ASTR 214 - General Astronomy Laboratory:
Radio Jove Data Analysis

In this laboratory exercise, you will analyze data taken at WKU's remote Radio Jove receiving station at Bell Observatory. Please read the following instructions carefully. Write your answers on separate paper to hand in. Include any graphs or other calculations as well.

Radio Jove is a NASA-supported project for education and outreach in radio astronomy. It uses fairly modest equipment to detect radio emissions from the planet Jupiter, from the Sun, and from our Galaxy.

The WKU Radio Jove experiment has been run for several ASTR 214 classes in the past few years. Its two dipole antennas capture radio emissions at a frequency of about 20 MHz (wavelength of 15 m) and translate it into a (very weak) electrical current, which is conducted along the coaxial feed line cables to the receiver, where it is amplified and converted into an audio-frequency signal that is recorded digitally. The data are retrieved every few days and then converted into an uncalibrated estimate of the received power by squaring the audio waveform signal (a measure of voltage vs. time) and computing its average over small, equal time intervals. Your task is to examine these data and report your findings for each of the questions below.

  1. The following links give some overview plots of uncalibrated received signal power vs. time. The power is plotted on a logarithmic scale, with power labeled every factor of 10 and unlabeled tickmarks shown for intermediate linear steps in power (e.g., 103, 104, 105 labeled, plus unlabeled ticks for 2×103, 3×103, ..., through 9×103, then 2×104, 3×104, ..., through 9×104, etc.).

    Each linked page shows a series of thumbnail plots, one per 24-hour day. Clicking on a thumbnail plot will show a larger version. Each 24-hour plot starts at 06:00 Universal Time (UT) = midnight Central Standard Time (CST).

    Each plot is labeled with the range of dates when it was observed. The vertical axis is uncalibrated received power as described above. The horizontal axis is Universal Time (a.k.a. Greenwich Mean Time) in hours, where midnight of the first "day" shown is 0, midnight of the second "day" is 24, etc. Central Standard Time (CST) is 6 hours behind UT. Although the radio equipment does not detect visible light, the "daytime" signal is generally much stronger than the "nighttime" signal, because the Sun excites the ionosphere during the day, allowing it to reflect undesirable contaminating signals, called "radio-frequency interference" (RFI), from over the horizon -- even from thousands of miles of miles away -- into the radio receiver. As a result, RFI is much more prominent in daytime than nighttime, although RFI can appear at night as well. The daytime RFI is a mixture of human-generated radio chatter and lightning strikes, which occur roughly 100 times per second in the tropics. Most of the radio signal power in the plots during the day is RFI, not sunlight or any other "celestial" emission. (The upper panel shows the raw data, while the lower panel has some of the worst RFI filtered out -- strong spikes lasting only a second or so -- but this still leaves a lot of "ambient" RFI.) Inspect a few plots and estimate, as a ratio, about how much stronger the typical daytime RFI is than the ambient night sky emission.

  2. The U.S. Naval Observatory's Rise + Set Time Calculator can be used to determine sunrise and sunset times for Bowling Green. Find the rise and set times for the Sun for the following two days, which are four weeks apart. (The USNO calculator does not allow dates that are more than a year or two off of the present, but you can use the current year without significant error, or if you want to be sure you have the exact answer, try timeanddate.com.) Report these times in UT hours, and comment on whether the base-level received radio signal power (not small RFI spikes) increases before or after sunrise, and also whether it decreases before or after sunset.

  3. As an example of more extreme RFI, these GIF + PDF plots show ambient daytime noise power in about the first 40% of the time, then switches to a calibration signal for about 1 minute, and then switches over to WWV, a time-announcing broadcast station in Fort Collins, Colorado. Report the ratio of WWV signal power divided by the ambient RFI noise in the first part of the plot, and also the ratio of the WWV power divided by the calibration signal power, which is similar to night sky power levels. Broadcast stations can easily overwhelm desired signals in radio astronomy!

