Published July 26, 2022 | Version v1.0
Dataset Open

A map of active cropland and short-term fallows across Northern Mozambique derived from PlanetScope data

  • 1. UCLouvain

Description

Overview

A map of smallholder-dominated landscapes covering the provinces Niassa, Zambezia, Cabo Delgado, and Nampula in Northern Mozambique. The map includes active cropland and short-term fallows as separate classes, as well as five land cover classes (herbaceous vegetation, open woodlands, closed woodlands, non-vegetated land, water). The map is based on PlanetScope mosaics and consequently comes at 4.77m spatial resolution.

The download contains the following files:

  • ps_lc_nmoz.tif / .qml: land cover map and associated QGIS style file
  • ps_lc_nmoz_probmargins.tif / .qml: probability margins and associated QGIS style file
  • training.gpkg: training samples with class labels
  • LICENSE.pdf: NICFI data program user license

Map accuracy

We conducted an area-adjusted accuracy assessment based on a stratified random sample, which yielded important insights regarding accuracies and error types. The area-adjusted overall accuracy of the map is 88.9%, but users should be aware of the most important error types:

  • Active cropland were overestimated, whereas local topographical depressions with moist soils, and regions with exposed soils/rocks and sparse vegetation cover were found to be falsely classified.
  • Short-term fallows were underestimated, particularly in regions with high growth rates and extensive land management, such as parts of the northern and north-eastern study region.

Further resources

The production of this map was made possible through the NICFI data program, providing the PlanetScope mosaics and the Google Earth Engine cloud computing platform for preprocessing of the satellite data and classification. As such, the use of the map falls under the NICFI data program license agreement included in the download. The code for preprocessing the PlanetScope mosaics is based on the Google Earth Engine Python API and made available at https://github.com/philipperufin/eepypr/.

We advise map users to read the preprint or the open access paper for detailed insights. In case of questions please consult these resources or contact the lead author of the work.

Notes

This work was supported by the FRS-FNRS, grant no. T.0154.21 and the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program (Grant agreement No 677140 MIDLAND). This research contributes to the Global Land Program. The authors would like to thank Yara Ubisse, Sá Nogueira Lisboa, and Dr. Almeida Sitoe for their invaluable help in preparing and conducting the fieldwork in Northern Mozambique in 2021.

Files

LICENCE.pdf

Files (18.0 GB)

Name Size Download all
md5:cfa14379f3514b955bcbf6fee5aa3048
213.4 kB Preview Download
md5:bfbcb9b653986e746ec2058a4f8c64a4
2.7 kB Download
md5:40ff3b18b75248906dc287fa5aca3fdd
2.2 GB Preview Download
md5:f72560dd7c460f305b046d44167b1b75
35.1 kB Download
md5:43e4a066b8a39ea8fbfab0702c1cc69b
15.8 GB Preview Download
md5:5298cfd121c2ffedb6dd07c21d9d8095
1.7 MB Download