AGRICULTURAL AND FOOD SCIENCE Agricultural and Food Science (2024) 33: 90–115 90 https://doi.org/10.23986/afsci.143577 Gypsum and structure lime amendments in boreal agricultural clay soils: Do climate emissions compromise water quality benefits? Ollikainen Markku1, Lötjönen Sanna1, Tikkanen Tommi2, Ala-Harja Venla3, Uusitalo Risto4 and Ekholm Petri2 1Department of Economics and Management, Latokartanonkaari 7, 00014 University of Helsinki 2Finnish Environment Institute (SYKE), Latokartanonkaaari 11, 00790 Helsinki 3Helcom, Katajanokanlaituri 6 B, 00160 Helsinki 4Natural Resources Institute Finland (Luke), Latokartanonkaari 9, 00790 Helsinki e-mail: markku.ollikainen@helsinki.fi We examine cost-effectiveness and social net benefits of using soil amendments, gypsum and structure lime, in reducing phosphorus loading while accounting for the climate emissions from both amendments. Recent field experiments and large-scale pilots in Finland and Sweden suggest that both gypsum and structure lime improve soil structure and can reduce total P loading from clayey fields but differ as soil amendments. While gypsum does not change soil pH, structure lime helps to adjust it to a desired level. Drawing on literature, gypsum is postulated to reduce both dissolved (25%) and particulate losses (50%) of phosphorus, while structure lime is postulated to reduce only particulate phosphorus (40%). Life-cycle analysis is applied to determine greenhouse gas emissions from both soil amendments. We examine 5 and 10 years impacts on phosphorus loss by choosing doses and their timing accordingly. Both amendments provide the highest water quality benefits on erodible soils or soils with high soil phosphorus. Accounting for climate issues drastically changes the picture. Greenhouse gas emissions from gyp- sum production are 14.43 kgCO2e ha-1, and those from structure lime from pristine materials are 1837 kgCO2e ha-1. Cost-effectiveness of P load reduction including carbon price of GHG emissions is 59 € kg-1P for gypsum and 122 € kg-1P for structure lime. At the national level (application to 0.54 Mha), differences in greenhouse gas emissions without soil emissions are huge and in favour of gypsum (0.048 Mt and 1.04 Mt). Structure lime from recycled zero- emission materials performs well but its supply is very limited. Key words: phosphorus loss, water quality, soil amendment, greenhouse gas emissions, cost-effectiveness, JEL clas- sification: Q16, Q18, Q53 Introduction The Baltic Sea suffers from severe eutrophication due to nutrient loads from industrial and municipal point sourc- es, and especially from agricultural nonpoint sources. For science and policy targeting water pollution, the hard- est challenge is to reduce nonpoint loads from agricultural fields. Despite many efforts in both Europe and the US, success has been modest (Shortle et al. 2021). The primary challenge for agricultural water protection policy relates to agricultural areas that exhibit excessive soil phosphorus (P) status because it takes decades to “mine” this “legacy P”. In search of more efficient and rapid agricultural management practices, some countries are pro- moting soil amendments either using gypsum (Finland, US) or structure lime (Sweden) to reduce P loading (e.g., Ekholm et al. 2012, Ulén and Etana, 2014, Kost et al. 2018). Soil amendments can be considered an interim meas- ure that allows time for the time-consuming lowering of legacy P. Both gypsum and structure lime have been found to improve the structure of clay soils and promote the formation of larger soil aggregates, thus reducing P loading from fields and keeping P available for crop growth. Although gypsum (CaSO4 ∙ 2H2O) has been used for soil improvement for centuries (Shainberg et al. 1989), its role as a wa- ter protection measure was discovered only recently (Zhu and Alva 1994, Aura et al. 2006). Although lime is a traditional agronomic means to keep pH at a desired level, the use of structure lime (a mix of CaCO3 and CaO or Ca[OH]2) is of a more recent origin. Berglund (1971), Bell (1996), Blomqvist and Berglund (2021) and Al-Mukhtar et al. (2010), among others, established the stabilizing impacts of structure lime on clay soils and its ability to re- duce P loading. As a historical anecdote we note that the farmer magazine Käytännön Maamies (no 4/1953, p. 14) advertised enriched calcium carbonate as follows: the new soil amendment is delivered in paper sacks and it contains 58% reactive lime (CaO). Laboratory and field experiments as well as large-scale pilots have suggested that both gypsum and structure lime have a considerable ability to reduce total P (totP) loading from clayey fields (e.g., Berglund 1971, Blomquist and Received 25 February 2024 / Accepted 20 June 2024 The Scientific Agricultural Society of Finland ©This is an open access article under the CC BY 4.0 M. Ollikainen et al. 91 Berglund 2021, Ekholm et al., 2012, 2024). While the reductive impacts of both materials are well established, measurements concerning the effects somewhat differ across field experiments. Nevertheless, the results pro- vide strong grounds for society to promote the use of gypsum and structure lime for water protection purposes. A feature not focused on thus far is the carbon footprint of gypsum and structure lime amendments. Given the extensive efforts to mitigate climate change in Europe, the carbon footprints of alternative management practices in all sectors, agriculture included, are receiving increasing attention. The EU is applying the “do no significant harm principle” to all environmental policies. This principle emphasizes the need to also consider the greenhouse gas (GHG) emissions from soil amendments. As is well known, the production of lime is a great source of GHG emissions (IPCC 2006, West and Marland 2002, Stork et al. 2014) , and this feature is also present in manufactur- ing structure lime; the emissions depend on the share of recycled materials in the production process. Gypsum in turn is a side-product of phosphoric acid production, and its carbon footprint depends on the allocation of emis- sions between the two products. Gypsum is also available from other sources, but those are less used in Europe. The presence of the climate mitigation aspect raises the question of whether climate damage is so large that it outweighs the benefits from reduced P loading. Furthermore, if the use of soil amendments turns out to be desir- able, on what type of field do soil amendments give the highest benefits? These questions constitute the research problem of this paper. Drawing on the literature reporting the impacts of gypsum and structure lime amendments, we assess the economics of reducing P loading using soil amendments by adding climate emissions from both materials as a new variable in the analysis. Gypsum and structure lime differ as soil amendments. While gypsum does not affect soil pH, structure lime helps to increase it. Thus, while using the former is predominantly an invest- ment in water protection, the latter is at the same time an investment in crop production and water protection, and when used it replaces conventional agricultural lime. We account for this difference in our analysis and also note that the government provides subsidies to farmers for using structure lime in Sweden and for using gypsum in Finland. In both countries, the ground for both subsidies lies on the expected water quality benefits. Our research questions are: i) what is the cost-effectiveness of gypsum and structure lime in reducing phospho- rus loads, ii) how does accounting for climate effects of both amendments change their cost-effectiveness ratios, and iii) what are the overall net social benefits of both amendments? We build our answer to these question on a social cost‒benefit analysis, but with the special feature that we also employ the principles of life cycle anal- ysis (LCA) in determining GHG emissions (Curran 2013, Habert 2013, Boardman 2014). In this framework, soil amendment is treated as an investment in water protection, providing an annual reduction in P loading during the duration of the investment. We use the results on P loading from the literature, especially from Sweden and Finland. For GHG emissions from using soil amendments, we provide a genuinely new analysis. Furthermore, we employ sensitivity analysis to shed new light on the management issues needed to ensure the successful use and targeting of gypsum and structure lime. Impacts of gypsum and structure lime amendment on phosphorus loading – a literature review Literature on the impacts of gypsum and structure lime comes mostly from Finland, Sweden and the US. Research in Finland and the US has predominantly focused on the impacts of gypsum and in Sweden on structure lime. In this section, we will review and discuss the effects of gypsum and structure lime on P loading. Effects of gypsum on P loading Gypsum, or calcium sulfate dihydrate (CaSO4 · 2 H2O), is a mineral that is available in nature and is also produced as a byproduct of phosphoric acid manufacturing (Ekholm et al. 2012). Another source of gypsum, widely used in the US, is formed by flue-gas desulfurization in coal-fired plants (Kost et al. 2018). When spread on the fields, gypsum is dissolved into calcium and sulfate ions. The subsequent increased ionic strength of the soil solution promotes the micro-aggregation of the soil particles and thereby reduces particulate phosphorus (PP) losses (Uusitalo et al. 2012). In addition, it increases the binding of phosphate on soil particles and favors the formation of Ca-P asso- ciations (Muneer and Oades 1989, Uusitalo et al. 2012), which lower dissolved reactive phosphorus (DRP) loss- es. Table 1 collects the main peer-reviewed contributions to gypsum and reports the main findings by field- and catchment scale studies on P loading from clayey soils. Agricultural and Food Science (2024) 33: 90–115 92 In Finland, Aura et al. (2006) were the first to report the reduction in particulate and dissolved P by gypsum in clay suspensions in Finnish soil. Pietola (2008) extended lab experiments to cover different soil types. She employed a dose of 4 g kg-1, which corresponds to the Finnish “standard” rate of 4 t ha-1. Interestingly, gypsum also reduced P loading in fine sands and organic soils, which has not received much attention lately. Uusitalo et al. (2012) per- formed an experiment on clay soils in which gypsum was applied on two fields (on both at 6 tn ha-1 and on one of them also at 3 tn ha-1 rate) and undisturbed soil columns were taken to the laboratory for artificial rainfall; a typi- cal Finnish gentle rain intensity of 5 mm h-1 was applied on two consecutive days. The 31-month mean impact of applying gypsum was drastic (Table 1), as the particulate P concentration in the through-flow water decreased on average by 60% (range in the individual rainfall simulations conducted in 3 subsequent years was 25 to 92% reduc- tion) and the dissolved reactive phosphorus decreased on average by 43% (range from 16% increase to 91% de- crease relative to untreated control soil). In addition, the concentration of dissolved organic carbon was reduced by one third. The effects gradually declined over the 31-month study period as gypsum leached through the soil profile. The set-up of this experiment is considered to show the effect of gypsum on mobilization of P forms from soil to runoff on the field. Shifting the focus to the catchment level, Ekholm et al. (2012) reported an experiment covering 101 ha of arable lands, 91% of which were amended with gypsum (4 t ha-1). Gypsum amendment reduced the losses of PP by 64% and DRP by 29% as a 2.5-year average. A large catchment scale study (SAVE) in the Savijoki River treated 1 492 ha of agricultural land with 4 tons gypsum per ha (Ollikainen et al. 2020). The SAVE II project continued the meas- urement of P reduction in the same pilot area, and the findings suggested a wide range: a 5-year average reduc- tion of 34–70% for PP and –3–25% for DRP, the range reflecting the observations in two neighbouring measure- ment areas (Ekholm et al. 2024). John Nurminen Foundation arranged an even larger pilot of 3500 ha in the River Vantaa catchment in 2019–2021. The preliminary results for the first 1.5 years suggested that gypsum amend- ment reduced PP by 53% and DRP and dissolved organic carbon to some extent (Vantaanjoen kipsihanke 2020). Outside Finland, only few field- or catchment scale studies have been performed on clayey soils and with several years long monitoring period. Thus, a rather robust assumption is that gypsum reduces PP loading on average by 50%. As the results concern- ing DRP reduction are more uncertain, we will use 25% as the benchmark DRP reduction but also examine the case of zero impact in the sensitivity analysis. The Savijoki pilot is the first to examine a sufficiently long duration of impacts and suggests that a time period of at least 5 years, although lessening with time (Ekholm et al. 2022). Similarly, Aakriti et al. (2023) provide an LCA based review of environmental and economic impacts of various products made of flue-gas desulfurization gypsum and Endale et al. (2014) provide an analysis of gypsum with the application of broiler litter. Effects of structure lime on P loading The term structure lime refers to products that add a reactive component, either quicklime (CaO) or hydrated (slaked) lime (Ca[OH]2), into conventional agricultural lime (CaCO3) (Ulén and Etana 2014). Alternatively, it may refer to side products that are composed of quicklime/slaked lime and calcium carbonate. Structure lime is not, however, a homogenous product, and the share of reactive components typically varies between 15% and 40%. For example, the structure lime currently used in Sweden