A method of quantifying the carbon footprint of livestock

Abstract

This disclosure relates generally to determining a change in soil organic carbon quantity in a geographical area, and determining a total carbon footprint of an agricultural area. A processor calibrates an in-silico model configured to simulate soil carbon sequestration within the geographical area using measurements received from one or more sensors placed at: a first location having environmental conditions similar to the geographical area; and / or a second location associated with the geographical area, the measurements being indicative of: (a) a level of carbon flux; and optionally (b) one or more environmental conditions; of the geographical area. The processor provides input data, and determines the change in soil organic carbon quantity based on an output of the in-silico model. The processor determines the total carbon footprint based on the change in soil organic carbon quantity and a livestock carbon footprint determined by applying a carbon accounting model to livestock data.

Description

Cross-Reference to Related Applications

[0001] The present application claims priority from Australian Provisional Patent Application No 2024901579 filed on 28 May 2024, Australian Provisional Patent Application No 2024903468 filed on 25 October 2024, and Australian Provisional Patent Application No 2025901233 filed on 10 April 2025, the contents of which are incorporated herein by reference in their entirety. Technical Field

[0002] This disclosure relates generally to determining a change in soil organic carbon quantity in a geographical area. This disclosure also relates generally to determining a total carbon footprint of an agricultural area. Background

[0003] Geographical areas, such as agricultural areas like farming properties, output greenhouse gas emissions, such as carbon dioxide and methane, due to animals within the geographical area, as the animals emit greenhouse gases through respiration, digestion of ruminants and their excreta. While greenhouse gases may also be emitted through other practices within the geographical area, such as electricity use and vehicle use, the greenhouse gases that result from livestock are emitted in large quantities, especially if the areas contain much livestock, such as farming properties and beef production areas.

[0004] However, greenhouse gas emissions can be offset through plant photosynthesis. Further, carbon in the atmosphere is removed and stored as soil carbon (e.g., soil organic carbon) through a process referred as soil sequestration. As such, the net (or overall) greenhouse gas emissions may not be significant or may even be carbon negative (i.e., the area draws in more greenhouse gases from the atmosphere that it produces). Therefore, accurately understanding the net emissions is particularly important for management of geographical areas, such as cattle farms, as they can manage the geographical area to modify their carbon footprint. For example, beef producers can implement practice changes to enhance soil carbon sequestration within a geographical area to offset their carbon emissions. Australian beef production relies heavily on optimising pasture and animal productivity to ensure sustainability, profitability and carbon neutrality.

[0005] Understanding the carbon footprint of a geographical area can be understood by determining the degree of soil carbon sequestration (i.e., how soil organic carbon quantity changes). Currently, the change in soil organic carbon (SOC) quantity in a geographical area can be measured through soil coring methods, in which a corer is physically inserted into the ground of the geographical area to retrieve a soil sample which is then further analysed. However, soil coring methods are limiting as they require a physical tool, which can be difficult in obtain in remote areas and may lead to inaccuracy. Further, soil coring is time consuming, which makes it difficult to implement practices changes quickly or “on the fly”. Even further, periodic soil coring has faced scrutiny due to variability and unreliable vendors. The soil carbon accounting through soil coring also relies on measuring changes in SOC stocks by comparing values at two points in time. While this method can detect whether SOC has increased or decreased, it does not reveal the underlying cause of change. In particular, it cannot differentiate between the changes driven by climate variability and those resulting from management interventions. As such, there is a high likelihood that SOC is overstated in this approach (sampling between two points in time). Soil coring results are also highly sensitive to both the sampling strategy and laboratory processing methods, leading to the risk of further significant variability in SOC estimates. In short, current coil carbon methods are expensive, inaccurate, and uncertain in the quantification of soil carbon levels.

[0006] Further, due to slow rate of change and high spatial variability of soil organic carbon, detecting changes in soil organic carbon through direct measurements within 5-10 years is typically infeasible due to high sampling effort required. Moreover, establishing a counterfactual (e.g., what would have occurred in the absence of management changes, which is useful for understanding project impacts) cannot be directly measured using soil coring.

[0007] As a result of these issues, it is difficult to accurately predict and quantify any changes in soil carbon quantity. It is also difficult to do this in a cost-effective manner. Moreover, these issues have meant that the financial viability of undertaking a soil carbon project and making a positive return on this investment is too risky for many producers to consider. Therefore, there is a need for a more accurate method for determining a change to soil organic carbon quantity in a geographical area. Further, there is a need for a method which is also cost effective. Such a method would also be able to provide an accurate determination of the total carbon footprint of a geographical area, such as an agricultural area like a farming property, such that practice changes can be implemented to modify the total carbon footprint.

[0008] Any discussion of documents, acts, materials, devices, articles or the like which has been included in the present specification is not to be taken as an admission that any or all of these matters form part of the prior art base or were common general knowledge in the field relevant to the present disclosure as it existed before the priority date of each of the appended claims.

[0009] Throughout this specification the word “comprise”, or variations such as “comprises” or “comprising”, will be understood to imply the inclusion of a stated element, integer or step, or group of elements, integers or steps, but not the exclusion of any other element, integer or step, or group of elements, integers or steps. Summary

[0010] Disclosed herein is a method and system for determining a change in soil organic carbon quantity of a geographical area. In particular, the disclosed method may be a computer-implemented method, in which an in-silico model of the geographical area is provided to accurately determine the soil sequestration in the geographical area. Moreover, disclosed herein is a method and system for determining a total carbon footprint of an agricultural area. Similarly, the disclosed method may be a computer-implemented method, which accounts for the soil sequestration and greenhouse gas emissions produced by livestock within the agricultural area to provide an accurate determination of the total carbon footprint of the agricultural area.

[0011] In a broad aspect of the present disclosure, there is provided a method for determining a change in soil organic carbon quantity and / or a level of soil organic carbon of a geographical area, the method comprising: providing an in-silico model of the geographical area calibrated using measurements received from one or more sensors, the in-silico model being configured to simulate soil carbon flux, such as soil carbon sequestration, within the geographical area, wherein the one or more sensors are placed at: a first location having environmental conditions similar to the geographical area; and / or a second location associated with the geographical area, the measurements being indicative of a level of carbon flux in the geographical area; providing input data indicative of current characteristics of the geographical area to the in-silico model; and determining the change in soil organic carbon quantity and / or the level of soil organic carbon of the geographical area based on an output of the in-silico model.

[0012] In another aspect of the present disclosure, there is provided a method for determining a change in soil organic carbon quantity and / or a level of soil organic carbon of a geographical area, the method comprising: calibrating an in-silico model of the geographical area using measurements received from one or more sensors, the in-silico model being configured to simulate soil carbon flux, such as soil carbon sequestration, within the geographical area, wherein the one or more sensors are placed at: a first location having environmental conditions similar to the geographical area; and / or a second location associated with the geographical area, the measurements being indicative of (a) a level of carbon flux; and optionally (b) one or more environmental conditions; of the geographical area; providing input data indicative of current characteristics of the geographical area to the in-silico model; and determining the change in soil organic carbon quantity and / or the level of soil organic carbon of the geographical area based on an output of the in-silico model.

[0013] According to the present disclosure, there is provided a system for determining a change in soil organic carbon quantity and / or a level of soil organic carbon of a geographical area, the system comprising a processor configured to perform any of the previously described methods.

[0014] According to the present disclosure, there is provided a method for determining a change in soil organic carbon quantity of a geographical area, the method comprising: calibrating an in-silico model of the geographical area using measurements received from one or more sensors, the in-silico model being configured to simulate soil carbon sequestration within the geographical area, wherein the one or more sensors are placed at: a first location having environmental conditions similar to the geographical area; and / or a second location associated with the geographical area, the measurements being indicative of: (a) a level of carbon flux; and optionally (b) one or more environmental conditions; of the geographical area; providing input data indicative of current characteristics of the geographical area to the in-silico model; and determining the change in soil organic carbon quantity of the geographical area based on an output of the in-silico model.

[0015] It is an advantage to calibrate the in-silico model using the measurements received from the one or more sensors, as the measurements are indicative of the carbon flux, which can thereby be accounted for in the in-silico model. This produces a more accurate representation of the geographical area and hence, provides a more accurate determination of the level of soil carbon of the geographical area.

[0016] In some embodiments, determining a change in soil organic carbon quantity comprises determining a level of soil organic carbon of the geographical area.

[0017] In some embodiments, , the measurements are indicative of: (a) a level of carbon flux; and (b) one or more environmental conditions; of the geographical area

[0018] In some embodiments, the method comprises repeating the steps of calibrating the in-silico model, providing input data to the in-silico model, and determining the change in soil organic carbon quantity upon receiving further measurements from the one or more sensors, such that the change in soil organic carbon quantity is quantified on a rolling basis.

[0019] In some embodiments, the input data comprises soil data corresponding to information regarding soil within the geographical area.

[0020] In some embodiments, the information regarding soil within the geographical area comprises information indicative of one or more of: soil texture; soil pH; and soil bulk density.

[0021] In some embodiments, the method further comprises receiving soil data of the geographical area based on a result of soil testing in the geographical area.

[0022] In some embodiments, the input data comprises plant data representing information regarding plants within the geographical area.

[0023] In some embodiments, the input data comprises pasture data representing information regarding pasture within the geographical area.

[0024] In some embodiments, the measurements received from the one or more sensors corresponds to one or more of: weather data representing information regarding weather conditions within the geographical area; and rainfall data representing information regarding rainfall within the geographical area.

[0025] In some embodiments, the method further comprises receiving the weather data and / or the rainfall data periodically; and determining the change in soil organic carbon quantity of the geographical area upon receiving the weather data and / or the rainfall data periodically.

[0026] In some embodiments, the in-silico model is indicative of a history and management of the geographical area.

[0027] In some embodiments, the in-silico model is a digital twin of the geographical area.

[0028] In some embodiments, the in-silico model is based on a time series biogeochemical model.

[0029] In some embodiments, the one or more sensors is a flux tower.

[0030] According to the present disclosure, there is provided a method for determining a total carbon footprint of an agricultural area, the method comprising: providing input data indicative of current characteristics of the agricultural area to an in-silico model of the agricultural area, wherein: the in-silico model is configured to simulate soil carbon sequestration within the agricultural area; the in-silico model is calibrated using measurements received from one or more sensors placed at: a first location having environmental conditions similar to the agricultural area; and / or a second location associated with the agricultural area; and the measurements are indicative of: (a) a level of carbon flux; and optionally (b) one or more environmental conditions; of the agricultural area; determining a change in soil organic carbon quantity of the agricultural area based on an output of the in-silico model; determining a livestock carbon footprint by applying a carbon accounting model to livestock data, the livestock data representing information of livestock within the agricultural area and the livestock carbon footprint being indicative of an output of the carbon accounting model; and determining the total carbon footprint of the agricultural area based on the change in soil organic carbon quantity and the livestock carbon footprint.

[0031] In some embodiments, the method further comprises determining a carbon footprint of a product derived from the livestock from the agricultural area based, at least in part, on the total carbon footprint of the agricultural area.

[0032] In some embodiments, the product is one or more of: wool; meat; and milk.

[0033] In some embodiments, the livestock data comprises liveweight of at least one animal in the livestock.

[0034] In some embodiments, the liveweight of the at least one animal is based on walk over weighing.

[0035] In some embodiments, the livestock data comprises liveweight gain of at least one animal in the livestock.

[0036] In some embodiments, the livestock data comprises data indicative of one or more of: total livestock numbers; species of livestock; livestock added; livestock removed; livestock weights; wool shorn; and livestock birthing rates.

[0037] In some embodiments, the method further comprises monitoring a selection of animals in the livestock to determine at least part of the livestock data in real time.

[0038] In some embodiments, monitoring each animal in the selection of animals in the livestock is based on a radio frequency identification device placed on each animal.

[0039] In some embodiments, determining the livestock carbon footprint comprises applying the carbon accounting model to one or more of: plant data representing information regarding plants within the agricultural area; fertiliser data representing information regarding fertiliser use within the agricultural area; energy consumption data; and fuel consumption data.

[0040] In some embodiments, the method further comprises generating a price grid matrix based on the total carbon footprint of the agricultural area.

[0041] In some embodiments, the method further comprises generating a report containing one or more of: the total carbon footprint of the agricultural area; the change in soil organic carbon quantity; the livestock carbon footprint; the carbon footprint of the product; and a further carbon footprint based on other sources of carbon within the agricultural area.

[0042] In some embodiments, the method further comprises providing a recommendation based on the generated report, the recommendation being indicative of a practice change to implement in the agricultural area to modulate the total carbon footprint of the agricultural area.

[0043] In some embodiments, the practice change comprises one or more of: pasture improvement; planting legumes; applying fertilisers; transitioning from cropping to permanent pastures; and time-controlled grazing.

[0044] In some embodiments, the total carbon footprint of the agricultural area; the change in soil organic carbon quantity; the livestock carbon footprint; and / or the further carbon footprint based on other sources of carbon is represented in carbon credits, Australian Carbon Credit Units (ACCUs) or carbon dioxide equivalent (CChe).

[0045] According to the present disclosure, there is provided software that, when executed by a computer, causes the computer to perform the method of any one of the preceding embodiments, or part thereof.

[0046] According to the present disclosure, there is provided a system for determining a change in soil organic carbon quantity of a geographical area, the system comprising: a processor configured to: calibrate an in-silico model of the geographical area using measurements received from one or more sensors, the in-silico model being configured to simulate soil carbon sequestration within the geographical area, wherein the one or more sensors are placed at: a first location having environmental conditions similar to the geographical area; and / or a second location associated with the geographical area, the measurements being indicative of: (a) a level of carbon flux; and optionally (b) one or more environmental conditions; of the geographical area; provide input data indicative of current characteristics of the geographical area to the in-silico model; and determine the change in soil organic carbon quantity of the geographical area based on an output of the in-silico model.

[0047] According to the present disclosure, there is provided a system for determining a total carbon footprint of an agricultural area, the system comprising: a processor configured to: provide input data indicative of current characteristics of the agricultural area to an in-silico model of the agricultural area, wherein: the in-silico model is configured to simulate soil carbon sequestration within the agricultural area; the in-silico model is calibrated using measurements received from one or more sensors placed at: a first location having environmental conditions similar to the agricultural area; and / or a second location associated with the agricultural area; and the measurements are indicative of: (a) a level of carbon flux; and optionally (b) one or more environmental conditions; of the agricultural area; determine a change in soil organic carbon quantity of the agricultural area based on an output of the in-silico model; determine a livestock carbon footprint by applying a carbon accounting model to livestock data, the livestock data representing information of livestock within the agricultural area and the livestock carbon footprint being indicative of an output of the carbon accounting model; and determine the total carbon footprint of the agricultural area based on the change in soil organic carbon quantity and the livestock carbon footprint.

[0048] Optional features provided in relation to the any method, equally apply as optional features to any other method, the software and the systems. Brief Description of Drawings

[0049] An example will be described with reference to the following drawings:

[0050] Fig. 1 illustrates a system for determining a change in soil organic carbon quantity of a geographical area.

[0051] Fig. 2 illustrates a system for determining a total carbon footprint of an agricultural area.

[0052] Fig. 3 illustrates a method for determining a change in soil organic carbon quantity of a geographical area.

[0053] Fig. 4 illustrates a method for determining a total carbon footprint of an agricultural area.

[0054] Fig. 5 shows specific locations that flux towers were placed in Eastern Australia during the experiments described herein.

[0055] Fig. 6a shows the cumulative NEE from flux tower (solid black line) versus modelled output for Location C for the model calibrated using Brigalow Catchment Study data only.

[0056] Fig. 6b shows the cumulative NEE from flux tower (solid black line) versus modelled output for Location E for the model calibrated using Brigalow Catchment Study data only.

[0057] Fig. 7a shows the cumulative NEE from flux tower (solid black line) versus modelled output for Location C for the model with fine-tuned calibration using flux tower data.

[0058] Fig. 7b shows the cumulative NEE from flux tower (solid black line) versus modelled output for Location E for the model with fine-tuned calibration using flux tower data.

[0059] Fig. 8a shows evapotranspiration measured using a flux tower versus evapotranspiration determined using an in-silico model for Location C.

[0060] Fig. 8b shows net ecosystem exchange (NEE) measured using a flux tower versus NEE determined using an in-silico model for Location C.

[0061] Fig. 8c shows the soil carbon measured using a flux tower versus soil carbon determined by an in-silico model for the Location C.

[0062] Fig. 9a shows evapotranspiration measured using a flux tower versus evapotranspiration determined using an in-silico model for Location E.

[0063] Fig. 9b shows net ecosystem exchange (NEE) measured using a flux tower versus NEE determined using an in-silico model for Location E.

[0064] Fig. 9c shows the soil carbon measured using a flux tower versus soil carbon determined by an in-silico model for Location E.

[0065] Fig. 10 shows that observed (dots) and simulated (line) aboveground biomass C (t C ha-1) at the Brigalow Catchment Study site in northeast Australia.

[0066] Fig. 11 shows the observed (dots) and simulated (lines) soil organic carbon (SOC) stock in the 0-0.3 m soil depth (t C ha-1) in the total (red) and inert (blue) carbon pools at the Brigalow Catchment Study site in northeast Australia.

[0067] Fig. 12 shows the observed and simulated aboveground biomass C (t C ha-1) across the flux tower trials in northeast Australia. Dots and error bars indicate mean values and standard errors.

[0068] Fig. 13 shows the observed and simulated aboveground biomass N (kg N ha-1) across the flux tower trials in northeast Australia. Dots and error bars indicate mean values and standard errors.

[0069] Fig. 14 shows the observed and simulated weekly average actual evapotranspiration (AET) (mm) during calibration (top) and validation (bottom) periods across the flux tower trials in northeast Australia.

[0070] Fig. 15 shows the observed (solid lines) and simulated (dotted lines) cumulative actual evapotranspiration (AET) (mm) during calibration (top) and validation (bottom) periods across the flux tower trials in northeast Australia.

[0071] Fig. 16 shows the observed and simulated weekly average net ecosystem exchange (NEE) (g C m-2) during calibration (top) and validation (bottom) periods across the flux tower trials in northeast Australia.

[0072] Fig. 17 shows the observed (solid lines) and simulated (dashed lines) cumulative net ecosystem exchange (NEE) (t C ha-1) during calibration (top) and validation (bottom) periods across the flux tower trials in northeast Australia.

[0073] Fig. 18a shows the observed and simulated soil organic carbon (SOC) stock in the 0-0.3 m soil depth (t C ha-1) across the flux tower trials in northeast Australia, shown against 1:1 line.

