Current climate variability of pastoral yield: a case study in Canterbury, New Zealand

Increased variability in the timing and yield of pasture growth is an often discussed but rarely quantified aspect of on-farm climate change impacts and adaptation. Here we assess the impacts of current climate change (indicatively the 2015-2025 period) on pasture growth and annual yield to provide insights into the significance of increased variability. We used a storyline and stochastic weather generator approach in a localised case study of two sites in Canterbury, New Zealand. The method uses detrended climate and river flow records in conjunction with event transition probabilities derived from Weather@home climate simulations to generate a suite of stochastic weather and flow realisations for application to an irrigation and pasture growth model. Our results suggest that climate change is already having an impact, with a 17.2% reduction in median pasture yield for dry-land farms and a 3.7%-5.5% increase for irrigated farms. These findings are in line with previous studies in New Zealand, which suggest a mean change in pasture yield due to climate change of -2% to 10% on irrigated farms for the period 2050-2100. Importantly we also estimate the current annual pasture yield variability, and find it to be 66% and 14% of the median yield for dry-land and irrigated farms, respectively. This represents an increase in variability of 0.1 - 0.5 tons DM ha-1 for the 25th-75th percentile range (an increase of 6-29%) and a 0.3-0.6 tons ha-1 for the 5th-95th percentile range (an increase of 4-12%) over the non-stationary historical record. The inter-annual variability reported here is significantly higher than the current change in average yield and the projected mean changes in other studies by 2050-2100. Our analysis of on-farm irrigation storage as an adaptation to climate change suggests that storage can provide a significant reduction in variability (13-41%), but that the effectiveness of the mitigation can be underestimated if the full weather/climate variability is not simulated. Our results were highly localised, but we found different outcomes between our two sites which are c. 20 km apart. In total, our work demonstrates a method to assess the impacts of climate change induced variability on pastoral systems and the value that such assessments provide. We suggest that changes in variability are as, if not more, consequential than the long-term change in average yield and that consequential changes in variability are already occurring, but are likely to be highly location-specific. Further we find that understanding the effects of changing variability is essential for effective climate change adaptation within the planning and management time horizons of relevance to the farming sector.