contains 15–20% hydrated lime, the remaining part be- ing CaCO3 (Blomquist 2021). In Finland, a lime containing about 19–25% reactive calcium was used by Kämäri et al. (2019) and Anttila et al. (2021) in two large pilots. Side-stream kilned lime from the forest industry, used in agriculture where available, typically contains higher shares of reactive lime (CaO/Ca[OH]2), up to over 40%, de- termined for Finnish side-stream structure lime products (unpublished data of Luke). Structure lime may thus be manufactured from pristine materials or from recycled materials, often as a combination of both. Table 1. Impact of gypsum on phosphorus loads on clay soils based on peer-reviewed long-term studies on a field- or catchment scale Source Type of experiment Dose t ha Reduction, % Duration of experiment, yearsPP DRP Uusitalo et al. (2012)a) field and lab 6 59–74 43–63 2.6 Uusitalo et al. (2012)b) field and lab 3 58 54 2.6 Ekholm et al. (2012) catchment 4 64 29 2.5 Ekholm et al. (2024)c) catchment 4 35–72 –3–25 5 DRP = dissolved reactive phosphorus, PP = particulate phosphorus; a) Reduction in concentrations, means over the 31-mo study (6 tn ha-1 dose was applied on two fields, 3 tn ha-1 on one field) b) Reduction in loads. c) The reported range indicates observed differences in two measurement areas in the SAVE pilot in the Savijoki River. M. Ollikainen et al. 93 When mixed with a clay soil structure, structure lime creates several reactions, including increases in the ion- ic strength of soil water, cation exchange, flocculation and agglomeration together with slower cementing and almost irreversible pozzolanic reactions (Kavak and Baykal 2012). While cation exchange may take place within a day, the completion of pozzolanic reactions may require a year at low temperatures (Al-Mukhtar et al. 2010). Be- cause pozzolanic reactions require very high pH, it is very uncertain if these occur at all in soil amendment use. The reactions listed result in a tighter binding of P and reduced erosion and P loading (Alakukku and Aura 2006). In general, the higher the clay content is, the higher the amount of the reactive component of the structure lime should be (Berglund 1977, Geranmayeh 2017). Structure lime effectively increases the soil pH level (e.g., Ene- si et al. 2023). According to a farmer guide (Ajosenpää et al. 2021), structure lime works best on fields where the aggregate structure is poor and electrical conductivity is low. Structure lime improves the workability of soil (Blomquist et al. 2023). Soil treated with structure lime is suggested to dry up more quickly in the spring and bet- ter maintains the aggregate structure, which improves the seedbed. The more even aggregate size distribution in the topsoil reduces evaporation, which is beneficial, especially during dry summers. Table 2 reports the findings from the Swedish and Finnish studies on the impacts of structure lime on clayey soils. Table 2 shows that Swedish studies performed at the field scale suggest that structure lime reduces PP losses by approximately 30–50%, while the effect on DRP varies from study to study. For example, Ulén and Etana (2014) found a 44% reduction in PP loss but no reduction in DRP loss in a field with 60% clay. However, in a field with 25% clay but a high soil P status, they found a reduction of 47% of DRP loss but no reduction in PP. Svanbäck et al. (2014) obtained a 42% reduction in PP loading in drainage flow relative to the benchmark, but no reduction in DRP. To study whether placing CaO on the subsoil could reduce P leaching through macropores, Andersson et al. (2016) found in a three-year lysimeter study that PP loading was reduced by 50% and DRP remained unaffected in two clayey subsoils, whereas DRP was reduced in a sandy soil that showed no effect on PP. Berglund et al. (2017) examined the effects of three different structure lime products in a two-year field study. A pure slaked lime product reduced PP loading by 45%, a mixture of slaked lime and agricultural lime by 50% and a recycled product by 35%. Norberg and Aronsson (2022) found that the common dose of 8 t ha-1 of structure lime (with about 20% of CaO) reduced total P loading in the second and third years after treatment by approximately 45% and 26%, respectively. A dose of 16 t ha-1 gave a higher retention for the third year (approximately 38%). Some Swedish studies have estimated the performance of structure lime on the basis of aggregate stability tests (e.g., Blomquist et al. 2018). For example, using a dose of 8 t ha-1, structure lime (with 15–20 % share of slaked lime (Ca[OH]2) has been found to increase aggregate stability for 1–2.5 years by 15–35% (Blomquist 2021). There is strong spatial variability in the performance, and the stabilizing effect likely depends on, e.g., clay content and mineralogy, soil organic matter content and initial pH (Blomquist and Berglund 2021). However, such improve- ments in tested water-stable aggregates are not directly translated to P loss mitigation, as aggregate stability does not necessarily correlate with actual load reduction (Norberg and Aronsson 2022). Table 2. Impact of structure lime on phosphorus loads on clay soils Source Type of experiment Dose t ha-1 (equivalent to CaO) Reduction, % Duration of experiment, yearsPP DRP Ulén and Etana (2014) field topsoil Exp. 1: 5 Exp. 2: 2 Exp. 1: 44 Exp. 2: 0 Exp. 1: 0 Exp. 2: 47 Exp. 1: 6 Exp. 2: 3 Svanbäck et al. (2014) field drainage flow 5 42 0 6 Andersson et al. (2016) lysimeter subsoil 5 50 0 3 Berglund et al. (2017) field and lab topsoil 3 35–50 0 2 Valkama and Mikkilä (2018) catchment runoff 3.5–8 seasonal variation 13–78 seasonal variation 0 4.5 Kämäri et al. (2019) catchment runoff 1.2 0 0 2 Norberg and Aronsson (2022) field drainage flow 1.6; 3.2 28a) 3 CaO = calcium oxide, i.e., the reactive component; DRP = dissolved reactive phosphorus; PP = particulate phosphorus; totP = total phosphorus; a) Mean reduction in totP over three years. Agricultural and Food Science (2024) 33: 90–115 94 Overall, experimental periods have been too short to provide a concise estimate of the duration of the impact of structure lime amendment and the application rate (Norberg and Aronsson 2022). The effects of structure lime may be long lasting mainly due to a pozzolanic reaction, which is potentially irreversible if such reactions occur. In her survey, Geranmayeh (2017) concluded 10 years as the estimate of the duration of the impacts, with the reduction for PP being 30%. The Finnish studies on structure lime amendments are of more recent origin and largely not yet published in scientific journals. The first results are from laboratory and field experiments. Alakukku and Aura (2006) found in an artificial rainfall study on undisturbed soil columns treated with soft lime kiln dust, a byproduct from the pro- duction of burnt lime, that lime amendment clearly decreased the concentrations of PP and DRP in surface runoff regardless of cultivation type (no-till, stubble cultivation, ploughing). The results from two catchment-scale stud- ies have been reported. Kämäri et al. (2019) applied structure lime (5 t ha-1 equivalent to 1.1 t CaO ha-1) to a 41 ha loamy clay soil. The concentrations of PP in runoff were already low before the treatment and were not affected by structure lime. The DRP concentrations may have decreased, but due to infrequent sampling, the finding was un- certain. In another catchment-scale study, Valkama and Luodeslampi (2020) found that structure lime (3.5–8 t ha-1) decreased PP losses most strongly one year after application. The seasonal reductions ranged from 13% to 78%, and after four years, there was no effect anymore. Even though the concentration of water-extractable P in soil increased up to threefold, possibly due to lime-induced elevation in soil pH, the losses of DRP did not increase. Finally, a large pilot on structure lime is underway, and the first results on PP are promising (Anttila et al. 2021). Drawing on the literature, we adopt the following estimates for the impacts. In the baseline case, we assume that structure lime from pristine material reduces PP loading by 40% and has no impact on DRP loading. To account for variation in measurements in Table 2, we resort to sensitivity analysis for lower (30%) and higher (50%) estimates of PP reduction. Despite small differences in the measured effect on P leaching for structure lime from pristine (40%) and recycled materials (35%), we assume that they exhibit the same effectiveness in reducing P loading. Finnish studies suggest 5 years as the duration of the P-reducing impacts (Valkama and Luodeslampi 2020), while some Swedish studies suggest that the effect might last as long as 10 years if larger amounts of structure lime are applied (e.g., Svanbäck et al. 2014, Ulén and Etana 2014, Geranmayeh 2017). In our main analysis, we postulate that structure lime has an impact over 5 years when 1.2 t ha-1 of reactive lime is used. As a hypothetical alterna- tive, we also examine a case where a 10-year-long duration of the impact can be obtained. In this case, we follow the dose reported in Ulén and Etana (2014) and Svanbäck et al. (2014) and assume that this effect is obtained when 5.5 t ha-1 of reactive lime is used. We acknowledge that such large quantities may not be practical, unless a soil is extremely acidic, but added this scenario as a hypothetical one. Most structure lime used in agriculture in Finland is manufactured from pristine sources, mainly kiln dust. Addi- tionally, a very limited amount of structure lime from recycled materials is available. Therefore, and unfortunately, structure lime from recycled materials does not provide a feasible option to reduce Finnish P loads at the large scale in the short run. Therefore, henceforth, we develop our calculations for structure lime from recycled mate- rials in Appendices 1 and 3 but comment on the results throughout the text. In the following analysis, structure lime refers to that originating from pristine material, unless otherwise stated. P runoff in the Archipelago Sea catchment area We apply the above reduction rates of gypsum and structure lime to a representative field parcel in southwest Finland. By assumption, the farm cultivates barley under typical agronomic conditions in a field with 15.9 mg l -1 as the soil test P value (extraction with pH 4.65 ammonium acetate; Vuorinen and Mäkitie 1955), which denotes the average soil P value in the catchment area of the Archipelago Sea. We employ the P runoff functions from Uusitalo and Jansson (2002) and Saarela et al. (1995) to determine P loads at the field edge with and without soil amendments. For both structure lime products, we assume that a 1.2 t dose of reactive lime (i.e., CaO/Ca[OH]2) per hectare has the measured P load reduction for 5 years. For the 10-year effectiveness, we use a 5.5 t/ha appli- cation of reactive lime drawing on Table 2. Assumptions concerning erosion and precipitation are based on data in the Archipelago Sea catchment area (values are given in Table A2, Appendix 2). Table 3 reports the baseline P runoff (no soil amendments) and the runoff under soil amendments at the field edge. M. Ollikainen et al. 95 Phosphorus runoff from the average field in the Archipelago Sea is rather high, clearly exceeding the Finnish av- erage of 1.1 kg ha-1 y-1 (Tattari et al. 2017). Gypsum provides an almost 50% reduction in totP loads, and struc- ture lime results in a reduction of slightly more than 30%. Differences in PP reduction between the amendments are rather small. The most important difference between soil amendments is that gypsum also reduces DRP, but structure lime does not. Although smaller, DRP is readily available for algal growth and thus contributes dispro- portionately to eutrophication. Determination of GHG emissions and costs from gypsum and structure lime amendments We assess GHG emissions in terms of carbon dioxide equivalent (CO2e) emissions. It is a metric measure that helps one to compare emissions from various GHGs on the basis of their global-warming potential (GWP), by convert- ing amounts of other gases to the equivalent amount of carbon dioxide (CO2) with the same global warming po- tential. Carbon dioxide-equivalent emissions and costs from both gypsum and structure lime amendments result from three sources: production, transport and spreading. Additionally, the application of agricultural lime within structure lime causes emissions from soil, which are, however, largely displayed since conventional agricultural lime is not applied for fields treated with structure lime (soil emissions are omitted here but included in Table 5 to calculate the net emissions from cultivation). We employ an LCA analysis to determine the GHG emissions from the production of gypsum and structure lime (for details, see Appendix 1). For gypsum we apply the economic allocation principle for GHG emissions from the production of phosphorus acid. For structure lime, we apply emissions reported in Nordkalk’s sustainability reports 2021 and 2022 for quicklime (Nordkalk 2021, 2022). Those emissions are very close to the average Eu- ropean assessments of GHG emissions. Nordkalk manufactures a small amount of structure lime from lime kiln dust (LKD). Economic allocation of emissions suggests that GHG emissions from this product are close to zero. This amount is, however, so small that it would not be realistic option for a nation-wide soil amendment of arable lands for water quality purposes. Given the variety of raw material sources for structure lime, we distinguish be- tween two extremes assuming that structure lime is manufactured either using pristine materials only or alterna- tively using solely recycled materials. Many available products lie between these extremes, and their impacts can be