01 Load GAMMA results

Pipeline: examples/01_load.toml. Tutorial: nbs/Tutorials/CLI/CampiFlegrei/01_load.ipynb.

Input

A GAMMA processing directory with co-registered SLCs and geocoding products:

gamma/
├── rslc/       YYYYMMDD.rslc, YYYYMMDD.rslc.par   (co-registered SLCs, one per date)
├── geocoding/  <geo>.hgt, <geo>.diff_par, <geo>.lt_fine, <geo>.lv_theta, <geo>.lv_phi
└── DEM/        dem_seg_par

Ask the user for the directory, the reference date (reference, the date the rslc stack is flattened to) and the date prefix of the geocoding files (geo) if they are not obvious from the file names. The defaults of the file are the ones of the sample data set, Campi Flegrei (data/CampiFlegrei/gamma, decision 0033).

moraine run examples/01_load.toml --workdir WORK --var gamma=/path/to/gamma --var reference=20200707 --var geo=20200707

GAMMA must be on PATH (which base_calc phase_sim_orb geocode create_offset).

Steps and outputs (WORK/raw/)

step	command	output
rslc	load-gamma-flatten-rslc	rslc.zarr (nlines, width, nimages) complex64, flat earth and topography removed
lat_lon_hgt	load-gamma-lat-lon-hgt	lat.zarr, lon.zarr float64, hgt.zarr float32, (nlines, width); nan where 0 in GAMMA
look_vector	load-gamma-look-vector	theta.zarr, phi.zarr (elevation / orientation angle)
range	load-gamma-range	range.zarr slant range distance
metadata	load-gamma-metadata	meta.toml (dates, wavelength, heading, pixel spacing, perpendicular baselines)
web_mercator	transform	e.zarr, n.zarr web mercator coordinates (EPSG:3857) for map plots
rslc_pyramid	ras-pyramid	rslc_pyramid/ to check the interferograms, decimated: SLC pixels cannot be averaged (their scatterer phases differ), so the whole scene is speckle; zoom in, or look at the mean pyramid of the n2f interferograms of 02

Sample data (Campi Flegrei, 981 x 4160 x 92, 2026-10-08): 162 s, 126 s of it phase_sim_orb for each date (64 threads); the pyramid takes 21 s.

Expected results (sample data)

result	sample value	sane range
rslc.zarr	complex64 (981, 4160, 92), chunks (1000, 1000, 1)	the number of dates in rslc/
rslc_pyramid amplitude	p01 0.023, p50 0.29, p99 1.56	p50 roughly 0.1-1
rslc_pyramid nan_fraction	0.40 (the Gulf of Pozzuoli, zero in the coregistered SLCs)	small, unless water or the footprint is masked
meta.toml dates	92, 12 days apart	all dates of rslc/, in order

Figures

Quicklooks of the sample data run of 2026-10-08:

Sequential interferogram 0 of the raw stack (quicklook raw/rslc_pyramid --show intf_seq --index 0): noisy single look phase with structure on land, nan on the gulf.

Checks

• moraine info raw/rslc.zarr: shape (nlines, width, number of dates in rslc/).
• moraine info raw/rslc_pyramid: amplitude p50 around 0.1-1. A nan_fraction near 1 means the GAMMA files did not match; on the sample data 0.40 is the sea.
• moraine quicklook raw/rslc_pyramid --show intf_seq --index I -o intf.png: sequential interferograms. Expect noisy phase with visible fringes / structure in coherent areas; pure uniform noise everywhere suggests a wrong reference date or unflattened data.
• raw/meta.toml: dates must list all dates in order.

Problems

• sh: base_calc: command not found: GAMMA is not on PATH.
• A step fails on a missing *.par / geocoding file: check geo / reference.
• The scratch directory (raw/scratch) keeps GAMMA intermediates; it can be deleted after the run.