# Grid-electricity data quality assessment

Generated: 2026-06-19
Scope: Decide whether the consolidated 2020–2024 Nigeria grid generation datasets are good enough to materially improve NEFDB's location-based grid emission factors.

## 1. Confidence in the reconciled national generation mix

| Year | Reconciled total (GWh) | Reconciled gas (GWh) | Reconciled hydro (GWh) | Confidence | Why |
|---|---:|---:|---:|:---:|---|
| 2020 | 38,000.00 | 29,484.16 | 8,341.99 | **High** | Total from Ember/OWID; hydro from Nigerian Energy Data Bank plant records for Kainji, Jebba, Shiroro and NESCO. Only year with full official plant-level hydro dispatch. |
| 2021 | 39,200.00 | 29,804.02 | 9,153.81 | **Medium** | Total from Ember/OWID; hydro/solar/biomass/oil from IRENA. NEDB Kainji/Jebba records stop at 2020, so the hydro total is an international estimate, not plant-verified. |
| 2022 | 38,000.00 | 28,477.93 | 9,235.35 | **Medium** | Total from Ember/OWID (IEA corroborates 37,916 GWh); fuel split from IRENA. No Nigeria plant-level hydro. |
| 2023 | 40,920.00 | 31,117.21 | 9,426.41 | **Medium** | Total from Ember/OWID; fuel split from IRENA. EIA gives a higher total (42,500 GWh), creating a material source conflict that is noted but not adopted. |
| 2024 | 37,093.70 | 25,433.85 | 11,469.85 | **High** | Total and hydro from Guardian Nigeria summary of NERC 2024 official data. Hydro share of 30.92% is materially higher than Ember/OWID (~24.5%). |

Summary: two years (2020 and 2024) are **high** confidence; three years (2021–2023) are **medium** confidence because the hydro total is inferred from international databases rather than Nigerian plant dispatch records.

## 2. Confidence in plant-level hydro time series

| Plant | Years with usable data | Confidence | Notes |
|---|---:|---:|---|
| Kainji | 2020 only | **High** 2020; **Low** 2021–2024 | NEDB plant record validated by Okakwu et al. No annual records found after 2020. |
| Jebba | 2020 only | **High** 2020; **Low** 2021–2024 | NEDB plant record. No annual records found after 2020. |
| Shiroro | 2020–2021 | **High** 2020–2021; **Low** 2022–2024 | NEDB plant record; stops after 2021. |
| NESCO | 2020–2022 | **High** 2020–2022; **Low** 2023–2024 | NEDB record; plant-level unit label was corrected (labels said MWh but values were kWh). |
| Zungeru | None | **Low** | 700 MW commissioned June 2023; no verified annual GWh located. |
| Gurara | None | **Low** | Small plant; no verified annual GWh located. |

The Adoghe et al. (2023) plant-level figures for 2020 were rejected because they are an order of magnitude lower than NEDB/Okakwu and cannot be reconciled without further clarification.

## 3. Source conflicts and how they were resolved

| Conflict | Resolution | Remaining uncertainty |
|---|---|---|
| **2020 hydro:** NEDB plant sum = 8,341.99 GWh vs Ember/OWID = 7,690 GWh (8.5% lower). | Used NEDB plant-level total because it is official, plant-resolved and validated by Okakwu et al. | Small residual from difference between plant-sum and international-database estimates. |
| **2020 total:** Ember/OWID 38,000 GWh vs IRENA 35,427.85 GWh. | Retained Ember/OWID as the national total anchor; IRENA kept as candidate row. | IRENA undercounts total by ~2,570 GWh. |
| **2021–2023 hydro:** NEDB missing Kainji/Jebba/Shiroro/NESCO after 2020/2021/2022. | Used IRENA hydro totals for the national mix; plant-level rows left blank. | Hydro could differ if NERC/TCN dispatch differs materially from IRENA. |
| **2023 total:** Ember/OWID 40,920 GWh vs EIA 42,500 GWh. | Retained Ember/OWID for time-series consistency; EIA kept as candidate row. | ~1,580 GWh difference, directionally raising the true total. |
| **2024 hydro:** Guardian/NERC 11,469.85 GWh (30.92%) vs Ember/OWID 9,200 GWh (~24.5%). | Used Guardian/NERC official figure. | Press-sourced NERC summary; primary NERC annual report should be obtained when available. |
| **2024 total:** Guardian/NERC 37,093.70 GWh vs Ember/OWID 37,560 GWh. | Used Guardian/NERC official figure. | Difference is small (~466 GWh, 1.3%). |
| **Oil/biomass:** IRENA reports small oil generation (17–46 GWh/year) and biomass; Ember/OWID reports zero oil and near-zero biomass. | Used IRENA values for 2020–2023 and Ember/OWID for 2024. | Contribution to EF is negligible (<0.2% of generation). |

## 4. Should the national location-based EF be recalculated?

**Yes.** The current `location_based_ef.csv` and `grid_electricity.ts` record (`ng.grid.national.location.2020`, 0.5717 kgCO₂e/kWh) was built from an incomplete 2020 dataset that captured only about 7,282 GWh of gas and 76 GWh of hydro — roughly 20% of actual national generation. Kainji, Jebba and Shiroro were excluded because their units were misread.

The reconciled 2020 mix is materially better:

- Gas: 29,484.16 GWh (up from 7,281.71 GWh)
- Hydro: 8,341.99 GWh (up from 75.59 GWh)
- Total: 38,000 GWh (up from ~7,357 GWh)

Approximate recomputed 2020 EF using the same IPCC 2006 gas factor and 35% efficiency is **~0.45 kgCO₂e/kWh**, about **21% lower** than the current Phase 1 value. A 2024 recomputation using the NERC hydro share gives **~0.40 kgCO₂e/kWh**, about **30% lower**. These differences are material for corporate Scope 2 reporting and regulator submissions.

The national EF should be recalculated for **2024** (most recent official data, highest hydro share) and **2020** (most complete plant-level data). If only one year can be shipped now, use **2024** for current reporting and keep 2020 as a back-cast.

## 5. Should DisCo-level EFs be recalculated?

**Yes.** All current DisCo factors (`ng.grid.disco.*.2020`) use a 2020-derived regional gas/hydro split and Q2 2024 billed energy. The underlying national hydro share used to anchor those regional splits was wrong because the Phase 1 2020 hydro total excluded the large hydro plants. Recomputing with the reconciled national mix — especially the 2024 NERC hydro share — will change the implied regional generation mixes and lower DisCo EFs that are currently hydro-heavy.

DisCo EFs should be recalculated using the **2024** national generation mix, because the billed-energy denominator is already from Q2 2024. The same regional allocation coefficients can be retained, but they should be re-normalised against the corrected national fuel shares.

## 6. Recommendation

**ENRICH FACTORS**

The consolidated datasets materially improve the data foundation for 2020 and 2024. The current Phase 1 national location-based EF under-represents hydro by more than 100×, producing a factor that is too high. Recomputing with the reconciled generation mix will give a more accurate, defensible EF and better DisCo factors.

### Required follow-up work
1. Recompute `location_based_ef.csv` and `location_based_factor.json` for 2024 (and optionally 2020) using the reconciled generation mix and the existing IPCC 2006 stationary-combustion factors.
2. Recompute DisCo-level factors for 2024 using the corrected 2024 national gas/hydro split.
3. Add uncertainty bands and update confidence metadata.
4. Obtain 2021–2024 NERC quarterly dispatch data or Elec-T dashboard records to upgrade 2021–2023 from medium to high confidence.