approximated using these two figures (see Fig. A1 in Appendix 1). Table 4 collects GHG emissions and costs related to the use of soil amendments when the aim is to achieve either 5 or 10 years’ long reductive impact on the P loss (detailed calculations of emissions from production and trans- port are reported in Appendix 1). For the 5-year impact, gypsum and structure lime are applied at the beginning of the first year. For the 10-year impact, gypsum is applied in the first and sixth years, while structure lime is ap- plied only in the first year but with an increased application rate (5 t reactive lime instead of 1.2 t). See Figure 1 for a schematic illustration. Table 3. P runoff at the field edge with and without gypsum and structure lime amendments in the catchment of the Archipelago Sea P runoff, kg ha-1 y-1 No amendment Gypsum Structure lime DRP 0.50 0.37 0.50 PP 1.31 0.66 0.79 totP 1.81 1.03 1.28 DRP = dissolved reactive phosphorus; PP = particulate phosphorus; totP = total phosphorus Table 4. GHG emissions without soil emissions and costs in terms of net present value (NPV) for gypsum and structure lime: production, transport spreading and harrowing (5 and 10 years). Gypsum, 5 years (1 application) Gypsum, 10 years (2 applications) Structure lime, 5-year duration Structure lime, 10-year duration Emissions, kgCO2e ha-1 Costs, € ha-1 Emissions, kgCO2e ha-1 Costs, € ha-1 Emissions, kgCO2e ha-1 Costs, € ha-1 Emissions, kgCO2e ha-1 Costs, € ha-1 Production/Price 14.4 72.6 28.9 135 1837 173 8420 791 Transport 54.2 129 108 239 9.79 23.2 44.9 106 Spreading 19.6 26.0 39.2 48.4 19.6 26.0 19.6 26.0 Harrowing - - - - 39.2 70.6 39.2 70.6 Total 88.3 227 177 423 1906 292 8524 994 Agricultural and Food Science (2024) 33: 90–115 96 Starting with GHG emissions in Table 4, differences between gypsum and structure lime are considerable. Struc- ture lime causes more than 21 times higher emissions for a 5-year impact, most of the emissions stemming from the production stage. The difference in emissions is even higher for a 10-year impact. For gypsum, the economic allocation of emissions from manufacturing phosphorus acid yields 3.61 kgCO2e t-1, giving 14.43 kgCO2e ha-1 when 4 tons are applied. Emissions from structure lime production (with 24.6% as the share of quicklime, CaO), are much higher, 367.4 kgCO2e t-1, giving 1837 kgCO2e ha-1 with 5 tons applied. Gypsum has slightly higher emissions from transport due to a longer transport distance. In cost calculations, we use 18.15 € t-1 as the price of gypsum (Ollikainen et al. 2020), and 34.50 € t-1 as the price of structure lime (KM 2022) (see also background data in Table A1 and A2 of Appendix 2). As structure lime substi- tutes for agricultural lime (see Section 3, Agronomic data), agricultural lime is not applied to fields receiving struc- ture lime. The cultivation costs are therefore reduced by the amount that would have been needed to purchase agricultural lime (9 € t-1 with application of 4.5 t ha-1), along with the spreading cost. When structure lime reduces P loads for 10 years, agricultural lime is applied on year 6 to keep the soil pH at the desired level, and these costs are included in the analysis (see also Fig. 1). For transport, we employ an identical unit cost of 0.071 € km-1; thus, costs differ because of distance, while spreading costs are 26 € ha-1 (Ollikainen et al. 2020). For the cost of disc harrowing of structure lime, we use 35.3 € ha-1, which was the average price for 2018 (Palva 2021) and multiply this by two. Thus, we conclude that the application costs are lower for gypsum, mostly because structure lime has a higher price than gypsum and it must to be harrowed twice to ensure the P loss reducing effect. We provide a similar analysis for structure lime from recycled materials (detailed calculations are given in Appen- dices 1 and 3). Recycled structure lime is assumed to reduce P loss by 40%, i.e., the effectiveness is the same as for structure lime from pristine materials. Provided that the manufacturer’s reported LCA data on GHG emissions are correct, structure lime from recycled materials causes low GHG emissions, 79 kgCO2e ha-1 (5-year duration of the impact) and 126 kgCO2e ha-1 (10-year duration with a higher dose). Its costs are the second lowest (238 € ha-1) for the 5-year impact on P loss but quite high (742 € ha-1) for the 10-year impact. Framework and data Net benefits and cost-effectiveness We consider gypsum and structure lime amendments as an investment in water protection (see Fig. 1 for the full outline of the investment and other decisions). Both investments are costly; thus, the environmental benefits should be higher than the investment costs. Let denote annual (short-term) profit from ordinary, baseline culti- vation (i.e., private profits), dpz(P) denote the eutrophication damage from total P runoff (with dp as the marginal damage) and dce indicate climate damage from GHG emissions from cultivation comprising emissions from fossil fuels and the soil (with dc as the marginal damage). Therefore, the annual social net benefits (B) from cultivation without investment in soil amendments are as follows: (1) Equation (1) gives the benchmark social net benefits from agricultural choices as a sum of private profits and en- vironmental damages relating to them in the absence of water protection investment. In its simplest form, pri- vate profit, , is given as = pf(F)-cF-M, where p is the crop price, f(F) denotes the yield response function, F re- fers to combined N, K and P fertilizer, c is the cost of fertilizer, and M is the fixed cost of cultivation (see Shortle et al. 2021). The time horizon of our analysis is 5 and 10 years, reflecting the above discussion on how long the soil amendments affect P loading. Thus, the net present value of baseline cultivation over the given years (i.e., NPV of social net benefits), that expresses the current value of future benefits and costs, provides our benchmark (t denotes years and r the discount rate): (2) An important feature of both soil amendment materials is that they do not take land out of cultivation and cause additional costs via reduced yields. Structure lime amendment may sometimes increase yields, whereas gypsum amendment is neutral in terms of yields. We denote soil amendment in general by X. Thus, X is either gypsum (G) or structure lime (S). Private profits from cultivation with soil amendment are denoted . Furthermore, we express P loads as a function of soil amendments zt(P(X)) and all GHG emissions related to soil amendment treat- ment by e1(X). Cultivation emissions not related to soil amendments, as in Eq. (1), are denoted with et. The invest- 𝐵𝐵𝐵𝐵0 = 𝜋𝜋𝜋𝜋� − 𝑑𝑑𝑑𝑑𝑃𝑃𝑃𝑃𝑧𝑧𝑧𝑧(𝑃𝑃𝑃𝑃) − 𝑑𝑑𝑑𝑑𝐶𝐶𝐶𝐶𝑒𝑒𝑒𝑒 𝜋𝜋𝜋𝜋� 𝜋𝜋𝜋𝜋� 𝐵𝐵𝐵𝐵𝐴𝐴𝐴𝐴 = ∑ 𝐵𝐵𝐵𝐵0(1 + 𝑟𝑟𝑟𝑟)−(𝑡𝑡𝑡𝑡−1)ℎ 𝑡𝑡𝑡𝑡=1 , h = 5,10 𝜋𝜋𝜋𝜋𝑡𝑡𝑡𝑡(𝑋𝑋𝑋𝑋) 𝜋𝜋𝜋𝜋� M. Ollikainen et al. 97 ment cost related to the use of soil amendments is denoted with . The social net benefits from soil amendment and cultivation are given as follows: (3) where X denotes either gypsum, G, or structure lime, S. We assume that structure lime may increase crop yields and thereby also private profits ( ), but this is not the case for gypsum ( ). We will quantitatively compare Equations (2) and (3) to see how well gypsum and structure lime perform as soil amendments relative to the baseline. Interpretation of this comparison is straightforward: if BG and BS exceed BA, the use of soil amendments is socially desirable. The above analysis also provides information to develop the required cost-effectiveness of the measures. They focus solely on investment costs and environmental benefits. A simple cost-effectiveness ratio is given in Equation (4a): , (4a) where we naturally account for the annual reduction in P loading during the five or ten years. When damage from additional GHG emissions relative to the baseline is considered, the cost-effectiveness ratio is transformed to the following: (4b) Equation (4b) indicates that damage from GHG emissions increases the costs of using soil amendments. Agronomic data and specification of the case study To determine cost-effectiveness and net benefits, we need to characterize revenue from crops and costs of culti- vation. Furthermore, we need to make assumptions on how soil amendments enter farmers’ management plans and impact crop yields. The use of gypsum does not change soil pH and, thus, can be treated as independent of farmers’ choices con- cerning crops and pH. YARA conducted a series of field experiments to examine how gypsum impacts crop yields (Ylivainio 2020). No yield penalties were found, but in fields with obvious sulfur deficits, gypsum increased crop growth. For certain crops (e.g., potato and sugar beet), calcium in pH-neutral gypsum can increase the yield qual- ity. Despite this, we conduct the calculation assuming no yield increase for gypsum. This is in line with a US meta- analysis, which did not show any negative or positive effect of gypsum on yield (Kost et al. 2018). Individual stud- ies have suggested that in some cases, the yield can be improved (e.g., Toma et al. 1999), but this seems not to be a general pattern. Analysis of structure lime requires some consideration. Depending on the crop, farmers aim at crop-specific pH levels. Farmers use agricultural lime to increase or maintain soil pH to desired levels. Liming improves the availa- bility of nutrients and in acidic soils it increases yields. For instance, Uusitalo et al. (2012) applied agricultural lime on two field plots differing in P status and tillage depth and from which soil columns were taken to the laboratory for leaching studies spanning 7, 19 and 31 months after lime amendment. They did not find any statistically sig- nificant difference in the concentration of PP of percolation water from soil columns treated with agricultural lime as compared with unlimed controls through rainfall simulations. However, the concentration of DRP was 10–14% higher in the lime-treated unit than in the controls. In Finnish agriculture, farmers make maintenance liming (2–5 t ha-1) roughly every five years. Given the rather high difference in prices of agricultural lime and structure lime, farmers most likely do not use structure lime for pH maintenance as a substitute for agricultural lime unless the soil structure is very labile. In the pH management plan, structure lime has the same yield effects as agricultural lime, and in addition, it provides additional benefits from improved soil structure. Studies on the effects of structure lime on crop yields have suggested various outcomes over the years and dif- ferent areas. The estimates vary from a 13% reduction to slight increases in crop yields (Svanbäck et al. 2014, Berglund et al. 2017, Blomquist et al. 2018, Ajosenpää et al. 2021, Norberg and Aronsson 2022). Blomqvist et al. 𝐵𝐵𝐵𝐵𝑋𝑋𝑋𝑋 = −𝐼𝐼𝐼𝐼𝑋𝑋𝑋𝑋 − 𝑑𝑑𝑑𝑑𝐶𝐶𝐶𝐶𝑒𝑒𝑒𝑒1(𝑋𝑋𝑋𝑋) + � [𝜋𝜋𝜋𝜋𝑡𝑡𝑡𝑡(𝑋𝑋𝑋𝑋) − 𝑑𝑑𝑑𝑑𝑃𝑃𝑃𝑃𝑧𝑧𝑧𝑧𝑡𝑡𝑡𝑡�𝑃𝑃𝑃𝑃(𝑋𝑋𝑋𝑋)� − 𝑑𝑑𝑑𝑑𝐶𝐶𝐶𝐶𝑒𝑒𝑒𝑒𝑡𝑡𝑡𝑡 ℎ 𝑡𝑡𝑡𝑡=1 ](1 + 𝑟𝑟𝑟𝑟)−(𝑡𝑡𝑡𝑡−1), 𝜋𝜋𝜋𝜋(𝑆𝑆𝑆𝑆) ≥ 𝜋𝜋𝜋𝜋� 𝜋𝜋𝜋𝜋(𝐺𝐺𝐺𝐺) = 𝜋𝜋𝜋𝜋� 𝑚𝑚𝑚𝑚 = 𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝑚𝑚𝑚𝑚𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 𝑃𝑃𝑃𝑃 − 𝑟𝑟𝑟𝑟𝐼𝐼𝐼𝐼𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑐𝑐𝑐𝑐𝐼𝐼𝐼𝐼𝑟𝑟𝑟𝑟𝑐𝑐𝑐𝑐𝐼𝐼𝐼𝐼 𝑚𝑚𝑚𝑚 = 𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝑚𝑚𝑚𝑚𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 + 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑚𝑚𝑚𝑚𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝐼𝐼𝐼𝐼 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑐𝑐𝑐𝑐𝑚𝑚𝑚𝑚 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝐼𝐼𝐼𝐼𝑚𝑚𝑚𝑚𝑒𝑒𝑒𝑒𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝑒𝑒𝑒𝑒𝑐𝑐𝑐𝑐𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑐𝑐𝑐𝑐𝑚𝑚𝑚𝑚 𝐼𝐼𝐼𝐼𝑐𝑐𝑐𝑐𝑒𝑒𝑒𝑒𝑠𝑠𝑠𝑠 𝑑𝑑𝑑𝑑𝑚𝑚𝑚𝑚𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝑑𝑑𝑑𝑑𝑚𝑚𝑚𝑚𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 𝐼𝐼𝐼𝐼𝑐𝑐𝑐𝑐𝐼𝐼𝐼𝐼𝑑𝑑𝑑𝑑𝑠𝑠𝑠𝑠 𝑃𝑃𝑃𝑃 − 𝑓𝑓𝑓𝑓𝐼𝐼𝐼𝐼𝑑𝑑𝑑𝑑𝑟𝑟𝑟𝑟𝑐𝑐𝑐𝑐𝐼𝐼𝐼𝐼𝑒𝑒𝑒𝑒𝑐𝑐𝑐𝑐𝐼𝐼𝐼𝐼 Agricultural and Food Science (2024) 33: 90–115 98 (2018) reported results from an experiment where the yield effect varied between –4 and +3% on average over four years. For our analysis, especially important is the finding by Blomquist et al. (2018) concerning barley yields; they concluded that crop yield responses were inconsistent, with changes in spring barley grain yield of ±10%. Decreased availability of micronutrients through binding in limed soil can possibly explain the observed yield de- creases. Yield increases were likely attributable to a finer tilth (Blomquist et al. 2018). Drawing on these ambiguous results, we assume that farmers use structure lime for water protection and main- tenance of soil pH. Therefore, in the baseline, we do not postulate any yield effect above what would have been achieved using ordinary agricultural lime. Nevertheless, we check in the sensitivity analysis for the case where structure lime is assumed to increase yields more than agricultural lime. We employ a crop yield function estimated for barley to determine the yields (see Appendix 1 Table A2). For emissions from cultivation, we use an estimate for emissions from soil, machinery and fertilization from Lötjönen and Ollikainen (2019). Figure 1 provides a schematic illustration of our approach. The farmer’s private choice provides the baseline for the comparison of soil amendments. In year 1, the farmer invests in soil amendments. In the private solution and under gypsum amendment, agricultural lime is used to change pH, followed by four years of ordinary cultivation, and in year 6, another investment is made, followed again by four years of cultivation. Structure lime amendment does not require agricultural lime in year one but does in year 6 in the case of a 10-year duration of the impact. Table 5 collects net revenue from cultivation and all employed values for climate and water