Dimensionally Stable Sensors for a Continuous Monitoring ... - BioOne

HOUSEHOLD AND STRUCTURAL INSECTS

Dimensionally Stable Sensors for a Continuous Monitoring Program to Detect Subterranean Termite (Isoptera: Rhinotermitidae) Activity NAN-YAO SU1 Ft. Lauderdale Research and Education Center, University of Florida, Institute of Food and Agricultural Sciences, Ft. Lauderdale, FL 33314

J. Econ. Entomol. 95(5): 975Ð980 (2002)

ABSTRACT A dimensionally stable sensor composed of a closed-cell polyethylene sheet on which a silver particle circuit was painted and sandwiched between two spruce stakes was tested for use in a monitoring program to detect subterranean termites. Sensors were connected to a datalogger for continuous monitoring of sensor circuit breakages over 12 mo, and were manually inspected monthly to assess sensor performance. The mean monthly sensor accuracy for three Þeld test sites was 98.7%, with most false responses caused by early timing of the monthly inspection when termites entered the station before damaging the sensor circuits. Mean sensor longevity (the time for a sensor circuit to break in the absence of termites) of the dimensionally stable sensors was 11.7 mo; a substantial improvement over the 4.4-mo longevity recorded previously for wooden sensors. KEY WORDS Isoptera, closed-cell polyethylene, Sentricon, termite baits

ONE RECENT DEVELOPMENT for subterranean termite control is to manage termite populations with a monitoring-baiting program such as the Sentricon Termite Colony Elimination System (Dow AgroSciences, Indianapolis, IN). Unlike soil insecticide barriers, the system relies on routine inspection to detect termites (Su 1994), and insecticide baits are used only when termites are detected. This arrangement drastically reduced pesticide use for control of subterranean termites. Approximately 1 g of hexaßumuron (active ingredient for the Sentricon system) is sufÞcient to eliminate a colony of subterranean termite (Su 1994). An on-going monitoring program following colony elimination continues to protect the structure by early detection of newly invading termite populations. The manual monitoring currently used for the system, however, is labor intensive, costly, and potentially disruptive to termites that have recently initiated feeding in the station. To improve efÞciency, sensors composed of conductive circuits of silver particle emulsion painted directly on wooden stakes were tested for their potential in detecting termite feeding activity (Su et al. 2000). The wooden sensors were further incorporated into a datalogger to automatically monitor termite feeding activity using a computer and telephone communication (Su 2001). With these sensors, termite activity was transformed to an electric signal when termites fed on the wooden sensors and broke the conductive circuits. Under wet environments, however, moisture-induced wood expansion could break 1

E-mail: nysu@uß.edu.

the circuits, resulting in false positive responses, i.e., circuit breakage in the absence of termites. The use of wood as the sensor medium thus yielded a short sensor longevity of ⬇4 Ð 6 mo in previous studies (Su et al. 2000, Su 2001). In this study, a polymer foam sheet was used as the sensor medium to improve sensor longevity. This article reports the Þeld performance of this dimensionally stable sensor.

Materials and Methods Dimensionally Stable Sensors. As shown in Fig. 1, the sensor is composed of a closed-cell polyethylene sheet (2.5 by 15 by 0.2 cm) on which a silver particle circuit was painted with a conductive pen (Chemtronics, Kennesaw, GA) and sandwiched between two spruce (Picea sp.) stakes (2.5 by 16 by 0.5 cm). The sensor was inserted into a Sentricon station housing along with an extractor and a monitoring device (spruce stake, 2 by 19 by 1 cm) (Fig. 1). Several sensors were wired together using 20-AWG (0.41-mmdiameter) electric cable to form a continuous loop of an electric circuit that was then connected to the I/O (input/output) outlet of a datalogger (CR10X, Campbell ScientiÞc, Logan, UT) as described previously (Su 2001). Stations placed at 2- to 10-m intervals at each site were grouped into Þve zones, and the continuous loop of sensor circuit for each zone was connected to one of the Þve I/O outlets of the datalogger. The numbers of stations in each zone varied from 3 to 12. The datalogger was programmed to apply 2,500 mV current through the circuit loop of each sensor zone

