Trial protocol: measuring a live-microbial input on arid farmland

What this document is

This is the protocol, published in full, that we ask a grower to run before deciding anything. It sets out a paired treatment and control design on a working commercial farm, the baseline panel to measure before the first application, the two application regimes, three measurement checkpoints, and the rules for reading a result that is noisier than most agronomic trials.

The evidence state, stated once. The composition and the microbiology of the product are laboratory work and are published on the evidence page: chemistry S26/3362, culture microbiology M26/9725 and DNA sequencing M26/9726, all from NviroTek Wynland Laboratories on sample AF4_1, with an independent interpretation by Elrica de Necker (SACNASP 139526). No yield trial has completed. Nothing in this document reports a Magic Power field result, because there is not one to report.

What follows is therefore a method, not a finding. Section 8 collects what the published literature reports for biological inputs generally, with citations, so that a grower designing a trial knows what magnitude of effect is worth powering for.

01The microbiome problem in arid-land agriculture

Arid and semi-arid soils are not depauperate by accident. They are the slow output of climates where mean annual precipitation rarely exceeds 400 mm, evapotranspiration regularly exceeds 1,500 mm, and net primary productivity, the carbon input that feeds belowground life, is structurally low. Under those conditions, soil organic carbon (SOC) stocks in the upper 30 cm typically range from 0.3 to 1.2 percent by mass in cultivated land, against 2 to 4 percent in temperate systems with comparable management (Plaza et al. 2018; Lal 2004). The microbial biomass carbon (MBC) that those SOC stocks support is correspondingly small, often below 200 mg C per kg of dry soil, against 500 to 900 mg C/kg in productive temperate cropland (Bardgett and van der Putten 2014).

When these soils are drawn into intensive irrigated agriculture, three reinforcing pressures accelerate microbial contraction.

The first is carbon starvation. Synthetic mineral fertilizer delivers nitrogen, phosphorus and potassium to the plant but adds no organic substrate to the soil. With residue removal common through grazing, burning or mechanical clearing, the carbon flux that would otherwise feed heterotrophic bacteria, fungi and protists is interrupted. Microbial communities respond by contracting, simplifying, and shifting toward stress-tolerant, less metabolically diverse assemblages (Fierer 2017). Functional redundancy, the property that protects soil ecosystems against perturbation, falls.

The second is salinization. Irrigation water in arid regions typically carries 0.3 to 2.0 dS/m of dissolved salts. In a soil with too little rainfall to flush them, those salts accumulate in the root zone over years to decades. Above approximately 2 dS/m saturation extract conductivity (ECe), microbial respiration is measurably suppressed; above 4 dS/m the suppression becomes severe and the nitrifying community in particular contracts sharply (Rath and Rousk 2015). Many plant-growth-promoting rhizobacteria and arbuscular mycorrhizal fungi tolerate moderate salinity poorly. Salinity suppresses yield directly through osmotic stress on the plant, and it also suppresses the biology that would otherwise help the plant tolerate that stress.

The third is physical degradation. Without a steady supply of microbially produced binding agents (glomalin from arbuscular mycorrhizae, microbial extracellular polymeric substances, fungal hyphae), soil aggregates weaken (Rillig 2004). Compaction follows. Infiltration rates decline. A surface seal forms after each irrigation. Water that should have penetrated to the root zone runs off, evaporates, or perches in a thin band that supports neither deep rooting nor microbial activity.

The compounding outcome is what growers describe as tired soil: more input for the same yield, more water for the same biomass, and a season eventually arriving in which the yield drops despite full-rate fertilization. A higher synthetic dose does not answer that. Whether a live-microbial input answers it on your soil is what the trial below is for. (For a plain-language treatment of the same mechanisms, see What’s actually alive in your soil.)

02What the product is, and is not

Be precise about the material being trialled, because the trial design depends on it.

Magic Power is the liquid biological residue of a closed-loop recirculating aquaculture system. On sample AF4_1 the laboratory measured total nitrogen at 421.5 mg/L, potassium at 216.8, phosphorus at 16.5, calcium at 105.6, magnesium at 34.6 and sulphur at 29.3 mg/L, at pH 7.96 and electrical conductivity 374 mS/m (NviroTek Wynland Laboratories, report S26/3362, page 1 of 2, reported 10 July 2026). A litre therefore carries about 0.42 g of nitrogen. A kilogram of urea carries about 460 g. This is not a nutrient product and the trial should not be designed as though it were.

