The Forecaster’s Mindset: Learning Nature’s Rules

Many business coaches indicate that a Growth Mindset is essential to adapt, learn and develop new skills. This way of thinking has been key to the development of our weather forecast system.

Is a mindset transition also needed in a smarter world of predictions?

As the technical developer, I always objectively question the status quo. The public sentiment is that climate change happens, the planet warms, and that we cannot do anything about the consequences. The fact is that the global warming trend will continue even in a zero-emission scenario. We must not only accept that climate change is real, but act now to lessen the future impacts before waiting to see the outcomes.

Food production can be increased using agri-tech. Real-time crop monitoring from drones and satellite imagery are already used to make disease predictions for preventative spraying, soil health is now better maintained by agronomists and fertilisers are applied more effectively using variable-rate application. Crop growth rates can be modelled using weather data, and the supply of fresh produce can be optimised to better meet supermarket demand.

As agricultural production reaches a plateau, innovation is now essential to make radical change.

Our goal to tackle the challenge of food security requires viewing the world as a ‘system’, with vulnerabilities and risks posed by changing patterns of seasonal weather. Some aspects of daily weather risks (probabiliities) such as air-frost, summer heat waves, and heavy rainfall are predictable on longer timescales and are valuable to growers even if their timing is not precise. With growing seasons shifting by several weeks from year to year, seasonal forecasts are vital to any growers success. They can be used to optimise the timings of frost protection, drilling or the propagation of high-value crops within glasshouses.

Rain Fields Machinery Agriculture

The official definition of a forecast, according to the Cambridge dictionary, is a “statement of what is judged likely to happen”. It is now widely recognised that a good forecast delivers on two essential features: skilful predictions, and confidence estimates. A forecaster is skilful if (on average) their predictions provide insights about the future that could not be achieved by chance alone. By comparing the accuracy of our predictions to past events, our margin of forecast skill is up to 18% above chance: and the confidence or bounds of outcomes are well constrained, retrospective hindcasting using the past 25 years of observations.

The predictive technology trend.

The trend to use technology to manage our lifestyles and behaviours is creating a more deterministic culture. Serendipity, once an important aspect of our experiences, is now uncommon. Its not surprising therefore that I’m constantly told that it is impossible to predict the long-term weather. This is not strictly true. International weather centres provide better forecasts than could have been imagined a decade ago. The challenge is not forecasting the long-term weather, but managing the uncertainty that makes deterministic forecasting techniques invalid beyond 7 days.

We need to shift our thinking from the short-term weather forecast mindset, which implies that we can precisely know the future, embrace uncertainty, and invent new ways of clearly presenting these insights to growers.

The biggest risk is doing nothing.

Machinery hire is expensive and it can not be moved into the fields if the soil is too wet or hardened by severe frosts. To advise on these risks medium-term (6 to 15 days) forecasts are currently used, updated daily. Weather forecasts of this type also prove useful for protecting flowering fruit trees in the spring. The problem arises in long-term planning adversely affected by changing seasonal weather patterns. While short to medium-term forecasts are often used for crop protection, farmers are now exposed to the risks of longer term weather variability for standard or fixed-date operations.

Our seasonal forecasts will soon provide information on summer rainfall in addition to cold weather risks (most useful for 2 to 10 week growing decisions) helping to reduce crop wastage, input and labour costs, time in the field, soil and environmental damage, while increasing crop performance and maintaining better farm incomes.

So what makes Weather Logistics different?

While we cannot say what will happen on a specific day, we provide lower and upper ranges in our seasonal outlooks. We deliver weather outlooks as daily metrics (specific to weather event exposure) at a farm location (field-level advisories). Agricultural experts using our weather data can perform reliable cost-benefit assessment to make better informed farming decisions. Whether it be frosts or heat waves, we provide you with a ‘error bar’ specific to your farm.