Overpeck Park is on the outskirts of New York City. The area is a highly urbanized region of Bergen County, New Jersey where the New Jersey Turnpike and I-80 intersect (Figure 1(A)). The park is in Zip Code 07660 and mainly serves residents from 10 other surrounding zip codes (Figure 1(B)). The zip codes occupy 29 square miles. At the time, total population in the zip codes was 236,397 and, depending on the zip code, population density ranged between 3686 people/sq. mile and 15,987 people/sq. mile (average 9777 people/sq. mile) (Table 1). The area was majority minority inhabited and considerably wealthy since median house values were, depending on zip code, $370,300 to $723,400. This observation contrasts with most instances where landfills tend to be placed in low-income areas. However, this must be interpreted in the context of the area having been settled by farmers and therefore historically less well-off compared to, back then, a vibrant New York City commercial neighborhood nearby. Median household incomes ranged between $79,476 and $137,847. The population was relatively middle aged (median age of 39 to 46) and household average size was 2 - 3 people. The community was quite active throughout the day as evidenced from area traffic into and out of the public park.

Figure 1. Location of Overpeck Park (Panel A) and collection of zip codes in its vicinity (Panel B). Note: Z1 to Z7 in panel A refer to Zone 1 - 7. Maps in the 2 panels are not on the same scale.

Visitor to the park enjoy walking trails, biking, various sports courts (for soccer, football, baseball, softball, tennis, pickleball), children’s playgrounds, catch-and-release fishing, and picnic areas. The park also features an amphitheater, athletic fields, boating, and equestrian activities. It is open to the public from 6 am to 10 pm throughout the year.

Table 1.Statistics and demographic characteristics of the study area.

Note: ¹Includes waterways; ²[] represents average value; ³N/A=Not applicable. Source: URL https://www.unitedstateszipcodes.org/ (accessed 03/16/2025).

Park management identified a need for several methane detectors in each of the seven (7) buildings on the premises. For purposes of this project, the buildings were coded as Zones 1 through 7. The monitored spaces in each building (zone) generally included mechanical rooms, a workshop area, public restrooms, storage rooms, lunchrooms, office spaces, and dressing rooms. Depending on the size and usage of the building, the number of detectors manufactured by Macurco Gas Detection Inc. (Model Macurco TXP_T40) installed in each building ranged between 3 and 10. The TracXP TXP-T40 combines multiple sensor types, including electrochemical, catalytic bead, and infrared sensing technologies, for Class 1, Division 1, or Division 2 toxic and combustible area gas monitoring.

Within Zones 2 - 7, individual detectors were connected in a daisy chain fashion (Figure 2) using 18/3 low voltage cable shielded (Cat# 5628; Syston Cable, Chino, CA USA). The first detector in each series within the zone was ultimately connected to a Wireless Modbus Stick (WBS; Cat# MBS-CBBL, SignalFire Wireless Telemetry, Marlborough, MA, USA; Figure 3(A)) using a 3 Cond 18 Gauge shielded wire (Cat# E2203S.30.86; Carol Prysmian Group, Manchester, NH, USA). The manufacturer guidelines outlined under Figure 2 legend notes were followed to ensure robust communication between the WBS and the Wireless Gateway Stick (WGS). For each zone, the WBS was mounted on the rooftop to facilitate wireless communication with a Macurco TXP-C64 (N4X, FIB, DUAL RS-485 PORTS, COM REL, C1D2; (Cat# 88-7000-0300-00)) controller located in Zone 1 at the southwestern end of the park. The controller was pre-configured and housed in a TXP-C64 analog (16) input board. The roof of the building in Zone 1 where the controller was located accommodated 4 wireless gateway stick (WGS; Cat# GWS-CBBL, SignalFire Wireless Telemetry, Marlborough, MA, USA; Figure 3(B)). On the roof, the wireless gateway sticks were installed 6 ft (1.8m) apart to facilitate communication between the 6 WBS in Zone 2 - 7 and the controller in Zone 1 (Figure 4). The gateway sticks were connected to the controller as presented in the wiring diagram (Figure 5). Likewise, the notes following Figure 5 legend provided guidance to ensure robust communication between the WBS and WGS.