What the same sample carries is biology. The culture panel reports a total plate count above 30,000,000 cfu/g, which is the method’s upper reporting ceiling rather than a measurement, with yeast at 10 cfu/g and mould at 3,200 cfu/g (report M26/9725, all three results inside the laboratory’s schedule of accreditation). Sequencing resolved 952 bacterial taxa, 516 of them to species, at a Shannon index of 4.49, dominated by Pseudomonas at 19.7 percent of reads, Aeromonas at 9.8, Comamonas at 9.1, Flavobacterium at 7.9 and Acinetobacter at 7.8 (report M26/9726, tabled in the interpretation report at page 9).

The interpreting scientist’s own conclusion is the honest frame for a trial: the sample is "a living, dilute biological ferment with moderate nitrogen and potassium concentrations", the bacterial profile is "rather a broad environmental/fermentation mix than a purpose-built beneficial-microbe inoculant", and on its own the product "will therefore not replace a complete fertility program" (Eco Veritas interpretation report AF4_1, pages 12 and 13).

So the hypothesis under test is nutrient-use efficiency and soil function, not substitution. A trial designed to show that the product replaces mineral nitrogen will fail, and it will fail for arithmetic reasons rather than biological ones.

03Pre-trial baseline measurement

A trial that does not start from a measured baseline cannot end with measured conclusions. Invest in baseline before you change a single input.

Field design

Designate, at minimum, a paired treatment and control block within a single field of uniform soil texture, slope and irrigation infrastructure. For statistical power, prefer four blocks of each with randomized assignment if logistics allow. Block size should be agronomically meaningful: 0.5 hectare per block for open-field row crops, 0.1 hectare for high-value greenhouse, one full row pair for perennial systems. Document GPS coordinates, area, irrigation specification and rotation history for each block.

Our own trials programme asks for 5 hectares minimum with a paired control, a baseline soil test at the grower’s cost, and consent to publish the result either way. That gate is on the trials page.

Sampling protocol

Sample soils two to four weeks before the trial begins, after the previous crop is removed and before any pre-plant tillage or amendment. Take 0 to 15 cm and 15 to 30 cm depths as separate composites. Within each block collect at least 15 sub-samples on a regular grid or W-pattern, mixed to one composite per depth per block. Use a clean stainless or chrome-plated auger; do not use galvanized tools where micronutrient analysis is required. Bag in clean polyethylene, label with block ID, depth and date, and ship to a certified soil laboratory within 48 hours. For microbial assays refrigerate at 4 degrees C and ship cold. Do not freeze.

The baseline panel

• Soil organic carbon by Walkley-Black or, preferably, dry combustion. Report as percent C by mass.
• Microbial biomass carbon by chloroform-fumigation extraction (Vance et al. 1987). Report in mg C per kg dry soil.
• Basal soil respiration by 24-hour CO2 efflux at 25 degrees C and 60 percent water-holding capacity. Report in mg CO2-C per kg per day.
• Aggregate stability by wet-sieving for mean weight diameter of water-stable aggregates, or the slake test as a field-friendly proxy.
• Infiltration rate by single-ring or double-ring infiltrometer, 15 cm minimum diameter, three points per block, in mm per hour at quasi-steady state.
• Volumetric water-holding capacity at -33 kPa and -1500 kPa by pressure plate, on undisturbed cores where possible. The difference is plant-available water.
• Electrical conductivity of the saturated paste extract, in dS/m.
• pH of the saturated paste, calibrated meter.
• Available macro and micro nutrients on the analyte panel standard for your region.

A minimum-viable kit: a stainless auger, a 15 cm single-ring infiltrometer, a calibrated pH and EC meter, a soil thermometer, sample bags, a marker and a cooler.

Yield and input baseline

Separately, document the previous three to five years of yield per hectare, fertilizer input by type and rate, irrigation volume, and where available kilograms of marketable yield per cubic metre of irrigation. Soil tests without yield context generate confusion rather than insight.

04Application regimes

Two regimes exist and they are priced and measured differently. Say which one you are running before you start, because a trial that mixes them cannot be read.

Regime A, the season programme

This is the regime the product is sold against and the one our own trials use.

• 15 to 25 litres per hectare per season, banded in the starter fertilizer line at planting (FishIt integration guide, section 1, In-Furrow Plant-Food Stabilization, co-blending recipe per hectare). The guide is a customer integration memo written by JB van den Berg for Van Den Berg Farming. It is not a trial result.
• Diluted at least 1:10 with water, placed about 5 cm to the side of or below the seed, never in direct seed contact. The reasons are the ammonium at 244 mg/L and the salt load at EC 374 mS/m, either of which will scorch germinating seed and young roots when the concentrate lies against it (Eco Veritas interpretation report, page 11, section 5.1).
• A season cap of 30 litres per hectare where fertigation events follow the in-furrow band. That cap is a programme design figure, not a measured threshold.

