**1. Introduction**

296 Remote Sensing of Biomass – Principles and Applications

Valdez R, M.; González&De Los Santos, H. M. (2006). Estimación de cobertura arbórea

Wackernagel, H. (2003). Multivariate *Geostatistics: an Introduction With Applications*, Third

Wackernagel, H.; Bertino, L.; Sierra, J.P. & González del Río, J. (2002). Multivariate Kriging

Wang, Ch. (2006). Biomass allometric equations for 10 co-occurring tree species in Chinese

Wulder, M. (1998). Optical remote-sensing techniques for the assessment of forest inventory and biophysical parameters. Progress in Physical Geography, 22, pp. 449–476. Zewdie, M.; Olsson, M. & Verwijst, Th. (2009). Above-ground biomass production and

Zheng, D.; Heath, L.S. & Ducey, M.J. (2007). Forest biomass estimated from MODIS and FIA data in the Lake States: MN, WI, and MI, USA. Forestry 2007, 80, 265-278. Zheng, D.; Rademacher, J.; Chen, J.; Crow, T.; Bresee, M.; Le Moine, J. & Ryu, S. (2004).

temperate forests. Forest Ecology and Management 222 (2006) 9–16

40(3): 383-394 p. ISSN 1405-3195.

Shaarawi, (Eds.), 57-75, Springer-Verlag, London

Edition, Springer, Berlin

(2009) 421–428

mediante imágenes Satelitales multiespectrales de alta resolución. Agrociencia

for Interpolating With Data From Different Sources, In: *Quantitative Methods for Current Environmental Issues*, C.W. Anderson, V. Barnett, P. Chatwin & A.H. El

allometric relations of Eucalyptus globulus Labill. coppice plantations along a chronosequence in the central highlands of Ethiopia. Biomass and Bioenergy 33

Estimating aboveground biomass using LANDSAT 7 ETM+ data across a managed landscape in northern Wisconsin, USA. *Remote Sensing of Environment*, 93, 402–411

Regarding the three objectives of European Union (EU) energy policy (secure supply, competitiveness and environmental protection) the EU Commission published the Communication entitled 'White Paper: Energy for the Future - Renewable Sources of Energy' (EU Commission, 1997; Mourelatou & Smith, 2004). This document, which was the starting point for the European promotion and development of renewable energy, stated the objective that 12% of energy production in 2010 would come from renewable sources. In Spain, this objective was recognized by the government in 1999 in the Plan to the Promotion of Renewable Energies (PPRE). To achieve this objective, the PPRE focused on increasing the use of biomass, identifying forest residues as one of the biomass sources. Specifically, this document established an increase in the use of forest residual biomass to 450 000 tonnes of petroleum equivalent per year (TPE/year) (IDAE, 1999).

Among the different sources of renewable energy, this chapter focuses on forest residual biomass (FRB). This term refers to branches, foliage, and unmerchantable stem tops that are commercially unsuitable in terms of the timber obtained from regular operations in forest management or in timber exploitation (Esteban et al., 2004; IDAE, 1999; Velázquez, 2006). Following the 'complete-tree concept,' the term 'branches' include the wood and bark of live and dead branches; 'foliage' refers to all leaves-needles, new shoots and reproductive organs; and the 'unmerchantable stem top' is the upper section of the stem that is left unutilized in logging operations due to its small diameter and high degree of branching (the bottom stem diameter of this unmerchantable top usually ranges from 5 to 10 cm) (Hakkila & Parikka, 2002). The treatments commonly applied to these residues in Spain include controlled burning, stacking within the forest and, less commonly, splintering to improve their incorporation into the soil (IDAE, 2005a, 2007; Velázquez, 2006).

