Detailed case study: assess impact of less livestock and change in diet on soil organic carbon

Scenario modelling for assessing impacts of policy changes and socio-economic effects on ecosystem services of soils Deliverable D6.4 Detailed case study: assess impact of less livestock and change in diet on soil organic carbon Due date of deliverable: Month 57 of EJP SOIL Actual submission date: 31.10.2024 GENERAL DATA Grant Agreement: 862695 Project acronym: EJP SOIL – SIMPLE Project title: Scenario modelling for assessing impacts of policy changes and socioeconomic effects on ecosystem services of soils Project website: www.ejpsoil.eu Start date of the SIMPLE project: November 1st, 2022 Project duration: 24 months Name of lead contractor: AGROSCOPE Funding source: This project has received funding from the European Unions’ Horizon 2020 research and innovation programme under grant agreement No. 862695 EJP SOIL. Type of action: European Joint Project COFUND DELIVERABLE NUMBER: D6.4 DELIVERABLE TITLE: Detailed case study: assess impact of less livestock and change in diet on soil organic carbon DELIVERABLE TYPE: Report WORK PACKAGE N: WP6 WORK PACKAGE TITLE: Land-use change DELIVERABLE LEADER: Flanders Research Institute for Agriculture, Fisheries and Food, EV-ILVO, Belgium AUTHOR: DOI: Kaat Mertens, Ioanna Panagea, Greet Ruysschaert 10.5281/zenodo.14142841 DISSEMINATION LEVEL: CO/PU ABSTRACT In this study, the impact of an adjustment in the composition of cattle feed on the soil organic carbon stocks in Flanders (Belgium) was investigated. Silage maize is the most common fodder crop, and even the most common crop cultivated in Flanders. However, silage maize does not have a high contribution to the SOC storage in arable soils compared to some other crops, such as cereals and fodder legumes. In this study, we have investigated the impact on SOC stocks of replacing silage maize by winter barley, with straw residues removed after harvest, on 10% of the parcels where it is currently cultivated. The impact of the shift in fodder crops was simulated with a Roth-C based model. The Belgian Soil Organic Carbon Calculator (BeSOCC) integrates the Roth-C model, with an approach for the initialisation and the calculation of carbon inputs by crops and fertilization developed specifically for Belgium, Flanders that builds further on earlier work for developing a digital decision support tool for farmers to simulate the evolution of the SOC in arable land. The scenario was compared with a business-as-usual (BAU) scenario. The BAU scenario was developed based on the management data of 5 years (2018-2022). Due to missing crop and fertilization data, simulations could be performed for 52.3% of the arable fields. The simulations were performed at parcel-level and afterwards aggregated to NUTS3 level. The results indicate that an increase of the SOC storage will occur in all NUTS3 areas if there is a shift in the fodder crops. This increase is on average 0.013 Mg C. ha-1. Yr-1 and ranges between 0.003 and 0.025 Mg C. ha-1. Yr-1 . Table of Contents List of Tables ............................................................................................................................................ 4 List of Figures ........................................................................................................................................... 4 List of acronyms and abbreviations ......................................................................................................... 4 1. Introduction .................................................................................................................................... 5 2. Methodology .................................................................................................................................. 5 2.1. BeSOCC model ....................................................................................................................... 5 2.2. Baseline scenario ................................................................................................................... 7 2.3. Cattle feed adjustment scenario .................................................................................................. 8 3. Results and discussion .................................................................................................................... 9 4. Conclusion .................................................................................................................................... 