Regime B, continuous fertigation

Where the grower runs a continuous fertigation line rather than a starter band, the product is injected at a concentration of up to 1 part product per 1,000 parts irrigation water. Read that as an injection concentration and nothing else. The annual volume per hectare is set by how much water the crop takes, not by the ratio, and it should still be held inside the season cap above. The figure "1 litre per 1,000 litres" is not interchangeable with the season programme and does not convert into a per-hectare dose.

Foliar

We publish no foliar rate. The safety data sheet lists foliar application under uses advised against, and phytotoxicity, pathogen and residue testing is commissioned but not complete. Foliar sits on the trials page as a protocol under test, with the conditions the interpretation report sets out.

Handling and timing rules that travel with every regime

• Wear gloves and eye protection when handling concentrate. The community includes Aeromonas and Citrobacter, which the interpretation report classifies as a hygiene concern rather than a plant-pathogen risk (page 12).
• On edible grain and feed, apply early in the season or to the soil, not close to harvest. The reason is the opportunistic bacteria together with aflatoxin-forming Aspergillus DNA detected at 9 reads of about 13,900 fungal reads (page 11, section 5.2, and page 10, section 4.3). DNA presence is not viability and not toxin.
• Molybdenum was reported at 1,690 ug/L with copper below detection. On pasture that imbalance is worth monitoring for copper deficiency in ruminants (page 12).
• The interpretation report recommends adding copper to correct the imbalance. Do not add molybdenum to the tank; one integration memo suggests it and it contradicts the laboratory interpretation directly.

Irrigation system requirements

Three constraints hold whichever regime you run.

First, the dosing pump must deliver the concentrate at a stable ratio: a Venturi injector, Dosatron, or proportional injection pump rated for biological products, with stainless or PVC wetted parts.

Second, chlorinated source water is incompatible with live-microbial inoculation. Free chlorine above 0.5 ppm at the emitter will reduce viable counts in the dosed water. Where municipal chlorinated water is the only source, install granular activated carbon upstream of the injection point, or hold the water in an open reservoir for 24 hours. Chloraminated water needs ascorbic acid dosing or a dedicated chloramine filter; neither holding nor activated carbon alone is enough.

Third, inject during the middle two-thirds of the irrigation event so the dosed water reaches the wetted zone and is followed by enough clean water to clear the lines. End every event with a five to ten minute clean-water flush.

Storage

Store in shade, ventilated, not frozen, not in direct sun, closed between uses. Shelf life is not yet measured: the safety data sheet claims 24 months with no stability data behind it, and a viable-count series at 3, 6 and 12 months is commissioned. Until that series reports, treat product age as an uncontrolled variable and record the production date of every drum used in the trial.

05Measurement checkpoints

Three checkpoints at three, six and twelve months from the first application, plus an end-of-season agronomic summary.

Three months

The response, if there is one, is biological and not yet visible at the plant. Re-sample at the same depths and locations as the baseline and run:

• Microbial biomass carbon. Usually the first metric to move in the literature on biological inputs.
• Basal respiration. Read it beside MBC as the metabolic quotient qCO2.
• A slake test in the field. Cheap, repeatable, and legible to the people doing the work.

Do not expect bulk SOC to have moved at three months; it is a slow-response variable and a three-month change is rarely above the detection threshold of standard methods. Do not expect infiltration or yield to have moved.

Six months

Re-run the three-month panel and add:

• Aggregate stability by wet-sieving, as mean weight diameter.
• Infiltration rate at three points per block. Variance here is high; three points is the minimum that means anything.
• Soil EC. Watch it rather than expect it to fall.

For annual crops the six-month checkpoint usually falls within or after one harvest. Record yield, marketable fraction, and qualitative observations on disease pressure, root mass and crop uniformity.

Twelve months

Run the full baseline panel a second time, against the matched control rather than against the baseline alone. The control is what protects you from reading a good season as a treatment effect.

Reporting

At each checkpoint, file a written report comparing each treatment block against its matched control, in absolute and percentage change. Include photographs of the slake test, the cores and the crop. Track water input volume per block. Anecdote without measurement is noise, and measurement without context is too.

06Reading the result

A field trial is not a meta-analysis. Variance at the individual-farm level is substantial.

