Harvestly

Where every number comes from.

Thirty-nine numbered references, each labelled by how far we checked it. Links open the publisher or the copy we read. Checked 9 to 12 October 2026; a link can still change.

How the engine uses them

01 How we label

Three labels.

  • Verified Read at the publisher or its official record.
  • Secondary Read through a mirror, a news report or a search summary.
  • Cited Cited in concept note v3 (sources checked 9 October 2026). Not re-opened in the 12 October check.

02 Wording we chose

Four places where we chose our words with care.

  • “USDA cut its forecast by half.” The 51% is a cut against USDA’s earlier forecast of 4.5 Mt, not a fall against last year.
  • Insects. The Kenyan storage-loss study counts losses to two insects. It does not rank insects against moisture or mould, and its authors say its total looks high. We do not quote the total.
  • Hermetic bags. About one in six Kenyan maize farmers uses them: a focus-group estimate of farmers, not a household count 2.
  • White maize. The tables are for US corn. A study of US white and waxy corn found each type needs its own curve 36. That is not Kenyan white maize either.

03 The references

References.

  1. 1
  2. 2

    De Groote, Muteti and Bruce (2023). On-farm storage loss estimates of maize in Kenya using community survey methods. CIMMYT repository 10883/22998.

    Secondary The CIMMYT site blocked us; read through a Mexican national repository mirror. Focus groups in 121 communities, 1,439 farmers: hermetic bags 16%. It counts losses to two insects only and does not rank insects against moisture.

  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8

    USDA Foreign Agricultural Service, GAIN report KE2026-0017, “Dry Weather Decimates Kenyan Harvest”, 28 August 2026 (copy hosted by APEDA).

    Secondary Read in full via the APEDA copy; the usda.gov original was unreachable (“Access Denied”). The 51% is a cut against USDA’s earlier 4.5 Mt forecast, not a fall against last year, which is why this site says “cut its forecast by half”.

  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14

    Precision Development. MoA-INFO.

    Cited Published user counts and crop coverage differ between sources, so this site quotes no number.

  15. 15
  16. 16
  17. 17
  18. 18

    Johnson (1992). Determining safe storage moisture content of maize: salt granule technique. CSIR Food Research Institute, Ghana.

    Cited Secondary, unverified. The PDF could not be opened from here (SSL error, retried), so the 15% salt-test line is an unverified rule of thumb, and no amounts or times from the paper are used.

  19. 19
  20. 20
  21. 21
  22. 22

    Ng’ang’a et al. (2016b). Efficacy of hermetic and woven bags in on-farm maize storage, Makueni County, Kenya. Food Security 8: 621–633.

    Secondary The old Purdue PDF link returned 404; this DOI link replaces it. Weight loss in woven bags was 41.2% over the trial; hermetic bags lost under 1%. The week-7 timing is not in the abstract and is our reading.

  23. 23
  24. 24

    Oracle. Always Free Autonomous AI Database (limits).

    Verified Also: stopped for 90 days in total may be reclaimed; about 5 minutes to wake after a restart.

  25. 25
  26. 26
  27. 27

    Africa’s Talking Help Centre. USSD sessions and menu limits.

    Cited A USSD session lasts 180 seconds, but Safaricom ends it after 30 seconds without input (Airtel 60), so any menu flow must be quick. Price and shortcode terms are not on this site because they are not settled.

  28. 28
  29. 29

    Open-Meteo. Terms of Use (free API for non-commercial use; CC BY 4.0 credit).

    Cited The free tier is for non-commercial use. A pilot or production service needs the paid plan or another weather source, and Open-Meteo must be credited with a link.

  30. 30
  31. 31

    Ndegwa et al. Effectiveness and Economics of Hermetic Bags for Maize Storage (randomised trial). AgEcon Search 212524.

    Secondary Abstract only. The county (reported as Nakuru in secondary summaries) and the loss figures (12.4% and 3.9% over 4 months) were not confirmed in the full text.

  32. 32
  33. 33
  34. 34

    People Daily (27 September 2025). Rains, high drying costs put 7.5m bags of Trans Nzoia maize at risk.

    Cited A 2025 report whose body we could not read. It is not the current outlook, and this site does not use its figure.

  35. 35
  36. 36

    Bartosik and Maier (2005). ASAE Paper 056081: moisture curves of white and waxy corn differ from the ASAE corn parameters.

    Secondary Abstract-level reading. US white and waxy corn, 12–20% moisture, 0–25 °C; the numeric parameters were not seen.

  37. 37

    Office of the Data Protection Commissioner (Kenya), FAQs on registration, and the Data Protection (Registration of Data Controllers and Data Processors) Regulations, 2021 (the Legal Notice number is unconfirmed).

    Verified Verified: the ODPC FAQ, read at the publisher. It says a small controller (turnover under KES 5 million and fewer than 10 staff) need not register unless it processes data for a listed purpose. Agriculture is not on that list. Not confirmed: the Regulations’ Legal Notice number (265 of 2021). The Notice itself returned an error from here on 12 October. The Act applies either way.

  38. 38

    NASA Langley Research Center, Prediction Of Worldwide Energy Resources (POWER): Daily API, data sources and referencing. Used for recent weather at a rounded ward centre (T2M, RH2M, PRECTOTCORR).

    Verified Read and called on 12 October 2026. Acknowledgement: “The data was obtained from National Aeronautics and Space Administration (NASA) Langley Research Center's Prediction Of Worldwide Energy Resources (POWER) project funded through the NASA Earth Science Division.” A value covers a grid cell of about 55 km by 70 km and the newest day is about 3 days old. We found no validation against Kenyan stations, so it is shown only as an estimate. Data version reference: “The data was obtained from the POWER Project's Daily 2.10.0 version on 2026/10/12.” Not endorsed by NASA.

  39. 39

    NASA Earthdata, Global Imagery Browse Services (GIBS) API, with the Worldview snapshot tool (MODIS Terra true colour). One dated image of Trans Nzoia on the Pilot page.

    Verified Acknowledgement: “We acknowledge the use of imagery provided by services from NASA's Global Imagery Browse Services (GIBS), part of NASA's Earth Science Data and Information System (ESDIS).” The snapshot tool worked on 12 October 2026 but is not on the official GIBS API page, so the image was fetched once and is stored as a file. Not endorsed by NASA.