# Field Protocol

## Site selection
Choose three site classes:
- corridor sites
- control sites
- background/reference sites

## Recommended pilot coverage
- Hammond
- Whiting
- East Chicago
- one or two comparison sites outside the heavy corridor

## Suggested measurements
- broadband EMF meter
- magnetic field meter / gaussmeter if available
- time-stamped notes
- GPS location
- optional infrasound or audio logging if the project expands
- keep ELF magnetic-field readings and any RF readings in separate columns, units, and instruments
- record the instrument frequency band or model so later analysis does not mix unlike measurements
- record the meter response class (true-RMS, linear-response, flat-response, or waveform-capable) so later analysis does not compare unlike instruments as if they were the same thing
- record measurement mode at every stop: stationary spot, walking scan, or body-worn
- record the NIOSH method family separately when relevant: spot measurements, walkthrough surveys, area monitoring, source surveys, or personal monitoring
- if you run a walkthrough-style screening pass, treat it as a continuous-monitoring screen; keep it separate from stationary spot, walking-scan, and body-worn samples, and use it to rank sites and choose follow-up points rather than as a personal exposure estimate
- NIOSH’s walkthrough-survey guidance treats that route screen as a practical hazard-surveillance tool to be evaluated with personal measurements, so keep the screening and personal layers separate
- for any screening pass, record the route-level summary metric actually used, such as mean, max, or a Hi-5-style peak screen, plus the device class and sampling interval
- record `sampling_interval_seconds` and whether the reading is a TWA, rolling maximum, or brief-peak screen so later analysis does not compare unlike meters as if they were the same thing
- if the instrument logs at a coarser interval such as EMDEX Lite-style peak capture, treat the maximum as a method-specific screen unless rapid changes are known to be impossible at that site
- if peak detection matters, use the same device class and sampling interval across corridor and control sites, or flag the difference explicitly in the notes
- do not treat spot measurements as if they captured temporal variation; if time variation matters, use repeated sampling or a different method family
- if the instrument uses a single-axis probe, take sequential x/y/z readings on a nonconductive stand for repeatable spot measurements
- note whether the setup is stationary or moving, because stand-based spot measurements suppress motion-gradient fields and are not comparable to moving-person or scan data
- for utility-adjacent or industrial equipment sites, record operating state (on/off, load level, maintenance, quiet/active) at each stop
- for utility-adjacent stops, also log `source_state` separately from `source_class` so transient conditions like energized, idle, under maintenance, active construction, AMI meter exchange, or brief interruption do not get blended into the baseline
- for utility-adjacent stops, also log `construction_context` separately from `source_state` so AMI exchange, system-upgrade construction, and routine service activity do not get collapsed into one bucket
- for utility-adjacent stops, also note access constraints separately from both exposure and source context: blocked driveway, blocked lane, sidewalk detour, right-of-way work zone, locked gate, or no-access stop
- for utility-adjacent stops, also note `workzone_location` separately from `access_constraints` and `access_method` so public right-of-way work, yard work, street work, or a behind-the-gate stop do not get blended together
- for utility-adjacent stops, also note whether nearby equipment is energized, loaded, throttled, idle, or under maintenance, plus any visible substations, transformer pads, storage equipment, or buried-utility constraints
- if an identifiable source is present, estimate and log `source_distance_cm` or `source_distance_m` from the measurement point, using a repeatable reference point rather than a guess
- for close-proximity point sources, predefine a simple distance ladder such as contact, ~30 cm, ~61 cm, and beyond when the geometry and safety allow it; use the same ladder at corridor and control sites so apparent gradients are comparable instead of improvised
- if walls, equipment, or safety constraints block that ladder, record the closest normal worker standing position and the obstruction instead of inventing a fake distance
- if a source is being monitored directly, also log `source_dwell_seconds` so analysis can separate proximity from time spent near the source
- if an AMI rollout reveals a meter-integrity or billing anomaly, log it as utility-ops context, not exposure
- for AMI visits, record whether the stop is electric AMI or gas AMI, whether a meter exchange was needed, and whether the stop was a module install, meter exchange, or door-hanger-only follow-up
- for AMI visits, record the actual notice window as route context: electric AMI installs are typically noticed about 30 days ahead, gas AMI module installs are typically noticed about 30 days ahead, and gas meter exchanges (if needed) are notified at least 14 days in advance; immediate-maintenance cases may skip the letter and use a door knock and door hanger instead
- for AMI visits, keep access requests separate from exposure: note whether the technician asked to enter because the meter was inside or behind a locked gate, and log any no-access or door-hanger-only outcome as workflow context, not as signal
- for AMI visits, if the stop involves a gas meter exchange, record whether a temporary shutoff and relight were required
- for AMI visits, if the route packet mentions the meter radio as low-power RF, log that as device context only; do not use it as a surrogate for ELF magnetic-field exposure
- for gas AMI exchanges, record whether a temporary gas shutoff or relight was required, and whether the technician requested access because the meter was inside or behind a locked gate
- for AMI visits, record the exact AMI subtype separately: electric AMI meter install, electric meter exchange, gas AMI module install, gas meter exchange, door-hanger-only, no-access, or immediate-maintenance follow-up
- for AMI visits, log install-time detail separately from exposure: electric AMI installs are typically under 15 minutes and usually do not require entry; gas AMI module installs are also short and normally do not require entry; if a gas meter exchange is needed, record the temporary shutoff and relight as separate workflow fields
