MODIS Burned Area (MCD64A1)¶

MODIS/Terra+Aqua Burned Area Monthly L3 Global 500m SIN Grid V061 (MCD64A1, combined Terra and Aqua MODIS monthly burned area at 500 m)

Being replaced

At a glance¶

Note on period: Starts in November 2000 (earlier data are not distributed because of unresolved Terra MODIS crosstalk and calibration issues). The early record uses Terra alone; Aqua observations were added once Aqua was operating. A CMR query on 2026-10-13 returned July 2026 as the latest month (produced 2026-09-10). June 2001 is heavily degraded by an extended Terra outage.

Overview¶

MCD64A1 is a monthly, global burned area product that combines 500 m Terra and Aqua MODIS surface reflectance with 1 km MODIS active fire detections and records, for every burned 500 m cell, the date on which it burned. Dynamic thresholds, trained on active fire observations, are applied to a burn-sensitive vegetation index from the shortwave infrared bands 5 and 7 and to a temporal texture measure. It is the most widely used reference burned area record, underpinning fire emission inventories such as GFED and FINN as well as carbon cycle and air quality research.

What changes with the successor¶

MCD64A1 C6 → MCD64A1 C6.1¶

Transition: 2021-11

The full record was regenerated as Collection 6.1 with improved Level-1B calibration. The burned area algorithm itself is unchanged from C6, and C6 (version 006) distribution ended on 31 July 2023.

Sources: MCD64A1 C6.1 User Guide · Final Notice: MODIS Version 6.0 Land Data Products Decommissioned on July 31, 2023

MODIS MCD64A1 C6.1 → VIIRS VNP64A1 C2 Planned¶

Transition: 2027

VNP64A1 adapts the MCD64A1 C6.1 algorithm to Suomi NPP VIIRS and keeps the same 500 m sinusoidal grid and layer set. Over 2012–2023 the monthly global totals correlate at 0.989, but VIIRS maps 2–6% more area every year, so the two records should not simply be spliced. And because Suomi NPP data delivery ends in November 2026, VIIRS continuity itself depends on Collection 3, which is planned to add NOAA-20.

Sources: Giglio et al. (2025), The NASA VIIRS burned area product, global validation, and intercomparison with the NASA MODIS burned area product · Transition from MODIS to VIIRS (Earthdata, 2025-05-02) · VNP64A1 V002 (LP DAAC DOI landing page)

Compare the two side by side

Available variables¶

The five science data sets (SDS) listed in the LP DAAC and LAADS file specifications (HDF-EOS2, HDF4). The layer names contain spaces (e.g. "Burn Date"), so use them exactly as written. There is no scale factor. In Google Earth Engine (MODIS/061/MCD64A1) the bands are named without spaces (BurnDate, Uncertainty, QA, FirstDay, LastDay).

instantaneous mean accumulated maximum minimum static — instantaneous: value at the time stamp · mean/accumulated/maximum/minimum: over the indicated interval · static: does not change with time

Monthly burned area layers (500 m)¶

The full variable list is in the official documentation.

How to get the data¶

• NASA Earthdata Search (LP DAAC, Earthdata Cloud) — Web search and HTTPS download / earthaccess / direct S3 access in AWS us-west-2
  • short_name MCD64A1, version 061. Files are per-tile HDF-EOS2 (.hdf) and need a free Earthdata Login. The S3 prefix is s3://lp-prod-protected/MCD64A1.061/, with temporary credentials from https://data.lpdaac.earthdatacloud.nasa.gov/s3credentials. MODIS distribution from the old LP DAAC Data Pool (e4ftl01) ended on 30 June 2025.
• AppEEARS — Area and point subsetting with reprojection (GeoTIFF, NetCDF4)
  • Convenient when you want only a region of interest on a lat-lon grid rather than raw tiles.
• LP DAAC product page (DOI landing page) — Documentation (User Guide, ATBD) and layer information
• University of Maryland burned area server (fuoco.geog.umd.edu) — SFTP — Plate Carrée GeoTIFF and shapefile subsets
  • The science team provides regionally reprojected GeoTIFFs and burned area shapefiles; the openly shared login details are in the User Guide.
• Google Earth Engine — Cloud analysis (MODIS/061/MCD64A1)
• NASA FIRMS (Burned Area mode) — Web map viewer
  • Monthly and annual burned area can be overlaid on VIIRS VNP64A1 for comparison.

Example code¶

Downloads the March 2025 tiles covering the eastern Korean Peninsula with earthaccess and counts burned pixels and area. HDF4 files are opened with rasterio built against GDAL with HDF4 support. Pick the layer from the subdatasets list; one cell is 463.31271653 m squared.