quality impacts. Given that more than 90% of the Finnish farms participate in the national agri-environmental scheme, we assume that farmers apply the allowed levels of N and P fertilization (N = 100 kg ha-1 and P = 5 kg ha-1), which imply a barley yield of 4413 kg ha-1. The tight restriction on P fertilizer use reflects a field with good soil P status. Structure lime contains agricultural lime the application of which causes emissions from soil. For a 5-year duration, the share of agricultural lime is 3.8 t ha-1 with soil emissions of 1672 kgCO2e ha-1. For a 10-year duration, the share of agricul- tural lime is 17.4 t ha-1 with soil emissions of 7663 kgCO2e ha-1. These emissions are largely displayed since con- ventional agricultural lime is not applied for fields treated with structure lime. Table 5 contains soil emissions from both the direct use of agricultural lime and from its share in structure lime. Gypsum (5 y) Baseline (10y) Gypsum (10y) Structure lime (10y) St ru ct ur e lim e (5 y) G yp su m (5 y) Ba se lin e (5 y) Year 1 Agricultural lime Barley cultivation Year 1 Agricultural lime Gypsum (4 t/ha): manufacturing, transporting, purchasing, spreading Barley cultivation Year 1 Structure lime (5 t/ha with 1.2 t/ha reactive lime; 22.9 t/ha with 5.5 t/ha reactive lime): manufacturing, transporting, purchasing, spreading, harrowing x2 Years 2-5 Barley cultivation Years 2-5 Barley cultivation Years 2-5 Barley cultivation Years 7-10 Barley cultivation Years 7-10 Barley cultivation Years 7-10 Barley cultivation Year 6 Agricultural lime Barley lti ti Year 6 Agricultural lime Gypsum (4 t/ha): manufacturing, transporting, purchasing, spreading Year 6 Agricultural lime Barley cultivation Fig. 1. Schematic illustration of the 5-year and 10-year calculations M. Ollikainen et al. 99 Farmer revenue comes from selling crops. We also include the social valuation of agricultural landscapes by including a support component devoted to less favourable areas (LFAs) on the revenue side. Costs from soil amend- ments are expressed on a per hectare basis, and gypsum is approximately 65 € ha-1 cheaper than structure lime with 5-year effectiveness. The net revenue from cultivation with agricultural lime application is 281 € ha-1 and in other years 348 € ha-1. A further notable feature is that cultivation itself causes very high GHG emissions, 6018 kgCO2e ha-1 with agricultural lime and 3446 kgCO2e ha-1 without. Soil amendments increase emissions slight- ly (gypsum) or substantially (structure lime, including soil emission from agricultural lime within the product). Owing to a lack of data, we focus only on CO2 emissions from soil amendments, although they may also impact N2O emissions (Nadeem et al. 2020). We use these figures as data to solve for the cost-effectiveness properties of both soil amendments, where the costs are related to the obtained reduction in P loads. The costs comprise the annual present value of the direct Table 5. Annual GHG emissions, water quality impacts and net revenues over different years (see Appendices 1 and 2 for details and references) Baseline, no soil amendment, 5-year and 10-year durations Gypsum, 5-year and 10-year durations Structure lime, 5-year duration Structure lime, 10-year duration Costs, € ha-1 Fertilizer 174 174 174 174 Variable 101 101 101 101 Fixed 362 362 362 362 Agricultural lime (purchase and spreading) 67 67 - 67* Total costs, year 1, € ha-1 703 703 637 637 Total costs, year 6, € ha-1 703 703 - 703 Total costs, other years, € ha-1 637 637 637 637 Soil amendment, total, € ha-1 (see Table 4) 0 227 292 994 Revenue, € ha-1 Crop yield 768 768 768 768 LFA 217 217 217 217 Total revenue, € ha-1 985 985 985 985 Net revenue, year 1, € ha-1 281 54 56 -646 Net revenue, year 6, € ha-1-1 281 54 - 281 Net revenue, other years, € ha-1 348 348 348 348 GHG emissions, kgCO2e ha-1 Soil emissions 1855 1855 1855 1855 Management practices 489 489 489 489 N2O due to mineral N fertilization 671 671 671 671 Mineral N fertilizer manufacture 432 432 432 432 Agricultural lime (production and soil) 2572 2572 - 2572* Agricultural lime in structure lime (soil) - - 1672 7663 Soil amendment (see Table 4) - 117 1906 8524 Total GHG emissions, year 1, kgCO2e ha-1 6018 6106 7024 19633 Total GHG emissions, year 6, kgCO2e ha-1 6018 - - 6018 Total GHG emissions, other years, kgCO2e ha-1 3446 3446 3446 3446 Nutrient runoff, kg ha-1 N runoff 15 15 15 15 DRP runoff 0.50 0.37 0.50 0.50 PP runoff 1.31 0.66 0.79 0.79 totP runoff 1.81 1.03 1.28 1.28 *Only on year 6 Agricultural and Food Science (2024) 33: 90–115 100 cost of soil amendments or of the sum of the direct cost and climate damage from GHG emissions. We then em- ploy the above framework to assess the social net benefits of both soil amendments. Finally, note that both soil amendments also create important environmental co-benefits. Gypsum reduces the loss of particulate and dis- solved carbon loads, thus contributing to water quality, climate mitigation, and agricultural productivity. Under current limited knowledge, any comparison of climate processes taking place in soil or in water does not warrant a monetary analysis of these benefits. In a similar vein, by possibly increasing yield, structure lime may increase carbon sequestration in soil, but in this case, the size of this effect is hard to quantify. Cost-effectiveness and net benefits of soil amendments Cost-effectiveness in reducing P loading Table 6 condenses the cost-effectiveness analysis for P load reduction by considering only the treatment costs of soil amendments (denoted by I) and alternatively the sum of treatment and increased climate costs from green- house gas emissions (I + GHG), both defined in Equations (4a) and (4b), respectively. For this analysis, the treat- ment cost of structure lime is decreased by the purchase and spreading costs of agricultural lime (transport cost is negligible). We report ratios for 5- and 10-year impact durations; costs are given as present values. Table 6 shows a clear difference between the soil amendments in favour of gypsum. The difference in the cost- effectiveness figures (euros per reduced totP) between the soil amendments results from the differences in treat- ment costs and in the reduction of totP loads. This difference decreases if structure lime amendment increases yields beyond that of agricultural lime and this is accounted as a reduction in the investment cost. Accounting for GHG emissions as a negative externality increases the costs of both soil amendments and the wedge between structure lime and gypsum. Private and social net benefits Table 7 determines net private profits and ex post net social benefits (i.e., social net benefits resulting from the privately optimal management which omits water and climate externalities) from cultivation in the absence and presence of soil amendments as a net present value (NPV). To determine the social benefits in monetary terms, we need to value eutrophication and climate damages. As monetary estimates, we employ 35 € t-1 CO2e for cli- mate damage (based on Tol 2011) and, using nitrogen equivalents, Ne, to aggregate P and N, 11.6 € kg -1Ne for eutrophication damage (Gren 2001, value modified to year 2023 using a 2-% discount rate). Table 6. Cost-effectiveness (€ kg-1 P reduction) of gypsum and structure lime amendments based on annual treatment costs only (I) and annual treatment costs and increased climate costs (I+GHG) Reduction in totP load, kg ha-1 y-1 Cost-effectiveness (€ kg-1 P) I I + GHG Gypsum (5 y) 0.78 58.2 59.0 Gypsum (10 y) 0.78 54.2 55.0 Structure lime (5 y) 0.53 86.0 121.7 Structure lime (10 y) 0.53 176.6 318.4 Table 7. NPV of private profits and social net benefits over 5 or 10 years with a 3% discount rate (€ ha-1) Private profits (NPV) Social net benefits (NPV) P GHG P + GHG Baseline, 5 years (€ ha/5 years) 1575 862 916 203 Gypsum, treatment on year 1 (€/ha/5 years) 1347 942 685 280 Structure lime, duration 5 years, €/ha/5 years) 1349 843 596 90 Baseline, duration 10 years (€/ha/10 years) 2933 1606 1705 378 Gypsum, treatment on years 1 and 6 (€/ha/10 years) 2510 1755 1277 522 Structure lime, duration 10 years (€/ha/10 years) 2006 1064 34 –908 M. Ollikainen et al. 101 Notice first that in all cases, soil amendments reduce private profits compared to the baseline without soil amend- ments; thus, their use for water protection is not privately optimal unless they increase yields to offset the costs. The use of conventional agricultural lime (CaCO3) aims precisely at higher yields. It helps to adjust pH to the de- sired level and make nutrients more easily available for crops. Gypsum does not change pH but it contains a small share of P and lots of S, thus, it may increase yields only if the field is suffering from sulfur deficiency. Structure lime increases soil pH and may increase yields more than agricultural lime that it replaces. Unlike farmers, soci- ety also considers damage from P loads and GHG emissions. We account only for the P loads in the second col- umn and only for the GHG emissions in the third column. The second and fourth columns are the most relevant for our analysis. Cultivation with gypsum amendment provides a higher net present value of social net benefits than the baseline when only P runoff or both P runoff and GHG emissions are included. The social net benefits of structure lime always remain below the baseline. To see if structure lime would become socially acceptable, we need to resort to a sensitivity analysis. We restrict the sensitivity analysis only for the case of a 5-year duration of the impact, as structure lime has negative net benefits and thus performs very poorly for a 10-year duration. Regarding structure lime produced from recycled materials (reported in Appendices 1 and 3), it performs well for a 5-year duration, with the highest private profits after baseline, and the second highest social welfare when the P load or both the P load and GHG emissions are considered. For a 10-year duration, it performs weakly, but better than structure lime from pristine material. Regardless, structure lime from recycled materials provides an interesting alternative. Unfortunately, its availability is so restricted that it cannot be relied on for national plans. Sensitivity analysis: the impact of exogenous variables The above results depend on many assumptions related to environmental damage. With regard to water protec- tion, an assumption on the rate of erosion is crucial, as structure lime mostly reduces erosion, but reduced ero- sion also makes the most of gypsum’s impact on P loss. The soil P value especially affects the runoff of DRP but also PP (Uusitalo and Aura 2005). Also, we need to account for variation in the measured P reduction of both amendments, especially concerning DRP in the case of gypsum. Furthermore, valuation of social damage from eutrophication and climate damage plays a crucial role. Recall that in the above calculations, erosion was set at 1.1 t ha-1, and the soil P value was set at 15.9 mg l-1 in the P runoff functions. The climate damage estimate was 35 € t-1CO2e, an undervaluation in the presence of current policies, and runoff damage was set to 11.6 € kg-1Ne. Table 8 provides relevant results for lower and higher values of erosion and soil P and higher values for damage estimates. In the analyses, only one parameter is changed at a time, everything else being the same as in the ba- sic calculations above. Sensitivity analysis helps us characterize under what conditions soil amendments could play an especially import- ant role. For gypsum, a lower 34% reduction in PP and no reduction in DRP would make its social net benefit fall below the baseline, where no action is taken for water protection, while the higher reduction of PP by 50% makes it an extremely beneficial water protection measure. With a 2% yield increase above what agricultural lime would do, or a higher, 50% PP reduction, the net benefits from structure lime also become higher than the baseline. Concerning the state of arable soils (lower erosion or lower soil P status) and for the eutrophication damage es- timate used, the social net benefit of gypsum and structure lime decrease. However, with lower erosion gypsum still provides higher social net benefits than the baseline. For low erosion or low soil P status, the baseline P run- off is rather small, making soil amendments less effective (total P runoff without soil amendments: 1.54 kg ha-1 for low erosion and 1.15 kg ha-1 for low soil P value). For other cases, where erosion and soil P status are high, soil amendments play an important role. Both gypsum and structure lime amendments outperform the baseline, while the social net benefits from gypsum remain the highest. This suggests that soil amendments should be used in fields with high soil erosion and high soil P values (total P runoff without soil amendments: 2.14 kg ha-1 for high erosion and 2.37 kg ha-1 for high soil P value). Damage estimates complement the analysis. The higher the marginal damage of nutrient loading, the higher the benefits from soil amendments. Combining this finding with the above result, we conclude that the higher the P loads are, the larger the area that should be treated with soil amendments. Higher climate damage reduces ben- efits from gypsum and structure lime, as both entail a slight increase in GHG emissions over the baseline. Nev- ertheless, P loading and GHG emissions combined still suggest that gypsum performs better than structure lime and the baseline. As a final note, when soil P is approximately 8 mg l-1, that is, quite close to the identified social optimum in the presence of P loads (Iho 2010), the use of soil amendments does not increase social welfare. Agricultural and Food Science (2024) 33: 90–115 102 Finally, if the prices of fuels