Appendix A. Individual conceptual networks representing agricultural management and ecosystem services provision.

The relationships between agricultural management and eight ecosystem services (ES) provided by Pampean agroecosystems were represented in individual conceptual networks (Figs. 1 to 8). The eight conceptual networks developed in this work contained five types of nodes, and four types of logical links between nodes (see Section 2.3.). These logical links present capital letters in order to easily explain each conceptual network.

1. Supporting Service: Elements cycling - C Balance

Fig. 1 Conceptual network representing functional relationships between agricultural management and provision of the Supporting service: Elements cycling - C balance. Capital letters represent the logical links between nodes. Legend: circles meaning input variables; rounded-squares meaning decision variables; squares meaning state variables; triangles meaning ecosystem processes and diamond meaning ecosystem service provision indicators. Tº: temperature, and Pp: rainfall

It is generally known that C inputs in soils consist of crop residues and roots, and sometimes additions of soil organic amendments; while C loss is caused by humus and residue mineralization, in conditions where soil erosion and C leaching are minimal (C leaching is not an important cause of soil organic carbon (SOC) losses in Pampean agroecosystems (Roberto Álvarez, personal communication)) (Oorts and others 2006). The interaction between temperature and rainfall regulates SOC through the influence of soil organic matter (SOM) mineralization (Fig. 1, Relations A and B) (Roberto Álvarez and Raúl Lavado, personal communication). High temperature reduces SOC because of intense SOM mineralization, while there is no linear answer for rainfall (Fig. 1, Relations A, B and F) (Álvarez and Lavado 1998). However, it is widely accepted that, in general, rainfall has the same effect as temperature (Fig. 1, Relations B and F) (Roberto Álvarez and Raúl Lavado, personal communication). Additionally, crop species, their growth rate and yield determine the amount and type (i.e., quality) of crop residue (including crop roots) (Fig. 1, Relations D and I) (Ernst and others 2002) which change SOC in surface soil layers, specially in no-tillage systems (Álvarez and Lavado 1998). Surface soil layers have greater C amounts because of the input of crop residue from harvested plants (Álvarez and Lavado 1998). Generally, legume species (e.g., soybean) have higher mineralization rates than gramineous species (e.g., wheat, maize) due to lower C/N relations (Fig. 1, Relation E) (Ernst and others 2002).

Another conditioning factor of SOC reduction is erosion vulnerability which is higher in continuous cropping systems, principally by 1) removing C from one site and depositing it elsewhere, and 2) promoting soil degradation and then reducing productivity (Fig. 1, Relation G) (Martínez-Mena and others 2008). However, it can be assumed that SOC movement is dependent on the topographic position (Haydée Steinbach and Roberto Álvarez, personal communication). Soils under no-tillage reduce both eolic erosion in semiarid sites, and hydric erosion in sites with great slopes (Monzon and others 2006).

2. Supporting Service: Elements cycling - N Balance

Fig. 2 Conceptual network representing functional relationships between agricultural management and provision of the Supporting service: Elements cycling - N balance. Capital letters represent the logical links between nodes. Legend: circles meaning input variables; rounded-squares meaning decision variables; squares meaning state variables; triangles meaning ecosystem processes and diamond meaning ecosystem service provision indicator. Tº: temperature, and Pp: rainfall

Soil nitrogen (N) availability is modulated by four main factors: SOM mineralization, crop residue, fertilization regime and N losses (Cassman and others 2002). N mineralization through SOM is a very important supply source due to its usage availability (Fig. 2, Relation G), increasing or decreasing crop yield (Fig. 2, Relation J) (Bono and Álvarez 2007). The increase in soil moisture content increases mineralized N (Fig. 2, Relations B and F) (Helena Rimski-Korsakov, personal communication). This increase is a direct consequence of higher microbial activity, until the concentration of oxygen in the soil becomes a limitation for the microorganisms (Navarro and others 1991). Moreover, SOM is not only affected by mineralization but also by crop residue disposal on soil surface layers (Fig. 2, Relation E) (Ernst and others 2002), as it occurs for SOC in no- and reduced tillage systems. N fertilization can increase the amount of soil N pools which will be available for crops (Fig. 2, Relation K) (Abril and others 2007). However, N excedent can also be immobilized by microorganisms, resulting in a non linear effect (i.e., increase or reduce) of the application (Fig. 2, Relation K) (Cassman and others 2002; Portela and others 2006). Furthermore, N losses by denitrification, volatilization or leaching are the main causes for the low efficiency in the use of N, and therefore they affect available N in soil (Fig. 2, Relation L) (Abril and others 2007). Because of the low degree of these losses during the whole crop growth cycle (Álvarez and Grigera 2005), they can be grouped all together under the name of N losses (Roberto Álvarez, personal communication).

