Experimental design

The Smart Grassland Systems Project aims to address one of the greatest challenges facing Irish farmers today – how to produce high quality/high yielding grassland forage in an environmentally sustainable way. In UCD’s Research Farm a large field experiment was established to explore these pressing issues.

Eight pasture mixtures were selected using a constrained simplex-centroid design with three functional groups; grasses, legumes and herbs. There was a constraint imposed on the design so that there was at least 40% grass in each mix, therefore no more than 60% legume or herb (see diagram, after Cornell 2002). This was repeated at three levels of species richness, with 1, 2 or 3 species per functional group (24 mixtures). The design was replicated at four different nitrogen (N) rates (0, 45, 90, 135 kg N/hectare/year).


By using this experimental design we can pin-point multi-species mixtures that can produce high-yielding, quality forage and support sustainable agriculture, such as requiring less fertiliser, supporting farmland diversity or reducing weed invasion.
This design also allows us to explore some key ecological theories too, such as the effect of species versus functional group redundancy on ecosystem services. These ecosystem services could be, for example, primary production, biodiversity support value, invasion resistance.
References
Cornell, J.A. (2002) Experiments with Mixtures: Designs, Models, and the Analysis of Mixture Data, 3rd edition. Wiley, Chichester.