Towards local bioeconomy: A stepwise framework for high-resolution spatial quantification of forestry residues

https://doi.org/10.1016/j.rser.2020.110350Get rights and content

Highlights

  • CamBEE: A framework for high-resolution spatial quantification of forestry residues.

  • Uses open-source spatial data & presents results with uncertainties.

  • A metric for deciding the spatial resolution for such assessments is provided.

  • Exemplified results for France reveal 8.4 Million t DM y−1 of forestry residues.

Abstract

In the ambition of a transition from fossil carbon use, forestry residues are attracting considerable attention as a feedstock for the future bioeconomy. However, there is a limited spatially-explicit understanding of their availability. In the present study, this gap has been bridged by developing a generic framework “CamBEE”, for a transparent estimation of aboveground primary forestry residues. CamBEE further includes guidelines, based on standard uncertainty propagation techniques, to quantify the uncertainty of the generated estimates. CamBEE is a four-step procedure relying on open-access spatial data. The framework further provides insights on the appropriate spatial resolution to select. In this study, the proposed framework has been detailed and exemplified through a case study for France. In the case study, primary forestry residues have been spatially quantified at a resolution of 10 m, using spatial and statistical data on forest parameters (net annual increment, factor of basic wood density, biomass expansion factors, etc.). The results for the case study indicate a total theoretical potential of 8.4 Million Mgdry matter year−1 (4.4–13.9 Million Mgdry matter year−1) available in France, the equivalent of 161 PJ year−1. The case study validates that the CamBEE framework can be used for high-resolution spatial quantification of PFRs towards integration in local bioeconomy.

Introduction

Facing the emergency to limit the global mean temperature to 1.5 °C above pre-industrial levels [1], additional releases of fossil carbon must be prevented, while sinks must be induced to achieve so-called carbon neutrality [2,3]. By using biogenic carbon to supply materials, chemicals, food, and energy services, bioeconomy is one option to reach the balance between anthropogenic emissions and sinks. In particular, the use of biomass residues has generated growing interest as a feedstock not associated with land-use changes [[4], [5], [6]], Hamelin et al. [4] highlighted a total (or theoretical) residual biomass potential of 8500 PJ year−1 for EU-27 + Switzerland, with forestry residues representing ca. 37% of this potential.

Forestry residues can be divided into two categories [7]; (i) Primary forestry residues (PFRs), and (ii) Secondary forestry residues. PFRs are defined as residues that are left after logging operations (branches, stumps, treetops, bark, sawdust, etc.). In contrast to PFRs, secondary forestry residues are by-products and co-products of industrial wood-processing operations (bark, sawmill slabs, sawdust, wood chips, etc.). This study focuses solely on aboveground PFRs (Fig. 1). Accordingly, belowground stumps and roots are excluded, given the heavy concerns related to the environmental [8,9] and economic sustainability of harvesting these [10].

When it comes to spatial quantification of PFRs, the methods have generally improved over the last ten years, but due to a lack of standardization, there is still a significant concern in the user community on the available methods and practices [11,12]. Spatial quantification of PFRs is here defined as quantifying the amount of PFR generated per unit area along with positional attributes (Supplementary Table 1). Besides the lack of standardization on methods and definitions, the availability of spatially-explicit forest data on forest classes, species, annual increment, expansion factors, etc. is a major limiting factor for such assessments. This is especially true for regions with diverse physiography [13]. One well-known forest inventory data is FAO's “Forest Resource Assessments” (FRAs) [14], which provides spatial data on forest statistics. Although valuable, the spatial data provided by FRAs are at best at the country scale, hence unsuitable for high-resolution mapping for use in developing local deployment strategies. The data countries supply to the FAO for the FRA typically stem from National Forest Inventories (NFIs), which provide comprehensive data on forestry [15] at sub-regional, regional, or country scales. Despite major harmonization efforts, the aggregation of these national data in FRA reports still present severe discrepancies regarding nomenclature, baseline variables, etc. [[16], [17], [18]]. For example, the Swiss NFI only measures trees with the diameter at breast height (d.b.h.) ≥ 12 cm for quantifying growing stock [19]. In contrast, this threshold is 0 cm (hence including the complete stem top) in the reference definition of Cost Action E43 [20], a landmark in Europe for harmonized NFI definitions [16].