[0074] Fig. 18b shows the observed and simulated soil organic carbon (SOC) stock in the 0-0.3 m soil depth (t C ha-1) across the flux tower trials in northeast Australia over time.

[0075] Fig. 19 shows the simulated soil organic carbon dynamics at flux sites (Location E in Central Queensland and Location C in South Queensland, Australia) with control, time-controlled grazing and legume incorporation grassland management strategies up to 2050 using recent climate conditions.

[0076] Fig. 20 shows the results from the calibrated in-silico model for different combinations of above ground biomass (AG) and below ground biomass (BG) increases.

[0077] Fig. 21 shows the cumulative net carbon position of an agricultural area in an experiment described herein.

[0078] Fig. 22 shows a quantification of the animal and soil carbon emissions with legume / grass pasture.

[0079] Fig. 23 shows a comparison of net ecosystem exchange (NEE) between a site where no practice change was implemented and a site where a practice change was implemented.

[0080] Fig. 24 shows the model simulated SOC versus the observation SOC data for the on-farm trials in the Cowra Trough located in central New South Wales (NSW).

[0081] Fig. 25 shows a comparison between simulated and observed SOC in the 0-30 cm topsoil across the Cowra Trough and OAI grazing trial sites in NSW.

[0082] Fig. 26 shows an example of the procedure for the initialisation of the in-silico model.

[0083] Fig. 27 shows an example of the procedure for the demonstration of the in-silico model.

[0084] Fig. 28 shows an example of scenario analysis conducted for the case study which compares the Soil Organic Carbon of Buffel and Desmanthus Legumes using long term weather patterns over a 25 year period.

[0085] Fig. 29 is an example of scenario analysis conducted for the case study which compares the above ground biomass of Buffel and Desmanthus Legumes using long term weather patterns over a 25 year period.

[0086] Fig. 30 illustrates an example of a method for determining a change in soil organic carbon quantity and / or a level of soil organic carbon of a geographical area.

[0087] Fig. 31 illustrates another example of a method for determining a change in soil organic carbon quantity and / or a level of soil organic carbon of a geographical area. Description of Embodiments

[0088] This disclosure provides methods and systems to address the issues of current soil coring methods. In particular, this disclosure provides a method to accurately predict and quantify any changes in soil carbon quantity and do this in a manner that is much lower cost than current methods. This method achieves this by using an in-silico model (i.e., a computer model) which is configured to simulate soil carbon sequestration (or soil carbon flux) within a geographical area. As such, a change in soil organic carbon quantity of the geographical area can be determined based on an output of the in-silico model. Most in-silico models which simulate soil carbon sequestration are also computationally efficient, meaning that minimal computing power and resources are required to operate such models.

[0089] However, some in-silico models are not always accurate. For example, some in-silico models tend to be accurate only for a particular climate (i.e., environmental conditions). In particular, some in-silico models are only accurate for the climate where they are developed, as the developers can only validate the models based on their local conditions. Such models may not necessarily translate to climates which are vastly different from the climate they were originally developed or based. For example, Australia is the driest of all inhabited continents, with considerable rainfall and temperature variability both across the country and from year to year. As such, some in-silico models may not be applicable for Australia’s harsh and variable climate. Therefore, there is a need for in-silico models which accurately or more accurately represent the climate of the geographical area being considered.

[0090] An in-silico model may be calibrated by using long-term soil organic carbon data, such as data of a geographical area spanning around 10, 15, 20, 25 or 30 years, for example. The long term data may provide an indication of carbon variability due to seasonal changes and climate change. However, it can be difficult to obtain such data given the long time span required. Further, the availability of long-term soil organic carbon data is often limited, particularly for new management practices, and also may not well represent the typically large spatial variation in pasture systems. Moreover, this data may only be available for a small number of geographical areas.

[0091] That being said, this disclosure shows that measurements received from one or more sensors may be used to calibrate the in-silico model and produce a high accuracy of a soil organic carbon prediction over time. In particular, the measurements may be used to calibrate and / or validate the in-silico model’s carbon cycling with its continuous carbon flux measurements. The disclosure also shows that calibration of the in-silico model can account for spatial heterogeneity to account for the lack of availability of long-term soil organic carbon data. This is a surprising and unexpected result, as one may expect that the long-term data would be sufficient to accurately calibrate a model. However, as will be discussed in this disclosure, these measurements from the one or more sensors significantly increase the accuracy of the in-silico model, while also being able to be gathered in a shorter time frame, compared to the long-term data. Such measurements can also be easier to obtain than long-term data using one or more sensors, compared to soil coring techniques, for example. Moreover, such measurements do not need to be obtained directly from the geographical location of interest, but rather the measurements may be associated with the geographical area and / or have similar environmental conditions to the geographical area, as will be discussed. This greatly increases the applicability and generality of the disclosed method and system, while enabling the in-silico model to be adapted to different and / or new geographical areas. Such calibration has never been done, particularly for a location with a harsh and variable climate such as Australia.

[0092] The disclosed methods and systems address these issues associated with long-term data by calibrating an in-silico model of the geographical area based on a level of carbon flux and one or more environmental conditions of the geographical area. In particular, measurements being indicative of a level of carbon flux of the geographical area; are used to calibrate the in-silico model. In some embodiments, the measurements are indicative of one or more environmental conditions of the geographical area. As such, the calibrated in-silico model accurately represents the geographical area under consideration. More particularly, the in-silico model can provide a real-time or periodic (e.g., daily, weekly, monthly or yearly) representation of the geographical area, thereby providing a real-time or periodic prediction of the change in soil organic carbon quantity of the geographical area, such as over a particular time period. As such, the disclosed methods and systems may provide a “rolling” real-time or periodic quantification of changes in soil organic carbon. In some embodiments, the in-silico model is initially calibrated using longterm study data, then further calibrated (or fined-tuned) using the local measurements corresponding to the geographical area. As will be shown in this disclosure, such further calibration with such measurements significantly increases the accuracy of soil organic carbon quantity in the geographical area.

[0093] Further data may be provided to the in-silico model to provide more information about the conditions and implemented practices of the geographical area. In particular, information regarding the soil within the geographical area may be used by the in-silico model to provide a more accurate representation of the soil carbon sequestration occurring within the geographical area. Moreover, weather can affect the soil carbon sequestration process, thereby affecting the change in soil organic carbon quantity. For example, rainfall can cause carbon to be transferred from the soil to the atmosphere, while wind can cause the carbon in the atmosphere to be transported to a different location. Moreover, water movement throughout the geographical area (through evapotranspiration, for example) can also affect soil sequestration and information regarding the water movement can also be provided to the in-silico model to accurately represent the soil sequestration of the geographical area.

[0094] By combining the accurately calibrated in-silico with a model of the carbon production of the livestock within a geographical area, the total carbon footprint of the geographical area can be determined. In particular, the calibrated in-silico can accurately determine the change in soil organic carbon quantity in the geographical area. This can be used as a real time indicator of carbon footprint that producers can share with their supply chain confirming that the meat or wool produced is quite possibly carbon neutral, for example. In other words, accurately quantifying soil carbon sequestration can assist in achieving carbon neutrality in agricultural areas, such as beef operations. This is because the disclosed methods can quantify a change in soil organic carbon quantity or a total carbon footprint in near real-time (as opposed to waiting 5, 10, 15 or 20 years to obtain the results from traditional soil coring methods on how much soil carbon has been sequestered).

[0095] The disclosed methods and systems may be integrated into a toolkit to provide soil carbon modelling with pasture and animal productivity data to a user, such as agricultural production managers like beef production managers, for example. This toolkit may provide producers with insights into feed availability, pasture dynamics, and stocking densities, aiding in profitability and sustainability. By predicting Average Daily Gains (ADG) and sales strategies, the toolkit can adapt to market changes while considering environmental factors like rainfall. The toolkit may also be tailored towards management goals, including carbon offsetting and market access. As such, the toolkit can provide indicators of profitability, sustainability, and environmental stewardship in a geographical area, which is beneficial for livestock farmers, for example.

[0096] Integrating with soil carbon management, the toolkit discussed above may provide indicators to producers enabling them to make informed decisions, optimise resource use and receive economic returns. The toolkit may highlight the point of diminishing returns in pasture utilisation, feed availability, and livestock management, offering insights that could also consider cattle prices for decision-making. By incorporating natural capital and economic factors, the toolkit enables producers to balance management, environmental, sustainability and profitability aspects in real-time and future scenarios.

[0097] For example, in Australian rangelands, managing uncertainties like rainfall timing is important to decision-making based on pasture and stock market prices. Pasture productivity directly impacts stock weight gains, ground cover, and soil health, all integral to effective and sustainable management. Understanding the point of diminishing returns across various fields helps optimise operations, ensuring wise investments while maintaining profitability. For instance, forecasting feed availability and livestock productivity over the next 14-28 days (or 90 days) aids in determining specific operational activities.

[0098] It is noted that the disclosed methods and systems are particularly directed towards determining (i.e., quantifying) “soil organic carbon”, which includes the once-living matter from plants, dead leaves, roots, and soil microbes. This is opposed to inorganic carbon, which is mineralbased and much less responsive to management. However, the disclosed methods and systems may also be applicable to determining (i.e., quantifying) inorganic carbon within a geographical area. Moreover, the disclosed methods and systems may also be applicable to determining (i.e., quantifying) total soil carbon within a geographical area, which includes both organic and inorganic carbon. That being said, this disclosure is directed towards implementing practice changes within a geographical area to modulate soil sequestration within the geographical area. As such, the disclosed methods and systems are preferably directed towards “soil organic carbon”, rather than inorganic carbon or total soil carbon.

[0099] The disclosed methods and systems may also provide the following benefits: • relatively low cost; • can provide forward-looking predictions tailored to specific contexts, accounting for anticipated conditions on the ground including weather and management. These predictions are useful for decision making e.g., producers choosing whether to participate in carbon farming; • can be used to make short-term predictions over timescales where direct sampling would not detect change, thereby supporting crediting on an annual basis; • can simulate what would happen without specific management interventions, while keeping all other conditions equal to those in the project simulation; and • can be used in strategic decision-making (e.g., informing on ideal project locations and scales) and are viewed as a cost-effective means to lessen or optimise soil sampling for direct field validation.

[0100] The disclosed methods and systems may also be used to determine changes in water quantities and the like in a geographical area, by using an in-silico calibrated to simulate evapotranspiration within the geographical area. Hence, disclosed methods and systems may also be used to determine water use efficiency of a geographical area. Some in-silico models may be capable of (or configured to) simulate soil sequestration and evapotranspiration and hence, may be used to determine a change in soil organic carbon quantity, a change in water quantity (such as a quantity of water stored in soil), or both.

[0101] Calibrating the in-silico according to the present disclosure may address the scarcity of available data in some geographical areas. Australia is one example of such geographical areas where long-term datasets for model calibration are scarce. The measurements from one or more sensors (such as high-resolution Eddy Covariance Flux Tower data, for example) may be short term data in comparison to long-term datasets, thereby addressing the scarcity of SOC datasets in regions such as Australia. In other words, the measurements from the one or more sensors may reduce in-silico model’s dependence on long-term datasets. As will be discussed later in this disclosure, the calibration of an in-silico model using the measurements from one or more sensors according to the present disclosure provides a robust model for Australian grazing systems, validated in both the calibration region (Brigalow bioregion) and in an independent, unseen area (South Western Slopes NSW bioregion) (RMSE < 10%). This demonstrated the calibrated model’s generalisation ability and applicability for many different geographical areas.

[0102] The use of the one or more sensors (such as Eddy Covariance Flux Towers, in some examples) can reduces spatial variability and thereby the uncertainty in soil carbon quantification. This means smaller soil carbon gains can be credited enabling more projects to become viable and facilitate greater market penetration. This also significantly reduces the need for intensive soil sampling to account for spatial heterogeneity, which is difficult in soil coring. It was shown that the uncertainty of results generated by the calibrated in-silico model according to the present disclosure, and often lower than, those associated with soil sampling. Specifically, the uncertainty around the soil sampling estimate was determined to be around ±111 C / Ha compared to ±4 t C / ha for the calibrated in-silico model, using a 95% confidence interval. Moreover, in some embodiments, the calibrated in-silico model represents a digital twin of the geographical area (such as a farming property), enabling the separation of climate and management effects which cannot be achieved through soil coring.

[0103] In some embodiments, the disclosed method may provide continuous, high-frequency data. Unlike soil coring, which provides a single snapshot in time (up to 5 years apart) at a specific point on the geographical area in question, the one or more sensors may operate continuously, capturing continuous data on carbon fluxes over a 20-50 ha area for a defined period in time. This approach enables better and faster model calibration, and reduced reliance on costly soil sampling that is vulnerable to spatial and temporal variation.

[0104] In some embodiments, the disclosed method may be highly sensitivity to changes in soil organic carbon quantity. In particular, the calibrated in-silico model may be highly sensitive and can detect small changes in carbon flux that will likely be missed by coarse or infrequent soil sampling. This contributes to a better understanding of why soil is gaining or losing carbon, which can be fed back to the farmer in near real-time. This level of understanding and actionable information cannot be achieved through soil coring or statistical models. System for determining a change in soil organic carbon quantity

[0105] Fig. 1 illustrates system 100 for determining a change in soil organic carbon quantity of geographical area 110. Fig. 1 is one example of a configuration of system 100. However, system 100 is not strictly limited to this configuration and this may be one possible embodiment of system 100. It is noted that system 100 of Fig. 1 is only meant to illustrate an example and a preferred system which is capable of performing the disclosed method.

[0106] Geographical area 110 may be any type of geographical region, such as an agricultural area (i.e., an area where agricultural activity occurs). For example, the geographical area may be a property such as a privately or publicly owned property, a facility such as a red meat production facility, a farm or another equivalent type of geographical area. However, geographical area 110 is not limited to being an owned property and may simply be a region of interest, for example. Although geographical area 110 is depicted as a rectangle in Fig. 1, geographical area 110 may be any type of shape, such as a polygon or abstract shape. Geographical area 110 may also be of any size.

[0107] System 100 comprises one or more sensors. The one or more sensors are placed at: a first location having environmental conditions similar to geographical area 110; and / or a second location associated with geographical area 110. A location associated with geographical area 110 may simply be a location within geographical area 110. A location having environmental conditions similar to geographical area 110 may also be a location within geographical area 110. In other examples, a location associated with geographical area 110 may be a location near or adjacent to (e.g., within less than about 10, 5, 2 or 1 kilometre, or any range therein, of) geographical area 110, but not necessarily within geographical area 110 (although the one or more sensors may be placed with geographical area 110). A location having environmental conditions similar to geographical area 110 may be a location that is distant (e.g., at least about 10, 20, 50, 100, 200, 500, 1000, 5000 or 10000 kilometres, or any range therein, distant) from geographical area 110. A location associated with geographical area 110 may have similar environmental conditions or similar carbon flux levels as geographical area 110. The term “environmental conditions” is used throughout the application to encompass a range of environmental or climatic conditions (e.g., ambient air temperature, wind conditions, barometric pressure, gas concentrations, annual rainfall levels, soil type, oxygen levels, humidity, pasture type), as are well known in the art. The term “similar environmental conditions” refers to a physical location where the one or more sensors may be placed to make measurements, such as of a level of carbon flux and / or one or more environmental conditions, and may refer to a single field or different fields, or other types of geographical areas or locations where one or more climatic conditions are comparable with the geographical area in question. However, both the first and second locations may be within geographical area 110. The one or more sensors may be configured to capture measurements indicative of: (a) a level of carbon flux; and (b) one or more environmental conditions; of geographical area 110. Each of the one or more sensors may be a location that is distant (e.g., at least about 10, 20, 50, 100, 200, 500, 1000, 5000 or 10000 kilometres, or any range therein, distant) from another one or more one or more sensors. The one or more sensors may be distributed at one or more locations based on a density of sensors to area, which may be optimised for calibration of the in-silico model. For example, there may be one or multiple sensors for about each 1 square kilometre. There may also be one or multiple sensors for about each 10, 20, 50,100, 200, 500, 1000, 5000 or 10000 square kilometres, or any range therein, or combination thereof.

[0108] In some examples, “carbon flux” may refer to the change in atmospheric carbon (such as carbon dioxide, methane etc). More specifically, “carbon flux” may refer to the change in local atmospheric carbon, localised about geographical area 110. For example, “carbon flux” may be the vertical turbulent flux of carbon within atmospheric boundary layers. The level of carbon flux indicated by the measurements captured by the one or more sensors may be an indicator of soil sequestration and hence, the measurements captured by the one or more sensors may be an indicator of a change in the soil organic carbon quantity. Locations that have similar environmental conditions and / or are associated with geographical area 110 may have similar levels of carbon flux as geographical area 110 and hence, the measurements from the one or more sensors at locations that are either similar and / or associated with geographical area 110 can indicate a level of carbon flux within geographical area 110.

[0109] In Fig. 1, system 100 comprises sensors 111, 112. However, it is noted that in some examples, system 100 may only comprise one sensor. Moreover, system 100 may comprises more than two sensors, in some examples. Sensor 111 and sensor 112 may be associated with geographical area 110. However, sensors 111, 112 may be associated with or at a location having environmental conditions similar to geographical area 110. Environmental conditions may include water quality, air quality, humidity, overall or average temperature and weather such as overall or average rainfall, or the like.

[0110] As depicted in Fig. 1, sensor 111 is placed outside of geographical area 110, whereas sensor 112 is placed within geographical area 110. As such, sensors 111, 112 are placed at a location associated with geographical area 110. As sensor 112 is placed within geographical area 110, sensor 112 is placed at a location having environmental conditions similar to geographical area 110. However, sensor 111 may not necessarily be placed at a location having environmental conditions similar to geographical area 110. In some examples, the in-silico model of the geographical area is initially calibrated using the measurements received by sensor 111 (e.g., at a first location distant from geographical area 110, but having similar environmental conditions) and then further calibrated using the measurements received by sensor 112 (e.g., at a second location within or adjacent geographical area 110).