- for AMI visits, record the notice window as route context: electric AMI installs are typically noticed about 30 days ahead, gas AMI module installs are typically noticed about 30 days ahead, gas meter exchanges are notified at least 14 days in advance, and immediate-maintenance cases may skip the letter and use a door knock and door hanger instead
- for AMI visits, keep access requests separate from exposure: note whether the technician asked to enter because the meter was inside or behind a locked gate, and log any no-access or door-hanger-only outcome as workflow context, not as signal
- for AMI visits, if the route packet mentions the meter radio as low-power RF, log that as device context only; do not use it as a surrogate for ELF magnetic-field exposure
- for gas AMI exchanges, record whether a temporary gas shutoff or relight was required, and whether the technician requested access because the meter was inside or behind a locked gate
- for AMI visits, keep the subcontractor name when available so QUES, InfraSource, NIPSCO staff, and other crews do not get merged into one generic field note
- if the stop is near active NIPSCO work, note whether it looks like an AMI meter exchange, system-upgrade construction, or routine service activity
- if the stop appears to involve preliminary utility work, document it explicitly: locating or marking underground utilities, photos/video of property state, or other pre-construction prep should go in `preliminary_work`, not exposure
- for East Chicago / Hammond / Whiting routes, check the current EPA Grand Calumet remediation projects page before leaving so active construction and monitoring windows stay tagged as route context
- for East Chicago / Hammond / Whiting routes, log `cleanup_phase` separately from `source_state` and `construction_context` with values like `late-2026 mobilization`, `active construction`, `monitoring`, or `completed`
- for East Chicago / Hammond / Whiting routes, also log `cleanup_subarea` separately from `cleanup_phase` so Junction Reaches, Eastern 5 Miles, and Lake George Canal East/Middle do not get collapsed into one cleanup bucket
- for East Chicago / Hammond / Whiting routes, split `cleanup_subarea` by actual project footprint where possible: Junction Reaches should follow the IHC south of Columbus Drive, the West Branch between Indianapolis Boulevard and the IHC, and the East Branch between the IHC and Kennedy Avenue; Lake George Canal should stay split into Middle vs. East
- for East Chicago / Hammond / Whiting routes, keep `managed_corridor_context` separate from `cleanup_phase` and `source_state` so active remediation, managed lands, and source operation do not get merged into one bucket
- for Hobart Marsh / Deep River routes, check the current NWI UWFP work plan before leaving and log the Hobart Marsh Conservation Area NIPSCO ROW separately from Duck Creek Tributary restoration and from exposure readings
- for Hobart Marsh / Deep River routes, also note that Duck Creek Tributary Phase 2 finished on 2024-11-11 and keep that completion timing separate from exposure values and from the HMCA ROW geometry
- for Hobart Marsh / Deep River routes, log Duck Creek Tributary Phase 2 completion as restoration context — the 2025 work plan says that construction finished on 2024-11-11 after the 2024 season — so route context does not get mistaken for EMF signal
- for Hobart Marsh / Deep River routes, treat the Hobart Marsh Conservation Area NIPSCO ROW as corridor context; the work plan says the ROW spans roughly 300 acres and 5.5 miles of NIPSCO fee-owned electric transmission ROW and gas transmission easements through managed natural areas, within a broader managed/protected area of about 2,000 acres
- for Hobart Marsh / Deep River routes, treat the NIPSCO ROW and restoration work as managed-corridor context unless a direct source is identified
- for Hobart Marsh / Deep River routes, distinguish Duck Creek Tributary restoration work from the NIPSCO ROW and log it as corridor context
- for Urban Waters routes, use the current EPA Program Contacts page as the first coordination check; Victoria Wittig, Ph.D. is the Northwest Indiana Urban Waters Ambassador at Purdue University Northwest in Hammond
- for Hobart Marsh / Deep River routes, add `managed_corridor_context` as a dedicated field so the NIPSCO ROW, DNR-managed lands, and Duck Creek restoration timing do not get fused into one bucket
- if the same hotspot keeps showing unusual ELF readings, add one waveform-characterization spot with a waveform-capable meter so static-field strength, harmonics, and non-sinusoidal features can be checked explicitly
- when you do that waveform follow-up, mount the probe on a fixed stand if possible, use the same placement each time, and keep the probe as close as safely possible to the worker location or source-adjacent standing position, ideally within about 0.3 m
- when you do that waveform follow-up, document the exact capture configuration and keep it stable: the NIOSH industrial-magnetic-field workflow used 512 data points per axis at 6142.5 Hz, which produces an 83.35 ms window covering five 60 Hz cycles; use that or an equivalent validated waveform setting
- when you do that waveform follow-up, record the exact meter model, usable frequency range, sample window length, sampling frequency, stand height, and probe-to-target proximity so the diagnostic sample stays interpretable
- when you do that waveform follow-up, flag any out-of-range capture, baseline shift, non-integral-cycle window, or motion artifact before the sample is promoted into the main dataset
- keep waveform-characterization readings separate from ordinary ELF magnitude spot checks and only use them on a small subset of repeated hotspots or source-adjacent stops
- for AMI visits, capture service-workflow metadata separately from source and exposure: advance notice, door hanger, access requested, meter exchange required, temporary gas shutoff, and relight required
- for East Chicago / Hammond routes, tag cleanup-window context separately from source class and construction context, especially Junction Reaches construction beginning Fall 2026 and Lake George Canal monitoring through 2027 with possible 2028 follow-on remediation
- if you need a single AMI scale reference for route planning, record whether it came from the NIPSCO overview page or the electric/gas subpage so the rounded program counts stay traceable