0022-0493/02/0975Ð0980$02.00/0 䉷 2002 Entomological Society of America

976

JOURNAL OF ECONOMIC ENTOMOLOGY

Vol. 95, no. 5

Fig. 1. A dimensionally stable sensor composed of a closed-cell polyethylene sheet on which a silver particle circuit was painted and sandwiched between two spruce stakes is inserted into a Sentricon station housing along with an extractor and a monitoring device.

and record the return voltage every 2 h. A sudden and sustained decrease in return voltage ⬍2,500 mV signiÞed circuit breakage of at least one sensor within the zone. The bi-hourly return voltage readings were stored in the datalogger memory and downloaded using a telephone modem every 4 d for 12 mo to a host computer located at Ft. Lauderdale, FL. Testing Sites and Termite Survey. The dimensionally stable sensors were tested at the same three sites as described in the previous Þeld experiments with the computerized remote monitoring system (Su 2001). Sensors were installed at all three sites in late February 1998. The one residential site (SHL) in Ft. Lauderdale, FL, had active subterranean termites (Reticulitermes species) before 1996, but no live termites were found during the testing period. Sixteen Sentricon stations were installed in soil along a partial perimeter of the structure and were divided into Þve groups. Sensors placed in the stations of these zones were connected to a datalogger mounted on the exterior wall. The second site (VER) was a parking structure at the Florida Medical Entomology Laboratory, University of Florida, in Vero Beach, FL. The structure was heavily infested by a colony of the eastern subterra-

nean termite, Reticulitermes flavipes (Kollar), during 1995Ð1997 (Su 2001). Beginning in March 1997, wooden sensors were installed at VER using 27 Sentricon stations, and this colony was eliminated by hexaßumuron baits by August 1997, following the detection of termite activity by the wooden sensors connected to the datalogger (Su 2001). In February 1998, 37 Sentricon stations containing dimensionally stable sensors were placed in soil surrounding the structure. The sensors were grouped into Þve zones and connected to a datalogger mounted on the interior wall of a room. The third site (VR0) was the administrative building of the Florida Medical Entomology Laboratory. Landscape materials such as railroad ties near VR0 were heavily infested by another colony of R. flavipes. Wooden sensors were tested using this colony, but it was not baited and left viable as in the previous study (Su 2001). Activity of this R. flavipes colony was monitored with underground monitoring stations (plastic collars 17 cm in diameter and 15 cm in height) containing feeding blocks (six spruce boards [7 by 13 by 2 cm] nailed together) as described by Su and Scheffrahn (1986). Thirty-one Sentricon stations containing dimensionally stable sensors were placed in the soil

October 2002

SU: DIMENSIONALLY STABLE SENSOR FOR SUBTERRANEAN TERMITES

977

Fig. 2. Bi-hourly voltage records from Þve zones of sensors for site VER for 12 mo. A signiÞcant decline of voltage from 2,500 mV indicates circuit breakage of sensor(s) in a zone. Date axis between 2 consecutive months represent date of the earlier mo.

around the building and near the railroad ties, and were grouped in Þve zones and connected to a datalogger mounted on the exterior wall of the building. In May 1998, sensors infested with termites were removed and replaced with bait tubes containing paper matrix impregnated with 0.5% novißumuron (Dow AgroSciences). Novißumuron is a chitin synthesis inhibitor referred to as an “experimental active ingredient” by Getty et al. (2000) for use against Þeld colonies of Reticulitermes species in Northern California. During the baiting period, cables connected to the infested sensors were detached and reconnected to the datalogger to restore circuit continuity of the zone, thus the baited station was not monitored by the datalogger. Substantially consumed baits (⬎50% visual estimate) were replaced with fresh baits. When termite activity was no longer observed in any Sentricon stations or underground monitoring stations, baits were replaced with fresh sensors that were reconnected to the circuit loop of each zone for continuous monitoring of termite activity by the datalogger. Data Collection and Analysis. All stations were manually opened monthly to examine the presence of

