RD_013: PVGIS national solar specific-yield by NZ region (16 regions)

Source

https://re.jrc.ec.europa.eu/pvg_tools/en/tools.html — original source (opens in a new tab; the file is not redistributed)

PVGIS national solar specific-yield by NZ region (16 regions)

Reference dataset — PVGIS-ERA5; now validated 13/15 regions vs NIWA ground normals (CR_025)

Underlying data: EU JRC PVGIS (PVGIS-ERA5). The 16-region CSV was compiled via the project retrieval pipeline on 2026-05-31 (compiling operator unrecorded — not authored by the project owner). One representative centre per region; 1 kWp crystSi, north-facing, tilt = round(latitude), 14% system loss. Now VALIDATED against NIWA ground-station normals (1991–2020) for 13/15 centres within ±10% (CR_025; resolves RT_207) — was only ~2% (Lower Moutere/Nelson). PVGIS runs ~+5% high vs NIWA (conservative-haircut candidate); Wellington (+16.4%) and Dunedin (+10.8%) are flagged where the coarse grid misses local cloud/topography.

Summary

A single-methodology PVGIS-ERA5 simulation of solar specific yield for all 16 NZ regions, with annual and full 12-month outputs on identical parameters (1 kWp crystalline-silicon, north-facing, tilt = round(latitude), 14% system loss), so the table is directly comparable across regions. It is the national extension of the CR_010 Nelson-Tasman pilot figure and resolves RT_138 — giving the NI energy model a defensible kWh/kWp/yr anchor for any NZ region plus the monthly columns needed to size for the winter minimum that drives battery design. RD_013

Key claims

- claim: "Annual specific yield ranges from 1,246 kWh/kWp/yr (Southland, Invercargill) to 1,573 kWh/kWp/yr (Wellington) — a 1.26× spread; Tasman/Lower Moutere is 1,389 kWh/kWp/yr, inside CR_010's independent 1,350–1,380 band."
  source_location: "annual_yield_kWh_kWp column"
- claim: "Implied capacity factor (annual yield ÷ 8,760 h): Tasman ~16.0%, Wellington ~18.0%, Southland ~14.2% — every modelled region is at or above the 14% national CF assumption inherited from OT_001/CR_009."
  source_location: "computed from annual_yield_kWh_kWp"
- claim: "Every one of the 16 regions records its monthly minimum in June; per-region June output (Tasman 78.4, Southland 61.7, Otago 67.4 kWh per kWp) is the binding value for off-grid/battery winter sizing."
  source_location: "winter_low_month / winter_low_kWh columns"
- claim: "Summer-high:winter-low monthly ratio widens southward: Auckland 1.74×, Tasman 1.84×, Wellington 2.06×, Southland 2.17× — South Island designs need proportionally more storage/backup than the North at equal annual yield."
  source_location: "summer_high_kWh / winter_low_kWh columns"
- claim: "Method: PVGIS-ERA5, 1 kWp crystSi, north-facing (azimuth 0°), tilt = round(latitude) (36° Northland to 46° Otago/Southland), 14% system loss; validated against NIWA SolarView at Lower Moutere within ~2%."
  source_location: "CSV header lines 1–2"
- claim: "PVGIS year-to-year variability for the pilot (Lower Moutere, north-facing, 40° tilt) is a standard deviation of 31.24 kWh on the 1,389 kWh/kWp/yr mean — a ~2.25% coefficient of variation — giving a P90 (bad-year, P50 − 1.28·SD) of ~1,350 kWh/kWp/yr, i.e. P90/P50 ≈ 0.971. NZ solar is stable inter-annually, so the annual bad-year solar penalty is mild (~3%); the dominant solar resilience risk is seasonal (winter / cloudy spells), not the annual P90."
  source_location: "PVGIS v5_2 PVcalc API E_y / SD_y, lat −41.3 lon 173.15, retrieved 2026-06-10"

Regional solar yield — model data (② pv_yield_kwh_per_kwp)

From RD_013 (annual kWh/kWp/yr · June low kWh per kWp):

RegionYieldJun lowRegionYieldJun low
Northland1,38882.7Tasman1,38978.4
Auckland1,47186.3Nelson1,43679.2
Waikato1,33480.4Marlborough1,51287.1
Bay of Plenty1,37082.7West Coast1,31373.4
Gisborne1,48386.2Canterbury1,46375.6
Hawke’s Bay1,43581.3Otago1,33067.4
Taranaki1,54484.2Southland1,24661.7
Manawatū-Whanganui1,39073.3Wellington1,57379.4

Neobiome Intelligence relevance

Generalises the D01 solar supply side from the Tasman pilot to all of NZ. The model can now look up pv_yield_kwh_per_kwp for any region rather than assuming the Tasman figure or a flat 14% CF. The Tasman row (1,389) lands inside CR_010’s independently-derived band, giving cross-method corroboration. The monthly columns are the load-bearing addition: every region troughs in June, so winter coverage/backup can be sized against the binding month, and the southward-widening summer:winter swing means South Island designs need proportionally more storage. RD_013

Tag for the model: pv_yield tagged sourced for the pilot (Tasman, NIWA-validated); the other 15 regions are sourced-but-unvalidated (PVGIS-ERA5). Tilt is round(latitude) for annual-max yield — for off-grid battery-backed designs CR_010 recommends a winter-biased 40–45°, so these annual-max yields slightly overstate winter output (see gaps). RD_013

Bad-year band (D16): the PVGIS year-to-year SD gives P90 ≈ 0.971 × P50 (2026-06-10 enrichment) — the model’s pv_yield_p90 column is filled at 0.971 × pv_yield for all 16 regions, which unlocks the D16 bad-year SSI run. The ratio is derived at the pilot site; per-region inter-annual variability (likely higher in cloudier West Coast/Southland) could refine it later, but the headline is that NZ solar’s annual bad-year penalty is small — seasonal/winter risk is the binding resilience constraint, not the annual P90. Resolves RT_100. RD_013

Research targets

Documents to retrieve

  • NIWA SolarView national / per-region runs — to validate the 15 non-Tasman PVGIS rows against ground-station + terrain data (only Lower Moutere was validated here). Free at data.niwa.co.nz/solarview. New target RT_207. Brent (2020) Solar Atlas of NZ is the peer-reviewed cross-check — now ingested as LIT_071 (RT_101 resolved 2026-07-17) — note LIT_071 validates the satellite GHI dataset (annual rMBD 3.2%, winter JJA 8.3%), not the PVOUT kWh/kWp layer, so it corroborates RD_013’s dataset accuracy rather than directly cross-checking the specific-yield rows.

Research gaps

  • Seasonal-basis reconciliation — RD_013’s monthly summer:winter ratio for Tasman (~1.84×) differs from CR_010’s 3.7–4.5× PVGIS solar-angle ratio and 1.5× NIWA sunshine-hours ratio (different bases). The engine needs one documented winter-sizing basis. (Engine-design note, not a retrieval target.)
  • Winter-optimised tilt — these yields use round(latitude) (annual-max); a re-run at off-grid winter-biased tilt (40–45°) would better serve battery sizing. (Engine-design note.)

Connections

Links to

Sources (3): CR_010 · CR_025 · LIT_071

Referenced by