Detection of historical changes in pasture growth and attribution to climate change

Few studies consider historical trends in biological systems in relation to changes in climate. These detection and attribution studies are particularly challenging in agricultural systems where other factors (such as management) are changing over time. Here we consider changes in pasture yield (net herbage accumulation [NHA]) over the period 1960-2004 in a dataset from a trial in New Zealand where management (grazing protocol and fertiliser application) was constant over time. We used 2 approaches: a statistical approach looking for trends in, and correlations between, NHA and climate variables and a process-based modelling approach where combinations of variables were held constant at their starting values or allowed to change with time enabling us to isolate the impact of individual factors. There was a significant positive trend for NHA in spring over the period and positive trends in rainfall and atmospheric CO2 concentration; soil nitrogen (N) also increased over time. The statistical approach was useful for identifying trends but was unable to resolve the driving variables. Modelling identified CO2, soil properties and their interaction as the most influential variables. The calculated impact of CO2 was a 0.21% increase in NHA ppm CO2-1; this compares to a value of 0.19% from a FACE (free air carbon dioxide enrichment) experiment with a similar type of management and pattern of pasture production. The results instill confidence in experimental estimates of the CO2 fertilisation effect, particularly at low levels of CO2 enrichment, and provide evidence that climate change impacts are already in progress.