3. Supporting Service: Water cycling - Soil water balance

Fig. 3 Conceptual network representing functional relationships between agricultural management and provision of the Supporting service: Water cycling - Soil water balance. Capital letters represent the logical links between nodes. Legend: circles meaning input variables; rounded-squares meaning decision variables; squares meaning state variables; triangles meaning ecosystem processes and diamonds meaning ecosystem service provision indicators. Tº: temperature, and Pp: rainfall

In Pampean agroecosystems, water supply for crops is determined by nine variables: 1) evaporation, 2) runoff, 3) soil structural stability, 4) soil texture, 5) aquifer depth, 6) soil depth, 7) presence of weeds/fallow/cover crops, 8) irrigation, and 9) rainfall (Fig. 3, Relations M, N, O, P, Q, R, S, T and C). These variables, in general, increase or affect water supply for crops. For instance, no-tillage systems leave crop residue on the soil surface and, therefore, soil evaporation is clearly decreased (Fig. 3, Relations E and F) (Monzon and others 2006). Relative soil evaporation rates directly influence the amount of soil water retained which will be used by the crop (Fig. 3, Relation M) (O´Leary and Connor 1997). Stubble mulch protects the surface soil from erosion and runoff, and increases water storage by minimising surface sealing and enhancing infiltration, as well as by directly reducing evaporation (Fig. 3, Relations G, J, N and O) (O´Leary and Connor 1997). Moreover, irrigation not only increases water supply for crops (Fig. 3, Relation T) but also affects runoff, depending on the amount of water irrigated and crop residue on soil surface (Fig. 3, Relation L) (Olga Heredia, personal communication). Systems under no-tillage can increase soil water accumulation during fallows (Fig. 3, Relation S), and thereby offer the potential for affecting crop yield in Pampean agroecosystems (Olga Heredia and Francisco Bedmar, personal communication) (Fig. 3, Relation U).

Soil depth is related with the ability of roots to explore soil profile and to absorb water stored there (Fig. 3, Relation R); on the other hand, aquifer depth can be defined by characterizing the average depth fluctuation of water table in different regions (Fig. 3, Relation Q) (Esteban Jobbágy, personal communication). This is specially important in sandy soils (Claudia Sainato, personal communication). Finally, weeds can be burned to avoid evaporation as well as the establishment of cover crops (Fig. 3, Relation S) (Olga Heredia and Silvina Portela, personal communication).

4. Supporting service: Soil conservation – Soil structural maintenance

Fig. 4 Conceptual network representing functional relationships between agricultural management and provision of the Supporting service: Soil conservation – Soil structural maintenance. Capital letters represent the logical links between nodes. Legend: circles meaning input variables; rounded-squares meaning decision variables; squares meaning state variables; triangles meaning ecosystem processes and diamonds meaning ecosystem service provision indicators. Tº: temperature, and Pp: rainfall

Structural stability is defined as soil capacity to preserve the system of solids and pore space, when subjected to different external disturbances (e.g., tillage) (Taboada and Micucci 2002). Its loss is the critical factor which determines structural deterioration. This deterioration is evidenced by the formation of surface crusts, higher rates of runoff and soil loss due to erosion, as well as reduced water storage (Taboada and Micucci 2002). Soil structural stability is clearly affected by land use, which is in turn positively associated with crop residue, total organic C concentration and the forms of organic C (Fig. 4, Relations I and J) (Caravaca and others 2004). The close association found between structural stability, labile carbon and microbial biomass confirms both their importance in the mineralization process and their ability as aggregate cementitious (Fig. 4, Relations G and H) (Urricarriet and Lavado 1999). According to the first statement, SOM decomposition may be limited by pore size distribution due to the localization of SOM in pores inaccesible to microorganisms, a limited nutrient supply to microorganisms and restricted predation of those microorganisms (Miguel Taboada and Roberto Casas, personal communication). Furthermore, soil structural stability is one of the most important characteristics affecting crop yield (Fig. 4, Relation K) because it affects root penetration, water storage capacity, and air and water movement in soil (Fig. 4, Relation O) (Aparicio and Costa 2007).

