Premium
Cultivar coefficient stability and effects on yield projections in the SPUDSIM model
Author(s) -
Fleisher David H.,
Haynes Kathleen G.,
Timlin Dennis J.
Publication year - 2020
Publication title -
agronomy journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.752
H-Index - 131
eISSN - 1435-0645
pISSN - 0002-1962
DOI - 10.1002/agj2.20070
Subject(s) - cultivar , growing season , canopy , calibration , yield (engineering) , agronomy , environmental science , mathematics , gene–environment interaction , crop , coefficient of determination , atmospheric sciences , horticulture , biology , botany , statistics , biochemistry , materials science , metallurgy , geology , gene , genotype
Crop model calibration refers to the process where values are obtained for a subset of model parameters that represent cultivar traits. Calibrated parameters, however, are just as likely to compensate for limitations in model structure as well as reflect true phenotypic characteristics. This confounding of genetics with production environment limits model accuracy particularly for climate assessments and can result in crop parameter values that are location dependent. We evaluated the calibration stability of the potato ( Solanum tuberosum L.) model SPUDSIM for two cultivars grown in two sites in the United States sites using a location specific (R1) and cross‐location (R2) methodology. Differences between R1 coefficient values ranged from 4 to 17% between the two sites depending on cultivar. This likely reflected over‐coupling of canopy expansion rate within the model structure as well as an over‐sensitivity to temperature and photoperiod. Broader site differences were observed for growth coefficients, particularly when compared with the R2 method. Validation year end‐of‐season yields (17% error or less) and in‐season RMSE (<36.1 g plant −1 ) were similar for both methods. The R2 calibration overpredicted yields by 5% or less as compared with R1, with a maximum discrepancy of 0.35 Mg ha −1 in climate forecasted impacts. Consistent differences between R1 and R2 calibrated models in rates of yield decline per decade were not observed for cultivar, location, or climate scenario, and averaged 0.034 Mg ha −1 yr −1 for R1 and 0.038 for R2. Thus, calibration for these cultivars and locations was relatively stable in SPUDSIM for the temperate climate evaluated.