Farm Data Sensing

Problem Statement

Data sensing of soil and environmental factors has the intended outcome of reduced economies of scale for managing more diverse-integrated agroecosystems and increased feedback loops - which creates a citizen culture of open science as fun and social, and inquiry as part of citizenship and encourages observation/investigation/involvement in adaptive management. The cultural influence of the open approach is of utmost importance - lowering the cost and increasing accessibility so that this kind of work is no longer relegated just to governments and multinational corporations. In my view open access to the systems that support a regenerative food and fuel supply is crucial to self determination.

Why?

For what scale are you talking about enabling? Diverse crop systems on the 40 acre farm scale, 1 acre, or what?

I don't understand this fully, as integrated agroecosystems by default are effective in resilient resource management, making the 'engineered approach = sensors and point feedback' of the larger scale moot. Can you explain?

This is compared to our work on equipment. Strict engineering requirements and complexities are reduced by overbuilding, modularity, lifetime design - creating a much more resilient mechanical system. It appears that similar principles of robustness apply to agroecology - where 'sensors and feedback loops' are a layer that is already addressed by design of the agroecosystem in the first place.

Answer

Hi Marcin,

This gets into a meaty topic. Our challenge with agroecosystems are many fold. I agree that there is the ideal of a self regulating, resilient agroecosystem, but in practice steady state is really a range within a regime, and there are many states of performance within a regime, not all of which are optimal. (ball in basin model from resilience thinking and reference to CS Holling's adaptive cycle).

1) there are so many variables in biologically based (plant animal, plant soil, climate plant etc., interactions) agricultural systems that it is difficult to communicate about these systems and to transfer knowledge about them from farmer to farmer and from farm to farm and yet there is a clear need for far greater experimentation.

2)The state of our science is such that we don't really understand the root and soil level nutrient cycling and extraction process very well, and yet we know it functions well in certain instances and there are general practices that encourage soil building etc.

3) while each farm has unique microclimates and soil, plant animal interactions that should be accounted for, broader trends also effect management decisions - so both micro and macro resolution are important

4) we have not even begun wide spread localization of varieties and implementation of more complex rotations and a real methodology around f on-farm trials. Ideally every farm becomes a research farm - with environmental data being one of the primary public goods produced along with clean water, habitat, other services as well as agricultural goods.

5) the current model of replicated trails on research farms is both labor intensive, expensive and not very accurate at predicting on-farm performance in conditions other than sterilized conventional farming methods. Trails are often done in conditions with poor soil health etc, and so much of the complexity is missed in published papers. Trials have historically been constructed to oversimplify systems to make the data collection and statistical analysis easier. lower cost sensors and increase in computing power changes that model, but this has yet to be incorporated into the culture.

6) collecting on-farm data and records has been expensive and erratic, prone to error and often doesn't get done. Variables in human observation, and the limits to our observational accuracy and capacity are also limitations. There are lots of examples, but we can only see in a limited spectrum, and while we are good at seeing patterns, our point of view is limited by our limited field of view and perspective. We are also limited by our time to manually observe.

7) adaptive management in general and of nitrogen in particular is sensitive to small environmental changes.

8) beyond on-farm management many of the environmental factors extend into the commons/public domain. All the major nutrient cycles have wider public and environmental services, which are important to quantify for evaluating the positive (or negative) effect on these systems. Quantifying these flows will enable greater cultural understanding and shift values about farming practices and compensation for farmers for services other than agricultural production. This includes all the major nutrient cycles - Carbon, water, nitrogen, phosphorus, potassium and associated services including water and air filtration, water storage, and regulation, carbon sequestration, soil stabilization, habitat development. Flows of these services from one farm to another are important to recognize where practices are working and where they are not.

With inexpensive sensors, and imaging we have the ability for the first time in human history to not only record environmental data at a high resolution but to inexpensively share it, log it, aggregate it etc. We many be able to gain individual insights from our own data, but by aggregating and analyzing it across a landscape we can also identify broader patterns and trends. So to answer your question about scale, it is less about scale than about a network which may have high and low density areas.

This has a clear overlap with equipment too in the culture of adaptation and experimentation. Just like good documentation and observation of these biological systems can be fed through a broader network, so too can observations of the tools that are interacting with those biological systems. One farmer's observation of the effect of a yeoman's plow design compared with another may not only be transmitted through the same platform that the soil conditions and other environmental data which puts those observations in context. What works in a wet clay soil, in a wet year with low root density many have quite a different effect in a rocky sandy soil, with dense grass rhizomes etc.. The culture of observation and sharing and network context is what is most important. I hope this helpful in clarifying the approach and relevance to the broader project.