[0111] Sensor 111 and sensor 112 may be the same type of sensor or different types of sensors. Sensors 111, 112 are configured to provide measurements being indicative of a level of carbon flux of geographical area 110. In some embodiments, the measurements are indicative of one or more environmental conditions of geographical area 110. In some embodiments, the measurements are indicative of a level of carbon flux; and one or more environmental conditions of geographical area 110. In some examples, sensor 111 may be configured to provide measurements being indicative of a level of carbon flux, and sensor 112 may be configured to provide measurements being indicative of one or more environmental conditions, or vice versa. Sensor 111 and sensor 112 may also capture other data such as weather conditions, precipitation, amount of sunlight, humidity, wind speed etc. Sensor 111 and sensor 112 may also measure high- frequency wind and scalar atmospheric data series, gas, energy, and momentum, which may be used to provide flux data. Sensor 111 and sensor 112 may measure the vertical exchange of CO2 and other trace gases, as well as energy and water vapor. Sensor 111 and sensor 112 may also be an ultrasonic anemometer or an infrared gas analyser, for example.

[0112] In some embodiments, one or both sensors 111, 112 are flux towers (or part of a flux tower). More specifically, sensor 111 and sensor 112 may be an Eddy Covariance Flux Tower. As such, sensor 111 and sensor 112 may be configured to detect turbulent eddies in the air surrounding geographical area 110. Turbulent eddies in the air surrounding geographical area 110 are swirling, chaotic motions of air that occur in turbulent atmospheric conditions. Turbulent eddies may cause turbulent flow of the air surrounding geographical area 110 (as referred to as simply as turbulence). Detecting the turbulent eddies may be used to measure the vertical exchange of CO2 and other trace gases, as well as energy and water vapor, for example. Flux towers may measure the net ecosystem exchange (NEE) and actual evapotranspiration (AET) of geographical area 110 which may be used to provide information regarding SOC dynamics by tracking carbon inputs (e.g., from photosynthesis) and outputs (e.g., respiration). In particular, flux towers may provide information indicative of small changes in soil carbon levels, without the cost of invasive soil testing and laboratory analysis, thereby providing data that may be more sensitive to minor soil organic carbon changes. In other words, flux towers systems are highly sensitive and can detect small changes in carbon flux that will likely be missed by coarse or infrequent soil sampling.

[0113] Flux towers may also continuously or periodically measure gas exchange between the land surface (soil and vegetation) and the atmosphere. Flux towers may continuously monitor greenhouse gas movement over extensive ranges, from hundreds of square meters to tens of square kilometres. Flux towers are advantageous as they enable ecosystem data to be collected in months rather than the decades it takes using soil coring, for example. Data from flux towers may also be used to validate the in-silico model after calibration.

[0114] Flux towers are a research tool for ecosystem monitoring, and can capture an immense amount of information, relatively cheaply, and over large areas. The latter is particularly useful for capturing variation at a paddock scale, encapsulating impacts of animal movement, plant species diversity and distribution and microrelief impacts on hydrology. However, translating this vast amount of data into usable metrics for monitoring agricultural sustainability may be difficult, especially when it is desirable to have these metrics periodically, such as on a “rolling” basis (i.e., periodically and continuously), or even in real-time, as this would require a vast amount of data processing. In some cases, the amount of data processing required to achieve the desired “rolling” basis, or real-time quantification, may require significant computing power and hence, may not be feasible, in terms of computing time and costs.

[0115] The systems and methods disclosed herein address this problem as the in-silico model can be calibrated on historical and / or current / contemporary flux tower measurements and hence, predictions regarding soil sequestration (such as a change in soil organic carbon quantity) can be made without having to continuously process the vast amount of data received from the flux towers. In essence, calibrating the in-silico model using the flux tower measurements provides information regarding soil sequestration without requiring the complex data processing. As such, the disclosed systems and methods can determine quantities, such as the change in soil organic carbon quantity, on a “rolling” or periodic basis, or even in real-time.

[0116] In some embodiments, sensors 111, 112 may be one of: a phenocam, a temperature sensor, a humidity sensor, a wind sensor, a rainfall sensor, an imaging sensor or another similar type of sensor or any combination thereof. In essence, sensors 111, 112 may be any type of sensor configured to provide measurements of an agricultural, environmental or another similar type of factor. Preferably, sensors 111, 112 may provide any one of the following measurements (or combination thereof) which can be used to calibrate the in-silico model. These measurements can be divided into three main categories: 1. Productivity: • Gross Primary Productivity (GPP); • plant production, which can be measured using the measurement of photosynthesis via the drawdown of CO2; • the total amount of carbon that enters the system. Stored as above-ground biomass (AGB = pasture shoots), below-ground biomass (roots) and soil carbon (through the transfer of shoots / roots to soil organic carbon); • total net primary productivity (NPP), which can be determined by measuring cumulative above ground biomass at certain points across a season, then partitioning the total NPP from other measurements into the above and below ground components; • pasture (shoot) biomass growth, which may be measured on a daily basis using sensors, such as flux towers and / or phenocams; • root growth, which can be predicted by combining this data with root / shoot ratios from soil coring; and • soil carbon (and carbon use efficiency), which can be given by subtracting shoots plus roots from total NPP. 2. Carbon: • carbon entering the soil via NPP, which needs to stay there to be classed as soil organic carbon; • Net Ecosystem Exchange (NEE), which is the balance between CO2 entering the ecosystem via photosynthesis, and out via ecosystem respiration. Ecosystem respiration can be broken into autotrophic (from plants) and heterotrophic (microbial breakdown of carbon for energy); and • Soil carbon sequestration, which can be given by partitioning NEE from CO2 stored in pasture shoots and roots. 3. Water • water input, which can be measured through a rain gauge and output through ecosystem transpiration and evaporation (i.e., evapotranspiration (ET)); • water stored in the soil, which can be measured as plant-available water content from soil moisture probes connected to the sensor, such as a flux tower; • the ratio between rainfall in and ET out (R / ETr), which is how much of the rainfall entered the soil and was used by the plants. This can be used to predict how much of the rainfall is being lost to surface runoff (a potential indication of low infiltration rates) or deep draining below the root zone. For example, overgrazing pastures can lead to shallow roots and low R / ETr, and deep-rooted legumes accessing moisture deeper in the profile will have a higher R / ETr; • rainfall efficiency, which can be calculated by comparing NPP with rainfall. This provides how many kg of biomass was grown per mm of rainfall. Overgrazing or shallow rooted pastures will reduce this number, which may limit plant access to water deeper in the profile or increasing runoff; • water-use efficiency (WUE) is the ratio between ET and NPP / NEE. This is how much water is lost per kg of carbon grown or sequestered. Large areas of bare soil / low ground cover or pasture in stage 1 growth can lead to high ET rates with little carbon sequestered (low ratios). Plants in growth stage 3 can remain green, photosynthesising and using water but not actively growing. Maintaining pasture in stage 2 can optimise WUE.

[0117] The measurements received by the one or more sensors described herein may be obtained over a set period of time, such as a sufficient amount of time in which to appropriately calibrate the in-silico model. By way of example, the measurements received by the one or more sensors may be obtained over a period of time of at least about 30 days (e.g., about or at least about 30, 50, 60, 90, 100, 120, 150, 180, 200, 210, 240, 250, 280, 300, 310, 340, 350, 365, 400, 450, 500, 550, 600, 650, 750, 800, 850, 900, 950 or 1000 days or any range therein), at least about 60 days, at least about 90 days, at least about 120 days, at least about 150 days, at least about 180 days, at least about 210 days, at least about 240 days, at least about 270 days, at least about 300 days, at least about 330 days or at least about 360 days. In particular examples, the measurements received by the one or more sensors are obtained over a period of time of between about 30 to about 500 days, between about 30 to about 1000 days, between about 40 to about 400 days, between about 50 to about 350 days, between about 60 to about 300 days, between about 70 to about 250 days, between about 80 to about 200 days, between about 100 to about 200 days, between about 100 to about 500 days or between about 100 to about 1000 days. More particularly, the measurements received by the one or more sensors may be obtained over at least 2, 3,4, 5 etc contiguous seasons (e.g., at least from the start of Spring to the end of Summer, at least from the start of Summer to the end of Autumn, at least from the start of Autumn to the end of Winter or at least from the start of Winter to the end of Spring).

[0118] It is noted that gross primary production (GPP) is the amount of chemical energy, typically expressed as carbon biomass, that primary producers create in a given length of time. Some fraction of this fixed energy is used by primary producers for cellular respiration and maintenance of existing tissues (i.e., “growth respiration” and “maintenance respiration”). The remaining fixed energy (i.e., mass of photo synthate) is referred to as net primary production (NPP), which is thereby given by: NPP = GPP - respiration [by plants]. Net primary production is the rate at which all the autotrophs in an ecosystem produce net useful chemical energy. Net primary production is available to be directed toward growth and reproduction of primary producers. As such it is available for consumption by herbivores. This disclosure also references net ecosystem production (NEP), which is the difference between gross primary production and total ecosystem respiration, represents the total amount of organic carbon in an ecosystem available for storage, export as organic carbon, or nonbiological oxidation to carbon dioxide through fire or ultraviolet oxidation.

[0119] System 100 comprises device 120, which may be a smartphone, computer, tablet, server device, or any other similar device. Device 120 comprises processor 121. Device 120 comprises memory 122, which comprises non-volatile memory 123 and / or volatile memory 124. Processor 121 may communicate with memory 122 by communicating with non-volatile memory 123 and / or volatile memory 124. Non-volatile memory 123 is a non-transitory computer readable medium and may be an optical disk drive, hard disk drive, solid-state drive, flash memory, storage server, cloud storage or another equivalent type of memory. Volatile memory 124 may be cache, RAM or another equivalent type of memory.

[0120] Memory 122 may store data to be retrieved for later use. For example, memory 122 may store parameters indicative of the in-silico model. In other words, memory 122 may store parameters which define the in-silico model. Memory 122 may also store measurements provided by sensors 111, 112. Memory 122 may store the output of the in-silico model. More specifically, memory 122 may store the change in soil organic carbon quantity provided by the in-silico model. The data thereof may be stored in memory 122 in the form of a JSON format file, XML format file or another equivalent data format file.

[0121] The methods described herein may comprise applying one or more trained machine learning models. These machine learning models may be stored on memory 122 by storing the weights that from the respective models, for example. Memory 122 may also store any output values calculated by processor 121 applying the one or more trained machine learning models, or any other variable or data necessary to perform such methods described herein.

[0122] Software, that is, an executable program stored on non-volatile memory 123 causes processor 121 to perform methods for determining a change in soil organic carbon quantity. While the singular of “processor” is used herein, it is meant to also encompass multiple processors that are individually or together configured (e.g., programmed) to perform the methods disclosed herein. As such, processor 121 may refers to multiple central processing units (CPUs) and / or graphical processing units (GPUs) that are configured to collectively perform the methods disclosed herein. Once executed, the software may cause processor 121 to calibrate an in-silico model using measurements received from one or more sensors, provide input data indicative of current characteristics of geographical area 110 to the in-silico model and determine the change in soil organic carbon quantity of geographical area 110.

[0123] Device 120 may be remotely located from sensors 111, 112. More specifically, device 120 may be at a location that is not associated with geographical area 110. Moreover, the device 120 may be at a location without similar environmental conditions to geographical area 110. As such, device 120 (more specifically, processor 121) may communicate with sensors 111, 112 via antenna 125. Antenna 125 may communicate with sensors 111, 112 through a wireless connection, such as by using a Wi-Fi network according to IEEE 802.11. The Wi-Fi network may be a decentralised ad-hoc network, such that no dedicated management infrastructure, such as a router, is required or a centralised network with a router or access point managing the network. In some examples, antenna 125 may communicate using a short-range wireless technology standard, such as Bluetooth. In some cases, device 120 may communicate with sensors 111, 112 using a wired connection, such as using Ethernet.

[0124] Device 120 (more specifically, processor 121) may also communicate with a server (not shown) to retrieve data or may retrieve data by communicating with an application programming interface (API) using a transmitter, transceiver or antenna 125 or a wired connection. For example, device 120 may communicate with the Australian Bureau of Meteorology (BoM) to retrieve weather data regarding geographical area 110 by interacting with the BoM’s API. Device 120 may also communicate with remote sensors, such as satellites (either directly or indirectly) to retrieve data, such as satellite imagery of geographical area 110.

[0125] The server may also perform some operations that can be performed by processor 121. For example, the server may store the in-silico model and the server may have one or more processors configured to provide input data to the in-silico model, such that the in-silico model generates an output on the server. As such, a user may upload an input data to device 120, processor 121 may then communicate the input data to the server, where the server provides the input data to the in-silico model (i.e., the server evaluates the in-silico model on the input data) to generate an output. The server may then communicate the output to processor 121. The server may also be configured to determine the change in soil organic carbon quantity based on the output generated by the in-silico model, for example. Processor 121 may communicate with the server via one or more APIs, for example.

[0126] System 100 comprises monitor 126, which may be configured to display information to a user. For example, monitor 126 may display the data relating to geographical area 110 graphically or the change in soil organic carbon quantity of geographical area 110. In another example, monitor 126 may display a graphical representation 127 of the in-silico model or aspects of the in-silico. In this way, processor 121 communicates the data to monitor 126 via the input / output (I / O) port 128. Monitor 126 may also display a satellite view of geographical area 110, such as the overall geographical region with geographical area 110 displayed (such as through an annotation, for example). Although I / O port 128 is shown as a single entity, it is to be understood that any kind of data port may be used to receive data, such as a network connection, a memory interface, a pin of the chip package of processor 121, or logical ports, such as IP sockets or parameters of functions stored on non-volatile memory 123 and executed by processor 121.

[0127] Software may provide a user interface that can be presented to the user on a monitor 126. The user interface is configured to accept input from the user, via a touch screen or a device attached to monitor 126 such as a keyboard or computer mouse. Device 120 may also include a touchpad, a connected touchscreen, a joystick, a button, and a dial. In an example, monitor 126 may display multiple instances of input data indicative of current characteristics, and the user may choose one of the multiple instances of input data for methods 300, 400 to be performed on, by processor 121. The user may also provide the input data indicative of current characteristics or other information regarding geographical area 110 (e.g., associated livestock data or information) by interacting with the touch screen or inputting the selection with a keyboard or computer mouse. The user input may also be provided to the I / O port 128 by the monitor 126. Users may also edit input data by interacting with the touch screen or inputting the selection with a keyboard or computer mouse.

[0128] The user interface may also provide users with other capabilities. For example, a user may display a satellite view of a geographical region and may select (or draw) geographical area 110 (i.e., the area that is of interest to the user). The user may also input geographical coordinates to define geographical area 110, such as the corners of geographical area 110. Selecting (or defining) geographical area 110 on the user interface may cause processor 121 to create location data regarding the location of geographical area 110, such as the geographical coordinates occupied by geographical area 110. Given the location data, weather data relevant to geographical area 110 may be obtained via an integration to publicly available daily data from the Bureau of Meteorology (BOM) and / or site-specific weather data, for example. Similarly, given the location data, soil chemical and physical property data relevant to the area selected may be obtained via an integration to relevant publicly available data. System for determining a total carbon footprint

[0129] Fig. 2 illustrates system 200 for determining a total carbon footprint of agricultural area 210. Fig. 2 is one example of a configuration of system 200. However, system 200 is not strictly limited to this configuration and this may be one possible embodiment of system 200. It is noted that system 200 of Fig. 2 is only meant to illustrate an example and a preferred system which is capable of performing the disclosed method. System 200 is similar to system 100. As such, elements of system 200 may be similar or equivalent to (or may perform similar or equivalent functions to) elements of system 100.

[0130] Agricultural area 210 may be a geographical region where agricultural activity occurs (e.g., cultivating soil, producing crops, and raising livestock and in varying degrees the preparation and marketing of the resulting products). Agricultural area 210 may be a property such as a privately or publicly owned property, a facility such as a red meat production facility, a farm or another equivalent type of agricultural area 210. In particular, agricultural area 210 comprises livestock 215. Livestock 215 are domesticated animals raised in an agricultural setting in order to provide labour and produce diversified products for consumption such as meat, eggs, milk, fur, leather, and wool. For example, livestock 215 may comprise sheep, goats, cattle, deer, buffalo, camelids or other similar types of animals.

[0131] System 200 comprises one or more sensors. The one or more sensors are placed at: a first location having environmental conditions similar to agricultural area 210; and / or a second location associated with agricultural area 210. The one or more sensors are configured to capture measurements indicative of: (a) a level of carbon flux; and (b) one or more environmental conditions; of agricultural area 210.

[0132] In Fig. 2, system 200 comprises sensors 211, 212. However, it is noted that in some examples, system 200 may only comprise one sensor. Moreover, system 200 may comprises more than two sensors, in some examples. Sensors 211, 212 are similar to sensors 111, 112 of system 100. However, sensors 211, 212 may also be configured to collected information or data from livestock 215. For example, sensors 211, 212 may monitor livestock 215. More specifically, sensors 211,212 may track when livestock 215 is within a certain region of agricultural area 210. In other examples, sensors 211, 212 may measure one or more properties of livestock 215. For example, sensors 211, 212 may be a weight sensor, such as a walking over weigh sensor. Moreover, sensors 211,212 may be capable of directly or indirectly measuring a level of methane production or output by the livestock 215. Sensors 211, 212 are suitably capable of determining an identity of an individual animal of livestock 215, such as by the presence of an RFID tag or similar associated therewith (e.g., in an ear tag).

[0133] System 200 comprises device 220, which is similar to device 120 in system 100. More specifically, device 220 comprises processor 221, memory 222, which comprises non-volatile memory 223 and / or volatile memory 224. Software, that is, an executable program stored on nonvolatile memory 223 causes processor 221 to perform methods for determining a total carbon footprint. Once executed, the software may cause processor 221 to provide an in-silico model of agricultural area 210, determine a change in soil organic carbon quantity of agricultural area 210, determine a livestock carbon footprint and determine the total carbon footprint.

[0134] Similar to device 120, device 220 comprises antenna 225 for remote (e.g., wireless) communication with sensors 211, 212 and other devices. System 200 comprises monitor 226, which may be configured to display the data relating to agricultural area 210 graphically or the change in soil organic carbon quantity of agricultural area 210, for example. In another example, monitor 226 may display a graphical representation 227 of the in-silico model or aspects of the in-silico. In this way, processor 221 communicates the data to monitor 226 via the input / output (I / O) port 228. Monitor 226 and I / O port 228 have similar functionality to monitor 126 and VO port 128 of Fig. 1, respectively. Method for determining a change in soil organic carbon quantity

[0135] Fig. 3 illustrates method 300 for determining a change in soil organic carbon quantity of geographical area 110. Fig. 3 is to be understood as a blueprint for a software program and may be implemented step-by-step, such that each step in Fig. 3 is represented by a function in a programming language, such as, but not limited to, Python, FORTRAN, C++ or Java. The resulting source code is then compiled and stored as computer-executable instructions on non-volatile memory 123, which causes processor 121 (or multiple processors or a distributed computing architecture) to perform method 300.

