Sloan Digital Sky Survey Telescope Technical Note 19970220
Matthew Buffaloe
University of Washington
During the summer months, a regular plague of moths occurs in the mountains of the southwest deserts of the United States. Apache Point Observatory is located 2800 m (9200 feet) above sea level in the Sacramento Mountains of south-central New Mexico. The intensity of the plague varies from year to year; near the peak of a bad year, hundreds of moths are captured per day inside the observatory buildings.
The moths have a variety of behaviors. Most troublesome to us is that they seek dark tight spaces. The telescope axes at the observatory are moved by friction between a motor-driven capstan and a large drive disk, to which the telescope is connected. The capstan-disk interface is apparently attractive to moths. When the telescope moves, moths are often crushed at the interface. The resultant contamination may cause encoder errors, causing the telescope to point and track poorly until the surfaces are thoroughly cleaned. The result is lost data and telescope time. In other locations, moths interfere with the operation of shutters and other mechanisms.
The goal of the study reported herein was to study moth behavior in an attempt to develop means of reducing their impact on observatory operations. A moth abatement system must be compatible with several constraints.
Figure 1: Images of moths. The images on the right show crack-seeking behavior. The upper right image shows a number of moths that have sought the space behind the blinds between the window handle and frame.
Trapping moths allows regular sampling of the population for species identification. Also, the comparison of various trap designs gives insight into moth behavior.
A series of cardboard boxes were constructed with varied orifices. These boxes were simply placed around the lower portion of the 2.5m telescope enclosure at varying heights. Box #1 had a 0.2 inch slit, oriented along the length of the box. Box #2 had a circular hole with a diameter of 0.3 inch. Box #3 had a corner cut off. Box #4 had 0.3 inch slit in the side with interior walls running the length of box closing off the open space. Box #5 was similar to #4 except that the slit was on the top of the box. Box #6 was similar to #1 except that the slit was oriented with the width of the box. Box #7 had a 0.3 inch square hole in the top with the inside left open.
The traps were placed in the lower level of the Sloan Digital Sky Survey (SDSS) 2.5-m telescope enclosure unless noted otherwise. The enclosure is braced by pairs of angle iron placed so that there is a 10 mm gap between the angle iron. These gaps contained tens of moths most of the summer of 1996, a light moth year.
The boxes were placed on the floor. The next morning, they were checked for moths. One moth was found in each of boxes #1 and #4. The boxes were checked again three days later. There were no moths in any of the boxes. After another three days the boxes were checked and moths were found under the boxes. The boxes were then turned to conceal the openings. The next morning no moths were found in or around them.
Due to the lack of success with the boxes simply set around the room, selected boxes were affixed directly to the angle iron inhabited by the moths. Box #1 was suspended under the angle irons with the slit of the box matching that between the angle iron. Boxes #4 and #5 were set similar to #1 except they were placed on top of the angle iron. Box #6 was set on top of the angle iron with its slit oriented vertically just to the side of the gap in the angle irons. Six days later, these traps were removed and put into plastic bags. They were then placed in the kitchen freezer to kill the moths without damaging them. When the box traps were removed from the freezer boxes #1 and #6 had no moths in them. Box #4 had one moth in it, while box #5 had nine moths.
A week later another attempt was made to trap these insects. Boxes #1 and #6 were cut to allow a night light to be inserted into each of the boxes. Box #1 was lined with foil tape to create a shiny reflective surface, while box #6 was lined with black cloth tape to give a dull, dark surface. These boxes were set, with lights on, in the operations building, in a dark room. They were checked the next two mornings with no success.
Having found that the design of a space consistently attractive to moths was much more difficult than assumed, a moth sample was obtained by searching for them in existing hiding places and catching them by hand.
Once a suitable method had been established an entomologist at the New Mexico State Department of Agriculture was contacted. A series of samples were taken to be identified. The sample of moths collected in early July in the 2.5m telescope support building were identified as Order Lepidoptera, Family Noctuidae, Agrotis ypsilon. Samples collected in late July and early August, from the lower portion of the 2.5m telescope, contained the same species. The larvae of this moth eats a variety of broad-leaf plants and grasses. They have a life cycle of approximately 30 days. This species winter in the southern U.S. and Mexico. As spring temperatures rise the moths migrate north. It is possible that these moths live in the area around Observatory all year.
