AN ECONOMIC-EVALUATION OF ALTERNATIVE GENETIC-IMPROVEMENT STRATEGIES FOR FARM WOODLAND TREES

A cost-benefit analysis (CBA), focusing on the Net Present Value (NPV) of a current genetic improvement programme for broadleaved trees was performed using Monte Carlo simulation, with an add-on software packag e ('@RISK') specifically designed to take account of the uncertainty a ssociated with long-term projects. The CBA was undertaken by evaluatin g the total cost of achieving a given estimated genetic gain via each of the breeding strategies considered. The estimated values of genetic gain were then expressed in terms of the increased value of timber ou tput. Cash flows were based on current estimated tree establishment co sts and anticipated productivity of the four tree species included in the programme (ash, Fraxinus excelsior; sycamore, Acer pseudoplatanus; wild cherry, Prunus avium; and sweet chestnut, Castanea sativa), when grown primarily for a timber crop. The results of the NPV analysis in dicated that tree improvement could be cost-effective for small geneti c gains, but that current breeding strategies differed markedly in the ir cost-effectiveness. Improvement scenarios based on conventional sel ection and testing techniques, such as simple mass selection and recur rent selection (seed orchards), were found to be the most cost-effecti ve at a discount rate of 6 per cent. In contrast, tree improvement sce narios based on clonal techniques consistently ranked lowest, despite the much higher genetic gains achieved. The use of clonal techniques w as found to be particularly hard to justify with broadleaved tree spec ies of relatively low timber value. Overall, with the current state of broadleaved timber markets in the UK, and the current areas being pla nted, investment in basic genetic improvement of high-value timber spe cies appears financially worth while. The estimated direct additional financial benefit to growers, if new planting is undertaken with impro ved stock as opposed to unimproved stock, is estimated To range from p ound 38 ha(-1) with Simple Mass Selection to pound 100 ha(-1) with Sim ple Recurrent Selection.