13 List of references ................................................................................................................................... 15 List of Tables Table 1: Parcel history and associated division of the organic carbon stocks over DPM, RPM, BIO and HUM ........................................................................................................................................................ 6 Table 2: Clay content and basic bulk density for each texture class ...................................................... 6 Table 3: Monthly average climatological data for Flanders as used in the BeSOCC model ................... 8 Table 4: Results of Business-as-usual (BAU) scenario and cattle feed adjustment scenario ................ 13 List of Figures Figure 1 : ∆SOC-stocksyearly per NUTS3 level in Flanders for the Business-as-usual scenario. ................ 9 Figure 2: ∆SOC-stocksyearly per NUTS3 level in Flanders for the cattle feed adjustment scenario. ....... 10 Figure 3:The SOC accrual of the cattle feed adjustment scenario per NUTS3-level. ............................ 11 List of acronyms and abbreviations SOC Soil Organic Carbon BeSOCC Belgian Soil Organic Carbon Calculator DPM Decomposable Plant Material RPM Resistant Plant Material HUM Humified Organic Matter BIO Microbial Biomass IOM Inert Organic Matter BAU Business-as-usual 1. Introduction In Flanders, after grass, silage maize is the most important component in the feed composition on many livestock farms and the most common fodder crop used as a basis for cattle feed (ILVO, 2024). It is also the most cultivated crop in Flanders’ arable fields. It is cultivated in more than 11000 farms occupying an area of about 120 000 ha (Agentschap Landbouw en Zeevisserija, 2024). In Flanders forage maize is a high-energy, high-starch, low-protein forage crop, commonly used for cattle, often in combination with complementary high protein crops like alfalfa. The adoption of silage maize is rising because of its high energy content and easy digestibility among ruminants. However, despite its numerous advantages, silage maize has a limited capacity to contribute to soil organic carbon (SOC) storage compared to other crops (Kan et al., 2022). SOC storage is important for maintaining soil health and increasing the SOC stocks could aid in the mitigation of climate change by sequestering additional carbon. Replacing silage maize by another fodder crop could lead to an improvement of the SOC storage. However, completely replacing silage maize is impractical for many farmers, who rely on it as a staple forage crop. Therefore, this study examines a more realistic approach to enhancing SOC storage by replacing silage maize with winter barley on 10% of the parcels currently dedicated to it. To assess the impact of this partial shift on SOC stocks evolution in Flanders, we used a Roth-C-based model. 2. Methodology To estimate the impact of an adjustment in composition of cattle feed, a scenario was developed to simulate a shift in fodder crops cultivated in Flanders. The scenario was then simulated with a Roth-C based model developed specifically for Flanders (BeSOCC) and compared to a baseline scenario. 2.1. BeSOCC model The Belgian Soil Organic Carbon Calculator (BeSOCC) integrates the Roth-C model, with an approach for the initialisation and the calculation of carbon inputs by crops and fertilization developed specifically for Belgium, Flanders and builds further on earlier work for developing a digital decision support tool for farmers to simulate the evolution of the SOC in arable fields (Beirinckx et al. 2023; UGent and BDB, 2006; Verlinden et al., 2013). The Roth-C model is a model that simulates the turnover of organic carbon in the top-layer of non-saturated mineral soils (Coleman and Jenkinson, 2014). The model considers different parameters, such as clay content, temperature, monthly rainfall, soil cover and carbon input by crops and organic amendments to calculate the SOC stock. The model divides the SOC stock in four active pools (i.e., DPM, RPM, BIO and HUM) and one inactive pool (i.e., IOM). Each active pool decomposes according to a first-order process with their own characteristic rates. The incoming carbon from organic fertilization and plants are divided over the DPM and RPM pools. The division is based on the DPM/RPM ratio which gives an estimation of the decomposability of the incoming organic matter. At initialization