1 Introduction

Soil quality is a major determinant of crop productivity, farm resilience, and environmental quality of agricultural ecosystems (Karlen et al. 1997; Kumar and Karthika 2020; Stevens 2018). Current unsustainable farming practices threaten soil quality worldwide (Koch et al. 2013). Subsoil compaction caused by heavy field traffic, loss of soil organic matter, acidification, decline of biodiversity, nutrient losses, and erosion are examples of such threats (Kumar and Karthika 2020; McBratney et al. 2014; Ros et al. 2022; Yang et al. 2020). Deteriorated chemical, physical, and biological soil functions will likely have an adverse impact on crop production in the long term. Changes in the climate, like increased occurrence of droughts and floods, and a growing world population may further add to these threats at global and regional levels, making it more challenging to achieve food security (Amundson et al. 2015; McBratney et al. 2014; Wall and Smit 2005).

Sustainable soil management has to overcome these threats by meeting present productivity needs without compromising soil quality for future generations (adapted from Smith and Powlson 2007). Various studies have shown that good soil management practices can boost soil quality and mitigate the adverse impact of agriculture on the environment. Examples include practices to increase soil organic matter (SOM), to minimize synthetic inputs, to reduce soil disturbance, to keep soil covered, and to diversify crops and crop rotations (Adetunji et al. 2020; Castellazzi et al. 2008; Silva et al. 2021). Given the long history of efforts to embed improved soil management measures into mainstream agriculture, there are abundant studies that deal with barriers that limit the adoption of sustainable soil management in agriculture. Currently, the implementation of sustainable soil management is not evident due to the following six key limiting factors.

  1. (1)

    Insufficient insights into the relationship between soil quality and farmers’ production management, namely the complete set of physical and non-physical inputs made by the farmer (Brady et al. 2015; Kik 2021a). For example, soil compaction from field traffic is difficult to measure (Rücknagel et al. 2015) and its impact on crop yield and quality is not always clear.

  2. (2)

    Interactions among soil quality indicators (Bouma 2014; Stevens 2018). Trade-offs exist where improvement of one soil quality indicator might come at the expense of another. For example, cover crops have a beneficial impact on soil structure, SOM, and reduce nitrate leaching, but they might serve as a host crop for plant-parasitic nematodes, thereby reducing crop yield (Adetunji et al. 2020; Puissant et al. 2021).

  3. (3)

    Trade-offs between long-term soil quality (> 10 years) and annual farm income (Bos et al. 2017; Kik 2021a; Stevens 2022). Pressure on short-term profit resulting from current narrow crop margins makes it difficult to invest in long-term soil quality, in particular when farmers do not own the fields they cultivate (Stevens 2022). Potential gains from increased soil quality are uncertain and typically manifest in the long term. For example, a higher organic matter input from compost, crop residue incorporation, or cover crops will increase SOM, but it is costly in the short term (Kik et al. 2024).

  4. (4)

    Trade-offs between agricultural productivity and environmental quality (O’Sullivan et al. 2018; Schulte et al. 2014; Stevens 2018). Optimizing soil quality for agricultural productivity might come at the expense of environmental quality. For example, an economically optimal crop nitrogen (N) fertilization level might come at the cost of high N-losses to the environment (Silva et al. 2021).

  5. (5)

    Proper implementation of sustainable soil management strongly depends on the initial soil quality and current production management. Therefore, production management decisions must be tailored at farm level to effectively achieve soil quality targets (Young et al. 2021). If these target values are not defined and the impact of production management is highly site-dependent, general implementation is undesirable (Hannula et al. 2021). For example, the impact of carbon sequestration strongly depends on the current carbon saturation level of the soil (Lessmann et al. 2022; Moinet et al. 2023).

  6. (6)

    Lock-in, being defined as “a self-reinforcing mechanism that reproduces the status quo and impedes change” (Weituschat et al. 2022). Farmers’ decisions are driven by a wide range of factors including technical developments, policies, and economic context. These factors can interact to create a cognitive lock-in, in which current production management is maintained (Weituschat et al. 2022).

At the farm-level, achieving a sufficient yearly income and long-term continuity of the farm are prime goals for farmers (Kay et al. 2012). Regarding these goals, the implementation of sustainable soil management can be considered an economic problem. Considering sustainable soil management as a socio-economic problem, the challenge is to implement the right production management decisions at the farm-level that ensure (1) profit for a sufficient yearly farm income and (2) long-term preservation of soil quality. Such a challenge is typically addressed via bio-economic modeling, which is among the most popular farming system redesign methods (Janssen & van Ittersum 2007). Integrative bio-economic models evaluate trade-offs and synergies between different farm management strategies, but most of these models only make tenuous references to soil quality (Schreefel 2022). Examples of such studies can be found in Bos et al. (2017), Dogliotti et al. (2005), and Mandryk et al. (2014). According to Schreefel (2022), models that focus on the assessment of soil quality and its multifunctionality largely lack integration of the environmental impact and socio-economic impact at farm level. They therefore integrated the Soil Navigator soil assessment model (Debeljak 2019) into the FARMdesign bio-economic farm model (Groot et al. 2012), thereby optimizing the multifunctionality of soil using nutrient flows and soil organic matter as soil quality indicators. The Soil Navigator uses qualitative decision rules to assess the effects of management on soil functions. However, making the right production management decisions at farm level requires quantitative relationships between management and soil quality to ensure sufficient farm profit and preservation of soil quality. Therefore, our aim is to build on Schreefel (2022) approach by using the FARManalytics bio-economic modeling approach developed by Kik et al. (2024)

The aim of this study is to optimize production management on representative Dutch arable farms to maximize farm profit while increasing long-term soil quality. We define three sub-objectives:

  1. 1.