## Sampling schedule
- run the same route at the same times each day when possible
- capture at least one workday and one off-peak window
- keep a fixed dwell time at each point
- record weather, nearby operations, traffic, and any obvious source changes
- take a second reading at a few points in each session so transient spikes and meter drift are visible instead of imagined away
- for each stop, note likely source class: utility line, transformer/substation, industrial equipment, rail/traffic, or background
- do a quick pre-route walkthrough or recon where needed so congested utility corridors, access limits, and obvious source changes are not discovered halfway through the session
- check the current-year Urban Waters work plan during route planning when a stop may overlap a managed corridor or restoration site, then log that context separately from source class
- for Hobart Marsh / Deep River routes, check the current NWI UWFP work plan during route planning so restoration and ROW context stay separate from source class
- for spot measurements, keep probe height and orientation consistent across repeated readings; if the instrument supports it, use the same mount or stand each time
- keep measurement mode consistent within a session and do not pool stationary spot readings with walking scans or body-worn samples
- log a short corridor-context tag when relevant, such as Grand Calumet AOC, East Chicago mixed-use cleanup area, or Urban Waters corridor, so location context stays separate from source class
- log access constraints and construction context immediately while on site so route issues do not get reconstructed from memory later
- if a utility crew is doing preliminary marking or property documentation, log that as `preliminary_work` so it does not get mistaken for a source-state change
- when planning East Chicago or Hammond routes, check the current Grand Calumet remediation projects page and log any cleanup-window context separately from source class

## QA/QC
- verify meter calibration status before use
- repeat a few points twice per session
- log battery level, settings, and operator name
- flag anything that looks like a transient spike or setup error
- if possible, include one background check away from obvious sources before and after the route

## Safety and permissions
- use public areas or get permission first
- do not trespass for data
- avoid interpreting access denial as data