The hurdle to widespread adoption of satellite imagery in precision farming.

Satellite imagery-based maps must be intuitive and actionable – a case study on precision nitrogen fertilization. 

The experts believe that the farmers must adopt precision farming and that can be supported by cost-effective, easy-to-use, satellite imagery based agricultural application maps, but the adoption of recommendation maps are not as high as the industry anticipates. To address this adoption barriers, the industry must offer maps to the farming operations, which are intuitive and actionable. What does that actually mean?

Commonly used analogy, which helps to understand the problem: Let’s say a blood test report says that the patient’s blood sugar level is 200 mg/dL. An average patient probably would not understand much from that report. Now, let’s assume that the report also says that a blood sugar level, less than 140 mg/dL is normal. With this new information, the report makes a bit more sense and the patient now can understand that his blood sugar level is high. But, to prescribe a solution, additional questions must be answered like what the patient’s state was, when the (e.g., during fasting) blood sample was taken, patient’s age, sex etc.

This is also true for agricultural recommendation services. The Ag Tech industry must offer measurable plant health information along with the benchmark information so that the farmers can understand the status and the needs of their crops, better.

A case study on nitrogen application: The goal of the study was to find out – how various satellite-imagery based information impact farmer's farming decisions. To find out, we have provided two sets of maps to two groups of farmers. Both map sets provided information on the plant’s health that can be utilized to make better and faster N fertilization decisions.

Group A: First group of farmers has received NDVI maps (Normalized difference vegetation index) to perform Variable Rate (VR) nitrogen application (see figure below). The NDVI is a dimensionless index that shows the density of greenness (crop health) on an area of land. NDVI values for plants range between 0 to 1, where 1 represents high and healthy biomass and 0 means dead and low/no biomass. We also have explained the farmers, how to best utilize the NDVI maps to spread the fertilizer better – to vary the fertilizer rates according to the variability of the NDVI values across the field.

From the farmers, we have learned two different strategies for VR N app. using NDVI:

1. The plants with high NDVI values should receive high N rate and the plants with low NDVI should receive less. The underlying assumption is that the field areas with poor plant health/low density/low biomass (low NDVI) have less demand for N fertilizer.
2. The plants with low NDVI value should receive high N fertilizer rate because the low NDVI values may represent the plant nutrient deficiency, thus the plants need more N fertilizer to catch up the healthier plants.

Both strategies make sense, but which strategy is more appropriate and under which condition? The discussion continued and the ambiguity is still there and at the end there are two groups of farmers who support two different strategies. The takeaway message from that discussion is that there was no consensus, and probably such ambiguity is limiting the majority of the farmers adopting NDVI maps for VR N application. There are ways to resolve such ambiguity by conducting soil test and plant analysis to find out which strategy is appropriate for which field (or parts of the field), but that makes the NDVI map utilization process for VR N fertilizer application, time intensive and cost inefficient.

NDVI map is simple to read and good for field scouting but offers complex information, which is hard to interpret thus makes it less actionable for N fertilization because NDVI represents combined information on biomass and stresses.

Group B: The second group of farmers was provided with two maps, initially (see the figure below). One showing the dry matter (DM) distribution in t/ha and another one showing the nitrogen uptake in kg/ha. The idea was to show them the plant growth (biomass accumulation) and N uptake, separately. We quickly realized that the N-uptake map is much more intuitive because the farmers cloud relate to that and engage in productive discussions like “I usually provide 170 N kg/ha to grow my winter wheat and the map shows that the plants have already taken up 70 kg/ha N, which means the plant still needs 100 kg/ha N that needs to be distributed over the next two applications”. On the contrary, group A farmers had struggled to decide the N rate in relation to the dimensionless NDVI index values.

Moreover, it was also interesting to observe that the DM and N uptake distribution are not the same and that raised the question why? The explanation is that the higher DM does not necessarily mean that the N uptake will be higher and that may be due to a variable supply of N by the soil or variability in the plant’s ability to take up N due to some other limiting factors like water availability or total biomass accumulation. Here, the VR fertilization strategy is – areas with high DM + high N uptake needs less fertilizer but the areas with high DM + low N uptake needs more fertilizer. This strategy was convincing and there was a consensus. Now the question is how to determine the fertilization rate, precisely?

To address that, as a next step, we have presented the farmers with two more maps (see the figure below). One showing the current, total N demand (N critical, the benchmark information) of the plants for optimal growth, derived from the dry matter map. And another map is showing the difference between the N demand and N uptake maps, which we call N shortage map. The N shortage map represents the amount of N, the plant still need to grow optimally, and this map is the suitable for VR N fertilizer application.

For the field (shown above), the NDVI map and DM map are somewhat correlated, but this might not be true for other fields. Now, if we notice the NDVI map and the N shortage map, we can observe that there is not much correlation between the two maps. So, applying N fertilizer based on NDVI map might not be optimal in every case. On the other hand, the N shortage map provides measurable information thus makes it actionable. However, we have recommended an average 42 kg/ha N rate for this field, but the farmer had decided to apply 60 kg/ha N. Without our information the farmer would have applied 80 kg/ha. So, by using this maps, the farmers have cut 20 kg/ha N – meaning additional revenue for his farm.

The DM and N uptake maps, along with the benchmark (N critical) information, offers the comprehensive overview to the farmers that helps them to understand and assess the N fertilization need better and makes the N shortage map much more actionable.

If you are interested to know more about this or how we do it, you can listen to this podcast.

Conclusion: Our farmers are brave in adopting new technologies. They care about the environment and understand the very concept of sustainability. Because Farming is a business of sustainability, which is usually carried out over generations. So eventually the farmers will adopt the precision farming techniques in every field of our planet, and we can ensure faster adoption by providing them intuitive and actionable information therefore assist them in growing our every-day food.