and fertilizers increase, they reduce the profitability of all alternatives. Accounting for P damage only, a higher fuel price resulting in a 10% increase in transportation costs decreases the NPV from gyp- sum by 13 € ha-1 (1.4%) and from structure lime by 2 € ha-1 (0.3%). Fuel prices affect gypsum the most, as it has the longest transportation distance, but gypsum still yields the highest NPV. A higher fertilizer price would reduce the NPV in all cases by the same amount since fertilization rate is equal. We provide sensitivity analysis of the cost-effectiveness ratios presented in Table 6 using the same parameter changes as above in Table 8. The resulting cost-effectiveness range (considering only the investment cost and in- vestment costs amended with changes in GHG emissions; see details in Table A6 in Appendix 4) for gypsum is 42.5–103.2 € kg-1P and for structure lime 68.8–177.8 € kg-1P. Thus, the range is quite large for both amendments. *) If structure lime increases crop yields by 2%, it provides higher social net benefits than the baseline for the marked cases. Table 8. Sensitivity results (€ ha-1), NPV of private profits and social net benefits over 5 years with a 3% discount rate (values exceeding baseline NPV are bolded) Private profit NPV Social net benefit NPV P GHG P + GHG Changed P runoff reduction and yield increase Baseline (totP load 1.81 kg ha-1) 1575 862 916 203 Gypsum, lower P reduction (PP 34%, DRP 0%) 1347 811 685 149 Gypsum, higher P reduction (PP 72%, DRP 25%) 1347 1056 685 394 Structure lime: 30% PP reduction 1349 791* 596 39 Structure lime: 50% PP reduction 1349 895 596 142 Structure lime, 2% yield increase 1421 915 669 163 Lower erosion (875 kg ha-1) Baseline (totP load 1.54 kg ha-1) 1575 968 916 309 Gypsum 1347 995 685 333 Structure lime 1349 906* 596 154 Higher erosion (1375 kg ha-1) Baseline (totP load 2.14 kg ha-1) 1575 733 916 74 Gypsum 1347 878 685 216 Structure lime 1349 765 596 13* Lower soil P status (8 mg l-1) Baseline (totP load 1.15 kg ha-1) 1575 1122 916 463 Gypsum 1347 1096 685 434 Structure lime 1349 1038 596 285 Higher soil P status (25 mg l-1) Baseline (totP load 2.37 kg ha-1) 1575 643 916 -16 Gypsum 1347 805 685 143 Structure lime 1349 667 596 -86* Higher climate damage (50 € t-1CO2e) Baseline (totP load 1.81 kg ha-1) 1575 862 633 -79 Gypsum 1347 942 402 -4 Structure lime 1349 843* 274 -232 Higher runoff damage (15 € kg-1Ne) Baseline (totP load 1.81 kg ha-1) 1575 653 916 -6 Gypsum 1347 823 685 161 Structure lime 1349 695 596 -58* M. Ollikainen et al. 103 Scaling up the use of soil amendments To assess the overall potential of soil amendments in reducing P loading under the “do no significant harm” prin- ciple, we scale gypsum and structure lime applications to a national level. Table 9 collects the P load reductions and GHG emissions (without soil emissions) in one scenario. The potential area of gypsum application in Finland is approximately 1.05 Mha (Ekholm et al. 2022), covering all cultivated fields on mineral soils and excluding catch- ment areas around lakes. The total potential area (Ekholm et al. 2022) comprises of the following areas in different catchments: 196 000 ha for the Gulf of Finland, 225 000 ha for the Archipelago Sea, 263 000 ha the Bothnian Sea and 370 000 ha for the Bothnian Bay (based on Huttunen, unpublished data). In the Savijoki gypsum pilot, 51.5% of fields in the area were amended with gypsum by farmers (Ollikainen et al. 2020). The figure reflects the fact that not all soils are suitable for gypsum, and not all farmers were willing to participate (Ollikainen et al. 2020). Thus, we employ 51.5% of the potential area in each catchment for the scaling of both soil amendments. The potential area for structure lime may be somewhat higher than that for gypsum because it includes fields upstream of lakes. On the other hand, most clay soils are in the coastal areas in Finland, which lack numerous lakes. For the Archipelago Sea catchment area, the baseline loading is 1.81 kg totP ha-1 according to the model and chosen parameters used in our calculations. For the other catchments (Gulf of Finland, Bothnian Sea and Bothnian Bay), we use a specific annual load of 1.1 kg totP ha-1, averaged from three studies looking at agricultural load carried by agricultural riv- ers and streams (Rankinen et al. 2016 [1.1 kg ha-1], Röman et al. 2018 [0.83 kg ha-1], Tattari et al. 2017 [1.26-1.58 kg ha-1]). The baseline annual total P loading is thus 680 tons (total areas based on Huttunen, unpublished data). The required amount of soil amendments is considerable if they are used as the only material over the entire land area. Small differences in P load reductions scale up to much larger differences in the achieved P loading reduction measured at the field edge. Gypsum reduces national P loads almost one-third more than structure lime. This results partly from a higher reduction in PP loading and its ability to reduce DRP loading. This finding is important for P policy. This suggests the use of gypsum and structure lime in combination, with structure lime targeting fields unsuitable for gypsum. For instance, amending 0.54 Mha soils by gypsum and complementing it on the no-go areas, approximately 0.45 Mha, would half the need for structure lime but would reach 0.99 Mha. Here, we have not accounted for the higher availability of DRP to algae (Iho et al. 2023), which further suggests the use of gypsum as a primary choice. The Archipelago Sea is Finland’s only hot spot in Helcom’s loading list, and in a recent report, Fleming et al. (2023) suggested 126 t as the maximum upper limit for its total P loads. As total P loading to the sea area is reported at 470 t y-1, the required reduction is 344 t y-1. Applying gypsum on all fea- sible fields in the catchment would result in 90 t y-1 reduction and structure lime to 61 t y-1 reduction, which are considerable shares of the required reduction. Looking at GHG emissions shows that gypsum causes only negligible climate impacts, whereas structure lime is a considerable GHG emission pulse. A comparison to emissions from transport illustrates the climate impacts. Light-duty vehicles produced 5.3 MtCO2e in 2021 in Finland (Statistics Finland 2023). Emissions from structure lime treatment on the 0.54 Mha (soil emissions omitting) are equal to approximately 16% of these emissions, while those of gypsum would be only 1%. Turning the differences from the production of soil amendments to emissions from driving, emissions from gypsum production per ha are equal to emissions of light-duty vehicles traveling only 87 km, while emissions from structure lime production per ha are as high as emissions from a 11020 km drive (calculations based on the Autokalkulaattori of the Finnish Climate Change Panel, https://www.ilmasto- paneeli.fi/autokalkulaattori/). This scale emphasizes that the only source of structure lime that is in line with the “do no significant harm” principle is structure lime from recycled materials. Therefore, a final note on the use of soil amendments relates to their availability. The supply of gypsum and structure lime from pristine materials is adequate at the national level. The availability of structure lime from recycled materials depends on the supply of pre-used material from the forest industry. That capacity is annually very limited, preventing the large-scale use of this structure lime variant. Table 9. Reduction in P loading (tons per year) and associated GHG emissions without soil emissions (million t) when gypsum or structure lime is applied to 543 000 ha Gypsum Structure lime, 5-year effectiveness Structure lime, 10-year effectiveness Quantity applied, Mt 2.2 2.7 12.4 DRP reduction, t year-1 34 0 0 PP reduction, t year-1 273 218 218 TP reduction, t year-1 306 218 218 GHG emissions, Mt CO2e year-1 0.048 1.04 4.63 Agricultural and Food Science (2024) 33: 90–115 104 Discussion Our study relies on the existing literature examining the reductive impact of soil amendments on P loads. There is still uncertainty concerning the impact and its duration for both amendments, gypsum and structure lime. In general, controlled lab experiments show the highest impacts, and shifting to field experiments and pilot areas increases the number of uncontrolled variables, leading to lower impacts. Most studies of gypsum demonstrate reduction of both PP and DRP loading but the duration of their reductive impact differ. Some studies on structure lime find reduction of both PP and DRP but many do not find any DRP reduction. We resolved these ambiguities in the literature by using reduction rates of 50% (PP) and 25% (DRP) for gypsum and 40% (PP) and 0% (DRP) for structure lime and by adopting 5 years impact as the base case. The impacts of uncertainty were checked by a sensitivity analysis. It led to a crucial finding that the desirability of soil amendments as a water protection invest- ment is lost if the reduction of P load remains below 30% for PP and is 0% for DRP. Most studies, however, have demonstrated higher impacts, providing support for the use of soil amendments. Our cost-effectiveness results ranged between 54–59 € kg-1P for gypsum and 86–177 € kg-1P for structure lime when GHG emissions are not included. These figures can be compared to other measures available for agriculture. Omitting gypsum, the reduction in DRP is possible only in a longer time span and under consistent and tight P fer- tilization restrictions that facilitate mining of excess soil P reserves. Buffer strips reduce PP loading in the short run (Uusi-Kämppä and Jauhiainen 2010). Expanding the size of buffer strips to achieve 40% and 50% PP reduction at the field edge would cost 116 € kg-1P (40%) and 161 € kg-1P (50%) via lost farm income (see Appendix 5 for details of the calculations). This is double relative to the costs of gypsum but stays within the range for structure lime. Another measure suggested for agriculture is to improve the drainage water management (DWM) (Frankenberger et al. 2023). DWM may provide a win-win solution, as good drainage improves crop growth and at the same time reduces especially N but also P losses from fields to waterways. The review of Frankenberger et al. (2023) sug- gested approximately 0.1 kgP ha-1 as the reduction in total P (0.04 kg ha-1 for DRP and 0.06 kg ha-1 for PP). These figures are low when compared to both soil amendments. Frankenberg et al. (2023) determine the cost-effective- ness ratio for N but not for P. Hjerppe and Väisänen (2015) provide a comprehensive analysis for multiple water protection measures under the Finnish circumstances (five different rivers). They provide a large range for costs. Cost-effectiveness of drainage water management ranges between 500–3000 € kg-1P, while cost-effectiveness of buffer strips is about 200 € kg-1P. Thus, our estimations for both soil amendments are lower than those. Sihvonen et al. (2020) apply a dynamic optimization framework with a simultaneous choice of N and P fertilization, the use of gypsum and including cover crops in the presence of legacy P. The socially optimal solution reduces soil P reserves from the privately choice (36 mg l-1) to the social optimum (8.6 mg l-1), thus, reducing especially DRP loads over a 120-year time horizon. While these results are not directly compatible with our results, it is interest- ing to note that gypsum plays a short run role in their model. Their work demonstrates well how important it is to have a consistent policy to reduce legacy P over time to complement the use of soil amendments. Finally, our calculations are the first one to provide estimates of GHG emissions in the context of soil amendments. We built our analysis on the work by Stork et al. (2014), Kämäri et al. (2019), Anttila et al. (2021) and IPCC (2006) for structure lime and on the standard LCA approach with economic allocation principle for gypsum (Belboom et al. 2015). One should note, however, that local production conditions differ. Therefore, GHG emissions from gyp- sum and structure lime differ depending on their origin. For instance, gypsum may come directly from mining or from flue gas desulfurization. Also, structure lime can be produced from multiple sources from pristine materials to several times recycled material. Furthermore, the carbon price plays important role for climate damages of both soil amendments. We employed a modest value 35 € t-1CO2e. Higher carbon price would considerably reduce the social desirability of using structure lime from pristine materials. For structure lime the share of recycled materials turned out to be crucial. Our work facilitates a linear description for emissions from structure lime, assuming almost zero emissions to manufacture structure lime from recycled materials (see Figure A1 in Appendix 1). This helps one to adjust GHG estimates to better reflect cases in which recycling of raw materials increases in the production of structure lime. M. Ollikainen et al. 105 Conclusions We examined the role and targeting of two soil amendments, gypsum and structure lime, in reducing P loads from arable fields. Recent field experiments and large-scale pilots in Finland and Sweden suggest that both gypsum and structure lime can reduce total P loading from clayey fields. While Sweden has promoted the use of structure lime, Finland has done the same with gypsum. In the spirit of the EU’s green transition, we followed the “do no significant harm” principle and included GHG emissions in the analysis. Thus, we asked if the use of soil amendments is so- cially desirable when benefits from reduction of P loads and climate damages from GHG emissions are considered. We condense our key findings on water protection as follows. First, the cost-effectiveness ratios for gypsum and recycled structure lime are in general lower than those