Conditional reliability of gridded rainfall datasets in Nigeria

Segun Ezekiel Akorede , Chioma Suzan Dibani

Gridded rainfall datasets support Nigerian research and climate services, but published agreement depends on the rainfall property, release and validation design. This critical narrative review audits 18 Nigerian gauge-reference evaluations and two contextual comparator studies. The core register contains five complete-text, five passage-only and eight abstract-only records. An archived supplementary search maps 93 result appearances to 52 screened records; historical search provider details remain unavailable, limiting reproducibility of retrieval. Domain-based appraisal separates source access from the assessability of product identity, reference support, matching, diagnostics and provenance. No core entry establishes station-and-date independence across the relevant product-input pathways; this unresolved status does not demonstrate actual overlap. Removing abstract-only evidence retains conditional CHIRPS leads and a passage-supported monthly comparison favouring MSWEP v2.2 and IMERG Final v6, while several zonal and local product orderings lose their supporting records. Restrictions on complete texts further narrow the recommendations. Reported annual spatial agreement, monthly climatology and daily correlation test different quantities and cannot form a common accuracy ranking. Six pairs across four national NiMet-based studies share stated reference-year windows, indicating possible dependence that warrants station-level checks. A worked appraisal traces ARC2 gauge inputs and distinguishes zone-average event detection from station-level performance. Practical guidance specifies the statistics, evidence base, and checks required for accumulated rainfall, drought, wet days, extremes, and trends. The principal contribution is a traceable account of which Nigerian recommendations survive source-access restrictions and which remain limited by unresolved methods or observational provenance.