Query (optional)   in Class  All Allele Author Colleague Collection Assembly Gene Gene Class Gene Product Germplasm Image Journal Keyword Library Locus Map Map Data Marker Pathology Polymorphism Probe Protein QTL Rearrangement Reference Sequence Species Trait Trait Study Two Point Data  

---

GrainGenes Reference Report: JID-96-209

[Submit comment/correction]

Reference	JID-96-209
Title	Response of durum wheat cultivars to Algerian environments - III. GIS-based definition of cultivar recommendations
Journal	Journal of Agriculture and Environment for International Development
Year	2002
Volume	96
Pages	209-226
Author	Annicchiarico P Bellah F Chiari T Delli G
Abstract	The definition of cultivar recommendations should ideally be extended to all sites within a target region. Grain yields of 24 durum wheat cultivars grown for 2 years in 14 Algerian locations were modelled by Additive Main effects and Multiplicative Interaction (AMMI) with one genotype x location (GL) interaction principal component (PC 1) and by factorial regression including winter mean temperature (across January and February) and annual rainfall as site covariates. Both models were interfaced with a Geographic Information System (GIS) to model nominal yield responses of genotypes as a function of site mean values of relevant environmental variables. For AMMI modelling, this required the preliminary definition of a multiple regression equation for predicting the scaled PC 1 score of sites. The equation, selected by a step-wise multiple regression analysis performed on 6 environmental variables, included winter mean temperature and annual rainfall, had both partial regression coefficients significant at Pless than or equal to 0.01, and explained 81% of the variation in site PC 1 scores observed on test years. After spatial modelling of yield responses, areas that showed the same pair of expected top-yielding genotypes were grouped to form one subregion for cultivar recommendation. Both up-scaling procedures allowed for a fine-tuned geographical definition of subregions. The factorial regression-based technique indicated 3 major recommendation domains, i.e. GTA Dur+Simeto for low-elevation, mild-winter locations (29% of the potential durum wheat cropping area), GTA Dur+Chen's for moderately cold- and drought-prone sites (53% of the area), and GTA Dur+Ofanto for severely cold- and drought-stressed sites (13% of the area). The AMMI-based technique implied only the first 2 recommendation domains, and displaced the first subregion somewhat westward relative to up-scaled factorial regression results. The 2 techniques provided same recommendations on about 55% of the area, but the consistency was higher (77%) when considering individual recommended cultivars. Results for test sites suggested that the consistency did not decrease as a consequence of the temporal up-scaling (i.e. the modelling as a function of long-term rather than test year climatic data) but, rather, tended to increase compared with conventionally-modelled data. The generated information can enlarge the scope for site-specific targeting of cultivars by extension services and assist national seed production and distribution systems
Keyword	[ Hide all but 1 of 17 ] crop yield cultivar environment environmental factors genetic variance genotype genotype environment interaction geographical information systems hard wheat multiple regression precision agriculture principal component analysis rain seed production temperature variety trials winter

GrainGenes is a product of the Agricultural Research Service of the US Department of Agriculture.