5. Regulating Service: Climate regulation – N2O emission control

Fig. 5 Conceptual network representing functional relationships between agricultural management and provision of the Regulating service: Climate regulation – N2O emission control. Capital letters represent the logical links between nodes. Legend: circles meaning input variables; rounded-squares meaning decision variables; squares meaning state variables; triangles meaning ecosystem processes and diamonds meaning ecosystem service provision indicators. Tº: temperature, and Pp: rainfall

Although denitrification is only part of direct N2O emissions from soils, it is the most studied process in contrast with nitrification occurring in unsaturated soils, among other conditions (Fig. 5, Relation P) (Laura Yahdjian, personal communication). Thus, the main factors controlling denitrification are: soil pH, soil texture, nitrate concentration, C availability, aeration and moisture content (Guo and Zhou 2007). However, the major factors to consider, in terms of N2O production in Pampean agroecosystems, are available N in soil and moisture content (in this case, rainfall) (Fig. 5, Relations N and O) (Palma and others 1997; Ciampitti and others 2005). For instance, it is known that the presence of actively growing plants limits the denitrification process in comparison with those treatments without plants, due to reduced water availability and to lower levels of nitrates in soil, to a lesser extent (Sainz Rozas and others 2004). Once the crop is harvested and crop residue remains on the surface, soluble C concentration is associated with denitrification (Fig. 5, Relation E); this is because bacteria biomass capable of denitrification is probably controlled primarily by C availability under aerobic conditions (Fig. 5, Relation M) (Miguel Taboada, personal communication), while emissions occur mainly during anaerobic conditions (Fig. 5, Relation O).

6. Regulating Service: Water purification - Groundwater contamination control

Fig. 6 Conceptual network representing functional relationships between agricultural management and provision of the Regulating service: Water purification - Groundwater contamination control. Capital letters represent the logical links between nodes. Legend: circles meaning input variables; rounded-squares meaning decision variables; squares meaning state variables; triangles meaning ecosystem processes and diamonds meaning ecosystem service provision indicators. Tº: temperature, and Pp: rainfall

Nitrate (NO3) leaching is one of the main causes for groundwater contamination (Fig. 6, Relations N and O) (Abril and others 2007; Claudia Sainato and Olga Heredia, personal communication). However, Mugni and others (2005) measured NO3 concentration in four Pampasic streams and concluded that it was relatively modest compared to intensively cultivated basins in Europe and North America. Consequently, there is a slow N enrichment of water resources in Pampean agroecosystems (Portela and others 2006). Water quality is reduced not only by N fertilization (Fig. 6, Relation J) (Rimski-Korsakov and others 2004; Abril and others 2007), but also SOM mineralization through several years removes great amounts of NO3 towards aquifers (Portela and others 2006; Helena Rimski-Korsakov and Raúl Lavado, personal communication) (Fig. 6, Relation G). N fertilization could also be indirectly inducing soil NO3 leaching, by altering the ability of plants root system to acquire N from soil and net mineralization rate from organic N pools (Cassman and others 2002). Furthermore, fertilization in excess of crop requirements or water excedent, such as rainfall events (Fig. 6, Relation M) or irrigation (Fig. 6, Relation K), increase the probability of soil NO3 leaching (Costa and others 2002; Rimski-Korsakov and others 2004; Vergé and others 2007). It is important to clarify that the Pampa region has low N inputs through rainfall (Portela and others 2006). Other factors affecting soil NO3 leaching are particle size distribution, soil porosity and the ocurrence of preferential flow paths. These causes can be grouped under soil texture, which is another important factor because of its ability for retaining water (Fig. 6, Relation L) (Taboada and Micucci 2002).

7. Regulating Service: Regulation of biotic adversities

Fig. 7 Conceptual network representing functional relationships between agricultural management and provision of the Regulating service: Regulation of biotic adversities. Capital letters represent the logical links between nodes. Legend: circles meaning input variables; rounded-squares meaning decision variables; squares meaning state variables; triangles meaning ecosystem processes and diamonds meaning ecosystem service provision indicators

In Pampean agroecosystems, crop environment is determined by: 1) tillage system, 2) crop protection, 3) sowing density, 4) sowing date, 5) fertilization, 6) genotype selection, and 7) irrigation (Fig. 7, Relations B, A, C, D, E, F and G). These variables affect not only crop yield but also species composition and abundance of plant and animal community, and beneficial species (Fig. 7, Relations H, J and L) (Emilio Satorre and Elba De la Fuente, personal communication). Beneficial species as well as crop environment and crop changes affect species composition and abundance/incidence of pests, diseases and weeds (Fig. 7, Relations K, L and N). The latter reduces crop yield and affects natural pest mitigation of ecosystems (Fig. 7, Relations I and O). The presence of weeds influences the presence of diseases and diversity and abundance of two insect types: pests, with negative consequences for cropping systems, and their natural enemies (Altieri 1999). Generally, a high density of weeds is counter-productive because they reduce crop yield and its quality (Albrecht 2003).