CLASSIFICATION: MODEL ESTIMATE · Confidence: Low. Every number in this chapter is a mass-balance thought experiment built from the published USDA dip formula and from soil-arsenic concentrations measured in Australia and the US Southeast, not in California. No California dip-site soil has ever been tested. These figures state what is plausible, not what is. They could be wrong by an order of magnitude (~10×) in either direction. Nothing here asserts that any California soil is contaminated. Grades: [A1/A2] for inputs; the model output is MODEL ESTIMATE throughout.
Computed by scripts/arsenic_mass_balance.py (assumptions stated inline; reproducible).
| Quantity | MODEL ESTIMATE |
|---|---|
| Elemental arsenic per 8-lb charge | 6.1 lb (dip-fluid strength ~1,450 mg As/L) |
| Arsenic standing in one full vat at a time | 3,000 gal → 36 lb; 4,000 gal → 48 lb |
| Lifetime throughput per vat (poured through, replenished over 3-10 yr) | ~145-1,200 lb As; central ~470 lb |
| Arsenic in the ground per heavily-used vat (vat + apron + drain zone at 500-3,000 mg/kg) | ~100-500 lb As (low 32 / high 857) |
| Statewide ceiling (total white arsenic ever purchased by the program) | ~33,000-165,000 lb (15-75 t) As |
| Statewide near-vat concentrated fraction | ~6,600-66,000 lb (3-30 t); central ~10 t |
| Implied number of vats | ~50-150 (short, small program; Florida had 3,000+) |
The statewide figure is bounded by an independent cross-check: a throughput analysis (USGS Arsenic Historical Statistics DS140 + the USDA cattle-drag-out method) caps total white arsenic ever used by the whole California program at ~15-75 t elemental As across ~6-8 seasons. No more arsenic can sit in California soil than was ever purchased and poured. Two independent methods agree on ~3-30 t concentrated near vats, central ~10 t; the remainder dispersed thinly across rangeland via cattle drag-out or was discarded as spent fluid. The earlier ~20-100 t statewide figure assumed 200-1,000 vats and is superseded, it was inconsistent with the ceiling.
The per-site numbers are the ones that matter for screening any single location. A dip site at the empirical anchor of 500-3,000 mg/kg would sit far above every yardstick:
| Yardstick | Value | A dip site (500-3,000 mg/kg) is… |
|---|---|---|
| EPA residential screening level (RSL) | 0.68 mg/kg | ~700-4,400× |
| California DTSC residential | 0.11 mg/kg | far above |
| California regional background | ~11-12 mg/kg | ~40-270× background |
Because California background arsenic already sits near a 1×10⁻⁴ cancer risk, the state regulates arsenic against background, which is precisely why a point source rising 40-270× above background is the thing a soil test would reveal [A2]. Arsenic is an IARC Group 1 human carcinogen (skin, lung, bladder) and does not degrade; the real-world concern is chronic, not acute, with children the sensitive receptor via incidental soil ingestion, and the inhalation pathway largest during mass grading/earthmoving.
Mass grading neither creates nor destroys arsenic: it can dilute a hot spot (lowering mg/kg while spreading the mass) or relocate it intact. None of this is knowable for any specific site without (a) locating the vat and (b) testing the soil, the two things never done in California. The material toxicity of the chemical in the bag (one 8-lb charge ≈ tens of thousands of theoretical adult lethal doses) is context only, it is not an exposure or harm estimate; soil is dilute, arsenic sorbs to Fe/Al oxides, and is only partly bioavailable. This entire chapter is a MODEL ESTIMATE and is not a measurement.
Source registry: S-USDA-C174 [A1]; S-NSW-DIP [A2]; DTSC HHRA Note 11, EPA RSL [A2]; USGS DS140
[A1]. Model output: MODEL ESTIMATE, Confidence Low, reproducible via
scripts/arsenic_mass_balance.py. Full grading in chapter 73.