Salinity is rising in two of our greenhouses. Here is what we are doing about it.
Following our April 7 soil analysis, elevated EC readings in GH 6 and GH 3 have triggered immediate corrective action. This post details the specific steps our farm team is taking to bring salinity levels back within safe thresholds before the next growing cycle.
The mean EC across all 24 plots was 231.67 μS/cm — within acceptable range overall — but the spread from 34 to 866 μS/cm reveals that uniform management is no longer sufficient. Every plot will be treated according to its own measured soil condition going forward.
Salinity is rising in two of our greenhouses. Here is what we are doing about it.
Our 7 April soil analysis flagged elevated EC levels in GH 6 and GH 3. This article details the corrective protocol our farm team is implementing to restore safe soil conditions ahead of the next planting cycle.
This post is a direct follow-up to “Inside our greenhouses: what the soil is telling us” (7 April 2026). We recommend reading that piece first for full context on the soil analysis methodology and dataset.
The problem
What the data revealed — and why it demands action
Soil electrical conductivity is one of the most sensitive indicators of plant stress risk. When EC rises above the 400 μS/cm threshold, the ionic concentration in the soil solution begins to compete with plant roots for water, effectively creating osmotic stress even when the soil appears moist. In extreme cases, this impairs nutrient uptake, stunts growth, and reduces yield.
Our April 7 survey recorded EC values ranging from as low as 34 μS/cm to a peak of 866 μS/cm across 24 valid plot observations. Two plots stood out as high-priority concerns: GH 6_2 at 866 μS/cm, the only very-high reading in the dataset, and GH 3_1 at 740 μS/cm, which also falls firmly in the high-EC category. Both plots recorded correspondingly elevated nitrogen, phosphorus, and potassium levels, strongly suggesting localised over-fertilisation or salt accumulation from irrigation water.
GH 6_2 peak EC
GH 3_1 EC reading
400 μS/cm threshold
The response
Four corrective actions now underway
Rather than a uniform farm-wide response, our approach is calibrated to the specific condition of each affected plot. The following four actions are being implemented in sequence.
GH 6_2 and GH 3_1 are being irrigated with clean, low-salinity water to flush excess dissolved salts downward through the growing medium and away from the root zone. Volume and duration are being calibrated to avoid waterlogging.
Both plots will be retested after leaching is complete. Planting will not resume in those plots until EC readings fall below the 200 μS/cm non-saline threshold. Results will be published as a follow-up update.
Blanket fertiliser application across the farm has been suspended. GH 1, which recorded the lowest nutrient values in the entire dataset, will receive targeted inputs based on its measured deficits — while GH 6 inputs are paused until EC normalises.
Moisture content ranged from 20.4% to 51.5% across plots. This spread points to uneven water distribution. Delivery rates and timing across all six greenhouse units are being reviewed, with focus on GH 5 which recorded the driest plot in the dataset.
The lesson
Why uniform management was no longer enough
One finding from the April analysis has changed how we think about greenhouse management across the entire farm: the near-perfect correlation (r > 0.99) between EC, nitrogen, phosphorus, and potassium. These parameters do not vary independently. They move as one factor across the spatial distribution of plots, which means that wherever nutrient levels are too high, salinity is too high — and wherever they are depleted, the soil is likely undersupplied.
This makes plot-level data not just useful but essential. A single mean EC value across six greenhouses — 231.67 μS/cm in our case — looks perfectly acceptable. But it conceals a range of 34 to 866 μS/cm that tells a completely different story about what is actually happening at the plant-root level. Averages, in soil management, are insufficient guides for action.
Every plot will be managed according to its own measured condition. Uniform inputs on non-uniform soil is a formula for waste — and, at the extremes, for crop loss.
What comes next
Building toward a continuous monitoring system
The April 7 survey was a single-session, cross-sectional snapshot. Its value is significant but limited by the small within-greenhouse sample size of four plots per unit. The ANOVA conducted on between-greenhouse EC differences returned a non-significant result (F = 1.404, p = 0.270), not because the differences are small, but because high within-group variability reduces statistical power when n is low.
Our immediate next step is to expand the sampling protocol to at least six to eight observations per greenhouse per session, and to establish a regular monthly measurement schedule. This will give us a time-series view of how EC, nutrient levels, and moisture evolve across growing cycles, enabling proactive — rather than reactive — management decisions.
The goal is not perfect soil. The goal is informed soil. Every measurement round brings us closer to growing conditions that are not just controlled, but continuously understood.