NEOBIOME

LIT_120: Bolinger & Bolinger (2022), Land Requirements for Utility-Scale PV

Properties4
typesource
source_typeliterature
year2022
topicsenergy_generation

9 min read

LIT_120: Bolinger & Bolinger (2022), Land Requirements for Utility-Scale PV

Source

doi:10.1109/JPHOTOV.2021.3136805 — original publication (opens in a new tab; the file is not redistributed)

Summary

Peer-reviewed empirical study (Bolinger & Bolinger, 2022, IEEE Journal of Photovoltaics; DOI 10.1109/JPHOTOV.2021.3136805) that measures the land occupied by utility-scale ground-mounted PV in the United States and reports updated power density (MW-DC per acre) and energy density (MWh/year per acre) benchmarks. Using ArcGIS polygons drawn around satellite imagery of every US ground-mounted PV plant >5 MW-AC that came online 2007-2019 (736 plants, 35.5 GW-DC), the authors derive 2019 reference densities of 0.35 MW-DC/acre for fixed-tilt and 0.24 MW-DC/acre for single-axis tracking. For Neobiome Intelligence its load-bearing contribution is exactly this fixed-tilt figure: 0.35 MW-DC/acre converts to ~0.865 MW/ha ≈ 865 kWp/ha (rounded to ~870 kWp/ha) — the authoritative primary behind the ~870 kWp/ha ground-mount density proxy already cited (second-hand, via this paper) by the JRC agri-PV report ingested in the same batch. The single most important qualification, stated verbatim by the authors, is that the polygons capture array / direct-impact area only (panels plus inverter pads), NOT the total leased or owned site — so ~870 kWp/ha is a ceiling density and a lower bound on the land a real project must secure. It is an international proxy (data_quality: high = source authority, not NZ-transferability): US irradiance, latitude, and site-quality distributions differ from NZ, and the density must be flagged as a proxy wherever it is applied.

Key claims

- claim: "2019 reference POWER densities (array/direct-impact area, DC basis): 'Power density: 0.35 MWDC/acre (2.8 acres/MWDC) for fixed-tilt and 0.24 MWDC/acre (4.2 acres/MWDC) for tracking.' Conversion here (1 acre = 0.404686 ha): 0.35 MW-DC/acre ≈ 0.865 MW/ha ≈ 865 kWp/ha (rounded to ~870); 0.24 MW-DC/acre ≈ 0.593 MW/ha ≈ 593 kWp/ha. The 0.35/fixed-tilt figure is the primary origin of the ~870 kWp/ha ground-mount density proxy. In-source reciprocal is 2.8 acres/MW-DC (NOT 2.86 — that is a downstream rounding)."
  source_location: "Conclusions and Next Steps, Slide 13 (PDF p.13)"
- claim: "2019 reference ENERGY densities: 'Energy density: 447 MWh/acre (2.2 acres/GWh) for fixed-tilt and 394 MWh/acre (2.5 acres/GWh) for tracking. These are median values—there is obviously a range around the median, depending on latitude, site resource quality, etc.'"
  source_location: "Conclusions and Next Steps, Slide 13 (PDF p.13)"
- claim: "The method measures ARRAY / direct-impact area only, EXPLICITLY EXCLUDING the total site: 'Our polygons focus on the area directly occupied by the arrays (and any associated nearby equipment, such as inverter pads) – NOT on the total leased or owned area of the site... Therefore, only the direct/array area provides usable information about power and energy density.'"
  source_location: "We focus on the area occupied by the arrays..., Slide 7 (PDF p.7)"
- claim: "Worked example of the array-vs-total-site gap — Copper Mountain 3 (Nevada): '345 MWDC and 255 MWAC (DC:AC = 1.35); Array (red) = 945 acres; Total (yellow) = 1,375 acres; Array occupies 69% of the site.' (i.e. the direct array footprint was only 69% of the leased/owned land; total-site density at this plant ≈ 345 MW / 1,375 acres ≈ 0.62 MW/ha ≈ 620 kWp/ha vs the ~900 kWp/ha array density.)"
  source_location: "We focus on the area occupied by the arrays..., Slide 7 (PDF p.7)"
- claim: "The authors' own de-rating instruction for users wanting total-site land: 'Users can de-rate our numbers to suit their own local conditions—e.g., if only 75% of your site is buildable, just multiply our numbers by 75%.'"
  source_location: "We focus on the area occupied by the arrays..., Slide 7 (PDF p.7)"
- claim: "Sample and scope: 'Our sample consists of 736 plants totaling 35.5 GWDC (27.0 GWAC) that came online from 2007-2019 across 38 (of 50) states... This sample includes 92% of all utility-scale (i.e., ground-mounted and >5 MWAC) PV plants that came online over this 13-year period... our analysis focuses on the period from 2011-2019.'"
  source_location: "Overview of our sample—in numbers, Slide 8 (PDF p.8)"
- claim: "Method: 'We used ArcGIS to draw polygons around satellite imagery (from Google Earth and Maxar/Digital Globe) of each plant's PV array(s) and to calculate the polygons' acreage,' over 'the universe of ground-mounted PV plants >5 MWAC.'"
  source_location: "What we did, Slide 4 (PDF p.4)"
- claim: "Density has risen over the study window (so pre-2019 benchmarks understate current density): 'Median power density (MWDC/acre) increased by 52% (fixed-tilt) and 43% (tracking) from 2011 through 2019... Median energy density (MWh/year/acre) increased by 33% (fixed-tilt) and 25% (tracking) from 2011 through 2019.'"
  source_location: "Conclusions and Next Steps, Slide 13 (PDF p.13)"
- claim: "Why fixed-tilt is denser than tracking (ground-coverage ratio): 'Fixed-tilt's higher power density reflects its higher ground coverage ratio (GCR): ~0.40-0.50 GCR for fixed-tilt versus ~0.25-0.40 GCR for tracking.'"
  source_location: "Power density has steadily increased over time, Slide 10 (PDF p.10)"

