Remote Sensing for Terrestrial Biogeochemical Modeling

Keywords

carbon cycling, ecosystem modeling, evapotranspiration, hyperspectral imaging, model data assimilation, net primary production, vegetation properties.

Introduction

Remote sensing and biogeochemical modeling share a highly complementary nature, whichhasled to a growing number of applications that involve some degreeof coupling between the two. Whereas remote sensing represents the only means by which landscape and vegetation properties can be sampled over large and contiguous portions of the Earth's surface, models focus on the underlying biogeochemical processes that regulate the flow and storage of water, carbon and nutrients, often over much longer time scales than can be considered through remote sensing alone. The aims and goals of biogeochemical modeling are wide ranging and include studies of carbon (C) and water cycling, analyses of nitrogen (N) enrichment and leaching to aquatic ecosystems, and trace gas transfers from soils to the atmosphere. Biogeochemical models are also used as part of larger, integrated modeling environments of regional and global biosphere–atmosphere interactions, biogeography, and climate change (e.g., Sellerset al. 1997, Cramer et al. 2001). More recently, biogeochemical modeling has taken a role in decision support for conservation, management, and policy development (Potter et al. 2006).

Remote sensing can provide biogeochemical models with information on vegetation type, leaf area index (LAI), canopy height, the fraction of absorbed photosynthetically active radiation (fPAR), light-use efficiency (LUE), leaf N concentration, pigments and other biochemical compounds to simulate plant growth and mortality. Other remote sensing-related inputs have included temperature, precipitation, solar radiation levels, and soil moisture. In this chapter, we review a number of approaches through which remote sensing data can be applied to the detection of vegetation properties, and discuss the trade-offs of various methods with respect to biogeochemical modeling. Given the breadth of the topic, our goal in preparing this chapter was not to provide a working manual of all remote sensing-model integration methods available. Instead, we sought to summarize important overall strategies for vegetation detection into a framework that involves the types of instruments used, the ecological properties they can be designed to detect, and the manner in which those properties can be utilized by models.

Models Utilizing Remote Sensing

This section provides a brief overview of major categories of ecosystem biogeochemistry models that can be driven or guided by remote sensing. In the interest of simplicity, we describe each type of model as being distinct from one [Page 412]another, but readers should be aware that the boundaries between them are often blurry, and hybrid approaches are also available.