However, this biomass segment can also be used as a source of energy in heating applications (fuel for domestic or industrial stoves and boilers) and in the generation of electricity (replacing fossil fuels in power stations) (Asikainen et al., 2002; IDAE, 2005b), with the majority of the residue currently being utilized for the latter in Spain (IDAE, 2007). The benefits associated with this energy-related use can be divided into two types: environmental and socio-economic. The first benefits are generated in the production phase, as the recovery and elimination of FRB reduces the risk of forest fires and their severity (IDAE, 2005a; Velázquez, 2006), and also because the implementation of silviculture can

Using Remote Sensing to Estimate a Renewable Resource: Forest Residual Biomass 299

Thus, remotely sensed data have not only facilitated an increase in the speed, cost efficiency, precision, and timeliness of inventories, but they have also allowed the construction of maps of forest attributes with spatial resolutions and accuracies that were not feasible even a few years ago (McRoberts & Tomppo, 2007). These advances have made remotely sensed data the primary source for biomass estimation (Lu, 2006). However, no studies have yet presented a technique for biomass estimation that is consistent, reproducible and entirely applicable at regional scales (Muukkonen and Heiskanen, 2005; Powell et al., 2010), especially in forests located in areas of irregular topography and characterized by heterogeneity in species composition, complex stand structures and environmental conditions (Brown et al., 1999; Lu & Batistiella, 2005; Lu, 2006; Mallinis et al., 2004;). Those characteristics are inherent features of Mediterranean forests (Shoshany, 2000), where not many studies have been carried out. The objective of this chapter is to explain a methodology developed to estimate the amount of FRB potentially suitable for renewable energy production in the pine forests of Mediterranean areas at regional scale, using satellite images and forest inventory data. It is intended, therefore, to eliminate a major barrier to the use of this renewable source of energy. In turn, by using a plain methodology, it is intended that the method developed can be adopted by decision makers and land managers for both forest management and regional planning, considering that energy planning is a major component of land management. For this study, we used FRB data obtained from allometric equations applied to the Second Spanish National Forest Inventory (NFI-2) (dependent variable) and spectral data from Landsat 5 TM imagery (independent variable). In order to avoid the effects of forest heterogeneity in the establishment of accurate predictive models, different methods were

The methodology applied in this study was developed in the framework of geographic information technologies (Remote Sensing and Geographical Information Systems, GIS) as information sources and tools for forest management. The software Erdas Imagine was used to process the Landsat images, ARC-GIS was used for the management of ancillary

The methodology proposed and developed in this study was divided into five phases or steps. Firstly, FRB data was obtained by means of field work and pre-existing forest data. Secondly, suitable satellite images were selected and different treatments were applied in order to convert the data to a suitable format for quantitative analysis and to guarantee the validity of the results. Thirdly, three different methods were used to relate field data and radiometric information in order to overcome the difficulties involved in estimating forest parameters in heterogeneous Mediterranean forests. The fourth phase focused on developing and running regression models to map FRB in the study area, and to evaluate and compare the results at pixel level obtained using the applied extraction methods. Finally, the best model was selected for application to a recent image to quantify the amount

The study area is the Province of Teruel, which is located in the northeastern Iberian Peninsula (Figure 1). This province was selected in the context of the LIGNOSTRUM project

tested to extract the spectral data from the Landsat images.

information, and SPSS was employed for statistical analyses.

**2. Materials and methods** 

of FRB at present time in the study area.

**2.1 Study area** 

improve the health of forests (Eriksson et al., 2002; Raison, 2002). Environmental benefits are also obtained in the application phase. Considering only the combustion process, the energy produced is almost neutral in terms of carbon dioxide, as the amount of CO2 emitted to the atmosphere due to FRB combustion is the same than the fixed in the FRB formation (net primary production). In addition, sulphur and chlorine emissions are very low. Consequently, taking into account only the combustion process, the generation of energy from FRB does not contribute to an increase in greenhouse gasses (Hakkila & Parikka, 2002; IDAE, 2005a). The social and economic benefits can be separated into two categories: (i) those that occur at the national scale, associated with a reduction in the use of nonrenewable fossil fuels (Domínguez, 2002; Eriksson et al., 2002; IDAE, 2005a), and (ii) those that occur at the local scale, as the utilization of a waste product leads to increasing harvests, transportation, and utilization of forest residues in power stations, which also leads to an increase in rural employment. These considerations are important for rural areas in which the level of unemployment and depopulation is a public policy issue, and where increased employment can help to support a population that can maintain the natural environment (Borsboom et al., 2002; Domínguez, 2002; IDAE, 2005a, 2007).

