Afforestation or intense pasturing improve the ecological and economic value of abandoned tropical farmlands

Increasing demands for livelihood resources in tropical rural areas have led to progressive clearing of biodiverse natural forests. Restoration of abandoned farmlands could counter this process. However, as aims and modes of restoration differ in their ecological and socio-economic value, the assessment of achievable ecosystem functions and benefits requires holistic investigation. Here we combine the results from multidisciplinary research for a unique assessment based on a normalization of 23 ecological, economic and social indicators for four restoration options in the tropical Andes of Ecuador. A comparison of the outcomes among afforestation with native alder or exotic pine, pasture restoration with either low-input or intense management and the abandoned status quo shows that both variants of afforestation and intense pasture use improve the ecological value, but low-input pasture does not. Economic indicators favour either afforestation or intense pasturing. Both Mestizo and indigenous Saraguro settlers are more inclined to opt for afforestation.

confidence limits according to Eq. (4). d. Key element "Soil quality" is given a very high 14 subjective weight, w sub , of 5 instead of 1 to express a strong preference for food production, 15 while all other key elements have a weight of 1, given Eq. (9). This puts intense pasture at 16 first place for the ecological assessment. See Supplementary Table 22  material and Methods main text). Thinnings after 12 and 16 years with 40% of N r removed each time; mortality rate = 2% of N r . 69 N r = number of remaining trees; N d = cumulative number of dead trees; Dbh = mean diameter at breast height Ba = basal area; h t = mean total height; v t = total 70 volume without bark; v c = commercial volume without bark; bm a = above-ground biomass; bm b = below-ground biomass; bm t = total biomass; C = carbon 71 sequestration; LAI = leaf area index; N rm = number of trees removed.   93   94  95  96  97  98  99  100  101  102  103  104  105  106  107  108  109  110  111  112  113  114  115  116  117  118  119  120  121  122 Supplementary      ecosystems. Any differences detected between the natural forest and the abandoned 296 pastures are difficult to assess because they may be due purely to inherent differences 297 in the two types of habitats, or simply to plant age.

298
In summary, clearing of the forest for agricultural land use causes an extraordinary 299 decline in vascular plant and moth diversity, which may or may not be paralleled by soil

312
The nutritive value of the grasses commonly used here is rather low, and therefore, the colonists who arrived in the study area during the 20 th century. As agro-pastoralists 37,38 , 327 they participate in both a market economy (cattle ranching for cheese, milk and meat 328 production) and a subsistence economy (crop production, horticulture and cattle 329 ranching for subsistence needs). Cattle ranching is the most important market activity, 330 and from the farmers' point of view, the most profitable. The main product drawn from 331 cattle ranching is cheese (quesillo -an unsalted fresh white cheese), which is sold in the 332 local markets. Only farmers who have good access to roads can sell milk to regional 333 producers of dairy products. Despite both ethnic groups practicing a similar land-use     The total area of repasturisation was 4,500 m 2 , and was subdivided into 4 x 4 m plots.   Table 1). Three grazing rounds per year were 399 simulated, with a higher number of cattle than in the low-input option to insure sufficient 400 trampling of the fern. High yields have been achieved on long-standing pasture on 401 favourable sites using a fertilisation protocol to replace the nutrients removed by grazing 402 (50 kg N and 10 kg P ha -1 y -1 ; 1.25-fold increase from 9.02 Mg DM ha -1 y -1 to 11.2 Mg 403 DM ha -1 y -1 ) 1,43 . In these cases, the levels of protein, N, P and Ca were also much 404 higher than in our research pastures (Supplementary Table 1 Table 5). The movement of soil water within each layer is simulated 508 using the Richards equation. Water leaving the soil column is directed to the ground 509 water using a Dirichlet boundary condition with a constant negative pressure. Slope in 510 the study area is considered using a hydraulic gradient suitable for a slope of 10%.

511
It should be stressed that CMF is directly coupled with the SoBraCo-model using a   Table   546 12), here it is modified according to activities undertaken during the 20-year model run 547 in a particular management option (e.g. timber extraction, grazing), as defined by the 548 requirements of the economic assessment ( Supplementary Fig. 1). For tree plantations, 549 the LAI scenarios are certainly very conservative. In the first years, when trees have not 550 yet formed a closed canopy (achieved after 3 to 6 years), no LAI and thus, no 551 production of biomass is modelled, although production levels similar to those in the 552 abandoned pastures are to be expected. We are aware that this leads to an 553 underestimation of the performance of the afforestation areas, but accept this. Thus, the 554 better performances obtained for the afforestation options represent conservative 555 estimates and underline the robustness of our assessment. In the MC analysis, the PDF 556 form and relative deviation (in %), shown in Supplementary Table 13 are applied to the 557 respective mean LAI scenario value at each time step t as shown in Supplementary Fig.   558 1.

