Supplement of Towards agricultural soil carbon monitoring, reporting, and verification through the Field Observatory Network (FiON)

Abstract. Better monitoring, reporting, and verification (MRV) of the amount,
additionality, and persistence of the sequestered soil carbon is needed to
understand the best carbon farming practices for different soils and climate
conditions, as well as their actual climate benefits or cost efficiency in
mitigating greenhouse gas emissions. This paper presents our Field
Observatory Network (FiON) of researchers, farmers, companies, and other
stakeholders developing carbon farming practices. FiON has established a
unified methodology towards monitoring and forecasting agricultural carbon
sequestration by combining offline and near-real-time field measurements,
weather data, satellite imagery, modeling, and computing networks. FiON's
first phase consists of two intensive research sites and 20 voluntary pilot
farms testing carbon farming practices in Finland. To disseminate the data,
FiON built a web-based dashboard called the Field Observatory (v1.0,
https://www.fieldobservatory.org/, last access: 3 February 2022). The Field Observatory is designed as an online service
for near-real-time model–data synthesis, forecasting, and decision support
for the farmers who are able to monitor the effects of carbon farming
practices. The most advanced features of the Field Observatory are visible
on the Qvidja site, which acts as a prototype for the most recent
implementations. Overall, FiON aims to create new knowledge on agricultural
soil carbon sequestration and effects of carbon farming practices as well as
provide an MRV tool for decision support.



S1 Calibration protocol details
The parameters of the BASGRA_N model (https://github.com/PecanProject/pecan/tree/v1.7.2/models/basgra/src) were screened with a global sensitivity analysis by varying the parameter values in their prior range. Priors are probability distributions encapsulating the likely values of model parameters that are formulated according to information coming from previous analyses, the literature, meta analyses, expert opinion, ecological theory etc. In this study, priors are chosen according to values provided by model developers and our expert opinion (also see Höglind et al., 2016;Hjelkrem et al., 2017, Huang et al., 2021. The top 20 parameters (see Table S1.1 below) that contribute most to model net ecosystem exchange (NEE) and leaf area index (LAI) output uncertainties were chosen for targeting in the calibration. Both the modeling and calibration approaches as implemented in PEcAn software are transparent, scalable and state-of-the-art techniques (Fer et al., 2018;. Calibration is done by assimilating both the LAI and NEE data streams simultaneously. The influence of NEE time-series were corrected for effective sample size as described in Fer et al. (2018). Due to the coarser time step and smaller sample size of the LAI data, no additional autocorrelations were applied for this data stream. alternative scenario). We restart the model from the states as of 13 June 2021 (which we archived) and run it forward with the 15-day weather forecast drivers available to us at the time (which we also archived). In other words, we start the model from its retrospective states and we don't use the now observed weather drivers for this period but the 15-day weather forecast to mimic a real-life scenario. The available forecast horizon for the farmer in this case is the next 15 days as of 14 June (until and including 28 June).
3 Figure S2.1 Alternative scenario of delaying (right column) the harvest date (red vertical lines) leads to a slightly larger C sink within the 15-day period while resulting in less yield. LAI stands for leaf area index and NEE for net ecosystem exchange where negative values indicate uptake and positive release of CO2 from the ecosystem to the atmosphere.
Delaying the harvest results in ~14% less yield for the farmer (Fig S2.1, top row). Both leaf area index (LAI) development ( Fig S2.1, middle row) and net ecosystem exchange (NEE; Fig S2.1, bottom row) during the week following the harvest imply that delaying the harvest may have some adverse effects compared to the real case. While leaving the grass intact longer during the 14-21 June period contributes to the overall C sink within this 2-week period (+0.08 vs alternative -1.16), it is not clear (from the perspective of carbon storage) whether the amount (even if we account for yield difference as a C output from the system) justifies delaying harvest on the whole at the expense of yield. Looking at this picture, considering the next 15-day forecast on 14 June, the farmer could be advised to keep the harvest date as 14 June in this case study.
To observe post-hoc what will happen to the C sequestration at this period later in the season, we run the model forward until the end of our predefined growing season, 31 October and calculate the C balance over the 14 June to 31 October period.
Note that this is just a forward run without being informed by the observations during this period, therefore this should be taken as a relative comparison. When we consider only the fluxes between the ecosystem and atmosphere, the actual harvest time was more favourable, i.e. it acted as a higher sink than the alternative case (-1251 kgC ha-1 vs. alternative -861 kgC ha-1, Fig. S2.2). If the carbon removed as yield is also considered, the actual harvest day was again more favourable (+288 kgC ha-1 vs. alternative +453 kgC ha-1, Fig. S2.2). This suggests that keeping the harvest time as it is might have indeed been preferred. While this is a demonstration, it exemplifies types of experiments that can be designed through this framework. Here we used a static and arbitrary one week delay and considered only yield, LAI and NEE variables, but more dynamic, systematic and longer-term explorations are also being devised. In addition, other constraints, such as the practical flexibility in the agricultural activities of an average farm should also be taken into account.