  4. The calibration signal has an effective brightness of 35,355 K in radio astronomical "brightness temperature" intensity units. This can be used to determine the brightness of astronomical sources in the data, like the Galaxy. Measure the calibration signal brightness in the power units used in the previous step. Then measure the brightness of typical night-sky emission in a at least two of the following 24-hour plots, which are cleaner of night-time RFI than most, and also begin at 18:00 UT = noon CST, to make night measurements easier: Report these measurements, and then use the calibration brightness temperature to determine the brightness of the ambient "night" sky at 20 MHz. Where free of RFI, this is primarily synchrotron emission from our Milky Way! This is what Karl Jansky first observed in the 1930s that led to the birth of radio astronomy. NOTE: Brightness temperature is NOT real temperature, but it represents the temperature a blackbody would need to produce radiation of the observed intensity. This is often a convenient way of expressing intensity in radio astronomy.

  5. Use the same calibration technique to estimate the brightness of a Solar flare detected by the WKU RJ receiver at 14:12 UT on 2014 October 27 and plotted here: Ignore the narrow RFI spikes and look for the broad "three fingers" emission -- it's quite bright compared to the ambient noise! For fun, you can also try playing the WKU RJ audio recording of the flare event. Note in this case we are obtaining an instrumental quantity called antenna temperature, which is like brightness temperature but is reduced if the object subtends a small angle compared to the resolution of the radio telescope -- i.e., it's a point source. Report your finding for the Solar flare's antenna temperature.

  6. The RJ antenna has a resolution solid angle of about 2.2 steradians (120° × 60°). Yes, that's big! Use it with the formulae given on the brightness temperature page to determine (a) the Solar flare's specific intensity in watts per square meter per hertz per steradian [W m-2 Hz-1 sr-1] and (b) its specific flux in Janskys, a common unit of flux in radio astronomy. (The formulae correct for the difference between antenna and brightness temperature.)

  7. Compute the radio flare's spectrally-integrated radio flux in watts per square meter [W m-2] by assuming it has a spectral bandwidth of 10 MHz. Compare this integrated radio flux to the integrated X-ray flux of 10-4 W m-2 at the peak of the flare. Which of these is easier to detect? (Explain what you mean by the word "easy" in this context!)

  8. Following the Solar flare, the daytime RFI diminishes greatly, due to X-rays from the Sun exciting the ionosphere so much that it becomes opaque to radio waves that it would normally reflect from over the horizon. This event is called a Sudden Ionospheric Disturbance (SID), and can be seen as an abrupt drop in the daytime signal strength lasting about an hour, beginning around 14:15 UT and lasting until roughly 15:15 UT. Estimate the noise power during the SID and compare it to typical night-time levels. Is a big SID like this a good analog of a Solar eclipse at optical wavelengths?

  9. Assuming the flare radio emission spans 10 MHz, use the distance to the Sun to estimate the total radiated power in watts at the peak of the flare. Then visually estimate the duration of the flare and use this to obtain the total energy yield of the flare at radio wavelengths.

  10. Jupiter storms come in two common types: "long" L-bursts of several seconds each, which sound like ocean waves, and "short" S-bursts that can last only milliseconds and sound like popcorn (try playing the audio samples at the Radio Jove website). Power from a sample L-burst storm is shown here over a 40-minute interval (GIF + PDF) and in a close-up of some brighter features (GIF + PDF). Measure the brightness of some L-bursts and convert them to Jansky units. Then assume a bandwidth of 10 MHz and a reasonable distance to Jupiter to determine the power output of an L-burst if it radiates isotropically. Given the duration of the burst, estimate the equivalent energy yield involved.

  11. EXTRA CREDIT: The duration of an emission event can be used to constrain its physical size, since it cannot vary coherently over a larger scale than light can propagate over during the length of the event. Thus, the width of an L-burst spike in time units, multiplied by c, indicates the maximum sensible size for the emitting region (although it can also be smaller). Use this technique to estimate the maximum possible size of the region emitting the L-bursts, and comment on which of the following possible areas might be responsible. Use this same technique to comment on the possible source of S-bursts lasting only a few milliseconds.