    Describe nine existing farms in the Netherlands and calculate the impact of their current production management on soil quality and farm economics.

  2. 2.

    Optimize production management on these farms using different scenarios to achieve highest profit while increasing soil quality

  3. 3.

    Explore alternative production management on these farms for (1) policy scenarios implementing stricter water quality regulations, and (2) field beans as an alternative crop.

We selected the Netherlands as a case study region (Fig. 1). The Netherlands is a suitable case study region to evaluate potential sustainable farm management strategies because of its high demand for agricultural products and its intensive land use due to high land prices. We assume the farmer to be the primary actor and decision-maker with regard to sustainable soil management (Kik 2021b). Moreover, we assume farmers to be financially rational decision-makers. In this study, we focus solely on agricultural productivity and do not consider other ecosystem services as a potential additional source of farm income.

Fig. 1
figure 1

Soil management on a Dutch arable farm (picture: Maarten Kik)

Full size image

2 Methods

2.1 Model description FARManalytics

FARManalytics is a bio-economic modeling approach that allows to analyze and optimize farmers’ production management to maximize farm profit while preserving or increasing soil quality. Production management decisions included crop rotation, cover crops, manure, fertilizer, and crop residue management. Definitions and illustrations of these decisions can be found in Appendix 1. For an extensive description of the model, we refer to Kik et al. (2024). The two main inputs of FARManalytics are a soil quality indicator set (Soil quality) and an economic calculation framework (Farm economics). The key features or FARManalytics are explained in the “FARManalytics” section.

2.1.1 Soil quality

The soil quality indicator set consists of 15 chemical, physical, and biological indicators based on studies by de Haan et al. (2021) and Ros et al. (2022). Examples of these indicators are phosphorous (P) availability, soil organic matter (SOM), subsoil compaction vulnerability, and plant parasitic nematodes. Indicators and their respective target ranges are used to assess soil quality at a specific location at a certain point in time. Based on the indicators, constraints set requirements on farmers’ production management to ensure the indicators stay in or move towards the target range. For example, if the indicator potassium (K) availability is below the minimum target, the constraint K fertilization advice is formulated in such a way that the K availability moves towards the target range. If K availability is within the target range, K fertilization is formulated in such a way that the K availability remains in the target range. To apply the constraints, quantitative rules on the impact of production management decisions on soil quality were developed. These entail for example the K content of fertilizers and K removal via harvested crops in order to calculate the K fertilization advice.

2.1.2 Farm economics

The economic calculation framework calculates the contribution of production management decisions to farm income. The indicator for farm income is profit, which is the gross income minus all costs, including opportunity costs for own labor and capital (Kay et al. 2012). Farm profit is calculated as the sum of profit generated through crops and selling crop residues minus costs for cover crops, manure, and fertilizer. Farm overhead costs are neglected, the calculated profit therefore is a profit on crop enterprise, being the total profit on farm level excluding costs that cannot be attributed to the considered production management decisions. The calculation framework uses Activity-Based-Costing to accurately attribute fixed costs towards the respective production management decisions (Drury 2008).

2.1.3 FARManalytics

FARManalytics is a bio-economic model with the scope at farm level. At farm level, farms are assumed to be homogeneous in their soil type and production management. The temporal scale is the time length of a crop rotation, typically 8 to 12 years. Within this time, we assume production management (PM) to be static. Input prices and output prices, i.e., crop prices are assumed static. Yields are determined using a target-oriented approach, which implies that yields are static and do not respond to changes in soil quality or production management (van Ittersum and Rabbinge 1997). Figure 2 presents a conceptual outline of the FARManalytics model.

Fig. 2
figure 2

Overview of the main inputs (farm economics and soil quality) and conceptual outline of the FARManalytics bio-economic model. The objective of FARManalytics is to optimize farmers’ production management (PM) to achieve maximum farm profit while maintaining or increasing soil quality

Full size image

FARManalytics consists of two modules: PM calculator and PM optimizer. In the PM calculator, the production management decisions for crop rotation, cover crops, manure and fertilizer application, and crop residue management are fixed inputs. The PM calculator allows to gain insight into whether current production management keeps soil quality indicators in their target range. The PM optimizer is an optimization model that selects the most appropriate production management fulfilling the soil quality constraints (Soil quality) while generating highest farm profit (Farm economics). Figure 2 presents an outline of the PM optimizer module. The first process in the PM optimizer is to generate all feasible crop rotations using ROTAT+ (Dogliotti et al. 2003). The input for ROTAT+ is a list with crops and the respective plant and harvest dates, maximum crop frequency, and minimum period of return. The second process is to optimize the selection for cover crops, manure and fertilizer application, and crop residue management within every feasible crop rotation using Mixed-Integer-Linear-Programming. The output from the soil quality indicator set (Soil quality) and farm economics (Farm economics) is used as an input. The result of process two is a collection with all feasible crop rotations with optimized cover crop, manure, fertilizer, and crop residue choices. Process three is to select the best production management set in this collection: the set achieving maximum profit while fulfilling all soil quality constraints.