[0136] Processor 121 calibrates 301 an in-silico model of geographical area 110 using the measurements received from the one or more sensors (such as sensors 111, 112). The in-silico model is configured to predict soil carbon sequestration within geographical area 110. In other words, the in-silico model is configured to simulate soil carbon sequestration within geographical area 110. As will be appreciated, “in-silico” refers to a process performed by a computer (more specifically, a computer processor). As such, “in-silico model” refers to a model (or representation) provided by a computer (more specifically, a computer processor), as opposed to physical model, for example. In particular, the in-silico model may be a digital representation of geographical area 110. For example, the in-silico model may be simulation of geographical area 110, which may simulate the carbon flux and / or environmental conditions over a period of time. In essence, the in-silico model is a predictive model of geographical area 110.

[0137] In some embodiments, the in-silico model is indicative of a history and management of geographical area 110. In some examples, data regarding information about the history and management of geographical area 110 (e.g., historical data) may be provided to the in-silico model. In other words, the input data provided to the in-silico model may comprises data indicative of history and management of geographical area 110. For example, the input data may comprise satellite imagery of geographical area 110. In particular, the input data may comprise satellite imagery of geographical area 110 over time, which provides an indication of the history of geographical area 110. In another example, the input data may comprise records of the management of geographical area 110.

[0138] In some embodiments, the in-silico model is based on a biogeochemical model. More specifically, the in-silico model may be based on a time series biogeochemical model. As such, the in-silico model may simulate fluxes of carbon and / or nitrogen between the atmosphere, vegetation, and soil over time. The in-silico model may also simulate the fluxes of water between the atmosphere, vegetation, and soil (i.e., simulates evapotranspiration) over time. In some embodiments, the in-silico model may comprise a weather model that simulates the weather within geographical area 110 based on historical weather data.

[0139] In a specific example, the in-silico model may be a digital twin of geographical area 110. In this context, a digital twin is a digital representation of geographical area 110 contextualized in a digital version of its environment. The in-silico model may be based on DayCent (Daily Century model). DayCent is a daily time series biogeochemical model used in agroecosystems to simulate fluxes of carbon and nitrogen between the atmosphere, vegetation, and soil to produce daily fluxes of nitrogen gases, CO2 flux from heterotrophic soil respiration, soil organic carbon and nitrogen, net primary production, and H2O and NOs' leaching. The in-silico model may be a particular version of DayCent. For example, the in-silico model may be DayCent CABBI (Center for Advanced Bioenergy and Bioproducts Innovation). CABBI researchers have modified DayCent to add a “GrassTree” crop type to better represent large bioenergy crops and grasses in the model. These can be either annual or perennial and leaves and stems can be managed separately, so representing crop harvest is more flexible. DayCent-CABBI is useful as it can be effectively calibrated using Eddy Covariance (EC) flux data; namely net ecosystem exchange (NEE) and actual evapotranspiration (AET). However, other in-silico models may be used in some embodiments including, but not limited to, the Rothamsted Carbon Model (RothC), the DeNitrification-DeComposition (DNDC) model and other variations of DayCent.

[0140] In some embodiments, the in-silico model may be a mathematical model. In particular, the in-silico model may be a mathematical model defined by parameters. The parameters may be specific to geographical area 110. In some examples, the in-silico model may be a machine learning model. A machine learning model is understood to be a model that receives input and generates an output based on the input. The machine learning model may be of an architecture, such as, but not limited to, a neural network, for example. In general, machine learning models are ‘trained’ to leam and recognise patterns in an input and provide an output that is a prediction based on the training it has undergone. Training involves updating weights or parameters (as referred to as hyperparameters) of the machine learning model, which define the machine learning model, to minimise a loss value, thereby creating a trained machine learning model (in other words, a machine learning model trained to generate an output). This may involve a gradient descent and backpropagation method.

[0141] Processor 121 may calibrate the in-silico model by modifying the parameters that define the in-silico model. More specifically, processor 121 may modify the parameters that define the in-silico model based on the received measurements. In essence, calibrating 301 an in-silico model of geographical area 110 using the measurements enables geographical area 110 to be more accurately represented by the in-silico model, as the climate can be accounted for by calibrating the model. As the in-silico model is configured to simulate soil carbon sequestration within geographical area 110, calibrating 301 the in-silico model provides a more accurate prediction of soil carbon sequestration or loss, such as that occurring over time with changes in environmental conditions and management practices (e.g., livestock). Processor 121 may calibrate the in-silico model by performing an algorithm or mathematical operation. Processor 121 may also calibrate the in-silico model by machine learning techniques, such as, but not limited to, cross validation, gradient descent and backpropagation.

[0142] For example, processor 121 may calibrate the in-silico model by modifying the parameters based on predetermined parameters. The predetermined parameters may be based on the received measurements from the one or more sensors. In other words, the in-silico model may be parameterised by these predetermined parameters. In this sense, the predetermined parameters may essentially be input into the in-silico model. The predetermined parameters may be stored on memory 122 or may be stored externally, such as on a database or external server. The predetermined parameters may be based on a “look-up” table, for example. The predetermined parameters may be parameters of the in-silico model which best represent measurements or experimental data.

[0143] It is noted that calibrating 310 the in-silico model may involve recalibrating an in-silico model. For example, an in-silico model may be provided which is already calibrated to some extent. Calibrating 301 the in-silico model may thus be re-calibrated or further calibrating the in-silico model, such as fine tuning the calibration of the in-silico model. In some examples, processor 121 may calibrate 310 the in-silico model by calibrating the in-silico using long-term study data. Alternatively, or additionally, processor 121 may calibrate 310 the in-silico model by calibrating the in-silico model (e.g., fine-tuning the in-silico model) using measurements from the one or more sensors. The measurements from the one or more sensors may be based on flux tower measurements, for example.

[0144] Processor 121 may receive the measurements from one or more sensors directly or indirectly, which are then used to calibrate 301 the in-silico model. For example, the measurements may be stored on memory 122 and processor 121 may receive the measurements by retrieving the measurements from memory 122. The measurements may also be stored on an external (remote) server and processor 121 may receive the measurements via antenna 125. The measurements may correspond to a net flux of carbon e.g., carbon dioxide and methane, in geographical area 110. It is noted that “carbon flux” may not necessarily be greenhouse gases which contain carbon. In some embodiments, “carbon flux” may refers to the flux of one or more greenhouse gases, including carbon dioxide, methane, nitrous oxide, fluorinated gases and water vapour. Environmental conditions may comprise air quality, and weather including rainfall, sunlight (e.g., cloud coverage, UV index), humidity, temperature and other similar factors. The measurements may also be indicative of net eco exchange (NEE) and actual evapotranspiration (AET). Processor 121 may also receive the measurements directly from the one or more sensors, as the measurements (data) is generated, for example.

[0145] Processor 121 then provides 302 input data indicative of current characteristics of geographical area 110 to the in-silico model. In other words, processor 121 may evaluate the in-silico model on the input data indicative of current characteristics of geographical area 110. Providing 302 the input data to the in-silico model and / or evaluating the in-silico model on the input data may be considered to be invoking, initialising, calling, applying or the like. Processor 121 may generate an output upon providing 302 input data to the in-silico model. The output of the in-silico model may be a value indicative of soil carbon sequestration (or losses) and / or current soil carbon levels in geographical area 110. More specifically, the output of the in-silico model may be indicative of a prediction of soil carbon sequestration (or losses) and / or current soil carbon levels.

[0146] The current characteristics are indicative of the characteristics of geographical area 110 at the time where the change of the soil organic carbon quantity or level is determined. As such, the change of the soil organic carbon quantity that is determined using method 300 is indicative of the current characteristics of geographical area 110 and their estimated or predicted effect on changes in soil carbon levels as indicated by the in-silico model. The current characteristics may be current soil characteristics, current weather such as current temperature, humidity, rainfall, or the like. It should be noted that “current” is subjected to slight variation. As such, the input data may not necessarily represent the exact current characteristics of geographical area 110. For example, there may be a slight change in weather from the time that the current characteristics are measured to the time that the current characteristics are provided as input to the in-silico model. However, the measured characteristics can still be considered as current characteristics in this situation.

[0147] In some examples, processor 121 may provide input data indicative of characteristics of geographical area 110 to the in-silico model, where the characteristics are captured at a time point other than the current time point (e.g., at the time when the in-silico model is initialised). For example, processor 121 may provide input data indicative of previous characteristics of geographical area 110 to the in-silico model. A user may also wish to make a prediction based on characteristics they could implement in the future. As such, processor 121 may provide input data indicative of future characteristics of geographical area 110 to the in-silico model. The remainder of method 300 may then be carried out using the specified characteristics to make a prediction with regards to the specified characteristics.

[0148] The input data provided to the in-silico model may comprise a variety of different data indicative of current characteristics of geographical area 110. For example, the input data may comprise image data of geographical area 110. In another example, the input data may comprise soil data indicative of information regarding the soil of geographical area 110. In a further example, the input data may comprise weather data indicative of information regarding the weather (such as current or historical weather conditions) of geographical area 110. In other examples, the input data may comprise schedule data indicative of information regarding scheduled events within geographical area 110, such as such as planting, growth start and end, fertilization, and the like.

[0149] There may be a number of different ways to invoke the in-silico model, which all fall within the meaning of “providing input data to the in-silico model” as recited throughout this disclosure. For example, providing input data to the in-silico model (or evaluating the in-silico model on the input data) may involve calling an API routine to send the input data to a server and the server then performs the calculations according to the in-silico model and returns the results.

[0150] In other examples, “providing input data to the in-silico model” may involve issuing a command to local hardware, such as a local chip or device, which has the in-silico model stored thereon and provides a command interface to interact with the model. Integrated circuits, such as FPGAs, can be used where flexibility, speed, and parallel processing capabilities are desired. In such an embodiment, the integrated circuit may be part of device 120 of system 100 and may be considered as a “processor” or “processing unit”, similar to processor 121. Other implementations, such as application specific integrated circuits (ASIC) are equally useable. It is also possible to have a local copy of the in-silico model available so that the calculations are performed by the main processor of the local machine. Other local, remote or distributed implementations (such as cloud computing environments) are equally useable.

[0151] Processor 121 then determines 303 the change in soil organic carbon quantity of geographical area 110 based on an output of the in-silico model. Determining 303 the change in soil organic carbon quantity may be considered to be quantifying, predicting, estimating or the like. In some examples, the change in soil organic carbon quantity may correspond exactly to the output of the in-silico model. In other examples, processor 121 may apply additional processing to the output of the in-silico model (such as applying an algorithm or mathematical operation to the output) to determine 303 the change in soil organic carbon quantity. The change in soil organic carbon quantity may be indicative of the change in soil organic carbon quantity between the data used to last calibrate the in-silico and the time when the current characteristics of geographical area 110 were measured.

[0152] In some embodiments, after calibrating 301 the in-silico model, processor 121 may receive further measurements from one or more sensors (e.g., sensors 111,112). For example, the further measurements may be similar to the measurements used to calibrate the in-silico (i.e., further measurements may be indicative of: (a) a level of carbon flux; and (b) one or more environmental conditions; of geographical area 110). As such, processor 121 may re-calibrate (or update) the in-silico model after receiving the further measurements. More specifically, processor 121 may repeat the steps of calibrating 301 the in-silico model, providing 302 input data to the in-silico model and / or determining 303 the change in soil organic carbon quantity upon receiving further measurements from the one or more sensors. In particular, processor 121 may update the in-silico model upon receiving the further measurements such that the in-silico model is updated and hence an updated change in soil organic carbon quantity is provided in real-time. As such, the change in soil organic carbon quantity can determined in real time. More particularly, the change in soil organic carbon quantity, inclusive of the level of soil organic carbon, can determined on a rolling, continuous or periodic (e.g., daily, weekly, monthly or yearly) basis.

[0153] Processor 121 may also receive further measurements periodically. For example, processor 121 may also receive further measurements on a daily basis, such as at the same time each day. As such, processor 121 may determine the change in soil organic carbon quantity of geographical area 110 on a “rolling basis”. As such, processor 121 may repeat the steps of calibrating 301 the in-silico model, providing 302 input data to the in-silico model and / or determining 303 the change in soil organic carbon quantity upon receiving further measurements from the one or more sensors, such that the change in soil organic carbon quantity is quantified on a rolling basis.

[0154] In some embodiments, the input data comprises soil data corresponding to information regarding soil within geographical area 110. In other words, processor 121 provides 302 input data to the in-silico model by providing the soil data to the in-silico model. In one example, processor 121 may apply (or evaluate) the in-silico model on the soil data. In a further example, processor 121 may modify the parameters that define the in-silico model based on the soil data. As such, the determined change in soil organic carbon quantity is based on information regarding the soil in geographical area 110. Processor 121 may receive the soil data of geographical area 110 based on a result of soil testing in the geographical area. For example, the soil testing may be soil coring and hence, the soil data may be based on the results of the soil coring. Soil data may also be obtained from a publicly available database such as CSIRO’s Soil Landscape Grid of Australia or site-specific sources. As such soil testing, such as soil coring testing, may be carried out so as to assist with calibration of the in-silico model described herein. By way of example, soil testing may be utilised to determine a baseline reference level of soil carbon and then optionally periodically thereafter (e.g., every 0.5, 1, 2, 5, 10, 15 or 20 years or any range therein) as required. Such subsequent soil testing may be utilised to facilitate confirming the accuracy of the in-silico model and / or allow for any re-calibrating or further calibrating thereof as required. In some examples, processor 121 may determine the input data using an application programming interface (API). For example, processor 121 may use the Australian National Soils Information System (ANSIS), or the Terrestrial Ecosystem Research Network (TERN) to determine input data, or part thereof. Processor 121 may also determine input data from the Bureau of Meteorology (BoM) using an API, for example.

[0155] The soil data may comprise information about one or more properties or characteristics of the soil in geographical area 110. For example, the soil data may comprise information regarding one or more of: soil type, soil quality and soil moisture. More specifically, the information regarding soil within the geographical area comprises information indicative of one or more of: soil texture; soil pH; and soil bulk density. These soil characteristics can affect to soil sequestration process for carbon and hence, incorporating the soil data into the in-silico model can provide a more accurate determination of the change in soil organic carbon quantity, and more particularly the level of soil organic carbon, of geographical area 110.

[0156] In some embodiments, processor 121 may receive image data of geographical area 110. For example, processor 121 may receive satellite image data corresponding to a “birds’ eye” view of geographical area 110. The image data may capture the soil in geographical area 110 and / or the vegetation / plants in geographical area 110. For example, processor 121 may receive the image data from a phenocam or similar placed within geographical area 110. The soil data may be based on the image data. For example, the image data may provide an indication of one or more soil characteristics, such as soil quality. Processor 121 may extract the soil data from the image data. For example, processor 121 may extract the soil data by applying an algorithm, mathematical model or a machine learning model to the image data. The image data may be stored on memory 122 as Joint Photographic Experts Group (JPEG) format, RAW image format or a similar / equivalent image format, for example. In some examples, the image data may be an RGB image, a multispectral image, a hyperspectral image, an infrared image, a point cloud or a similar / equivalent image type.

[0157] In some embodiments, the input data comprises plant data representing information regarding plants within geographical area 110. The plants in geographical area 110 may include trees, bush, grass, crops, pasture (i.e., plants that are grazed by domesticated livestock) or the like. In essence, the reference to “plants” in this disclosure may refer to any organic matter capable of performing photosynthesis. Similar to providing the in-silico model based on soil data, the plant data may be provided as input into the in-silico model. In another example, processor 121 may apply the in-silico model on the plant data. In a further example, processor 121 may modify the parameters that define the in-silico based on the plant data. As such, the determined change in soil organic carbon quantity is based on information regarding the plants in geographical area 110.

[0158] The plant data may comprise information about one or more properties or characteristics of the plants in geographical area 110. For example, the plant data may comprise information regarding one or more of: plant types, plant quality, plant quantity or plant coverage. These plant characteristics can affect the soil sequestration process and hence, incorporating the plant data into the in-silico model can provide a more accurate determination of the change in soil organic carbon quantity of geographical area 110. Processor 121 may extract the plant data from the image data which captures geographical area 110 and, more specifically, captures the plants within geographical area 110. For example, processor 121 may extract the plant data by applying an algorithm, mathematical model or a machine learning model to the image data.

[0159] In some embodiments, the input data comprises pasture data representing information regarding pasture within geographical area 110. Pasture may include grasses and legumes, both native and cultivated, that are feed for livestock (ruminants) in geographical area 110 such as cattle, horses, sheep, and goats. The information regarding pasture within geographical area 110 may include information indicative of pasture quality, pasture type (i.e., pasture composition), pasture quantity, pasture growth rate, and the like. Processor 121 may extract the pasture data from the image data which captures geographical area 110 and, more specifically, captures the pasture within geographical area 110. For example, processor 121 may extract the pasture data by applying an algorithm, mathematical model or a machine learning model to the image data.

[0160] In some embodiments, the input data comprises weather data and / or rainfall data. For example, the measurements received from the one or more sensors correspond to weather data representing information regarding weather conditions within geographical area 110, and rainfall data representing information regarding rainfall within the geographical area. For example, the one or more sensors (i.e., sensors 111, 112) may be a rain gauge, humidity sensor, wind sensor, temperature sensor or combination thereof. Hence, the measurements received from the one or more sensors may correspond to rainfall, humidity, wind speed and wind direction or temperature. In other examples, processor 121 may receive weather data or rainfall data via an integration to publicly available daily data, such as from the Bureau of Meteorology (BOM) and / or site-specific weather data.

[0161] In some embodiments, processor 121 may receive the weather data and / or the rainfall data periodically, such as on a daily, weekly or monthly basis. Processor 121 may determine 303 the change in soil organic carbon quantity of geographical area 110 upon receiving the weather data and / or the rainfall data and hence, processor 121 may determine 303 the change in soil organic carbon quantity periodically, such as on a daily basis. In other embodiments, processor 121 may receive the weather data and / or the rainfall data in real-time and hence, processor 121 may determine 303 the change in soil organic carbon quantity in real-time based on the real-time weather data and / or rainfall data.

[0162] Some environmental factors may be stochastic and unpredictable. In particular, the effects of the climate change may affect some environmental factors such as, but not limited to, precipitation and temperature. Hence, in some embodiments, processor 121 may model some stochastic, unpredictable or random factors using an algorithm, which may then be used an input to the in-silico model. For example, processor 121 may model annual precipitation and / or annual temperature using an algorithm to model random nature of these environmental factors resulting from climate change. The algorithm may be a Monte-Carlo algorithm, or another type of stochastic or randomisation algorithm.