Furthermore, one vital consideration for spatial quantification is the scale and spatial resolution that is being used. For geographic information system (GIS) applications, it is a common practice to merge datasets through the use of different functions, such as map overlay. But if datasets do not have harmonized spatial resolutions, the result of this operation becomes erroneous. In simple words, spatial scale refers to the size of a land area or geographic distance studied. However, there are various categories of scale, such as geographic, operational, measurement, and cartographic, and it is essential that a clear distinction is made amongst them. The definition of these terminologies is provided in Supplementary Table 1. Most studies on spatial quantification do not consider spatial scale as an important parameter; instead, in general, the studies directly use certain spatial data related to the available methods without questioning the uncertainties associated with them. For example, the use of local forest inventory statistics on a global or continental assessment would lead to a large amount of uncertainty in the analysis due to inconsistencies in geometry, attributes, and semantics owing to scale differences [21]. In fact, when it comes to high-resolution mapping of residual biomass resources at country scales, there is no real discussion on what resolution can be considered as useful for bioeconomic purposes.

Globally, very few studies have attempted to spatially quantify the potential of residual biomass from forestry and none of them report the uncertainty associated with the results. A brief critical summary of some of these studies is presented in Table 1.

As highlighted in Table 1, researchers have adopted several methods for estimating PFRs. The diversity of these methods is an indicator of the lack of standardization for such assessments. Furthermore, the approaches listed chronologically in Table 1 also emphasize how technological improvements in remote sensing have changed the data requirements over the years, with recent methods [12,25,31] relying on actual remote sensing measurements than purely being based on certain assumptions and models [22,23]. Despite this, almost all of the methods listed in Table 1 require field data from forest inventory or statistics for either initializing models or for validating the results. Field data are essential for robust estimation of PFRs, as the quantity of PFRs is primarily dependent on the harvesting practices, which varies spatially due to several factors such as forest type, harvesting policies, machinery, management practices, etc. Out of all the methods presented in Table 1, only those of [12,28] seem promising in the context of large geographic-scale high-resolution mapping. The other methods are either too conceptual, relying primarily on ancillary data from the literature [23], or would be unfeasible to implement for a country scale assessment due to expensive data and processing time [31].

The method developed in Refs. [12] is an approach to spatially estimate the quantity of logging residues by using remote sensing products. In their study, a set of multi-year harvest maps and a set of forest attribute maps along with inventory data from NFI were used, and finally, this data was aggregated per hectare of forest area. As reported by the researchers [12], the advantage of using this approach is to have an estimate on the future availability of the residues at a country scale. However, there are several constraints for the application of this method in different regions of the world such as, availability of long-term (decadal) harvest maps, availability of specialized products like “CanLaD” (a comprehensive dedicated regional database), etc., which limits the reproducibility of their approach in other regions. The shortcomings of this approach can be overcome by using the method provided by Vis and Van den Berg [30], also referred to as the BEE approach in this study. The BEE approach is the most transparent among all methods listed in Table 1, and amongst the only ones to describe the intermediate steps it uses, making it replicable and adaptable. Another advantage of using this technique is the use of both forest cover map and forest statistics in the intermediate steps of calculation. This eliminates the need for other assumptions, such as regression parameters [22] or allometric equations [24]. Consequently, the use of this method will likely better represent the actual ground condition with a lower degree of variability (depending on the quality of the input statistical data) compared to the other techniques listed in Table 1. Even with a constraint like limited availability of inventory statistics, this method can be used to estimate the quantity of PFRs based entirely on remote sensing data products (Forest cover map from CORINE Land Cover and Net Primary Productivity data from MODIS) with a certain degree of accuracy. Additionally, due to the discrete and flexible nature of the BEE approach, uncertainty accounting can easily be introduced in the intermediate steps of calculation. This would allow a significant improvement of the current method by generating final quantification results with a confidence interval. Ergo, the present study fully acknowledges and builds upon the efforts and knowledge developed within the BEE method.

From a theoretical viewpoint, spatial estimations are representations, and all representations of reality have their complexities that cannot be measured directly, which may be termed as potential error or uncertainty. For spatial assessments, uncertainty quantification not only gives a quantified idea about the confidence in calculated results (as ranges); it may also help to understand the spatial pattern of deviation, and ultimately lead to better decision making for the mobilization of the assessed resources. However, there is a severe dearth of studies incorporating uncertainty analysis in spatial quantification of forest residues; in fact, no literature reporting uncertainty in spatial quantification of forest residues has been found.

In an endeavor to bridge these gaps, this study aims to develop a replicable method for high-resolution spatial quantification of aboveground PFRs at the pixel level along with uncertainty quantification and to illustrate how to use it for a concrete case study. The study at hand also proposes a scheme for selecting the optimal resolution for the spatial quantification of PFRs based on several performance criteria. The methodology produced in this study intends to equip bioeconomy policymakers with reliable spatially-explicit estimates of this key stream for the future bioeconomy. Besides the investment decision itself, these high-resolution estimates could support decisions such as the siting of a bio-refinery unit.