2.2 Case farms and data collection

The case farms were participating in the “Farmers Network for Soil Sampling” initiated by Wageningen Plant Research. In 2019, nine out of in total 16 farmers in this network responded positively to a call to participate in this study. In 2020, we collected the current production management data on all farms, which is presented in Table 1. Soil samples were obtained from each farm, being available from routine soil sampling events done every 4 years. We also gathered information on current production management decisions, including crop rotations, manure and fertilizer application (type, quantity, timing, and application technology), and types of cover crops. This data was sourced from farm management systems and through semi-structured interviews with the farmers. In 2021, we made an economic analysis based on current production management. Economic data was obtained through structured interviews with the farmers using templates. The templates were completed based on estimations made in consultation with the farmers, as much of the required economic data was not readily available. For the economic data such as costs of inputs, crop prices, and crop yields, we used averages from 2019, 2020, and 2021. Following completion of the analysis, the results were reviewed with the farmers in a semi-structured format. By combining soil sample data with current production management information, we identified soil quality limitations and challenges. Again, the results were shared and discussed with farmers in a semi-structured interview. In 2023, we presented the FARManalytics scenario results to the farmers. In an unstructured interview, we also asked for their input on a scenario addressing a specific challenge on their farm.

Table 1 Farm characteristics and current production management decisions for nine arable case study farms from the Netherlands. Detailed manure applications can be found in Appendixes 2 and 4. Crop acronyms: CA carrots, GB green beans, GC grass clover, GS grass seed, KC corn (kernel), PE peas, SC silage corn, S seed onions, SP seed potato, SB sugar beets, SP spring barley, SPI spinach, STP starch potato, WB winter barley, WP ware potato, WW winter wheat. Cover crop mixes: Mix 1 A. Strigosa, vetch, clover, white radish; Mix 2 A. Strigosa, flax, vetch, clover, peas; Mix 3 oats, Vetch, Phacelia; Mix 4 = Oats, yellow mustard; Mix 5 = clover, vetch, flax, sorghum, peas
Full size table

2.3 Scenarios

We defined four farm-level scenarios to study the impact on soil quality and farm economics (Farm scenarios). Subsequently, we study the impact of two policy developments on soil quality and farm economics (Stricter water quality regulations and Field beans as part of protein transition sections).

2.3.1 Farm scenarios

The selected farm scenarios consist of a baseline scenario, a soil quality tactical scenario, and soil quality strategic scenario, which were the same for all farms. The farm strategic scenario consists of a farm-specific challenge. The precise definition of the farm scenarios is as follows:

  • Baseline: Continuation of current production management during one complete rotation. Calculated with the FARManalytics module PM calculator.

  • Soil quality tactical: Maximize farm profit with soil quality constraints applied. Calculated with the FARManalytics module PM optimizer. Decision variables are cover crops, manure, fertilizer, and crop residue management because these are typical production management choices that can be changed in the tactical dimension.

  • Soil quality strategic: Additional decision variable compared to “soil quality tactical” is crop rotation.

  • Farm strategic: Maximize farm profit with soil quality constraints applied for a farm-specific challenge. Decision variables are equal to the “soil quality strategic” scenario. Inputs and constraint settings ware made in consultation with the farmers. Table 2 details the farm-specific challenges for each farm.

Table 2 Challenges to be modeled with bio-economic modeling approach FARManalytics in the farm strategic scenarios, which include a farm-specific challenge to maximize profit while preserving or increasing soil quality
Full size table

2.3.2 Stricter water quality regulations

The first policy scenario is the stricter water quality scenario. In it, we explore the impact of two recent policy measures aimed at improving surface and ground water quality in agricultural ecosystems in the Netherlands. We only chose to implement this scenario on the farms S, CR, S-CR, L, and CSW-1 because these farms represent different soil types and are considerably different in their current production management. The studied policy measures are as follows: (1) the Dutch 7th Nitrate Directive Action Program, which intends to achieve the objectives formulated in the European Union (EU) Nitrate Directive; and (2) the EU Derogation Grant for 2023–2025, which includes a gradual phasing out of the exemption for the Netherlands to use more animal manure than 170 kg N ha−1. Both policy measures have been implemented because the current water quality is below the standard defined in the EU Water Framework Directive, despite efforts made in the past to achieve this objective. The combined measures specified in the 7th Action Nitrate Directive Program and the EU Derogation Grant have a major impact on farmers’ production management and subsequently profit. Changes to the current production management might be required to comply with legislation, especially for farms on sandy and loess soil. Various measures will be implemented gradually, with full implementation by 2027. For this scenario, we selected the full implementation outlined for 2027 because farmers will ultimately have to comply with these requirements. Table 3 presents the exact measures implemented in the water quality scenario.

