THE CLIMATE KELPIE BLOG: A framework to explore climate risky decisions

Posted by BCG on 29th March 2022

Peter Hayman, SARDI Climate Applications and Barry Mudge, Mudge Consulting

Over the last four years the Forewarned is Forearmed project has worked with Industry Reference Groups from grains, northern red meat, southern red meat, dairy, wine grapes and sugar. These groups are made up of leading farmers and advisers from around the country. It is not surprising given the spread of industries and regions that during the life of the project there has been first-hand experience of floods, fires, drought, untimely rain at harvest, heat waves and severe frost damage. Early in the project each group was asked to prioritise their main weather and climate risks. Examples of unfavourable events include drought, frost and heat spikes on grains and wine grapes and excessively wet conditions influencing harvest in crops or impacting animal health, pugging paddocks and restricting operations in livestock enterprises.

The Bureau of Meteorology is the main research partner in the Forewarned is Forearmed project and most of the project resources have been allocated to the underlying science and effective presentation of forecast products to address these extreme weather and climate events.  See Climate Kelpie for a description of the products.  The project has also allocated resources to the question of how the forecast information can be used in decision making.

Farmers are regularly making climate risky decisions such as topdressing crops and pastures with fertiliser, spraying crops for foliar disease, sourcing extra harvesting equipment to prepare for a wet harvest, adjusting the stocking rate of sheep and cattle for the coming season or changing the ration for dairy cows before an expected heatwave.  Although these decisions have different context, they all share the same structure where a choice has to be made prior to the event and one choice is preferred under a certain set of future climate conditions and another choice is preferred in other conditions. Many of these climate risky decisions involve tens or hundreds of thousands of dollars.

Our aim is to have clear conversations with farmers and advisers about their climate risky decisions and sometimes it helps to get this out of our heads and onto a whiteboard or a spreadsheet. The logic of these decisions can be helped by 1) clarifying the favourable and unfavourable climate conditions, 2) identifying choices and 3) considering the outcomes that arise from the combination of climate and choices.  Note that we are addressing the logic of the climate risky decision before we consider forecasts. We also recognise that farming involves many complex factors that can’t be included in a simple framework. Our experience is that getting the simple logic down generates useful discussion about the complexities. This process doesn’t have to be used for routine decision making, it is designed to generate conversations and thinking that can be used to confirm or challenge rules of thumb.   

Step 1 Clarifying the climate risk by asking what is the unwanted or adverse set of climate conditions? It is useful to be more specific about the risk. Rather than just saying “a wet harvest”, it is helpful to think about the likelihood or extremity of the risk.  It is unlikely that half the harvest periods in the past have been problematic so rather than wetter than average (or median) it is better to say as wet or wetter than the 1 in 5 or 1 in 10 harvest period. This does not need to be overly precise, sometimes it is easier to think in terms of likelihood of the risk (decile 9 and 10) than trying to be too specific about an exact threshold in mm or wet days.

Step 2 Identify the choices – to lean towards optimism or caution

In almost all risky decisions it is useful to consider a more cautious choice that reduces the chance of loss and a more optimistic choice that is potentially more profitable and hopes for a favourable climate. The example of a more cautious approach to a wet harvest is to employ a contractor or purchase extra equipment. Part of slowing down and thinking is not jumping to which decision is best but laying out the options and then deciding.  It might be that it is obvious to lean towards caution, however, managers will point to shortfalls of being overly cautious or overly optimistic.

Step 3 Considering the outcomes – rewards and regrets

We can consider the simplest situation of four outcomes arising from two choices (caution or optimism) and two climate states (adverse or favourable).

Without perfect knowledge of the future climate, the possibility of regret is unavoidable. By considering these four possible futures the decision can be obvious. The higher the possible regret, the greater the value of information about the coming climate. If the level of reward or regret is trivial, the value of the forecast is trivial.

Step 4 Introduce forecasts of adverse weather and climate  

Imperfect forecasts of adverse weather or climate events can be a hit (True warning or True non- warning) or a miss (Failure to warn or False Alarm).  Because of the previous steps clarifying the links between decisions, climate states and outcomes, we can match these reward and regrets to the forecast.

A contingency table includes forecasts, action and outcomes. If we take the example of excessively wet conditions and assume that a farmer acts on the forecast.  If there is a forecast for wet conditions, the farmer may consider extra harvest resources through employing more labour, accessing more equipment or employing a contractor. With a forecast of normal or dry harvest the farmer will proceed with their usual equipment and labour.

The aim of this simple framework is to bring clarity to the verbal description of the risky decision. The best outcome is the bottom right-hand corner where there has been no expense of the extra resources such as the contractor. If conditions are excessively wet, forecasters and farmers hope to be in the top left hand corner of being forewarned in time to be forearmed. It is inevitable that an imperfect forecast will involve failures to warn and false alarms. Comparing the regret of caution (cost of unnecessary protection) and the regret of optimism (loss and damage) provides some guidance on the use of uncertain forecasts.

A farmer, looking at the above table asks the reasonable question of how often the forecast lands in each of the four boxes. Although thumbs up and thumbs down is a step toward ranking the outcomes, calculating the economic outcomes for each box can provide guidance on the required hit rate for the forecast to be useful and the percent chance of wetter conditions that should spur action. We have developed a spreadsheet framework called Rapid Climate Decision Analysis which considers the profit across a wider range of climate states (usually 10 deciles).   This will be discussed in a future Climate Kelpie article.

The Forewarned is Forearmed project is supported by funding from the Australian Government Department of Agriculture, Water and the Environment as part of its Rural Research and Development for Profit program.

Leave a Reply