Drivers of land use efficiency and trade embodied biomass use of Finland 2000–2010

Ecological Indicators 77 (2017) 348–356

Contents lists available at ScienceDirect

Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind

Drivers of land use efficiency and trade embodied biomass use of Finland 2000–2010 Laura Saikku ∗ , Tuomas J. Mattila Finnish Environment Institute, P.O. Box 140, FI-00251 Helsinki, Finland

a r t i c l e

i n f o

Article history: Received 22 April 2016 Received in revised form 14 February 2017 Accepted 15 February 2017 Keywords: HANPP Environmentally extended input-output analysis IPAT International trade Finland

a b s t r a c t Globally human pressure on the biosphere is increasing, in spite of increases in land use efficiency. The pressure consists of land use and potential degradation. Human appropriation of net primary production (HANPP) is emerging as an indicator, which combines the dual aspects of biomass use and land degradation. Recently HANPP has been used to map the increasing dependence of European countries on biomass imports and the conflicting processes of increased yields and increased consumption. However large overview studies could be complemented with indepth analysis into the causes of changes in individual countries and economic sectors. This allows the analysis of the macroeconomic drivers of change and the responses in sectors to these drivers. In this study we decomposed the HANPP of Finland including imports for the years 2000–2010 using IPAT and applied input-output analysis to look at sectoral land use efficiency in that time period. Finland is a country with intensive biomass trade, and with a very high per capita HANPP. During the study period the sum of domestic and embodied in imports HANPP of the Finnish economy decreased from 76 Mt C/a to 62 Mt C/a (−1% annually on average), while the HANPP related to imports increased from 12 Mt C/a to 14 Mt C/a. The overall trend was that of declining exports and increasing domestic consumption. Of the economic sectors wood harvesting and processing dominated HANPP results, followed by residential construction, animal production and energy supply. In terms of HANPP, most of these decreased, but housing and energy production increased considerably from 2002 to 2010. At the macroeconomic level domestic biomass use per unit of value added decreased (−2.2%/a) as did the amount of HANPP per unit of biomass (−1.1%/a) reflecting increased economic efficiency in land use. In contrast, GDP/capita (+1.3%/a), population (+0.4%) and the share of outsourced HANPP (+0.6%) resulted in increased consumption-based HANPP (from 21 Mt C in 2002 to 27 Mt C in 2010). Our results underline the importance of including international trade and consumption in interpreting overall change in regional HANPP. © 2017 Elsevier Ltd. All rights reserved.

1. Introduction Human pressure on the biosphere is increasing: the global human appropriation of net primary production (HANPP) grew from 6.9 Gt C in 1910–14.8 Gt C by 2005, (0.8% annually on average) (Krausmann et al., 2013). Net primary production (NPP) is a key process for life on Earth and through examining its appropriation the possible impact of land use on the health of the ecosystem can be estimated. HANPP has emerged as a key indicator for quantifying the dual aspects of land use: biomass harvest and potential land degradation. HANPP integrates the appropriation of biomass

∗ Corresponding author. E-mail addresses: laura.saikku@ymparisto.fi (L. Saikku), tuomas.mattila@ymparisto.fi (T.J. Mattila). http://dx.doi.org/10.1016/j.ecolind.2017.02.021 1470-160X/© 2017 Elsevier Ltd. All rights reserved.

harvest or burning (harvested HANPP, HANPPharv) and the productivity changes resulting from conversion of natural ecosystems to managed lands (HANPP related to land use change, HANPPluc). Globally approximately half of the human appropriation of ecosystem productivity is contributed by harvests and the other half by land use-induced productivity changes and human-induced fires (Haberl et al., 2007). Globally, HANPP results are dominated by agriculture (Krausmann et al., 2013), but considerable variation exists among regions. In general, population density, the stage of industrialization and international trade have all been identified as key determinants for HANPP (Krausmann et al., 2012; Krausmann et al., 2013; Teixidó-Figueras et al., 2016). The HANPP of a given region depends on a combination of factors: while a low population density results in low biomass consumption, it may also include harvesting for export purposes. And while higher levels of indus-

L. Saikku, T.J. Mattila / Ecological Indicators 77 (2017) 348–356

trialization generally increase land use efficiency (yield) they also increase trade and possibly consumption (Krausmann et al., 2012). In Europe the general trend is a high level of HANPP, stagnated by the conflicting processes of cropland intensification, reducing agricultural and grazing land and increasing settlement and forest areas, but with considerable differences among countries (Gingrich et al., 2015). Traditionally HANPP has been used to analyze the extent of human impact compared to a natural reference state in a region. A recent development has been the inclusion of embodied HANPP (eHANPP) of imported products (Erb et al., 2009). This allows the analysis of the impacts of international trade between regions, the potential externalization of biomass production and the linkages between production and consumption. Recent eHANPP studies have highlighted the increasing dependence of Europe from outsourced biomass production from other regions (Kastner et al., 2015). Therefore the consumption-based approach, accounting for the total global impacts occurring from economic consumption within a country, is becoming increasingly important also for land use, as it has become for evaluating climate change (Peters and Hertwich, 2008). In an area-specific approach, HANPP serves as an indicator of land-use impacts on a defined area, and the consumption-based embodied HANPP approach allows assessment of impacts related to individual products or the aggregate consumption of countries (Haberl et al., 2014). With increased international trade, embodied HANPP should be a key component in identifying the drivers of increasing HANPP at the country or regional level. Finland represents a sparsely populated industrialized country with a forest dominated landscape (Luke, 2014). Finland both exports and imports considerable amounts of biomass (Sandström et al., 2014). Expressed in land area the biomass imports corresponded to 55% of the domestic land area, amounting to 134,000 km2 in 2010 (Mattila and Saikku, submitted). Domestic HANPP is high, approximately 50% of the potential net primary production (NPP) but has decreased 75 Mt C to 62 Mt C in 1990–2010 (Saikku et al., 2015). In a comparison of 28 EU member states, Finland ranked the highest in production-based per capita HANPP and sixth in respective consumption-based accounting (Kastner et al., 2015). The overview studies have not provided insight on why the Finnish HANPP is so high and have so far not estimated the amount of HANPP embodied in imports. The aim of this study was to apply IPAT-decomposition to the Finnish HANPP dataset for the years 2000–2010, to quantify the changes over time in eHANPP and to identify the key drivers behind the overall change. This study examines the reasons behind the high Finnish HANPP and tests the applicability of IPAT-decomposition analysis and input-output analysis in identifying these drivers. The paper is structured as follows. First we describe the methodologies and their application, especially the estimation of embodied HANPP for imported commodities, the application of IPAT-decomposition and disaggregating the results to sector level. Then we present the results on the embodied HANPP, the decomposition to IPAT factors and the disaggregation to economic sectors. Finally we discuss the implications of this analysis to the development of resource efficiency in Finland, the usefulness of the analysis methods, and the use of HANPP as a resource efficiency indicator.

