Proof-first only counts if the method holds — a measurement is only as trustworthy as the way it was taken. Here is exactly how each reading in the baseline is made, and where each method's limits lie.
Every measurement on this site should trace back to something real: a reading we took, with the site, date, and method noted, or a credible published source. If we can't back it up, we leave it out — even when it would read well. This page is the method half of that promise.
We measure four things, each with a protocol we can repeat season after season at the same marked spots. Repeatability is the point: a comparison over time is only honest if the second reading was taken the same way as the first. So we write the method down, and we say where it falls short.
None of these methods is new, and that is deliberate. They are established field techniques, chosen because one person can run them carefully without a lab. We would rather run a well-understood method well than reach for one we can't yet do justice to.
Each measurement area, how we take the reading, and where the method falls short — stated plainly.
We estimate organic matter by loss-on-ignition: a dried soil sample is weighed, burned off, and re-weighed, and the mass lost stands in for organic content. The limit: this is an estimate, not a lab-grade carbon assay, and clay-bound water can bias it — so later years add a lab panel to anchor the field readings against a known standard (the Haney soil-health test is one well-recognized option).
We time infiltration with a single ring driven at three replicate points per site: how long a set volume of water takes to soak in, reported as the median of the three. The limit: a single ring reads faster than a double-ring rig, so we treat the number as a relative trend at a site over time, not an absolute rate we could hand off as hydraulic conductivity.
We count plant diversity in a fixed 0.25 m² quadrat frame (50 by 50 cm) set down at marked points along a transect: how many species share each frame. The limit: a frame that small samples a small area and misses rare or patchy plants, so the counts run conservative and are read alongside site notes rather than as a full inventory.
We read the biodiversity a working farmer can see without a lab. Dung beetles first — sorted by the three functional groups (dwellers, tunnelers, rollers) working a fresh pat, since it is they who bury dung, cycle its nutrients into the soil, and break the parasite and fly cycle, which makes them a sensitive read on pasture health. It is also why we hold a standing goal of keeping avermectin dewormers (ivermectin and its kin) off the pasture: the residue passes through a treated animal into its dung and kills the beetles that work it. An animal that genuinely needs treatment is drawn off, treated apart, and off-ramped from the herd rather than returned, so the chemistry never reaches the ground we are counting beetles on. Then pollinators counted on a timed walk, and birds by presence and call. The limit: indicator counts are a coarse stand-in for whole-system biodiversity and shift with weather and time of day — good for direction, not for precision.
The baseline is a program that grows, not a one-time survey. Year 0 records what one operator can measure rigorously without a lab; later years layer in lab panels and advisor-run tests as the operation can support them. That staging is part of the honesty: we would rather publish a modest reading we actually took than a sophisticated one we couldn't yet stand behind.