Skip to content

MODIS vs VIIRS on the Same Fire

Running one fire through both sensors is the fastest way to see what MODIS retirement will change in your own numbers. The example uses the first five days of the Big Grass Fire in southeastern Oregon, from July 23, 2026.

Pull Both Sensors

from firms import get_fires, high_confidence, daily_counts

BOX = (-118.3, 41.9, -116.4, 44.6)
START = "2026-07-23"

modis = get_fires("MODIS_SP", BOX, 5, START)
noaa20 = get_fires("VIIRS_NOAA20_SP", BOX, 5, START)
noaa21 = get_fires("VIIRS_NOAA21_SP", BOX, 5, START)

counts = daily_counts([high_confidence(f) for f in (modis, noaa20, noaa21)])
print(counts)

daily_counts() returns one row per date and one column per source.

Map Them Side by Side

import matplotlib.pyplot as plt

fig, axes = plt.subplots(1, 2, figsize=(12, 6), sharex=True, sharey=True)
for ax, frame, title in [(axes[0], modis, "MODIS (1 km)"), (axes[1], noaa20, "VIIRS NOAA-20 (375 m)")]:
    ax.scatter(frame["longitude"], frame["latitude"], s=4, c=frame["frp"], cmap="inferno")
    ax.set_title(f"{title}: {len(frame)} detections")
    ax.set_xlabel("Longitude")
axes[0].set_ylabel("Latitude")
fig.savefig("big_grass_modis_vs_viirs.png", dpi=150, bbox_inches="tight")

Reading the Difference

Expect VIIRS to return many more points than MODIS for the same fire. Each VIIRS pixel covers about a seventh of the ground a MODIS pixel does, so one MODIS detection can match several VIIRS ones. A higher VIIRS count describes the same fire at finer detail, and it doesn't mean more fire.

Check the acquisition times too, because Terra's morning pass produces MODIS detections that no VIIRS satellite in FIRMS can match. Early growth that shows up only in the MODIS panel is the morning gap NASA is still working to cover after Terra retires.

Detection counts, then, can't be swapped between sensors as if they measured the same thing. Migrating Off MODIS explains how to handle that in a long-term record.