2. Materials and methods 2.1. Estimating HANPP embodied in imports The methods for calculating HANPP and eHANPP are well established, but in this study we used two modifications to those. The first change was the consideration of reduced NPP in forestry as

349

Table 1 The terminology used in this paper. Term

Abbreviation

Explanation/consists of

Human Appropriation of net primary production

HANPP

Production-based HANPP/domestic HANPP Impact

harvested HANPP (used + unused extraction) + HANPP related to land use change HANPP related to the imported goods in the producing country HANPP related to exported goods including both the HANPP from Finland and from countries from which goods were imported prior to processing and re-export HANPP embodied in imports + domestic HANPP − HANPP embodied in exports HANPP from Finnish industries primary production

I

Affluence Population Consumption

A P C

domestic HANPP + HANPP embodied in imports GDP/capita

Technology Outsourcing

T O

HANPP embodied in imports HANPP embodied in exports

Consumption-based HANPP

used biomass/GDP (=used extraction/GDP) domestic HANPP/used biomass (domestic HANPP + HANPP embodied in imports)/domestic HANPP.

HANPPluc , according to Saikku et al. (2015). The previously calculated dataset with the forestry induced HANPP was used as such, and the details of the methodology and its changes compared to the standard method are described in Saikku et al. (2015). The changes in quantifying the import embodied HANPP are described in this chapter. The main terminology used in this paper is described in Table 1. The commonly used approach in HANPP studies is to estimate the embodied HANPP through bilateral trade and biophysical data (Kastner et al., 2015). A problem with this approach is that highly processed goods and services are not included (Hubacek and Feng, 2016). The second change to the standard methodology was to use life cycle assessment to convert material flows into primary biomass flows for the imports of processed goods. The first stage was to link the import commodity flows to available life cycle inventories (see Supplementary information SI for the raw material equivalents of traded biomass and HANPPluc). Ecoinvent 2.2. and 3.0 databases (Ecoinvent, 2015) were used to find 240 life cycle inventory datasets, which corresponded to the imported commodities. The dataset was complemented through individual LCA studies on animal products (Williams et al., 2006), fish (Ytrestøyl et al., 2011; Nielsen et al., 2003) and beverages (Coltro et al., 2006; Ntiamoah and Afrane, 2008; Amienyo et al., 2013; Brewers of Europe, 2002; Gazulla et al., 2010). The original data on Finnish imports was in the combined nomenclature (CN) classification. For analysis it was then aggregated to the Classification of Product by Activity (CPA) level and then to a customized 180 product aggregated resolution (ETTL) used in previous studies (Koskela et al., 2011). A correspondence table was used to link the LCI inventories to the products of the import statistics. Used and unused flow of imports was compiled based on this data. For the harvested HANPP, used and unused extraction was calculated. The above mentioned life cycle inventories included the resources extracted during the life cycle of the product. For the purposes of this study, only biotic raw materials were included. This covered wood, crops and fibres, which were aggregated into a total biomass used extraction figure. Unused extraction was added to

350

L. Saikku, T.J. Mattila / Ecological Indicators 77 (2017) 348–356

these through expansion factors. For services and products which could not be linked to specific studies, the European-wide used and unused resource extraction factors were used (Eurostat, 2014). Finally HANPPluc was added using factors provided by (Alvarenga et al., 2015) to the used and unused extraction to give a dataset of embodied HANPP for each of the imported commodity. As the factors of Alvarenga et al. were expressed per land area, they had to be converted into per mass basis. As a first step the factors were converted into per monetary unit basis by multiplying them with the direct land use requirement of each industry in the EXIOBASE inputoutput model (E in Eq. (1)). As this provided a list of factors for each product from each country, the data was averaged to Finnish import mix by using the shares of import from different country for each commodity (obtained for the year 2010 from the EXIOBASE model economic flow table). Finally the factors were multiplied by the price of the commodity in 2010 in Finland using customs statistics data (Customs, 2016). The calculated and averaged HANPPluc factors and the unused and used extraction coefficients are available as supporting information. A full trade analysis would include the embodied HANPP in both imports and exports. Adding the imports to the domestic HANPP and substracting the HANPP embodied in exports would then give a consumption-based HANPP account (Erb et al., 2009; Haberl et al., 2009). For the purposes of this study, we calculated a full consumption-based HANPP analysis only for the years 2002 and 2010. A detailed input-output model was available for those years in order to quantify the HANPP embodied in exports throughout their life cycle (Section 2.2). For the remaining years from 2000 to 2010 we calculated only the domestic and import embodied HANPP. This represented the impact of all economic activities in Finland, including production for both domestic and export markets. 2.2. Quantifying exports and consumption-based HANPP with input-output analysis Imported HANPP was quantified with a combination of trade statistics and LCA studies, but the impact of exports on the overall HANPP balance was calculated through input-output analysis. Environmentally extended input-output (EEIO) analysis describes the economy and it’s interactions with the environment in a single equation (Wood and Lenzen, 2009): q = E(I − A)−1 y,

(1)

