Validation et optimisation d'une méthode d'indice de risque de perte de phosphore

Authors: Therrien, Karine
Advisor: Gallichand, JacquesParent, Léon-Étienne
Abstract: Surface water quality impairment is primarily caused by phosphorus (P) lost from surrounding agricultural fields. This knowledge triggered the development of P losses predicting tools taking into account transport and source factors. In the United States, a great number of P index (PI) approaches has been developed and adopted. Because these PIs are not necessarily applicable in the province of Québec, a P index specific to Québec has been developed : the Phosphorus Risk Index (PRI). The objective of this research was to validate and possibly improve the PRI method using P losses measured on nine experimental agricultural plots located on the IRDA farm in Saint-Lambert-de-Lauzon. For each plot, P losses and the amounts of water from runoff and tile-drainage were continuously measured during a two-year period (2001-2002) using automated systems. The total P losses were on average 540 g ha-1 from which 95% was exported via the subsurface drainage system. For each plots, the selected P loss value of each of the PRI components were multiplied by the weight assigned to each components and summed to obtain the final PRI value which is associated to one of the five P loss ratings (i.e. very low, low, medium, high and very high). Results indicate that the measured total P losses and PRI values showed a correlation coefficient of 0.63. In order to improve the relationship between P losses and the IRP, the Trust-region algorithm of the SAS none-linear programming (proc. NLP) was used to optimize weights and P loss potential values of the components. Optimization resulted in a correlation coefficient of 0.92. Modifications to the method kept the additive structure of the PRI method. Results were only valid for the Saint-Lambert plots. Generalization across Québec would require experiments on a range of soil, climate, and agronomic conditions. Results from this research indicated possible improvement in the predictive accuracy of the PRI method. Keywords: Phosphorus index, phosphorus losses, optimization, risk assessment, source factors, transport factors.
Document Type: Mémoire de maîtrise
Issue Date: 2007
Open Access Date: 13 April 2018
Permalink: http://hdl.handle.net/20.500.11794/19654
Grantor: Université Laval
Collection:Thèses et mémoires

Files in this item:
SizeFormat 
24468.pdf585.37 kBAdobe PDFView/Open
All documents in CorpusUL are protected by Copyright Act of Canada.