# Research Note — 2026-05-31 ## Confirmed facts - NIOSH’s EMF manual treats walkthrough surveys, spot measurements, area monitoring, source surveys, and personal monitoring as distinct exposure-assessment families, and says spot measurements do not capture temporal variation. [^1] - NIOSH’s waveform-measurement paper says ordinary ELF magnitude readings can miss static-field, harmonic, and non-sinusoidal features near worker locations, and that waveform-capable spot measurements are needed for a complete magnetic-field assessment. [^2] - NIPSCO’s electric AMI page says electric AMI installs are phased, typically take less than 15 minutes, may require access if the meter is inside or behind a locked gate, and may leave a door hanger when the customer is not home. [^3] - NIPSCO’s electric AMI page also says meter exchanges can be needed to complete a module install and may be handled by a subcontractor or NIPSCO employee. [^3] - NIPSCO’s gas AMI page says gas AMI installs are phased, may require access if the meter is inside or behind a locked gate, may leave a door hanger, and meter exchanges can require a temporary gas shutoff plus a relight. [^4] - EPA’s Grand Calumet River AOC page says Junction Reaches construction is scheduled to begin in Fall 2026, and Lake George Canal monitoring continues through 2027 with possible 2028 follow-on remediation. [^5] ## Practical takeaway - The pilot now has a cleaner split between screening and exposure: route-screen metrics should stay screening-only, while repeated hotspots can get a rare waveform-capable follow-up without contaminating the main ELF tables. - AMI activity should stay in route metadata, not exposure metadata: electric vs gas, install vs exchange, access requested vs no access, and relight/shutoff vs none. - Grand Calumet cleanup timing still belongs in context fields, not as a signal bucket. ## What changed in the project files - Tightened `file 'nwi-emf-pilot/02-field-protocol.md'` with waveform-follow-up QA/QC and probe-handling notes. - Tightened `file 'nwi-emf-pilot/03-analysis-plan.md'` so route-screen metrics and waveform diagnostics stay separate from the main exposure tables. ## Update — 2026-05-31 ### Confirmed facts - The 2025 NWI UWFP work plan says the ambassador role is hosted through Purdue University Northwest, and it explicitly names the Hobart Marsh Conservation Area NIPSCO ROW in the Hobart Marsh and Deep River section. [^8] - The same work plan describes Duck Creek Tributary restoration as part of the Hobart Marsh and Deep River area, so that work should be logged as corridor context rather than exposure signal. [^8] - EPA's current Grand Calumet River AOC page says Junction Reaches construction is scheduled to begin in Fall 2026 and Lake George Canal monitoring continues through 2027 with possible follow-on remediation in 2028. [^9] ### Practical takeaway - Hobart Marsh / Deep River stops should now be tagged with separate managed-corridor and utility-row context fields. Otherwise the ROW and restoration work will masquerade as signal — which is exactly how bad data gets dressed up as insight. - Grand Calumet cleanup timing still belongs in route context, not the exposure table. ## Update — 2026-05-31 (local context refresh) ### Confirmed facts - The 2025 NWI UWFP work plan says Duck Creek Tributary Phase 2 construction finished on 2024-11-11 after the 2024 season, so that work should be logged as restoration / corridor context rather than exposure signal. [^8] - The same work plan says the Hobart Marsh Conservation Area NIPSCO ROW is about 300 acres and 5.5 miles of NIPSCO fee-owned electric transmission ROW and gas transmission easements through managed natural areas, so its geometry belongs in route context. [^8] - EPA’s Grand Calumet River AOC page still says Junction Reaches construction is scheduled to begin in Fall 2026 and Lake George Canal monitoring continues through 2027 with possible 2028 follow-on remediation, so those windows remain route context, not EMF signal. [^9] ### Practical takeaway - The next pass should keep Hobart Marsh / Deep River geometry, restoration timing, and NIPSCO ROW status in dedicated context fields. Nice place to stop pretending the map is the measurement. - Grand Calumet cleanup timing still belongs in context fields, not the exposure table. ## Update — 2026-05-31 (AMI workflow refresh) ### Confirmed facts - NIPSCO’s electric AMI page says the electric rollout is phased over about three years starting in March 2024 for roughly 490,000 electric residential and business customers, with advance mail notice and door-hanger follow-up when access fails. [^3] - NIPSCO’s gas AMI page says the gas rollout is phased over several years for roughly 870,000 gas customers, with advance mail notice, door-hanger follow-up when access fails, and a temporary shutoff plus relight when a meter exchange is required. [^4] ### Practical takeaway - AMI stops should now carry `advance_notice_lead_time` and `visit_duration` as route-context fields, because those details explain access patterns without pretending to be exposure signal. - The clean split is still the point: AMI workflow is logistics, not EMF. ## Update — 2026-05-31 (Hobart Marsh / Deep River corridor refresh) ### Confirmed facts - The 2025 NWI UWFP work plan says Duck Creek Tributary Phase 2 construction was completed on 2024-11-11, so the project is a finished restoration context for route tagging rather than an active construction signal. [^8] - The same work plan says the Hobart Marsh Conservation Area NIPSCO ROW covers about 300 acres and 5.5 miles of NIPSCO fee-owned electric transmission ROW and gas transmission easements, and that NIPSCO initiated planning and restoration efforts there in 2021. [^8] - The work plan also says Hobart Marsh is managed by the Indiana DNR Division of Nature Preserves and that the managed/protected land complex totals about 2,000 acres, so the corridor context is broader than the utility ROW alone. [^8] - EPA’s current Grand Calumet River AOC page still says Junction Reaches construction is scheduled to begin in Fall 2026 and Lake George Canal monitoring continues through 