Tropentag, September 17-19, 2013, Stuttgart-Hohenheim
“Agricultural development within the rural-urban continuum”
Virtual Crop Modelling for Technology Impact Assessment -
Lessons from a Potato Crop Growth Model
Ulrich Kleinwechter1, Senthold Asseng2, Manuel Gastelo3, Joe Ritchie2, Gerald C.
1International Potato Center (CIP), Social and Health Sciences Global Program, Peru
2University of Florida, Agricultural & Biological Engineering Department, United States
3International Potato Center (CIP), Genetics and Crop Improvement Global Program, Peru
4International Food Policy Research Institute (IFPRI), Environment and Production Technology Division,
Eﬀorts are under way in the agricultural research for development community to employ
crop growth simulation models for the ex-ante assessment of the impacts of agricultural
technologies, in particular improved crop varieties. The approach, dubbed “virtual crop
modelling”, consists in the manipulation of genetic coeﬃcients of the cultivars contained
in crop growth models to simulate productivity eﬀects of genetic improvements brought
about by future breeding eﬀorts and assesses the impacts of these improvements under
ﬁeld conditions. However, the suitability of the models for this purpose is not yet proven.
In an eﬀort to explore the suitability of crop growth models for virtual crop modelling,
the DSSAT-SUBSTOR potato growth model, a model which simulates potato growth as
a function of abiotic, but not of biotic factors, is used to carry out a sensitivity analysis
of changes in genetic coeﬃcients of a modern tropical potato cultivar. Yields are simula-
ted across a wide range of genetic coeﬃcients of the crop model at sites located in four
contrasting tropical, subtropical and temperate environments.
Simulation results show that the model responds to parameter changes as expected and
that it gives meaningful results in all tested environments. Site characteristics appear to
play an important role for potential simulated yield levels and for the responses of the model
to changes in genetic coeﬃcients. However, there is substantial unexploited potential for
productivity growth from improved management practices. This potential appears to be
greater than the beneﬁts from genetic improvements that appear likely today.
The analysis shows that, by helping to identify yield constraints for diﬀerent environ-
ments, crop modelling can make substantial contributions to ex-ante impact assessment of
agricultural technologies. Virtual crop modelling, however, is only a viable option if crop
growth is actually constrained by those genetic factors which are represented by adjustable
genetic coeﬃcients of a given crop model. Furthermore, the scope of many crop models
is limited to abiotic yield-limiting factors and an extension towards biotic factors, e.g. a
linkage with disease models, might be necessary for some environments.
Keywords: Crop improvement, crop modelling, DSSAT, ex-ante impact assessment, potatoes, SUB-
Contact Address: Ulrich Kleinwechter, International Potato Center (CIP), Social and Health Sciences Global Pro-
gram, Apartado 1558, 12 Lima, Peru, e-mail: email@example.com