This note defines a national equal-area projected coordinate reference system (CRS) for Mozambique, based on the Lambert Azimuthal Equal Area projection (EPSG method 9820) centred on the country's territorial midpoint at 18°30′S, 35°30′E.
Mozambique spans two UTM zones (36S and 37S), so national raster products carry a zone seam, and conformal projections do not preserve area, biasing zonal statistics. A single national equal-area CRS removes the seam and makes pixel area exact everywhere.
All projections considered here preserve area exactly, so the comparison is one of shape (angular) distortion. Over the area of use the proposed CRS has a median angular deformation of 0.16° and a maximum of 0.59°, and 0.37° over the onshore territory. The common continental fallback for the region, Africa Albers (ESRI:102022), reaches about 7° over northern Mozambique because its standard parallels are placed for the whole African continent; area is preserved there too, so area statistics computed in it are not biased.
Benchmarked against the equal-area CRSs registered by other national agencies (European Environment Agency, USGS, Geoscience Australia), the proposed CRS sits comfortably inside the distortion envelope those bodies accepted for their own official CRSs.
The CRS is in operational use for full-scale national multitemporal mosaics of Sentinel-1, Sentinel-2 and Landsat imagery and derived hydrological products of the Mozambique hydrogeological mapping programme, led by the Direcção Nacional de Gestão de Recursos Hídricos (DNGRH) and financed by the World Bank. LNEG executes the remote-sensing component under consulting services contract DNGRH-346682-CS-QCBS.
This document is the authoritative defining source for the CRS and supports its registration in the IOGP EPSG Geodetic Parameter Dataset.
Reproducibility: every number and figure in the note is produced by the computation scripts included in this deposit; the note quotes their output. The scripts read the national boundary, which is verified by SHA-256 before use, so the statistics cannot change silently with an upstream dataset revision. Angular deformation is computed two independent ways, from the PROJ Tissot factors and from the singular values of the local Jacobian built from central geodesic differences, and the two are asserted to agree. The results were reproduced independently on Linux with Python 3.12 and on Windows with Python 3.13, using pyproj 3.7.2 with PROJ 9.5.1.