559
To properly analyse the output variance of the model, we include more than 3,000  Table 15). With regard to the hydrological ratings, small 567 changes in the overland flow are observed (Supplementary Table 16). While the first 568 and second ranks remained the same for Pinus and Alnus, changes occur only in one 569 case for the indicator overland flow which is characterized by very small differences in 570 mean values. Here, the third best option is now low-input pasture changing rank 571 position with abandoned pasture. However, no change is found for the area-specific 572 discharge. The integrated P k rankings do neither change for the "less is better" option 573 nor for the "more is better option". "more is better" or the "less is better" option becomes visible (Supplementary Table 18).

584
Only the low-input and intense pasture options changed rank position.

585
In summary, the integrated assessment scheme for climatic and hydrological indicators 586 proves to be very robust against parameter uncertainties and oscillations in forcing  The soil quality indicators for plant growth presented in Supplementary Table 7 are well   592 known I) to vary in response to land-use change 43 , II) to support plant productivity 60 and 593 III) to contribute to soil biodiversity 61 . The variation of indices when using 95 % 594 confidence limits is reported in Supplementary Table 19. where t is a point in time; T is the period of 20 years; r i,t is the net revenue of option, i, at 603 time, t; q is the discount factor, d is the discount rate (either 5 or 8 %); and t* indicates 604 the payback period, defined as the time needed until cumulative discounted net 605 revenues cover the up-front costs.

606
The results obtained are a good fit to the range of values achieved in other studies 64 . 607 We concentrate on provisioning services when investigating the potential value of a 608 land-use option, as financial consequences are crucial drivers of land-use decisions 65 .

609
As labour plays an important role in both forestry and agriculture, a daily wage of 10 610 US$ is assumed. In the case of afforestation, the harvestable volume (m³ ha -1 without 611 bark) of commercial and non-commercial timber is the production quantity of saw timber 612 and fuel wood respectively (Supplementary Table 8). Costs for site preparation and 613 planting are also taken into account (603 US$ ha -1 for A. acuminata, and 570 US$ ha -1 614 for P. patula). These costs include marking, distribution of saplings, removal of bracken, 615 seedling prices (A. acuminata: 0.3 US$; P. patula: 0.27 US$ per seedling), labour used 616 for planting, and tools. In addition, expenses for the installation (60 US$ ha -1 ) and 617 annual maintenance (40 US$ ha -1 ) of fire breaks and extraction lines are considered.

618
Revenues minus these costs are calculated with respect to both the quantity and quality 619 of timber extracted (costs: 26.7 US$ m -³ for saw timber and 7 US$ m -³ for fuel wood;  Agriculture Organization (faostat.fao.orgs). The production of milk and meat is based on 625 the number of cows per hectare (low-input: 0.4 cows ha -1 ; intense: 1.1 cows ha -1 ), and 626 allows for a minimum pasture preparation phase of either 24 (low input pasture use) or 627 18 (intense pasture) months. Expenses for site preparation and planting sum to 950 628 US$ ha -1 for low-input and 1,400 US$ ha -1 for intense pasture. Costs included in these 629 figures are labour for mechanical (low-input) or chemical (intense) bracken removal, 630 tools, purchase, transport and planting of mother tussocks (32,400 ha -1 ) and, in the 631 case of intense pasture, fertilisation. While the purchase price of cows is excluded from 632 the expenditures considered, costs for infrastructure such as fencing (year 1; 6; 11; 16), 633 as well as expenses for veterinary care (13 US$ cow -1 y -1 ) are also calculated.

634
Assuming a proportion of cows for milk production and cows for meat production of 1:1, 635 production costs (costs for milking: 13 US$ ha -1 for low-input and 34 US$ ha -1 for 636 intense pasture) versus revenues (0.34 US$ l -1 for milk and 1.9 US$ kg -1 for meat are 637 calculated (Supplementary Table 8). Finally, discounted returns (5% and 8%) and 638 payback periods are calculated for all options.

639
To obtain a reference scenario, we also simulate the economic coefficients of the 640 currently prevailing form of land-use in the study region, which is low-input pasture after  Table 9). From year 2 onwards, the same 647 net revenues are considered as those assumed for low-input pasture. to 5 (lowest rank). Two scenarios are then tested -one in which farmers rehabilitate the 663 abandoned areas using their own means (without subsidies) and a second in which 664 farmers receive financial support for major inputs (e.g. seedlings, fertiliser, labour) from 665 external agencies (with subsidies). Performance indices (P i ) are calculated from R i 666 values which represent the total number of responses indicating a specific option as 667 either the best or the second best alternative (see Table 8