Introduction
The Industrial Revolution and intensification of agriculture have in some cases led, for the first time since the emergence of permanent settlements and agriculture more than 12,000 years ago, to economic activities that profoundly influence the ecosystem to the point where environmental stability and geographic political security are jeopardizedReference Rees and Wackernagel1, Reference Scialabba and Muller-Lindenlauf2. Thus, the World Commission on Environment and Development (the Brundtlandt Commission) coined the definition of sustainable development in 1987—it is development that satisfies the needs of current generations without compromising the needs of future generations3. In recent years, numerous tools and methods have emerged that are supposed to determine sustainable development on the level of single enterprisesReference Veleva, Hart, Greiner and Crumbley4 as well as on a higher, societal levelReference Lenzen and Murray5, Reference Chen, Jiang, Yang, Chen, Ji and Zhou6. One of these tools is the environmental or ecological footprintReference Rees and Wackernagel1. It aims at estimating the biologically productive area needed to produce materials and energy used by the population of a certain region (city, state and world). The calculated area is compared to the area available to a certain population or individual, called the biocapacity, which presents the productive land and/or water of a region. In cases where the ecological footprint is greater than the biocapacity, human consumption exceeds the natural carrying capacityReference Haberl, Erb and Krausmann7. Data for the ecological footprint are usually based on statistical data; in the case of agriculture, yearly statistics of individual countries from the Food and Agriculture Organization (FAO) of the United Nations are used. The drawback of such data lies in their inherent inaccuracy, making the footprint less useful for evaluating smaller units, e.g., single farms.
Other tools based on actual and/or real data are more appropriate to evaluate individual production processes. A framework for applying such evaluation methods is life cycle assessment (LCA), which considers the environmental burden caused by a product, a production process or any activity to provide servicesReference Curran8. It takes into account the technological processes of all activities along the life cycle, from the provision of basic materials to transportation into and from the production unit to the production process itself and finally the use phase of any product and its safe disposal. It is based on an eco-inventory identifying all material and energy flows exchanged with the environment along the whole life cycle. These flows are then evaluated with an appropriate ecological evaluation method. The result can be interpreted on a per-unit-of-product basis (kg) or equivalent area (ha), where areas used outside of the production unit are includedReference van der Werf, Tzilivakis, Lewis and Basset-Mens9. One drawback of this approach is the limited comparability of the results as they critically depend on the scope of the LCA, which may differ from study to study, even for the same products or services.
Research in the area of the ecological footprint or LCA in agriculture is still developing. In this paper, we will apply ecological evaluation using the LCA framework and compare the production of field crops in different production systems by the Sustainable Process Index® (SPI)Reference Narodoslawsky and Krotscheck10–Reference Sandholzer and Narodoslawsky13. This evaluation method has been customized for agriculture, e.g., by introducing an algorithm to account for the impact of seed production. We used experimental data from a systems comparison field trial over 3 years; therefore, results reflect conditions in real-life situations and farming systems. The main question we posed was: how sustainable are the production systems most commonly used today (exemplified by Slovenian wheat and spelt production) and how can they be improved to increase sustainable food production for future generations?
Materials and Methods
Long-term field trial
The experimental site is located at the University Agricultural Centre of the University of Maribor in Pivola near Hoče (46°28′N, 15°38′E, 282 m a.s.l). The annual mean air temperature of the area is 10.7°C; where the mean monthly minimum is in January at 0.4°C and the average monthly maximum is in July at 20.8°C. Average annual rainfall in the area is around 1000 mm. Sixty 7 m×10 m experimental field plots were established on a dystric cambisol (deep) [average pH value 5.5 (0.1 M KCl solution), soil soluble P at 0.278 g kg−1 and soil soluble K at 0.255 g kg−1 in ploughing soil layer], and are maintained within two different five-course crop rotation designs (Table 1). In one rotation there are typical crops for this region [2 years of red clover–grass mixture, winter wheat (Triticum aestivum L.), white cabbage (Brassica oleracea L. var. capitata L. f. alba), oil pumpkins (Cucurbita pepo var. styriaca Greb.]; in the other one there is an alternative crop rotation [2 years of red clover–grass mixture, spelt (Triticum spelta L.), red beet (Beta vulgaris L.), false flax (Camelina sativa L.)]. Since we wanted to compare the same crops in each year of the trial, this resulted in three different combinations of each crop rotation (Table 1). Two years prior to the beginning of the trial a red clover–grass mixture was grown on site and the whole experimental plot was managed according to organic farming standards for 6 years before the trial started in 2007. Four production systems+control plots were arranged in a randomized complete block split-plot design with four replicates, where there were three main plots in each farming system and replicate, which were then split into the two different crop rotations. The farming systems differed mostly in plant protection and fertilization strategies and are defined by the valid legislation and standards (Table 2)—conventional (CON)14, integrated (INT)14, 15, organic (ORG)14, 16, 17, biodynamic (BD)14, 16–18 farming system and control14 plots, where no fertilization/plant protection was used. Basic soil cultivation, sowing and harvesting dates and methods were identical among experimental plots and were performed on the same dates and in the same manner to adjacent fields (Table 2). Also, the same varieties were used in all farming systems under study (wheat ‘Antonius’ and spelt ‘Ebners Rotkorn’), of conventional origin for CON and INT systems and of organic origin for ORG, BD and control systems.
