DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación
Artículo científico (JournalArticle)
Adapting the CROPGRO model to simulate alfalfa growth and yield
14 Pags.- 6 Figs.- 6 Tabls. Available freely online through the author-supported open access option.
Copyright © 2018 by the American Society of Agronomy. This is an open access article distributed under the CC BY-NC-ND license., Despite alfalfa’s global importance, there is a dearth of crop simulation models available for predicting alfalfa growth and yield with its associated composition. The objectives of this research were to adapt the CSM-CROPGRO Perennial Forage Model for simulating alfalfa growth and yield and to describe model adaptation for this species. Data from six experimental plots grown under sprinkler irrigation in the Ebro valley (Northeast Spain) were used for model adaptation. Starting with parameters for Bracharia brizantha, the model adaptation was based on values and relationships reported from the literature for cardinal temperatures and dry matter partitioning. A Bayesian optimizer was used to optimize temperature effects on photosynthesis and daylength effects on partitioning and an inverse modeling technique was employed for nitrogen fixation rate and nodule growth. The calibration of alfalfa tissue composition was initiated from soybean composition analogy but was improved with values from alfalfa literature. There was considerable iteration in optimizing parameters for the processes outlined above where comparisons were made to measured data. After adaptation, the Root Mean Square Error and d-statistic of harvested herbage averaged across 58 harvests (yield range: 990–4617 kg ha–1) were 760 kg ha–1 and 0.75, respectively. In addition, good agreement was observed for Leaf Area Index (LAI) (LAI range: 0.1–6.7) with d-statistic of 0.71. Simulated belowground mass was within the range of literature values. The results of this study showed that CROPGRO-PFM-Alfalfa can be used to simulate alfalfa growth and development. Further testing with more extensive datasets is needed to improve model robustness., This work was funded by the Ministry of Economy and
Competitiveness of the Spanish Government through the research
grants (AGL2013-48728-C2-2-R). We thank this Ministry for
awarding Wafa Malik a predoctoral fellowship and financial support
for research abroad internship at the University of Florida., Peer reviewed