termites and status of sensors. Sensor accuracy was assessed by comparing the site inspection results and the datalogger records. Results were assigned to one of four categories: true negative (TN), circuit intact in the absence of termites; true positive (TP), circuit breakage in the presence of termites; false negative (FN), circuit intact in the presence of termites; and false positive (FP), circuit breakage in the absence of termites. Accuracy rate (A) was calculated as the percentage of both true categories within all categories, or A ⫽ 100*(TN ⫹ TP)/(TN ⫹ TP ⫹ FN ⫹ FP). The monthly results for the entire 12-mo test period were analyzed for signiÞcant differences between true (T ⫽ TN ⫹ TP) and false (F ⫽ FN ⫹ FP) responses for all sites with the WilcoxonÕs signed-ranks test (SAS Institute 1998). Sensor longevity was calculated as the mean time before a sensor produced a false positive response, namely sensor circuit breakage caused by factors other than termite feeding. Only sensors installed at the beginning of the test were included in calculating sensor longevity. This excludes sensors removed earlier or placed later due to baiting and thus were tested ⬍12 mo.

978

JOURNAL OF ECONOMIC ENTOMOLOGY

Table 1. Site

VER

Site total T/F

VRO

Site total T/F

Vol. 95, no. 5

Performance for a dimensionally stable sensor system tested in two sites (VER and VRO) for 12-mo period Month

True

False

TN

TP

FN

FP

1 2 3 4 5 6 7 8 9 10 11 12

35 35 34 35 35 35 35 35 35 35 35 35

0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0

0 0 1 0 0 0 0 0 0 0 0 0

1 2 3 4 5 6 7 8 9 10 11 12

28 28 21 21 20 31 31 31 31 30 30 31

2 2 0 1 0 0 0 0 0 0 0 0

1 1 6 0 0 0 0 0 0 0 0 0

419

333

1

11

0 0 1 0 1 0 0 0 0 1 0 0

% accuracy (mean) 100 100 97 100 100 100 100 100 100 100 100 100 (99) 97 97 75 100 95 100 100 100 100 97 100 100 (97)

Wilcoxon ranked test Z

P

⫺3.059

0.0022

⫺3.059

0.0022

All sensors at another site SHL registered TN results with 100% accuracy throughout the test period. TN, true negative, circuit intact in the absence of termites; TP, true positive, circuit breakage in the presence of termites. FN, false negative, circuit intact in the presence of termites; FP, false positive, circuit breakage in the absence of termites. % Accuracy ⫽ 100*(TP ⫹ TN)/(TP ⫹ TN ⫹ FP ⫹ FN). Standard normal distribution (Z) and signiÞcance level (P) between true (T ⫽ TN ⫹ TP) and false (F ⫽ FN ⫹ FP) responses.

Results and Discussion In the absence of termites, no circuit breakages were recorded from sensors installed at site SHL throughout the 12-mo test period, resulting in 100% true negative (TN) response. Sensors at this site were exposed to periodic wet conditions from rainfall and landscape irrigation, but the occasional wet conditions did not appear to impact the sensor integrity. Results at VER were similar to those of SHL. As described in the previous study (Su 2001), an R. flavipes colony infesting this structure was eliminated using hexaßumuron baits in 1997 following the detection of termite activity in wooden sensors. With the exception of one sensor whose circuit was separated at the cable connection 3 mo after installation (Fig. 2, FP for zone 1), the rest of the sensors correctly indicated the absence of termite activity by reporting the ⬇2,500 mV return voltage throughout the 12-mo period (Fig. 2, TN). Soil at site VER was muddy and often saturated with water, and some of the sensors at this site were occasionally soaked in muddy water. Unlike the wooden sensors tested previously at this site (Su 2001), the sensors used in this study were not affected by these extreme wet conditions. The polyethylene sheet used as the medium for the current sensor was dimensionally stable enough to maintain circuit integrity for the entire 12-mo test period. The mean accuracy rate for sensors used at site VER for the 12-mo period was ⬎ 99%, with signiÞcantly (Z ⫽ ⫺3.059, P ⫽ 0.0022) more true (419) than false (1) records (Table 1).