Despite the benefits outlined above, previous studies conducted before the subsequent revision of PPRE, the Plan to Renewable Energies 2005–2010 (PRE), showed that the use of FRB had a long way to go in achieving the anticipated objectives. In the PRE analysis, the lack of specific methodologies to assess the regional-scale quantity of FRB was identified as one of the main problems (IDAE, 2005a). This is a fundamental concern because power stations that use FRB require knowledge of the amount of resource available (Esteban et al., 2004; IDAE, 2005b, 2007; Velázquez, 2006). To overcome this problem, a methodology is needed that quantifies the potential of this forest resource in a given area. This is an important matter considering that the new Plan to Renewable Energies in Spain 2011-2020 (currently being written using the provisions of Directive 2009/28/EC) set a final energy consumption target of 20% derived from renewable energy, doubling the previous mark in the 2010 Plan (MITYC, 2010).

Several studies have demonstrated the effectiveness of satellite images in estimating forest variables, including biomass. These studies were carried out using both passive (optical) and active (radar and lidar) sensors. Focusing on optical sensors, different experiments have been conducted using multispectral or hyperspectral coarse spatial resolution data (i.e. Anaya et al., 2009; Muukkonen & Heiskanen, 2007) and fine spatial resolution data (i.e. Gonzalez et al., 2010; Proisy et al., 2007); although the most frequent experiments have been performed using medium spatial resolution data, mainly using Landsat Thematic Mapper (TM) or Enhanced Thematic Mapper Plus (ETM+) data (Lu, 2006; Powell et al., 2010). Radar data has also been used extensively, establishing significant correlations between biomass and radar backscatter at C-band (Kurvonen et al., 2002) and L-band (Austin et al., 2003; Lucas et al., 2010), with the latter being more consistent and robust. Recently, forest biomass assessment research has focused on using polarimetric synthetic aperture radar interferometry (PolInSAR) and laser scanning systems (lidar) data, mostly by means of airborne sensors. Koch (2010) offers a complete review of the recent advances and future developments in this issue.

Images derived from remote sensing register continuous and complete information across a landscape and such images can be obtained at frequent intervals. These characteristics help to overcome some of the problems associated with inventory methods exclusively based on field work, interpolation techniques and GIS (Franklin, 2001; Lu, 2006; Salvador & Pons, 1998a). Thus, remotely sensed data have not only facilitated an increase in the speed, cost efficiency, precision, and timeliness of inventories, but they have also allowed the construction of maps of forest attributes with spatial resolutions and accuracies that were not feasible even a few years ago (McRoberts & Tomppo, 2007). These advances have made remotely sensed data the primary source for biomass estimation (Lu, 2006). However, no studies have yet presented a technique for biomass estimation that is consistent, reproducible and entirely applicable at regional scales (Muukkonen and Heiskanen, 2005; Powell et al., 2010), especially in forests located in areas of irregular topography and characterized by heterogeneity in species composition, complex stand structures and environmental conditions (Brown et al., 1999; Lu & Batistiella, 2005; Lu, 2006; Mallinis et al., 2004;). Those characteristics are inherent features of Mediterranean forests (Shoshany, 2000), where not many studies have been carried out.

The objective of this chapter is to explain a methodology developed to estimate the amount of FRB potentially suitable for renewable energy production in the pine forests of Mediterranean areas at regional scale, using satellite images and forest inventory data. It is intended, therefore, to eliminate a major barrier to the use of this renewable source of energy. In turn, by using a plain methodology, it is intended that the method developed can be adopted by decision makers and land managers for both forest management and regional planning, considering that energy planning is a major component of land management.

For this study, we used FRB data obtained from allometric equations applied to the Second Spanish National Forest Inventory (NFI-2) (dependent variable) and spectral data from Landsat 5 TM imagery (independent variable). In order to avoid the effects of forest heterogeneity in the establishment of accurate predictive models, different methods were tested to extract the spectral data from the Landsat images.