Figure 2. Daisy chain wiring of T40s (Notes: 1. Showing typical 2 wire Modbus RS-485 connection from TXP-T40s to Wireless Modbus Stick. The T40s must be set to SLAVE. 2. Wiring must be a 2 wire twisted pair with shield. Sheild should be landed on the “S” terminal at the TXP-T40 Relay/Modbus Boards. 3. Modbus RS-485 loop can continue to the next device otherwise it needs to be terminated at the last T40 via the TR1/TR2 jumpers in the A position. 4. Jumpers need to be set to position B (Dual Channel) to allow each preceding T40’s channel information to pass through the loop to the Wireless Modbus Stick. 5. TXP-T40 Relay/Modbus Board must be specified to connect the T40 to the Wireless Modbus Stick using Modbus RS-485).

Four WGS were deemed necessary because the manufacturer recommended transmission of only 60 packets per minute per gateway device. This value practically limited the number of detectors per gateway stick. During the designing phase, the manufacturer tested a rate of 5 packets per minute from a TXP-T40 transmitter and estimated the requirement of no more than 12 gas detectors per gateway stick. Since the whole project comprised 41 detectors, it was estimated each Modbus data wireless would require 4 gateway sticks. The technical specifications for WBS and WGS are presented in Table 2.

Figure 3. SignalFire Modbus antenna (Panel A) and Gateway stick (Panel B) used at Overpeck Park.

Figure 4. Layout of the 4 gateway sticks on top of the building (Zone 1) where the controller is located.

Figure 5. Wiring diagram of four gateway sticks on the Zone 1 roof connection to the controller. (Notes: 1. Two wireless Gateway Sticks to each com port. 2. Two local TXP-T40s connected through analog 4 – 20 mA to the Analog Input Board. 3. In this configuration, the Gateway/Modbus Breakout boards are not shown. They were connected later for diagnostic purposes – see Figure 9(B)).

Table 2.Manufacturer-specified technical specifications for the Gateway and Modbus stick.

Within each building (i.e., zone), the system required 12 - 24V DC. The electric power mains were stepped down from 115 V to 12 VDC or 24 VDC using a multi-point system MAC6Amp-4-115VAC, 60 Hz Power Supply-10 Amp or MAC10Amp-4-115VAC, 60 Hz Power Supply-6 Amp (Manufacturer: Altronix® Brooklyn, NY, USA) unit, respectively. For each building, a horn/strobe TXP-Combo alarm red strobe, 103 - 115 DB horn, 12 - 24 VDC, IP65, GP (Cat# 93-9926-R000-10) and TXP-Combo alarm base, 12 - 24 VDC, IP65 (Cat# 93-9926-0000-12) was installed on top of the controller in Zone 1 or on top of the power supply unit (PSU) at each of the 6 buildings (i.e., Zones 2-7). The horn/strobe was wired as presented in Figure 6. The horn/strobe combo was programed to be triggered whenever methane levels exceeded a specified setpoint on any of the methane detector in the daisy chain connected to the system within the respective zone.

This measurement and others are deliberate, using specifications that anticipate your paper as one part of the entire journals, and not as an independent document. Please do not revise any of the current designations.

Figure 6. Horn/strobe wiring schematic (A) and landing location on the controller (B).

Within each building, electric and communication cable was installed to accommodate the methane detector system design and manufacturer specifications. Most of the wire installation was through electric conduit pipe to the ceiling and similar confined spaces per electrical code requirements (Figure 7(A) andFigure 7(B)). Final connection to the respective detector included connecting electric power not only to the detector base but also to the display board (Figure 7(C)). Once completed, the system at each building was powered on to display the detector home screen (Figure 7(D)).

Figure 7. Wiring in each building (Panels A and B) and the connection points on each detector (C) in the daisy chain to attain complete electrical and communication circuitry (D).

A splash guard with a calibration port (p/n 88-C50G-0000-00) was installed on each methane detector prior to mounting the detector in place. Figure 8 displays the geometry of the detector-splash guard fitting. As is evident from the geometry, air movement around and into the sensor housing is unimpeded, even when the splash guard is fitted on the detector. Because the methane detectors were mounted at 2.7 m (9 ft) or higher, each detector had a dedicated splash guard to make routine bump testing and calibration exercises conveniently manageable.

Figure 8. Geometry around the sensor when the splash guard is inserted.