The paper is structured as follows: Section 2 presents the scope of the study and defines the key terms used. Section 3 details the generic stepwise methodological framework proposed for high-resolution spatial quantification of PFRs. Section 4 deals with the exemplification of the method. In section 5, the results of the study have been critically discussed, focusing mainly on the methodological framework and uncertainty associated with different data sources. The discussion section also includes a comparison of the results with other studies and how this work can be improved further. Based on these, the study concludes how the proposed methodology with uncertainty quantification is a definite step forward towards providing a transparent, replicable, and harmonized methodology for spatial quantification of PFRs.

Section snippets

Scoping

An essential consideration in biomass resource assessment is the type of biomass potential being estimated. The BEE method [30], as well as several other studies, e.g., by Hamelin et al. [4], and Greggio et al. [29], distinguish five types of biomass potentials, i.e., theoretical potential, technical potential, economic potential, implementation potential, and sustainable potential. The type of biomass potential being assessed largely determines the approach, methodology, and data requirement [

Stepwise framework for quantification of primary forestry residues

Table 1 highlighted that efforts have already been undertaken to develop methodologies to spatially estimate the PFRs. The approach used here builds upon the existing efforts, and in particular, upon the landmark BEE handbook technique presented in Vis and Van den Berg [30], also used as an underlying framework for many of the studies carried out post-2010 (Table 1).

However, unlike [30], an in-depth focus on the specific PFR stream and high spatial resolution is made, data uncertainty and its

Study area

To exemplify the proposed technique, a case of Metropolitan France was presented (Mainland European France, including Corsica), herein referred to as France. France is administratively divided in the following hierarchy: Regions (13) > Departments (96) > Communes (36,569), though other administrative or political areal units also exist. In terms of the latest European Nomenclature of Territorial Units (NUTS), Regions correspond to the NUTS level 1 level while Departments correspond to NUTS

Discussion

The work presented in this study bridges the gap between spatial sciences, forestry, and bioeconomy by providing an easily replicable spatially-explicit framework “CamBEE” for estimating PFRs at high-resolution. By being at the intersection of geographical and “bioeconomic” science, it is also the first attempt to discuss questions such as the appropriate resolution for these types of assessments. The CamBEE framework can work across wide ranges of geographical scales and spatial resolution

Conclusions

The essential contributions of the present study are summarized as:

  • A generic, transparent and stepwise methodology for spatial quantification of PFRs at the pixel level incorporating associated uncertainties was developed. The applicability of the proposed methodology was tested and demonstrated for the case of Metropolitan France. The results indicated not only that the framework is easy to use, but also that about 8.4 Million Mgdry matter year−1 of PFRs, the equivalent of 161 PJ year−1, are

CRediT authorship contribution statement

S.K. Karan: Conceptualization, Data curation, Formal analysis, Visualization, Investigation, Methodology, Software, Writing - original draft. L. Hamelin: Conceptualization, Funding acquisition, Investigation, Resources, Supervision, Writing - review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This work was carried out within the framework of the research project Cambioscop (https://cambioscop.cnrs.fr), partly financed by the French National Research Agency, Programme Investissement d’Avenir (ANR-17-MGPA-0006) and Region Occitanie (18015981). The authors gratefully acknowledge Dr. Antoine Collin of IGN for insightful discussions, advices, and the access to BD Forêt version 2 data and to Dr. Laurent Polidori and Dr. Jordi Inglada of CESBIO for sharing the land cover data of

References (81)

  • B.P. Weidema et al.

    Data quality management for life cycle inventories—an example of using data quality indicators

    J Clean Prod

    (1996)
  • N. Martinkus et al.

    A comparison of methodologies for estimating delivered forest residue volume and cost to a wood-based biorefinery

    Biomass Bioenergy

    (2017)
  • K. Blennow et al.

    Forest owner motivations and attitudes towards supplying biomass for energy in Europe

    Biomass Bioenergy

    (2014)
  • P. Stjepan et al.

    Private forest owners' willingness to supply woody biomass in selected South-Eastern European countries

    Biomass Bioenergy

    (2015)
  • Summary for policymakers

  • C. Le Quéré et al.

    Global carbon budget 2014

    Earth Syst Sci Data

    (2015)
  • UN

    (2015)
  • D. Tonini et al.

    Environmental implications of the use of agro-industrial residues for biorefineries: application of a deterministic model for indirect land-use changes

    GCB Bioenergy

    (2016)
  • P. Ruiz et al.

    The JRC-EU-TIMES model. Bioenergy potentials for EU and neighbouring countries

    (2015)
  • J.D. Walmsley et al.

    Stump harvesting for bioenergy - a review of the environmental impacts

    Forestry

    (2010)
  • T. Persson et al.

    Stump harvesting for bioenergy: a review of climatic and environmental impacts in northern Europe and America

    WIREs Energy Environ

    (2018)
  • A. Thivolle-Cazat et al.

    La mobilisation de la ressource forestière aujourd’hui et demain

    (2014)
  • J. Barrette et al.