Of the 31 sensors installed at site VR0, termites infested three within a month (Table 1). Circuits of two infested sensors were damaged by termites and thus reported true positive (TP) responses in early March (Fig. 3, one each in zones 2 and 5). The closedcell polyethylene sheet on which the circuit was placed was not a feeding substance for termites, i.e., no nutritional value, but termites readily chewed the sheet and circuit when feeding on the sandwiching wooden stakes (Fig. 4). The circuit of the third infested sensor (Fig. 3, zone 3) was not broken because the infestation was in the early stage and termites had not yet reached the sensor circuits, thus resulting in a false negative (FN) response (Table 1). Similar FNÕs were also reported from previous studies when wooden sensors were used (Su et al. 2000, Su 2001). If inspected later and termites were allowed to infest further, the circuit would probably have been broken by termites. Results of the second month were similar to those of the Þrst month, two TP and one FN with 97% accuracy (Table1). In May 1998, 3 mo after installation, six stations (two in zone 3, and four in zone 5) were infested with termites, but none of their circuits were broken (Fig. 3, FN). Again, the monthly on-site inspection uncovered the infested stations before termites damaged the circuits. Moreover, one sensor in zone 2 reported circuit breakage in the absence of termites in late May (Fig. 3, FP), resulting in 75% accuracy (Table 1). All infested sensors (6 FN) were replaced with Recruit II and baited stations were disconnected from the datalogger. One station in zone

October 2002

SU: DIMENSIONALLY STABLE SENSOR FOR SUBTERRANEAN TERMITES

979

Fig. 3. Bi-hourly voltage records from Þve zones of sensors for site VR0 for 12 mo. TN, true negative, circuit intact in the absence of termites; TP, true positive, circuit breakage in the presence of termites, FN, false negative, circuit intact in the presence of termites; FP, false positive, circuit breakage in the absence of termites. Date axis between 2 consecutive months represent date of the earlier mo.

Fig. 4. Termites readily broke the circuit (TC) of a dimensionally stable sensor at site VR0 when feeding on the sandwiching wooden stakes.

980

JOURNAL OF ECONOMIC ENTOMOLOGY

5 reported correctly the presence of termites in June (fourth month), and received Recruit II (Table 1), but the sensor circuit in one station in zone 1 broke in the absence of termites in July (Fig. 3, FP). Termites continued to consume bait until August 1998 (sixth month), after which no termite activity has been recorded from this site. This R. flavipes colony was probably eliminated after consuming seven bait tubes containing novißumuron bait during the 3-mo baiting period. In agreement with the results of Getty et al. (2000) with Reticulitermes species in Northern California, novißumuorn appeared to eliminate the R. flavipes colony faster than hexaßumuron. With the exception of one sensor in zone 2 that was broken in the absence of termites (FP), all other sensors correctly reported the absence of termite activity after this colony was eliminated in August (Fig. 3), thus resulting in 100% accuracy for most of the 6 Ð12 mo postelimination period (Table 1). The 12-mo mean accuracy for sensors used in site VR0 was 97%, with most of the false negatives (eight cases FN) caused by the early timing of on-site inspection where termites were found in the stations before sensors were damaged (Table 1). For all practical purposes, such false negatives are probably not a problem because the stations will be inspected quarterly or bi-annually in commercial use. The three cases of false positive (FP) recordings were due to separations of cables from sensor circuits, most likely by the inadvertent force on the cable during handling. Even with the 11 cases of false results, there was signiÞcantly (Z ⫽ ⫺3.059, P ⫽ 0.0022) more true response than false response at site VR0 (Table 1). The overall accuracy rate for all three sites (including all true responses for 16 stations at site SHL for 12 mo) during the 12-mo test period was 98.7%, and the mean sensor longevity was 11.7 mo, a substantial improvement over the longevity of 4.4 mo for wooden sensor tested previously (Su 2001). There are other devices for detecting termites, including an acoustic emission detector (Scheffrahn et al. 1993), electronic odor devices (Lewis et al. 1997), and more recently TERMATRAC (Termatrac, Queensland, Australia) that uses microwave reßection to sense insect movement behind walls and/or in wood. One critical issue in detection technologies is the ability of a device to differentiate signals unique to termite activity from nontermite related events. The acoustic emission detector (Scheffrahn et al. 1993), for example, was programmed to register vibrations at frequency ranges that are produced when wood Þbers are torn by termites. Because few other insect species feed on wood, circuits on the wooden sensors used in the previous study (Su 2001) were readily broken when termite fed upon the sensors. Other biotic factors such as fungal decay or wood boring beetles did not impact the wooden sensors, but moisture-induced expansion broke circuits of the wooden sensors, causing numerous false positive events. When depending on termite feeding to trigger the detection signals, a sensor needs to satisfy two opposing characteristics, namely the sensor needs to be

