Calibration par intelligence artificielle de la détermination des besoins en chaux des sols acides
|Authors:||Dupre, Lenz Clove Richardly|
|Advisor:||Khiari, Lotfi; Gallichand, Jacques|
|Abstract:||In Quebec, in all Canadian provinces and in 18 American states, the method for determining the lime requirements of agricultural soils is based on pHₛₘₚ analysis. However, this method poses accuracy problems for soils with low lime requirements. Thus, this method is not in phase with the agronomic reality of a majority of Quebec soils with low needs and maintained each year by liming practices. The objective of this study was to develop a new procedure for evaluating lime requirements based on routine soil analyses (pMᴡₐₜₑᵣ, MO, Pₘ₃, Kₘ₃, Caₘ₃, Mgₘ₃, Alₘ₃)) and their calibration by supervised learning on soil acidity titration curves. To carry out this study, 270 titration curves were elaborated from laboratory experiments, then these curves were parameterized and the covariates resulting from this parameterization are predicted by artificial intelligence from basic physicochemical characteristics of the analyzed soils. This research project was carried out in such a way as to take into account the reality of agricultural soils in Quebec as much as possible. The sampling was done in 9 regions of Quebec. The results obtained prove that it is possible to parameterize a soil titration curve and to predict its lime needs from its routine physico-chemical parameters.|
|Document Type:||Mémoire de maîtrise|
|Open Access Date:||20 September 2021|
|Collection:||Thèses et mémoires|
All documents in CorpusUL are protected by Copyright Act of Canada.