Estimating The Carbon Footprint Of Digital Agriculture Deployment: A Parametric Bottom-Up Modelling Approach

TL;DR

This paper tackles the problem of estimating the carbon footprint of large-scale digital agriculture deployments by introducing a bottom-up, parametric inventory and a simplified impact assessment that accounts for embodied and use-phase emissions. It models multiple technological systems deployed across a territory with a non-homogeneous distribution of farm sizes, enabling direct comparison of deployment pathways under realistic farm-size heterogeneity. Two France-wide case studies (dairy cattle and cereal crops) show that device diversity leads to heterogeneous footprints and that more complex devices do not always yield net benefits due to higher embodied emissions and greater device counts on larger farms. The approach enables first-order footprint estimation to inform policy and farm-level decision-making, while highlighting opportunities to extend the framework to additional environmental indicators and broader farming contexts.

Abstract

Digitalization appears as a lever to enhance agriculture sustainability. However, existing works on digital agriculture's own sustainability remain scarce, disregarding the environmental effects of deploying digital devices on a large-scale. We propose a bottom-up method to estimate the carbon footprint of digital agriculture scenarios considering deployment of devices over a diversity of farm sizes. It is applied to two use-cases and demonstrates that digital agriculture encompasses a diversity of devices with heterogeneous carbon footprints and that more complex devices yield higher footprints not always compensated by better performances or scaling gains. By emphasizing the necessity of considering the multiplicity of devices, and the territorial distribution of farm sizes when modelling digital agriculture deployments, this study highlights the need for further exploration of the first-order effects of digital technologies in agriculture.

Figures (7)

• Figure 1: Illustration of the deployment of various TS over a distribution of farms of different sizes for a given use-case.
• Figure 2: Schematized view of the process to assess the environmental impacts of a digital agriculture case study. # represents a number. Inputs excluded, parallelograms indicate computing processes and rectangles indicate process outputs. Overlapping rectangles are used to represent a multiplicity of individual results.
• Figure 3: Schematized plots of mixed TS modelling with $J = 3$ TS. In this figure $s$ represents the set of all farm sizes, $\forall s_i, s_i \in s$, $a_j$, $b_j$, $w_j$ models initial TS weighting function $\overline{t_j}(s)$ and $t_j(s)$ TS final function after normalization.
• Figure 4: Power consumption (a) and GHG emissions (b) for cattle use-case using French dairy cattle farm distribution (51,000 farms) and electrical carbon footprint for one year. (a) and (b) left-column: Full deployment of each technological system using farm size distribution. ($\diamond$): GHG emissions using an average farm size extrapolation approach. (b) right-column: Two different scenarios mixing the considered technological systems. "Low PC" mix deploys more systems using Connected Collars (TSCC), "High PC" mix deploys more systems using PC for computer vision (TSPC). Details related to mixed scenarios are available in Supporting Information S2 and figure's data is available in Supporting Information S3.
• Figure 5: Power consumption (a) and GHG emissions (b) for crop use-case using truncated French cereal farm distribution (65,223 farms) and electrical carbon footprint for one year. (a) and (b) left-column: Full deployment of each technological system using farm size distribution. ($\diamond$): GHG emissions using an average farm size extrapolation approach. (b) right-column: A scenario mixing the three considered technological systems. Details related to mixed scenario are available in Supporting Information S2 and figure's data is available in Supporting Information S3.