[0163] In some embodiments, processor 121 may determine 303 a change in soil organic carbon quantity by determining a level of soil organic carbon of the geographical area. For example, processor 121 may determine a level of soil organic carbon (a quantity of soil organic carbon) at two different time points using, at least part of, method 300. Processor 121 may determine 303 the change in soil organic carbon quantity based on the level of soil organic carbon for both time points, using a difference , ratio or another mathematical formula. In other embodiments, processor 121 may determine a level of soil organic carbon, rather than determining a change. As such, method 300 may also be considered to be a method for determining a level of soil organic carbon. Method for determining a total carbon footprint

[0164] Fig. 4 illustrates method 400 for determining a total carbon footprint of agricultural area 210. Fig. 4 is to be understood as a blueprint for a software program and may be implemented step-by-step, such that each step in Fig. 4 is represented by a function in a programming language, such as, but not limited to, Python, FORTRAN, C++ or Java. The resulting source code is then compiled and stored as computer-executable instructions on non-volatile memory 223, which causes processor 221 (or multiple processors or a distributed computing architecture) to perform method 400. It is noted that embodiments described above in relation to method 300 may also be embodiments of method 400.

[0165] A carbon footprint is a calculated value or index that makes it possible to compare the total amount of greenhouse gases that an activity, product or company adds to the atmosphere. Carbon footprints are usually reported in tonnes of emissions (CCh-equivalent) per unit of comparison. Such units can be for example tonnes CCh-eq per year, per kilogram of protein for consumption, per kilometre travelled, per piece of clothing and so forth. However, it is noted that “carbon footprint” may not necessarily refer to the footprint of greenhouse gases which contain carbon. In some embodiments, “carbon footprint” may refers to the flux of one or more greenhouse gases, including carbon dioxide, methane, nitrous oxide, fluorinated gases and water vapour. As such, carbon footprint may also be referred to as “greenhouse gas footprint”. The reference to “total carbon footprint” is a combination of sources or sinks of carbon (e.g., the change in soil organic carbon quantity and the carbon footprint resulting from livestock). However, the reference to “total carbon footprint” in this disclosure may or may not necessarily account for all possible carbon sources or carbon sinks within agricultural area 210.

[0166] Processor 221 provides 401 input data indicative of current characteristics of agricultural area 210 to an in-silico model of agricultural area 210. The in-silico model is similar or equivalent to the in-silico described with reference to method 300. In particular, the in-silico model in method 400 is configured to simulate (e.g., predict) soil carbon sequestration, and thus soil carbon levels, within agricultural area 210. Moreover, the embodiments discussed above in relation to the in-silico of method 400 are also embodiments of the in-silico model of method 400.

[0167] The in-silico model is also calibrated using measurements received from one or more sensors (such as sensors 211, 212) placed at a first location having environmental conditions similar to agricultural area 210; and / or a second location associated with agricultural area 210. The measurements are indicative of: (a) a level of carbon flux; and optionally (b) one or more environmental conditions; of agricultural area 210. As such, step 401 of method 400 is similar or equivalent to step 302 of method 300. Processor 221 then determines 402 a change in soil organic carbon quantity of agricultural area 210 based on an output of the in-silico model. Step 402 of method 400 is similar or equivalent to step 303 of method 300.

[0168] Processor 221 then determines 403 a livestock carbon footprint by applying a carbon accounting model to livestock data. The livestock data representing information of livestock 215 within agricultural area 210. Further, the livestock carbon footprint is indicative of an output of the carbon accounting model. The carbon accounting model provides a prediction of the carbon footprint produced by livestock 215 from their grazing behaviour, live weight gain and excreta, for example. The carbon accounting model may be an algorithm or mathematical model or combination thereof. The carbon accounting model may also be a trained machine learning model. As such, the carbon accounting model may be defined by parameters and hence, the carbon accounting model may be stored on memory 222 by storing the parameters that define the model.

[0169] The greenhouse gas emissions generated by livestock production and other farm-related operations can be measured by conducting a “carbon account”, which may be provided by the carbon accounting model. A carbon account allows producers to calculate their current greenhouse gas emissions and also helps them to understand how greenhouse gases and carbon management can impact enterprise productivity. As such, the carbon accounting model is a calculation tool for modelling greenhouse gas emissions from an agricultural area. The information that may be used to perform a carbon account for livestock (i.e., information that may be input into the carbon accounting model) may include: a livestock inventory (births, deaths, sales, purchases, weights and weight gain, pregnancy status), an inventory of purchased inputs, carbon in vegetation and (potentially) soil, or the like. An example of a model that can be used as the carbon accounting model described herein is the Full Carbon Accounting Model (FullCAM) fuHcam), which is incorporated by reference in its entirety herein. In particular, FullCAM may also be used to generate abatement estimates for vegetation Methodology Determinations (methods) under the Australian Carbon Credit Units (ACCU) program, previously known as the Emissions Reduction Fund (ERF).

[0170] Processor 221 then determines 404 the total carbon footprint of agricultural area 210 based on the change in soil organic carbon quantity and the livestock carbon footprint. Determining 404 the total carbon footprint may be considered to be quantifying, predicting, estimating or the like. Processor 221 may apply an algorithm or mathematical operation to the change in soil organic carbon quantity and the livestock carbon footprint to determine 404 the total carbon footprint. For example, processor 221 may simply add the change in soil organic carbon quantity to the livestock carbon footprint to determine 404 the total carbon footprint. In another example, processor 221 may calculate a weighted sum of the change in soil organic carbon quantity and the livestock carbon footprint to determine 404 the total carbon footprint. In some examples, further input data may be used to quantify the soil carbon sequestration. This further data may include farm management data e.g., what has farmer done on the area. For example, the farm management data may include data indicative of mechanical movement of soil (e.g., blade ploughing).

[0171] In some embodiments, processor 221 may determine a carbon footprint of a product derived from livestock 215 from agricultural area 210 based, at least in part, on the total carbon footprint of agricultural area 210 over a particular period of time. A user (rather than processor 221) may also determine a carbon footprint of a product derived from livestock 215 from agricultural area 210 based, at least in part, on the total carbon footprint of agricultural area 210, such as over that period of time required to generate, grow or produce the product in question. The carbon footprint of the product may be similar or equivalent to the total carbon footprint of agricultural area 210. Processor 221 may apply an algorithm or mathematical operation to the total carbon footprint to determine the carbon footprint of the product. For example, processor 221 may apply a factor (such as a multiplicative factor) to the total carbon footprint to determine the carbon footprint of the product. In some embodiments, the product is one or more of: wool, meat, and milk.

[0172] In some embodiments, the livestock data comprises data indicative of one or more of: total livestock numbers; species of livestock; livestock added; livestock removed; livestock weights, including liveweight gain over time; wool shorn; livestock birthing rates; or the like. The livestock data may also include date of cattle entry and exit from paddock, class of cattle, liveweight gains whilst in paddock if known). In some embodiments, the livestock data comprises liveweight of at least one animal in livestock 215. Liveweight refers to the weight of an animal. More specially and with reference to cattle farm, liveweight refers to the weight of an animal (such as cattle) before it has been slaughtered and prepared as a carcass. In some embodiments, processor 221 may receive the liveweight of the at least one animal in livestock 215 from a sensor in agricultural area 210. In other words, a sensor (such as sensor 211, for example) may measure the liveweight of an animal in livestock 215 and transmit the measurement to processor 221.

[0173] In some embodiments, the liveweight of the at least one animal is based on walk over weighing. For example, the sensor (such as sensor 211, for example) may which measures the liveweight may be walk over weighing sensor. A walk over weighing sensor enables easy weigh measurements of animals on a regular basis. Animals voluntarily weigh themselves by walking over the scales, with no intervention required. In some embodiments, the livestock data comprises liveweight gain of at least one animal in livestock 215. Processor 221 may determine the liveweight gain based on the liveweight of livestock 215 at two different points in time. For example, processor 221 may determine the liveweight gain by calculating the liveweight gain using a ratio of the difference between the liveweight and the difference of the time between the liveweight measurements.

[0174] In some embodiments, processor 121 may monitor a selection of animals in livestock 215 to determine at least part of the livestock data in real time. The selection of animals in livestock 215 may be a subset of animals in livestock 215 or may be all animals in livestock 215. In some examples, monitoring each animal in the selection of animals in livestock 215 is based on a radio frequency identification device (RFID) placed on each animal. The RFID may comprise a sensor that monitors data of the animal and transmits the data to processor 221 via antenna 225, for example. The sensor may track properties / animals across their properties. The data received from the sensor may be indicative of movement data. As such, the sensor may be one or more accelerometers, for example. The sensor may also be a location sensor, such as a global positional system (GPS) device. As such, the data transmitted to processor 221 may comprise location data of the animals in the selection of animals. The location data may be used to track animal movements from property to property or from paddock to paddock within agricultural area 210, for example.

[0175] In some embodiments, processor 221 may determine 404 the livestock carbon footprint by applying the carbon accounting model to one or more of: plant data representing information regarding plants within agricultural area 210; fertiliser data representing information regarding fertiliser use within agricultural area 210; energy consumption data; and fuel consumption data (such as fuel consumed by vehicles used within agricultural area 210). This provides a total carbon footprint that represents all (or most) possible sources of carbon and all (or most) possible carbon sinks within agricultural area 210. In some examples, processor 221 may determine 404 the livestock carbon footprint based on a tree carbon number. The tree carbon number may be based on a number of trees and / or type of trees planted on agricultural area 210. Processor 221 may perform a tree carbon calculation (e.g., by using a tree carbon model such as FullCAM).

[0176] In some embodiments, processor 221 may generate a price grid matrix based on the total carbon footprint of agricultural area 210. A price matrix is a strategic tool used in pricing decisions, enabling businesses to determine prices based on varying criteria such as quantity, customer type, or product version. It is a framework that systematically organizes different price points for products or services, often displayed in a grid format. In some examples, processor 221 may generate a Carbon Neutral (CN) Price Grid Matrix, which can be used by producers, processors and retailers can purchase certified carbon neural meat or other products that are derived from livestock 215 in agricultural area 210.

[0177] In some embodiments, processor 221 may generate a report containing one or more of: the total carbon footprint of agricultural area 210; the change in soil organic carbon quantity; the livestock carbon footprint; the carbon footprint of the product (i.e., a product derived from livestock 215 from agricultural area 210); and a further carbon footprint based on other sources of carbon within agricultural area 210 (e.g., vehicles, fertiliser, electricity and plants such as trees). The report may also comprise one or more visual representations of the mentioned quantities, such as a chart, graph or the like.

[0178] The report may also be provided with data insights such as trends observed or characteristics of the mentioned quantities. Processor 221 may generate the trends observed or characteristics of the mentioned quantities by applying an algorithm or a mathematical model / operation, such as a trained machine learning model, on the mentioned quantities and generate an output indicative of the trends observed or characteristics of the mentioned quantities. The report may also be provided with the uncertainty of the measured quantities or the in-silico model. Processor 221 may generate the uncertainty by applying an algorithm or a mathematical model / operation, such as a trained machine learning model, on the mentioned quantities and generate an output indicative of the uncertainty, for example.

[0179] The report may also comprise the price grid matrix or other information regarding financial aspects about agricultural area 210 that are related to the total carbon footprint or other mentioned quantities. The report may be useful for a manager of agricultural area 210 to make decisions about the management of agricultural area 210. For example, based on the report, the manager may enact a change to reduce the total carbon footprint of agricultural area 210. The generated report may be presented to a user on monitor 226, for example. The report may be provided as a CSV file, XML file, or another equivalent file type, for example.

[0180] In some embodiments, processor 221 may provide a recommendation indicative of a practice change to implement in agricultural area 210 to modulate the total carbon footprint of agricultural area 210. The recommendation may be based on the generated report or the previously mentioned quantities (e.g., the total carbon footprint of agricultural area 210; the change in soil organic carbon quantity; the livestock carbon footprint; the carbon footprint of the product; and a further carbon footprint based on other sources of carbon within agricultural area 210). The recommendation may also be provided on the generated report. For example, the practice change comprises one or more of: pasture improvement; planting legumes; applying fertilisers; transitioning from cropping to permanent pastures; time-controlled grazing, pasture rotation, pasture spelling and the like. Processor 221 may generate the recommendation by applying an algorithm or mathematical model / operation, such as a trained machine learning model, on the mentioned quantities and generate an output indicative of the recommendation.

[0181] In some embodiments, the previously mentioned quantities (e.g., the total carbon footprint of agricultural area 210; the change in soil organic carbon quantity; the livestock carbon footprint; the carbon footprint of the product; and a further carbon footprint based on other sources of carbon within agricultural area 210) may be represented in carbon credits or Australian Carbon Credit Units (ACCUs) or carbon dioxide equivalent (CChe). The representation in carbon credits may be provided on the generated report. As such, the disclosed method can provide an estimate of an amount of carbon credits that could be earned or are required for agricultural area 210. Moreover, the generated report may provide a recommendation on how a practice change can be implemented within agricultural area 210 to alter / modulate soil carbon levels and / or a carbon footprint thereof and thereby earn a certain amount of carbon credits. Such carbon credits may then be applied to a product derived from livestock within agricultural area 210 and may be indicated by a label, stamp, tag or the like (e.g., a barcode or QR code) associated with the product (e.g., a whole animal carcass or selected cut of meat). Alternatively, such carbon credits may be held as an asset or sold to the market as required.

[0182] One carbon credit represents a reduction, avoidance or removal of one metric Tonne of carbon dioxide or its COie. A carbon credit can be bought or sold after certification by a government or independent certification body. In particular, the ACCU program encourages people and businesses to run projects that reduce emissions or store carbon, for example by: using new technology upgrading equipment changing business practices to improve productivity or energy use changing the way vegetation is managed. As a result, the ACCU program encourages the practice changes discussed previously in order to reduce carbon emissions.

[0183] As such, the disclosed methods and systems may improve the financial returns of soil carbon projects for producers by reducing the cost and increasing the accuracy of soil carbon quantification. This may result in a higher and more timely production of ACCUs at a lower cost which will markedly increase the financial returns of soil carbon projects for producers. Achieving this may provide large areas of additional land for soil carbon projects for the betterment of land productivity and decarbonisation. Moreover, the application of ACCUs or carbon credits to products, such as meat, derived from the agricultural area may attract a price premium by virtue of being a carbon neutral product. Experiments

[0184] Experiments to test the performance of the methods disclosed herein will now be discussed. In these experiments, at least one of the one or more sensors is a flux tower. More specifically, at least one of the one or more sensors is an Eddy Covariance Flux Tower. In these experiments, the DayCent CABBI model was used as the in-silico model, which was continuously calibrated using measurements received from the flux towers are the data was received. In other words, in-silico model calibration at the core sites continued as the data became available. DayCent CABBI model is a modified version of DayCent that was developed by the Center for Advanced Bioenergy and Bioproducts Innovation based on the DayCent-Photo model version which simulates the seasonally variable photosynthetic capacity in the calculation of gross primary production (GPP).

[0185] For the following experiments, seven flux towers were rented and were placed in different locations in Eastern Australia. More specifically, the seven flux towers were placed at different locations in Central Queensland and Central New South Wales, which experience unique and harsh environmental conditions. The specific locations of these flux tower can be seen in Fig. 5. In particular, Fig. 5 shows the sites located in the Brigalow Belt bioregion, northeast Australia consisting of one long-term trial (LTTS) for model calibration and the seven flux tower sites (Location C, Location D, Location A and Location B in the south and Location E, Location F and Location G in the north). Some of the locations were used as a control test site, while other locations were experimental sites where at least one practice change was implemented to observe the difference in soil sequestration that occurred as a result of the practice change.

[0186] For example, Desmanthus was planted at some locations to observe the effect it has for soil sequestration. The deep tap roots of the Desmanthus, as well as its nitrogen delivery and strong drought tolerance provide great productivity improvements and soil carbon sequestration. The results herein showed that the Desmanthus paddocks sequestered an additional 0.3+ tonnes of carbon per hectare per year, which in combination with the increased weight gain and methane reduction properties of Desmanthus, can create a pasture system where the carbon drawdown is greater than the carbon emissions from the animals. In other areas, time controlled grazing (TCG) was implemented. The specific locations and a description of the site use is provided in Table 1. More detail about these sites is also provided below.

[0187] The calibration of the in-silico model was undertaken in two steps: 1. Using a long-term research site that has long term (30+ years) or relevant measurements to provide an initial calibration of the model. For the experimental results provided herein, the results of the Brigalow Catchment Study (https: / / www.catchmentstudv.com / ) was used to provide an initial calibration of the model. Given that the Brigalow Catchment Study was a long-term study, the initial calibration of the in-silico model enabled the model to simulate the ecological development of the area since it was first colonised. As such, using the long-term research site data to initially calibrate the model enabled the model to be indicative of a history and the management history of the agricultural areas observed in these experiments; and 2. A second fine-tuning calibration of the model using flux tower data at a more site-specific level. The fine-tuning calibration using the flux tower data enabled the in-silico model to more accurately resemble the specific locations and more accurately model the soil sequestration at these locations (i.e., the specific agricultural areas observed during these experiments).