What a working trial looks like

The strongest signal is internal consistency across metrics. MBC, basal respiration, aggregate stability and infiltration tend to move together. SOC and water-holding capacity move more slowly and should trend in the same direction. Yield is the noisiest metric in the first season and should not be the sole basis of any conclusion. If five biological and physical metrics rise modestly and yield is statistically indistinguishable from control, the trial is working. Soil response precedes plant response, often by a season.

When to adjust the protocol

If at six months no biological response is measurable, examine four causes in order: source-water chlorine or chloramine residual at the emitter; dilution error at the dosing pump; salinity above 4 dS/m masking the response; carbon depletion severe enough to need a supplemental organic amendment alongside the input. Change one thing and re-measure at the next checkpoint.

If at twelve months the biology has moved and yield has not, the limiting factor is most likely outside the soil-microbiome envelope: water deficit, mineral macronutrient deficiency, salinity, variety choice or pest pressure. Escalate the agronomic diagnosis. This protocol addresses one part of a system and does not replace the rest of agronomy.

On replicate variance

Read central tendency, not extremes. A treatment block that responds well while its pair responds barely at all is not a failed trial; it is a working trial with an unidentified confounder in the second block. Investigate spatial heterogeneity, irrigation uniformity and historical management before discounting anything. Soil is heterogeneous at the metre scale, which is the reason replication exists.

07Common pitfalls

Five recurring failure modes account for most disappointing trials. Each is avoidable.

Rescuing a treatment block in season one

The most common pitfall is the urge to add synthetic nitrogen back into a treatment block when yields run behind expectation. A panic dose of urea or ammonium sulfate loads salt onto rebuilding biology and confounds the trial at the same time. A block that received both is no longer a treatment block. Hold the protocol.

Wrong dilution

Two failures recur. The first is a fixed-rate metering pump on a variable irrigation flow, which produces undosed and overdosed segments inside a single event; use proportional injection. The second is mixing the concentrate into a stock tank of high-EC soluble synthetic, where the salinity strips out a large part of the microbial population before it reaches the field. Inject separately, downstream of any soluble-fertilizer injection point.

Anaerobic storage

This is an aerobic biological liquid. Sealed anaerobic storage at high temperature shifts the community toward fermentative anaerobes within weeks, producing off-odours, gas pressure in the container and a reduced effect on application. Store in shade, vent the container between uses, record production dates.

Freeze damage

Freezing lyses a substantial fraction of the viable population. On sites with sub-zero winter exposure, store concentrate in a temperature-controlled facility through winter. Outdoor unheated storage is fine in subtropical and tropical sites and not in continental temperate ones.

Chlorinated source water

Free-chlorine residuals above the 0.5 ppm tolerance are a common and invisible cause of a flat result. Test source water for free and combined chlorine before the trial starts. Do not assume well water is chlorine-free; some farms top up well storage with chlorinated municipal water during dry months without recording it.

08What the published literature reports for biological inputs generally

None of the figures in this section describe Magic Power. They describe the wider class of live-microbial and biofertilizer inputs in the peer-reviewed literature, and they are here so that a grower can power a trial sensibly and know what size of effect is worth designing for. Each carries its citation. Our own literature review treats them at length.

Yield. Schütz et al. (2018) synthesized 171 studies across cereals, legumes and vegetables and reported a mean grain-yield increase of about 16 percent in cereals, with a wide confidence interval and soil organic carbon and baseline phosphorus availability as the strongest moderators: inoculants performed better on carbon-poor and phosphorus-limited soils. Rubin, van Groenigen and Hungate (2017) examined plant growth and yield responses across a large body of study cases and found effect sizes shrinking as fertility rose. Both findings point the same way. The more degraded the starting point, the larger the expected response, and a saturated high-input system is the wrong place to look for one.

Variance and failure. A non-trivial fraction of published trials report no significant yield effect. Bashan et al. (2014) catalogue persistent commercialization failures traced to low cell viability at point of use, carrier incompatibility and co-formulated chemistries. The literature supports a benefit on average. It does not support a guarantee in any one season, soil and crop combination.

Water. Augé (2001) established across a large review that arbuscular mycorrhizal symbiosis improves leaf water status and transpiration efficiency under drought. Rillig (2004) and Six et al. (2004) set out the aggregate-stabilization mechanism that carries the soil-side effect. Reported field improvements across that literature fall in the range of 10 to 40 percent for infiltration and 5 to 25 percent for water-holding capacity, over multi-season horizons and with wide variance.