Table 3 Measures resulting from the 7th Action Nitrate Directive Program (ND) and Derogation Grant 2023–2025 (DER) and their implications on farmers’ production management. 1Break crops are crops with an expected positive impact on soil quality and water quality. A list of break crops is available from RVO (2022). 2Catch crops are cover crops that take up residual N after cultivation of a main crop to prevent N leaching. A list of allowed catch crops is available from RVO (2023). 3Winter crops are crops that do not require planting a catch crop because they have a low amount of residual N. A list of winter crops is available from RVO (2023). Acronyms: N nitrogen, P phosphorous
Full size table

In the water quality scenario, we used crop rotation, cover crops, manure, fertilizer, and crop residue management as decision variables. The measures resulting from stricter water quality regulations as presented in Table 3 were implemented as additional constraints. A detailed description of the implementation of these constraints in the model can be found in Appendixes 3 and 5.

2.3.3 Field beans as part of protein transition

In this scenario, we explored field beans as an alternative crop to meet the increasing demand for high-quality plant-based protein sources (Augustin and Cole 2022; Ofoedu et al. 2022). Field beans offer advantages over other protein crops like soybeans, peas, lentils, and lupines in the Netherlands. Field beans demonstrate the highest yield potential in the regions’ cooler, humid climates (Timmer and Toren 2022). The harvest time of field beans in August or early September fits better compared to the late harvest of soybeans in October. Ongoing research indicates promising advancements in field bean cultivars, leading to higher yields and improved quality (Augustin and Cole 2022).

Despite current field bean cultivation in the Netherlands (Statistics Netherlands 2020), critical information such as input costs, crop yields, and output prices remains scarce. Field beans exist in two forms: winter field beans (planted in autumn) and spring field beans (planted in spring) (Jensen et al. 2010). Winter beans excel in yield potential but are vulnerable to winter foraging (Van Overveld, Limagrain Netherlands, personal communication, 25 May 2023).

We estimated costs, yields, and prices for field beans as follows. Direct costs, such as seeds and crop protection, were assumed consistent across farms (Roothaert, Limagrain Netherlands, personal communication, 25 May 2023). Field operation costs such as mechanization, energy, and labor were individually calculated for each farm. Yield data was derived from recent trials in the Netherlands (Prins et al. 2018, 2019; Prins and Timmer 2017; Timmer and Toren 2022), with average yields set at 5000 kg ha−1 for spring beans and 7000 kg ha−1 for winter beans. Given the absence of registered field bean prices, we estimated prices based on average feed prices from 2019 to 2022 (Wageningen Livestock Research 2023). Energy and protein content values of field beans were used to calculate an estimated price of 0.32 kg−1. (Limagrain Nederland 2023). Based on the average yield and average price, we calculated a standard revenue. We conducted sensitivity analyses varying from −25 to +50% of standard revenue to assess the viability of field beans in the cropping plan. Given that field beans are leguminous, they require no additional nitrogen fertilization and can supply approximately 75 kg N ha−1 to subsequent crops (Limagrain Nederland 2023).

3 Results

The results are structured as follows: “Farm scenarios” to “Field beans as part of protein transition” sections are structured as follows. First, we discuss the results of the four farm-level scenarios (Impacts of farm scenarios). Next, the results of stricter water regulations and of field beans as an alternative crop are discussed in the “Impact of stricter water quality regulations” and “Impact of stricter water quality regulations” sections, respectively.

3.1 Impacts of farm scenarios

Figure 3 shows the results of the soil quality indicators in the baseline scenario.

Fig. 3
figure 3

Results of soil quality indicators in the baseline scenario for nine case farms. “x” indicates a lower limit for the indicator. “-” indicates an upper limit for the indicator. Blanks imply that an indicator is not applicable on the soil type. For P, both a lower limit (P advice) and an upper limit (P legal norm) apply. The unit for the indicators M. chitwoodi, P. penetrans, R. solani, S. sclerotiorium, and V. dahlae is the potential loss in crop revenue (PRL) in € ha−1. Acronyms: N nitrogen, P phosphorous, K potassium, S sulfur, Mg magnesium, CRA crumbling ability, SV slaking vulnerability, WEV wind erosion vulnerability, SCI subsoil compaction index, OM organic matter

Full size image

Overall, the case farms demonstrate commendable performance, with only a few indicators showing limitations. Moreover, most of these limitations can be addressed through simple and practical production management decisions, such as increasing potassium fertilizer, sulfur fertilizer, or lime application. Figure 3 shows that P input is a limiting factor on multiple farms due to the upper bound P norm being lower than the lower bound P fertilization advise. This implies that farmers cannot apply P according to the agronomic recommendations, because it will exceed the legal norm. All farms exceed the target set by the subsoil compaction index, although farms with a higher share of cereals in the crop rotation (CN, CO, CSW1) come close to the target value. Compared to crops such as potatoes, sugar beets, and seed onions, cereals pose a lower risk on subsoil compaction due to the lower number of crop operations, the smaller harvest volume, and the earlier harvest time. Input of organic matter is a limiting factor primarily due to the required increase in SOM for improving cation exchange capacity on sandy soils (S, S-CR), or due to the reduced slaking vulnerability on loess and clay soil (L, CR, CSW1, CSW2).

Figure 4 shows the current costs and profits for the baseline scenario on all farms.

