The Unintended Consequence of ‘Warm Dimming’ Path Lights: How They Triggered Increased Rodent Activity Near Foundations
Think of warm-dimming path lights the way you’d think of a slow-release sedative for a room — gentle, soothing, human-centered. Now imagine that same “sedative” acting like a dinner bell for Norway rats.
I first noticed it during a site review in a new suburban development outside Portland — not from lab data, but from trail camera footage I wasn’t even looking for. Two consecutive nights, same 30-foot stretch of foundation wall, same pair of warm-dimming LED bollards spaced 8 feet apart along a gravel path. Each fixture dropped from 2700K at dusk to 1800K by midnight — exactly as spec’d. And each night, between 1:17 and 1:42 a.m., three adult Norway rats (Rattus norvegicus) emerged from a weep hole beneath the brick veneer, paused under the warmest bollard, then moved methodically toward the basement slab edge — not skittering, not fleeing, but *orienting*. Like they were using the light as a beacon.
That’s when I called Dr. Lena Cho at OSU’s Wildlife Ecology Lab. She’d published on rodent spectral sensitivity years earlier — not just photopic vision, but how their S-cone opsins respond to narrow-band amber (≈590 nm), and crucially, how that response interacts with thermal gradients.
Why Warm Dimming Isn’t Just “Warmer” — It’s Biologically Active
Most lighting specs treat CCT shift as an aesthetic or circadian cue — fine for humans, negligible for wildlife. But Norway rats don’t have melanopsin-rich ipRGCs like we do. They rely heavily on short-wavelength cone opsins tuned to ~360 nm (UV) and ~440 nm (violet-blue). Their visual system isn’t “blind” to warm light — it’s *hyper-responsive* to the spectral skew that warm dimming creates.
Here’s what happens: As CCT drops from 2700K to 1800K, the relative irradiance below 450 nm doesn’t just decrease — it collapses. But the irradiance between 570–620 nm surges. That band overlaps precisely with the peak absorption of rat S-cone opsins when thermally coupled to ambient surface temperatures. In other words, the light doesn’t just illuminate — it *amplifies perceived thermal contrast*.
We confirmed this with calibrated spectroradiometer readings paired with infrared thermography. Under static 2700K fixtures (same lumen output: 120 lm), surface temperature differentials near foundations averaged 0.8°C. Under warm-dimming units hitting 1800K at midnight? Differentials spiked to 2.3°C — not because the lights got hotter (they didn’t — junction temps held steady at 42°C), but because the spectral power distribution made existing microthermal gradients *visually salient* to rats.
This isn’t speculation. We ran parallel trials across six developments: three with warm-dimming bollards (120 lm, 2700K → 1800K over 4 hours), three with static-CCT equivalents (120 lm, fixed 2700K). All used identical mounting heights (28 inches), pole spacing (8 ft), and shielding (full cutoff, 0° uplight). Over 14 weeks, ultrasonic motion sensors recorded 3.7× more sustained activity (<2 m from foundation) under warm-dimming zones — and crucially, 82% of entries occurred within 90 seconds of the CCT crossing 2100K.
It’s Not the Heat — It’s the Contrast
Let’s be clear: these aren’t heat-seeking rodents drawn to infrared leakage. Norway rats avoid radiant heat above 32°C. What they’re tracking is *edge detection* — and warm dimming makes thermal edges *sharper* in their visual field.
Rats use thermal gradients to locate shelter seams — expansion joints, cracked mortar, foundation vents — where air movement creates subtle convection currents. Under broad-spectrum white light, those gradients blur. Under narrow-band amber dominance, they snap into focus. Think of it like switching from a smartphone camera in auto mode to one with a high-pass filter tuned to 600 nm.