A sample had been collected between the summers of 1995 and 1996. They were collected from the backs of various drawers and window sills. The collection contained 9 of the Agrotis Ypsilon. There were also 4 Bulia deducta, of the Family Noctuidae.
Three larger species were also captured at the observatory. The first was Hyles lineata of the Family Sphingidae. The second was Coloradia pandora Blake of the Family Saturniidae. The third remains unidentified, but is believed to be of the Family Saturniidae, as well. None of these three species were noted more than twice around the observatory.
It has been noted that during storms moth activity increases. This is especially true during hail storms. It is also believed that moths are sensitive to ultrasonic, due to the fact that they are reportedly a food source of bats. The jangling of keys has produced a sound that caused moths to become agitated. Due to the sensitive optics involved with the telescopes on site, if the vibrations were to be investigated they had to be dealt with very carefully.
One experiment along this line of inquiry was to strike various portions of the building and look for a response from the moths. This was carried out in the lower portion of the 2.5m telescope enclosure. A mallet with a rubber tip was used to prevent any damage to the building. With the mallet the building was struck systematically, beginning with the walls. The moths showed no sign of annoyance. The process then moved to the wall beams. Again the moths were unaffected by the hammering. This carried on to the floor and the beams containing the moths, themselves. The moths seemed to be uninterested until the hammering was within a few inches of directly outside their position between the support beams.
On the same afternoon as the hammering experiment, a similar test was performed. Various loud stomping and clapping noises were made. This was done throughout the room. As above, the moths remained unaffected until disturbance was made quite near to them. Anything further than foot away did nothing.
Based upon the fact that moths are preyed upon by bats the observatory acquired a number of ultrasonic bug repellents approximately a year ago. These devices, once plugged into an electrical outlet, are supposed to emit an ultrasonic sound that is disturbing to moths. Unfortunately, these devices have yet to have any noticeable affect on the moth population around the observatory.
A signal generator and a small self-amplified speaker were set-up in the lower portion of the 2.5m telescope enclosure. The speaker was aimed up, in the general direction of a large cluster of moths between the beams. The amplitude was set at 2 Vpp so as not to damage the speakers. The volume was changed by controls on the speaker itself. The frequency of the signal was varied from 200 Hz to 200 kHz. The sounds produced no affect on the moths, except in the range of 10 kHz. At this frequency the target cluster became excited for a moment, but very quickly settled back down. After 19 kHz no audible signal could be detected from the speakers. The speaker's effectiveness was questioned but not investigated. This data is shown in Table 1.
Table 1: Moth frequency responses. A range of sounds were created using a signal generator attached to a speaker. Moth response to each sound was noted.
Amplitude Frequency Reaction 2 Vpp 0.2 kHz none 0.3 kHz very slight 0.4 kHz startled but settled soon 0.5 kHz none 0.6 kHz 1 moth noted moving briefly 0.7 kHz none 0.8 kHz none 0.9 kHz none 1.0 kHz none 1.1 kHz none 1.3 kHz none 1.5 kHz none 3.0 kHz none 5.0 kHz none 10.0 kHz disturbed but not moving 11.0 kHz disturbed but not moving 12.0 kHz disturbed but not moving 14.0 kHz disturbed but not moving 16.0 kHz none 18.0 kHz none 20.0 kHz none 40.0 kHz none 60.0 kHz none 80.0 kHz none 100.0 kHz none 200.0 kHz none
Due to the interesting response noticed at a frequency of approximately 10 kHz, the equipment was used to create a chirping sound. With the signal generator a frequency modulation mode was used. The amplitude was set as previous, while the frequency was set at 10 kHz. The span was set at either 1 kHz or 2 kHz. The rate was set at 5, 10, 15, 20, or 100 Hz. Beyond momentary agitation, the moth's showed no concern toward the signals. This test was also run at 15 kHz with spans of 1 and 3 kHz and a rate of 100 Hz. This again did nothing to the moths.
A number of casual observations have been made by the staff of Apache Point Observatory. One of the more interesting items involves the blowing of air over moths. It has mentioned a number of times that this has a profoundly disturbing affect.
This observation led to a conceptual deterrent system. This system involved a system of air hoses running to all of the drive assemblies. These hoses would release jets of compressed air at prescribed intervals.