Fig. 4
figure 4

Profit or costs from current production management decisions (baseline scenario). All components above the x-axis represent profits, all components below represent costs. Farm acronyms: S sand, S-CR sand river clay, L loess, CO clay organic, CN clay north, CR clay river, CP clay polders, CSW 1 clay southwest 1, CSW-2 clay southwest 2

Full size image

Figure 4 shows considerable differences in profit between the farms in the baseline scenario. The profit of farm loess (L) is €1493 ha−1, whereas this is €79 ha−1 for farm clay polders (CP). The differences between these farms are mainly caused by crop profits and land costs. For example, land costs on clay polders are €1318 ha−1, while land costs on loess soil are only €557 ha−1. Additional variation in the baseline scenario is caused by differences in costs and profits of cover crops, manure, crop residues, and fertilizers. For example, on-farm clay rivers (CR) the total cost of all these production management decisions is €125 ha−1, compared to €612 ha−1 for farm clay organic (CO). Figure 4 illustrates that manure can either be a cost or a profit. Because of the high manure surplus in the Netherlands, slurry is often available at a premium. Therefore, farms that mainly apply slurry (S, S-CR, L, CR) show a profit for manure. Manure costs on farm clay organic (CO) are high because this farm is required to buy expensive organic manure.

Figure 5 shows the total profit for the baseline, soil quality tactical, soil quality strategic, and farm strategic scenarios. A detailed overview of all production management decisions in all scenarios can be found in Appendixes 2 and 4.

Fig. 5
figure 5

Profit resulting from optimization of production management with bio-economic model FARManalytics for different scenarios on arable farms in the Netherlands. Farm acronyms: S sand, S-CR sand river clay, L loess, CO clay organic, CN clay north, CR clay river, CP clay polders, CSW 1 clay southwest 1, CSW-2 clay southwest 2

Full size image

For almost all farms, we observe an increase in profit when production management is optimized, subject to all soil quality constraints. There are substantial differences in the potential of profit increases between farms, varying from € 47 ha−1 for a farm on river clay (CR) up to €704 ha−1 for an extensive farm on clay (CSW2). For most of the farms, switching to other crops in the cropping plan leads to an increase in profit (Fig. 6A). For farms on loess soil (L), the soil quality strategic scenario results in a decline in crop profit, which implies the crop rotation in the baseline scenario does not match the agronomic constraints in the soil quality strategic scenario. The clay farm in the North (CN) and the farm on river clay (CR) already have an optimal crop rotation, as evidenced by the fact that a further increase in profit in the soil quality strategic scenario compared soil quality tactical scenario is not possible.

Fig. 6
figure 6

A Change in costs or profit from production management decisions in the soil quality strategic (ss) scenario compared to the baseline scenario (b for nine arable farms in the Netherlands. B Change in costs or profit from production management decisions in the farm strategic (fs) scenario compared to the baseline scenario(b) for nine arable farms in the Netherlands. Farm acronyms: S sand, S-CR sand river clay, L loess, CO clay organic, CN clay north, CR clay river, CP clay polders, CSW 1 clay southwest 1, CSW-2 clay southwest 2

Full size image

Aside from cropping plan changes, increases in profit come from (1) reduction in fertilizer costs; (2) increase in profit from crop residues; and (3) lower manure costs. A decrease in fertilizer costs ranging from €19 to €182 ha−1 is possible on all farms, with the exception of farm CO, which does not use synthetic fertilizer. In the soil quality strategic scenario, crop residues (cereal straw) are sold on all farms that do not sell crop residues in the baseline scenario (S, S-CR, L, CO, CSW1, and CSW2). Profit decreases in the soil quality tactical scenario for farm S, because in order to meet the soil quality target, more cover crops have to be grown than in the baseline scenario. However, in the soil quality strategic scenario, profit increases compared to the baseline scenario: by optimizing the cropping plan with more silage corn instead of starch potatoes, soil quality targets can be met while increasing profit.

Figure 6A and B show how the different production management decisions contribute to a change in profit for the soil quality strategic and farm strategic scenarios compared to the baseline scenario.

Figure 6B shows larger changes in crop profit compared to Fig. 6A, which implies more crop rotation changes in the farm strategic scenario. The changes in costs and profit for crops, manure, fertilizer, and crop residue management in the farm-strategic scenario are similar to the changes in the soil quality strategic scenario. For the farms sand (S), sand-river-clay (S-CR), loess, clay polders (CP), and clay southwest 1 (CSW1) the farm strategic scenario resulted in a larger increase in profit than the soil quality strategic scenario. Exchanging land with dairy farmers, which was a challenge studied in the farm strategic scenario for farms on sandy soil and clay polders (S, S-CR, CSW1) resulted in a considerable increase in profit compared to the baseline scenario. Switching from low-yielding seed onions to better-yielding plant onions and celeriac on clay farm southwest 1 (CSW1 increased the profit with €250 ha−1 year−1 compared to the baseline scenario. Growing seed onions in the farm strategic scenario on the farm clay nord (CN) only resulted in a minor increase in profit compared to the baseline scenario. On farm CSW2, a higher share of sugar beets in the rotation, at the expense of the more profitable crops carrots and onions, resulted in lower total profit compared to the soil quality strategic scenario. The implementation of cut-and-carry fertilizer as an alternative for expensive organic solid manure on clay organic (CO) resulted in higher costs, decreasing profit compared to the soil quality strategic scenario. On the farm clay river (CR), the farm strategic scenario centered on potential profit without ware potatoes in the cropping plan due to labor constraints. This resulted in a decrease of profit compared to the baseline and soil quality strategic scenarios, as ware potatoes are profitable for this farm.

