minutes. Joe used his watch to measure the time between sunset and the setting of the top of the twilight glow. He then used the chart below, which is based on the Meinel's method, to estimate the height of the aerosol layer for his latitude. The heights he measured were in the range of those caused by major volcano eruptions.

Height of Aerosol Layer from Twilight Time

3. Reduced Sunlight.

While the two methods above supported Joe's claim that the satellite was wrong, he wanted more proof. So he built a simple sun photometer from an opamp, a gain resistor and a green LED. Here's the circuit he built:

Reduced Sunlight
The LED is a green or red emitter that functions as a photodiode that detects a narrow band of green or red light without the need for a filter. In operation, green or red wavelengths of sunlight create a photocurrent in the LED that's transformed into a voltage by the opamp. Most any single supply opamp will work. For optimum results, use an LT1006. (Jameco stocks the SMT version.) Resistor R1 determines the gain of the amplifier. R1's value depends on the sensitivity of the LED to direct sunlight. Start with a few megohms and reduce the resistance if the output voltage from the opamp saturates in full sunlight. While C1 isn't absolutely necessary, including it will help prevent unwanted oscillation. Use a polypropylene capacitor for best results.

Joe used a file to flatten the end of the LED to make it less sensitive to pointing. He then installed the LED at the end of a hollow tube that served as a collimator that admitted only the direct light from the sun when the collimator was pointed at the sun. He knew the collimator was properly pointed at the sun when its shadow disappeared.

Joe calibrated his sun photometer using the Langley method, in which a series of measurements of direct sunlight is made over the course of a morning as the sun rises in the sky. He then entered into a spreadsheet the natural log (ln) of each measurement and the air mass at which the measurement was made. (The air mass (m) is approximately the reciprocal of the sine of the sun's angle over the horizon (ÿ) or m = 1/sinÿ. Thus, the air mass directly overhead is 1.)

Joe then made a scatter plot of the data, with the air mass being plotted on the horizontal (x) axis and the ln of the signal on the vertical (y) axis. His data points formed a straight line. The intercept with the vertical (y) axis was the natural log of the signal his instrument would have measured at m = 0, which is the top of the atmosphere.

Joe used the extraterrestrial signal inferred from his Langley calibration at m = 0 to determine the optical depth of the atmosphere overhead. Details are given under "Going Further" below.

The bottom line was that Joe measured an aerosol optical depth much higher than that expected for his area. This also supported his conclusion that the satellite was not detecting volcanic aerosols that were still present. Joe sent his findings to the satellite scientist, whose assistant agreed that the satellite instrument was in trouble.

Going Further

Search the Web to see more images of volcano twilights. Paintings of colorful twilights caused by the eruption of Krakatau are in Krakatau 1883, (The Volcanic Eruption and Its Effects by Tom Simkin and Richard S. Fiske) (1983, Smithsonian Institution Press).

More details about the LED sun photometer and its use are here. Many school teachers and students have built the VHS-1 LED sun photometer. A description of global network of LED sun photometers. The development of LED sun photometers is given here.

Background

I began making LED sun photometers in 1989 and was using them every day when Pinatubo erupted on June 15, 19 91. This Puzzler is based on a true story that illustrates my experiences using these and other simple sun photometers to validate satellite measurements from space. Usually, the satellite instruments I have checked are working well. But I've found a string of issues with various ozone satellites, and I also found the aerosol problem described in this Puzzler. The first satellite error I confirmed became my first publication in Nature (F. M. Mims III, Satellite Monitoring Error, Nature, 361, 505, 1993).

Back to the Electronics Puzzler >>