⚙️ How-To Wind turbines Power plants & electricity Notebook By

How to Match USGS Wind Turbines to EIA-860 Wind Plants

On this page

Problem statement

Two federal datasets describe the same wind farms from different ends. The US Wind Turbine Database places each turbine on a map with its model and size. EIA-860 lists each wind plant's generators as their owners report them, with capacity, operating date and the plant code that links to monthly generation in EIA-923. Joined, they give generation per turbine model, capacity factors by rotor size, or a map of EIA plants drawn from real turbine positions. USGS stores the EIA plant code on most turbines as eia_id, so the join looks trivial. It mostly works, but wind farms built in phases are split differently on each side. This page joins version 9.1 to EIA-860 for 2025 and measures where the two agree.

Quick answer

Total the turbines by eia_id, total EIA-860's wind generators by Plant Code, and outer-join the two:

e = w.groupby("Plant Code").agg(name=("Plant Name", "first"), st=("State", "first"), mw_eia=("Nameplate Capacity (MW)", "sum"),
                                turbines_eia=("Number of Turbines", "sum"))
u = wt.groupby("eia_id").agg(turbines=("case_id", "size"), mw_usgs=("t_cap", lambda s: s.sum(min_count=1) / 1e3))
j = e.join(u, how="outer")
both = j.dropna(subset=["mw_eia", "turbines"])
print(f"EIA wind plants {len(e):,}; eia_id values {len(u):,}; in both {len(both):,}; EIA only {j.turbines.isna().sum()}; USGS only {j.mw_eia.isna().sum()}")
print(f"capacity within 1 MW: {((both.mw_usgs - both.mw_eia).abs() <= 1).mean():.1%}; turbine count equal: {(both.turbines == both.turbines_eia).mean():.1%}")
EIA wind plants 1,371; eia_id values 1,348; in both 1,276; EIA only 95; USGS only 72
capacity within 1 MW: 73.3%; turbine count equal: 79.2%

EIA-860 lists 1,371 operating wind plants and the turbine file uses 1,348 different plant codes. 1,276 appear in both. For those, the turbines' summed capacity is within 1 MW of EIA's nameplate capacity at 73.3% of plants, and the turbine counts agree exactly at 79.2%. Where they differ, the next steps show why.

Log-log scatter of plant capacity in EIA-860 against the USGS turbines' sum for each plant code, most points on the diagonal and a scatter of red points away from it. Log-log scatter of plant capacity in EIA-860 against the USGS turbines' sum for each plant code, most points on the diagonal and a scatter of red points away from it.
Plant capacity in the two sources.

Where the two disagree

1. Turbines without a plant code

none = wt[wt.eia_id.isna()]
print(f"turbines with no eia_id: {len(none):,}, {none.t_cap.sum() / 1e3:,.1f} MW known; by project year:")
print(none.p_year.value_counts().sort_index().tail(4).to_string())
turbines with no eia_id: 5,051, 2,510.6 MW known; by project year:
p_year
2023.0      20
2024.0      82
2025.0    1666
2026.0     751

5,051 turbines have no plant code, and nearly all are from 2025 and 2026 projects: 1,666 and 751 of them. A plant reaches EIA-860 once it is operating, a year's data appears the following September, and USGS adds the code when it fills in the rest of the record. So the newest turbines are simply too new to link.

2. The largest gaps

gap = both.assign(diff=both.mw_usgs - both.mw_eia).dropna(subset=["diff"]).sort_values("diff")
print(gap[["name", "st", "mw_eia", "mw_usgs", "turbines_eia", "turbines"]].head(3).round(1).to_string(index=False))
print(gap[["name", "st", "mw_eia", "mw_usgs", "turbines_eia", "turbines"]].tail(3).round(1).to_string(index=False))
                name st  mw_eia  mw_usgs  turbines_eia  turbines
  Great Prairie Wind TX  1027.0    361.0         365.0     128.0
    Wagon Wheel Wind OK   598.4      3.4         176.0       1.0
NedPower Mount Storm WV   264.0      6.0         132.0       3.0
                       name st  mw_eia  mw_usgs  turbines_eia  turbines
     Raymond Wind Farm, LLC TX   200.2    401.4          91.0     184.0
King Creek Wind Farm 1, LLC TX   184.0    396.6          47.0      96.0
         Red Cloud Wind LLC NM   350.0   1055.6         124.0     377.0

The largest gaps come from phases, unfinished records and codes on the wrong project. Great Prairie Wind has a third of its turbines under its own code and the rest under two others (Example 1). Wagon Wheel's other 161 turbines are in the file with no code and no capacity yet. Mount Storm's are most likely the 77 turbines filed as "unknown Grant County". At the other end, Raymond carries 93 turbines of its neighbour West Raymond, and King Creek 1 those of King Creek 2. The 377 turbines under Red Cloud's code all belong to a different project, Western Spirit. Matched plant by plant, one misfiled phase counts as two errors, though the turbines are all there.

Diverging bars of the twelve largest plant-level gaps, from Great Prairie Wind at about minus 670 MW to Red Cloud at plus 706 MW. Diverging bars of the twelve largest plant-level gaps, from Great Prairie Wind at about minus 670 MW to Red Cloud at plus 706 MW.
The largest gaps between the turbine file and EIA-860.