of other measures reducing P runoff from agriculture. Including GHG emissions, if structure lime is manufactured using pristine material, its cost-effectiveness ratio is high compared to gypsum. Second, the social net benefits (accounting only damage from P load) from reduced P loading by gypsum and by structure lime with at least a 2% increase in crop yields exceed the current agricultur- al baseline, showing that society should promote their use. Again, including GHG emissions clearly reduces the social net benefits for structure lime from pristine materials. Third, sensitivity analysis emphasizes targeting the use of both soil amendments on erodible soils or high soil P values. For sites with low erosion or low soil P values, benefits from soil amendments remain below the baseline. Accounting for climate issues does not substantially change the above results. GHG emissions from soil amendments increase the cost-effectiveness ratio and lower social net benefits. This impact is minor for gypsum and recycled structure lime but considerable for structure lime from pristine materials. If the prices of fuels and fertilizers increase, they reduce the profitability of all alter- natives, but the case of higher gasoline prices in transport reduces the profitability of gypsum the most, as it has the longest transportation distance. Drawing on the analysis, we make the following conclusions. Both soil amendments provide high potential to reduce P loads to waterways especially on fields that have high soil P values or high erodibility. Their use is not, however, profitable for farmers without government subsidies under current agri-environmental policies. Thus, active promotion of amendments is economically justified. Accounting for climate impacts of both amendments reduces both cost-effectiveness and net social benefits of both amendments. Nevertheless, gypsum and struc- ture lime from recycled materials perform well even when GHG emissions are taken into account, thus accounting for GHG emissions and climate damage does not hinder their use. In contrast, structure lime from pristine materials has a high carbon footprint. There are no environmental grounds for promoting the use of structure lime from pristine materials on a large scale for water protection unless its GHG emissions are radically reduced. Thus, for the case of structure lime from pristine materials, climate emissions do compromise water quality benefits. The literature forming the basis of our estimates for the reductive impact of soil amendments on P loads covers only a limited combination of agri-environmental mosaic found in boreal lands. With increasing knowledge on the effect of soil amendments on PP and DRP, it is possible to get more valid assessments of their potential role in attaining lower agricultural P loads and improved ecological state of coastal waters, considering the climate impact. An equally important question is how to create incentives to strengthen environmentally friendlier agri- cultural practices. Acknowledgements This work is a joint product of two projects: SAVE II, funded by the Ministry of the Environment, and Shared Wa- ters, funded by the Finnish Cultural Foundation. The authors gratefully acknowledge the funding. Comments by two anonymous reviewers are gratefully acknowledged. References Aakriti, Maiti, S., Jain, N. & Malik, J. 2023. A comprehensive review of flue gas desulphurized gypsum: Production, properties, and applications. Construction and Building Materials 393: 131918. https://doi.org/10.1016/j.conbuildmat.2023.131918 Ajosenpää, T., Anttila, L., Ekholm, P., Heikkinen, J., Jaakkola, S., Kaseva, A., Kämäri, M., Kääriä, J., Luodeslampi, P., Malmilehto, S., Muurinen, S., Rasa, K., Soinne, H., Talola, S., Uusi-Kämppä, J. & Uusitalo, R. 2021. Kipsi, kuitu ja rakennekalkki-Opas viljelijöille. https://jukuri.luke.fi/handle/10024/551094 Alakukku, L. & Aura, E. 2006. Zero Tillage and Surface Layer Liming Promising Technique to Reduce Clay Soil Erosion and Phos- phorus Loading. ASABE 2006 Annual International Meeting, July 9-12, 2006, Portland, Oregon. Paper number 062191. https://doi.org/10.13031/2013.20749 Agricultural and Food Science (2024) 33: 90–115 106 Al-Mukhtar, M., Lasledj, A. & Alcover, J.-F. 2010. Behaviour and mineralogy changes in lime-treated expansive soil at 20 °C. Ap- plied Clay Science 50: 191–198. https://doi.org/10.1016/j.clay.2010.07.023 Andersson, H., Bergström, L., Djodjic, F., Ulén, B. & Kirchmann, H. 2016. Lime placement on subsoil as a strategy to reduce phos- phorus leaching from agricultural soils. Soil Use and Management 32: 381–389. https://doi.org/10.1111/sum.12290 Anttila, L., Kämäri, M., Ekholm, P. & Mikkilä, E. 2021. Rakennekalkkikäsittelyn vaikutukset valumavesissä - lupaavia havaintoja Eurajoen pilottialueilta. Vesitalous 4: 8–12. (in Finnish). Aura, E., Saarela, K. & Räty, M. 2006. Savimaiden eroosio. MTT:n selvityksiä 118: 32. (in Finnish). Belboom, S., Szöcs, C. & Léonard, A. 2015. Environmental impacts of phosphoric acid production using di-hemihydrate process: A Belgian case study. Journal of Cleaner Production 108: 978–986. https://doi.org/10.1016/j.jclepro.2015.06.141 Bell, F.G. 1996. Lime stabilization of clay minerals and soils. Engineering Geology 42: 223–237. https://doi.org/10.1016/0013-7952(96)00028-2 Berglund, G. 1971. Kalkens inverkan på jordens struktur. Stenciltryck 46: 14 p. https://pub.epsilon.slu.se/5797/1/Berglund_G_110617pdf (in Swedish). Berglund, G. 1977. Mikroaggregatanalysen som testmetod vid strukturkalkning. Stenciltryck 102: 120 p. https://pub.epsilon.slu.se/5616/1/berglund_g_110223.pdf (in Swedish). Berglund, K., Etana, A. & Simonsson, M. 2017. Ekokalk: Strukturkalkning för förbättrad markstruktur och minskade fosforförluster i ekologisk odling? https://fou.jordbruksverket.se/fou/dl/Fil-004417 (in Swedish). Blomquist, J. 2021. Effects of structure liming on clay soil. Acta Universitatis Agriculturae Sueciae 86: 170 p. Blomquist, J. & Berglund, K. 2021. Timing and conditions modify the effect of structure liming on clay soil. Agricultural and Food Science 30: 96–107. https://doi.org/10.23986/afsci.103422 Blomquist, J., Englund, J.-E., Sjöberg, C., Kårhammer, J., Svensson, S.-E., Pettersson, E., Keller, T. & Berglund, K. 2023. Structure lim- ing reduces draught requirement on clay soil. Soil and Tillage Research 231: 105703. https://doi.org/10.1016/j.still.2023.105703 Blomquist, J., Simonsson, M., Etana, A. & Berglund, K. 2018. Structure liming enhances aggregate stability and gives varying crop responses on clayey soils. Acta Agriculturae Scandinavica, Section B - Soil & Plant Science 68: 311–322. https://doi.org/10.1080/09064710.2017.1400096 Boardmann, A. E., Greenbergm D.H., Vining, A.R. & Weimen, D.L. 2014. Cost-benefit analysis: Concepts and practice. Fourth Edi- tion. Pearson Education, Inc., Upper Saddle River, New Jersey, 07458. Campo, F.P., Tua, C., Biganzoli, L., Pantini, S. & Grosso, M. 2021. Natural and enhanced carbonation of lime in its different applica- tions: A review. Environmental Technology Reviews 10: 224–237. https://doi.org/10.1080/21622515.2021.1982023 Curran, M.A. 2013. Life Cycle Assessment: A review of the methodology and its application to sustainability. Current Opinion in Chemical Engineering 2: 273–277. https://doi.org/10.1016/j.coche.2013.02.002 Ekholm, P., Ollikainen, M., Ala-Harja, V., Begum, K., Huttunen, M., Järvenranta, K., Kiirikki, M., Kuosa, H., Lötjönen, S., Riihimäki, J., Taskinen, A., Tikkanen, T. & Yli-Halla, M. 2022. Peltojen kipsikäsittely fosforikuormituksen hallinnassa - pilottina Savijoen valuma- alue. Suomen ympäristökeskusen raportteja 32. http://urn.fi/URN:ISBN:978-952-11-5506-2 (in Finnish). Ekholm, P., Ollikainen, M., Punttila, E., Ala-Harja, V., Riihimäki, J., Kiirikki, M., Taskinen, A. & Begum, K. 2024. Gypsum amendment of agricultural fields to decrease phosphorus losses - Evidence on a catchment scale. Journal of Environmental Management 357: 120706. https://doi.org/10.1016/j.jenvman.2024.120706 Ekholm, P., Valkama, P., Jaakkola, E., Kiirikki, M., Lahti, K. & Pietola, L. 2012. Gypsum amendment of soils reduces phosphorus losses in an agricultural catchment. Agricultural and Food Science 21: 279–291. https://doi.org/10.23986/afsci.6831 Elonen, P., Erjala, M., Hakkola, H., Heinonen, R., Heinänen, E., Hiivola, S.-L., Jaakkola, A., Jokinen, R., Kivistö, J., Korkman, J., Kuisma, P., Köylijärvi, J., Melén, A., Mälkönen, E., Puustinen, M., Rajala, J., Saarela, I., Sallasmaa, S., Tanska, T. & Österman, R. 1991. Kalki- tusopas. Maaseutukeskusten Liiton julkaisuja Tieto tuottamaan: 78555. https://jukuri.luke.fi/handle/10024/471491 (in Finnish). Endale, D.M., Schomberg, H.H., Fisher, D.S., Franklin, D.H. & Jenkins, M.B. 2014. Flue Gas Desulfurization Gypsum: Implication for Runoff and Nutrient Losses Associated with Broiler Litter Use on Pastures on Ultisols. Journal of Environmental Quality 43: 281–289. https://doi.org/10.2134/jeq2012.0259 Enesi, R.O., Dyck, M., Chang, S., Thilakarathna, M.S., Fan, X., Strelkov, S. & Gorim, L.Y. 2023. Liming remediates soil acidity and improves crop yield and profitability-A meta-analysis. Frontiers in Agronomy 5. https://doi.org/10.3389/fagro.2023.1194896 Ervola, A., Ollikainen, M. & Mikkola, J. 2012. Agriculture and climate change: The socially optimal production, land use, and GHG emissions. Food Economics 9: 10–24. https://doi.org/10.1080/16507541.2012.695120 Eurofins 2017. Viljavuustutkimuksen tulkinta. Eurofins. https://cdnmedia.eurofins.com/european-east/media/1818630/viljavuus- tutkimuksentulkinta2017teroprint.pdf (in Finnish). Finnish Food Authority 2020. Ympäristökorvauksen sitoumusehdot 2020. Ruokavirasto. https://www.ruokavirasto.fi/tuet/ maatalous/peltotuet/ymparistokorvaus/ymparistokorvauksen-sitoumusehdot/ymparistokorvauksen-sitoumusehdot-2020/ (in Finnish). Fleming, V., Berninger, K., Aikola, T., Huttunen, M., Iho, A., Kuosa, H., Niskanen, L., Piiparinen, J., Räike, A., Salo, M., Sarkkola, S. & Valve, H. 2023. Rannikkovesien ravinteiden kuormituskatot ja kuormituksen vähentämisen keinoja: Loppuraportti Valtioneuv- oston selvitys- ja tutkimustoiminnan julkaisusarja 2023:45. https://julkaisut.valtioneuvosto.fi/handle/10024/165047 (in Finnish). Frankenberger, J., McMillan, S. K., Williams, M. R., Mazer, K., Ross, J. & Sohngen, B. 2023. Drainage water management: A review of nutrient load reductions and cost effectiveness. Journal of the ASABE 0(0), 0. https://doi.org/10.13031/ja.15549 Geranmayeh, P. 2017. Strukturkalkning i stor skala-Vad krävs och vad kostar det? BalticSea2020. https://balticsea2020.org/im- ages/Bilagor/Strukturkalkning-i-stor-skala_Geranmayeh-2017.pdf (in Swedish). M. Ollikainen et al. 107 Gren, I.-M. 2001. International Versus National Actions Against Nitrogen Pollution of the Baltic Sea. Environmental and Resource Economics 20: 41–59. https://doi.org/10.1023/A:1017512113454 Habert, G. 2013. A method for allocation according to the economic behaviour in the EU-ETS for by-products used in cement in- dustry. The International Journal of Life Cycle Assessment 18: 113–126. https://doi.org/10.1007/s11367-012-0464-1 Hjerppe, T. & Väisänen, S. 2015. A practical tool for selecting cost-effective combinations of phosphorus loading mitigation meas- ures in Finnish catchments. International Journal of River Basin Management. https://www.tandfonline.com/doi/abs/10.1080/ 15715124.2015.1012516 Iho, A. 2010. Spatially optimal steady-state phosphorus policies in crop production. European Review of Agricultural Economics 37: 187–208. https://doi.org/10.1093/erae/jbq009 Iho, A., Valve, H., Ekholm, P., Uusitalo, R., Lehtoranta, J., Soinne, H. & Salminen, J. 2023. Efficient protection of the Baltic Sea needs a revision of phosphorus metric. Ambio 52: 1389–1399. https://doi.org/10.1007/s13280-023-01851-2 Intratec 2022. Commodity Price Database. https://www.intratec.us/products/primary-commodity-prices IPCC 2006. 2006 IPCC Guidelines for National Greenhouse Gas Inventories, Prepared by the National Greenhouse Gas Inventories Programme. (Eggleston, H., Buendia, L. Miwa, L., Ngara, T.& Tanabe, K. eds.). IGES. https://www.ipcc-nggip.iges.or.jp/public/2006gl/ Kämäri, M., Ekholm, P., Röman, E., Ahonen, E., Seppälä, M., Markula, R., Kiirikki, M. & Urkko, J. 2019. Rakennekalkituksen vaiku- tus Pakkalanjärven fosforikuormitukseen - #RAKAVA. Loppuraportti. 13.12.2019 (p. 58). https://www.syke.fi/fi-FI/Tutkimus__ke- hittaminen/Tutkimus_ja_kehittamishankkeet/Hankkeet/Rakennekalkin_vaikutus_Pakkalanjarven_fosforikuormitukseen_RAKAVA (in Finnish). Kavak, A. & Baykal, G. 2012. Long-term behavior of lime-stabilized kaolinite clay. Environmental Earth Sciences 66: 1943–1955. https://doi.org/10.1007/s12665-011-1419-8 Keskinen, R., Ketoja, E., Heikkinen, J., Salo, T., Uusitalo, R. & Nuutinen, V. 2016. 