[0188] Input data including weather data was then provided to the calibrated in-silico model (i.e., the in-silico model was evaluated on the input data) to simulate the soil sequestration in the abovementioned locations to determine the changes in soil organic carbon quantity. More specifically, the input data included: • soil data, which was obtained from the Soil and Landscape Grid of Australia (SLGA) (https: / / esoil.io / TERNLandscapes / Public / Pages / SLGA / ); • climate data, in which data from the SILO climate database (https: / / www.daia.qld.gov.au / daiaset / silo-cliniate-database) was used; and • pasture data, which specified that pasture included Buffel grass (the pasture data was calibrated using Brigalow Catchment Study data). Gathering measurements from one or more sensors

[0189] Seven field trials with flux tower measurements were conducted in the improved grazing grasslands located in the south and north of the Brigalow Belt bioregion in Queensland, Australia. These pasture systems were dominated by C4 grass species, typically Buffel grass. The soil at these sites was classified as a Vertisol following the USDA, characterised by cracking clay soils and gilgai microrelief. The climate in the region is humid subtropical (Koppen-Geiger classification: Cfa) with mean annual precipitation of 561-684 mm and mean annual temperature of 19.6-22.0°C across the sites. Characteristics of the field sites are summarised in Table 1. SUBSTITUTE SHEET (RULE 26) Region South Queensland (Southern Brigalow Belt) Central Queensland (Northern Brigalow Belt) Site code Location A Location B Location C Location D Location E Location F Location G Latitude -28.38 -28.33 -28.11 -28.12 -24.29 -24.08 -24.21 Longitude 150.56 150.61 150.58 150.58 148.72 148.70 149.83 Treatment TCG TCG Control Legume Control Legume TCG Flux data start 2 / 11 / 2021 1 / 02 / 2019 15 / 05 / 2021 15 / 05 / 2021 15 / 05 / 2021 1 / 03 / 2021 24 / 11 / 2021 Flux data end 29 / 10 / 2023 28 / 04 / 2022 30 / 09 / 2023 29 / 09 / 2023 1 / 10 / 2023 12 / 09 / 2022 26 / 09 / 2023 Mean annual temperature (°C) 20.1 20.1 20.0 20.0 22.1 22.4 22.6 Mean annual precipitation (mm) 548 548 538 538 561 554 638 Bulk density (gcm‘3) 1.28 1.13 1.32 1.48 1.48 1.15 1.22 PH 8.5 8.6 7.7 8 7.3 8.7 7.7 Clay content (%) 42.3 46.1 50.2 45.3 26.6 51.7 51.5 Table 1: Bioclimatic and soils (0-0.3 m) information for the seven flux sites located in the Brigalow Belt bioregion, northeast Australia. WO 2025 / 245577                                   PCT / AU2025 / 050561

[0190] The control sites (Location C and Location E) were managed by conventional continuous grazing without fertilisation or irrigation. An improved pasture management practice to sequester C into the soil, time-controlled grazing (TCG) was implemented at the Location A, Location B, Location G sites, which involves using smaller paddocks (15-30 ha) that are heavily stocked at 1015 adult-equivalent per hectare for short periods, followed by long rest periods that allow the pasture to recover and regenerate before grazing is applied again. Carefully timed grazing may enhance grass production and rotational grazing may result in higher SOC stocks compared to continuous grazing. This may be attributed partially to the positive impacts of high-intensity grazing on surface litter incorporation and belowground biomass turnover, including trampling effects.

[0191] A tropical legume, Desmanthus (Desmanthus spp., cv. Progardes), was introduced at the Location D and Location F sites. Leguminous plants supply N to grass by fixing N from the atmosphere and thus increase pasture productivity and C inputs to the soil. Furthermore, legumes typically have longer tap root systems that may allow them to access deep moisture and nutrients out of reach of more shallow root grass pasture species, resulting in greater productivity and ground cover during extended dry periods. A deep root system may also be more likely to sequester carbon at depth in the soil profile, which is less likely to be accessed and decomposed by soil microbes.

[0192] At the flux sites, soil, plant biomass and flux tower measurements were conducted. Soil cores were sampled to 0.3 m at approximate annual increments from a 30 ha area around the flux tower at each site. A coring method (in particular, the Balanced Acceptance Sampling method) was used to derive 24 coring points within the flux footprint which were repeatedly sampled at 1 ±7 m GPS accuracy across sampling events. These soil cores were cut to fixed depth intervals (00.1, 0.1-0.2 and 0.2-0.3 m or 0-0.1 and 0.1-0.3 m), weighed and dried to 105 °C, to determine field water-content and bulk density. The samples were ground and sieved to a fine < 2 mm fraction and any significant gravel fraction was recorded. Sub-samples were analysed for SOC and soil total N by the dry combustion method using a CNS-2000 analyser. A subset of 4-6 composites per depth was bulked according to strata defined by the Empirical Bayesian Kriging method using the observed SOC stocks. For each composite sample, SOM was fractionated into POM and MAOM by physical and chemical procedures, followed by C and N content measurements using Isotope Ratio Mass Spectrometry (20-22 Sercon Limited, UK). Soil pH and particle size distribution of each composite sample were analysed in a 1:5 (w / v) water extract and by the hydrometer method, respectively.

[0193] Aboveground biomass was measured up to four times a year using quadrat cuts between 0.25 and 1.00 m2. All biomass cuts were at 0.1 m above the ground to separate new growth from historic litter / standing stubble. The remaining biomass 0.1 m above the ground was quantified twice during the experiments and the average amount was applied to all the biomass measurements per site. The plant samples were either wholly processed or sub-sampled prior to drying at 60 °C in the laboratory to determine standing aboveground dry matter. Plant C and N contents were analysed by the dry combustion method using CNS-2000 analyser.

[0194] The flux towers were installed in 2019-21 and measured continuously until 2022-23 (approximately two years or longer at each site as summarised in Table 1). Turbulent (carbon, water, and energy) fluxes were measured at 20Hz using a standard configuration of; a LI-COR LL 7500DS CO2 / H2O infrared gas analyser, Gill WindMaster Pro and LI-COR SmartFlux 3 unit, mounted at approximately 3.4 m above ground level across sites. Additional soil and meteorological measures were collected; soil heat flux (HFP-01, Hukseflux, Delft, Netherlands), soil temperature & volumetric water content (Hydraprobe II, Stevens, Portland, Oregon, USA), surface net all-wave radiation (NR-Lite2, Kipp & Zonen, Delft, Netherlands), air temperature & relative humidity (HMP155 combined thermometer & hygrometer, Vaisala, Vantaa, Finland), precipitation (TR525-M, Texas Electronics, Dallas, Texas, USA), up- and down-welling solar and thermal radiation flux (CNR4, Kipp & Zonen, Delft, Netherlands) or where this unit was not fitted - LI-COR LL200R Pyranometer and LI-COR LL190R Quantum sensors were installed (LLCOR Biosciences, Lincoln, Nebraska, USA). All soil and meteorological measurements were collected on LI-COR Data Acquisition Module (DAqM) units at 1Hz with 30-minute averaging intervals. The 20 Hz flux measurements were processed to 30 min covariances utilising covariance maximization to compensate for time lag, and spike removal following. These measurements were then post-processed using PyFluxPro (version 3.4) following the OzFlux processing protocol, resulting in gap-filled daily net ecosystem exchange (NEE) and actual evapotranspiration (AET). The in-silico model

[0195] The DayCent-CABBI model used in the experiments described herein consists of several sub-models, including an SOC sub-model that differentiates SOC into active, slow, and passive pools each with either the original first-order (FO) or new Michaelis-Menten (MM) kinetics that accept C and N inputs from above and below-ground plant litter pools as well as exogenous sources. The MM sub-model simulates microbial mediated decomposition and microbial necromass recycling along with mineral interactions modified from those in the FO sub-model. In the MM sub-model, most C from other pools is now routed through the live microbe and dead microbe pools as material passes through the decomposition process before entering either slow or passive soil pools, though some lignin from decomposing structural litter and dead wood can bypass microbial processing and flow directly to the slow pool, and some microbial necromass flows directly to the passive pool. The slow and passive soil pools may be now considered POM and MAOM pools, respectively. The updated parameters in fix.100 (e.g., psls3, ps2s3 and varat parameters) represent these pools and C flow. The MM sub-model maintains a similar pool structure, pool properties, soil texture effects, and lignin effects on carbon flows from the original DayCent decomposition function but adds surface and soil dead microbe biomass pools and replaces the surface and soil active pools with surface and soil microbe biomass pools. This implementation method may have allowed the new MM sub-model to include microbial explicit processes without substantially increasing the complexity of the DayCent model.

[0196] Together with the plant, soil water, soil N sub-models, the DayCent-CABBI model enables simulation of GPP, plant and soil respiration, net ecosystem exchange, evapotranspiration, soil C, N and water dynamics, allowing direct comparison with EC flux tower measurements. Input into the in-silico model and configuration

[0197] Soil inputs (bulk density, soil texture and pH) (0-1.0 m) for the DayCent-CABBI model were calculated from the initial soil sampling event at each site. The historic climate data were retrieved from Scientific Information for Land Owners (SILO) climate stations and interpolated gridded database. For future projection under climate change, the moderate- and high-emissions scenarios - that is Representative Concentration Pathways 4.5 and 8.5 (RCP4.5 and RCP8.5 respectively) - were considered. In the experiments described herein, high-resolution (10 km) daily climate change projections for Queensland were applied using dynamical downscaling of CMIP5 global climate models provided by Terrestrial Ecosystem Research Network (TERN).

[0198] The DayCent-CABBI model was initialized using the measured baseline SOC stock (00.3 m) and the measured ratio of organic C in MAOM as a proportion of total SOC (including PyOM at the LTTS site) as the passive pool. The relative proportion of the passive pool to SOC ranged from 72% to 87% (on average 80%) across the flux sites. The measured MAOM CN ratio ranged from 11.0 to 12.4 (on average 11.8) and was also used to initialize the passive SON pool. Then, an approximate cropping history simulation was run after land clearing (typically since 1950’s to 70’s) following the farmer’s record at each site to reproduce the current masses of labile (“active” and “slow” in the model) soil C.

[0199] Grazing management in DayCent-CABBI was assumed to continuously remove a proportion of aboveground biomass over a month. For the control and legume treatments at each site, annual continuous grazing was implemented by assuming to remove 10%-20% and 5%-10% of green and dead aboveground biomass respectively except for the winter months (May-August) when 5% and 8% respectively was removed. For the TCG treatment, a grazing event was assumed to remove 20% of both green and dead aboveground biomass over five days, considering the increased density of livestock during the short time window. This TCG event was introduced approximately four times per year, which resulted in three months of the rest period. Considering the potential trampling effects of TCG, the conversion of green biomass into dead biomass was increased by 10% and incorporation of dead biomass, surface litter and root biomass into the soil was increased by 20% at each TCG event. For the legume treatment, the parameter determining symbiotic N fixation maximum for grass (snfxmx(l)) was set at 0.03 compared to 0.005 for the control, following the previous modelling exercise on biological N fixation by legume. In addition, the parameters of C allocation to roots (cfrtcw(l) and cfrtcw(2)) were increased to 0.75 and 0.45, respectively, to reflect the deeper root system of introduced legume. Calibrating and validating the in-silico model

[0200] More specifically, the DayCent-CABBI model was calibrated using the long-term aboveground biomass, SOC and SOC fraction datasets at the Brigalow Catchment Study. In this long-term trial, Buffel, a typical grass in the Brigalow Belt bioregion including the flux tower trial sites, was sown and managed with grazing, representing the baseline grass parameter set. Since the grass community included Buffel but moderately differed across the flux tower sites, the key grass parameter (prdx(l)) and site conditions were further refined by fitting the first half of the time sequence for AET and NEE measured using the EC flux tower at each site. The model was then validated against the second half of the time sequence for AET and NEE as well as the observed aboveground biomass C and SOC stock data across the flux tower sites. This is further detailed in the following sections.

[0201] The calibrated in-silico model was cross-validated against other trial data (Brigalow Catchment Study) to ensure longer-term predictions are accurate. Specifically, to validate the in-silico model, the follow methodology was performed: a. Calibrating Buffel grass parameters of the DayCent CABBI model to fit biomass observed as well as soil organic carbon (SOC) parameters (mainly DEC4, DEC5(2) and PS1S3(2)) to fit total SOC, mineral-associated organic carbon (MAOC) with pyrogenic organic carbon (PyOC) (passive) and soil organic nitrogen (SON) down to 30 cm; b. Refining grass and SOC parameters per trial to fit biomass and SOC; and c. Comparing (thereby validating) the output of the in-silico model with flux tower measurements.

[0202] In particular, it was shown that the in-silico model (i.e., the DayCent) initially calibrated using Brigalow Catchment Study data overestimated the NEE (hence, underestimated the NEP). Fig. 6a shows the cumulative NEE from flux tower (solid black line) versus modelled output for Location C for the model calibrated using Brigalow Catchment Study data only. Fig. 6b shows the cumulative NEE from flux tower (solid black line) versus modelled output for Location E for the model calibrated using Brigalow Catchment Study data only.

[0203] It was thought that the overestimated NEE was possibly because (1) production is greater but allocated more to roots; (2) mixture of C4 & C3 results in negative NEE during the winter too; and (3) flux tower measurements include C going below 30 cm soil depth (outside of the boundary assumed in the model). Therefore, it was thought to fine-tune the calibration of the model to (1) increase productivity to enhance overall C inputs; (2) provide more C allocation to roots to keep the same aboveground biomass; and (3) modify temperature response to have some production during the winter.

[0204] Using the above calibration method, the in-silico model (i.e., the DayCent CABBI model) was calibrated using the data obtained from the flux towers at the specified locations, after being initially calibrated using the Brigalow Catchment Study data. The results of this fine tune calibration with the flux tower data are shown in Fig. 7. Fig. 7a shows the cumulative NEE from flux tower (solid black line) versus modelled output for Location C for the model with fine-tuned calibration using flux tower data. Fig. 7b shows the cumulative NEE from flux tower (solid black line) versus modelled output for Location E for the model with fine-tuned calibration using flux tower data.

[0205] In comparison with Figs. 6a and 6b, it can be seen that the fine-tuning calibration of the in-silico model with the flux tower data significantly improves the accuracy of the in-silico model. These results suggest the significance of local calibration on an in-silico model using flux tower data in determining changes in soil organic carbon. Moreover, these results highlight that long term research site data (e.g., data from the Brigalow Catchment Study) may not be enough data or may not be specific enough to accurately capture the environmental conditions of the locations observed in these experiments.

[0206] Further data was obtained using the fine-tuned in-silico model for Location C and Location C. Fig. 8a shows the flux tower evapotranspiration (ET) (dots) versus modelled output for Location C. Fig. 8b shows the net ecosystem exchange (NEE) flux tower versus model output for Location C. Fig. 8c shows the soil carbon measured (solid black) versus modelled output for Location C. Fig. 9a shows the flux tower evapotranspiration (ET) (dots) versus modelled output for Location E. Fig. 9b shows the net ecosystem exchange (NEE) flux tower versus model output for Location E. Fig. 9c shows the soil carbon measured (solid black) versus modelled output for the Location E.

[0207] Initial results demonstrate the models perform well in the long-term. While the model may not have captured the highly dynamic changes, the flux tower measurements were observed over the measurement period from drought to a triple La Nina event, which is a 1 in 50-year event. In essence, these results show that SOC at the start of trial can be achieved by locally calibrating SOC parameters with reasonable ranges. The experiments also showed that, since the contribution of soil heterotrophic respiration is relatively small already, adjusting the slow and passive SOC parameters is unlikely to make much difference for a short-term. Modifying decomposition rates of active SOC pool and surface litter would also have limited effect.

[0208] Further calibration / validation of slow and passive SOC parameters can be done by comparing with more fractionation data from flux sites. There may also be correspondence between the model’s three pools and different fractionations (with or without PyOC). Fitting NEE may use further adjustments in grass parameters and management practices, basically decreasing NEE (increasing net ecosystem production (NEP)). Calibration with the Brigalow Catchment Study dataset

[0209] The calibration of the in-silico model (i.e., the DayCent-CABBI model) using the longterm Brigalow Catchment Study data enabled the model to capture the growth pattern of Buffel grass (RMSE: 0.66 t C ha-1, nRMSE: 41.2%), which ranged from 0.3 to 3.6 t C ha-1 across years and growing stages, which is shown in Fig. 10. In particular, Fig. 10 shows that observed (dots) and simulated (line) aboveground biomass C (t C ha-1) at the Brigalow Catchment Study site in northeast Australia. The adjusted growth parameter on overall productivity (prdx(l)) was 0.5. The CN ratio for assimilated aboveground biomass was set to range from 20-60 (pramn(l, 1) and pramx(l, 1)) to 35-120 (pramn(l, 2) and pramx(l, 2)) in response to aboveground biomass up to 120 g biomass C m-2 (biomax). The range of C allocation to roots was set to 0.3-0.5 (cfrtcn(l) and cfrtcn(2)) and 0.35-0.65 (cfrtcw(l) and cfrtcw(2)) in response to nitrogen and water stresses.

[0210] For SOC simulation, the maximum decomposition rate of the passive and slow SOM pools (dec4 and dec5(2)) was highly influential and adjusted to 0.0011 and 0.20 (Table 2). The simulated passive SOC was constant over the observation period at around 411C ha-1 and agreed well with the upper end of the observed inert fraction of SOC, which is shown in Fig. 11. In particular, Fig. 11 shows the observed (dots) and simulated (lines) soil organic carbon (SOC) stock in the 0-0.3 m soil depth (t C ha-1) in the total (red) and inert (blue) carbon pools at the Brigalow Catchment Study site in northeast Australia. The simulated active + slow SOC slightly declined in the first few years and increased back to the initial level of SOC stock over ~10 years in the range of 47-53 t C ha-1, resulting in a close agreement in total SOC between observation and simulation (RMSE: 1.811 C ha-1, nRMSE: 3.6%), as shown in Fig. 11. Fine-tuning and validation with the flux tower trial datasets

[0211] After refining the grass parameters using the NEE data at the flux tower sites, the parameter prdx(l) varied from 0.5 to 0.7. The scaling factor for potential evapotranspiration (fwloss(4)) also slightly varied from 0.5 to 0.55 by fitting AET across the flux tower sites.

[0212] There was reasonable agreement between simulated aboveground biomass C and the observed values across the flux tower trials during the time window from 2020 to 2023 (0.88 t C ha-1 of RMSE and 55.6% of nRMSE), which is shown in Fig. 12. In particular, Fig. 12 shows the observed and simulated aboveground biomass C (t C ha-1) across the flux tower trials in northeast Australia. Dots and error bars indicate mean values and standard errors.

[0213] Aboveground pasture biomass measurements at the Location A and Location B sites covered high inter-annual rainfall variability and the response of aboveground pasture biomass ranging from 0.2 to 2.8 t C ha-1 was well simulated by the model. Within-year variation was not fully captured at some sites such as Location E, but the average productivity was simulated without bias. Aboveground biomass N was simulated with a comparable performance as biomass C, resulting in 11.7 kg N ha-1 of RMSE (nRMSE: 48.3%), which is shown in Fig. 13. Fig. 13 shows the observed and simulated aboveground biomass N (kg N ha-1) across the flux tower trials in northeast Australia. Dots and error bars indicate mean values and standard errors.

