When I think about precision agriculture my usual thoughts are around the continuous learning about the variability in a field. There are many layers of information or data that can be gathered about each field. And from these layers we can begin to make decisions about potential productivity and applications of crop inputs. But when we think of all the layers available, are we concentrating on the most efficient processes? Enter satellite imagery. And specifically, in-season imagery that offers a progression of images of the developing crop.
For all the promises of its use we have yet to realize much value at the farm level. Why is this? The most recent survey of ag retailers suggests that is not because we are not trying. And most certainly some retailers have had success in offering this service to farmers. But its usage trails other types of precision ag offerings that are more expensive rivals.
I’m sure that some of the reason is that we have recently come through several years of unprecedented margins for most farmers. This afforded many the luxury to try new services they had never used before. Most imagery could be acquired from sources that made it relatively easy for them to view images of their fields. As I speak to farmers almost all of them have tried imagery services in the recent past. However, almost all of them also say that they have struggled to find the value in those services.
Retailers also have struggled with the value of in-season imagery. There are always a few successful discoveries in the imagery each year that can be attributed to some significant net returns. But they seem to come almost by accident instead of the result of an engrained mindset of using imagery as a scouting regimen. And the most recent Precision Ag survey of retailers by Purdue University reflects their difficulty in making any margin from its use.
One of the most frequent reasons that I hear is to blame for imagery not meeting expectations is the temporal resolution of the images that are presented. In other words, we don’t get images frequently enough to make use of them. And indeed, that can be a problem. If you don’t have an image in the timeframe that allows some action, what value is it other than the getting bad news right before harvest about your yield robbing condition? Or maybe, if you remember, you can monitor that area or condition in subsequent years? But don’t we have yield maps for that? You get the point.
And what about the time it can take to process the data from the satellite sensors into an image on any given platform? FarmersEdge has quantified their process in the chart to the right.
Two to three days of processing is typical. But added into the process can be the communication necessary for a crop scout to ground truth the image and report back to the farmer. A delay at any step will render the insights from the imagery historical reference instead of actionable data. Some retailers have abandon the use of satellite imagery for field scouting for this reason alone. They do not get the number of images they need at the right times to add enough value to their scouting services. Mostly they are relying on directed aerial imagery.
So how do we turn the corner in the usefulness and value perceptions of in-season satellite imagery? I believe there are two significant developments in the marketplace that will change our perceptions for good this growing season. One is the launch of the dove network of satellites by Planet Labs. The second is the development of new tools of analysis using the various spectral bands to deliver more unique images.
Planet Labs is already streaming data from its new constellation of dove satellites. When they are finished launching the complete constellation it will increase the number of available images available during the growing season dramatically.
FarmersEdge offers traditional NDVI maps but has also developed two new maps for comparison using all the spectral bands to present maps that highlight variability in different ways. Map one shows the greatest variability in the field and is assigned a percentage of variability that exists on the map. Map two shows the next, subtler factor of variability in the field and is also assigned a percentage. The purpose and usefulness of the maps are many, but there are two most significant that will change our perceptions of in-season imagery forever.
Firstly, the images will show variability that has yet to be detected in a normal NDVI image. This gives the agronomist or farmer an earlier warning of where to scout and alerts them of a field challenge much sooner than ever before. Secondly, Field Variability Map 2 will show what is developing as the next potential factor of difference in the field. You may ask “how do I use these two maps together?” Think of it this way, the major variability in the field the day the image is taken will be reflected in FVM1, the next biggest variability will be reflected in FVM2. For example, patches of insect damage when it starts might only appear in FVM2, but if left untreated, this insect damage may increase and on the next image may be the biggest cause of variability in the field and moves into FVM1.
This growing season we will have both the increased frequency of images to work with, and a new way of processing the data to give us earlier detection of variability within a field. How exciting! I have not personally been more optimistic about the value of in-season imagery delivering value to the farmer than I am this year due to these two new developments.
I encourage everyone in the practice of consulting with farmers to take a close look at these new developments in the in-season imagery field of services to see if greater value for farmers can be achieved this year.