Extension and Evaluation of Intercropping Field Trials Using Spatial Models

Heike Knoerzer,Bettina U. Mueller,Buqing Guo,Simone Graeff-Hoenninger,Hans-Peter Piepho,Pu Wang,Wilhelm Claupein

DOI: https://doi.org/10.2134/agronj2009.0404

IF: 2.65

2010-01-01

Agronomy Journal

Abstract:Intercropping has often been considered as a secluded cropping system within one field. However, in African and Asian countries, where intercropping is widespread, the system can be looked on at a much larger scale: small fields alternate as strips with different crops grown on them, turning the collection of fields into a kind of unplanned intercropping. The more fragmented the agricultural landscape, the more relevant the borders can become. Traditionally, statistical analysis of intercropping experiments has been done by a simple analysis of variance without taking spatial models into account. But especially strip intercropping experimental arrangements lack in randomization as the cropping system imposes alternating strips. Thus, spatial variability and its effect on yield have to be regarded differently. Two different features of intercropping experiments were studied in the present paper: statistical peculiarities of intercropping designs and the border effect which is a key component of intercropping performance. Field trial results from Germany and China indicated that intercropping showed significant border row effects within the first four rows. For statistical analysis, different spatial models were added to the baseline model to account for the spatial trend and to check whether or not standard models are suitable for analyzing intercropping experiments. The results showed that for the German experiment the baseline model fitted well in the year 2008 and a common analysis of variance seemed to be well suited. However, for the Chinese experiments and the German experiment in the year 2009 the spatial models improved the model fit.