In Central and Eastern Europe, where sowing usually takes place sooner than in the South, predictions of wheat flowering are reliable as early as November. Elsewhere, accuracy increases as the end of the growing season approaches.
In agriculture, such data can be put to good use when choosing which variety of crops to plant when, and deciding on appropriate agro-management practices, while limiting environmental impact in the process.
The authors of the study also suggest that policymakers can rely on accurate predictions to draw up market-stabilisation strategies in order to stave off sudden price increases. Similarly, forecasts can come in handy for adapting trade patterns, and planning ad hoc payments in advance.
Extreme climate events
Crop-related weather forecasts become more and more difficult as extreme events such as droughts or heavy rainfall become increasingly common. This is why the study also looked into predictive skill concerning such phenomena.
When it comes to droughts, some results were encouraging. Substantial predictive skill was detected as early as February, when at least half of the wheat growth cycle is still ahead.
However, drought predictions concerning the period between heading and maturity - an especially sensitive phase in wheat growth - were rather disappointing. Predictive skill early on in the season was limited, and in some regions (particularly the UK and Ireland) did not get any better even as harvest was already looming. This was mainly due to atmospheric mechanisms involving fickle extratropical jets.
As for excessive wetness - a particularly relevant factor in disease occurrence - the forecasts came up short. They were even “dangerously useless” for some periods and regions.
2018 provided a rare event when droughts and wetness appeared at the same time on the European continent, making it an especially tall order to get forecasts right. Indeed, they struggled to provide skilful predictions before as late as June.
Room for improvement
The scientists behind the study have a number of recommendations for further improving climate forecasts for crops:
- very-high-spatial-resolution simulations can enhance predictions at the local to regional level.
- Summer predictions for Central Europe could be improved by taking into account spring sea surface temperature anomalies in the North Atlantic.
- Closer cooperation with farmers could also be beneficial.
This latter point was already heeded when Clisagri - a new agro-climate service tool - was set up.
Clisagri was created with the help of farmers and agronomists, who provided useful advice on how different weather events affect yields, crop diseases, storage and fertilisation.
Insights from Clisagri already inform the crop yield estimates of the JRC’s own MARS (Monitoring Agricultural Resources) team in a qualitative manner, with further plans to integrate them quantitatively.
The latest MARS bulletin forecasts above-average crop yields, in spite of the frequent rains and heatwaves that afflicted Europe during this summer.
A previous study found the MARS crop yield forecasting system to be quite accurate, with reliability, just like in the case of SEAS5, increasing as the crop season progresses. Yet its authors also called attention to increasingly variable and sometimes unprecedented climatic conditions.
Crop forecasters are getting better at making predictions. But they are also chasing a more and more unpredictable climate.