Where q = HANPP caused by y [t] E = direct HANPP intensity vector [t/D ] I = identity matrix A = input-coefficient matrix [D /D ] y = economic final demand vector [D ] In order to account for the imports, the model matrices were constructed in a two-region setting. The domestic flows were quantified through input-output datasets for 2002 and 2010 (146 sector-by-sector model), while the imports were quantified through trade statistics for the inputs (A) and life cycle assessment for the intensities (E). This kind of integrated hybrid IO-LCA analysis (Suh and Huppes, 2005) has been applied in EEIO (Koskela et al., 2011). While the LCA-based inventory is subject to cut-off errors (Suh and Huppes, 2005), it also has the benefit of being highly disaggregated and allowing the matching of products within the broader classification to the corresponding LCA inventory data (Koskela et al., 2011). Eq. (1) allowed the calculation of HANPP of different final demand categories (total final demand, domestic consumption, investments and exports). When total final demand was used, the results included domestic HANPP and all of the imported HANPP. With domestic consumption and investments, the results repre-

sented a consumption-based HANPP account. The input-output model allowed the disaggregation into exports, investments and consumption for the years 2002 and 2010. In addition to trade analysis, the input-output model was used to identify the key sectors in the economy in terms of HANPP and value added. The identification was based on the “total flow analysis” (Szyrmer, 1992; Koskela et al., 2013; Wood and Lenzen, 2009), which quantifies the total impact an industry has on the economy directly and through its supply chain. The method has been used earlier to identify sectors with a high influence to climate change and raw material requirements (Koskela et al., 2013), but this was the first application to HANPP. Comparison of total flow based HANPP and value added describes the economic “HANPP efficiency” of a sector and its supply chain, i.e. it’s capacity to generate value added at the cost of increasing NPP appropriation. 2.3. Decomposing drivers of HANPP In theory, the input-output dataset could be used to decompose the change into components (i.e. changes in HANPP intensity, technological structure and final demand) (Mattila, 2012; Dietzenbacher and Los, 1998). However, in during or study period the sector classification changed twice and no price indexes were available for the detailed sector classification. A workaround would have been to aggregate the dataset to a level, where price indices would be available, however, losing sectoral resolution in the process. In addition, the input-output based decomposition would have been applicable for only the two years 2002 and 2010. Because of limitations in the input-ouput data, the macro-level drivers of change related to the development of environmental change were analyzed with the help of the ImPACT identity (Waggoner and Ausubel 2002). ImPACT is a straightforward and transparent method used for describing the historical development and predicting the effects of changes in population, affluence (GDP/capita), technology (environmental impact/use) and the intensity of consumption (use/GDP) on environmental impact. As ImPACT is a mathematical identity, any number of factors can technically be added to the equation. Here we assessed the historical development in HANPP following the basic ImPACT factors, but included also international trade, according to the equation: I = P × A × C × T × O, where I is the impact as domestic HANPP summed up with HANPP embodied in imports, P denotes the size of the population, A is affluence as GDP/capita, C denotes consumption as used biomass/GDP, T is technology as domestic HANPP/used biomass and O is outsourcing as the ratio of domestic HANPP and HANPP embodied in imports per domestic HANPP (see also Table 1). We analyzed the annual changes in the drivers and impact I, represented by respective lowercase letters. The annual percentage changes in the drivers contribute to the change in impact I. Growth rates were calculated using Eq. (2), ␣i,j er = ˛i+1,j

(2)

transforming to Eq. (3), r = ln(˛i+1,j ) − ln(˛i,j ),

(3)

where ␣ represents the value of the ImPACT variable in one year, i represents the year, e the natural logarithm, j the variables p, a, c and t, and r the annual growth rate. The sum of the growth rates of j equals the change in the impact, and each component’s share of the total can be determined. Population data and Gross domestic product was obtained from Statistics Finland (2015). Data on total HANPP and harvested biomass (harvested HANPP) was taken from Saikku et al. (2015).

L. Saikku, T.J. Mattila / Ecological Indicators 77 (2017) 348–356

351

Fig. 1. Embodied HANPP in Finnish imports, in carbon (Mt). Domestic HANPP, in carbon (Mt).

Fig. 2. Embodied HANPP related to Finland’s international trade in 2002 and 2010. Consumption-based HANPP = HANPP embodied in imports + consumption of domestic HANPP + HANPP in investments. Production-based HANPP (=domestic HANPP) = consumption of domestic HANPP + HANPP in investments + HANPP related to exports.

3. Results 3.1. HANPP embodied in international trade A clear increasing trend in the HANPP embodied in Finnish imports was observed during 2000–2008, when embodied HANPP increased from 11.6 Mt C/year in 2000 to 17.4 Mt C/year in 2008 (Fig. 1). In 2009, wood imports from Russia diminished remarkably, resulting in a significant reduction of HANPP embodied in imports. Embodied HANPP in imports was around 14.1 Mt C/year in 2010. The HANPP embodied in imports consisted mainly of harvested HANPP, and varied between 10.6 Mt C and 15.9 Mt C, in 2000–2010.The HANPP embodied in imports due to land use conversion (HANPPluc ) increased steadily from 0.9 Mt C to 1.6 Mt C, respectively, contributing 8–12% to total imported HANPP. Detailed examination at the product level for 2002 and 2010 showed that different wood products, such as pulpwood, pulp, paper, paperboard and forest chips, played a large role in the harvested HANPP embodied in imports. In 2002, these covered 68% and in 2010 58% of the harvested embodied HANPP, amounting to 8.6 Mt C/year and 7.3 Mt C/year in 2002 and 2010, respectively. These were followed by dairy products (1% and 5%), meat products (4% and 4%), vegetable oils and fats (0% and 3%), grain mill products (3% and 3%) and berries and fruits (2% and 3%) in 2002 and 2010, respectively. Regarding HANPPluc , two product groups covered 49% of the total:

vegetable oils and fats and processed and preserved fruit and vegetables amounted to 0.42 Mt C/year and 0.30 Mt C/year in 2010, respectively. The calculation based on the input-output model show how HANPP is attributed into domestic consumption, investments and exports (Fig. 2). Consumption-based HANPP, or the sum of consumption and investments, increased by 29% from 2002 to 2010, while the production-based HANPP decreased. Consumptionbased HANPP was at much lower level than the production-based, respectively, indicating that Finland exported more embodied HANPP than it imported. As the amount of exports decreased faster than imports, the consumption-based HANPP increased. Of the consumption-based HANPP, imports for domestic consumption and investments remained at a steady level, while the consumption of domestic HANPP increased by 29%. The largest increase in domestic HANPP for domestic consumption was in the housing sector, indicating increased use of wood for construction and heating. 3.2. Drivers of HANPP The sum of domestic and HANPP embodied in imported goods (I) decreased from 84.5 Mt C/year in 2000 to 76.3 Mt C/year in 2010 (−1.0% annual average decrease rate) (Fig. 3, Table 2). When looking at the drivers of this studied impact (I), the economic

352

L. Saikku, T.J. Mattila / Ecological Indicators 77 (2017) 348–356

Fig. 3. Effect of population (p), affluence (a), economic intensity of biomass use (c), domestic HANPP per used biomass (t) and outsourcing (o) on the development of Impact (i) in Finland in 1990–2010.

