Modelling phosphorus inputs and dynamics in a large tropical semiarid basin.

Abstract

This study proposed a simple process-based model to predict total phosphorus (TP) inputs and dynamics in a large tropical semiarid basin in Brazil (19,015 km²), where the Castanhão dam, the largest non-hydropower reservoir in Latin America, is located. The model solved on a monthly basis the water balance and TP dynamics along the river system, considering complete mixing at the cross-section, steady-state regime, and a piston-type transport with first-order decay. Both primary and secondary data were used to validate the model, with deviations of up to about 20%. A linear increase in the flow rate along the river was observed, indicating that transmission gains were dominant. TP concentration increased with river flow rate and a consistent decay along the system was observed. TP from point sources (PS) was assumed constant. TP from non-point sources (NPS) decreased with time (R² = 0.90) due to the continuous wash-off process. NPS was the dominant load, decreasing respectively from 99 to 61% of the total load from high to low flows. TP load varied from 1 to 141 ton/month, with an average of 34 ton/month. This resulted in an average TP yield of 21.2 kg km^-2 year^-1, which is significantly lower than the national average due to the peculiarities of the region: lower precipitation, intermittent rivers, and high-density reservoir network. All the simulated TP values were above the mandatory limit of 0.05 mg/L. PS reduction was only effective for relatively low flows, while NPS reduction was the most relevant remediation measure for both high and low flows. The impact of climate change on TP concentration also yielded TP > 0.05 mg/L for all the projected scenarios, with the input loads to the Castanhão reservoir spanning from 23 to 266 ton/month. The proposed model can also be applied to other regions with different hydroclimatic conditions and land uses.