the BeSOCC model uses the parcel history to distribute the initial SOC stock over the DPM, RPM, BIO and HUM pool (Table 1). Table 1: Parcel history and associated division of the organic carbon stocks over DPM, RPM, BIO and HUM compartments in the BeSOCC model. Parcel history %OC in DPM %OC in RPM %OC in BIO %OC in HUM Recently torn (< 6 years ago) permanent grassland 1 47 1 51 Arable field 1 15.5 1.5 82 In the BeSOCC model the incoming C coming from crop residues (including roots and exudates) and organic fertilizers is calculated using look-up tables where each crop and fertilizer type is linked to a specific C-content and DPM/RPM ratio. The look-up table for the crops was recently updated by Maenhout et al. (in preparation) and is also used for the local CAP eco-schemes. In the list each crop is assigned a specific C-content value a crop supplies to the soil based on findings from field experiments, expert knowledge or allometric functions. The DPM/RPM ratios were calculated based on the method of Dechow et al. (2019). The list was extended to include more crop types cultivated in Flanders by assigning crops with no values to similar crops that do have a value in the study by Maenhout et al. (In preparation). The look-up table for crops can be found on Zenodo (10.5281/zenodo.13929423). The look-up table for the fertilizers was first developed by the University of Ghent and soil service of Belgium (Ugent and BDB, 2006) and later updated and extended by the University of Ghent and Flemish Land Agency (Verlinden et al., 2013). Similar as for the crops, the look-up table for the organic fertilizers links the fertilizer type to a C-content, DPM/RPM ratio and a percentage of organic C going directly into the HUM pool. The look-up table for the organic fertilizers can be found on Zenodo (10.5281/zenodo.13988649). The soil data required for the Roth-C model is the clay content and basic bulk density. The basic bulk density ( kg dm-³) is the mass of soil per volume unit under the hypothetical conditions that the SOC would be 0%. The clay content and basic bulk density are used for the initialisation of the SOC stock distribution over the Roth-C pools and for the decomposition of SOC stock in the active Roth-C compartments. The BeSOCC model uses a look-up table to derive these input parameters from the texture class (Table 2). This table was developed by a study by Sleutel et al. , 2006. Table 2: Clay content (clay%) and basic bulk density (SGb) (at SOC of 0%) for each texture class as used in the BeSOCC model (Sleutel et al. 2006). Main texture class Average clay percentage (%) Average SGb (kg dm-³) Sand 4,7 1,55 Sandy loam 9,0 1,33 Loam 14,0 1,40 Clay 23,9 1,40 2.2. Baseline scenario For the baseline scenario, business-as-usual (BAU) management practices were simulated for each arable field in Flanders. To develop the BAU scenario the crop rotation and fertilization of the 2018-2022 period, obtained from the LPIS data (Agentschap Landbouw en Zeevisserijb, 2024) and the fertilization allocation model (BAM; Van Opstal et al. 2014) respectively. The LPIS data contain the crops and cover crops cultivated on each agricultural parcel each year, as well as information that indicates if a field is recently (< 6 years) converted from permanent grassland. The boundaries of agricultural parcels can change every year. For the baseline scenario the field boundaries of 2022 were chosen and for the years 2018-2021 the dominant crops within these boundaries were selected. The LPIS data do not contain information on the straw removal after harvest of a crop. Thus, in the baseline scenario it was assumed that for all cereals the straw residues are removed after harvest. The crops’ sowing and harvesting month are required for a simulation with the BeSOCC model. A look-up table with each crops’ sowing and harvesting months was developed for Flanders’ crops by combining data required from the Agency of Agriculture and Fisheries and studies by Verlinden et al. (2013) (Demeter) and Cecelja et al. (2019) (C-factor studie) (10.5281/zenodo.13929370). The fertilization allocation model (BAM) was developed by the Flemish Environment Agency. The model considers the total amount of fertilizer used per farm and per year and the crops at plot level and based on expert rules make a rational estimate of the amount and type of fertilizer applied to each plot, as well as the time of application. It considers the available information on fertilizer production, fertilizer