The Amazon is fertilised by a dry lakebed in Chad
Roughly 28 million tonnes of Saharan dust settle on the Amazon basin each year, carrying on the order of 22,000 tonnes of phosphorus — a quantity comparable to what the forest loses to rain and rivers. A large share of it lifts off the Bodélé Depression, the floor of a lake that dried up thousands of years ago.
The number comes from seven years of satellite lidar, which can see the dust layer in vertical cross-section as it crosses the Atlantic. The Bodélé is unusually productive because it is diatomite — the remains of freshwater algae from ancient Lake Mega-Chad — and diatom skeletons are light, friable and rich in phosphorus.
The system keeping the wettest place on Earth productive is a dead lake five thousand kilometres away, on the other side of an ocean. Ecological boundaries are a convenience of maps, not a property of systems.
CALIPSO fires a laser downward and times the backscatter, which gives the vertical structure of the plume rather than a flat image. Multiply the observed dust column by an assumed phosphorus fraction and you get a flux. The lidar is precise; the phosphorus fraction is sampled from a handful of ground campaigns, and it is where most of the uncertainty in the headline number lives.
Year-to-year variation in the transported mass approaches a factor of two, tracking Sahel rainfall. Both greening and further drying of the Sahel would change the delivery, in directions that are not simply opposite. A rainforest's nutrient budget turns out to be coupled to a drought index on another continent.
“The Amazon depends on Saharan dust” is stronger than the evidence.
Closing the phosphorus balance requires estimates of both loss and internal recycling, and some studies find Amazonian recycling considerably more efficient than the simple budget assumes. What is well established is the delivery. What is asserted is the dependency.
Basin-scale measurements showing internal phosphorus recycling large enough to make the atmospheric input marginal.