About NINOBOT's sources

Who we are

一般社団法人 にのラボ (GIA NINOLAB Japan) is a Japanese non-profit association formed to publish interactive Stories that make climate and policy information actionable for the public. The editorial and technical work behind each Story is carried out in partnership with 合同会社 TMP Climate, a specialist climate consultancy that works alongside government agencies, research institutes, and operational forecasters. TMP's regional specialists — covering different parts of the world — author and review the Story content in the bot's corpus.

What you will see cited

Citation markers in the bot's answers fall into a small number of categories:

Explainers

Short, plain-language articles about how El Niño works, how it is measured, and how its impacts propagate. Written by the TMP team.

Regional summaries

Country- or region-level briefs on what past El Niño events have meant for local climate, water, agriculture, and energy. Each is reviewed by the TMP regional specialist for that area.

Forecast bulletins

Current outlooks from operational centres such as NOAA's Climate Prediction Center, the Japan Meteorological Agency, the International Research Institute for Climate and Society, and ECMWF. Refreshed on each centre's publication schedule.

Methodology notes

Documentation of the data, definitions, and analytical methods behind the bot's answers.

Academic papers

Peer-reviewed climate science indexed and chunked for retrieval, with attribution to authors, journal, and DOI.

Maps and reference tables

Supplementary visual and tabular reference material drawn from authoritative datasets such as ERA5, GADM, and HydroBASINS.

How sources are developed

Content authored by the TMP team — explainers, regional summaries, and methodology notes — goes through a structured workflow. Each document carries machine-readable metadata recording author, last-updated date, version, lifecycle stage, content origin, and review status. Drafts are not retrieved by the bot. Before a draft becomes active, it is reviewed by a TMP specialist with relevant expertise in the topic or region. Active sources are version-locked across our six launch languages: an update in one language schedules updates in the others. External sources — forecast bulletins and academic papers — are ingested with attribution intact, so every cited chunk records its publisher, source URL, and authority tier.

The in-house empirical analysis

Some of NINOBOT's most useful answers — those that begin with "based on our in-house analysis of historical El Niño impacts" — come from an analysis we computed specifically for the bot. The analysis uses 47 years (1979–2025) of ERA5-Land reanalysis data at roughly 10-kilometre resolution, produced by the European Centre for Medium-Range Weather Forecasts and distributed through the Copernicus Climate Data Store. ENSO state is classified using the Relative Oceanic Niño Index (RONI) maintained by NOAA's Climate Prediction Center; RONI replaced the older Oceanic Niño Index in February 2026 as the operational ENSO classification index. The analysis produces calibrated qualitative impact tags — low or typical, moderate, or high — for every administrative polygon and sub-basin in our spatial reference layers. The full methodology document is part of the corpus and can be opened directly from any citation that points to it.

Why we think this is reliable

Three commitments anchor reliability. First: every TMP-authored source is reviewed by a regional or topical specialist before it enters the corpus, and that review is recorded as part of the source's metadata. Second: the bot is built to abstain. When retrieval does not surface grounded evidence above a confidence threshold, the bot says so rather than improvising — silence is preferred to invention. Third: every claim the bot makes is linked back to a citable chunk you can read for yourself. We make our reasoning visible. Where we are uncertain, we say so.