vignettes/preprocessing_cropland.Rmd
preprocessing_cropland.RmdA key feature of mapspamc is to allocate the crop
statistics to areas identified as cropland on a map - the cropland
extent. As described in the section on Input
data, we use a so-called synergy cropland map to account for the
uncertainty in the location of cropland. The synergy cropland map
combines different cropland products to construct a single cropland map
that presents the mean and maximum available crop area per grid cell as
well as a ranking that measures the agreement between the various
cropland products for each grid cell. A ranking of 1 means that all
sources agree there is cropland in a grid cell, while a higher rank
signals disagreement.
The package contains a script to create a synergy cropland map using
global land use products that are included in the maspamc database:
COPERNICUS, ESACCI, ESRI, GLAD and MODIS (see the database documentation
for details). The script is located in the
02_3_pre_processing_cropland folder. Before the synergy
cropland map can be created the user needs to clip target country from
the global datasets and ensure the maps have the right format
(i.e. desired extent, resolution and coordinate reference system).
Scripts to do this for each global land use product are provided in the
same folder.
For the Malawi example, we can make use of an existing global synergy
cropland map (SASAM) that was prepared for SPAM2010, which can be
selected by running select_sasam.R. For the Ethiopia
country example we used information from four datasets that represent
global cropland around the period 2015: GLAD, COPERNICUS, ESACCI and
MODIS. The user needs to run select_*.R first to process
the global maps. The script create_synergy_cropland_map.R
combines the selected land use maps and creates a country synergy
cropland map.
In order to create the synergy map, a ranking table is needed that
indicates, which (combination) of the inputs maps is regarded as the
most reliable. In case all the global cropland maps signal a grid cell
contains cropland the ranking is 1 (highest and most reliable). For all
other combinations the user needs to prepare a ranking table. The table
for Ethiopia is added to the mapspamc
database(synergy_cropland_table/synergy_cropland_table_2015.xlsx)
and contains a ranking from 1 to 16 (see below) because it is based on
four maps. For Ethiopia, we decided on the following order of increasing
importance when creating the scoring table: (1) GLAD, (2) COPERNICUS,
(3) ESACCI and (4) MODIS. GLAD receives the highest score because it has
the highest original resolution (30 meter) and is therefore assumed to
be the most accurate, followed by COPERNICUS (100 meter), ESACCI (300
meter) and MODIS (500 meter).
| agreement | rank | glad | copernicus | esacci | modis |
|---|---|---|---|---|---|
| 4 | 1 | 1 | 1 | 1 | 1 |
| 3 | 2 | 1 | 1 | 1 | 0 |
| 3 | 3 | 1 | 1 | 0 | 1 |
| 3 | 4 | 1 | 0 | 1 | 1 |
| 2 | 5 | 0 | 1 | 1 | 1 |
| 2 | 6 | 1 | 1 | 0 | 0 |
| 2 | 7 | 1 | 0 | 1 | 0 |
| 2 | 8 | 1 | 0 | 0 | 1 |
| 2 | 9 | 0 | 1 | 1 | 0 |
| 2 | 10 | 0 | 1 | 0 | 1 |
| 2 | 11 | 0 | 0 | 1 | 1 |
| 1 | 12 | 1 | 0 | 0 | 0 |
| 1 | 13 | 0 | 1 | 0 | 0 |
| 1 | 14 | 0 | 0 | 1 | 0 |
| 1 | 15 | 0 | 0 | 0 | 1 |
| 0 | 16 | 0 | 0 | 0 | 0 |
If the user is interested in creating a more recent synergy cropland map (e.g. for around 2020), we recommend using the following combination of cropland products, which are available in the mapspamc database: ESRI (2020), GLAD (2019), COPERNICUS (2019) and ESACCI (2020), and modify the 2015 ranking table accordingly. Alternatively, and if available, the user could add a national land use product and create a ranking with five different inputs that goes up to 32.