Table 2 The average annual percentage changes in domestic HANPP and HANPP embodied in imports, and related macro-level drivers in 2000–2010 in Finland.

2000–2005 2005–2010 2000–2010

i

p

a

c

t

o

−0.2% −1.8% −1.0%

0.3% 0.5% 0.4%

2.3% 0.4% 1.3%

−3.0% −1.4% −2.2%

−1.3% −0.9% −1.1%

1.5% −0.3% 0.6%

i = Changes in domestic HANPP + HANPP embodied in imports, p = changes in population, a = changes in affluence, c = changes in economic intensity of biomass use, t = changes in domestic HANPP per used biomass, o = changes in outsourcing.

intensity of used biomass (C) decreased by 2.2% annually on average, and the technology as domestic HANPP needed to produce the used biomass (T) also decreased by around 1.1% annually on average. In absolute terms, biomass used per every GDP produced 0.114 kg/D , and HANPP generated per every unit of biomass used amounted to 3.4 t C/t C of in 2010, respectively. GDP/capita (A) increased by around 1.3% annually on average, while the population (P) increased slightly. The ratio of domestic HANPP and HANPP embodied in imports to the domestic HANPP, outsourcing (O), increased by around 0.6% annually on average. Thus, appropriation of net primary production was increasingly outsourced abroad in relative terms. The average annual improvement observed in impact (I) during the ten-year period was largely affected by a decrease in 2009. Between 2008 and 2009, impact decreased due to decrease in affluence (A), decrease in domestic biomass use per GDP (C) and also due to observed negative value for outsourcing (O). The changes in the used biomass directly influenced technology (T) driver: most of the domestic HANPP is due to HANPPluc , which changes slowly. Therefore, an annual decrease in used biomass will instantly lead to an increase in T.

3.3. Domestic and embodied HANPP disaggregated to economic sectors The intensity of HANPP per GDP, improved from 2002 to 2010, but different sectors presented very different trajectories (Fig. 4). These were analyzed through the total flow analysis (Section 2.2), which presents the overall embodied HANPP of each industry, both from within industry and from within the industry’s supply chain. It should be noted, that since the industries are strongly inter-

linked, our results include double-counting and are not additive (e.g. the HANPP of silviculture is also included in sawmilling, pulp and paper manufacture and construction). However, our results demonstrate the relative influence of an industry on the whole economy level HANPP and GDP (Fig. 4, Table 3). Comparing the total flow HANPP between 2002 and 2010, HANPP decreased in the silviculture, sawmilling and pulp manufacturing sectors, but increased in the construction and housing sectors. Most of the HANPP and GDP was generated by a few key sectors. Many sectors which had a large influence on GDP, had a very low impact on HANPP.

4. Discussion 4.1. Drivers in domestic HANPP and HANPP embodied in imports Finnish domestic HANPP was 11.6 t C/capita/year in 2010. Internationally the level is high. In the Western industrial region, HANPP per capita remained more or less stable since 1980 and was 3.5 t C/capita/year in 2005 (Krausmann et al., 2013). It must be noted that 50% of Finnish HANPP was estimated to be due to HANPPluc in forestry, a parameter that has been disregarded in previous studies (Saikku et al., 2015). In Saikku et al. (2015) it was found that since natural forests have more standing stock biomass than commercial forests, the cycling of root and leaf biomass is much larger, which more than offsets the slower growth rate of natural forests. Therefore the commercial management of forests was found to reduce NPP, even though it increased the amount of harvestable wood products. Even if the HANPPluc of forestry is not considered, some studies show that Finland still ranks number one in production-based per capita HANPP among the EU countries (Kastner et al., 2015). Our results support these earlier findings,

L. Saikku, T.J. Mattila / Ecological Indicators 77 (2017) 348–356

353

Fig. 4. Influence of different sectors to the whole economy HANPP (domestic and embodied in imports) and gross domestic product (GDP) analyzed with total flow analysis. Open shapes represent sectors in 2002 and dark shapes represent sectors in 2010. The dashed line is the average HANPP intensity of sectors in 2002 (0.66 (t/D /year), the straight line is the average intensity in 2010 (0.47 t/D /year) respectively. The arrows describe the change from 2002 to 2010 in the five most important sectors regarding HANPP in 2010. The ten most important sectors are marked with triangular shapes.

Table 3 Sectors with the largest total flow of HANPP in the Finnish economy Mkg/a (Intensity of HANPP per unit of GDP in parenthesis t/D a).

Silviculture and forestry Sawmilling of wood Manufacture of pulp, paper and paperboard Construction of residential and non-residential buildings etc. Letting and operation of dwellings Manufacture of products of wood Growing of crops Animal production (excluding reindeer and fur animals) Electric power generation, transmission and distribution Manufacture of dairy products

however, the HANPP was decreasing 2000–2010 when measured with production-based methods or when including the HANPP embodied in imports. Based on the ImPACT analysis, the annual decrease of 1% in domestic and imported HANPP was mostly due to reduced biomass consumption of the Finnish economy as well as due to improved efficiency related to the HANPP generated per biomass used. Increases of GDP had the opposite effect, increasing also Impact. Moreover, the results of this study showed that a decrease in domestic HANPP occured simultaneously with an increase of imported embodied HANPP, as Finland was outsourcing biomass production during the time period. The decrease of in domestic and imported HANPP over the ten-year period was largely affected by a decrease in several drivers in 2009. Between 2008 and 2009, the economic downturn reduced GDP/capita. Due to economic conditions, exports also diminished greatly, as shown in the outsourced HANPP variable. Also domestic biomass was used less. On the other hand, during the period of stable economic growth in 2000–2005, barely any absolute decoupling between the appropriation of net primary production and GDP could be observed.