For this system some hardware was acquired earlier this summer. An electric timer and a small three way air valve had been ordered. Unused small rubber tubing with an inside diameter of 1/8 inch was located on site for the plumbing. An air gun and array of nozzles were ordered to facilitate testing of a prototype. Upon inspection of the timer it was noted that it would not give differing on and off times. To solve this problem another identical timer was purchased to put in series with the first. When inspecting the drives of the various telescopes it was noted that no matter which telescope this system was to be implemented on, sharp corners were to be navigated. To accommodate this, 3/16 inch outside diameter, thin wall stainless steel tubing was acquired.
Throughout the summer a variety of tests were run to answer questions mainly regarding effectiveness and efficiency. One of the more interesting tests began as an inquiry into the movement habits of the moths in the lower portion of the 2.5 m enclosure. To begin, the room was divided into sections. Moth populations were counted for each section. At this point there were 108 moths found in the room. With the use of the air gun all of the known moths were forced out of the beam gaps they were inhabiting. After this disruption a periodic population count was made over the next 11 days. The population on the eleventh day was 29. This accounts for just over 1/4 of the original, known, population. A thorough search of the room was conducted with no more moths to be found. Another note of interest is that while the disruption was occurring, one of the staff was in the upper portion of the 2.5 m enclosure. No surge of moths coming from below was noticed.
To further look into the effectiveness of the system, a test was devised to gain some knowledge regarding the strength of a moth. The main test apparatus consists of a clear piece of PVC pipe with a 1 7/8 inch outside diameter. At one end of the tube is attached a funnel. Through this funnel is inserted an air nozzle. For the experiment the steel tubing was used due to its use in the planned system to be implemented. At the far end of the tube is a perforated cap. This is to allow reasonably free air flow while preventing the moth from escaping. This setup is shown in Figure 2. While monitoring the moths distance from the nozzle, air jets were directed at them. The pressure of the compressed air was varied to give different air speeds.
A couple of interesting observations were made. First, these creatures are very strong. They were able to handle upwards of 80 psi so long as they were more than five- six inches from the nozzle. Once they moved within the prescribed distance they were dislodged by 50 psi. It should be noted that these moths showed signs of annoyance with as little as 20-30 psi. The second observation of interest had to do with the moths reaction when experiencing a jet. The moths had a tendency to "hunker down" and wait until the jet had stopped.
Figure 2: Air jet test setup. Testing apparatus used to gain information regarding strength of moths. The same equipment was used to relate pressure of the compressed air and the velocity of the jet experienced by the moths.
With the same setup as above a test was performed to relate the pressure of the compressed air and the velocity experienced by the moths. A series of holes were drilled into the PVC tube to permit the insertion of a pitot tube and an anemometer. These holes were drilled at six inch intervals starting at 12 inches and going to 30 inches. When a hole was not in use it was taped over to maintain continuity. This information concerning velocity was then used to estimate flow rate. This information is given in Table 2.
The anemometer measurements are sensitive to air pressure differences. Due to the high altitude of the observatory, the reading from the anemometer had to be corrected. The values in Table 2 reflect this correction.
Table 2 : Flow velocity and volume. For each gauge pressure the flow velocity was measured at each of the four ports drilled into the pipe. An average of the four velocities was calculated. From the average flow velocity, flow volume was calculated in both english and metric units. Pressure Tube Pitot Corrected Average Flow Flow port velocity anemometer velocity rate rate psi ft/min ft/min ft/min ft^3/min m^3/min 20 1 291.2 251.3 254.8 3.39 0.0959 2 230.2 241.3 3 243.6 243.8 4 291.2 245 30 1 318.9 375 337.9 4.49 0.127 2 318.9 352.5 3 318.9 360 4 308.8 350 40 1 371.2 475 406.3 5.4 0.153 2 356.6 450 3 356.6 446.3 4 356.6 437.5 50 1 493.7 581.3 515.8 6.86 0.194 2 493.7 555 3 482.8 550 4 424.4 545 60 1 524.9 622.5 581.8 7.74 0.219 2 544.7 620 3 554.3 615 4 563.8 608.8 70 1 642.9 697.5 660 8.78 0.249 2 651.1 690 3 642.9 683.8 4 591.3 680 80 1 675 750 727.7 9.68 0.274 2 727.9 743.8 3 727.9 736.3 4 720.6 740
Once all of the hardware had been assembled a prototype system was constructed. The system is controlled by the two previously mentioned timers set in series. The first timer was set to cycle on and off at seven seconds. The second timer gives a burst of one second at the beginning of every on cycle. This gives a total off time of 13 seconds with an on time of one second. The solenoid was plugged into the second timer. The air was supplied by the main site compressor. The regulator on the air outlet was set to 45 psi. This outlet was then connected to a reservoir tank to prevent excess back pressure effects. The solenoid was then connected to the reservoir tank. A single rubber tube was connected to the solenoid. This tube was then divided into three tubes. The assembled hardware for the prototype is shown in Figure 3. This contraption was set on the altitude drive of the 3.5 m telescope. The three rubber tubes were pointed into separate portions of the drive. Figure 4 shows the upper two tubes pointing into the drive assembly. Figure 5 shows the third last tube pointing up into the drive from below. This prototype ran for about two weeks with no ill effects noted. In addition there was no report of moths being run over by the altitude drive. This was not conclusive because only a small number of moths had been run over all summer. A week after removing the prototype system the wheel was checked and at least two carcasses had been squashed. These insects were squashed after the removal of the air jets.