3.2 Impact of stricter water quality regulations

Figure 7 shows the total profit for the scenario where production management is optimized subject to stricter water quality regulations. In addition to the “baseline” results, this figure also presents the outcomes of the “soil quality strategic” scenario, which represents total profit in an optimized situation without stricter water quality regulations.

Fig. 7
figure 7

Total profit for the water quality scenario, a scenario in which production management is optimized subject to policy measures to improve water quality. For comparison, the total profit from the baseline situation and total profit for the soil quality strategic scenario in which production management is optimized without water quality restrictions is included. Farm acronyms: S sand, S-CR sand river clay, L loess, CO clay organic, CN clay north, CR clay river, CP clay polders, CSW 1 clay southwest 1, CSW-2 clay southwest 2

Full size image

From Fig. 7 follows that for all farms the implementation of water quality restrictions is feasible but results in a decline in profit compared to the soil quality strategic scenario, a scenario without water quality. However, by optimizing their production management, farms S, CR, and CSW2 can increase their total profit compared to the baseline scenario while meeting both soil quality targets and implementing water quality restrictions. Despite the drastic measures (such as compulsory break crops) for the farm on sandy soil (S), the farmer can increase profit due to land exchange with dairy farmers. On clay soil (CR, CSW2), the impact of water quality restrictions is substantially less compared to sand and loess. On farms S-CR and L, profit decreases compared to the baseline scenario. This implies that despite the optimization of production management, implementation of water quality restrictions goes at the expense of income. Both farms have to comply with all water quality measures (Table 3). Farm S-CR has to grow more break crops, which are not very profitable. On farm L, the main driver for the loss in profit is the non-cultivated buffer zones.

3.3 Impacts of field beans as part of protein transition

Figure 8 presents the total profit for the field bean scenario and the results of the sensitivity analysis. For comparison, also the results of the baseline and the soil quality strategic scenarios (the optimized scenario without field beans) are included. The production management decisions in the field bean scenario can be found in Appendixes 3 and 5.

Fig. 8
figure 8

Profit for scenarios that explore the potential of field beans as an alternative crop on arable farms in the Netherlands. In the field bean base, field beans are added with their crop profit based on their current revenue. fb−25%, fb+25%, and fb+50 are the results of a sensitivity analysis with −25%, +25%, and +50% of the revenue compared to the base. The total profit from the baseline scenario and the soil quality strategic scenario (crop rotation) optimized without field beans) is included for comparison. Farm acronyms: S sand, S-CR sand river clay, L loess, CO clay organic, CN clay north, CR clay river, CP clay polders, CSW 1 clay southwest 1, CSW-2 clay southwest 2

Full size image

Figure 8 follows that based on the current yield and current prices of field beans in the field beans baseline, field beans can be considered a serious alternative crop in the cropping plan of all considered farms since profit increases compared to an optimized scenario without field bean (soil quality strategic scenario). However, if the revenue of field beans decreases with 25%, total profit does not increase compared to the soil quality strategic scenario, which implies that based on that revenue field beans are not an interesting alternative. Based on the sensitivity analysis “fb+25%” and “fb+50%” the share of field beans in the rotation increases on farms S, S-CR, and CR. On farms S and CR, the crop frequency of field beans approaches the maximum of 0.2. On farms L and CSW2, the share of field beans remains equal compared to the field bean baseline.

4 Discussion

4.1 Impact of production management

4.1.1 Soil quality bottlenecks and economic performance in the baseline

The evaluated farms perform relatively well regarding their impact on soil quality: most of the soil quality indicators are above their target. Ros et al. (2022) report similar findings using the Open Soil Index framework across agricultural fields in the Netherlands. Current soil quality bottlenecks are organic matter input, subsoil compaction, and nutrients. Mandryk et al. (2014) found that improving soil organic matter content was one of farmers’ most important goals. Van den Akker and Hoogland (2011) confirm subsoil compaction to be a major bottleneck in soil quality in the Netherlands. Finally, Ros et al. (2022) found insufficient sulfur availability in 49% of the Dutch agricultural fields. The baseline shows considerable differences in economic performance between farms, which aligns with farm income monitoring (Wageningen Economic Research 2022). For instance, in 2021, the income gap between the top 20% and bottom 20% of farmers was €74,000.