3. Plants in only one file

eo = j[j.turbines.isna()].sort_values("mw_eia", ascending=False)
print(eo[["name", "st", "mw_eia"]].head(4).to_string(index=False), f"\nEIA-only capacity: {eo.mw_eia.sum():,.1f} MW")
uo = j[j.mw_eia.isna()]
print(f"USGS codes not among EIA's operating wind plants: {len(uo)}, {int(uo.turbines.sum()):,} turbines")
                        name st  mw_eia
Clines Corners Wind Farm LLC NM   325.0
           Tecolote Wind LLC NM   272.0
       Lundgren Wind Project IA   270.9
         Pryor Mountain Wind MT   239.8 
EIA-only capacity: 2,728.4 MW
USGS codes not among EIA's operating wind plants: 72, 1,389 turbines

95 EIA plants, 2,728.4 MW, have no turbines under their code. Some are phases whose turbines USGS filed under a sibling's code, like Pryor Mountain Wind, whose turbines sit under code 59758 next door. Others are projects USGS has not reached yet. The 72 codes USGS uses that are not among EIA's operating wind plants hold 1,389 turbines: phases EIA reports under another code, and plants since retired or reported under a different technology.

Matching phases together

The data used on this page

The US Geological Survey's US Wind Turbine Database, version 9.1 of 28 September 2026 (Hoen and others, doi:10.5066/F7TX3DN0), public domain, and the wind table (3_2_Wind_Y2025.xlsx) of the US Energy Information Administration's Form EIA-860 for 2025, released in September 2026. Both were saved on this site on 19 October 2026, because USGS replaces its file every quarter and EIA revises the annual files.

from pathlib import Path
import zipfile

import pandas as pd
import requests

Path("data").mkdir(exist_ok=True)
# USGS US Wind Turbine Database v9.1 (28 September 2026) with its metadata and changelog, Census 2024 cartographic
# boundary counties and the EIA-860 2025 wind table. Saved 19 October 2026: USGS replaces the file every quarter.
# Live: https://energy.usgs.gov/uswtdb/assets/data/uswtdbCSV.zip and https://www.eia.gov/electricity/data/eia860/xls/eia8602025.zip
ARCHIVE = "usgs-uswtdb-v9-1-2026-10-19.zip"
if not Path(f"data/{ARCHIVE}").exists():
    Path(f"data/{ARCHIVE}").write_bytes(requests.get(f"https://data.spatialworkflow.io/{ARCHIVE}", timeout=600).content)
Z = zipfile.ZipFile(f"data/{ARCHIVE}")
wt = pd.read_csv(Z.open("uswtdb_V9_1_20260928.csv"), dtype={"t_fips": str})
w = pd.read_excel(Z.open("3_2_Wind_Y2025.xlsx"), sheet_name="Operable", header=1)
w = w[pd.to_numeric(w["Plant Code"], errors="coerce").notna()]       # the sheet ends in a note row
print(f"USWTDB v9.1: {len(wt):,} turbines; EIA-860 2025 operable wind generators: {len(w):,} at {w['Plant Code'].nunique():,} plants")
USWTDB v9.1: 77,839 turbines; EIA-860 2025 operable wind generators: 1,573 at 1,371 plants

Example 1: one complex, three codes

print(j[j.name.fillna("").str.contains("Great Prairie") | j.index.isin([64666, 64667])][["name", "mw_eia", "turbines_eia", "turbines", "mw_usgs"]].to_string())
                          name  mw_eia  turbines_eia  turbines  mw_usgs
Plant Code                                                             
64665.0     Great Prairie Wind  1027.0         365.0     128.0   360.96
64666.0                    NaN     NaN           NaN      75.0   211.50
64667.0                    NaN     NaN           NaN     107.0   301.74

Great Prairie Wind in Texas is one plant of 1,027 MW and 365 turbines to EIA. USGS splits it into three phases with their own codes, 128 turbines under EIA's code and 182 under two codes EIA doesn't use. Together the three hold 310 turbines and 874.2 MW. Matched code by code, it shows as one plant missing two thirds of its capacity and two plants unknown to EIA. Grouping codes that share a project name stem before comparing, or comparing at a coarser level, avoids that.

Example 2: compare by state

st = pd.DataFrame({"eia_mw": w.groupby("State")["Nameplate Capacity (MW)"].sum(),
                   "usgs_mw": wt.groupby("t_state").t_cap.sum() / 1e3}).dropna()
st["diff_pct"] = (st.usgs_mw / st.eia_mw - 1) * 100
print(st.sort_values("eia_mw", ascending=False).head(6).round(1).to_string())
print(f"states within 5%: {(st.diff_pct.abs() <= 5).sum()} of {len(st)}")
     eia_mw  usgs_mw  diff_pct
TX  43942.4  43928.4      -0.0
OK  13594.4  12942.0      -4.8
IA  13411.1  13654.5       1.8
KS   9713.7   9696.6      -0.2
IL   8461.1   8478.9       0.2
CA   6478.8   6022.6      -7.0
states within 5%: 36 of 43

By state the two sources agree far better: Texas's totals differ by 14 MW out of almost 44 GW, and 36 of the 43 states with wind are within 5%. California's turbine total is 7.0% below EIA's because 1,604 of its turbines have no capacity in the turbine file. Oklahoma's 4.8% is mostly Wagon Wheel, whose 161 turbines are still to be described. Plant-level mismatches mostly cancel within a state, so state and county totals from either source can be trusted much more than plant-level joins.

Bars of the percentage difference between USGS and EIA-860 capacity for the 16 largest wind states, all within 5% except California at minus 7.0% and Indiana at minus 6.1%. Bars of the percentage difference between USGS and EIA-860 capacity for the 16 largest wind states, all within 5% except California at minus 7.0% and Indiana at minus 6.1%.
By state the two sources agree.