35-year trends of acidity and soluble nutrients in cultivated soils of Finland. Geoderma Regional 7: 376–387. https://doi.org/10.1016/j.geodrs.2016.11.005 Kiirikki, M., Rantanen, P., Varjopuro, R., Leppänen, A., Hiltunen, M., Pitkänen, H., Ekholm, P., Moukhametshina, E., Inkala, A., Ku- osa, H. & Sarkkula, J. 2003. Cost effective water protection in the Gulf of Finland. Focus on St. Petersburg. The Finnish Environ- ment 632: 1–55. KM 2022. Kalkkitaulukko. Käytännön Maamies 8. (in Finnish). Kost, D., Nester, J. & Dick, W.A. 2018. Gypsum as a soil amendment to enhance water quality by reducing soluble phosphorus concentrations. Journal of Soil and Water Conservation 73: 22A–24A. https://doi.org/10.2489/jswc.73.1.22A Lankoski, J. & Ollikainen, M. 2006. Suojakaistat ja maatalouden ympäristöpolitiikka. In: Virkajärvi, P. & Uusi-Kämppä, J. (eds.) Lai- tumen ja suojavyöhykkeiden ravinnekierto ja ympäristökuormitus. Maa- ja elintarviketalous 76: 187–204. https://jukuri.luke.fi/ handle/10024/461862 (in Finnish). Lankoski, J., Ollikainen, M. & Uusitalo, P. 2006. No-till technology: Benefits to farmers and the environment? Theoretical analysis and application to Finnish agriculture. European Review of Agricultural Economics 33: 193–221. https://doi.org/10.1093/erae/jbl003 Lehtonen, H. 2001. Principles, structure and application of dynamic regional sector model of Finnish agriculture. Helsinki Univer- sity of Technology. https://urn.fi/urn:nbn:fi:tkk-002993 Lötjönen, S. & Ollikainen, M. 2019. Multiple-pollutant cost-efficiency: Coherent water and climate policy for agriculture. Ambio 48: 1304–1313. https://doi.org/10.1007/s13280-019-01257-z Luke 2019. Directory of statistics. Natural Resources Institute Finland. https://www.luke.fi/en/statistics/directory-of-statistics Mikkola, H.J. & Ahokas, J. 2009. Energy ratios in Finnish agricultural production. Agricultural and Food Science 18: 332–346. https://doi.org/10.23986/afsci.5958 Muneer, M. & Oades, J.M. 1989. The role of Ca-organic interactions in soil aggregate stability .II. Field studies with 14C-labeled straw, CaCO3 and CaSO4.2.H2O. Soil Research 27: 401–409. https://doi.org/10.1071/sr9890401 Mustonen, E. 2022. Sadonkorjuun jälkeen on vuorossa viljavuustalkoot. Käytännön Maamies 6. (in Finnish). Nadeem, S., Bakken, L.R., Frostegård, Å., Gaby, J.C. & Dörsch, P. 2020. Contingent Effects of Liming on N2O-Emissions Driven by Autotrophic Nitrification. Frontiers in Environmental Science 8. https://doi.org/10.3389/fenvs.2020.598513 Norberg, L. & Aronsson, H. 2022. Mitigating phosphorus leaching from a clay loam through structure liming. Acta Agriculturae Scandinavica Section B - Soil & Plant Science 72: 987–996. https://doi.org/10.1080/09064710.2022.2138528 Nordkalk 2021. Sustainability Report 2021. https://www.nordkalk.com/wp-content/uploads/2022/05/REPORT-NORDKALK-sus- tainability-report_2021.pdf Nordkalk 2022. Sustainability Report 2022. https://nordkalk.com/wp-content/uploads/sustainability-report/NORDKALK_sustain- ability_report_2022_FINAL_interactive.pdf Ollikainen, M., Kosenius, A.-K., Punttila, E., Ala-Harja, V., Puroila, S., Iho, A. & Ekholm, P. 2020. Gypsum amendment of arable fields as a water protection measure - farmers’ experience, phosphorus reduction potential and associated costs drawn from a large scale pilot. Agricultural and Food Science 29: 383–394. https://doi.org/10.23986/afsci.88902 Palva, R. 2021. Konetyön kustannukset ja tilastolliset urakointihinnat (2018) (457; TTS:n julkaisuja, p. 16). TTS Työtehoseura. htt- ps://tts.fi/wp-content/uploads/2023/10/Konetyon_kustannukset_ja_tilastolliset_urakointihinnat.pdf (in Finnish). Pietola, L. 2008. Gypsum-based management practices to prevent phosphorus transportation. NJF Report, Nordic Associa- tion of Agricultural Scientists, NJF Seminar 401, Uppsala Sweden 22.-23.9.2008 NJF Report 4: 78–82.https://orgprints.org/id/ eprint/16046/1/fosfor1.pdf Agricultural and Food Science (2024) 33: 90–115 108 ProAgria 2019. Tuottopehtori. https://www.webwisu.fi/tuottopehtori/ (in Finnish). Puustinen, M., Turtola, E., Kukkonen, M., Koskiaho, J., Linjama, J., Niinioja, R. & Tattari, S. 2010. VIHMA-A tool for allocation of measures to control erosion and nutrient loading from Finnish agricultural catchments. Agriculture, Ecosystems & Environment 138: 306–317. https://doi.org/10.1016/j.agee.2010.06.003 Rankinen, K., Keinänen, H. & Cano Bernal, J.E. 2016. Influence of climate and land use changes on nutrient fluxes from Finn- ish rivers to the Baltic Sea. Agriculture, Ecosystems & Environment 216: 100–115. https://doi.org/10.1016/j.agee.2015.09.010 Röman, E., Ekholm, P., Tattari, S., Koskiaho, J. & Kotamäki, N. 2018. Catchment characteristics predicting nitrogen and phospho- rus losses in Finland. River Research and Applications 34: 397–405. https://doi.org/10.1002/rra.3264 Saarela, I., Järvi, A., Hakkola, H. & Rinne, K. 1995. Fosforilannoituksen porraskokeet 1977-1994. Vuosittain annetun fosforimäärän vaikutus maan viljavuuteen ja peltokasvien satoon monivuotisissa kenttäkokeissa. Maatalouden tutkimuskeskus tiedote 16. (in Finnish). Sallasmaa, S. 1991. Peltojen happamuustilanne ja kalkin käyttö. Kalkitusopas. Tieto Tuottamaan 55: 4–9. (in Finnish). Shainberg, I., Sumner, M.E., Miller, W.P., Farina, M.P.W., Pavan, M.A. & Fey, M.V. 1989. Use of Gypsum on Soils: A Review. In: Stew- art. B.A. (ed.). Advances in Soil Science 9: 1–111. https://doi.org/10.1007/978-1-4612-3532-3_1 Shortle, J., Ollikainen, M. & Iho, A. 2021. Economics and Policy for Water Pollution Control. In: Shortle, J., Ollikainen, M. & Iho, A. (eds.). Water Quality and Agriculture: Economics and Policy for Nonpoint Source Water Pollution. Springer International Publish- ing. p. 17–74. https://doi.org/10.1007/978-3-030-47087-6_2 Sihvonen, M., Lintunen, J., Valkama, E. & Hyytiäinen, K. 2020. Management of legacy nutrient stores through nitrogen and phospho- rus fertilization, catch crops, and gypsum treatment. Natural Resource Modeling 33: e12289. https://doi.org/10.1111/nrm.12289 Simmelsgaard, S. 1991. Estimation of nitrogen leakage functions-Nitrogen leakage as a function of nitrogen applications for dif- ferent crops on sand and clay soils. Nitrogen Fertilizers in Danish Agriculture-Present and Future Application and Leaching, Insti- tute of Agricultural Economics Report 62. Simojoki, A., Jaakkola, A. & Alakukku, L. 1991. Effect of compaction on soil air in a pot experiment and in the field. Soil and Tillage Research 19: 175–186. https://doi.org/10.1016/0167-1987(91)90085-C Statistics Finland 2023. Greenhouse gas emissions in Finland 1990 to 2021. National Inventory Report under the UNFCCC. Submis- sion to the European Union. 15 march 2023. https://stat.fi/media/uploads/tup/khkinv/fi_nir_eu_2021_2023-03-15.pdf Stork, M., Meindertsma, W., Overgaag, M. & Neelis, M. 2014. A Competitive and Efficient Lime Industry-Cornerstone for a Sustain- able Europe. The European Lime Association. https://www.eula.eu/wp-content/uploads/2019/02/A-Competitive-and-Efficient- Lime-Industry-Technical-report-by-Ecofys_0.pdf Svanbäck, A., Ulén, B. & Etana, A. 2014. Mitigation of phosphorus leaching losses via subsurface drains from a cracking marine clay soil. Agriculture, Ecosystems & Environment 184: 124–134. https://doi.org/10.1016/j.agee.2013.11.017 Tattari, S., Koskiaho, J., Kosunen, M., Lepistö, A., Linjama, J. & Puustinen, M. 2017. Nutrient loads from agricultural and forested areas in Finland from 1981 up to 2010-Can the efficiency of undertaken water protection measures seen? Environmental Moni- toring and Assessment 189: 95. https://doi.org/10.1007/s10661-017-5791-z Tol, R.S.J. 2011. The Social Cost of Carbon. Annual Review of Resource Economics 3: 419–443. https://doi.org/10.1146/annurev- resource-083110-120028 Toma, M., Sumner, M. E., Weeks, G. & Saigusa, M. 1999. Long-term Effects of Gypsum on Crop Yield and Subsoil Chemical Prop- erties. Soil Science Society of America Journal 63: 891–895. https://doi.org/10.2136/sssaj1999.634891x Ulén, B. & Etana, A. 2014. Phosphorus leaching from clay soils can be counteracted by structure liming. Acta Agriculturae Scan- dinavica, Section B - Soil & Plant Science 64: 425–433. https://doi.org/10.1080/09064710.2014.920043 Uusi-Kämppä, J. & Jauhiainen, L. 2010. Long-term monitoring of buffer zone efficiency under different cultivation techniques in boreal conditions. Agriculture, Ecosystems & Environment 137: 75–85. https://doi.org/10.1016/j.agee.2010.01.002 Uusitalo, R. & Aura, E. 2005. A rainfall simulation study on the relationships between soil test P versus dissolved and potentially bioavailable particulate phosphorus forms in runoff. Agricultural and Food Science 14: 335–345. https://doi.org/10.2137/145960605775897713 Uusitalo, R. & Jansson, H. 2002. Dissolved reactive phosphorus in runoff assessed by soil extraction with an acetate buffer. Agri- cultural and Food Science 11: 343–353. https://doi.org/10.23986/afsci.5734 Uusitalo, R., Ylivainio, K., Hyväluoma, J., Rasa, K., Kaseva, J., Nylund, P., Pietola, L. & Turtola, E. 2012. The effects of gypsum on the transfer of phosphorus and other nutrients through clay soil monoliths. Agricultural and Food Science 21: 260–278. https://doi.org/10.23986/afsci.4855 Valkama, P. & Luodeslampi, P. 2020. Rakennekalkki ja ravinnekuitu - vaikutukset maatalouden vesiensuojelutoimina. RAKUVA- hankkeen loppuraportti (21/2020). Vantaanjoen ja Helsingin seudun vesiensuojeluyhdistys ry. https://www.vhvsy.fi/files/upload_ pdf/9453/Raportti%2021_2020%20Rakennekalkki%20ja%20ravinnekuitu%20loppuraportti.pdf (in Finnish). Valkama, P. & Mikkilä, E. 2018. Veden laadun sekä maan lämpötilan ja kosteuden seuranta. Rakenne-kalkituksen vaikutukset. Lohkon ominaispiirteet huomioiva ravinnekuormitusmallinnus ja sen kehittäminen (LOHKO-hanke) -hankkeen jatkohanke (LOHKO II). YM 104/481/2016 (1.1.2017-31.12.2018). https://www.mtk.fi/documents/20143/0/Veden+laadun%2C+maan+l%C3%A4mp%C3%B6ti lan+ja+kosteuden+seuranta+sek%C3%A4+rakennekalkitus_final.pdf/4a0eaa26-8d4b-64f2-74a9-a48c104d1d61?t=1563175451997 (in Finnish). van Breemen, N., Mulder, J. & Driscoll, C. T. 1983. Acidification and alkalinization of soils. Plant and Soil 75: 283–308. https://doi.org/10.1007/BF02369968 Vantaanjoen kipsihanke 2020. Vantaanjoen kipsihanke. Loppuraportti. https://johnnurmisensaatio.fi/wp-content/uploads/2023/09/ vantaanjoen-kipsihanke-loppuraportti_15.12.20.pdf (in Finnish). M. Ollikainen et al. 109 VTT 2020. October 6. LIPASTO yksikköpäästötietokanta. http://lipasto.vtt.fi/ (in Finnish). Vuorinen, J. & Mäkitie, O. 1955. The method of soil testing in use in Finland. Maatalouskoelaitoksen maatutkimusosasto. Agrogeologisia julkaisuja 63. 44 p. West, T.O. & Marland, G. 2002. A synthesis of carbon sequestration, carbon emissions, and net carbon flux in agriculture: Com- paring tillage practices in the United States. Agriculture, Ecosystems & Environment 91: 217–232. https://doi.org/10.1016/S0167-8809(01)00233-X Yara 2024. Yara Siilinjärvi. Yara. https://www.yara.fi/tietoa-yarasta/Yara-Suomi/toimipaikat/siilinjarvi/tuotantolaitos/ (in Finnish). Ylivainio, K. 2020. Unpublished data. Zhang, H., Liu, R. & Lal, R. 2016. Optimal sequestration of carbon dioxide and phosphorus in soils by gypsum amendment. Envi- ronmental Chemistry Letters 14: 443–448. https://doi.org/10.1007/s10311-016-0564-4 Zhu, B. & Alva, A.K. 1994. The effect of Gypsum amendment on transport of phosphorus in a sandy soil. Water, Air, & Soil Pollu- tion 78: 375–382. https://doi.org/10.1007/BF00483044 Appendix 1. GHG emissions from the production and use of gypsum and structure lime Emissions from gypsum production GHG emissions for gypsum production are determined using the economic allocation principle (Habert 2013). The economic allocation coefficient, ω, is defined as: , where p1m1refers to the mass value of the pri- mary product (market unit price times the produced/sold mass) and p2m2 to that of the side product. The coeffi- cient indicates the share of the sales of the side product in total economic benefits. For gypsum, GHG emissions relate to the production of phosphoric acid. Belboom et al. (2015) estimated that the production of fertilizer grade phosphoric acid (H3PO4) generates approximately 626 kgCO2e t-1 when the wet di-hemihydrate process is used. YARA calculations for the production in Siilinjärvi suggest a value slightly less than 600 kg CO2e t-1 (pers. communication, Soile Ylisuutari 2021). Therefore, 600 kgCO2e t-1 is used in the calcula- tions. The price of gypsum is 18.15 € t-1 and its annual sales are 80 000 t (Yara 2024, pers. communication, Guni- lla Stedt, 2024). The annual phosphorus acid production is 300 000 t and its price, taken from international data (Intratec 2022), is 800 € t-1. Thus, the coefficient is . Using this value for emissions yields 0.0361 × 600 = 3.61 kgCO2e t-1 gypsum produced. For the 4 t ha-1 application rate, the emissions from production of 14.43 kgCO2e ha-1. Emissions from structure lime production For structure lime, GHG emissions are calculated directly from the production of reactive and agricultural lime and their shares in structure lime (Anttila et al. 2021). For structure lime from pristine materials, we target a 1200 kg dose of reactive lime (CaO/Ca[OH]2) per hectare and use 24.6% as the share of CaO in the structure lime product (Kämäri et al. 2019). It requires that 5 t of struc- ture lime is spread on a hectare (Anttila et al. 2021). The average emission factor for quicklime production in Eu- rope is approximately 1092 kgCO2e t-1 (Nordkalk 2022, Stork et al. 2014). For agricultural lime (CaCO3), we use an emission factor of 131 kgCO2e t-1 (West and Marland 2002). Its share is 75.4% of the product. Using the relative shares in structure lime gives: 0.246 × 1092 + 0.754 × 131 = 367.4 kgCO2e t-1. Using the 5 t ha-1 application yields 1837 kgCO2e ha-1, while the 22.9 t ha-1 application yields 8420 kgCO2e ha-1. For structure lime from recycled materials, we also target a 1200 kg dose of reactive lime per hectare for a 5-year duration, but now the share of CaO in the structure lime product is 33.3%. Thus, the spread amount of struc- ture lime is 3.6 t for recycled materials. For a 10-year duration, the amount of reactive lime is 5.5 t ha-1 reflecting structure lime application of 16.5 t ha-1. Emissions estimate for recycled structure lime production is based on the manufacturer’s LCA showing that emissions are 3.65 kgCO2e t-1 (pers. communication Ossi Kinnunen 2023; aver- age moisture content of 17.75% from Mustonen [2022]). For an application rate of 3.6 t ha-1, emissions from pro- duction are 13.13 kgCO2e ha-1, which is slightly below the estimate for gypsum. Figure A1 presents a linear relationship between the production emissions of two extremes for cases studied here: structure lime produced only from pristine material or only from recycled materials. Given that structure lime products sold in the markets may contain different shares of pristine and recycled materials, this figure helps to assess the realized production emissions as a function of the share of recycled materials. 