background

What our experts have to say: Sencrop weather forecasts

What our experts have to say: Sencrop weather forecasts

Weather conditions and forecasts play a key role in every decision taken during the cultivation process. Are the conditions right for ploughing and sowing? Is there a risk of disease or pests? When should I spray to ensure maximum effectiveness and protection? When do my plants need to be irrigated? The weather data guides every decision you make and has a decisive impact on your yields.

šŸ’”
90% of crop losses are due to weather conditions, and 25% of these can be reduced if farmers have access to specialist weather services.

You need both accurate, local weather records and reliable weather forecasts to optimise your weather-related decision-making. Let's focus here on forecast data.

Integrating forecast models

Weather forecasting models are powerful tools that observe and transform data collected from a variety of sources - weather stations, satellites, radar, sounding balloons, aircraft and ships - into weather forecasts.

There are many forecasting sites and models at your disposal, and they often provide different information, sometimes complicating your decision-making.

The performance of a model depends on the parameters on which it is optimised, and we know that it is impossible for a model to be accurate in all situations and for all parameters. Each model therefore has its own advantages and limitations. Their reliability can vary depending on the region, the measurement or the time horizon. For example, a model may be very accurate in predicting the average temperature in a region, but less effective in anticipating night frosts.

Sencrop integrates around twenty weather models used by the main players in agricultural weather forecasting. These include GFS, ICON, Arome and Arpege. You can consult these different models and compare them at a glance using our model comparator.

Sencrop model comparator

Analysis of forecast data by Sencrop

Over and above the way in which a weather forecasting model is constructed, what interests us most is the relevance and accuracy of the information it provides.

At Sencrop, we analyse the data from our stations and those predicted by the weather models in order to calculate the statistical error between the observed values and the forecasts and obtain information on the performance of each model, at each station. To do this, we have access to a large observatory of weather events across our fleet of stations.

Analysis of the performance of forecasting models by Sencrop

Today, we have more than 35,000 connected weather stations, giving us a global and accurate view of territorial readings. It's the largest network of weather stations in Europe.

Sencorp network

Every week, more than 7 billion weather forecasts and over 21 million Sencrop hourly records are analysed. By analysing all the models on a daily basis, we are able to take a step back and assess their performance and accuracy.

To give you easy access to this information, the Sencrop application provides you with a classification of models in order of reliability and by type of measurement. This ranking is individual, for your locality.

Ranking of forecast models for a Sencrop station at specific time

The Sencrop forecast algorithm

We go even further, by offering you an automatic and continuous aggregation of the most reliable models for your locality and for each piece of weather data. Our aim? To simplify the way you consume forecast data, so that you always have quick and easy access to the most reliable forecasts possible.

How does it work? We aggregate the rankings presented above. We select the top 1 in the ranking for each metric (temperature, rainfall, humidity, wind) and for each time horizon. The horizon is the time separating the forecast from the observation, which allows us to have a different ranking and to adapt to the difficulty of predicting far into the future.

So the forecasting model used in your Sencrop forecasts can be different for each metric. And the model used for the same metric can be different for different time horizons.

Thanks to this visualisation, our users gain an average of 15% reliability on their forecasts, which is not insignificant when you consider how crucial the weather is in the agricultural sector.

It is therefore likely that there will be as many Sencrop forecasts as there are stations, depending on the results.

Of course, the weather remains unpredictable, but with Sencrop you save time and gain confidence in your day-to-day decision-making.

Discover the Sencrop solution

Create your profile, connect to a station close to your plots and benefit from a free trial, without obligation.

Try the application for 14 days