Soil properties that influence grassland production: their identification, production-indicator potentials and explanatory mechanisms

Soil properties that drive grassland production in the absence of mineral nitrogen applications and explanatory mechanisms thereof were sought in this study. Sixteen grassland farms varying in nitrogen inputs were selected in Counties Limerick and Clare, Ireland. Soil was sampled in 1995 and physical, chemical and biological properties analysed in the laboratory. Management data describing lime and fertiliser applications, grazing practices, soil disturbance, livestock and fodder production were obtained by interview. Grassland production for the year 1996/97 was measured as: stocking density; gross margin; and total energy required for livestock production, including maintenance, growth, offspring and milk and fodder sales, taking nongrassland feed energy into account. The botanical composition was determined in 1998. Data were not normally distributed therefore the research was conducted in an exploratory manner. Descriptive scatter graphs and statistics, bivariate nonparametric Spearman's rank correlation, multiple regression analyses, principle components analysis (PCA) and canonical correspondence analysis (CCA) were implemented. Four farms were in conversion to organic management, seven were organically managed and one was conventionally managed. Four farms had not had fertiliser applications. Microbial biomass, bacterial popUlations, and arginine ammonification activity that was interpreted as N-mineralisation and N-immobilisation activities in vitro, were not extreme in productive farms. Production measured as grassland energy correlated (P < 0.05) with: fertiliser inputs and mechanical disturbance; soil pH; depth to bedrock; soil plant-available magnesium and calcium contents, influences of which were direct. Root mass and N-mineralisation in vitro correlated inversely (P < 0.05) with production. Influential factors were highly correlated with each other (P < 0.05). Each influential factor, ordinated with variables correlated to it, gave rise to a production-related gradient of farms in PCA. A suite of N -mineralisation in vitro and its related variables reliably indicated grassland farm production in ordination space. Fertiliser and lime applications, vegetation, grazing, soil depth and metabiosis influenced bacterial populations and activities. Root mass and N-mineralisation rates in vitro were least in disturbed and mineral nitrogen fertilised soils. N-mineralisation in vitro was greatest in extensive farms and N-immobilisation in vitro indicated carbon abundance in botanically diverse swards. Soil microbial biomass, bacterial populations and activities behaved individualistically along a possible succession in grazed grassland. Based on observations, and on allosteric and catabolite repression of amino acid use, it was hypothesised that 'available carbon, AC, is used before organic nitrogen, ON' is ammonified. However, ammonification in vitro was uncorrelated with soil organic carbon or soil nitrogen contents or C:N ratio, because soil C and N analyses applied in the study may not have accounted for microbe-available resources. Soil organic carbon content was relatively high, but least in the conventional farm. In contrast to other studies in the literature, soil organic carbon content was not correlated with microbial biomass and arginine ammonification but may be above threshold values needed to facilitate biological function in the soils examined. Available forms of C and N ought to be considered as interdependent determinants of soil organic matter turnover rates subject to fertiliser use and grazing in grassland soils.