𝜔𝜔𝜔𝜔 = 𝑝𝑝𝑝𝑝2𝑚𝑚𝑚𝑚2 𝑝𝑝𝑝𝑝1𝑚𝑚𝑚𝑚1 + 𝑝𝑝𝑝𝑝2𝑚𝑚𝑚𝑚2 𝜔𝜔𝜔𝜔 = 18.15 ∗ 80 000 800 ∗ 300 000 + 18.15 ∗ 80 000 = 0.0361 Agricultural and Food Science (2024) 33: 90–115 110 Emissions from transport and spreading Emissions from transport depend on the distance and the size of trucks. For gypsum, we use the distance from the YARA Siilinjärvi factory to fields in southwest Finland, on average 450 km (Ollikainen et al. 2020). For structure lime, the distance is 65 kilometres, as it is produced at multiple sites in southern Finland. Emission factor for full trailers with a carrying capacity of 40 tons is 1.205 kgCO2e km-1 (VTT 2020). This leads to 54.23 kgCO2e ha-1 for gyp- sum, 9.79 kgCO2e ha-1 for structure lime with 5 years impact and 44.9 kgCO2e ha-1 for structure lime with 10 years impact. Emissions increase linearly with distance; for every 50 kilometres, emissions increase by 1.51 kgCO2e t-1. Spreading of soil amendments by manure or lime spreading equipment causes emissions from fossil fuels. For gypsum, only the spreading is needed. After spreading structure lime, the field must be harrowed twice within 48 hours to mix it fully with soil. Both spreading and harrowing consume 7.2 l ha-1 diesel (Mikkola and Ahokas 2009), with emissions from fuel 2.723 kgCO2e l-1 (VTT 2020). These figures yield 19.61 kgCO2e ha-1 for the spreading of soil amendments and 39.21 kgCO2e ha-1 for harrowing structure lime twice. Soil emissions A final note concerns how structure lime works when spread in soils. Upon reaction with CO2, CaO or Ca(OH)2 un- dergoes a partial (re)carbonation process (Campo et al. 2021). For soil stabilizing treatment at a construction site, Campo et al. (2021) reported carbonation rates of 37% and 80% in natural and enhanced (laboratory) CO2 atmos- pheres, respectively. In agricultural topsoils, where active microbial decomposition is maintained by continuous inputs of root exudates and dead plant material, the CO2 concentration in the soil gaseous phase is higher com- pared to atmospheric concentrations, especially in wet periods (Simojoki et al. 1991). In our case, the added re- active lime might be initially able to bind a considerable proportion of the CO2 lost in the calcination step. How- ever, in noncalcareous soils, CaCO3 is an unstable mineral, and in Finland, the maintenance of soil pH has been calculated to require lime amendment with CaCO3 at an annual rate of 300 kg ha-1 (Elonen et al. 1991, Sallasmaa 1991). The requirement for pH maintenance is due to the acidifying effects of ammonium-containing mineral fer- tilizers, natural soil forming processes and plant uptake of nutrient elements (see van Breemen et al. 1983). At pre- sent, the mean pH values of Finnish agricultural soils are approximately 6.0 in mineral soils and 5.5 in organic soils (Keskinen et al. 2016). Hence, we do not expect that carbonation of the active component of structure lime would be a permanent sink for CO2. Note, however, that the application of agricultural lime within the structure lime caus- es soil emissions of 440 kgCO2e t-1 (IPCC 2006). Gypsum may enhance carbonation of CO2 in alkaline soils (Zhang et al. 2016). However, we assume that carbon sequestration by gypsum is negligible in the slightly acidic Finnish soils. 0 50 100 150 200 250 300 350 400 0 20 40 60 80 100 kg CO 2e /t Share of recycled structure lime, % Fig. A1. Emissions from structure lime production as a function of the share of recycled materials M. Ollikainen et al. 111 Appendix 2. Data Table A1. Soil amendments: GHG emissions and costs Parameter Value Unit Reference Values employed for both gypsum and structure lime Transport cost 32.14 €/450 km Ollikainen et al. (2020) Gypsum Price 18.15 €/t Ollikainen et al. (2020) Used amount 4 t/ha Ollikainen et al. (2020) Transportation distance 450 km Ollikainen et al. (2020) Transport 32.14 €/t Ollikainen et al. (2020) Spreading 6.5 €/t Ollikainen et al. (2020) Structure lime Price 34.5 €/t KM (2022) Used amount, pristine, 5-year duration 5 t/ha Anttila et al. (2021) Reactive lime 1.2 t/ha 24.6% of structure lime, Kämäri et al. (2019) Agricultural lime 3.8 t/ha Used amount, recycled, 5-year duration 3.6 t/ha based on average share of reactive lime Reactive lime 1.2 t/ha 33.3% of structure lime Agricultural lime 2.4 t/ha Used amount, pristine, 10-year duration 22.9 t/ha based on average share of reactive lime Reactive lime 5.5 t/ha 24.6% of structure lime, Kämäri et al. (2019) Agricultural lime 17.4 t/ha Used amount, recycled, 10-year duration 16.5 t/ha based on average share of reactive lime Reactive lime 5.5 t/ha 33.3% of structure lime Agricultural lime 11 t/ha Transportation distance 65 km distance between Vampula and Lieto Transport 4.64 €/t based on transport distance and Ollikainen et al. (2020) Spreading 10 €/t assumed same as for gypsum Disc harrowing 35.3 €/ha Palva (2021) Agricultural and Food Science (2024) 33: 90–115 112 Table A2. Parameters used in the simulations Parameter Value Unit Reference Barley cultivation Barley market price 0.174 €/kg Luke (2019) NPK fertilizer price (YaraMila Y3) 0.4 €/kg ProAgria (2019) Share of N in NPK fertilizer 23 % ProAgria (2019) Variable cost 138 €/ha ProAgria (2019) Fixed cost 362 €/ha ProAgria (2019) Price of agricultural lime 9.0 €/t KM (2022) Used amount of agricultural lime 4.5 t/ha average, every five years N fertilization limit 100 kg/ha Finnish Food Authority (2020) P fertilization limit 5 kg/ha Finnish Food Authority (2020) LFA support payment 217 €/ha Pro Agria (2019) Barley N yield response Lehtonen (2001) Nutrient runoff PP runoff function, kgPP/ha based on Saarela et al. (1995) and Uusitalo & Jansson (2002) DRP runoff function, kgDRP/ha based on Saarela et al. (1995) and Uusitalo & Jansson (2002) N runoff function, kgN/ha Simmelsgaard (1991) P fertilization 5 kg/ha set to the limit according to Finnish Food Authority (2020) Soil phosphorus 15.9 mg/l Eurofins (2017) Erosion 1100 kg/ha chosen to represent average value in the study areaRunoff 247 mm Plant- and technology specific parameter 0.6 Puustinen et al. (2010) Plant- and technology specific parameter 2.2 Puustinen et al. (2010) N runoff at average fertilization 15 kg/ha Lankoski et al. (2006) Constant for N runoff, -0.7 Simmelsgaard (1991) Constant for N runoff, 0.7 Simmelsgaard (1991) PP reduction, gypsum 50 % literature review (this study) DRP reduction, gypsum 25 % literature review (this study) PP reduction, structure lime 40 % literature review (this study) DRP reduction, structure lime 0 % literature review (this study) GHG emissions related to barley cultivation Soil emissions 1855 kgCO2e/ha Ervola et al. (2012) Mineral fertilizer manufacture 4.32 kgCO2e/kgN Ervola et al. (2012) Soil N2O emissions from mineral fertilization 6.71 kgCO2e/kgN Ervola et al. (2012) Management practices (incl. tillage, harrowing, seeds, planting, weed control, harvesting, transport of harvest to drier) 489 kgCO2e/ha based on Ervola et al. (2012) Agricultural lime, production 131 kgCO2e/t West and Marland (2002) Agricultural lime, soil emissions 440 kgCO2e/t (IPCC, 2006) Agricultural lime, spreading 2 kgCO2e/t based on Ervola et al. (2012) Other parameters P to Ne (Redfield-ratio) 7.2 Kiirikki et al. (2003) Marginal runoff damage 11.6 €/kgNe based on Gren (2001), prolonged to year 2023 using 2 % interest rate Marginal climate damage 35 €/tCO2e based on Tol (2011) Δ[𝜁𝜁𝜁𝜁{250 ln(𝜎𝜎𝜎𝜎 + 0.01)𝑃𝑃𝑃𝑃 − 150}] 106 𝜓𝜓𝜓𝜓[𝜙𝜙𝜙𝜙0.021(𝜎𝜎𝜎𝜎 + 0.01𝑃𝑃𝑃𝑃) − 0.015] 100 𝜑𝜑𝜑𝜑𝑒𝑒𝑒𝑒𝑏𝑏𝑏𝑏𝑜𝑜𝑜𝑜+𝑏𝑏𝑏𝑏1 𝑛𝑛𝑛𝑛 100 M. Ollikainen et al. 113 Appendix 3. Results for structure lime from recycled materials Impacts on P load and GHG emissions The impact on the P load is the same as that of structure lime from pristine materials: 40% PP load and 0% DRP load (see Table 3 in the text). Table A3 provides data on GHG emissions without soil emissions drawing on the critical assumption that emissions from the production of recycled materials are close to zero. GHG emissions are lower than for other soil amendments for 5- and 10-year durations. The costs of a 10-year duration are higher than those of gypsum. Structure lime contains agricultural lime the application of which to the soil causes emis- sions. For a 5-year duration, the share of agricultural lime is 2.4 t ha-1 with soil emissions of 1056 kgCO2e ha-1. For a 10-year duration, the share of agricultural lime is 16.5 t ha-1 with soil emissions of 4840 kgCO2e ha-1. These emis- sions are largely displayed since conventional agricultural lime is not applied for fields treated with structure lime. Cost-effectiveness and social net benefits The cost-effectiveness ratio in Table A4 is better than that of structure lime from pristine materials but falls short of that of gypsum when only the P load or both P load and GHG emissions are considered. In the case of 5- year duration, it gives the highest private profits of all soil amendments and the second highest social welfare after gypsum when only the P load or both the P load and GHG emissions are considered. When targeted to 10-year duration, it performs weakly but still outperforms structure lime from pristine materials. Table A3. Emissions and costs from using recycled structure lime Structure lime, recycled materials, 5 y Structure lime, recycled materials, 10 y Emissions, kgCO2e ha-1 Costs, € ha-1 Emissions, kgCO2e ha-1 Costs, € ha-1 Production/Price 13.1 124 60.2 569 Transport 7.05 16.7 7.05 76.6 Spreading 19.6 26.0 16.9 26.0 Harrowing 39.2 70.6 39.2 70.6 Total 79.0 238 126 742 Table A4. Cost-effectiveness (€ kg-1 P reduction) of recycled structure lime based on treatment cost Reduction in totP load, kg ha-1 y-1 Cost-effectiveness (€ kg-1 P) I I + GHG Structure lime, recycled (5 y) 0.53 65.1 60.1 Structure lime, recycled (10 y) 0.53 129 177 Table A5. NPV of private profits and social net benefits over 5 or 10 years with a 3% discount rate (€ ha-1) Private profits (NPV) Social net benefits (NPV) P GHG P + GHG Structure lime, recycled, effective 5 y, € ha-1 5 y-1) 1404 898 758 252 Structure lime, recycled, effective 10 y (€ ha-1 10 y-1) 2044 1102 563 -379 Agricultural and Food Science (2024) 33: 90–115 114 Appendix 4. Sensitivity results for the cost-effectiveness ratios Table A6. Cost-effectiveness (€ kg-1 P reduction) of gypsum and structure lime amendments based on annual treatment costs only (I) and annual treatment costs and increased climate costs (I+GHG) Reduction in totP load, kg ha-1 y-1 Cost-effectiveness (€ kg-1 P) I I + GHG Baseline Gypsum 0.78 58.2 59.0 Structure lime 0.53 86.0 121.7 Gypsum, lower P reduction (PP 34%, DRP 0%) Gypsum 0.45 101.8 103.2 Gypsum, higher P reduction (PP 72%, DRP 25%) Gypsum 1.07 42.5 43.1 Structure lime: 30% PP reduction Structure lime 0.39 114.7 162.3 Structure lime: 50% PP reduction Structure lime 0.66 68.8 97.4 Lower erosion (875 kg ha-1) Gypsum 0.65 70.3 71.3 Structure lime 0.42 108.1 153.0 Higher erosion (1375 kg ha-1) Gypsum 0.94 48.1 48.8 Structure lime 0.66 68.8 97.4 Lower soil P status (8 mg l-1) Gypsum 0.51 88.7 89.9 Structure lime 0.36 125.6 177.8 Higher soil P status (25 mg l-1) Gypsum 0.99 46.0 46.6 Structure lime 0.63 71.2 100.8 Higher climate damage (50 € t-1CO2e) Gypsum 0.78 58.2 59.4 Structure lime 0.53 86.0 137.0 M. Ollikainen et al. 115 Appendix 5. Reducing PP runoff using buffer strips The PP runoff function with buffer strip is based on Saarela et al. (1995) and Uusitalo and Jansson (2002): where b is the share of the buffer strip from a 1 ha field, and α (based on Lankoski et al. (2006)) denotes the ability of the buffer strip to capture PP. Setting P = 5 and b = 0 gives the baseline PP runoff of 1.31 kg ha-1. To reduce PP runoff by 40%, i.e., to 0.788 kg ha-1, and keeping P fertilization at 5 kg ha-1, the value of b has to increase to 0.10. This means a 10-meter buffer strip on one side of a 1 ha square field. To reduce PP loading by 50%, i.e., to 0.656 kg ha-1, b needs to be 0.18, i.e., 18 meters. To evaluate the cost-effectiveness, private profits in the baseline and with the increased buffer strips are needed. In the baseline, private profits without the use of agricultural lime are 348 € ha-1 (see Table 5). The additional costs from buffer strip are the establishment and maintenance cost of 10 € ha-1 (Lankoski and Ollikainen 2006) and the reduced revenue from crop yield. Profits with a 10-meter buffer strip are 287 € ha-1, which gives a 61 € ha -1 re- duction in profits and further a cost-effectiveness of € kg-1P. Profits with an 18-meter buffer strip are 242 € ha-1, which gives a 106 € ha-1 reduction in profits and further a cost-effectiveness of € kg-1P. (1 − 𝑏𝑏𝑏𝑏𝛼𝛼𝛼𝛼)𝛥𝛥𝛥𝛥[𝜁𝜁𝜁𝜁{250 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙(𝜎𝜎𝜎𝜎 + 0.01) (1 − 𝑏𝑏𝑏𝑏)𝑃𝑃𝑃𝑃 − 150}] 106 348 − 287 1.31 − 0.788 ≈ 116 314 − 42 1.13 − 0.656 ≈ 161 Gypsum and structure lime amendments in boreal agriculturalclay soils: Do climate emissions compromise water qualitybenefits? Introduction Impacts of gypsum and structure lime amendment on phosphorus loading– a literature review Effects of gypsum on P loading Effects of structure lime on P loading P runoff in the Archipelago Sea catchment area Determination of GHG emissions and costs from gypsum and structure limeamendments Framework and data Net benefits and cost-effectiveness Agronomic data and specification of the case study Cost-effectiveness and net benefits of soil amendments Cost-effectiveness in reducing P loading Private and social net benefits Sensitivity analysis: the impact of exogenous variables Scaling up the use of soil amendments Discussion Conclusions Acknowledgements References