# pip install earthaccess rasterio numpy   (rasterio needs a GDAL build with HDF4 support)
import earthaccess
import numpy as np
import rasterio

earthaccess.login()  # free NASA Earthdata account
results = earthaccess.search_data(
    short_name="MCD64A1", version="061",
    temporal=("2025-03-01", "2025-03-31"),
    bounding_box=(127, 35, 130, 39),   # (west, south, east, north)
)
files = earthaccess.download(results, "mcd64a1")

for f in files:
    with rasterio.open(f) as src:
        sds = [s for s in src.subdatasets if s.endswith("Burn Date")][0]
    with rasterio.open(sds) as b:
        burn = b.read(1)
    burned = (burn >= 1) & (burn <= 366)       # 0 unburned, -1 unmapped, -2 water
    area_km2 = burned.sum() * 463.31271653**2 / 1e6
    print(f, burned.sum(), "pixels,", round(area_km2, 1), "km2")

Practical tips¶

• Burn Date is a date, not an area. Compute area as the number of pixels with values 1–366 times the cell area (463.31271653 m squared, about 21.5 ha), excluding 0, -1 and -2.
• The data are on a sinusoidal projection. Use nearest-neighbour resampling when moving to a lat-lon grid; averaging or bilinear interpolation mixes dates and creates spurious ones.
• A 500 m product misses many small fires. Against Landsat 30 m reference data (Giglio et al. 2025), MCD64A1 had 71.5% omission and 34.2% commission error, and GFED5 adds small fires to reach roughly 90% more burned area than MCD64A1. Underestimation is large where small wildfires and crop-residue burning dominate, as on the Korean Peninsula.
• The User Guide recommends treating burned area in croplands as low confidence. Do not rely on MCD64A1 alone for agricultural burning.
• June 2001 (extended Terra outage) and August 2020 (about two weeks of lost Aqua telemetry, partly offset by direct broadcast data) are degraded months.
• Under persistent cloud, Burn Date is -1 (unmapped) or the First Day–Last Day window is shorter than the month. Do not count missing data as unburned (check QA bits 1 and 2).
• Terra and Aqua orbit drift has changed observation times and pixel spacing in recent data. When examining trends in the 2020s, consider this change and compare with VIIRS VNP64A1, which overlaps from 2012.
• The production timestamp at the end of the file name changes with every reprocessing. Search by date and tile in scripts rather than hard-coding full file names.

Notices and change history¶

References for your paper¶

Dataset citation¶

MCD64A1 V061 (LP DAAC) — DOI 10.5067/MODIS/MCD64A1.061

@misc{mcd64a1_v061,
  author    = {Giglio, Louis and Justice, Christopher and Boschetti, Luigi and Roy, David},
  title     = {{MODIS/Terra+Aqua} Burned Area Monthly {L3} Global 500m {SIN} Grid {V061}},
  year      = {2021},
  publisher = {NASA EOSDIS Land Processes Distributed Active Archive Center},
  doi       = {10.5067/MODIS/MCD64A1.061}
}

Reference papers¶

Collection 6 burned area algorithm (citation recommended in the User Guide) — DOI 10.1016/j.rse.2018.08.005

@article{giglio2018mcd64,
  author  = {Giglio, Louis and Boschetti, Luigi and Roy, David P. and Humber, Michael L. and Justice, Christopher O.},
  title   = {The Collection 6 {MODIS} burned area mapping algorithm and product},
  journal = {Remote Sensing of Environment},
  volume  = {217},
  pages   = {72--85},
  year    = {2018},
  doi     = {10.1016/j.rse.2018.08.005}
}

Global validation of the Collection 6 product — DOI 10.1016/j.rse.2019.111490

@article{boschetti2019mcd64val,
  author  = {Boschetti, Luigi and Roy, David P. and Giglio, Louis and Huang, Haiyan and Zubkova, Maria and Humber, Michael L.},
  title   = {Global validation of the collection 6 {MODIS} burned area product},
  journal = {Remote Sensing of Environment},
  volume  = {235},
  pages   = {111490},
  year    = {2019},
  doi     = {10.1016/j.rse.2019.111490}
}

Acknowledgements¶

LP DAAC shares the data without restriction and encourages citation of the dataset DOI (EOSDIS Data Use and Citation Guidance). Cite Giglio et al. (2018) for the algorithm as well.

Information on the datasets used in this study was compiled with the help of the Korean Climate Data Navigator (KCDN) (Sim, 2026; https://doi.org/10.5281/zenodo.23300923).

Links¶

• LP DAAC MCD64A1 V061 (Earthdata catalog)
• MCD64A1 C6.1 User Guide
• MCD64 ATBD (C6)
• MCD64A1 file specification (LAADS)
• MODIS Land validation status (MCD64)
• MODIS/VIIRS Land product quality assessment (LDOPE) — known issues
• Aqua MODIS acquisition plan (orbit drift, LP DAAC)
• Terra MODIS acquisition plan (orbit drift, Earthdata)