2010

2002

Average annual change

53 771 (26) 20 696 (9) 17 679 (2) 9904 (0.6) 9441 (0.5) 6184 (3) 5050 (8) 4982 (3) 2696 (0.5) 2600 (1.5)

59 632 (30) 31 943 (13) 25 802 (3) 6679 (0.6) 1734 (0.1) 4742 (4) 6380 (11) 4972 (3) 1285 (0.4) 2385 (2)

−1% −5% −5% 5% 24% 3% −3% 0% 10% 1%

4.1.1. Reduced biomass use and increasing efficiency reduced HANPP During 2000–2010, GDP was increasingly produced at sectors other than those relying on the large-scale use of biomass. Although GDP increased, the use of biomass did not significantly increase. This was demonstrated by the gradually decreasing share of forest industries of the overall GDP. In 2002, the forest sector (silviculture and forestry, sawmilling, pulp and paper production and manufactured wood products) accounted for 6.8% of GDP and in 2010 the share came down to 4.0%. In Finland, the extraction of 1 ton of biomass resulted in a domestic HANPP of 3.5 tons in 2010. Globally, land use was more efficient: 1.6 tons of HANPP was appropriated per 1 ton of biomass used on average in 2005 (Krausmann et al., 2013). HANPP per unit area is larger for agriculture than forestry (339 vs. 215 g C/m2 /year), however, the area for forestry is very large in Finland, explaining most of the trends. Regarding forests, from 2000 to 2010 the domestic HANPP per used forest biomass improved from 4.3 to 3.6 t C/t C. This was due to increased forest growth, which approached the natural potential NPP implying reduced HANPPluc (Saikku et al., 2015).

354

L. Saikku, T.J. Mattila / Ecological Indicators 77 (2017) 348–356

In previous studies, HANPP due to land use change in forestry has been ignored; its inclusion in the used data source here explains also the large HANPP in forestry in Finland. Regarding agriculture, similar trends have not been observed, as the yields have stagnated. Finland also has low yields compared to the rest of Europe due to its shorter growing season. The sectors responsible for the increased efficiency were silviculture, sawmilling and pulp manufacturing sectors. These sectors had large reductions in both absolute HANPP and impact per GDP (Fig. 4). Export industries improved their efficiency more than sectors serving domestic consumption and investments (housing, residential construction) (Table 2). HANPP increased in the housing sector, but mainly due to factors related to calculation: changes occurred in the classification of the input-output tables used from 2002 to 2010. In 2002, firewood was considered to be part of household consumption, while in 2010 it was included in housing. This accounted for 60% of the HANPP at the housing sector in 2010. However, as the HANPP of housing increased by a factor of 5 from 2002 to 2010, housing was also actually using more biomass. For the construction sector, HANPP increased mainly due to a replacement of concrete and steel with wood structures. HANPP associated with wood product use in construction increased by 60% from 2002 to 2010. At the same time, HANPP associated with concrete and sand use (HANPPluc from excavation areas and biomass used for energy) decreased by 52%, while HANPP associated with steel use decreased by 34%. If Finland would seek to reduce its internationally very high HANPP level, the trends could be continued further. Much of the positive development was caused by the increased growth of forests, gradually approaching the potential NPP of natural forests, therefore decreasing the HANPPluc (see Saikku et al., 2015). Actions which would increase forest growth (e.g. fertilization, tree species selection, silviculture) would decrease the overall HANPP effectively. Also, moving the economy towards sectors with a lower HANPP/GDP ratio would increase economic growth and reduce HANPP. As most of the HANPP relates to wood products, this would imply for instance a shift from raw wood, construction materials and paper to higher value added products of bioeconomy. Currently, Finnish bioeconomy development relies on extraction of domestic biomass, whereas e.g. Dutch bioeconomy tends to be more product-oriented (Hildén et al., 2016). 4.1.2. Affluence and population as upward drivers of HANPP HANPP development in relation to population growth has resulted in contradictory results (Krausmann et al., 2009; Steinberger et al., 2010; Krausmann et al., 2013): in Finland, HANPP (domestic + embodied in imports) did not increase in alignment with the modest population growth; however, harvested HANPP remained fairly stable in relation to population. Population growth has been very modest in Finland. Food producing is directly linked to population, but in Finland agriculture plays a minor role (only 11% of domestic HANPP in 2010) compared to most countries. Furthermore, the forestry and wood processing sectors are predominantly export oriented in Finland; thus GDP is a much more powerful driver for HANPP than population. Some earlier global studies have estimated that sparsely populated countries have a low level of per capita HANPP (also when measured as a share of potential NPP), whereas per capita HANPP is high in countries with a high population density (Krausmann et al., 2009). However, this is not true regarding Finland where population density is relatively low, 15.9 inhabitants/km2 in 2010. Kastner et al. (2015) also showed that in the European Union, the sparsely populated regions, such as Finland, Lithuania, Estonia, Latvia, Sweden and Ireland, have the highest per capita levels of consumption-based and production-based HANPP. Productionbased HANPP in these countries is high especially due to high level

