Dear Chan,

Before generating the report, I would transform the raster from WGS-84 to a projection that has a metric unit (km or m). If the study area is not too large,bilinear or nearest neighbor transformation may be appropriate.

HTH,

Ákos

Ákos Bede-Fazekas , thank you for your reply. I am doing it at a global scale, do you have any idea which transformation method may be more suitable for continuous data?

You can try to use World Mercator or Pseudo-Mercator. In any case, others software can generate "square-area" in Geographical Projection

In case of continuous data, you can use bilinear interpolation. But note that in case of global coverage, transforming the raster from WGS-84 to another (e.g. Mercator) projection may increase the uncertainty of your data due to the interpolation. In this case a more precise way is using other tools that can calculate the metric area of raster cells or polygons created from raster cells.

All the steps (reclassification, transformation, area calculation) can be done in R using packages raster, sp, sf and rgeos.

HTH,

Ákos

If area is the critical property you want to preserve, in particular when working on near global scale (e.g. tropics and subtropics), I strongly would suggest an equal-area CRS, such as epsg6933 or similar projections. It simplifies the downstream process as you can measure the area by counting the cells and avoids latitudinal bias during calibration.

Rainer

You can calculate using Arc-GIS spatial analyst tool. But before that region-specific projection transformation required.

Also R packages raster sf, rgeos are useful.

https://r-spatial.github.io/sf/