Reading Your Results

The assessment gives you numbers and it tells you how much to trust them. The second part matters as much as the first.

The confidence badge

Every assessment carries a badge beside the site coordinates: High confidence, Medium confidence, or Low confidence. This is the trust signal, and it is the thing to read first.

On a low-confidence result the page also shows a red banner, Low-confidence estimate - verify before design use, with the percentage of your bore depth the estimate actually covers. Treat that banner as a stop sign, not a footnote.

The property tiles, and the units

The headline properties are shown in IP units with the SI value directly beneath them. Thermal conductivity, for example, reads 1.431 Btu/hr·ft·°F with 2.48 W/m·K underneath.

That matters when you move to a ground-loop design, because the design wizard asks for conductivity in W/m·K. Read the W/m·K number off the tile and type that. Do not convert by hand: the number you need is already on the screen, and the per-layer table and the PDF export carry both units as well.

The conductivity band

The estimate is a range, not a single figure, and the range is shown with the tile.

Each interval carries a band of its own, in the k range (W/m·K) column of the per-layer table below. It is not the tile's range repeated: the tile reports the band for the site, and the column reports what each formation contributes to it. Read the column when you want to know which layers made the site band as wide as it is.

What makes the band wide is the rock, not the number of logs. It starts from the published conductivity spread for each formation in the hole, then widens for three things: logs that disagree with each other, dry (unsaturated) intervals, and depth extrapolated below where log coverage runs out. Well count does not enter it - that feeds the confidence badge instead.

So a well-logged site in a formation with a naturally wide literature spread can carry a wide band, and a site with only a couple of logs through uniform limestone can carry a narrow one. Read the band and the badge together; they answer different questions.

Design against the low-conductivity end of the band, not the middle. Low conductivity means the ground carries heat away more slowly, and that is the case that sizes the loop field.

On-site TRT recommended

This callout appears above the results when the estimate is not solid enough to design against on its own. Its wording is direct: these are literature-based estimates, not a measured thermal-response test, and an on-site TRT would narrow the conductivity range and firm up loop-field sizing.

It fires on any one of several conditions, not only a wide band:

• no usable log coverage at the site
• a wide band relative to the estimate
• a large share of the bore depth extrapolated below where log coverage runs out
• nearby logs that disagree with each other
• signs of moving groundwater, meaning a high median yield in nearby wells through a saturated column

That last one is worth knowing, because it can fire on an otherwise tidy-looking estimate. Flowing groundwater carries heat away in a way a static-conduction model does not capture, and no amount of good log coverage tells you about it.

The estimate basis card

Estimate basis is not a trust rating. It is the provenance breakdown - what the estimate was actually built from, so you can judge it yourself:

• Nearby wells analyzed
• Control wells (lithology) - how many carried real formation detail
• Extrapolated below control - the share of your bore depth the estimate had to infer instead of reading from logs
• Static water level
• Est. residential depth
• Typical drilling method

Extrapolated below control is the row worth checking. At 0% every foot of the modeled bore was voted from real logs; at 40% you are designing against inferred rock for nearly half the depth. Log coverage is a separate figure - the Lithology coverage stat, and the number the low-confidence banner quotes. Some rows appear only when the data supports them.

Where coverage ends is not where the logs stop

This is the one number on the page that is easy to read as something it is not, so it is worth a paragraph on its own.

The estimate walks your bore a foot at a time and, at each depth, counts how many nearby logs still reach that far. While enough of them do, it votes the rock from those logs. The moment too few reach, it stops voting and fills the rest of the hole with one conservative rock assumption. That handover depth is what the percentage is measured from, and it is not the depth of your deepest nearby log.

Those two depths can be far apart, and the deeper one is usually the more comforting of the two. A site can have one log reaching 457 ft while coverage thins out at 279 ft. The estimate stops being evidence at 279, not at 457, and the 221 ft in between is inferred even though a log technically reaches through it. One log is not coverage.

Where a report names both, it says so plainly: the coverage depth first, then the deepest log as the separate fact it is. Read the coverage depth as the point this report runs out of evidence, not as anything your rig would hit.

Water matters here too. Saturated formations move heat considerably better than dry ones, so a site drier than typical for the area reads high and a wetter one reads low.

The per-layer table

Underneath the tiles is the formation-by-formation breakdown. It has eight columns, in this order:

Control agreement reads one percentage for a single interval and a span, such as 52-94%, for a row that merges several intervals the table cannot otherwise tell apart.

A dash in Control agreement has two causes, and the row itself tells you which. Where the row is marked Estimated beside the lithology on screen, or carries a * on its depth in the PDF, the interval sits below the depth nearby logs reach. No log covers it, so there is no agreement to measure. On an unmarked row the dash means the estimate carried no agreement figure for that interval. Either way the dash records the absence of a measurement rather than a measurement of zero. Reports generated before this release printed 0% on the marked rows, so an older report and a new one can disagree on the same interval.