of exports. The Finnish biomass-based forest industries are heavily export-oriented, and production-based HANPP is at much higher level compared to the consumption-based estimates calculated here. 4.2. Consumption-based HANPP increased while production-based HANPP decreased In Finland, besides production-based HANPP, also the consumption-based HANPP was found to be high. Furthermore, appropriation of net primary production was increasingly outsourced abroad and consumption-based HANPP increased. The increase in consumption-based HANPP from 2002 to 2010 can be explained through decreases in exports. The strong forest sector maintains a certain annual level of harvested forest biomass. As the exports started decreasing in 2008–2009, the harvested forest biomass was increasingly used for domestic consumption. Furthermore, the increased use of wood for energy production and construction has also been a national strategy for economic development. Increased consumption of wood biomass in Finland reduces potential for export production and also the supply of biomass to countries with less forest. The main consumption categories for HANPP were the same as for carbon footprint studies: housing contributed to 34% of the HANPP and food to 22%. Transportation, which is often high in household carbon footprint studies, had only a small impact on HANPP (4%), as the share of biofuels is minor. On the other hand, a quite large share of public services (15–23%) in the consumptionbased account may be untypical compared to other countries. Finland has a large share of public consumption of education, social services and health care arranged through the government. These services need biomass in their life cycle through construction and energy supply. The large role of public consumption is also a good leverage point for improving HANPP efficiency. As large purchasers, communities, cities and the government have influence on the manufacturing industries. Increased demand for goods with a lower “HANPP footprint” would create incentives for companies to improve efficiencies. 4.3. Discussion on the applied methods The ImPACT method, or IPAT in its basic form, identifies the key drivers of environmental impact. When applied to data for multiple time periods, it reveals pathways of changes in several macro-level forces as shown in this study. With the help of ImPACT, the goal of reducing environmental impact through improved technology and increasing dematerialization can be described (Waggoner and Ausubel, 2002). However, IPAT cannot be used to identify the sources of the impact in the society, and consequently the contribution of different sectors of the economy was conducted with the help of input-output analysis. Previous IPAT analyses have focused on domestic impacts, excluding the impact embodied in imports. Drivers of environmental pressures, with the focus on the embodied impacts of trade, are examined and discussed in detail in a recently published study (Teixidó-Figueras et al., 2016) using regression-based decomposition. Teixidó-Figueras et al. (2016) studied the drivers of HANPP between several countries with various different characteristics. We studied the change drivers in HANPP inside one country. The drivers of the study by Teixidó-Figueras et al. (2016) (income and active population share, urbanization and population density and climate and bioproductivity of land (potential NPP)), remained more or less unchanged in Finland during the studied time period, apart from changes in income. Krausmann et al. (2013) found that absolute values of HANPP per capita reflect the amount and mix of biomass products consumed per capita, which generally increase

L. Saikku, T.J. Mattila / Ecological Indicators 77 (2017) 348–356

with income and consumption of more HANPP-intensive products; along with the efficiency of biomass production relative to potential NPP; and with net biomass imports or exports. In our study, these variables are more or less presented. Most of all, we studied factors that may change between proximate years inside one country, i.e. HANPP efficiency of biomass production and the intensity of economic biomass use. In addition to including embodied impacts in IPAT, resource use was successfully examined for the first time using this decomposition method in our study. The Finnish HANPP per capita is high compared to other countries, both in production- and consumption-based analyses (Kastner et al., 2015). However, previous studies on consumptionbased HANPP have been based on specific methods for converting secondary products to primary products in order to estimate the material requirement of imports and exports (Kastner et al., 2015). The problem in these methods is often related to cut-off (Lenzen, 2001), where only a part of the supply chain is included in the assessment of embodied HANPP. Based on the previous assessment, the consumption-based HANPP was approximately 9 t dm/capita while the production-based HANPP was 15 t dm/capita (Kastner et al., 2015). Therefore the consumption-based estimate was 60% of the production-based one. In our findings the 2010 dataset gave similar results, with consumption-based HANPP being 53% of production, but in 2002 it was only 40%. This would indicate that the input-output-based methodology gives a higher estimate for the embodied HANPP in processed final products, such as paper, or energy-intensive products, which use wood in energy production in the supply chain. Input-output analysis includes such tertiary users of biomass better than process-based methods and therefore gives a more complete picture of the consumption-based HANPP. On the other hand, as input-output analysis is monetary-based, it allocates impacts based on their monetary value. In Finland the price for wood is different for different uses, with sawmilling logs having a price several times higher than the small diameter wood used for pulp manufacture. Therefore input-output analysis allocates a much larger share of forestry related HANPP to products, which use high value wood. In a material flow-based allocation, each unit of wood would have the same HANPP intensity. Both the consumption and production-based estimates of HANPP were much higher than those in the previous assessment (Kastner et al., 2015). This is related to the HANPPluc calculated for forests in the dataset used for this study: around 50% of Finnish HANPP was estimated to originate due to reduced NPP in managed forests compared to a natural reference state (Saikku et al., 2015). Finland was already a top ranking country in the 2007 comparison of EU countries’ HANPP intensity (Kastner et al., 2015). Including the HANPPluc increased the impact further, making Finland a hotspot for reducing HANPP at the European level. The results found for Finland are applicable to other countries with a high level of forest industry and export of forest derived products. 4.4. HANPP as a resource efficiency indicator Human appropriation of net primary production is one useful measure to estimate human impact related to land use (Haberl et al., 2007; Krausmann et al., 2013). HANPP measures resource use in terms of biomass carbon of plant growth. It measures how much of the net primary production (NPP) available per year for ecosystem processes is appropriated by humans, and how much of the trophic energy would be available for ecosystems in the absence of human influence. As a resource use indicator, in some regard HANPP is a similar indicator to material flow indicators such as direct material input (DMI) and raw material requirements (RMR). However, it focuses only on biomass, and it also attempts to include the virtual land consumption through NPP lost due to land use change. In that regard