4.1.2 Improving production management in the tactical dimension

In the soil quality tactical scenario, choices of cover crops, manure, fertilizer, and crop residue management were optimized within the farms’ current crop rotations. For seven out of nine farms, farm income increased substantially while achieving all soil quality targets except subsoil compaction. Preferred cover crops are A. strigosa, yellow mustard, and winter radish, which fit best in requirements regarding frost vulnerability, regrowth, rooting, and plant-parasitic nematode development. The most preferred manure types are cattle slurry and compost due to their economic value and composition. Cattle slurry, available either free or at a premium, offers high nutrient concentrations per unit of phosphorus (P), while compost supplies organic matter efficiently within P application limits. Solid manure is not preferred due to higher costs and lower nutrient availability. This is contradictory to farmers’ perceptions that solid manure is the best type of manure for the soil (Van Eekeren et al. 2009). A general conclusion from literature is that application of any type of manure is beneficial for soil quality compared to no manure being applied, due to the addition of carbon and base cations (Zavattaro et al. 2017). P and potassium (K) fertilizers were phased out. According to the advice of CBAV (2022), P and K recommendations can be fulfilled with the application of manure and compost alone. This is supported by the findings of de Vries et al. (2023) which demonstrate that manure and compost alone meet these nutrient needs nationally. Synthetic N fertilizer use decreased substantially compared to current practices, aligning with studies by Oenema et al. (2009), Silva et al. (2017), and Silva et al. (2021). A potential solution to more effectively apply N fertilizer might be by implementing best practices from peers (Lamkowsky et al. 2021). Based on Van Der Burgt et al. (2006), the FARManalytics model accounts for all possible sources of N, such as N supplying capacity from the soil, previous crops, and cover crops; this is not always done in practice due to the uncertainty surrounding the magnitude of these sources. Losses like volatilization, leaching, and denitrification during the seasons are partly ignored in FARManalytics; this might lead to an overestimation of plant available N by FARManalytics. On all farms that currently do not sell crop residues, crop residues can be used to generate additional income. Although crop residues have a widely acknowledged beneficial impact on soil quality (Klopp and Blanco-Canqui 2022; Turmel et al. 2015), alternative sources of nutrients and organic matter are widely available at low prices in the form of animal manure.

4.1.3 Improving production management in the strategic dimension

In the soil quality strategic scenario, changes in crop rotation were allowed. For three farms, profit did not increase compared to “soil quality tactical,” which implies that the current crop rotation is already optimal. However, for five out of nine farms, optimizing the crop rotation led to a substantial increase in profit by prioritizing the most profitable crops. Accurate crop profit calculation is crucial for this optimization (Mattetti 2022). For example, seed onions were excluded from the crop rotations on three farms because of their low profitability. In the farm strategic scenario, farm-specific challenges were studied. For three farms, the farm strategic scenario included land exchanges with dairy farmers. The land exchange resulted in even higher income increases compared to the quality strategic scenarios, indicating that land exchange could be a valuable approach to increase profit while achieving soil quality targets. Integrating arable and livestock production through joint land use approaches, such as integrated crop-livestock systems, shows potential benefits for farm income, soil quality, and environmental performance as recognized in the literature (Lemaire et al. 2014; Sekaran 2021).

4.1.4 Impact of stricter water quality regulations

In the water quality scenario, we optimized farmers’ production management with additional restrictions to preserve water quality. Implementation of these restrictions comes at a cost since profit is always higher in a scenario without water quality restrictions (soil quality strategic). However, when compared to the baseline scenario, three out of five farms could increase farm income while meeting soil quality targets and water quality restrictions by optimizing their production management. Belhouchette et al. (2011) studied the impact of implementation of the EU Nitrate Directive on French arable farming and found that farm income was not negatively affected, although substantial changes in production management were required. In contrast, Dellink et al. (2011) state that the economic performance of agriculture will suffer from more stringent water quality regulations.

4.1.5 Impact of field beans as alternative crop

Based on current yields and expected prices, field beans are an interesting alternative crop to replace part of cereals in the crop rotation, as highlighted by Jensen et al. (2010). However, current crop yield and the host status of field beans for soil-borne diseases are a point of concern. Field beans, being leguminous crops, can meet their own N needs and benefit subsequent crops (Ditzler et al. 2021; Jensen et al. 2010; Palmero 2022), and incorporating these benefits is crucial for profitable grain legume cultivation (Preissel et al. 2015). Sensitivity analysis shows that a substantially higher revenue (i.e., 50%) only has a limited impact on the optimal cropping plan. The share of field beans is close to the maximum (i.e., once every 5 years), and despite higher revenues, other crops such as potatoes and onions are more profitable.

4.2 Limitations and opportunities for on-farm decision support

In this section, we discuss the main challenges that need to be addressed before a model like FARManalytics can be applied for on-farm decision support.

4.2.1 Limited number of case farms

Case farms were selected from the Farmers Network for Soil Sampling. Farmers in this network already have a strong interest in sustainable soil management explaining the already good performance of soil quality indicators. Although the number of studied farms is limited, these case farms exhibit substantial diversity in soil type and farm setup, offering a representative overview of arable farms in the Netherlands. The wide variation in soil type and management styles underscores the need for customized solutions. Despite the network bias and small sample size, this study illustrates the added value of integrated bio-economic modeling of soil quality and farm economics. It suggests that even for front-runners, a substantial increase in farm income seems to be achievable while preserving soil quality. Future research should consider using larger samples of randomly selected farms to enhance generalizability. When extrapolating the findings of this study to arable farming in the Netherlands as a whole, we hypothesize that more intensive farms may face greater challenges in increasing farm income while meeting soil quality targets. Current production management of these intensive farms may push soil quality closer to the limit, leaving less room for improvement.

4.2.2 Spatial variability within farms

Currently, spatial allocation of production management is organized at farm level, assuming homogeneity in soil quality and field characteristics across the entire farm. We justify this decision based on balancing required input, model complexity, and result quality. However, two aspects have critical consequences. First, soil quality can vary substantially within a farm, meaning not every management decision is suitable for every field. For example, a field with low soil organic matter content might require more organic matter input and, hence, other management decisions than a field with high soil organic matter content. This is illustrated by Lessmann et al. (2022) and Moinet et al. (2023) who emphasize the importance of initial soil conditions in carbon sequestration potential. Second, we assume that available farmland can be flexibly allocated to different crops, whereas, in reality, field size and location are key drivers of crop allocation. For example, implementing a 12-year crop rotation on four 10-ha fields implies that every field has to be split into three parts, resulting in 12 fields of 3.33 ha. Castro et al. (2018) and Dury et al. (2012) recognize that many bio-economic farm models fail to sufficiently address spatial issues.