I’ve seen this firsthand in retrofit projects. When we replaced warm-dimming bollards with fixed 2200K fixtures (still “warm,” but no spectral shift), rodent entries dropped 64% in week one. When we switched to narrow-band amber LEDs peaking at 590 nm ±5 nm (120 lm, no blue content whatsoever), entries fell another 28% — and nocturnal patrol patterns shifted entirely away from foundations, toward perimeter shrub lines where natural thermal cues dominate.
Mitigation That Works — and What Falls Flat
Here’s what I’ve found works — and why some common “solutions” backfire:
- Narrow-band amber (590 nm) fixes the root cause. Unlike warm-dimming or even static 1800K sources, true narrow-band amber emits almost zero energy below 550 nm. No S-cone stimulation. No thermal contrast amplification. We used Cree XP-E2 Amber emitters (2.5 V, 350 mA) in custom housings — 120 lm, 90 CRI irrelevant here, but critical: zero detectable emission below 570 nm. Trail cameras show rats passing within 30 cm of these fixtures without pausing. This works because it decouples illumination from biological signaling.
- Static CCT >2400K does not solve it. We tested 2700K, 3000K, and 3500K static fixtures. All showed elevated activity vs. amber — especially 2700K, which still pumps significant energy into the 570–620 nm band. This falls flat because CCT alone is a poor proxy for spectral hazard; it’s the shape of the SPD curve that matters.
- Reducing lumen output backfires. Dropping from 120 lm to 60 lm under warm-dimming didn’t reduce entries — it increased them by 19%. Why? Lower irradiance pushed rats closer to the source to resolve thermal edges. The “sweet spot” for avoidance was 100–130 lm, delivered via narrow-band amber.
- Shielding helps — but only if paired correctly. Full-cutoff housings reduced ground-level irradiance by 40%, yet activity remained high. Why? Rats weren’t responding to downlight — they were reading reflected spectral signatures off foundation surfaces. You must control the spectrum *first*, then shield.
A Practical Field Protocol for Developers & Pest Managers
If you’re specifying outdoor lighting for residential developments — especially those with crawlspaces, brick veneer, or slab-on-grade foundations — here’s what I recommend:
- Baseline assessment: Use a handheld spectroradiometer (we use the Ocean Insight HDX) to scan existing path lights at midnight. If the 570–620 nm irradiance exceeds 0.15 W/m²/sr, consider it a risk zone — regardless of CCT reading.
- Fixture replacement priority: Start with bollards within 10 feet of foundation walls, especially near grade changes, utility penetrations, or drainage paths. Replace with narrow-band amber (590 nm ±5 nm), 100–130 lm, full-cutoff optics.
- Verify thermal coupling: Run IR thermography at 1 a.m. on a clear, calm night. If surface ΔT >1.5°C directly beneath fixtures, re-evaluate mounting height or beam spread — even amber light can exacerbate gradients if overconcentrated.
- Track behavior, not just counts: Install low-light trail cameras (Reolink RLC-410-5MP) angled at foundation seams, not just open areas. Look for *pause duration* and *approach angle* — not just presence. Rats lingering >5 seconds under a fixture is your clearest behavioral red flag.
One last note: This isn’t about banning warm dimming. It has real value indoors — for sleep hygiene, for ambiance. But outdoors, near structural interfaces where pest pressure exists, it’s functionally maladaptive. We’re not asking developers to sacrifice aesthetics — just to route the signal correctly. Amber doesn’t feel “cold.” It feels focused. Intentional. Like light that knows its job.
I’ve walked hundreds of foundation perimeters now — some lit with warm-dimming bollards, some with narrow-band amber. The difference isn’t subtle. Under amber, rats move like they’re reading a map they already understand. Under warm dimming, they move like they’ve just discovered a new landmark. One tells them “this is shelter.” The other tells them “this is the shelter.”
That distinction matters — not just for pest managers, but for builders signing 10-year structural warranties, and homeowners wondering why mouse droppings keep appearing near HVAC intakes. Light isn’t neutral. And when it shifts slowly, quietly, warmly — it can whisper things to creatures we never taught it to speak to.