    Forecasting the spatial distribution of logging residues across the Canadian managed forest

    Can J Res

    (2018)
  • J. Návar

    Measurement and assessment methods of forest aboveground biomass: a literature review and the challenges ahead

  • FAO

    Global Forest Resource Assessment 2020: Main Report

    (2020)
  • T. Gschwantner et al.

    Harmonisation of stem volume estimates in European national forest inventories

    Ann For Sci

    (2019)
  • E. Tomppo et al.

    National forest inventories

    (2010)
  • C. Vidal et al.

    The role of European National Forest Inventories for international forestry reporting

    Ann For Sci

    (2016)
  • T. Gschwantner et al.

    Comparison of methods used in European National Forest Inventories for the estimation of volume increment: towards harmonisation

    Ann For Sci

    (2016)
  • COST action E43 harmonisation of national inventories in Europe : techniques for common reporting

    (2008)
  • J. Zhang et al.

    Scale in spatial information and analysis

    (2014)
  • L.A. Wells et al.

    Spatial and temporal quantification of forest residue volumes and delivered costs

    Can J Res

    (2016)
  • L.A. Wells

    Spatial distribution and quantification of forest treatment residues for bioenergy production

    (2013)
  • L. Guindon et al.

    Annual mapping of large forest disturbances across Canada's forests using 250 m MODIS imagery from 2000 to 2011

    Can J Res

    (2014)
  • Guindon L, Villemaire P, St-Amant R, Bernier PY, Beaudoin A, Caron F, et al. Canada Landsat Disturbance (CanLaD): a...
  • M.W. Vis et al.

    Harmonization of biomass resource assessments. Volume I

    Best Practices and Methods Handbook

    (2010)
  • D. Domingo et al.

    Quantifying forest residual biomass in Pinus halepensis Miller stands using Airborne Laser Scanning data

    GIScience Remote Sens

    (2019)
  • European Commission

    A new EU forest strategy: for forests and the forest-based sector

    (2013)
  • Cited by (13)

    • Towards sustainable methane supply from local bioresources: Anaerobic digestion, gasification, and gas upgrading

      2022, Applied Energy
      Citation Excerpt :

      In addition to reducing or preventing fossil gas imports, renewable gas supports valorisation of local bioresources, in particular low-value residual bioresources, while increasing local energy production. To promote such a local bioeconomy for bio-based CH4 production, analysis of local bioresource availability is needed with respect to quantity, accessibility and quality (chemical, physical, and nutritional properties) [7–9]. As different bioresources are associated with different physico-chemical properties, this has profound importance both for the selection of conversion technology and for the environmental consequences associated with use of these bioresources for energy purposes.

    • Remote sensing-based biomass estimation of dry deciduous tropical forest using machine learning and ensemble analysis

      2022, Journal of Environmental Management
      Citation Excerpt :

      Upscaling these field measurements becomes unfeasible due to several factors such as high implementation costs, heterogeneity in forest species and sampling bias (Hill et al., 2013; Houghton et al., 2010; Vorster et al., 2020). In recent years, several studies have used remote sensing data to provide empirical evidence to forest dynamics at fine spatio-temporal scales (Karan and Hamelin, 2020; Singh et al., 2020; Valderrama-Landeros et al., 2021). These scientific advances illustrate that remote sensing is a viable alternative for ecosystem monitoring, alleviating some of the scaling and implementation limitations of field measurements.

    • Waste-to-nutrition: a review of current and emerging conversion pathways

      2021, Biotechnology Advances
      Citation Excerpt :

      The scope of categories is an expansion and harmonization of the streams described in Hamelin et al. (2019), also described in the SI and briefly reported here. The category “Wood-related residual biomass” shown in Fig. 1 includes primary forestry residues (defined in Karan and Hamelin, 2020), pruning residues, wood-processing wastes (e.g. sawdust) and some packaging waste (e.g. cardboard). The “Primary crop residues” category only includes straws, stalks and corn stover; tuber’s top fractions (e.g. potato leaves) are included within the “Green residual biomass” category along with garden- and park wastes (e.g. mowed grass).

    • Framework for consequential life cycle assessment of pyrolysis biorefineries: A case study for the conversion of primary forestry residues

      2021, Renewable and Sustainable Energy Reviews
      Citation Excerpt :

      In France alone, it is estimated that at least 2.7 Million Mg DM year−1 (~42 PJ) of logging residues (branches, treetops, leaves, needles and bark) could be harvested without reducing soil fertility [2]. Moreover, recent work indicates a theoretical potential (i.e. without considering environmental or economic constraints) of 8.4 Million Mg DM year−1 (4.4–13.9 Million Mg DM year−1) in the country [3]. Currently, most of logging residues are left on soil.

    View all citing articles on Scopus
    View full text