Crop rotation designs for years 2006–2010.
Farming systems under investigation in the field trial and differences between them.
BD, biodynamic; EC, European Council; GAP, good agricultural practice; INT, integrated farming; LU, livestock units.
SPIonExcel tool
The SPI, developed by Krotscheck and NarodoslawskyReference Krotscheck and Narodoslawsky11, is based on the assumption that a sustainable economy builds only on solar radiation as natural income. Most natural processes are driven by this income and the Earth's surface acts as the key resource for the conversion of solar radiation into products and services. Global surface area is a limited resource in a sustainable economy, and anthropogenic as well as natural processes compete for this resource. Therefore, area needed to embed a certain process sustainably into the ecosphere is a convenient measure for ecological sustainability; the more area a process needs to fulfill a service, the more it ‘costs’ from an ecological sustainability point of view.
Human activities exert impacts on the environment in different ways. On the one hand they need natural resources (e.g., to provide energy, material means of production like fertilizers, etc.), manpower and area for installations. On the other hand they produce emissions and waste besides the intended goods. Therefore, the SPI includes all these different aspects of ecological pressure on the environment and translates them into surface area required by the process. The conversion of mass and energy flows into area is based on two general ‘sustainability principles’Reference Sandholzer and Narodoslawsky13:
• Principle 1. Anthropogenic mass flows must not alter global material cycles; as in most global cycles (such as the carbon cycle) the flow to long-term storage compartments is the rate-defining step of these dynamic global systems; flows induced by human activities must be scaled against these flows to long-term stores.
• Principle 2. Anthropogenic mass flows must not alter the quality of local environmental compartments; here the SPI method defines maximum allowable flows to the environment based on the natural (existing) qualities of the compartments and their replenishment rate per unit of area.
Further details of this method would be out of scope for this paper; the basic algorithm used in this work is given below and the method is described in detail elsewhereReference Sandholzer and Narodoslawsky13, Reference Narodoslawsky and Krotscheck19.
The software SPIonExcel was developed to bring this methodology into an easily applicable form. It is available on the Internet (http://spionexcel.tugraz.at/) and calculates the ecological footprint of a process, product or service given an eco-inventory summarizing the flows to and from the environment over the life cycle in question.
For this paper, the SPIonExcel tool is modified to increase its applicability for agricultural systems, employing slightly different calculation methods compared to the original method (in particular by taking seed production into account) and using a detailed inventory and database for different production systems.
The modified SPIonExcel tool calculates a total ecological footprint (A tot) that is the area necessary to embed the whole life cycle generating a product (e.g., wheat) into the ecosphere. A tot is calculated from ‘partial footprints’ using the following equation:
where A l stands for the footprint of direct land use, A fp for the footprint fertilizer and pesticide, A m for the footprint derived for machinery use and A s for the footprint of seed use. Partial footprints were calculated directly from the experimental field trial data, except for the footprints of seed use, which were determined by using Equation 2 from the intermediate footprint (up to seed) of a production system:
where Y a stands for quantity (in this case the yield of grain) of a crop produced in 1 year and S a for the quantity of seed used for crop establishment in a year.