Substitution. Adesemoye et al. (2009) showed combined rhizobacterial and mycorrhizal inoculation in tomato allowing a 25 percent reduction in synthetic fertilizer with no yield penalty. That is a fertilizer-efficiency result, not a replacement result, and it is the shape of outcome this protocol is designed to detect.

Carbon. Treat soil-carbon claims conservatively. Lehmann and Kleber (2015) reframed soil organic matter as a continuum of decomposing fragments rather than a stable pool, and Schlesinger and Amundson (2019) and Powlson et al. (2014) show how shallow sampling, short horizons and absent bulk-density correction inflate reported sequestration. Measurable topsoil carbon gain over 5 to 10 years is supportable; a verifiable multi-tonne annual offset is not.

09References

Adesemoye, A. O., Torbert, H. A., and Kloepper, J. W. (2009). Plant growth-promoting rhizobacteria allow reduced application rates of chemical fertilizers. Microbial Ecology, 58(4), 921-929.

Augé, R. M. (2001). Water relations, drought and vesicular-arbuscular mycorrhizal symbiosis. Mycorrhiza, 11(1), 3-42.

Bardgett, R. D., and van der Putten, W. H. (2014). Belowground biodiversity and ecosystem functioning. Nature, 515, 505-511. https://doi.org/10.1038/nature13855

Bashan, Y., de-Bashan, L. E., Prabhu, S. R., and Hernandez, J.-P. (2014). Advances in plant growth-promoting bacterial inoculant technology: formulations and practical perspectives (1998-2013). Plant and Soil, 378(1-2), 1-33.

Fierer, N. (2017). Embracing the unknown: disentangling the complexities of the soil microbiome. Nature Reviews Microbiology, 15(10), 579-590. https://doi.org/10.1038/nrmicro.2017.87

Lal, R. (2004). Soil carbon sequestration impacts on global climate change and food security. Science, 304(5677), 1623-1627. https://doi.org/10.1126/science.1097396

Lehmann, J., and Kleber, M. (2015). The contentious nature of soil organic matter. Nature, 528, 60-68. https://doi.org/10.1038/nature16069

Plaza, C., Zaccone, C., Sawicka, K., Méndez, A. M., Tarquis, A., Gascó, G., Heuvelink, G. B. M., Schuur, E. A. G., and Maestre, F. T. (2018). Soil resources and element stocks in drylands to face global issues. Scientific Reports, 8, 13788. https://doi.org/10.1038/s41598-018-32229-0

Powlson, D. S., Stirling, C. M., Jat, M. L., Gerard, B. G., Palm, C. A., Sanchez, P. A., and Cassman, K. G. (2014). Limited potential of no-till agriculture for climate change mitigation. Nature Climate Change, 4(8), 678-683.

Rath, K. M., and Rousk, J. (2015). Salt effects on the soil microbial decomposer community and their role in organic carbon cycling: A review. Soil Biology and Biochemistry, 81, 108-123. https://doi.org/10.1016/j.soilbio.2014.11.001

Rillig, M. C. (2004). Arbuscular mycorrhizae, glomalin, and soil aggregation. Canadian Journal of Soil Science, 84(4), 355-363. https://doi.org/10.4141/S04-003

Rubin, R. L., van Groenigen, K. J., and Hungate, B. A. (2017). Plant growth promoting rhizobacteria are more effective under drought: A meta-analysis. Plant and Soil, 416, 309-323. https://doi.org/10.1007/s11104-017-3199-8

Schlesinger, W. H., and Amundson, R. (2019). Managing for soil carbon sequestration: let’s get realistic. Global Change Biology, 25(2), 386-389.

Schütz, L., Gattinger, A., Meier, M., Müller, A., Boller, T., Mäder, P., and Mathimaran, N. (2018). Improving crop yield and nutrient use efficiency via biofertilization: A global meta-analysis. Frontiers in Plant Science, 8, 2204. https://doi.org/10.3389/fpls.2017.02204

Six, J., Bossuyt, H., Degryze, S., and Denef, K. (2004). A history of research on the link between (micro)aggregates, soil biota, and soil organic matter dynamics. Soil and Tillage Research, 79(1), 7-31. https://doi.org/10.1016/j.still.2004.03.008

Vance, E. D., Brookes, P. C., and Jenkinson, D. S. (1987). An extraction method for measuring soil microbial biomass C. Soil Biology and Biochemistry, 19(6), 703-707. https://doi.org/10.1016/0038-0717(87)90052-6

---

No Magic Power field trial has yet completed. The first results will appear on the trials page.