[0214] Weekly-averaged AET was simulated close to observations with 0.74 and 0.86 mm of RMSE during the calibration and validation periods, respectively, which is shown in Fig. 14. Fig. 14 shows the observed and simulated weekly average actual evapotranspiration (AET) (mm) during calibration (top) and validation (bottom) periods across the flux tower trials in northeast Australia. At the cumulative scale, the RMSE for simulated AET was 104 and 136 mm (nRMSE: 16.8% and 20.2%) during the calibration and validation periods, respectively, which is shown in Fig. 15. Fig. 15 shows the observed (solid lines) and simulated (dotted lines) cumulative actual evapotranspiration (AET) (mm) during calibration (top) and validation (bottom) periods across the flux tower trials in northeast Australia. Weekly-averaged NEE agreed well between observation and simulation with 1.37 and 1.36 g C m-2 of RMSE during the calibration and validation periods, respectively, which is shown in Fig. 16. At the cumulative scale, the RMSE for simulated NEE was 0.93 and 1.16 t C ha-1 (nRMSE: 26.4% and 24.8%) during the calibration and validation periods, respectively, which is shown in Fig. 17.

[0215] There was good agreement between simulated and observed SOC stocks in the 0-0.3 m soil depth across the flux tower trials over 2-4 years, resulting in an RMSE of 2.81 t C ha-1 (nRMSE: 7.4%, which is shown is Fig. 18a. Due to the transition from dry to wet years from 2019 to 2023, the observed SOC stock at the Location A and Location B sites increased during this window by 0.72 and 1.75 t C ha-1 yr-1 and these increases were well simulated by DayCent (1.07 and 1.17 t C ha-1 yr-1, respectively), which is shown in Fig. 18b. It is noted that in Figs. 18a and b, dots and error bars indicate observed mean values and standard errors, and lines in Fig. 18b indicate simulated values. In terms of pasture management treatments, the Location G (TCG) and Location F (legume) sites demonstrated greater soil C sequestration by 1.55 and 2.96 t C ha-1 yr-1, respectively, compared to their control site (Location E, 1.23 t C ha-1 yr-1), which agreed with the simulated SOC accrual by 1.87 (Location G), 1.47 (Location F) and 0.92 (Location E) t C ha-1 yr-1. The observed SOC stock increased (on average) by 1.46 t C ha-1 yr-1 across the flux tower trials, and the model simulated comparable increases in SOC of 1.17 t C ha-1 yr-1. Climate change scenario analysis

[0216] Future soil C sequestration potential of the pasture management practices (i.e. control, TCG and legume) under the climate change scenarios (RCP 4.5 & RCP 8.5) up to 2050 was evaluated at the control sites (Location E in Central Queensland and Location C in South Queensland). Projected mean annual precipitation and mean annual temperature during 20242050 were 542 and 525 mm and 23.6 and 21.4 °C at the Location E and Location C sites respectively under RCP 4.5, and 623 and 592 mm and 23.8 and 21.5 °C under RCP 8.5. To account for the uncertainty associated with model parameters, a Monte Carlo simulation was conducted and the key grass parameter (prdx(l)) was sampled from the range of site-specific calibration for each treatment (n=100).

[0217] Results

[0218] The long-term forecast of SOC change (0-0.3 m) to 2050 under the two climate change scenarios found the largest increases in SOC in the TCG treatment, followed by the legume treatment and the control, which is shown in Fig. 19, at both Location E (Central Queensland) and Location C (South Queensland) sites. In particular, Fig. 19 shows the simulated soil organic carbon dynamics at flux sites (Location E in Central Queensland and Location C in South Queensland, Australia) with control (red), time-controlled grazing (blue) and legume incorporation (green) grassland management strategies up to 2050 under climate change scenarios, RCP4.5 (solid line) and RCP8.5 (dotted line). Lines and shaded areas indicate mean and 95% confidence intervals of parameter-induced uncertainty analysis. Between the climate change scenarios, RCP8.5 resulted in a larger SOC stock compared to RCP4.5 across pasture management strategies and sites. Across all the treatments and climate change scenarios, the increase in SOC stocks diminished over time.

[0219] Across the Location E and Location C sites, the change in SOC stock from 2024 to 2050 ranged from 2.6-6.6, 5.0-11.2 and 9.1-17.2 t C ha-1 under control, legume and TCG treatments respectively (Table 2). The change in soil carbon in MAOM were 0.2-0.4, 0.3-0.8 and 0.7-1.6 t C ha-1 under control, legume and TCG treatments respectively, accounting for 3%—15% of the SOC changes. The corresponding annual SOC accrual ranged from 0.10-0.25, 0.19-0.43 and 0.35-0.66 t C ha-1 yr-1 for the control, legume and TCG treatments, respectively. Compared to the control, the legume and TCG treatments resulted in an increase in SOC stock ranging from 2.4-4.0 and 6.5-9.6 t C ha-1 under RCP4.5 and 2.9-4.6 and 8.8-10.6 t C ha-1 under RCP8.5. SUBSTITUTE SHEET (RULE 26) Region Site Treatment Climate change scenario SOC change from 2024 (tCha-1) MAOC change from 2024 (tCha-1) SOC accrual (t C ha-1 yr-1) SOC change from the control (t C ha-1) Central Queensland Location E Control RCP 4.5 5.3 +0.8h 0.2 + O.Oi 0.2 Legume 9.3 ± 1.5e 0.3 + O.Oh 0.36 4.0 TCG 14.8 +0.5b 0.7 + O.Oe 0.57 9.6 Control RCP 8.5 6.6 + 0.8g 0.2 + O.Oi 0.25 Legume 11.2 + 1.6d 0.4 +0.1g 0.43 4.6 TCG 17.2 + 0.5a 0.8 + 0.0 c 0.66 10.6 South Queensland Location C Control RCP 4.5 2.6 + 0.5j 0.3 + O.Oh 0.10 Legume 5.0+ l.Ohi 0.6+0.1g 0.19 2.4 TCG 9.1 + 0.3e 1.3 + 0.0b 0.35 6.5 Control RCP 8.5 4.7 + 0.6i 0.4 +0.1g 0.18 Legume 7.6 + 1.2f 0.8 + 0.1d 0.29 2.9 TCG 13.5 +0.4c 1.6 + 0.0a 0.52 8.8 ANOVA P value Site < 0.001 < 0.001 Treatment < 0.001 < 0.001 RCP < 0.001 < 0.001 WO 2025 / 245577                 53                PCT / AU2025 / 050561 Table 2. Simulated soil organic carbon (SOC, in t C ha-1) change from 2024 to 2050 in the control, time-controlled grazing (“TCG”) and legume incorporation (“Legume”) treatments and under climate change scenarios (RCP4.5 and RCP8.5), soil carbon in mineral-associated organic matter (MAOC, in t C ha-1) change, average annual SOC accrual over the simulation period (t C ha-1 yr-1) and SOC change compared to the control treatment in 2050 at flux sites (Location E in Central Queensland and Location C in South Queensland, Australia). Different alphabet letters indicate a significant difference between combinations of site, treatment and climate change scenario in each variable. SUBSTITUTE SHEET (RULE 26) WO 2025 / 245577                 54                PCT / AU2025 / 050561 Predicting soil organic carbon pro jection

[0220] Using the in-silico model calibrated for Australian environmental conditions, a scenario analysis was undertaken of a large Central Queensland property for different biomass growth scenarios driven by different pasture species, grazing practices and soil nutrient correction. The in-silico model predicted soil organic carbon changes under different combination of above ground biomass (AG) and below ground biomass (BG) increases (driven by species, grazing and nutrients). Fig. 20 shows the results from the calibrated in-silico model for different combinations of above ground biomass (AG) and below ground biomass (BG) increases. Total carbon footprint of an agricultural area

[0221] Using the combination of the calibrated in-silico model and a carbon accounting model, the total carbon footprint of an agricultural area was determined in the following experiment. For the following experiments, the in-silico model was again the DayCent CABBI model, while the carbon accounting model was the “Sheep & Beef GHG Accounting Framework” (SB-GAF) provided by the Primary Industries Climate Challenges Centre (https: / / piece.org. au / resources / Tools.html), which is incorporated by reference in its entirety herein. However, it is noted that other models and frameworks may be used as the carbon accounting model.

[0222] Fig. 21 shows the cumulative net carbon position of an agricultural area in this experiment. These results can also be expressed in terms of the lightweight gain of the livestock animals within the agricultural area, to provide a better understanding of the carbon neutrality of the agricultural area. This is particularly important for beef operations, for example. Fig. 22 shows a quantification of the animal and soil carbon emissions with legume / grass pasture.

[0223] These results demonstrate the potential of a calibrated in-silico model that can provide a daily net carbon content. In particular, real time (or periodic) understanding of drivers of soil carbon enables optimised agricultural system management. Moreover, the net carbon position per kg beef grown (LW) was determined to be -19 kg, meaning that the agricultural area is carbon negative (i.e., the agricultural area removes more carbon from the atmosphere than it emits). The large amount of soil carbon as can be shown in Fig. 21 can be attributed to sequestration driven by improved pasture with legumes, better grasses and better soil nutrition.

[0224] The following assumptions were made when capturing the results mentioned above: •    Whole of life animal emissions = 2.8 T CO2-e; •    Animal life (slaughter animals) = 2.4 years; • Slaughter weight = 600 kgs; • Hectares per animal = 2; and • Soil carbon sequestration = 1.1 T CO2-e per hectare per year. Effect of practice changes

[0225] The results discussed above have already shown that implemented practice changes can change the amount of soil sequestration and hence, the total carbon footprint of an agricultural area. As such, the disclosed methods can provide accurate quantification of the impact of management practice changes. The disclosed method also enables scenario analysis of property changes to optimise returns. Fig. 23 shows a comparison of net ecosystem exchange (NEE) between a site where no practice change was implemented and a site where a practice change was implemented. Duration of measurements used to calibration the in-silico model

[0226] A study was conducted to determine an optimal duration of measurements used to calibrate the in-silico model. In particular, this study investigated the optimal duration of eddy covariance (EC) flux measurements to achieve reliable model validation for soil carbon sequestration. In this study, the DayCent-CABBI model was calibrated by fitting long-term aboveground biomass and SOC data from the Brigalow Catchment Study, then further refined using approximately one year of AET and NEE data measured using the EC flux tower at each site prior to the model verification flux measurements, similar to the other experiments discussed above. Eddy Covariance flux towers were installed in 2019-21 and measured turbulent (carbon, water, and energy) fluxes at 20 Hz continuously until 2022-24.

[0227] The results of this study indicated that uncertainty in model prediction decreased with longer measurement durations, indicating that that longer verification periods generally lead to greater stability in model prediction. However, to clarify, this stability is not necessarily a measure of model accuracy. The results indicate that an extended verification period improves the reliability of the model’s prediction (i.e., the consistency in which the same results are produced), but does not necessarily improve the model’s predictive performance. In other words, these results indicate not that the model itself performs better with a longer verification period—only that the verification outcome becomes more stable over time. As such, the accuracy of the model may not necessarily be dependent on a longer verification period. The results also indicate that a longer verification period may be useful to reliability predict measurements in the far future.

[0228] The study also highlights an advantage in the measurements spanning multiple seasons to capture environmental variability, particularly noting the impact of seasonal rainfall patterns on verification stability. The results indicate that testing data should span a sufficiently wide range of environmental conditions to reflect real-world variability. In particular, a verification duration should reliably cover the range of environmental conditions, such as temperature, insolation and precipitation, that a location experiences. For example, the results indicate that spanning multiple seasons may reliably cover the range of environmental conditions, such as between around 60-200 days. Using an in-silico model calibrated with measurements from a location associated with the geographical area

[0229] The calibrated in-silico model was used to model SOC changes in temperate pasture sites located in central NSW without further calibration from flux tower measurements within the region. The in-silico model was calibrated as previously described e.g., the model was ‘spun up’ to present day to get a baseline using the Brigalow Catchment study, and then fine-tuned using data collection within the corresponding geographical area (i.e., the Brigalow bioregion). The model, calibrated for the Brigalow bioregion, was then applied to an associated geographical area (i.e., central NSW), and the results of the in-silico model were compared to a historical dataset of soil coring data available through NSW Department of Primary Industries (DPI). No finetuning was performed using flux tower measurements from the associated geographical area (i.e., no flux towers were used at all on the NSW sites). Again, the DayCent CABBI model was used.

[0230] These results were provided by Meat and Livestock Australia (MLA), the Queensland University of Technology (QUT) and the NSW DPI. The soil cores used in the following results were provided in Badgery, W., Murphy, B., Cowie, A., Orgill, S., Rawson, A., Simmons, A., & Crean, J. (2020). Soil carbon market-based instrument pilot-the sequestration of soil organic carbon for the purpose of obtaining carbon credits. Soil Research,59(1), 12-23, which is incorporated by reference in its entirety herein. The following results are provided in Eckhard, R., Grace, P., Badgery W. (2025). Delivering Integrated Management System (IMS) options for CN30. Meat & Livestock Australia, Project Code P.PSH.1333, Retrieved February 28, 2025, from https: / / www.mIa.com.au / research-and-development / reports / 2025Zp.psh.1333—cn30-integrated-man a gement- s vstems / , which is incorporated by reference in its entirety herein.

[0231] The associated geographical area included sites at the Orange Agricultural Institute (OAI) in NSW and the Cowra Trough in the Central West region of NSW. This region is generally temperate, with year-round rainfall and hot summers. In particular, the Cowra Trough has an average annual rainfall of -600 mm and mean maximum and minimum temperatures of 25°C and 8°C respectively. Such geographical area is associated with the Brigalow bioregion. For example, similar environmental conditions may exist between the two geographical areas.

[0232] As will be discussed below, the results of this study demonstrate the generalisation potential of the calibrated model (i.e., the model does not necessarily need to be calibrated on data collected from the region of interest). As such, the result indicate the accuracy of an in-silico model calibrated using measurements from one or more sensors are placed at: a location associated with the geographical area.

[0233] Fig. 24 shows the model simulated SOC versus the observation SOC data for the on-farm trials in the Cowra Trough located in central NSW. Each data point in Fig. 24 represents a soil coring sample (e.g., SOC(l)) for a sampling event at a site within the Cowra Trough. The lines represent the in-silico model’s SOC prediction over time. For the Cowra Trough, simulations using the calibrated in-silico model indicate that even though the SOC measurements have indicated a decline in SOC at the majority of the Cowra Trough, the model trends show an increase in SOC in the sites where perennial pastures or improved grazing has been implemented. Continued declines in SOC stocks have been confirmed by the model in cropping systems. This is a good example of the value of simulation models to provide clearer trends in SOC change in contrast to measurements alone which may be misleading due to short term climate variability.

[0234] Fig. 25 shows a comparison between simulated and observed SOC in the 0-30 cm topsoil across the Cowra Trough (circle) and OAI grazing (triangle) trial sites in NSW. Similar to Fig. 25, each data point represents a soil coring sample (e.g., SOC(l)) for a sampling event at a site. Using the calibrated in-silico model, the simulated SOC stocks in the 0-30 cm layer agreed with the observed values with a very high degree of accuracy across the two NSW long-term trials (OAI and Cowra Trough), resulting in 2.43 t C / ha of RMSE (6.1% of nRMSE) with the range of observed SOC from 20.9 to 50.5 t C / ha. This tested the model’s performance without flux tower finetuning with very strong results. Case study

[0235] A case study demonstrating the disclosed method was performed for a grazing property located in the heart of the Brigalow Belt in Central Queensland. The latitude is slightly north of the Tropic of Capricorn and the average rainfall is 600mm. The soils in the project area are predominantly brown and grey clays with medium to high clay content. The predominant original vegetation was Brigalow. The original forest in the paddock was pulled and burnt in 1980. Buffel grass was then seeded and remained the dominant species. The paddock was blade ploughed in 1995 and in 2010 to control the Brigalow tree regrowth. The project area was due for blade ploughing (to control Brigalow regrowth) and seeding with legumes and high performing grasses.

[0236] A flux tower was installed in-situ for this process. By continuously measuring carbon fluxes and combining them with process-based modelling, the disclosed method could distinguish between changes in soil carbon driven by management and those caused by climatic variability. This ensures that only genuine, management-induced sequestration is credited, improving the integrity of the method and market confidence.

[0237] Fig. 26 shows an example of the procedure for the initialisation of the in-silico model. This process summarises each step at a high level which leads to the creation of the “digital twin” which can be used for scenario analysis, including feasibility, management change analysis and due diligence procedures. It is noted that this is only one example of the procedure for initialisation of the in-silico model, and variations of this procedure are equally possible according to the present disclosure. The example procedure shown in Fig. 26 was used in this case study, but is not limited to the particular property investigated in this case study.

[0238] Fig. 27 shows an example of the procedure for the demonstration of the in-silico model. This process summarises each step from calibration of the model through to its verification. Again, it is noted that this is only one example of the procedure for demonstration of the in-silico model, and variations of this procedure are equally possible according to the present disclosure. The example procedure shown in Fig. 27 was used in this case study, but is not limited to the particular property investigated in this case study.

[0239] It is noted that in some embodiments (not exclusively related to this case study), the in-silico model may simulate outputs, including SOC for different pasture options. In other words, the in-silico model may account for different practice changes and provide outputs (such as SOC) according to the different practice changes. This may include some of the pasture options or management changes previously discussed, e.g., addition of legumes to the geographical area.

[0240] In this case study, the DayCent-CABBI was used according to the procedure described in this disclosure. Fig. 28 shows an example of scenario analysis conducted for the case study which compares the Soil Organic Carbon of Buffel and Desmanthus Legumes using long term weather patterns over a 25 year period, aligned to an ACCU project. Desmanthus-Grass pasture delivers additional Nitrogen to the whole system, additional Carbon in deeper roots while being less impacted by rainfall variation. Grass only pasture is anticipated to produce lower Nitrogen and Carbon and is impacted more by rainfall variation.

[0241] Fig. 29 is an example of scenario analysis conducted for the case study which compares the above ground biomass of Buffel and Desmanthus Legumes using long term weather patterns over a 25 year period, aligned to an ACCU project. The results indicate indicates that Desmanthus-Grass pasture is anticipated to generate additional biomass compared to buffel, leading to improved soil fertility. Discussion of results Calibration and validation of DayCent in subtropical grasslands in Australia

[0242] The parameter set established for the 0-0.3 m soil depth in grasslands in northeast Australia was considerably different from the default model parameter set for the 0-0.2 m soil depth in the US. The parameters related to decomposition rates of C pools, especially dec4 and dec5(2), were shown to be highly influential on SOC simulation.

[0243] In the experiments described herein, refining the grass productivity parameter (prdx(l)) and maintaining the same SOC decomposition parameters across sites enabled the model to successfully replicate C fluxes observed from the network of EC flux towers (Figs. 16 and 17), as well as aboveground biomass (Fig. 12) and SOC stocks (Figs. 18a and b). This grass productivity parameter was highly influential on gross GPP and useful for fitting the observed NEE under different pasture management strategies (Figs. 16 and 17). In turn, the continuous NEE measurements using EC flux towers have a strong advantage in refining such influential grass parameters on top of low-frequency aboveground biomass measurements.