From the attained total ecological footprint, an additional overall footprint per unit was calculated, namely:
where a tot gives an appraisal of the ‘cost’ in terms of ecological sustainability of a given product or service by indicating how much surface area is needed to produce one unit of a product, in our case wheat or spelt grain.
The area derived from the above calculation can be related to the surface area that is statistically available to a person in a country, region or area (a inh), which can be obtained from statistical data. This relation then represents the fraction of the ‘sustainable ecological budget’ for a person consuming the product in question provided by a particular production system. This value is called the SPIReference Sandholzer and Narodoslawsky13:
As the number would be too small if given on a per-kg basis, it was multiplied by 1000 to give it on a per-ton basis and to better visualize differences between production systems.
The efficiency of a production system in providing a good or service is, however, better expressed through the ecological efficiency of production (EEP) calculated in Equation 5. It provides us with the information on how much of a good or service can be produced on 1 ha of surface area in 1 year with the process or system under study, embedding the provision of this good or service totally and sustainably in the ecosphere:
Data used
All work done on the trial in the years 2008, 2009 and 2010 was carefully monitored and recorded. Data collected from the field trial were transformed into tasks done in a system in 1 year and the time needed for those tasks (e.g., ploughing, seeding, harrowing, spraying, etc.). An example is given for wheat production in the year 2009 (Table 3). Because of the nature of the trial, in which not all operations could be done by machine (e.g., spraying and fertilization), real-life operational times were taken from the University Agricultural Centre Farm, where the experiment took place. The footprint was determined for 1 ha of area.
Sample technological chart for wheat (T. aestivum L. ‘Antonius’) production in the year 2009 with all inputs and machinery use noted.
Intensity of machinery use:
* light; **, normal; ***, heavy; HP, horsepower; BD, biodynamic. A detailed description of BD preparations can be found in Turinek et al.Reference Turinek, Grobelnik-Mlakar, Bavec and Bavec34.
Statistical analysis
Data for the yield, a tot, SPI and EEP were analyzed by multifactor ANOVA with production system and year as factors using Statgraphics Centurion (Version XV, StatPoint Technologies, Inc., Warrenton, VA) and were followed by least squares means comparisons after DuncanReference Hoshmand20. Values given within the paper are means±standard error (SE).
Results and Discussion
Yields
Yields of wheat and spelt varied among production systems and years (Table 4), with no significant interaction between factors. Highest yields of wheat were attained in the CON production system (4263 kg ha−1), the lowest in the control and ORG production systems (2467 and 2450 kg ha−1, respectively). CON wheat yields were similar to recent average Slovenian wheat yields21, but lower than those reported by other EU countries. BD and INT systems performed near the average of all farming systems; ORG and control wheat yields tended to be lower (Table 4).
Yields of wheat and spelt depending on production system and year.
1 Wheat yield is given for hulled grain, but spelt yield includes hulls.
2 Average value of each factor=100%.
Mean values±SEs are presented. Different letters indicate statistically significant differences at 95% probability (Duncan test). Levels of significance: n.s., non significant (P>0.05);
* P⩽0.05; ***P⩽0.001.
We see a more uniform picture with spelt yields, where differences among production systems are not as accentuated as in the case of wheat. Possibly this is due to the lower breeding modifications of spelt as compared to wheat and the somewhat unresponsive reaction to additional nitrogen fertilizer applications.
For both wheat and spelt, the influence of the production year on yields is significant, where lowest yields were attained in the year 2008 due to significant rain events and thus a delayed harvest (August 7) in that year. Yields in 2009 were above average; 2010, however, gave average yields of both grain crops.
Ecological footprint
The relatively large area appropriated by the CON and INT systems was mostly attributed to mineral fertilizer and pesticide use, while the smaller area appropriated by the ORG and BD systems was mostly due to machinery use (Table 5). For every hectare of CON wheat and spelt production, an additional 52–100 ha of surface area is impacted. The INT system did not perform any better, although it is publicized and advertised as nature-friendlier (compared to the CON system) and as a sustainable agricultural system15. Control plots appropriated the least area, which was still seven to eight times greater than the surface area used to plant the crops. In this sense there is great need for improvement in the current agricultural practice and the way we understand, till and work the soil. Furthermore, more efficient machinery and machinery use are a must in order to minimize the impact of agricultural production on the environment.