Although most programming was conducted by the manufacturer prior to shipping the equipment, some adjustments had to be made to suit local/field conditions. All necessary adjustments were done following manufacturer specifications as outline in the Comprehensive TracXP TXP-C64 Manual and TracXP-TXP-T40 Manual. After installation, each detector was bump tested with a methane gas standard (CH₄ 2.5% volume; 50%LEL; Macurco Gas Detection Inc.). The same standard gas was also used to calibrate each unit as specified by the manufacturer.

Using a CAT5 cable, the C64 controller was connected to the internet and programmed to transmit email notifications to several addresses. The email system was configured by logging into the device (based on IP address). The process provided a replica of controller functions on the laptop, enabling the desired programming including email recipient addresses and types of notifications using the menu ribbon (Figure 9(A)). A breakout board was installed between each gateway stick and the controller (Figure 9(B)). A USB DB25 cable was plugged into each breakout board and connected to a laptop to determine signal strength using the SignalFire Toolkit (Singal-Fire.com).

The template is used to format your paper and style the text. All margins, column

Figure 9. Replica of the controller displayed on a laptop screen after connecting for further programming (A). Four breakout boards corresponding to the 4 gateway sticks were installed at the controller for diagnostic purposes (B).

widths, line spaces, and text fonts are prescribed; please do not alter them. You may note peculiarities. For example, the head margin in this template measures proportionately more than is customary. This measurement and others are deliberate, using specifications that anticipate your paper as one part of the entire journals, and not as an independent document. Please do not revise any of the current designations.

In instances where gas was detected, the horn/strobe was triggered (see Results Section). These events led to short-term decisions manually power off the detectors in the affected building (i.e., zone) pending development of a proper response plan. The power shutdowns mostly coincided with precipitation events. Thus, rainfall data for a consecutive 108 days (i.e., March 6^th 2025 to June 22^nd 2025) from nearby Teterboro Airport (Zip Code 07608) reported by NOAA (Climate.gov; accessed August 12^th 2025) were retroactively compiled to further investigate the suspected environmental circumstances triggering the detected methane emissions. The TRX-T40 event log over the corresponding duration was retrieved by scrolling to the “System” function of the triggered detector. Under this function, each detector registers all the status changes and archives them by date and event type. The detector also tracks all “Sys boot” events each time the detector is powered on. However, it does not show how long the detector stays powered off. The archived information was used to develop a potential theory, based on cluster analysis, behind the intermittent gas emissions detected.

The Clustering is a powerful widely used tool in pattern recognition. Rainfall data and methane emission events at the location in question were coded as summarized in Table 3. The dates between 3/6/2025 to 6/22/2025 without a rainfall event were coded with NR (no rain) whereas those with a rainfall event were identified with an R as part of the ID. For each date, 10 variables/characteristics were annotated, resulting in a 10-digit algorithm comprised of “0” or “1”. More specifically, if the variable in Table 3 occurred/applied, a value of “1” was assigned whereas if it did not occur/apply for that specific date, the variable was assigned a zero (0). For example, if on date designated by month/date/2025 had no rain (i.e., NoRain) and none of the other variables listed in Table 3 applied, mm/dd/2025 algorithm was constructed of 1 (for NoRain) followed by 9 zeros: ending up with 10 digits. The resulting 10-digit algorithm for each day was subjected to cluster analysis (genomes.urv.cat/UPGMA) using Peason correlation coefficient (r) to generate the distance (d) values whereby:

The output dendrogram was exported to https://itol.embl.de/ for further analysis.

Table 3. Description of the identifier and variables used.

At the controller, the alarms were set at 2% lower explosive level (LEL) (Table 4). The same setpoint was also initially adapted at each detector. Over time, only one (i.e., Channel 11) in Zone 2 out of 41 detectors registered any methane emissions. None of the other 7 detectors in Zone 2 were activated, which suggested the emissions were localized to only one specific room within the building where Channel 11 was located. The room has a drainage pipe and a few cracks or fissures which may potentially allow the gas to seep into the room. The room is next to a parking lot with sewer drains from which methane was previously detected (Figure 11). Even then, emissions in the room where Channel 11 is located were detected during or shortly after precipitation events.

Table 4.Bump test, calibration and signal strength for 41 detectors around the park.