2027 with possible 2028 follow-on remediation. [^9] ### Practical takeaway - Hobart Marsh / Deep River stops should split the map layer into three separate notes: utility ROW geometry, restoration timing, and broader managed-corridor context. Otherwise the route description starts doing improv comedy with the data. - The next practical improvement is a dedicated `managed_corridor_context` or `corridor_geometry` field on the field log, so the NIPSCO ROW, DNR-managed lands, and Duck Creek restoration timeline do not get fused into one bucket. ## Update — 2026-05-31 (Grand Calumet subarea refinement) ### Confirmed facts - EPA’s current Grand Calumet River AOC page splits Junction Reaches into the Indiana Harbor Ship Canal south of Columbus Drive, the West Branch between Indianapolis Boulevard and the IHC, and the East Branch between the IHC and Kennedy Avenue; it also says construction is scheduled to begin in Fall 2026. [^9] - The same page says the Lake George Canal Middle and East project covers the East Portion from Indianapolis Boulevard to the CSX railroad bridge and the Middle Portion from the CSX bridge to the BP land bridge in Hammond, with monitoring continuing through 2027 and possible follow-on remediation in 2028. [^9] - The same page also says the Dune, Swale and Shelf Wetlands Restoration project is anticipated to be completed by 2026. [^9] ### Practical takeaway - `cleanup_subarea` is now worth keeping as a dedicated field for East Chicago / Hammond routes. If the project keeps one cleanup label for three different work footprints, the map will look tidy right up until it starts lying. [^7]: https://stacks.cdc.gov/view/cdc/188216/cdc_188216_DS1.pdf [^8]: https://www.epa.gov/system/files/documents/2025-05/2025-nwi-uwfp-work-plan.pdf [^9]: https://www.epa.gov/great-lakes-aocs/remediation-and-restoration-projects-grand-calumet-river-aoc ## Update — 2026-05-31 (context-field refinement) ### Confirmed facts - The 2025 NWI UWFP work plan separates Duck Creek Tributary restoration, Hobart Marsh managed land, and the Hobart Marsh Conservation Area NIPSCO ROW into distinct sections. [^8] - EPA’s Grand Calumet River AOC page separates Junction Reaches, Eastern 5 Miles, Lake George Canal Middle/East, and Dune / Swale restoration as separate project footprints and timelines. [^9] ### Practical takeaway - The next field-log pass should carry `managed_corridor_context` and `cleanup_subarea` as separate fields. That keeps ROW geometry, restoration timing, and active cleanup windows from getting mashed into one tidy little lie. ## Update — 2026-05-31 (AMI workflow fields) ### Confirmed facts - NIPSCO’s electric AMI page says customers are notified by mail about 30 days in advance, installations typically take less than 15 minutes, technicians may request access if the meter is inside or behind a locked gate, and a door hanger is left when the customer is not home or access fails. [^3] - NIPSCO’s gas AMI page says gas module installs are phased, meters may need to be exchanged, the work can require a temporary gas shutoff and relight, and a door hanger is left when the customer is not home or access fails. [^4] ### Practical takeaway - The next template pass should carry AMI subtype and service-workflow fields explicitly, because notice lead time, visit duration, access, shutoff, and relight explain route logistics without pretending to be EMF signal. - That keeps the pilot honest: logistics in one bucket, exposure in another. The data already has enough confusion without help from the spreadsheet. ## Update — 2026-05-31 (distance-ladder geometry refresh) ### Confirmed facts - NIOSH's metal-detector HHE measured magnetic fields inside the arch at four vertical points and outside at contact, 30 cm, and 61 cm, and when walls or equipment blocked those distances, the investigator measured at the closest location where a worker normally stood. [^6] - The same report used EMDEX II, Multiwave System II, and Holiday HI-3637 instruments; the waveform-capable Multiwave system captured static-field and sub-RF content, FFT spectra, polarization, and total harmonic distortion, and brief close-pass readings reached about 19.9 to 38.2 mG. [^6] ### Practical takeaway - The pilot should keep the contact / ~30 cm / ~61 cm ladder, but when geometry blocks it, the field log should record the closest normal worker standing position plus the obstruction. That keeps the stand-off story honest instead of pretending every site has open runway access. - If a pass-by source only spikes while moving through the field, that's a good candidate for a waveform-capable follow-up. The meter should match the question, not the other way around. ## Sources used - [^1]: https://stacks.cdc.gov/view/cdc/5254/cdc_5254_DS1.pdf - [^2]: https://stacks.cdc.gov/view/cdc/198316/cdc_198316_DS1.pdf - [^3]: https://www.nipsco.com/services/work-in-your-neighborhood/advanced-metering-ami/electric-ami - [^4]: https://www.nipsco.com/services/work-in-your-neighborhood/advanced-metering-ami/gas-ami - [^5]: https://www.epa.gov/great-lakes-aocs/remediation-and-restoration-projects-grand-calumet-river-aoc - [^6]: https://www.cdc.gov/niosh/hhe/reports/pdfs/1996-0113-2644.pdf ## Update — 2026-05-31 (AMI workflow specificity refresh) ### Confirmed facts - NIPSCO’s electric AMI page says the electric rollout is phased over three years starting in March 2024, with about 490,000 electric customers covered, roughly 30 days of advance mail notice, typical install time under 15 minutes, and access requests only when the meter is inside or behind a locked gate. [^3] - NIPSCO’s gas AMI page says the gas rollout is phased over several years for about 870,000 gas customers, that module installs also get about 30 days of advance mail notice, and that meter exchanges require at least 14 days of notice plus a temporary gas shutoff and relight. [^4] ### Practical takeaway - The field log should keep AMI subtype, advance notice, visit duration, access request, temporary shutoff, and relight as separate route-context fields. That is the difference between logistics and exposure — and spreadsheets love to pretend those are the same thing.