In between plant and animal breeding 1: economic weights


Unless you live in paradise and you have a single objective trait—so your whole breeding objective is “more X”—you have to take into account trade-offs between traits. On the genetics front we deal with this via genetic correlations, but on the value front we have to figure out how much is an extra unit of X worth compared to an extra unit of Y (at least in relative terms).

You might be thinking “I don’t use no stinking economic weights, mate”. Instead you could used desired gains, independent culling levels, etc. You are not using explicit economic weights, but there are undeclared, implicit weights in your selections.

There are a few things that make tree breeding economic weights a bit different from the ones used in crop and animal breeding:

  1. Time or, better, TIME! We are selecting trees to go to the breeding programme, which later will be deployed and go for a full rotation (7 years for pulp in the tropics or 15 in temperate environments, 25~40 years for solid wood). This create huge uncertainty for the value of traits in 15~50 years in the future. We did some work on dealing with economic uncertainty here:

Evison, D.C. and Apiolaza, L.A. 2015. Incorporating economic weights into Radiata pine breeding selection decisions. Canadian Journal of Forest Research 45(1): 135-140 (PDF).

  1. The relationship between wood properties and end products is harder to quantify, particularly if dealing with solid wood products. If you come from the animal world, imagine the difference between breeding for milk (that can be homogenized, a bit like pulp) and for meat (where mixing filet with rump is a major loss of value, a bit like solid timber).

, ,