use, fertilizer transport and fertilizer storage at farm level ,as well as the fertilization standards. (Van Opstal et al. 2014 and VMM 2016). The data obtained from the BAM model contains many fertilization applications with a dose of less than 1 Mg fertilizer/ha. A threshold for the dose of fertilizer was set at 1 Mg/ha, since this is an unrealistically low dose for organic fertilizers. The soil data was obtained from map layers available by the (sub)soil database of Flanders (DOV). The initial SOC stocks were obtained from the Soil Organic Carbon Stock Maps for Belgium with a resolution of 40m x 40m (DOVa, 2024). The texture classes were obtained from the digital soil map of the Flemish region which has a resolution of 1:20 000 (DOVb, 2024). Based on the selected field boundaries of 2022 the dominant texture class was chosen, and the mean SOC-stock was calculated for each field. The Roth-C model is unable to perform simulations on waterlogged soils and also our focus is on mineral soils. For the baseline scenario we assumed that, if the SOC is more than 10%, it is considered peat (Tits et al., 2020) and thus assumed to be waterlogged. For that a threshold at 10% was set for the SOC content in the baseline scenario. The climatological data needed to conduct simulations with the Roth-C model were obtained from the Joint Research Centre (JRC) Agri4Cast dataset (agri4cast.jrc.ec.europa.eu). This database contains daily observations from weather stations interpolated on a 25x25 km grid for the whole of Europe. Flander’s climatological data was derived from this dataset by calculating the daily average rainfall, air temperature and evapotranspiration of the whole region. For all three parameters the monthly averages of 30 years (1992-2021) were calculated and used as input parameters (Table 3). Table 3: Monthly average climatological data for Flanders as used in the BeSOCC model (derived from Joint Research Centre (JRC) Agri4Cast (agri4cast.jrc.ec.europa.eu)). Month Temperature (°C) Precipitation (mm) Evapotranspiration (mm) January 4.06 64.21 15.27 February 4.77 57.19 22.29 March 7.20 49.43 43.74 April 10.24 39.24 72.21 May 13.83 55.25 99.95 June 16.80 62.10 111.51 July 18.72 70.46 114.87 August 18.57 71.99 97.56 September 15.63 60.98 61.72 October 11,88 66.66 35.66 November 7,68 65.09 16.88 December 4,76 74.33 12.77 The 5-year management practices were simulated for each parcel over a period of 30 years. Due to missing C-inputs coming from some crops and fertilizers, simulations could be performed for 52.3% of the arable fields. The ∆SOC-stock (Mg C. ha-1. yr-1) was calculated for each parcel and then aggregated to a NUTS3level. 2.3. Cattle feed adjustment scenario In Flanders, the most common crop is silage maize, covering 37% of the region’s arable parcels. However, its contribution to the SOC storage in arable soils is relatively low compared to some other crops, such as fodder legumes and cereals, especially when the straw from these crops is left on the field after harvest (Panagea et al., submitted). In the investigated scenario, we assumed that some farmers will try to improve their soils by diversifying their crop rotation with crops with higher C inputs into the soil. It was decided to use a cereal, and more specifically winter barley as a replacement for silage maize on 10% of the silage maize parcels. Winter barley is a cereal commonly used as a fodder crop. We assumed that the straw residues of winter barley would be removed from the field after harvest, since this is common practice in Flanders. In the selected parcels the cover crops from the previous year were removed from the rotation as winter barley is normally sown in the previous year and harvested in August. According to Jeangros and Courvoisier (2019), it is not the most beneficial option to sow a cereal after the harvest of winter barley. In this scenario, the assumption was made that if a cereal was declared as an aftercrop after the replaced silage maize in a parcel, it would be removed from the crop rotation. However, if the following year a cereal was declared as main crop in the LPIS data, and thus sown the same year as winter barley’s harvest, it was not removed from the rotation. Each year in the five-year crop rotation, all the fields with silage maize were selected for this scenario and from these fields, 10% were randomly selected to replace silage maize by winter barley. This five-year crop rotation was then repeated for a duration of 30 years in the simulation