PM2.5 -22%
Finer particles inside the yard compared to the curb. Sanctuary #004 Montrose, summer 2025.
Nowak et al., 2014.
Labs · Methodology
Six of the lenses we instrument.
Numbers measure one face of a garden. A sensor cannot read what a Custode reads. But the numbers are still worth having, for honesty, for limits, for the conversation with city regulators and insurance tables that do speak in numbers. These are six instrumentations we currently run, site by site, with the citations and limitations stated plainly.
Instrumentation tiles
9 °F
Cooler to the touch than the nearest pavement, same location, same season.
Santamouris et al., 2017.
-3.4 dB
Quieter at the front door compared to a turf-and-driveway baseline across the street.
Van Renterghem, 2014.
+41%
More songbird species recorded at Sanctuary #004 than at the paired reference yard.
Narango et al., 2018.
17 species
Distinct pollinator species documented in one Holon yard over a single bloom season.
Xerces Society surveys, 2024-25.
14×
Stormwater infiltration rate compared to adjacent turf, same rain event.
EPA Heat Island Compendium updates, 2024.
What a sensor reads, and what it misses.
Every tile above is a partial instrument. The PM2.5 meter does not know the child whose asthma improved. The dB meter does not know that the quiet came with the mockingbird. The sensor reads what a sensor can read. The Custode reads the rest. Both of those readings are in the Dashboard, named as what they are.
The open notebook.
Raw data, calibration notes, and limitations are published on a rolling basis. We trust what we trust; we flag what we do not; what we outsource to researcher partners, we name.