Crop agriculture is expanding and intensifying across large areas of the tropics. In the Brazilian Amazon, pastures have replaced forests but now single-cropped soybeans have replaced pasture and double-cropped soybeans and corn have replaced single- cropped soybeans over millions of hectares. These transitions dramatically alter nutrient inputs to watersheds, but we know relatively little about how intensifying agriculture and increased fertilizer use influence surface water N and P concentrations and watershed export, nor whether responses in tropical cropland will resemble those in better-studied temperate croplands. Amazonian pastures generally receive little to no fertilizer, whereas soybean croplands receive about 50 kg P/ha but little N. Soybean-corn croplands receive about 10 kg/ha additional P and as much as 100 kg/ha N. We instrumented nine headwater watersheds (3 forest, 3 soybean single-cropped, 3 soy-corn double-cropped) in Mato Grosso, Brazil to measure water discharge, concentrations of dissolved N and P, and annual watershed exports of dissolved N and P. We also collected groundwater from wells located at the upper edge of the riparian slope and at the edges of crop fields in each watershed to determine dissolved N and P concentrations in groundwater arriving from uplands to riparian zones.
Results/Conclusions:
Mean water runoff was 3x greater from soybean and soybean-corn watersheds compared with forest watersheds. More than 90% of runoff in all watersheds occurred as base flow and <10% during storms. Streamwater nitrate concentrations were 1-6 uM and did not differ with land use. Ammonium and phosphate concentrations were low and also did not differ with land use. Annual nitrate export increased from 0.1 kgN/ha in forest to 0.3-0.4 kgN/ha in soybean and soybean-corn cropland, a difference caused solely by greater water runoff. Addition of N fertilizers in soybean-corn cropping did not increase dissolved N export. Phosphate export was <0.1 kg/ha in all watersheds. Extractable N and P in soil profiles indicated phosphate was removed by fixation in surface soils, whereas nitrate was removed by increased soil storage at 2-8 m. Dissolved N and P concentrations in streamwater and groundwater confirmed that soil storage and not riparian zone processes accounted for most nutrient removal. Watershed export was lower than in most intensive temperate croplands with similar fertilizer application rates. Future watershed N and P export from fertilized tropical cropland will likely depend on capacity of soil N and P storage mechanisms and cropland management that maintains vertical, groundwater-dominated hydrologic flowpaths.