Partial and total ecological footprints of wheat and spelt production for the years 2008–2010 (ha ha−1) for 1 ha of production area.
Mean values of 3 years are presented. Different letters indicate statistically significant differences between production systems at 95% probability (Duncan test).
Overall footprint of a product, SPI and EEP
Results for the a tot, SPI and EEP give an even more insightful picture, as yields are taken into the equation (Table 6). For all three parameters, production systems had a significant influence on the attained results for both wheat and spelt, where control, ORG and BD systems outperformed the CON and INT systems. Production year also significantly influenced the a tot and SPI for wheat production. Moreover, the interaction of production system and year was significant for wheat production. Reasons for these differences between years can be found in lower average yields in the year 2008, where inputs into the systems remained on a similar scale as in the following 2 years.
Overall footprint per unit (a tot), SPI and EEP for wheat and spelt production depending on production system and year.
Means±SEs are presented. Different letters indicate statistically significant differences for each factor and indicator separately at 95% probability (Duncan test). Levels of significance: n.s. – non significant (P>0.05);
* P⩽0.05; **P⩽0.01; ***P⩽0.001.
Ratios between farming systems, where control=1, provide us with a visual overview as to what influence production systems as such and yields have on the performance of farming systems under study (Fig. 1). Results indicate that higher yields in CON and INT systems can partly compensate for their high footprints. However, there still remains a 6:1 ratio for the a tot and SPI between the CON and control system, where EEP does not rise above the ratio 0.2:1 for CON:control, whereas it is 0.7–0.9:1 for the ORG/BD:control systems, respectively.
Ecological footprint, overall footprint per unit (a tot), SPI and EEP ratios (control=1) between production systems for 3 years of wheat (a) and spelt (b) production.
Currently, the ORG:CON farmed land ratios in the EU lie from 1:830 (Malta), to 1:15.4 (Slovenia) and up to 1:6.5 (Austria), with the EU-27 average amounting to 3.9% of the total agricultural area being managed organicallyReference Willer and Kilcher22. Where does that leave us in the future, when we take into account the results from this trial? One of the main objectives to organic farming is that it does not produce enough food to feed the whole population—now and in the futureReference Avery23. However, several research projects and reports have demonstrated otherwiseReference Pretty, Noble, Bossio, Dixon, Hine, Penning de Vries and Morison24, Reference Badgley, Moghtader, Quintero, Zakem, Chappell, Avilés-Vázquez, Samulon and Perfecto25. Even if yields in developed European countries, where CON industrial agriculture is now predominant, would be 5–20% lower with ORG and BD agriculture, population projections for developed countries in the next 50 years partly coincide with these lower yields26. Next to that, a large proportion of the currently produced grain goes toward feeding animals. In Slovenia alone, feedstuffs for animals are produced on more than three-quarters of the arable land21. High competition for grain and animal manures is also present in the developing ‘bio-gas’ sectorReference Oslaj, Mursec and Vindis27. In this sense there is a relatively great reserve in arable land, which could be used for food, instead of feed or energy production. Taking it a step farther from food production levels, what will happen when oil reserves get depleted? It is important to keep in mind that the relation between population and oil extraction is one of cause and effect. With greater use of mineral fertilizers and pesticides, production of which is based on conventional energy sources, higher yields were achieved in the past century and therefore more people could be fed from the same area than before. However, the downside of this advance in agricultural production is also visible in the results of this research—the high proportion of the final footprint going to mineral fertilizer and pesticide production, caused by using those conventional energy sources and therefore needing a large area to offset the high environmental burden. With abundant oil, a large population is possible—ignoring, of course, the fact that environmental degradation may diminish the human population. Without abundant oil, on the other hand, a large population will not be possibleReference Hanlon and McCartney28. Fossil oil renewal rates and quantities are much slower than the current usage rates and quantitiesReference Hanlon and McCartney28–Reference Leder and Shapiro30, and eventually the era of cheap fossil oil will come to an end. With this in mind we projected the magnitude of change, if all land for wheat and spelt production in Slovenia were converted to organic/biodynamic farming in 2050 (Table 7). Production levels would be lower by almost a third, the ecological footprint and a tot, however, would be lower by almost two-thirds (Fig. 2). Consequently, the EEP would rise threefold compared to the current situation. In 2009, around 170,000 tons of wheat were