i.hyper.methane computes a methane (CH4) matched-filter
enhancement map from hyperspectral imagery imported as 3D raster maps
(raster_3d) by
i.hyper.import.
The module selects all bands of the input cube whose wavelength falls within a SWIR spectral window (left_shoulder to right_shoulder, default 2120-2400 nm, bracketing the CH4 absorption doublet), then applies a background-covariance matched filter (Thorpe et al. 2013, Thompson et al. 2015) across that window:
alpha(x) = (x - mu)^T C^-1 t / (t^T C^-1 t)
where, for each pixel's radiance vector x over the selected bands:
By default, k is a synthetic Gaussian centered at absorption_wavelength (default 2298 nm, the strongest CH4 feature in this region) with full-width-at-half-maximum template_fwhm (default 40 nm) -- a simplified stand-in for a real gas cross-section curve. For a more physically faithful retrieval, supply absorption_spectrum: a two-column text file (wavelength_nm, absorption) with a real CH4 unit absorption spectrum (e.g. derived from HITRAN, or the "unit spectrum of methane absorption" used in matched-filter literature); the module interpolates it onto the input's own band wavelengths.
alpha is a relative enhancement statistic: values near 0 indicate no detected CH4 signature; larger positive values indicate stronger correlation with the absorption template, consistent with elevated methane along the sensor-to-surface path.
i.hyper.methane is part of the i.hyper module family. It
was built for flagging candidate methane hotspots (e.g. near oil/gas
infrastructure or palm oil mill effluent ponds) in hyperspectral scenes,
as a companion to the coarser-resolution, column-retrieval-based
Sentinel-5P methane mapping done outside GRASS by the ALMA Energy
toolset's 05_methane_hotspots script.
This module computes a matched-filter enhancement statistic, not necessarily a radiometrically calibrated ppm-m column retrieval. Its output is directly comparable in relative terms across a scene (larger alpha = stronger CH4 signature), and is substantially more sensitive and robust to background surface variability than a simple two-shoulder band-depth ratio, because it exploits the full spectral window and the scene's own background covariance rather than three isolated bands. If absorption_spectrum supplies a unit absorption spectrum in properly scaled physical units (e.g. from a full radiative-transfer model, as used operationally by EMIT's own CH4 plume product), alpha approximates a ppm-m enhancement; with the default synthetic Gaussian template it should be treated as a relative ranking statistic only.
Because the retrieval whitens against the scene's own background covariance, it is more susceptible to including plume-affected pixels within the "background" mean/covariance estimate for scenes with very large or pervasive plumes; for such cases consider excluding known plume regions before running the module, or iterating (mask flagged hotspots, recompute).
The module expects the input 3D raster to carry per-band wavelength
metadata, following the same convention as the rest of the i.hyper
family (an i.hyper.import hyper.json sidecar, or
wavelength/unit in each band's r3.support
history). If no band lies within tolerance nanometers of
absorption_wavelength, or if fewer than 5 bands fall within the
left_shoulder-right_shoulder window, the module fails
with G_fatal_error rather than silently proceeding with
insufficient spectral coverage -- this typically means the input's SWIR
coverage does not reach the CH4 absorption region near 2100-2400 nm.
Compute a methane matched-filter enhancement from a hyperspectral cube imported by i.hyper.import, using the default synthetic absorption template:
i.hyper.methane input=emit_scene output=emit_ch4_mf
Using a real unit absorption spectrum and a narrower matched-filter window:
i.hyper.methane input=emit_scene output=emit_ch4_mf \
absorption_spectrum=ch4_unit_absorption.txt \
left_shoulder=2150 right_shoulder=2350
i.hyper.import, i.hyper.continuum, i.hyper.indices, r.mapcalc, r3.support
Created for the i.hyper module family, ALMA Energy project