Skip to main content

How on-farm weather stations deliver greater accuracy - Blog

 

How on-farm weather stations deliver greater accuracy

This article was taken directly from The Progressive Farmer - April 2019 Edition

Recently, we analyzed the accuracy of "estimated" weather observations used in many precision agriculture applications. We then compared them against "known" observations from local, on-farm weather stations. Here's what we found.

The power of data
On the farm, data has many benefits. It can support timely, targeted decisions around planting, irrigation, spraying, fertilizing, disease management, insect outbreak predictions, yield estimates, and more.

While inaccurate data can cost you time, money, and yield, so can inapplicable data. For example, you may have noticed how rain and wind can vary greatly over relatively short distances. This can make forecasts for nearby locations unsuitable for spraying and fieldwork decisions. Likewise, soil and solar impact crop growth and stages, requiring specific awareness of what is going on directly above and below your plants.

The problem with distance
In precision agriculture, the most crucial weather parameters are precipitation and temperature. In particular, precipitation can vary greatly within a single mile.

In the United States, most official weather observations are gathered by the National Weather Service. The majority are farm airports located in or near urban areas. Virtually none of the official stations are located on farms. That's where estimates come in.

  • Most applications use one of the three methods to estimate precipitation and temperature:
    Use the nearest official station as a proxy.
    Use distance-weighted interpolation of official stations.
    Use radar or weather models.

For precipitation, many applications use radar-estimated rainfall, which is generally considered to be the most accurate. This produces spatially-consistent rainfall patterns.

 

The methodology
In our study, we compared estimated to actual data. To create the estimated data, we used spatial interpolation. This method is commonly used in environmental sciences to estimate values for unknown points by using data from nearby known points.

We then compared rainfall amounts and temperatures against readings from on-farm weather stations in our network of more than 5,000 systems across rural North America.

 

*Weather data gathered on-farm greatly improves accuracy throughout growing season. Precipitation is 20 percent more accurate; Growing Degree Days (GDD) 4 percent.*

 

What we found
The study showed that the estimated observations were much less accurate over the growing season. This is significant as many precision agriculture apps - such as crop growth stage, nitrogen use, and crop yield used plan fertilizer and chemical applications - depend heavily on weather data.

  • The results varied by weather parameter:
    There were significant differences between measured and estimated rainfall and temperatures.
    The average error in precipitation amounts was more than 3 inches (75mm) or 20 percent of the normal amount in the growing season. That translates into a similar 20 percent impact on yield.
    Precision agriculture apps that use estimated weather data to feed crop growth models for estimated plant stage, nitrogen use, or yield should expect errors of 20-25 percent based on weather data alone.

We found that the most accurate and relevant data is that gathered as close to the location as possible. Readings taken on site are ideal.

There is no substitute for an actual on-farm weather station. When the highly-accurate, on-site gathered data is paired with precision agriculture apps, better decisions can be made.

 

Learn more
Please visit www.dtn.com/ag-weather-station for more information on how to add an affordable on-farm weather station to your operation. You can also download a detailed white paper on our study.

 

This article was taken directly from The Progressive Farmer - April 2019 Edition

 

KP