consumed by the Slovenian population, and an additional 100,000 tons were used for animal feed. More than 45% of that wheat had to be imported21. This means that only to nourish people (in order to be self-sufficient) in 2050, twice as much arable land would have to be devoted to wheat production, assuming, of course, the same production levels as with current production techniques. But how can we tackle this issue in the future? Possible solutions can be sought in changing crop rotations and/or land use (as mentioned previously, more than 80,000 ha of land is currently devoted to maize for grain or silage production—mainly for animal feed), but foremost also improving and further developing current alternative agricultural production practices and techniques. In addition, Ewert et al.Reference Ewert, Rounsevell, Reginster, Metzger and Leemans31 argue that production levels of the main crops in Europe will rise in the following decades (owing to improved production techniques and a changed climate), and thus less land will have to be cropped to produce the same amount of food. As mentioned previously, efficient use of machinery and the invention of new forms of working the soil will be of crucial importance. Some good examples pointing toward the future can already be seen in practice. One of them is the Eco-Dyn System, where fuel use per hectare has been lowered to 20–30 litres ha−1 (it amounted to more than 90 litres ha−1 in our study); yields, however, remained stable around the average yields of GermanyReference Wenz32. Another example is the reinvention and improvement of the ridge-till systemReference Turiel33 in order to lower machinery use and improve the quality of soils and consequently the health of plants and quantity of produce. In both cases the farmers are, next to the technological innovations, using a biodynamic approach toward farm and soil management, which has been found to improve soil fertility in several studiesReference Turinek, Grobelnik-Mlakar, Bavec and Bavec34. However, all these improvements and approaches need further research and development in order to adapt them to different microclimatic, pedologic and cultural conditions.
Projected change of total yields, ecological footprint, overall footprint per unit (a tot), and EEP from 2010 (=100%) to 2050 for wheat and spelt production in Slovenia.
Land use for wheat and spelt production in Slovenia in 2010 with the corresponding ecological footprint and projected change with the eventual change of farming practice in 2015 and 2050.
1 Data for area obtained from Ministry of Agriculture, Forestry and Food of Slovenia on the basis of an official email inquiry, based on the number of farms that receive subsidies. Data for yields per hectare are taken from this trial, where organic wheat yields were lower than in previous studies. This is also the reason for the greater reduction in total yields in the 2015 and 2050 projection.
2 According to the Slovene Action Plan for Organic FarmingReference Hrustel-Majcen, Jurcan and Vrečko35), 20% of the utilizable agricultural area is planned to be converted to organic farming in 2015. Moreover, due to the planned change of the Common Agricultural Policy and the Slovene Agri-Environmental Programme after 2013, only integrated and organic (biodynamic) farming is under consideration to be subsidized in the future. The proportion of spelt in the total area is projected to increase with years, as spelt is more resistant to pests and diseases and is relatively undemanding and less susceptible to fluctuations in growing conditions. Also less fertilizer input is demanded.
3 Projection is based on the assumption that by 2050, conventional energy sources (oil, gas, coal and nuclear power) will be in decline and also expensive, thus a shift toward more sustainable, ecologically intensive and less energy-intensive agricultural systems will be a matter of need, rather than choice.
Conclusions
Our results indicate critical points in production of wheat and spelt in each production system, where greatest improvements could be achieved by abandoning mineral fertilizer and pesticide use in industrial farming systems. However, machinery use also needs attention in the near and distant future in all systems studied in order to improve all the recorded parameters and to get closer to sustainable farming systems from a productive and environmental point of view. All mentioned changes will, of course, have to be thought through carefully and support by government policies, economic incentives as well as grassroots activities will aid in making them successful.
The question that stakeholders in agriculture will have to ask in the following years is: Can we save and/or produce enough resources for the current and future generations, when we use or leave an impact on almost 80 ha of surface area to produce 1 ha of wheat (or in a similar size range any other crop)? Or do we have to rethink and above all change the way we farm, live and make decisions in order to survive on planet Earth?
Acknowledgements
The results presented in this paper are an output of the research project J4-9532: ‘The quality of food dependent on the agricultural production method’, funded by the Ministry of Higher Education, Science and Technology of Slovenia.
