The plant treated wood with pentachlorophenol—PCP—and chromated copper arsenate, and those two choices set the entire monitoring list. PCP is acutely toxic and capped in drinking water at one part per billion. Technical PCP made before the late 1980s carried dioxins and furans as manufacturing impurities, and those are among the most persistent and toxic compounds known: they don't break down on human timescales, they build up in living tissue, and the toxic-equivalent math exists precisely because trace amounts still count.

The arsenate side leaves arsenic and chromium in the soil, and arsenic is a known human carcinogen. None of this is the kind of contamination that quietly weathers off in a few years—which is the whole reason a place like this gets watched for decades, and why the list is exactly these chemicals and not something more convenient.

It has played out before. In Pensacola, the Escambia wood-treating plant (creosote and PCP) left dioxin contamination the EPA finally addressed by relocating about 358 households—one of the largest Superfund relocations in the country. In Texarkana, a subdivision called Carver Terrace was built on a former Koppers wood-treating site; by 1993 the government had bought out and relocated all 79 homes. Closer to home, Southern California Edison's Visalia Pole Yard ran the same PCP-and-creosote chemistry—down to the same octachlorinated dioxin impurity that turns up at Cloverdale—and pushed contamination roughly 100 feet into groundwater; it took years of aggressive thermal treatment to bring the water back to drinking standards before the site could be delisted.

On scale, Cloverdale is the mildest of the four. The contamination is smaller and more contained—no sixty-foot mound of poisoned soil, no neighborhood already built on it, and the dioxin that turned up here was low. All three of the others were federal Superfund sites; this one was handled by the state.

But each of those places ended the same way: either the people were moved off the ground, or the ground was cleaned to drinking-water standards before anyone signed off. Cloverdale is being signed off with the contamination still in the bedrock, the monitoring switched off, and the densest use yet drawn on top. The figure that should worry a neighbor isn't the parts per billion. It's the gap between how little was finished here and how much is planned.

◆

The Esmeralda plan proposes 605 homes, a school, senior care, and a park on this ground. Those are commitments measured in decades. The contamination beneath it is measured the same way.

Everything below is drawn from that file— groundwater sampling events between and . The map already shows where the wells are. What follows is when they were read, how far over the limit they ran, how deep it sits, whether it's actually breaking down, and what the closure record leaves out.

1 · When they looked

Every tick is one groundwater sampling event. The line above it is pentachlorophenol (PCP) at well MW‑37. Readings ran twice a year for a while, stopped for three years, thinned to one event in 2024, and none at all in 2025—then the No Further Action letter closed the case in January 2026 and ended the monitoring program. The last thing anyone measured at MW‑37 was 130 µg/L. on file

Source: GeoTracker EDF chemistry export, all water‑matrix samples at monitoring wells. Markers: covenant recorded 8/28/2013 (Sonoma Co. #2013088084); NFA letter 1/21/2026.

2 · How far over the limit

Four wells, PCP over time, on a log scale against the 1 µg/L drinking‑water limit (the dashed line). MW‑37 climbs through the final four events to 130 µg/L — 130× the limit—at closure. MW‑13 fell to non‑detect (attenuation does happen here). MW‑27 sat high for years, then declined—the last reading on record, in 2020, was about 0.1 µg/L. on file

detected PCP non‑detect (at reporting limit) 1 µg/L MCL

Source: EDF, parameter PCP, max value per event. Log y‑axis so the limit and the peaks fit one frame.

3 · The whole table at a glance

Every well that ever exceeded the limit (rows, worst at top) by year (columns). Color is the PCP reading against the 1 µg/L limit. Blanks are years that well wasn't sampled—note how the grid thins toward the right and stops at 2024. on file

Source: EDF, parameter PCP, max per well per year. Non‑detects shown as at‑or‑below limit.

4 · Whether it's actually breaking down

The closure rests on natural attenuation—the idea that PCP is degrading on its own. If it were, the parent compound would fall and its dechlorination daughters would build up. At MW‑37 the opposite shows: PCP is almost the entire chlorophenol mass at every sampling, the only daughters present are minor tetrachlorophenols, tri‑ and di‑chlorophenols are non‑detect, and the total is rising. Consistent with limited breakdown, not advanced cleanup. on file

PCP (parent) tetrachlorophenols tri‑/di‑chlorophenols (non‑detect)

Source: EDF, MW‑37, full chlorophenol suite events (2010, 2014, 2022, 2024).

5 · How deep it sits

PCP at the MW‑37 cluster by depth, from the 2010 characterization. The contamination is in deep bedrock groundwater, 80–88 feet down—beneath the ground proposed for a park and sports fields on the old log‑deck area. The covenant restricts shallow soil and bars new wells, but the plume is below all of that. on file

Depth‑resolved PCP: EDF discrete‑interval samples (MW‑37, 80 ft and 88 ft, Apr 2010). The overburden/bedrock band is schematic; the 5‑ft line is the covenant soil datum.

6 · What the closure leaves out

Dioxins and furans are the signature impurity of pre‑1987 technical PCP. They were measured here only in 2008—two rounds at two wells (MW-25 and MW-27)—and never again, and they don't appear in the 2026 closure record at all. The levels are low and the most toxic congener (2,3,7,8‑TCDD) was non‑detect; the detected mass is mostly OCDD, which is far less toxic. The point isn't a big number. It's that the one thing you'd most expect to track at a former wood‑treater was checked a single time, sixteen years before closure, and left out. on file

Source: EDF, method SW8290 (high‑resolution dioxin/furan), MW‑25 & MW‑27, May & Nov 2008, units pg/L. TEQ computed with WHO‑2005 toxic‑equivalency factors, non‑detects counted as zero. Reference: the 2,3,7,8‑TCDD drinking‑water limit is 30 pg/L.