Five of the columns print -- when they have nothing to show: Detail, both k columns, k range, and Control agreement. It means the estimate has no measurement for that interval, not that the value is zero.

For k range a whole column of dashes has two causes, and one test tells them apart. A dashed k range column on an older report means the report was saved before the column existed; running a new assessment fills it in. If a fresh run is still dashed, the site has no coded lithology on file, which is information about the site rather than about the report.

On a site with no coded lithology the tile still shows a conductivity band, and a deliberately wide one, because the estimate widens the band around its fallback value and always advises a TRT. A wide band with low confidence and a TRT callout above a dashed table is that path working as intended rather than a fault.

Detail also carries the one difference no other column shows. Consecutive intervals that match are drawn as a single row, so a lithology appearing twice means something changed. When that something is not visible anywhere else on the row, Detail names it in words, for example k range differs.

The PDF export prints the same table. In the export the per-layer breakdown moves to its own full-width page, with all eight columns in the order the screen uses. The depth profile keeps the formation strip and carries a line pointing to it. Where the rows carry conductivity bounds that line reads "Per-layer thermal properties, with conductivity ranges, follow on the next page." Where they do not, it reads "Per-layer thermal properties follow on the next page." One label differs too. The export abbreviates Control agreement to AGREEMENT so the column fits.

A legend under the table explains the dash, and it prints on both surfaces. It appears wherever some row carries an agreement figure. It opens by defining the column and the merged-row span, then ends with the part that explains a dash. The wording is the same on both surfaces; only the marker changes, because the two mark an extrapolated row differently. The legend ends:

In the export: An interval marked * shows -- because no nearby log reaches it, not because the logs conflict.

On screen: An interval marked "Estimated" shows -- because no nearby log reaches it, not because the logs conflict.

Under the table the export also lists every conductivity range the estimate allows, with the formations sharing each one. That line and the pointer answer the same question, so the export either promises ranges and prints them or does neither. A report whose rows carry no bounds gets the shorter pointer and no ranges line. The dash legend above follows the same shape on its own question, which is why it appears on a report with agreement figures and not on one without.

The rows themselves are the rows you saw. The screen and the export merge consecutive matching intervals through the same function, so the exported report shows exactly the rows on screen. The two new columns match the same way: k range and Control agreement are built on both surfaces by the same code, so a report cannot show one conductivity range or one agreement figure on screen and a different one on paper.

Sanity check the table against what you know locally. If it shows thick saturated sand where you have only ever drilled clay, trust your own experience and treat the estimate as suspect.

It also shows where the drilling changes. A shift from unconsolidated clay to consolidated limestone at a given depth is the point your rig changes method, and the narrative calls that out.

Rows below the coverage depth are marked Estimated, and their lithology reads Assumed rock rather than a formation name. That is the estimate telling you it has no logs down there and is assuming rock, not telling you it found rock. In particular, an Assumed rock row is not a second bedrock contact. Bedrock is wherever the geology estimate above puts it, and that is the only place on the page claiming to know.

Depth sensitivity

This card answers the question you are actually asking: how deep is worth drilling here.

At the top is a green recommendation callout naming the depth it suggests, and a Re-run at N ft button that re-runs the assessment at that depth in one click, so you do not have to retype the location.

Below it is a table comparing candidate depths, each with its effective conductivity, its range, and the ground temperature at that depth. A Best badge marks the one being recommended. Read down the conductivity column: the point where it stops climbing meaningfully is the point where extra footage stops paying for itself. Going deeper is not wrong; it is a decreasing return, and the table lets you price that tradeoff instead of guessing at it.

The card also carries a note when the drilling method changes with depth, which is where a deeper option stops being simply more of the same hole. The narrative explains the same choice in prose if you would rather read it than read the table.

When you have logged a test nearby

If a thermal-conductivity test of yours sits near the site, a green card appears above the results reading On-site TC test nearby, listing the test, its distance, its date, and its conductivity. With more than one in range the heading becomes N on-site TC tests nearby and it lists up to three. A test logged without a date shows No date, which is normal - the date is optional.

Read the second line, because the card has two states:

• Already calibrating this estimate means the test is anchoring the numbers. You will also see an Anchored by on-site TC test chip with the distance.
• Without that line, the test is merely nearby: close enough to tell you about, not close enough to be pulling the estimate toward it.

Anchoring does not automatically clear the TRT-recommended callout. An anchored estimate can still be flagged, and when both appear together the flag is telling you something the anchor does not answer.

Aquifer flow

The aquifer section is separate from the ground-property estimate, and it is the part that matters for open-loop work. Its yield figures are reported in gpm, with 25th and 75th percentile values. Those percentiles describe expected water yield. They are not the conductivity band, and the two should not be read together.

Related Topics

• Running an Assessment - produce the results
• Thermal-Conductivity Tests - anchor an estimate to a real measurement
• Ground-Loop Designs - size the loop field