4.2.3 Validation

Output validation and end-user validation are key concepts in model validation (Bockstaller and Girardin 2003). Output validation ensures that model results are realistic and reliable by comparing them with measured data (Groot et al. 2012). End-user validation assesses the usefulness of model results for decision support. (Bockstaller and Girardin 2003). We discussed results in an iterative way with the farmers, which contributed to both output and end-user validation. Farmers perceived the results as useful and confirmed the added value of integral decision support on soil quality. The most prevalent reasons for not adopting model suggestions concerned risks (e.g., of soil compaction when applying manure in spring) and changing circumstances (e.g., higher prices for crops in recent years than averages used in this study). A more thorough output validation can be achieved by comparing model-predicted soil quality indicators with long-term trial data, such as those described by Korthals et al. (2014) and Schrama et al. (2018). Regarding end-user validation, the current bio-economic modeling approach allows for the economic evaluation of scenarios, which provides decision support for farmers. However, when optimizing their production management, farmers take more economic aspects into account than just average profit, risk, and uncertainty. Therefore, we recommended doing a structured end-user validation in which farmers reflect on solutions provided by the model and indicate their reasons for implementing these suggestions or not.

4.2.4 Resource availability and product market

In our farm-level optimization, factors like market prices, product demand, climate factors, and regional impact are treated as exogenous variables (Dogliotti et al. 2005). However, production management decisions cannot be considered outside of their socio-economic context (Castro and Lechthaler 2022). For example, the preference for cattle slurry and compost in most scenarios assumes their availability and stable pricing despite potential increases in demand. Similarly, in the field beans scenario, we assume a demand based on current yields and expected prices, but the actual existence of this demand remains uncertain.

4.3 Implications

This study shows that management advice derived from the bio-economic modeling of soil quality and farm economics can increase farm income while improving soil quality. We first discuss the implications of our results at the farm level. Second, we outline the implications for policy makers.

4.3.1 Implications at farm-level

Even for the nine farmers who are actively concerned with sustainable soil management, our study indicates substantial potential to increase profit while improving soil quality. These findings suggest that similar results could be achieved for other arable farms in the Netherlands. The results from this study show how FARManalytics can offer alternative production management decisions to adapt to evolving legislations (e.g., water quality) and changes in product demand (e.g., field beans). These developments require a change of current production management and FARManalytics can help to make the right decisions. Ultimately, this ensures the long-term preservation of soil quality in a financially robust strategy. To yield useful and reliable results, it is crucial to tailor the FARManalytics model to farm-specific conditions, especially when making crop rotation decisions. Given the considerable variation in crop profitability among farms growing the same crop, accurate cost and revenue data are essential inputs for the model. A farmer’s expertise remains essential since FARManalytics does not cover crucial aspects like production risk, robustness of choices, spatial allocation of activities, and variation in fields.

4.3.2 Implications for policy makers

Insights from this study offer valuable information for policy makers and illustrate the importance of policy impact analysis. Broader application of the methods used in this study on a larger and less biased sample of farms will provide more insights. This study goes beyond existing studies like those by de Haan et al. (2021) and Ros et al. (2022) by assessing current soil conditions and their relationship with production management. This allows us to predict potential soil quality issues under different combinations of soil conditions and management strategies. For broader implementation, we recommended a thorough validation of the soil quality quantification in FARManalytics. The FARManalytics model can be extended with environmental performance indicators to perform policy impact analyses on policies aiming at improving environmental quality. For example, in the water quality scenario in this study, policy impact analysis would be beneficial for assessing the effects of measures on indicators like nitrate leaching. FARManalytics is also able to calculate economic returns based on production management decisions. Calculating the expected profit change for farmers to reach environmental targets could be useful for policy makers to create financial incentives needed to encourage change in production management.

5 Conclusions

The objective of this study was to optimize production management to achieve maximum profit while increasing long-term soil quality on nine case farms in the Netherlands using the bio-economic model FARManalytics. Nutrient management, subsoil compaction vulnerability, and soil organic matter input are current bottlenecks to soil quality in the Netherlands. In a scenario studying the impact of stricter water quality regulation policies, we found that these policies limit production management decisions, thereby reducing potential farm income. In a scenario studying the potential of field beans (Vicia faba) as an alternative crop, we found that it can potentially replace some cereals and positively contribute to farm income and sustainable soil management. However, field beans cannot yet compete with cash crops like potatoes or onions. Based on these scenarios, we conclude that FARManalytics is able to offer alternative production management strategies to comply with evolving legislation or market shifts tailored at farm level. We show that the novel bio-economic model FARManalytics allows the preservation of soil quality in a financially robust strategy by selecting appropriate production management. Even for front-runners, optimization can increase profit up to €704 ha−1 while meeting soil quality targets, except for subsoil compaction. These findings underpin the added value of integral modeling of soil quality and farm economics with FARManalytics.