^aRSSI = Received signal strength indicator; ^bTTL = Transistor-Transistor Logic; ^cN/A = Not applicable; ^d%LEL setpoint adjusted as to minimize routine activity interruption around the location due to strobe/horn (see details under Results and Discussion).

Figure 11.Floor drainage where Channel 11 is located (A) and in nearby parking lot with a history of methane emission (B).

To minimize disruption of park activities, the horn/strobe in Zone 2 was set to go off whenever more than 10%LEL methane was detected. This is as opposed to all the other detectors set at 2%LEL. It is worth noting that email notifications at the controller were left unchanged for Channel 11, set to dispatch email notifications above 2%LEL, just like all the other detectors in the same zone network.

Methane generation in landfills is a biological process that needs moisture [22]. Thus, moisture availability is an important factor in promoting waste decomposition. It is very likely that once water from rainfall (or snow) events somehow seeped into the capped waste underneath, the anaerobic biological was initiated, and the resulting fugitive gases migrated through fissures (or drainage system infrastructure) to the room in question. This conjecture as a plausible explanation for the intermittently detected methane in Zone 2; confined to Channel 11 was further scrutinized using cluster analysis. A subset of the dates and corresponding variable-based algorithms used for the analysis for 108 consecutive days showing the first 4 days and last the last day is shown in Table 5. Whereas heavy rainfall at or more than 12.7 mm (0.5 inches) occurred on 10% of the days, lower levels of rainfall were registered 32% of the days. No rainfall was recorded for 59% of the days in the 108 days duration. However, because rainfall effects can linger over time, even another 21% or 32% of days with low or no rainfall were cumulatively impacted over a 1 to 2 days duration as though they had received heavy rainfall. For cluster analysis, this latter was coded 1DHvyCum and 2DHvyCum, respectively. For Channel 11, email notifications about methane emission higher than the setpoints (i.e., >2%LEL at the controller and/or >10%LEL at the detector) was documented on 17 days (i.e., 16% of the days monitored). By comparison, archived information at the Channel 11 detector was triggered on 23 days (i.e., 21% of the days monitored). Occasionally, (i.e., 6 of the 23 days), the emission levels triggered the second level alarm (Al2 code) which signified emissions as high as 20%LEL.

Table 5.Sub-sample of the algorithm based on preset variables in Table 3.

Note: Column 1 represents mm/dd/2025 whereby mm=month, dd = day of month, R = Rain and NR = No rain.

The analysis revealed 25 clusters (Figure 12). At least 15 clusters were associated with days on which methane was detected.

Clusters with a solid red box are when the gas was detected as evidenced from strobe/horn, email notification and/or Channel 11 detector data log archive.Most methane emission incidents were on days with a heavy rainfall event (Cluster 1-4), low rainfall levels preceded by high rainfall events (Cluster 6), or with cumulative rainfall higher than 0.5 inches (12.7 mm) in 1-2 days after a rainfall event (Cluster 9-15).Cluster 17, 18, 20, 22, 23, and 24 comprised days without rain but followed cumulative rainfall of 0.5 inches (12.7 mm) or higher within a 1- to 2-day duration.Clusters 5, 7, 8, 19, and 21 represent inconclusive results as included days had all the characteristics of triggering the strobe/horn, but the electrical power had been turned off, disabling comprehensive data collection by the installed system. This is a key limitation which necessitates the system operators to develop a proper protocol for addressing alarm incidents instead of just shutting off the system.

The actual impact of methane detected inside the room in Zone 2 (Channel 11) is not completely clear as not much information about acceptable methane indoors is available in published literature. When measuring combustible gases such as methane, we use a slightly different scale; the lower explosive level (LEL). Overall, the explosive level for methane is 5% by volume (5% bv). The mixture is too lean to burn if less than 5% methane is present. But at 5%, methane can burn or explode if there is an ignition source. EPA Victoria [23] developed guidelines on action levels for methane at different monitoring locations in landfills (Table 6).

Figure 12. Relationship between rainfall events and the installed monitoring system response to methane (red boxes) at Channel 11. Note: T-Al 1 & T-Al 2 = Channel 11 archived data recorded alarm 1 ( > 10%LEL) and/or 2 ( > 20%LEL) methane; S&H = strobe and horn activity was visually documented. Dates with R had a rainfall event. Dashed box = Inconclusive (Electric power off).