355

HANPP gives a different focus on the national efficiency assessment. The estimation of NPP lost through land use change is highly sensitive to assumptions on the reference state selected to represent the natural state (Saikku et al., 2015). In life cycle assessment, the selection of a reference state has been increasingly acknowledged for land use and climate change impacts (Soimakallio et al., 2015). In this way, the HANPP measure is also analogous to other impact indicators with a reference system, e.g., on biodiversity (De Baan et al., 2013). Discussion between HANPP and LCA communities on the reference states could result in a more robust assessment. For example the potential natural NPP can be questioned as a reference baseline due to the fact that there may be multiple equilibria for the final state (Taelman et al., 2016). Furthermore, for transformation and permanent impacts, it remains unclear how long it takes to recover from land use change, which can be a highly dynamic process, and whether the land ever reaches the initial natural state (Koellner et al., 2013). Some authors suggest calculating the environmental impacts of any human intervention as the difference of the current situation and the reference state where the human intervention does not exist (Soimakallio et al., 2015). Others suggest comparing it to a dynamic relaxation state, where the ecosystem is allowed to recover and the annual potential recovery rate is used as the reference (Milà i Canals et al., 2007). In the case of Finland, the choice of a reference state had an influence of 50% of the overall HANPP, highlighting the importance of deciding on a robust reference state. Despite being a resource indicator, HANPP relates to other environmental dimensions as well. Human appropriation of net primary production reduces the amount of energy available to other species, influences biodiversity, water flows, carbon flows, energy flows within food webs and the provision of ecosystem services (Haberl et al., 2007). HANPP (above-ground processes) has been shown to inversely correlate with biodiversity, at least at levels of above 50% where HANPP has negative consequences in terms of biodiversity, as predicted by the species-energy hypothesis (Haberl et al., 2004). The suitability of NPP and HANPP as an ecological indicator has largely been discussed in earlier studies, such as Taelman et al. (2016). In the case of Finland, HANPP indicator could track several macro-level changes in land use, such as decreasing imports of biomass, increasing NPP levels in forests and an increasing dependence on domestic biomass for consumption, as well as the biotic resource efficiency of various economic sectors. HANPP indicator thus manages to integrate many different impacts into a single metric. 4.5. Concluding remarks In Finland, human appropriation of net primary production including domestic effects and impacts embodied in imports was at a relatively high level in 2000–2010. High HANPP indicates high resource consumption and relates to several environmental damages. Globally, HANPP has increased annually on average 0.8% in 1910–2005 (Krausmann et al., 2012). It would be essential to turn this trend globally. In Finland, The sum of domestic and HANPP embodied in imported goods prominently recently decreased: around −1.0% annually on average in 2000–2010. Regarding technology, HANPP per biomass produced decreased as actual forest NPP approached potential NPP levels. Regarding consumption, GDP was increasingly produced at sectors other than those relying on the large-scale use of biomass. However, the average annual improvement observed in total HANPP during the ten-year period was largely affected by an economic downturn in 2009: affluence, consumption and outsourcing heavily acted as downward forces for total HANPP between 2008 and 2009. If analyzing domestic production-based emissions only, the important implications of international trade are disregarded. Production-based HANPP in

356

L. Saikku, T.J. Mattila / Ecological Indicators 77 (2017) 348–356

Finland is high as Finland exports vast amount of biomass produced. Even though consumption-based HANPP remained at much lower level, our analysis surprisingly showed that the consumptionbased estimate for HANPP increased. Economic sectors related to wood consumption contribute the most to total production-based HANPP: both in absolute terms and also measured by economic intensity, followed by dairy products and the farming of animals. The largest contributors to the consumption-based HANPP were housing and food. Changes in the structure of the economy are slow and difficult to implement. Efficiency might improve more easily if forest NPP continues to grow and harvested biomass extraction is utilized more efficiently. As a resource use indicator, HANPP measures material flows in terms of biomass, but also attempts to include the virtual land consumption through NPP lost due to land use change, thus describing the environmentally impacts more comprehensively compared to traditional resource efficiency assessments. Acknowledgements Funding: This work was supported by the Academy of Finland project: Sustainable Use of Natural Resources and the Finnish Economy (SURE) [decision number 259555]. Ilmo Mäenpää from Oulu University provided data on biomass-related extraction for imports. Sampo Soimakallio is acknowledged for commenting on the paper. Two anonymous reviewers are greatly acknowledged for their helpful comments on the manuscript. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ecolind.2017. 02.021. References Alvarenga, R., Erb, K.-H., Haberl, H., Soares, S., van Zelm, R., Dewulf, J., 2015. Global land use impacts on biomass production—a spatial-differentiated resource-related life cycle impact assessment method. Int. J. Life Cycle Assess. 20 (4), 440–450. Amienyo, D., Gujba, H., Stichnothe, H., Azapagic, A., 2013. Life cycle environmental impacts of carbonated soft drinks. Int. J. Life Cycle Assess. 18, 77–92. Brewers of Europe, 2002. Guidance Note for Establishing BAT in the Brewing Industry Guidance Note for Establishing BAT in the Brewing Industry. CBMC (916-09-0 p.78.). Coltro, L., Mourad, A.L., Oliveira, P., Baddini, J., Kletecke, R.M., 2006. Environmental profile of brazilian green coffee. Int. J. LCA 11 (1), 16–21. Customs, 2016. Foreign Trade Statistics Databases (ULJAS-database) for Finland. www.uljas.tulli.fi. De Baan, L., et al., 2013. Land use impacts on biodiversity in LCA: a global approach. Int. J. LCA 18, 1216–1230. Ecoinvent, 2015. http://www.ecoinvent.ch (Retrieved 20.1.2015). Erb, K.-H.-, Krausmann, F., Lucht, W., Haberl, H., 2009. Embodied HANPP: Mapping the spatial disconnect between biomass production and consumption. Ecol. Econ. 69 (2), 328–334. Eurostat, 2014. Material flow accounts – flows in raw material equivalents. http:// ec.europa. eu/eurostat/statistics-explained/index.php/Material flow accounts - flows in raw material equivalents. Gazulla, C., Raugei, M., Fullana-i-Palmer, P., 2010. Taking a life cycle look at crianza wine production in Spain: where are the bottlenecks? Int. J. Life Cycle Assess. 15, 330–337. Gingrich, S., Niedertscheider, M., Kastner, T., Haberl, H., Cosor, G., et al., 2015. Exploring long-term trends in land use change and aboveground human appropriation of net primary production in nine European countries. Land Use Policy 47, 426–438 http://www.sciencedirect.com/science/article/pii/ S0264837715001374. Haberl, H., Shultz, N., Plutzar, C., Erb, K.H., Krausmann, F., et al., 2004. Human appropriation of net primary production and species diversity in agricultural landscapes. Agric. Ecosyst. Environ. 102, 213–218. Haberl, H., Erb, H., Krausmann, F., Gaube, V., Bondeau, A., et al., 2007. Quantifying and mapping the human appropriation of net primary production in earth’s terrestrial ecosystems. PNAS 104 (31), 12942–12947. Haberl, H., Erb, K.-H., Krausmann, F., Berecz, S., Ludwiczek, N., et al., 2009. Using embodied HANPP to analyze teleconnections in the global land system: conceptual considerations. Geografisk Tidsskrift-Danish J. Geogr. 109, 119–130.