with the BeSOCC model. The selection and simulation with the BeSOCC model was repeated five times. For each field the mean ∆SOC-stock, ∆SOC-stockyearly and SOC accrual of the five repeated simulations was calculated and then aggregated to a NUTS3-level. The ∆SOC- stock (Mg C. ha-1) is the change in SOC found after 30 years. The ∆SOC-stockyearly (Mg C. ha-1. Yr-1) is the annual change in SOC and is calculated with the following equation: ∆𝑆𝑂𝐶 𝑠𝑡𝑜𝑐𝑘𝑦𝑒𝑎𝑟𝑙𝑦 =∆SOC stock 30 The SOC accrual (Mg C. ha-1. Yr-1) gives the annual increase in SOC stock found in the cattle feed adjustment scenario in comparison to the baseline scenario (Don et al. 2023). It can be calculated with the following equation: 𝑆𝑂𝐶 𝑎𝑐𝑐𝑟𝑢𝑒𝑙 = ∆SOC stock𝑦𝑒𝑎𝑟𝑙𝑦,𝑐𝑎𝑡𝑡𝑙𝑒 𝑓𝑒𝑒𝑑 𝑎𝑑𝑗𝑢𝑠𝑡𝑚𝑒𝑛𝑡 − ∆𝑆𝑂𝐶 𝑠𝑡𝑜𝑐𝑘𝑦𝑒𝑎𝑟𝑙𝑦,𝑏𝑎𝑠𝑒𝑙𝑖𝑛𝑒 3. Results and discussion The simulated BAU scenario resulted in a mean ∆SOC-stock of 3.12 Mg C. ha-1 for all of Flanders. This indicates that if the current management practices will continue to be applied for the next 30 years there will be an annual increase in SOC storage of 0.104 Mg.ha-1.yr-1. However, Figure 1 illustrates that the ∆SOC-stock differs between the different NUTS3 regions. In the NUTS3 regions BE256 (District Roeselare) and BE213 (District Turnhout) we expect a decrease in SOC-stocks under BAU, while in BE242 (District Leuven) an increase is expected (Table 4). Figure 1 : ∆SOC-stocksyearly (Mg C. ha-1. yr-1) per NUTS3 level in Flanders obtained with a simulation with the Business-as-usual scenario. In the BAU scenario, silage maize is cultivated each year on an average area of 103,635 ha. In the scenario with a shift in fodder crops on average 9,150 ha of this area were replaced by winter barley. This shift in fodder crop rotation resulted in an increase of Flanders’ ∆SOC-stock of 3.51 Mg C. ha-1 after (1) (2) Maenhout P., De Boever M., Tits M., De Mets L., Salomez J., Sleutel S., Ruysschaert G., in preparation. Effective organic carbon input to soils by arable crops, grasses and cover crops – Available knowledge in Flanders, ILVO mededeling. Panagea, I., Quataert, P., Alonso-Ayuso, M., Bárcena, T., De Boever, M., Mariangela, D., Jacobs, A., Jensen, J., Seidel, F., Seitz, D., Spiegel, H., Vanden Nest, T., Don, A., Ruysschaert, G., submitted. Forage legumes in crop rotation increase soil organic carbon stocks–Evidence from 21 European long-term field experiments. European Journal of Soil Science. Poeplau C., & Don A. (2015). Carbon sequestration in agricultural soils via cultivation of cover crops–A meta-analysis. Agriculture, Ecosystems & Environment, 200, 33-41. https://doi.org/10.1016/j.agee.2014.10.024. Sleutel S, De Neve S, Beheydt D, Li C, Hofman G. (2006). Regional simulation of long-term organic carbon stock changes in cropland soils using the DNDC model: 2 Scenario analysis of management options. Soil Use Manage. 22, 352-361. https://doi.org/10.1111/j.1475-2743.2006.00045.x. Tits M., Elsen A., Deckers S., Bries J., Vandendriessche H. 2020. Bodemvruchtbaarheid van de Akkerbouwen Weilandpercelen in België en Noordelijk Frankrijg (2016-2019). Publicatie van de Bodemkundige Dienst van België. 235 pp. Universiteit Gent: Vakgroep Bodembeheer en Bodemhygiëne en Bodemkundige Dienst van België (2006). Ontwikkelen van een expertsysteem voor het adviseren van het koolstofbeheer in de landbouwbodems. Rapport in opdracht van Departement Leefmilieu, Natuur en Energie. 146p. Van Opstal M., Tits M., Beckers V., Elsen A., Van Overtveld K., Batelaan O., Van Orshoven J., Bries, J., Vandendriessche, H., Diels, J. (2014). Vernieuwde kwantificering van de verliezen van N en P vanuit de landbouw naar het oppervlaktewater. Eindrapport, mei 2014. Studie uitgevoerd in opdracht van de Vlaamse Milieumaatschappij door het Departement Aarden Omgevingswetenschappen (KU Leuven) en de Bodemkundige Dienst van België. Verlinden G, De Neve S, Postma R, Germeys D, Sleutel S, Hanegraaf M, et al. DEMETER: sustainable and integrated soil management to reduce environmental effects. In: D’Haene K, Vandecasteele B, De Vis R, Crappé S, Callens D, Mechant E, et al., editors. NUTRIHORT : nutrient management, innovative techniques and nutrient legislation in intensive horticulture for an improved water quality : proceedings. Merelbeke, Belgium: Institute of Agricultural and Fisheries Research (ILVO); 2013. p. 307–13. Vlaamse Milieumaatschappij, 2016. Bronnen van waterverontreiniging in 2016. Bronnen van waterverontreiniging 2016 — Vlaamse Milieumaatschappij (accessed 25 October 2024).