Neobiome Intelligence relevance

  • Primary origin of the ~870 kWp/ha ground-mount density proxy (RT_304). The fixed-tilt reference of 0.35 MW-DC/acre converts to ~865 kWp/ha ≈ ~870 kWp/ha — the authoritative primary behind the “global average ~0.87 MWp/ha” that the JRC agri-PV report (this same batch) cites as “Bolinger and Bolinger, 2022”. Single-axis tracking is lower at 0.24 MW-DC/acre ≈ 593 kWp/ha. These calibrate the exclusive-use ground-array density input against which agri-PV’s lower, spacing-driven density is compared LIT_120.
  • ⚠ Basis caveat — array footprint, not total site. The authors state verbatim that their polygons measure only the array / direct-impact area (panels + inverter pads), not the total leased or owned site. So ~870 kWp/ha is the densest figure — a lower bound on the land a project must secure and an upper bound on capacity-per-site-hectare. A real project’s total-site footprint is larger: at Copper Mountain 3 the array was only 69% of the leased site, implying a total-site density near ~620 kWp/ha at that plant (consistent with the ~650-750 kWp/ha total-site range in the solar-land-use cr_). When the model wants land to lease/own, de-rate ~870 by the buildable fraction (the authors’ own guidance) rather than using it raw LIT_120.
  • 🔴 Premise flag for the modeling work (flagged, NOT resolved). The engine’s interim exclusive-use pv_ground_kwp_per_ha = 150 (RT_304) sits ~4-6x below both the ~870 kWp/ha array density and the de-rated ~650-750 kWp/ha total-site density. On its face 150 looks far too low unless it deliberately encodes a very different basis (e.g. a heavily conservative usable-land fraction, or ground reserved only near dwellings). the modeling work should confirm the basis of 150 before wiring any ground-mount density. Not resolved here LIT_120.
  • ⚠ Proxy scope. US plants, 2007-2019. The physical densities are broadly transferable (module and racking technology is global) but US irradiance, latitude and site-quality distributions differ from NZ, and the paper’s cost/LCOE framing is US-specific. Apply as an international proxy and flag it; a NZ-specific ground-mount density (e.g. from the Lincoln/Massey ~1.5 ha/MW figure in this batch, or NZ fast-track consents) is the preferred NZ anchor where available.

