Defining drought from the perspective of Australian farmers
ABARES Working Paper
Published 11 November 2020
Authors: Neal Hughes, Wei Ying Soh, Chris Boult, Kenton Lawson
This study considers the definition and measurement of drought using ABARES farm survey data and the related farmpredict model. A new outcome-based drought indicator is developed for Australian broadacre farms which determines whether a given farm is ‘in-drought’ or not at a given point in time, based on the modelled effect of seasonal conditions on farm profits. This outcome-based indicator is compared against more traditional rainfall-based measures and farmer self-assessments of drought.
Farmer self-assessment data show differences in the perception of drought across farming types and regions (even after controlling for differences in climate conditions). The results also show a trend in drought perceptions over time, with current farmers less likely to self-assess as in-drought (for a given level of rainfall) relative to farmers in the past. This trend likely reflects farmers updating their expectations in-line with shifts in Australian rainfall related to long-term climate change.
The results show that the new outcome-based drought indicator more accurately represents the effects of climate on farm outcomes compered to simple rainfall measures. The farm profit indicator accounts for the circumstances of individual farms, including their mix of cropping versus livestock activity and their size and location. In particular, the farm profit indicator can account for rainfall timing effects (relative to the local cropping season) and the effects of drought on commodity markets (i.e., higher domestic grain and hay prices in drought years).
While these new drought indicators offer greater accuracy, they are more complex and careful consideration needs to be given to how they would be applied and communicated in practice. In particular, these indicators would require data and/or assumptions on farm characteristics and seasonal weather forecasts. Further, they offer assessments of drought for a specific context (Australian broadacre farms) and may not accurately reflect drought impacts in other farming sectors or the wider community.
Regardless, these new outcome-based drought indicators have a range of potential applications including serving as eligibility criteria or early warning systems for government drought programs or supporting new index-based (parametric) drought insurance products.
Long-standing problems with defining and measuring drought constrain the development of effective drought policy while also hampering drought research and adaptation. This study examines drought from the perspective of Australian farmers, drawing on data from the Australian Agricultural and Grazing Industry Survey and the related farmpredict micro-simulation model. Farmer drought self-assessment data are used to identify a range of factors beyond rainfall which influence farmers perception of drought. The data also show evidence of trends in farmer drought perception over time related to long-term shifts in the Australian climate.
The farmpredict model is then applied to develop an outcome-based drought indicator for Australian farms. This indicator measures the effects of seasonal weather conditions on farm profits, accounting for the characteristics of individual farms, and for the effects of drought on domestic commodity prices. The results show that this outcome-based drought indicator more accurately represents the effects of drought on farm profits than simple rainfall percentiles. This new indicator could help to address information problems which constrain both government drought programs and private drought insurance markets.