[0244] The validation results of DayCent-CABBI in the experiments described herein with 2.8 t C ha-1 of RMSE against 30-50 t C ha-1 of the SOC stock observed are similar to the latest US simulation studies using DayCent with RMSE ranging from 5 to 9 t C ha-1 with up to 100 t C ha-1 of the SOC stock observed. These experiments successfully demonstrate the applicability of DayCent-CABBI to simulate terrestrial C cycling in subtropical grasslands in Australia and serves as a template for the regional calibration of DayCent-CABBI combining long-term SOC data and EC flux tower data to examine the effect of pasture management strategies on SOC stocks. Pasture management strategies under climate change scenarios

[0245] The greater C sequestration simulated across the sites and treatments under RCP 8.5 compared to RCP 4.5 (Fig. 19 and Table 2) can be explained by the higher rainfall under RCP 8.5, promoting pasture growth and thus C inputs to the system in this region. Introduction of deep rooting legumes and TCG has the greater potential to sequester C for the long term (~25 years) under climate change scenarios, by up to 0.43 and 0.66 t C ha-1 yr-1 (Fig. 19, Table 2), which were about half of the short-term (2-4 years) accrual observed (1.46 t C ha-1 yr-1) and simulated (1.17 t C ha-1 yr-1) (Fig. 18b). These findings highlight the risk of overestimating long-term SOC sequestration potential based on short-term observations and the advantage of modelling approach to temporally extrapolate the experimental results accounting for the climate change impacts.

[0246] The estimated soil C sequestration was mostly by restoration of the POM degraded due to either cropping history or prior high-intensity grazing management, rather than increasing MAOM over the simulation period to 2050 (Table 2). The capacity of MAOM accumulation is primarily constrained by the soil texture is often considered as the significant factor in assessing the soil C sequestration potential.

[0247] However, there is potential for increasing POM pools and keeping POM over long time scales in many soils, and POM can be a direct precursor of MAOM. These experiments highlight the usefulness of pasture management practices targeting both MAOM and POM pools for practical and achievable soil C sequestration strategies.

[0248] The implementation of TCG management showed greater soil C sequestration based on the cumulative NEE observed at both paired sites (Location C vs Location A / Location B and Location E vs Location G) (Fig. 17), resulting in increases in the simulated SOC stock consistently across sites and climate change scenarios (Fig. 19). The effects of TCG are reflected in the higher grass productivity (prdx) and enhanced biomass turnover at each grazing event in addition to the short intensive grazing events with long rest periods. On the other hand, representation of the impacts of TCG with varying frequency and the trampling effects on grass growth and soil C cycling is still highly uncertain, requiring further data collection of precise grazing management and corresponding plant-soil interaction on C cycling.

[0249] The legume incorporation treatment at the Location D site did not achieve a greater soil C sequestration compared to its control site (Location C) based on the observed cumulative NEE or SOC changes (Figs. 17, 18a and b), likely due to a poor establishment of grass species and thus its low productivity compared to the introduced legume. The Location F site with legume incorporation resulted in a smaller NEE observed compared to its control site (Location E) (Fig. 17), while this soil C sequestration was attributed mainly to greater C allocation to deeper roots less susceptible to decomposition rather than additional N fixed. These contrasting paired sites resulted in a large variation in the fine-tuned grass productivity (prdx) in addition to the ability to fix N and the greater C allocation to roots to represent the legume incorporation in these experiments. The subsequent scenario analysis therefore showed a large uncertainty in the legume treatment (Fig. 19), highlighting the need for further investigation of the efficacy of legume incorporation on soil C sequestration. Furthermore, the deep rooting characteristics of leguminous plants emphasise the importance of incorporating subsoils in the model structure, which may be more suited to long-term C sequestration than topsoils. In turn, EC flux towers have a great advantage of continuously capturing all the C assimilation including belowground biomass in deep soils, which is otherwise extremely difficult by manual sampling, providing essential data for further model improvement. Measure, Model and Verification - MMV approach to soil C methodology

[0250] The results of the experiments described herein demonstrate the use of DayCent-CABBI to forecast soil C sequestration and the advantage of NEE and AET data from multiple EC flux towers to calibrate and validate the model.

[0251] The C flux measurements using EC flux towers overcome the high spatial variability in C cycling at the paddock scale (-30-100 ha), which otherwise requires a large number of soil core sampling over relatively long time intervals to be able to detect changes in the SOC stock. The high-frequency C flux data afforded by EC measurements is a powerful and relatively inexpensive, low-maintenance tool to constrain the C cycling of process-based models which typically have a daily time step (Figs. 16 and 17), allowing the refinement of specific processes such as photosynthesis and soil heterotrophic respiration in contrast to solely relying on relatively coarse temporal observations of changes in SOC stocks. Furthermore, refining remotely sensed GPP using the EC flux data has the potential to aid in the scaling of the combined flux modelling approach to the landscape scale. Model parameters associated with SOC decomposition can be generalised across a region using the observed SOC stocks and EC flux data, and the plant community growth parameters can be calibrated using the remotely sensed GPP. The results of the experiments described herein show the applicability of replacing large-scale soil core sampling efforts with a combination of biogeochemical models and EC flux towers, reducing the costs and increasing the scalability of soil C sequestration projects.

[0252] To assess soil C sequestration projects, the effect of management change on the SOC stock should be differentiated from its response to year-to-year climatic variability. Rainfall (as a surrogate for soil moisture) is considered as a primary driver of SOC variability over time, interacting with soil properties to mediate plant productivity and thus C inputs, which often masks the influence of the management change. Establishing paired control (or reference) sites within a region allows monitoring changes associated with the underlying climatic conditions rather than the use of a static baseline SOC sampling to evaluate soil C sequestration by the management change. However, establishing a paired control site with environmental conditions close to the project site is often difficult and also doubles the measurement costs while halving the opportunity to sequester C, highlighting the advantage of modelling approaches to generate such dynamic control. The climate change scenario analysis in the experiments described herein demonstrated the dynamic change in the SOC stock in the control, resulting in different amounts of soil C sequestration by the introduced management change when compared to conventional static baseline analysis (Fig. 19 and Table 2). These results therefore suggest the scenario analysis using the calibrated DayCent to disentangle the interaction effects between management and climate on the change in SOC stocks as a cost-effective solution.

[0253] Another important feature of this approach is uncertainty assessment via Monte Carlo simulation propagating uncertainties in parameters and inputs. In the experiments described herein, the DayCent-CABBI model was adapted to the Brigalow Belt bioregion of Australia with a single parameter set of SOM decomposition while attributing uncertainty to the grass productivity parameter. Calibration of the model with broader coverage of regions would require accounting for the uncertainty in the parameters related to SOM decomposition as well as further data collection and modelling efforts to reduce such uncertainty. Another probable source of uncertainty is the initial size of the various SOC pools, which are often estimated by long-term simulations or “spin-up” runs to ensure the modelled SOC stock is in an equilibrium state. However, this model initialization approach also involves large uncertainty in the long-term state of ecosystems and alternative model initialization methods have been suggested such as the use of measurable SOC pools and replicating the land use history of the site.

[0254] In the experiments described herein, the size of the various SOC pools was established using a combination of the baseline total SOC stock and the SOC fractions observed at each site together with the land use history simulation (Figs. 18a and b). Whilst in-situ observation is the ideal situation considering the large variation in the SOC fractions, their estimates are available across Australia and could be provided as defaults and used in combination with land use history simulations. These results therefore suggest this Measure, Model and Verification (MMV) soil C methodology as a cost-effective and scalable approach to evaluation of soil C sequestration projects. Further methods

[0255] Fig. 30 illustrates method 3000 for determining a change in soil organic carbon quantity and / or a level of soil organic carbon of geographical area 110. Fig. 30 is to be understood as a blueprint for a software program and may be implemented step-by-step, such that each step in Fig. 30 is represented by a function in a programming language, such as, but not limited to, Python, FORTRAN, C++ or Java. It is noted that embodiments in relation to method 300 of Fig. 3 or method 400 of Fig. 4 may be similar or equivalent embodiments of method 3000.

[0256] Processor 121 provides 3001 an in-silico model of geographical area 110 calibrated using measurements received from one or more sensors (e.g., sensors 111, 112). The in-silico model may be configured to simulate soil carbon flux, such as soil carbon sequestration, within geographical area 110. The one or more sensors may be placed at: a first location having environmental conditions similar to geographical area 110; and / or a second location associated with geographical area 110. The measurements may be indicative of a level of carbon flux in geographical area 110. Processor 121 may provide 3001 an in-silico model similar or equivalent to the in-silico model described earlier in the disclosure. Processor 121 may provide 3001 an in-silico model by retrieving or receiving the in-silico model, or part thereof (such as one or more parameters that define the in-silico model) from memory or a database, or external communication.

[0257] Processor 121 provides 3002 input data indicative of current characteristics of geographical area 110 to the in-silico model. Processor 121 determines 3003 the change in soil organic carbon quantity and / or the level of soil organic carbon of the geographical area based on an output of the in-silico model. 3002 and 3003 of method 3000 may be similar or equivalent to 302 and 303 of method 300, respectively. In some embodiments, the measurements received from the one or more sensors may be indicator of one or more environmental conditions of geographical area 110.

[0258] Fig. 31 illustrates method 3100 for determining a change in soil organic carbon quantity and / or a level of soil organic carbon of geographical area 110. Fig. 31 is to be understood as a blueprint for a software program and may be implemented step-by-step, such that each step in Fig. 30 is represented by a function in a programming language, such as, but not limited to, Python, FORTRAN, C++ or Java. It is noted that embodiments in relation to method 300 of Fig. 3, method 400 of Fig. 4 or method 3000 of Fig. 30 may be similar or equivalent embodiments of method 3100.

[0259] Processor 121 calibrates 3101 an in-silico model of geographical area 110 using measurements received from one or more sensors (e.g., sensors 111, 112). The in-silico model may be configured to simulate soil carbon flux, such as soil carbon sequestration, within geographical area 110. The one or more sensors may be placed at: a first location having environmental conditions similar to geographical area 110; and / or a second location associated with geographical area 110. The measurements may be indicative of a level of carbon flux of the geographical area 110. Processor 121 provides 3102 input data indicative of current characteristics of geographical area 110 to the in-silico model. Processor 121 determines 3103 the change in soil organic carbon quantity and / or the level of soil organic carbon of geographical area 110 based on an output of the in-silico model. 3101, 3102 and 3103 of method 3100 may be similar or equivalent to 301,302 and 303 of method 300, respectively In some embodiments, the measurements received from the one or more sensors may be indicator of one or more environmental conditions of geographical area 110.

[0260] It is noted that method 3000 and method 3100 may be used in conjunction with method 400 to determine a total carbon footprint of an agricultural area (such as agricultural area 410). For example, processor 121 may determine 404 the total carbon footprint of agricultural area 410 based on a livestock carbon footprint and a change in soil organic carbon quantity and / or a level of soil organic carbon determined from method 3000 or method 3100.

[0261] It will be appreciated by persons skilled in the art that numerous variations and / or modifications may be made to the above-described embodiments, without departing from the broad general scope of the present disclosure. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

Claims

1. A method for determining a change in soil organic carbon quantity of a geographical area,the method comprising:calibrating an in-silico model of the geographical area using measurements received from one or more sensors, the in-silico model being configured to simulate soil carbon sequestration within the geographical area, wherein the one or more sensors are placed at: a first location having environmental conditions similar to the geographical area; and / or a second location associated with the geographical area, the measurements being indicative of: (a) a level of carbon flux; and optionally (b) one or more environmental conditions; of the geographical area;providing input data indicative of current characteristics of the geographical area to the in-silico model; anddetermining the change in soil organic carbon quantity of the geographical area based on an output of the in-silico model.

2. The method of claim 1, wherein the method comprises repeating the steps of calibrating the in-silico model, providing input data to the in-silico model and determining the change in soil organic carbon quantity upon receiving further measurements from the one or more sensors, such that the change in soil organic carbon quantity is quantified on a rolling basis.

3. The method of claim 1 or 2, wherein the input data comprises soil data corresponding to information regarding soil within the geographical area.

4. The method of claim 3, wherein the information regarding soil within the geographical area comprises information indicative of one or more of:soil texture;soil pH; andsoil bulk density.

5. The method of claim 3 or 4, wherein the method further comprises receiving soil data of the geographical area based on a result of soil testing in the geographical area.

6. The method of any one of the preceding claims, wherein the input data comprises plant data representing information regarding plants within the geographical area.

7. The method of any one of the preceding claims, wherein the input data comprises pasturedata representing information regarding pasture within the geographical area.

8. The method of any one of the preceding claims, wherein the measurements received from the one or more sensors corresponds to one or more of:weather data representing information regarding weather conditions within the geographical area; andrainfall data representing information regarding rainfall within the geographical area.

9. The method of claim 8, wherein the method further comprises receiving the weather data and / or the rainfall data periodically; and determining the change in soil organic carbon quantity of the geographical area upon receiving the weather data and / or the rainfall data periodically.

10. The method of any one of the preceding claims, wherein the in-silico model is indicative of a history and management of the geographical area.

11. The method of any one of the preceding claims, wherein the in-silico model is a digital twin of the geographical area.

12. The method of any one of the preceding claims, wherein the in-silico model is based on a time series biogeochemical model.

13. The method of any one of the preceding claims, wherein the one or more sensors is a flux tower.

14. A method for determining a total carbon footprint of an agricultural area, the method comprising:providing input data indicative of current characteristics of the agricultural area to an in-silico model of the agricultural area, wherein:the in-silico model is configured to simulate soil carbon sequestration within the agricultural area;the in-silico model is calibrated using measurements received from one or more sensors placed at: a first location having environmental conditions similar to the agricultural area; and / or a second location associated with the agricultural area; andthe measurements are indicative of: (a) a level of carbon flux; and optionally (b) one or more environmental conditions; of the agricultural area;determining a change in soil organic carbon quantity of the agricultural area based on an output of the in-silico model;determining a livestock carbon footprint by applying a carbon accounting model to livestock data, the livestock data representing information of livestock within the agricultural area and the livestock carbon footprint being indicative of an output of the carbon accounting model; anddetermining the total carbon footprint of the agricultural area based on the change in soil organic carbon quantity and the livestock carbon footprint.

15. The method of claim 14, wherein the method further comprises determining a carbon footprint of a product derived from the livestock from the agricultural area based, at least in part, on the total carbon footprint of the agricultural area.

16. The method of claim 15, wherein the product is one or more of:wool;meat; andmilk.

17. The method of any one of claims 14 to 16, wherein the livestock data comprises liveweight of at least one animal in the livestock.

18. The method of claim 17, wherein the liveweight of the at least one animal is based on walk over weighing.

19. The method of any one of claims 14 to 18, wherein the livestock data comprises liveweight gain of at least one animal in the livestock.

20. The method of any one of claim 14 to 19, wherein the livestock data comprises data indicative of one or more of:total livestock numbers;species of livestock;livestock added;livestock removed;livestock weights;wool shorn; andlivestock birthing rates.

21. The method of any one of claims 14 to 20, wherein the method further comprises monitoring a selection of animals in the livestock to determine at least part of the livestock data in real time.

22. The method of claim 21, wherein monitoring each animal in the selection of animals in the livestock is based on a radio frequency identification device placed on each animal.

23. The method of any one of claims 14 to 22, wherein determining the livestock carbon footprint comprises applying the carbon accounting model to one or more of:plant data representing information regarding plants within the agricultural area;fertiliser data representing information regarding fertiliser use within the agricultural area;energy consumption data; and fuel consumption data.

24. The method of any one of claims 14 to 23, wherein the method further comprises generating a price grid matrix based on the total carbon footprint of the agricultural area.

25. The method of any one of claims 14 to 24, wherein the method further comprises generating a report containing one or more of:the total carbon footprint of the agricultural area;the change in soil organic carbon quantity;the livestock carbon footprint;the carbon footprint of the product; anda further carbon footprint based on other sources of carbon within the agricultural area.

26. The method of claim 25, wherein the method further comprises providing a recommendation based on the generated report, the recommendation being indicative of a practice change to implement in the agricultural area to modulate the total carbon footprint of the agricultural area.

27. The method of claim 26, wherein the practice change comprises one or more of:pasture improvement;planting legumes;applying fertilisers;transitioning from cropping to permanent pastures; and time-controlled grazing.

28. The method of any one of claims 25 to 27, wherein the total carbon footprint of the agricultural area; the change in soil organic carbon quantity; the livestock carbon footprint; and / or the further carbon footprint based on other sources of carbon is represented in carbon credits, Australian Carbon Credit Units (ACCUs) or carbon dioxide equivalent (CChe).

29. Software that, when executed by a computer, causes the computer to perform the method of any one of the preceding claims, or part thereof.

30. A system for determining a change in soil organic carbon quantity of a geographical area, the system comprising:a processor configured to:calibrate an in-silico model of the geographical area using measurements received from one or more sensors, the in-silico model being configured to simulate soil carbon sequestration within the geographical area, wherein the one or more sensors are placed at: a first location having environmental conditions similar to the geographical area; and / or a second location associated with the geographical area, the measurements being indicative of: (a) a level of carbon flux; and (b) one or more environmental conditions; of the geographical area;provide input data indicative of current characteristics of the geographical area to the in-silico model; anddetermine the change in soil organic carbon quantity of the geographical area based on an output of the in-silico model.

31. A system for determining a total carbon footprint of an agricultural area, the system comprising:a processor configured to:provide input data indicative of current characteristics of the agricultural area to an in-silico model of the agricultural area, wherein:the in-silico model is configured to simulate soil carbon sequestration within the agricultural area;the in-silico model is calibrated using measurements received from one or more sensors placed at: a first location having environmental conditions similar to the agricultural area; and / or a second location associated with the agricultural area; andthe measurements are indicative of: (a) a level of carbon flux; and optionally (b) one or more environmental conditions; of the agricultural area;determine a change in soil organic carbon quantity of the agricultural area based on an output of the in-silico model;determine a livestock carbon footprint by applying a carbon accounting model to livestock data, the livestock data representing information of livestock within the agricultural area and the livestock carbon footprint being indicative of an output of the carbon accounting model; anddetermine the total carbon footprint of the agricultural area based on the change in soil organic carbon quantity and the livestock carbon footprint.

Images