Vol. 95, no. 5

fragile enough to be damageable by termites but it needs to be durable enough to remain intact in the absence of termites. Results of this study showed that, unlike wood, the closed-cell polyethylene sheet was dimensionally stable against moisture and temperature, and was not damaged by fungi or other soil dwelling organism such as earthworms, ants, or beetles that were frequently found in the stations. Moreover, termites readily damaged the sheet and broke the circuits when feeding on nearby wood stakes. Other materials such as expanded polystyrene, expanded polypropylene, textured polyethylene, vinyl, polyol resin of polmeric diisocyanate, absorbent paper with polyethylene backing, cellular rubber, and sponge rubber, are also readily damaged by termite feeding activity yet are unaffected by humidity and temperature. These materials are potentially useful for producing dimensionally stable sensors such as that described in this study. Acknowledgments My gratitude to P. Ban and R. Pepin (University of Florida) for technical assistance, H. Puche (University of Florida) for assistance in statistical analysis, J. Perrier (University of Florida) for Þgure illustration, and R. H. Scheffrahn, and P. Bardunias (University of Florida) for review of the manuscript. This research was supported by the Florida Agricultural Experiment Station and a grant from Dow AgroSciences, and approved for publication as Journal Series No. R-08524.

References Cited Getty, G. M., M. I. Haverty, K. A. Copren, and V. R. Lewis. 2000. Response of Reticulitermes spp. (Isoptera: Rhinotermitidae) in northern California to baiting with hexaßumuron with Sentricon Termite Colony Elimination System. J. Econ. Entomol. 93: 1498 Ð1507. Lewis, V. R., C. F. Fouche, and R. L. Lemaster. 1997. Evaluation of dog-assisted searches and electronic odor devices for detecting the western subterranean termite. For. Prod. J. 47: 79 Ð 84. SAS Institute. 1998. StatView reference. SAS Institute, Cary, NC. Scheffrahn, R. H., W. P. Robbins, P. Busey, N.-Y. Su, and R. K. Mueller. 1993. Evaluation of a novel, hand-held, acoustic emissions detector to monitor termites (Isoptera: Kalotermitidae, Rhinotermitidae) in wood. J. Econ. Entomol. 86: 1720 Ð1729. Su, N.-Y. 1994. Field evaluation of hexaßumuron bait for population suppression of subterranean termites (Isoptera: Rhinotermitidae). J. Econ. Entomol. 87: 389Ð397. Su, N.-Y. 2001. A computerized system for remote monitoring of subterranean termites near structures. J. Econ. Entomol. 94: 1518 Ð1525. Su, N.-Y., and R. H. Scheffrahn. 1986. A method to access, trap, and monitor Þeld populations of the Formosan subterranean termite (Isoptera: Rhinotermitidae) in the urban environment. Sociobiology 12: 299 Ð304. Su, N.-Y., P. M. Ban, and R. H. Scheffrahn. 2000. Control of Coptotermes havilandi (Isoptera: Rhinotermitidae) with hexaßumuron baits and a sensor incorporated into a monitoring-baiting program. J. Econ. Entomol. 93: 415Ð 421. Received for publication 23 January 2002; accepted 25 March 2002.