The April 2026 Cambridge preprint (arXiv:2603.20897) analyzed each data center individually in isolation. It looked at >6,000 separate facilities worldwide and measured the temperature rise around each one independently. The authors explicitly state they focused on “the locations of the main AI hyperscalers” and computed temperature trends “within circular regions centered on each AI hyperscaler.”

There is zero discussion of:

• Clustering of multiple data centers
• Overlapping heat-island radii
• Cumulative or additive effects
• Whether two or more centers near each other would create a stronger combined heat island

What We Can Reasonably Infer

While the paper itself is silent, basic physics of heat islands (urban heat islands, industrial heat islands, etc.) strongly suggests the effects would be compounded in areas where the 10 km (6.2-mile) radii overlap:

• Each data center adds its own waste heat to the local atmosphere and land surface.
• In the overlapping zone, the temperature increases from each facility would add together.
• Example: – One center alone → +3.6 °F average rise at 2 miles away – Two centers 3 miles apart → the overlap area could see roughly +7.2 °F (or more in extreme cases) because the heat plumes reinforce each other.

This is exactly how traditional urban heat islands work when multiple heat sources (buildings, roads, factories) are close together — the effect is not linear but can be significantly amplified.

Why This Matters for Indiantown / Silver Fox 606

Public records already reference at least four data-center proposals being discussed along the same Fox Brown Road / Venture Park corridor. If two or more of these hyperscale facilities end up within a few miles of each other:

• The 6.5-mile heat island zones would heavily overlap.
• The compounded warming in the shared area would likely be greater than the single-center averages reported in the Cambridge study.
• Residents, farms, wetlands, and air quality in the overlap zone would face an even stronger microclimate shift.

The Cambridge study itself calls for more research on these “local microclimate zones” and their influence on communities. It does not give us the answer for clustered facilities — which is exactly why a site-specific cumulative heat-impact study should be required before any approvals in Indiantown.

Here is a clear, easy-to-read table showing the estimated average land-surface temperature increase at 1-mile increments from the epicenter of the Silver Fox 606 project (based on the April 2026 Cambridge University study, arXiv:2603.20897).

The study describes a step-function-like decay with the strongest effect near the facility and gradual fading out to 6.2 miles. The values below are derived directly from the reported averages and radial decay pattern (peak at center, ~+1 °C at 4.5 km / 2.8 mi, ~30 % intensity at 7 km / 4.3 mi, and detectable fading to 10 km / 6.2 mi). I have interpolated conservatively for each 1-mile band.

Distance from Epicenter	Average Temperature Rise (°F)	Intensity Level	Potential Exposure Notes
0 – 1 mile	+3.6 °F (up to +16.4 °F in extreme cases)	Peak	Highest impact zone – immediate vicinity of the facility
1 – 2 miles	+3.0 °F	Very Strong	Still very noticeable; significant daily heat increase
2 – 3 miles	+2.4 °F	Strong	Clear warming; affects most of central Indiantown
3 – 4 miles	+1.8 °F	Strong–Moderate	Noticeable on hot Florida days
4 – 5 miles	+1.3 °F	Moderate	Measurable but beginning to fade
5 – 6 miles	+0.9 °F	Moderate–Fading	Lower but still detectable year-round
6 – 6.5 miles	+0.5 °F	Fading but Detectable	Outer edge of the heat island; minor but persistent

Key Facts from the Study

• These are sustained average increases in land-surface temperature after the data center begins operations.
• The effect starts immediately and persists for years.
• It creates a new local microclimate that raises air temperatures, energy demand (more AC use), health risks, and stress on wetlands/ecosystems.
• If multiple data centers are built close together, these values would be compounded in the overlapping zones (the study did not model clusters).