Figure 3: Prototype hardware. This is all of the hardware used in the prototype air jet system. This shows the system just after being assembled, and just before installation.
Figure 4: Upper jets of prototype system. This photo shows two of the nozzles of the system. They are directed into two separate contact points on the upper portion of the altitude drive of the 3.5m telescope.
Figure 5: Lower jet of prototype system. This photo shows one of the nozzles of the system. It is directed up into the lower portion of the altitude drive of the 3.5m telescope.
In an attempt to more clearly prove the effectiveness of this system one more test was performed. The main idea behind the test was to attract moths to a specific surface. This surface was divided into two portions. One side would have a jet of air periodically blown across it. The first model of this test consisted of a foil lined box with a small light in it. This had a piece of plexi-glass for the surface. The light was to be the attractant. This was left overnight, in the base of the 3.5 m enclosure, and recorded on a video cassette. Due to poor lighting the video recording was completely unintelligible. This test was attempted again the next night using a video monitor, producing a white light. The recording of that night was much clearer. However, no moths were seen on the screen. In an attempt to test a larger population the test was moved to the lower portion of the 2.5 m enclosure. Again no moths visited the screen. The question was raised as to whether the air jet scared the moths away before they could be detected on the screen. For one night the air jet was disconnected. The screen was also adjusted to give a slight flashing appearance, in an attempt to be more appealing to the moths. Inspection if the recording showed, once again, no moths present.
While doing initial surveys of moths populations in the 2.5m enclosure it was noted that the moths were disturbed by the flashlight beam. When the light was pointed at them, they would quickly move out of the beam.
To further investigate this activity a desk lamp was affixed to one of the beams in the lower portion of the enclosure. The light was directed down upon a cluster of approximately 10 moths. Six of the ten moths moved away as soon as the light was turned on. By moving away, all the moths did was remove themselves from the direct beam of light. They were still in a well lit area. Within an hour, one more of the insects had moved. The next morning there were still three moths in the light, but they had moved out of the direct beam. By the third day only one moth remained and it was well toward the fringe of the light. On the fourth day two moths were present, but they were essentially out of the light. Since an incandescent lamp had been used for the experiment, heat was suggested as the driving force behind the moth exodus.
This question led to the use of a fluorescent lamp in the same experiment. The light was set over a cluster of 15-20 moths. Within 10 minutes of the light being turned on all of the insects had moved out of the beam. The lamp was then turned off. Two days later the light was moved over a new cluster of approximately 30 moths. As soon as the light was turned on the moths began moving. Within three minutes at least half of the cluster had moved off. By five minutes only approximately a third of the original population remained. After 10 minutes, six moths remained. After 1.5 hours all the moths had moved out of the light. A large cluster had formed toward the edge of the light. The light was then turned off. Two days later the light was turned on over a cluster moths. The experiment was neglected until the next morning. The cluster of moths had not moved out of the light.
