India Has Tested Its Soil Twice. The Second Time, a Machine Chose the Spot.
The Soil Health Card scheme has been measuring organic carbon on every card it issues since 2015 — more than 25 crore cards by 2026, the only national soil panel India has. In April 2023 the way samples are collected changed: the mobile app now fixes where an operator may sample, records the coordinates itself, and QR-links each sample to a geo-mapped lab "without any manual intervention." Before that, a person chose the spot and typed the location in. Any trend in soil carbon that crosses April 2023 is therefore comparing two collection regimes before it compares two soils — and the headline studies now circulating about declining organic carbon straddle exactly that seam without naming it. This is a methods note, not a finding about soil.
What changed in April 2023, in the ministry's own words
The Ministry of Agriculture's "Celebrating a Decade of Soil Health Cards" release of 18 February 2025 describes the upgraded system as in use since April 2023. Its own description of what the upgrade does is the whole basis of this note:
| Step | Before April 2023 (as implied by the release) | Since April 2023 (as the release describes it) |
|---|---|---|
| Where the sample is taken | Chosen by the collecting operator | The app restricts the operator's sample-collection region |
| Where the sample is recorded as being from | Location keyed in by hand | Latitude and longitude auto-captured by the app, "removing the need for manual entry" |
| Link from sample to result | Manual matching of sample IDs to lab results | A unique QR code per sample, linked to a geo-mapped lab; results flow to the portal "without any manual intervention" |
| How results are viewed | Tabular, by card | Portal integrated with a GIS layer so all test results are captured and seen on a map, down to cadastral boundaries |
The revamped portal and mobile application were formally inaugurated by the Union Agriculture and Rural Development ministers in early 2024 (PIB PRID 2012238), but the release marking the scheme's decade is explicit that the upgraded collection system had been operating since April 2023.
Why that is a seam, not just an upgrade
Each of those changes is an improvement in data quality. Together they also change what a "sample from district X" means. Before the seam, sample location carried two kinds of error the new system removes: the operator's choice of a convenient spot (a field edge, a road-adjacent plot, a farmer who asked), and transcription error in the typed location. After the seam, location is constrained and machine-recorded. If pre-2023 samples were systematically drawn from more accessible, more intensively farmed or otherwise unrepresentative spots, then a district's pre-seam organic-carbon average is not measuring the same population of soils as its post-seam average — and the difference between them is partly a difference in where the machine looked, not in what the soil holds.
This is the standard problem of any long-running survey whose sampling frame changes mid-series. It is why household surveys flag frame changes and publish bridging estimates. The Soil Health Card programme, as far as this piece could find, has published no such bridge, and the ministry's release presents the change purely as a modernisation.
The headlines that cross it
A six-year ICAR study coordinated by the Indian Institute of Soil Science, Bhopal, covering 2017 to 2023, has been widely reported this year as finding serious depletion of soil organic carbon in India's arable land, driven by imbalanced fertiliser use and climate change, with low organic carbon tied to micronutrient deficiency. This piece does not dispute the finding and does not reproduce the study's sample and district counts, which it could only find on exam-coaching sites and not in a primary ICAR document. The point is narrower: a 2017–2023 window ends inside the first year of the new collection regime. Whether the study drew on Soil Health Card samples, its own sampling, or both, any comparison of early-window and late-window carbon that used card data would be comparing across the seam. The study may well have handled this. The reporting of it does not mention it.
What this note could not do
The obvious next step is to pull district-level organic-carbon distributions from the portal's public Nutrient Dashboard for the years either side of April 2023 and look for a discontinuity — a step change in the mean or variance at the seam that no agronomic process would produce in a single season. This blog's source register verified in September 2026 that the dashboard's data endpoint is live. It could not be queried from the environment this piece was written in, so the test has not been run. Until it is, this remains a methods argument from the ministry's own description of its system, and should be read as exactly that.
What would fix this
A frame change in a survey is normal. What is missing is the bridge every other national survey publishes when it happens.
- Flag the collection regime on every record. A single field in the portal — pre-April-2023 or post — would let any analyst, including ICAR's own, keep the two populations apart without guessing from dates.
- Re-sample a fixed panel across the seam. Pick a few thousand pre-2023 sample points, revisit them under the new regime, and publish the paired results. That single exercise would show whether the old hand-chosen sites biased the carbon readings, and by how much — and would turn this note's argument into a measured correction factor, or dissolve it.
- Require studies that use card data to state their sampling frame. An ICAR study spanning 2017–2023 should say, in its abstract, whether it used card samples, whether any crossed the seam, and how it handled that. Reporting that omits this is not the study's fault; a publication norm would fix it.
- Open the district-level carbon tables as a dated download, not only a dashboard. The endpoint exists. A yearly CSV with the regime flag would let the test this piece could not run be run by anyone.
These are proposals from this blog, not recommendations any agency has adopted.
What this piece does and doesn't cover
- It makes no claim about the level or direction of India's soil organic carbon. The blocking check in this piece's own plan — obtain a district-level carbon table before saying anything about levels or trends — could not be passed, so the piece is the shorter methods note the plan specified for that case.
- The ICAR study's headline sample size (a figure in the hundreds of thousands of samples across several hundred districts) is deliberately omitted because it could not be traced past coaching-site summaries to IISS Bhopal or an ICAR publication.
- The pre-April-2023 process is inferred from what the release says the upgrade replaced ("removing the need for manual entry," "without any manual intervention"); the ministry has not published a side-by-side description of the old workflow.
- Cards-issued and samples-tested are different counts. One sample can generate cards for several farmers on a grid, and a card is reissued every three years under the current cycle; the 25.89 crore figure is cards, and this piece does not convert it to samples.
- Every figure came through this account's web-search tool. pib.gov.in and soilhealth.dac.gov.in were unreachable from this environment; PIB PRID 2104403 was captured in full by this blog's source register on 4 September 2026 and its description is quoted from that capture, while the cards-issued figure comes via secondary reporting of ministry statements.
Sources: PIB Release ID 2104403, Ministry of Agriculture & Farmers Welfare, "Celebrating a Decade of Soil Health Cards," 18 February 2025 (the April 2023 upgraded system; region-restricted sampling; auto-captured latitude/longitude; QR-coded samples; geo-mapped labs; GIS-integrated portal to cadastral level; 12 parameters including organic carbon) — captured directly by this blog's source register, 4 September 2026. PIB Release ID 2012238 on the inauguration of the revamped portal and mobile application. Soil Health Card portal, soilhealth.dac.gov.in, and its Nutrient Dashboard (endpoint verified live 4 September 2026; not queried). Cards issued: ministry figures as reported by All India Radio (over 25 crore, January 2026) and Vision IAS citing PIB (25.89 crore, July 2026). ICAR/IISS Bhopal organic-carbon study, 2017–2023, as reported in current-affairs coverage in 2026; primary document not located. Rattan Lal, "Soil Carbon in Agro-Ecosystems of India," Journal of the Indian Society of Soil Science 74(1), 2026, for the policy framing of soil carbon (open access, epubs.icar.org.in; not fetched from this environment).
v1.0.0 — 19 September 2026 — first published.
About this article: Researched, written and edited by Umashankar Triplicane Dwarakanathan, with AI research assistance; every figure is meant to trace to the primary source cited. See the Editorial Policy for how sourcing, AI use and corrections work.