Research targets

Research gaps

  • Informs (does not close) RT_304 — the ground-array density factor pv_ground_kwp_per_ha (engine interim 150). Supplies the international literature anchor for that factor: array-footprint ~870 kWp/ha fixed-tilt / ~590 kWp/ha tracking, de-rating to ~650-750 kWp/ha at total-site basis. RT_304 stays open: the two roof factors it also centres on (roof_usable_frac 0.40, pv_m2_per_kwp 6.0) are untouched, and the density figure needs a NZ-specific confirmation before it replaces the interim 150. No new RT opened (minimise-new-RTs).

Retrieval provenance

  • Upstream source: Bolinger, M. & G. Bolinger (2022), “Land Requirements for Utility-Scale PV: An Empirical Update on Power and Energy Density”, IEEE Journal of Photovoltaics, doi:10.1109/JPHOTOV.2021.3136805 (open access). The retrieved file is the authors’ own LBNL summary slide-deck (14 slides, presented 1 Feb 2022, hosted at emp.lbl.gov/publications/land-requirements-utility-scale-pv) — same authors, same figures as the peer-reviewed article.
  • Prepared by: retrieval pass on 2026-07-25, as part of the solar land-use / cool-store retrieval run (worklist folder the retrieval record held with the project; run summary RETRIEVAL_REPORT.md, solar primary S1).
  • Retrieval target: locate the authoritative primary behind the ~870 kWp/ha ground-mount PV density proxy (RT_304) and the array-vs-total-site land-use distinction the solar-land-use cr_ builds on.
  • Verification: the raw is the genuine authoritative LBNL PDF (not an AI synthesis). Every figure in Key claims was independently re-verified here via pdftotext -layout against the retrieved PDF and quoted verbatim with exact source_location. data_quality: high denotes source authority (peer-reviewed IEEE article + authors’ LBNL summary); it does NOT imply NZ-transferability — the density is an international proxy and is flagged as such on the page.

Notes

  • Retrieved artifact is the summary slide-deck, not the full article. The PDF is the authors’ 14-slide LBNL webinar summary (title-slide “Slide 1”, author Mark Bolinger, LBNL, 1 Feb 2022). It presents the identical figures and conclusions as the peer-reviewed IEEE J. Photovoltaics article (DOI 10.1109/JPHOTOV.2021.3136805) but in condensed form. Slide numbers 1-14 map 1:1 to PDF pages 1-14; source_location cites the slide title + PDF page.
  • Source is in ACRES; hectare figures are downstream. Every density in the paper is stated per acre (MW-DC/acre, MWh/acre, acres/MW-DC). The kWp/ha and MW/ha figures on this page are editorial conversions (1 acre = 0.404686 ha), clearly labelled as such. The verbatim acre figures are the authoritative values.
  • Do not write 2.86 acres/MW-DC. The in-source reciprocal of 0.35 MW-DC/acre is stated as 2.8 acres/MW-DC (Slide 13). The “2.86” seen in the solar-land-use cr_ is a downstream re-rounding, not in this source.
  • Array-only vs total-site is the crux. The ~870 kWp/ha proxy measures panels-plus-pads, not the land a developer leases or owns. Anywhere the model needs total land (siting, cost, land-competition), de-rate to total-site (~650-750 kWp/ha, or by the site’s buildable fraction) — the array density is a lower bound on land requirement.
  • Underlying peer-reviewed article not separately retrieved. The full IEEE J. Photovoltaics text was not downloaded (the LBNL summary deck carries all figures used here). If a reviewer needs the full methodology or per-year tables, the article (DOI above) is the retrieval target — not currently opened as an RT (the summary is sufficient for the density proxy).

Connections

Referenced by

EDT domains (1): D01: Renewable Energy & Storage Systems

Sources (1): CR_059

Graph View

Backlinks