Haberl, H., Erb, K.-H., Krausmann, F., 2014. Human appropriation of net primary production: patterns, trends, and planetary boundaries. Annu. Rev. Environ. Resour. 39, 363–391. Hildén, M., Bosman, R., Peck, P., 2016. Learning from country comparisons in renewing manufacturing towards a circular bioeconomy. TEKES Policy Brief: Renew. Manuf. 4. Hubacek, K., Feng, K., 2016. Comparing apples and oranges: some confusion about using and interpreting physical trade matrices versus multi-regional input-output analysis. Land Use Policy 50, 194–201, http://dx.doi.org/10.1016/ j.landusepol.2015.09.022. Kastner, T., Erb, K.-H., Haberl, H., 2015. Global human appropriation of net primary production for biomass consumption in the european union, 1986–2007. J. Ind. Ecol. 19 (5), 825–836 http://onlinelibrary.wiley.com/doi/10.1111/jiec.12238/ full. Koellner, T., de Baan, L., Beck, T., Brandão, M., Civit, B., et al., 2013. UNEP-SETAC guideline on global land use impact assessment on biodiversity and ecosystem services in LCA. Int. J. Life Cycle Assess. 18 (6), 1188–1202. Koskela, S., Mäenpää, I., Seppälä, J., Mattila, T., Korhonen, M.-R., 2011. EE-IO modeling of the environmental impacts of Finnish imports using different data sources. Ecol. Econ. 70 (12), 2341–2349. Koskela, S., Mattila, T., Antikainen, R., Mäenpää, I., 2013. Identifying key sectors and measures for a transition towards a low resource economy. Resources 2, 151–166, http://dx.doi.org/10.3390/resources2030151. Krausmann, F., Haberl, H., Erb, K.-H., Wiesinger, M., Gaube, V., et al., 2009. What determines geographical patterns of the global human appropriation of net primary production? J. Land Sci. 4 (1–2), 15–33. Krausmann, F., Gingrich, S., Haberl, H., Erba, K.-H., Musel, A., et al., 2012. Long-term trajectories of the human appropriation of net primary production: lessons from six national case studies. Ecol. Econ. 77, 129–138. Krausmann, F., Erb, K.-H., Gingrich, S., Haberl, H., Bondeau, A., et al., 2013. Global human appropriation of net primary production doubled in the 20th century. PNAS 10 (25), 10324–10329. Lenzen, M., 2001. Errors in conventional and input-output based life-cycle inventories. J. Ind. Ecol. 4, 127–148. Luke, 2014. Statistical Yearbook of Forestry 2014. Official Statistics of Finland. Finnish forest research institute www.metla.fi/metinfo/tilasto/julkaisut/vsk/2014/index.html. Mattila, T. Any sustainable decoupling in the Finnish economy? A comparison of the pathways and sensitivities of GDP and ecological footprint 2002–2005. Ecol. Indic. 16, 2012, 128-134Mattila, T., Saikku, L., How much land does Finland import? Estimating upstream land use with different MRIO and LCA data sources. Ecol, Econ. (submitted to). Milà i Canals, L., Bauer, C., Depestele, J., et al., 2007. Key elements in a framework for land use impact assessment within LCA. Int J. Life Cycle Assess. 12, 5–15. Nielsen, P.H., Nielsen, A.M., Weidema, B.P., Dalgaard, R., Halberg, N., 2003. LCA food data base. www.lcafood.dk. Ntiamoah, A., Afrane, G., 2008. Environmental impacts of cocoa production and processing in Ghana: life cycle assessment approach. J. Clean. Prod. 16, 1735–1740. Saikku, L., Mattila, T., Akujärvi, A., Liski, J., 2015. Human appropriation of net primary production in Finland during 1990–2010. Biomass Bioenergy 83, 559–567. Sandström, V., Saikku, L., Antikainen, R., Sokka, L., Kauppi, P., 2014. Changing impact of import and export on agricultural land use: the case of Finland 1961–2007. Agric. Ecosyst. Environ. 188, 163–168. Soimakallio, S., Cowie, A., Brandão, M., Finnveden, G., Ekvall, T., et al., 2015. Attributional life cycle assessment: is a land-use baseline necessary? Int. J. Life Cycle Assess. 20 (10), 1364–1375. Statistics Finland, 2015. Official Statistics of Finland (OSF): Annual National Accounts [e-publication]. Statistics Finland, Helsinki (ISSN=1798-0623. Available at: http://www.stat.fi/til/vtp/tau en.html). Steinberger, J.K., Krausmann, F., Eisenmenger, N., 2010. Global patterns of materials use: a socioeconomic and geophysical analysis. Ecol. Econ. 69 (5), 1148–1158. Suh, S., Huppes, G., 2005. Methods for life cycle inventory of a product. J. Clean. Prod. 13 (7), 687–697. Szyrmer, J.M., 1992. Input—output coefficients and multipliers from a total-flow perspective. Environ. Plan. A 24, 921–937. Taelman, S.E., Schaubroeck, T., De Meestera, S., Boone, L., Dewulf, J., 2016. Accounting for land use in life cycle assessment: the value of NPP as a proxy indicator to assess land use impacts on ecosystems. Sci. Total Environ. 550 (15), 143–156 http://www.sciencedirect.com/science/article/pii/ S0048969716300572. Teixidó-Figueras, J., Steinberger, J.K., Krausmann, F., Haberl, H., Wiedmann, T., et al., 2016. International inequality of environmental pressures: decomposition and comparative analysis. Ecol. Indic. 62, 163–173 (www.sciencedirect.com/science/article/pii/S1470160x15006731). Waggoner, P.E., Ausubel, J.H., 2002. A framework for sustainability science A renovated IPAT identity. PNAS 99, 7860–7865. Williams, A.G., Audsley, E., Sandars, D.L., 2006. Determining the Environmental Burdens and Resource Use in the Production of Agricultural and Horticultural Commodities. Main Report. Defra Research Project IS0205. Cranfield University and Defra, Bedford (Available on www.silsoe.cranfield.ac.uk, and www.defra.gov.uk). Ytrestøyl, T., Aas, T.S., Berge, G.M., Hatlen, B., Sørensen, M., Ruyter, B., Thomassen, M., Hognes, E.S., Ziegler, F., Sund V., Åsgård, T., 2011. Resource utilisation and eco-efficiency of Norwegian salmon farming in 2010. Nofima report Tromsa.