Borrow Source Approval With Straddling Atterberg Results
A contractor submits a borrow pit as the source of select fill for a 40,000-yd³ grading job. The specification requires PI of 15 or less and LL of 35 or less in the top 2 ft of subgrade, and PI of 25 or less below that. The submittal contains two samples: one at LL 31, PI 12, and one at LL 41, PI 22, both taken from the upper 4 ft of the same 15-ft face about 60 ft apart. It also includes a single Proctor curve and one gradation. The pit exposure shows three visibly different strata and a color change at about 8 ft. The contractor wants written source approval Friday so hauling can start Monday, arguing that the average of the two samples (LL 36, PI 17) is 'close enough' and will blend in the fill. No stockpiling by stratum is planned.
1. Can you approve the source on this submittal? What sampling would you require?
Read the worked answer
No. Two samples 60 ft apart on one face of a 15-ft pit with three visible strata cannot characterize 40,000 yd³, and the results are themselves the warning: PI 12 versus PI 22 straddles the 15 limit for the top 2 ft, which is real material variability rather than test scatter. A single Proctor curve and one gradation give no basis for a family of curves.
Require a sampling plan that treats each visible stratum separately per ASTM D75 practice, with samples from fresh faces at several elevations and locations, enough tests per stratum to state a range instead of an average, and a map of unsuitable zones. Only then can the engineer approve the pit outright, approve it conditionally, or approve it with depth restrictions. The average of two results is not a basis for accepting a material that must meet a limit in every load.
2. What can be done with the PI 22 material?
Read the worked answer
It is not automatically waste. The specification permits PI up to 25 below the top 2 ft, so that stratum can be used lower in the fill provided it is identified and controlled. That means stockpiling by stratum and directing hauling by stockpile, not mixing at random in the fill and hoping.
Blending the high-PI and low-PI materials is legitimate if it is designed: set a target ratio from the mass-weighted LL and PI of the two stockpiles, verify the blended Atterberg limits by test rather than arithmetic alone, and confirm the blended Proctor curve. Blending reliability depends on the pit run being uniform, so you test the blended product, not just the components.
What is not acceptable is placing PI 22 material in the top 2 ft, or letting the excavator blend strata by accident and then defending the result with the average.
3. What production testing and hold points would you recommend?
Read the worked answer
Recommend to the engineer of record: separate signed stockpiles by stratum; a full characterization set per stockpile (gradation, LL, PL and PI, moisture-density curve, plus a swell or strength test if specified); and a family of curves so a new maximum is available as the pit deepens.
Production testing at one full set per 2,000 yd³ or per 10 ft of face advance, whichever comes first, plus a technician's visual classification daily and at each stockpile change. Hold points: no hauling until a stockpile is characterized, and no placement in the top 2 ft from any stockpile not approved for that zone. Require stripping and separate stockpiling of topsoil and unsuitable material, and keep the pit floor drained so stockpiles do not gain moisture.
Issue the approval in writing as conditional, with the zone limits stated, so the criteria are enforceable at the fill and not just at the pit.