Geostatistical Estimation of Biomass Stock in Chilean Native Forests and Plantations

Abstract

There are a variety of approaches to estimate above ground biomass (AGB), which can be classified according to the data source being used (Koch and Dees, 2008): field measurement, remotely sensed data or ancillary data used in GIS-based modeling. Field measurements are based on destructive sampling or direct measurement and the application of allometric equations (Madgwick, 1994). Recently, remotely sensed data, from both passive and active sensors, have become an important data source for AGB estimation. In this chapter we will focus on the use of optical multispectral data such as TM/ETM+ to estimate AGB. Generally, biomass is either estimated via a direct relationship between spectral response and biomass using multiple regression, k-nearest neighbor, neural networks, inverse canopy models or through indirect relationships, whereby attributes estimated from the remotely sensed data, such as leaf area index (LAI), structure (crown closure and height) or shadow fraction are used in equations to estimate biomass (Wulder, 1998). Here, we discuss the use of remote sensing data of moderate spatial resolution as input to estimate AGB. Research has demonstrated that it is more effective to generate relationships between field measurements and moderate spatial resolution remotely sensed data (e.g., LANDSAT), and then extrapolate these relationships over larger areas using comparable spectral properties from coarser spatial resolution imagery (e.g., MODIS) (Steininger, 2000; Lu, 2005; Phua and Saito, 2003; Foddy el al., 2003; Fazakas et al., 1999; Roy and Ravan, 1996; Zheng et al., 2004; Mickler et al., 2002). In general terms, LANDSAT TM and ETM+ data are the most widely used data of remotely sensed imagery for forest biomass estimation, but data from other moderate spatial resolution sensors have also been used, including ASTER and HYPERION data. In this chapter we present approaches that are currently being developed in Chile. Specifically, we introduce methods for the estimation of AGB using medium spatial resolution satellite imagery and digital elevation models. The main objective is to create, calibrate and validate such methods for applications. We developed an alternative approach in the estimation of AGB using LANDSAT ETM + imagery and SRTM digital elevation models as covariates for geostatistical modeling. From the spatial perspective, AGB data correspond to an array of points in space (x, y), while covariates correspond to a set of data that has a large number of samples in geographic space (extracted from each pixel), some of which having overlap with the available AGB