Productivity of deciduous woody and fodder species in relation to air temperature and precipitation in a Mediterranean environment

Productivity and its relation to air temperature and precipitation of 11 deciduous woody fodder species was studied in a semi-arid Mediterranean environment with cold winters in Macedonia, Greece. The species studied included six shrubs: Amorpha fruticosa L., Carpinus orientalis Mill, Colutea arborescens L., Corylus avellana L., Fraxinus ornus L. and Ostrya carpinifolia Scop.; and four trees: Pirus amygdaliformis Vill., Quercus pubescens Wild., Quercus sessiliflora Salish. and Robinia pseudoacacia L., with two accessions, common and spineless. Fodder species were established as one-year-old seedlings in 1987 and studied through 1994. Productivity was measured by clearcutting of the aboveground annual biomass (DM) at the end of each growing season. Samples of biomass were hand separated to determine the grazeable material (leaves and twigs). In addition, monthly values of air temperature and precipitation were correlated with annual biomass through a multiple linear regression analysis. It was found that annual biomass was statistically different among species in all the study years with Robinia being the most productive of all the species tested followed by the other two legumes Colutea and Amorpha. Proportion of grazeable material ranged widely among species and years with an overall mean of 50%. Robinia produced grazeable biomass more than 1000 kg/ha followed by Amorpha and Colutea with about 800 and 700 kg DM/ha, respectively, while all the other species produced less than 500 kg DM/ha. Yearly variation in annual biomass was found to be affected much more by air temperature, especially during the beginning of the growing season (March), than by precipitation. Coefficients of determination (r ^sup 2^) were found to be high (0.50-0.80) in Amorpha, Pirus, Quercus and Robinia but only in the latter two species they were high enough (>0.70) so that the relations produced can be used for predictive purposes.[PUBLICATION ABSTRACT]