Based on Table 6, Victoria EPA recommended trigger action levels of 5,000 ppm inside buildings on landfill areas. Since 100%LEL represents 5% methane or 50,000 ppm methane, the equivalent action levels published by Victoria EPA [23], are converted to %LEL (Table 6; Column 4) to relate to the reading provided by the detector system at Overpeck Park. Based on that conversion, an action level of 10%LEL inside any of the buildings at Overpeck Park may elicit concern. This guidance is quite applicable and transferable to other industries such as mining, enclosed/underground parking lots/subway systems and similar confined spaces where methane emissions are of concern. To the authors’ knowledge, no other jurisdiction has published comparable guidelines. Based on this information, the 2%LEL setpoints in any of the of the buildings requiring remedial action at Overpeck Park may be overly cautious. For reference, action levels for other landfill-related design and management guidelines are also provided in Table 6.

Table 6.Landfill gas action levels.

Note: *Point of measurement is 50 mm above the landfill surface. **Point of measurement is 50 mm from the point of discharge. ***Intermediate cover areas do not have an engineered landfill cap and are not scheduled to receive waste during the next 3 months. (Source: EPA Victoria [23] with modification).

The equations Received signal strength indicator (RSSI) measures the power present in a received radio signal. In this instance, RSSI is a relative number indicated by a negative dBm value relating the strength from the Modbus antenna to the gateway stick. The higher the number, the better the signal. Thus, the closer to 0 dBm, the stronger the signal. The exact number varies between devices but generally -70 dBm and higher values usually equate to the antenna being in an excellent coverage area [24]. There is a point at which trying to obtain more signal delivered diminishing returns, because the quality of the connection is defined by more factors than just RSSI. At Overpeck Park, Zone 2-6 with RSSI -76 dBm to -59 dBm successfully communicated with the controller in Zone 1 (Table 4).

Voltage at the antenna was 13V for Zone 5 and 6 (Table 4) where a 60Hz power supply-6 Amp (Altonix) unit was installed. The other zones (i.e., 2, 3, 4, and 7) where a 60 Hz power supply-10 Amp (Altonix) unit was installed registered 26 – 28.6V. Irrespective of the voltage supply, message pulse check-in at 1-minute intervals was maintained. Transistor-transistor logic (TTL) of 6/7 or 7/7 were consistently achieved in all 6 zones communicating with Zone 1. TTL is a binary digital signal used in electronic circuits to represent an “off” or “on” state based on input and output voltages. Similar consistency of the mainboard and radio firmware were also attained. A sample RSSI output for gateway 3 (Zones 5 and 6) is presented in Figure 13.

Figure 13.Received signal strength indicator (RSSI) from Zone 5 and 6.

According to the manufacturer, the Modbus stick and Gateway antenna have a range of 4.8 km (i.e., 3 miles; Table 2). However, the park has natural vegetation which may reduce these specifications. Thus, signal strength from Zone 7 located 2.7 km (i.e., 1.7 miles) from the controller based on “line-of-sight” ranged between -92 dBm and -89 dBm (Table 4) and communication with Zone 1 was initially unsuccessful. Thus, the controller displayed a communication error for all channels in Zone 7 (i.e., Channel 35 to 43). To rectify the situation, a breakout was inserted between the Zone 7 antenna and the first detector in the daisy chain (i.e., Channel 35; Figure 14(A)). The breakout board was connected to a laptop and Zone 7 reprogrammed to use the antenna in Zone 5 as a booster/relay (Figure 14(B)). The modification successfully enabled communication between Zone 7 and the controller in Zone 1 as evidenced by RSSI (Table 7). The modification made GW4 on the roof of Zone 1 redundant.

Figure 14.Breakout board inserted between the antenna and first detector in the daisy chain (A). The antenna was reprogrammed (B) to use Zone 5 as a booster.

Figure 15. Controller system display after installation and troubleshooting were completed (Note: Grey spaces represent empty channels).

Table 7. RSSI from zone 7 after using antenna at Building E to relay communication signals.

Overall, system performance significantly improved, registering RSSI -81 dBM without compromising voltage, check-in intervals, TTL, mainboard as well as radio firmware (Table 7). The signal transmitted to the controller for all 41 channels stabilized for all methane gas detectors, eliminating all communication errors (Figure 15). The empty channels (in grey) can be used for future expansion of the system if needed.