Earlier in the year the observatory purchased a common insect zapper. The zapper consists of two black lights with a grid of wires hanging from between the lights. When a moth strikes the grid an arc is formed and the insect is electrocuted. Due to the dependency of the dielectric constant of air on pressure and the high altitude of the observatory, the grid would arc violently simply by plugging it in to the wall outlet. The manufacturer sent two capacitors to be hooked up to the grid, in order to try and reduce the voltage. These did not have enough of an effect. Instead a variable line voltage transformer was connected to reduce the voltage. This worked fine for the grid, however, the lights required full voltage. The zapper was rewired to accommodate this. The zapper was then installed in the intermediate level of the 3.5m telescope. A wire basket was hung from the zapper to collect the moths. This was necessary because of the fact that the zapper was hung on a beam that rotates with the telescope while the floor remains stationary. Without the basket moth carcasses could have been spread all over the floor of the level.
After a week of running the zapper, it was suggested that the insects are attracted to the infrared as opposed to visible or ultraviolet light. To test for this an incandescent bulb was affixed to the zapper above where one of the black lights would have been. The black and incandescent lights were alternated from night to night. There were also two periods for which no light was turned on. This was done to see how many moths would randomly run into the zapper. A table and graph of the carcass counts from night to night are shown in Table 3 and Figure 6.
Table 3: 3.5-m Telescope Moth Zapper. A bug zapper was set up in the intermediate level of the 3.5m telescope. The number of moths killed each day was recorded. Different lights were attached to the zapper to try and determine moth attraction. On two occasions no lights were turned on, to see how many moths would randomly run into the zapper. Date(Aug.) Light Moths Killed Moths/Day 1 Black 11 2 Black 19 5 Black 144 48 6 Black 14 7 Incandescent 9 8 Black *197 9 Incandescent 12 12 Black 73 24 13 Incandescent 25 14 Black 10 15 Incandescent *54 16 Black 25 19 Incandescent 48 16 20 None 1 21 Black 31 22 Incandescent 14 27 None 9 1 28 Incandescent 18 29 Black 12
Figure 6: 3.5-m telescope moth zapper. This is a graphical representation of Table 2.
The asterisks refer to days with heavy storms, which may account for the increased activity. The gaps in the data occurred during the weekends when the zappers were left unattended. The large counts after each weekend reflect multiple day accumulations of moth carcasses.
When the capacitors did not work to solve the arcing problem, the manufacturer sent a replacement model of a lower voltage. Upon receiving the new zapper it was tested and found that it arced as the first did. It was, however correctable with a single capacitor. This second zapper was hung in the Monitor Telescope enclosure. This was facilitated by the fact that the telescope had been taken apart for maintenance. With this zapper, the feasibility of running the zapper during the day and turning it off at night was tested. The moth carcasses were counted twice a day: once in the morning and again in the evening. This data is shown in Table 4 and Figure 7.
Table 4: Monitor telescope moth zapper. A bug zapper was set up in the observing level of the Monitor telescope. The number of moths killed by the zapper was counted at the end of each night and day. This was to determine the effectiveness of using a zapper to control moth population during the day, while the telescope would be inactive. Date(Aug.) Day/Night Moths Killed Moths/12hr 12 night 12 13 day 5 n 9 14 d 7 n 9 15 d *25 n 9 16 d weekend 28 4 19 d 3 n 4 20 d 10 n 1 21 d 5 n 2 22 d 2 weekend 28 3 27 d 1 n 1 28 d 3 n 1 29 d 0
Figure 7: Monitor telescope moth zapper. A graphical representation of Table 4.
In an attempt to add to our information regarding light related behavior a letter was posted on an internet newsgroup; sci.bio.entemology.misc. The response suggested a scientific company that sells insect light traps. This company was contact and a portion of their catalog was acquired. Upon inspection these traps were largely cleaner versions of a bug zapper.
Sometime in mid-July a dead bat was found in the lower portion of the 2.5m telescope enclosure. This implied the presence of bats in the area and suggested the possibility of natural predation as means of controlling the local moth population. The entomologist at the New Mexico Department of Agriculture was contacted. The question was also referred to the NMSU's Extension Wildlife Agent. It seems that while some bats do eat moths, not all species do. It was not clear whether those in the vicinity of the observatory preyed upon moths or not.
The bat population can often be locally increased by providing artificial bat houses. Bats are rather selective about their houses. They prefer high, stable temperatures in their roosts. Unfortunately, bats often find suitable habitats in or around human dwellings. Cases have been reported in which there were so many bats in the attic of a house that the owners could no longer insure the building. Coincidentally little, or no, decrease in the local insect population was noticed.
